i r ; I i ■■ 4 • i ■ ■.r? ^E«^^^*^B X m • .,. . . - . i ; i . d^^'^T'X o o o MBL/WHOl Library mfs 3 o ^L == n 5 X -i 00 LD s TOE( ^^^^ i_j CSL Th ENCYCLOPEDIA .f MICROSCOPY Edited by |— GEORGE L CLARK Research Professor of Anahjtical Chemistry/, Emeritus, Universihj of Illinois, Urbana, Illinois REINHOLD PUBLISHING CORPORATION, NEW YORK CHAPMAN & HALL, LTD., LONDON Copyright © 1961 by REINHOLD PUBLISHING CORPORATION All rights reserved Library of Congress Catalog Card Number 61-9698 Printed in the United States of America by The Waverly Press, Inc., Baltimore, Md. CONTRIBUTING AUTHORS Kenneth W. Andrews The United Steel Companies Ltd., England B Albert V. Baez Harvey Mudd College Thomas F. Bates Pennsylvania State University John H. Bender Los Alamos Scientific Laboratory James R. Benford Baiisch & Lomb, Inc. C. G. Bergeron University of Illinois A. Bergstrand Sabhatsbergs Hospital, Sweden F. P. BOHATIRCHUK University of Ottawa, Canada W. BOLLMANN Battelle Memorial Institute, Switzerland Willi AM. A. Bonner Southwestern Medical School of the Uni- versity of Texas R. BOEASKY University of Illinois D. E. Bradley Associated Electrical Industries, England Floyd Dunn University of Illinois J. Dyson Associated Electrical Industries, England N. A. Dyson University of Birmingham, England Wesley B. Estill Sandia Corporation M. L. Feeney University of California Medical Center F. Gordon Foster Bell Telephone Laboratories James A. Freeman U. S. Public Health Service William J. Fry University of Illinois Charles C. Fulton Department of Health, Education and Wel- fare L. K. Garron University of California Medical Center R. L. Gregory University of Cambridge, England Richard D, Cadle Stanford Research Institute George L. Clark University of Illinois Germain Crossmon Bausch & Lomb, Inc. D Norman L. Dockum General Electric Company H M. E. Haine Associated Electrical Industries, England Olle Hallen University of Goteborg, Sweden F. A. Hamm Minnesota Mining and Manufacturing Company Burton L. Henke Pomona College lU Contributing Authors L. L. Hundley Soiithwestern Medical School of the Uni- versity of Texas I Shixya Inoue W Dartmouth Medical School J. ISINGS Central Laboratory T.N.O., Holland Kazuo Ito Ja'pan Electron Optics Laboratory Co.. Ltd., Japan William Johnson The United Steel Companies Ltd., England Raymond Jonnard The Prudential Insurance Company of America B. E. Juniper Oxford University, England L. Marton National Bureau of Standards Ludwig J. Mayer General Mills, Inc. C. W. Melton Battelle Memorial Institute P. O'B. Montgomery Southwestern Medical School of the Uni- versity of Texas Erwin Muller Pennsylvania State University Jean M. Mutchler Linde Company N Joseph D. Nicol Michigan State University W. C. Nixon University of Cambridge, England J. Nutting University of Cambridge, England K Ernest H. Kalmus University of California P. M. Kelly University of Cambridge, England Charles J. Koester American Optical Company MoToi Kumai University of Chicago William R. Lasko United Aircraft Corporation Donald E. Lasko wski Armour Research Foundation of Illinois Institute of Technology K. Little Nuffield Department of Orthopedic Surgery, England M W. K. McEwEN University of California Medical Center Ronald H. Ottewill University of Cambridge, England D. H. Page British Paper and Board Industry Re- search Association, England H. H. Pattee Stanford University Ong Sing Poen Pomona College R Johannes A. G. Rhodin Bellevue Medical Center T. G. Rochow American Cyanamid Company George L. Royer American Cyanamid Company J. Salmon Laboratoire de Biologic Vegetate I, France IV Contributing Authors R. L. deC. H. Saunders Dalhousie University, Canada C. M. Schwartz Battelle Memorial Institute ' D. Scott National Engineering Laboratory, Scotland Michael Seal University of Cambridge, England K. C. A. Smith Pulp and Paper Research Institute of Canada, Canada John D. Steely Los Alamos Scientific Laboratory J. H. Talbot Transvaal and Orange Free State Chamber of Commerce, Africa V. J. Tenner Y Motorola Corporation w Ernest E. Wahlstrom University of Colorado Masaru Watanabe Japan Electron Optics Laboratory Co., Ltd., Japan Erwin K. Weise University of Illinois Alvar p. Wilska University of Arizona R. E. Wright Shell Chemical Company PREFACE With genuine pleasure and pride this "Encyclopedia of Microscopy" is presented as the fourth in a series of Reinhold contemporary^ integrated compilations of rapidly developing areas of science, following the "Encyclopedia of Chemistry" (1957), the "Encyclopedia of Chemistry Supplement" (1958), and the "Enc^'^clopedia of Spectroscopy" (December, 1960). Actually, "Spectroscopy" and "^Microscopy" were planned, projected, assembled and edited concurrently as twin volumes, but of necessity there was an interval of a few months in publication of the two, or a slightly delaj'ed birth, as it were, of the second twin. These two great instrumental techniques, valuable in so many disciplines, are so inti- mately related and interwoven that the simultaneous development of encyclopedias was the most logical procedure. Even though microscopy as a science is about three centuries old and spectroscopy only one, the common background and origins are well exemplified in the respective historical articles by Professor E. K. Weise. It may not be surprising, then, that one Preface was written originally for both Encyclopedias. This has appeared already in "Spectroscopy," and it is hoped that users of this volume will have the opportunity to read this more extended introduction to the pair of volumes, as well as to browse in a kindred science. The "Encyclopedia of Microscopy" is the fruit of the joint efforts of a trulj' international team of dedicated microscopists — English, Scotch, Canadian, South African, French, Ger- man, Swiss, Dutch, Swedish, Japanese and American — and it is this fact that gives such unique flavor, value and good will to the able and devoted coverage of a science which is as wide and boundless as the world itself. This Encyclopedia, of course, is a mosaic of 26 kinds of microscopj^ alphabetically arranged, and in most cases wdth numerous alphabetical subtopics under each. The numerous illus- trations in this picture book — diagrams of all kinds, photographs of microscopes and related instruments, and the micrographs made with them — speak eloquently' for themselves as superb art as well as science, heretofore unpublished in most cases. Carefully chosen lists of general and cross references, it is hoped, will add greatly to the value and usefulness of this volume to inquisitive students and laymen seeking information and illumination in this area which extends the powers of men's vision a millionfold or more, and to the experts who continue to build an ever-new science. The Editor is indebted personally to all the eminent scientists throughout the world who have contributed vitally important advice and encouragement, as well as one or more arti- cles based on experience and devotion in a specialized field. The entire list of authors pre- sented in the front pages is indeed a Roll of Honor. Special mention is due to colleagues at the University of Illinois — Professors E. K. Weise of the Department of Mining and Metal- lurgical Engineering, and R. Borasky, Director of the Electron Microscope Laboratory; and to the two loyal and able secretaries serving consecutively, Mrs. Ruth Tuite (1958-9) and Mrs. Claretta ]\Ietzger (1959-60), with whose help the entire task of planning, organiz- ing and editing has somehow been accomplished. The patience, enduring faith, guidance and technical aid by the publishing staff, especi- ally G. G. Hawley, Executive Editor, and Alberta Gordon, were indispensable factors in bringing both Encyclopedias to material fruition, and in looking ahead to new worlds of vu Preface science to bring between the covers of future members of this series of encyclopedias. The deep personal challenge and satisfaction to the Editor upon becoming a Professor Emeritus may somehow be reflected in the last paragraph of the Preface to the "Encyclopedia of Spectroscopy." George L. Clark Urbana, Illinois January, 1961 Vlll CONTENTS AUTORADIOGRAPHY 1 Autoradiography of Tissue, Norman L. Dockum 1 Shadow Autoradiography, E. Borasky 11 CHEMICAL MICROSCOPY 13 Alkaloids and Alkaloidal-Type Precipitatiox, Charles C. Fulton 13 Chemical Microcrystal Identifications, Charles C. Fulton 20 Forms of Microcrystals, Charles C. Fulton 37 History, Charles C. Fulton 38 Mixed Fusion Analysis, Donald E. Laskowski 42 Nitrogen-Bonded Radicals: Identification, Charles C. Fulton 46 Observing Microcrystals, Charles C. Fulton 51 Opium, Origin of, Charles C. Fulton 55 Purpose, Charles C. Fulton 56 Quinoline as a Reagent, /. M. Mutchler 57 Reagents for Microcrystal Identifications, Charles C. Fulton 58 Sympathomimetics and Central Stimulants, Charles C. Fulton 65 ELECTRON MICROSCOPY 72 Aerosols Containing Radioactive Particles, R. Borasky 72 Blood, James A . Freeman 77 Botanical Applications, D. E. Bradley 80 Cell Ultrastructure in Mammals, Johannes A. G. Rhodin 91 Ciliated Epithelia Ultrastructure, Johannes A. G. Rhodin 116 Colloids, Lyophobic, R. H. Ottewill 123 Crystal Lattice Resolution, G. L. Clark 145 Dislocations in Metals. See Transmission Electron Microscopy of Metals — Dislocations and Precipitation, p. 291 Electron Optics: Electron Gun and Electroalignetic and Electrostatic Lenses, M. E. Haine 147 Fibers (Textiles). See General Microscopy, p. 343 History of Electron Optics, L. Marlon - 155 Im.\ge Form,\tion Mechanism, L. Marlon 159 Ividney Ultrastructure, Johannes A. G. Rhodin 163 Leaf Surfaces, B. E. Juniper 177 Metals by Transmission, P. M. Kelly and J. Nutting 181 Microtomy. See General Microscopy, p. 385 Minerals, Thomas F. Bates 187 Paint Surface Replica Techniques, W. R. Lasko 200 ix Contents Pathology: Kidney, Anders Bergstrand 206 Plastics. See General Microscopy, p. 390 Pulp and Paper. See General Microscopy, p. 394 Reflection I, D. H. Page 220 Reflection II, Michael Seal 223 Replica and Shadowing Techniques, D. E. Bradley 229 Replicas, Removal from Surfaces, V. J. Tennery, C. G. Bergeron, and R. Borasky . 239 Resinography. See p. 525 Scanning, K. C. A. Smith 241 Selected Diffraction, J. H. Talbot 251 Snow Crystal Nuclei, Motoi Kumai 254 Special Methods, Masaru Watanahe and Kazuo I to 259 Specimen Preparation — Special Techniques at Los Alamos 270 A Modified Aluminum Pressing Replica Technique, J. H. Bender and E, H. Kalmiis 270 Preparation of Aerosols, E. H. Kalmus 272 Dispersion of Aerosols, J. H. Bender and J. D. Steely 272 Sorting and Fractography of Particles, J. H. Bender and E. H. Kalmus 273 Uncurling Carbon Replicas, W. B. Estill and E. H. Kalmus 274 Staining, Electron, R. Borasky 274 Tissues (Connective), Bones and Teeth, K. Little 276 Transmission Electron Microscopy of Metals — Dislocations and Precip- itation, W. Bollmann 291 Unsolved Problems, Franklin A. Hamm 307 Wear and Lubrication, D. Scott 308 Wilska Low- Voltage Microscopy, Alvar P. Wilska 314 ELECTRON MIRROR MICROSCOPY, Ludwig Mayer 316 FIELD EMISSION MICROSCOPY, Erwin W. Mailer 325 FLUORESCENCE MICROSCOPY, G. L. Clark 332 FLYING SPOT MICROSCOPY, P. O'B. Montgomery, Wm. A. Bonner, and L. L. Hundley 334 FORENSIC MICROSCOPY, Joseph D. Nicol 338 GENERAL MICROSCOPY 343 Fibers (Textile), J. I sings 343 Industrial Research, C. W. Melton and C. M. Schwartz 363 MiCROScopiSTS AND RESEARCH MANAGEMENT, Gcorge L. Royer 381 Microtomy, Olle Hallen 385 Plastics, J. I sings 390 Pulp and Paper, /. I sings 394 HISTORADIOGRAPH. See X-Ray Microscopy Contents INDUSTRIAL HYGIENE MICROSCOPY, Germain Grossman 400 INFRARED MICROSCOPY, G. L. Clark 411 INTERFERENCE MICROSCOPY 412 Fibers (Textile). See General Microscopy, p. 343 Industrial Research, Application to. See General Microscopy, p. 363 Instrument Classification and Applications, J. Dyson 412 Plastics. See General Microscopy, p. 390 Pulp and Paper. See General Microscopy, p. 394 Theory and Techniques, Charles J. Koester 420 LIGHT (OPTICAL) MICROSCOPY 434 Comparison Microscopes. See Engineering Microscopes, p. 438 Design and Construction of the Light Microscope, James R. Benford 434 Engineering Microscopes, G. L. Clark 437 Fibers (Textile). See General Microscopy, p. 343 Hardness Tests. See Engineering Microscopes, p. 438 Heating Microscopes. See Engineering Microscopes, p. 438 Industrial Research, Application to. See General Microscopy, p. 363 Introscope. See Engineering Microscopes, p. 438 Magnetography: The Microscopy of Magnetism, F. Gordon Foster 440 Measuring Microscopes. See Engineering Microscopes, p. 439 Optical Theory of the Light Microscope, James R. Benford 445 Origin and History, E. K. Weise 454 Particle Size and Shape Measurements and Statistics, Richard D. Cadlc 464 Plastics. See General Microscopy, p. 343 Projection Microscopes. See Engineering Microscopes, p. 438 Pulp and Paper. See General Microscopy, p. 343 Schmaltz Profile Microscope. See Engineering Microscopes, p. 438 Stereoscopic Microscope. See Engineering Microscopes, p. 439 METALLOGRAPHY 468 Industrial Research, Applications To. See General Microscopy, p. 343 Transmission Electron Microscopy of Metals — Dislocations and Precipi- tations. See p. 291 Wear and Lubrication. See Electron Microscopy, p. 308 MICROMETRON automatic microscope, G. L. Clark 468 MICRORADIOGRAPHY. See X-Ray Microscopy, p. 561 OPTICAL MINERALOGY, Ernest E. Wahlstrom 470 Petrographic Thin Sections, R. E. Wright 473 PHASE MICROSCOPY 476 Anoptral Microscope, A. Wilska 476 Fibers. See General Microscopy, p. 343 xi Contents Industrial Research, Applications to. See General Microscopy, p. 363 Plastics. See General Microscopy, p. 390 Pulp and Paper. See General Microscopy, p. 394 Theory and Microscope Construction. See Optical Theory of Light Mi- croscope, p. 445 POLARIZING MICROSCOPE 480 Basic Design and Operation. See Optical Theory of Light Microscope, p. 445 Design for Maximum Sensitivity, Shinya Inoue 480 Fibers (Textiles). See General Microscopy, p. 343 Industrial Research, Application to. See General Microscopy, p. 363 Plastics, See General Microscopy, p. 390 Pulp and Paper. See General Microscopy, p. 394 REFRACTION OF LIGHT, REFRACTOMETRY AND INTERFEROMETRY 485 Angle Refractometry, R. Jonnard 485 History of Light Refraction, R. Jonnard 494 Interferometric Methods, R. Jonnard 502 Refractometric Applications, R. Jonnard 515 RESINOGRAPHY, T. G. Rochow 525 STEREOSCOPIC MICROSCOPY 538 Basic Design, Operation and Use, J. R. Benford 538 Engineering Microscope. See Light (Optical) Microscopy, p. 434 "Solid-Image" Microscope, Richard L. Gregory 540 TELEVISION MICROSCOPE, G. L. Clark 542 ULTRAMICROSCOPY 544 Design and Operation. See Optical Theory of the Light Microscope, p. 445 ULTRASONIC ABSORPTION MICROSCOPE, Floyd Dunn and Wm. J. Fry 544 ULTRAVIOLET MICROSCOPY 548 Basic Principles and Design, J. R. Benford 548 Color Translating TV Ultraviolet Microscope, G. L. Clark 550 Image Formation by a Fresnel Zone Plate, Albert V. Baez 552 X-RAY MICROSCOPY 561 Bone Structure and Aging by Contact Microradiography 591 See Medico-Biologic Research by Microradiography, p. 591 Contact Microradiography, H. H. Pattee 561 Diffraction Microscopy, G. L. Clark 569 Eye Research Applications, W. K. McEwen, M. L. Feeney, and L. K. Garron ... 571 Xll Contents Geological, Mineralogical and Ceramic Applications of Microradiography, W. Johnson 574 Grainless Media for Image Registration, G. L. Clark 581 Histology by the Projection Microscope, R. L. deC. H. Saunders 582 Inter-Relation of Techniques for the Investigation of Materials, K. W. Andrews 586 Iron and Steel Applications of Microradiography, K. W. Andrews and W. Johnson 587 Medico-Biologic Research by Microradiography, F. Bohatirchuk 591 Microangiography with the Projection Microscope, R. L. deC. H. Saunders. . 627 MicROFLUOROScoPY. See Contact Microradiography, p. 561 Plant Microradiography, ./. Salmon 636 Point Projection X-Ray Microscopy, W. C. Nixon 647 Production of Continuous and Characteristic X-Radiation for Contact and Projection Microradiography, N. A. Dyson 653 Projection Microscopy, Ong Sing Poen 661 Reflection Microscopy (Kirkpatrick), G. L. Clark 672 Two-Wave (Buerger) Microscope, G. L. Clark 674 Ultrasoft X-Ray Microscopy, Burton L. Henke 675 Vascular and Dental Applications of Projection Microscopy. See Micro- angiography, p. 627 Xlll Autoradiography AUTORADIOGRAPHY OF TISSUE* This article discusses techniques used for autoradiography of human and animal tis- sues and, to a limited extent, plant tissues. Methods are outlined which will enable those relatively inexperienced in autoradi- ography to plan an experiment in\'olving alpha, beta and gamma emitters and to pro- ceed with the experiment confident of ob- taining meaningful autoradiograms. A cur- rent bibliography for the period 1954 to 1959 is that of Johnston (1). Boyd (2) has cited numerous literature references. Foreign journals often cany detailed papers relating to autoradiography which should not be overlooked as a source of information. The technique which produces an image on a photographic plate or film when radio- active material is opposed to it is called autoradiography. The result of the exposure of an emulsion to a radioactive specimen is called an autoradiogram. The autoradiogram supplies a graphic record of the sites of depo- sition of radioactive isotopes within or on a tissue and may be macroscopic, as in the case of plant leaves, or in some cases micro- scopic, at or below the cellular level. Au- toradiography of tissues containing alpha, beta and gamma emitters may be performed. Suggested Autoradiographic Tech- niques for Alpha Emitters; Pluto- nium In industries where contamination of per- sonnel with radioactive substances is pos- * Work performed under Contract No. AT(45- 1)-1350 between the Atomic Energy Commission and the General Electric Company. sible, low level detection procedures are necessary. Plutonium, an alpha (5.14 Mev) emitting radionuclide with some gamma ra- diation and a 24,300 year half-life, may be used as an example to illustrate one auto- radiographic technique by which either par- ticulate or soluble material may be graphi- cally localized. A 24-hour sputum sample (3) from a per- son known to have inhaled plutonium was taken and fixed in 10% formalin, as was a human biopsy of skin removed from a punc- ture wound in the hand. The samples were dehydrated in "Cellosolve" (glycol mono- ethyl ether), cleared in xylene, and em- bedded in paraffin. Sections were cut and floated on a water bath, transferred by a clean slide to a crystalhzing dish containing distilled water. Working under light filtered by a Wratten OA filter, a 5 /x NTA emulsion coated on a 1 x 3 inch slide with a thin gela- tin protective "T" coat was shpped under the section. The excess water was drained on to filter paper; the slides were then placed in a Hght-tight plastic box made for this pur- pose. A small vial of a desiccant (CaS04) lightly closed with a cotton plug was added. The appearance of the tracks from pluto- nium which was in solution is characterized in Figure 1 (a human sputum specimen), where individual alpha tracks proceed in a straight line. The appearance of the tracks in sputum when plutonium is in the par- ticulate form is illustrated in Figure 2, in which alpha tracks arise from a central point, with some random single tracks in the field. Thirty-three sections from the plutonium- AUTORADIOGRAPHY *# 50/i %^ 4 i i I "^yft Fig. 1. Autoradiogram of stained section of human sputum illustrating diffuse alpha-track pattern of plutonium in solution. {Courtesy Stain Technology^) . contaminated skin biopsy (4) were auto- radiographed by the above processing method. All of the two-hundred fifty indi- vidual skin sections were scanned in a thin window alpha detector. Radioanalysis showed that the entire skin biopsy specimen contained 0.0046 microcurie and the indi- vidual sections varied from 0 to 362 disinte- grations per minute. The appearance of one autoradiogram of a human skin section is illustrated in Figure 3, in which alpha tracks may be seen arising from particles. In addi- tion, random alpha tracks may be seen. Suggested Autoradiographic Tech- niques for Beta Emitters; I^^S Ru^**^, Sr«o, and P^^ I^^i, with an eight day half -life, is predomi- nantly a beta emitter with some gamma emission. It was administered to sheep in an acute and chronic feeding experiment (5) to determine the effect of radioiodine on the thyroid gland of grazing animals. The auto- radiographic response of the thyroids of some of these animals is of interest. Thyroid samples were routinely fixed in Bouin's solution for from 8 to 12 hours, fol- lowed by dehydration in "Cellosolve" and nor- mal parafl&n embedding. Adjacent sections were selected from the ribbons. One section was stained routinely with hematoxylin and eosin; the other was floated into a crystalliz- ing dish containing distilled water, from which it was floated onto a slide bearing a 5 fj. NTB emulsion plus a protective "T" coat. The slides, for animals which had received 100 nc P'^^ I.V. and which were sacrificed after 1 hour were then exposed for 5 days; the slides from animals receiving 100 nc I.V. in 0.9% saline and sacrificed after 4 hours were exposed for 53-^ days ; and slides for ani- FiG. 2. Autoradiogram of stained section of human sputum illustrating star formation of alpha tracks originating from plutonium particle in cen- tral position. {Courtesy Stain Technology^). ALTOKVDIOGRAPHY OF TISSUE ^.^•^ 100 /x Fig. 3. Autoradiogram of human skin showing distribution of plutonium and alpha tracks ap- proximately 800 M beneath stratum corneiun. Hema- toxylin and eosin preparation. (Courtesy Acta Radiological) . mals which had received 5 mc per day for 129 days were exposed for 9 days. The last animal was sacrificed during gestation and had previously suckled its dam for 4 months. The dam had also been receiving 5 mc a day, so that some additional I^^^ was received by the lamb through the milk. The autoradiograms were exposed in light- tight boxes containing "Drierite." The slides were warmed to room temperature and de- veloped in D-19 at 18°C for 5 minutes, water rinsed and fixed in x-ray fixer for 30 minutes. They were water- washed and stained with hematoxylin and eosin. Sections that adhere to the emulsion can usually be stained satis- factorily, if the staining procedure is not pro- longed to the point where overstaining of the emulsion is apparent. The appearance of the autoradiograms of sheep thyroid tissue from animals adminis- tered 100 /xc intravenously and sacrificed in 1 hour, illustrated in Figure 4, shows the grain response to be relatively uniform over both the colloidal areas and the epithelial cells. The colloidal area in animals admin- istered 100 jiQ. and sacrificed in 4 hours is shown in Figure 5. The grain distribution is more pronoimced over many of the col- loidal areas, indicating a variable uptake of P^^ by the colloid in some of the follicles. For animals fed 5 /xc a day for 129 days (Figure 6) there is an even distribution of grains over the colloidal areas in thyroid follicles which are greatly decreased in size from the normal. Areas showing no grain density indicate the edema surrounding the few remnant follicles that are present. The microfollicles register a limited grain den- sity. Little or no interstitial tissue is present. Si,— ■ 3P^ - ***.- 7^' ■* ■»^ •1* ■» r .i ■ 1 : • r^ \ * #.■ 4 « * 1 1. ; - » ^ ^jHKh ,v ■'•^ ■* ^ ^ , ^ ^ ■^ J r^ •-. „__ ^ w ' , # r^ B- *« «. ■> Pi i i -^ h- -50/i-^, 4," Fig. 4. Autoradiogram of hematoxylin and eosin stained sheep thyroid. Animal was adminis- tered 100 /iC of I'^i I.V. and sacrificed in one hour. Limited grain density increase over colloidal area and epithelial cells. AUTORADIOGRAPHY case the ruthenium was administered in an insokible particulate form. Ru^°^ particles were administered to mice by intravenous and intratracheal injection in 0.1 % aqueous dispersions of the wetting agent, Tween 80 (6). The animals were sacri- ficed after 100 to 420 days exposure to beta radiation. One experiment in autoradio- graphic quantitation was conducted on a single mouse into which 47 /xc of Ru^"® par- ticles, ranging in size from 0.5 to 0.8 micron, were injected into the tail vein of the mouse. A portion of these particles lodged in the lungs of the animal. The lungs were excised after 100 days exposure and fixed in 70% ethyl alcohol. Autoradiographic processing was standard, as outlined above. The re- sult was a series of sections through the entire lung; one section was stained with orcein, one unstained autoradiogram fol- » t-IOOft-) Fig. 5. Autoradiogram of sheep thyroid four hours following administration of 100 nc of I'^^ Hematoxylin and eosin stain. Specific increase in grain density over various colloidal areas indicates variable uptake by the follicles. It is important that the investigator be aware of the type of response to be expected after processing certain types of tissue and report the appearance of representative au- toradiograms that are taken to illustrate the uptake of the radioisotope at selected times following administration. Specifically, the autoradiographic response in the thyroids of the experimental animals mentioned is that of single grains diffusely covering specified areas of the thyroid. Other types of emul- sions will register the location of I^^^ as tracks rather than diffuse grains. Changes in the development procedures are necessary to register the I^^^ as tracks rather than grains. It is interesting to compare at this point the autoradiographic response of Ru^"®, an- other beta emitter having some gamma emission, and a half-life of one year. In this h*\ '4*' «• • V ^^ % **' ^ * i • • " " " ■ * ■• J * •» % 50/X- .-^^ :# ^ t Fig. 6. Autoradiogram of sheep thyroid 129 days following administration of 5 /xc of I"' per day. Negative grain density is apparent over areas of edema, with a limited uptake in the colloidal areas of the microfollicles. 4 AUTOKADIOGRAPIIY OF TISSUE lowed it, and this in turn was followed by a section stained by the Mallory method, to register connective tissue locations in rela- tion to the "spot diameter" regponse of the unstained autoradiograms. A "spot diam- eter" is the grain response of the emulsion to a single radioactive particle. The staining sequence was followed through the entire lung. All of the sections were measured for Ru^"^ content on a mica window beta counter. The insoluble ruthenium particles elicited a specific "spot diameter" response in the NTB emulsion. This was a well defined cir- cular area of developed emulsion grains over the precise location of the deposited par- ticles, as illustrated in Figure 7. Knowing the total activity density for a specific slide, the activity density of a single particle was taken as proportional to the diameter of the darkened area of the emul- sion. Thus, if a number of particles were present on a single slide, the total counting rate was compared with the sum of all of the darkened areas on that slide. Therefore, each particle could be assigned an activity density evaluation directly proportional to the size of the individual "spot diameter" measurement of the emulsion. The "spot diameter" darkenings of the grains of the emulsion varied in size and ranged from 8 fj, to 170 fjL. A quantitative straight -line func- tion exists between the sum of the spot diameters or a single-spot diameter and the total activity of each autoradiogram. This particular technique provides a means of measuring the radioactivity of a single radioactive particle or many particles within tissue when the exposure, processing and development are standardized. The par- ticles occur in a pattern of distribution such that the accumulated dose from nearby par- ticles is sufficient to initiate fibrosis. By means of the serial sections stained specifi- cally for connective tissue, and the autora- diograms, it is possible to reconstruct the tissue in depth and study the relationship ^ 4 GRAIN DENSITY Fig. 7. Stained lung-autoradiogram prepara- tion illustrating deposition of Rui^^Qa particles with resultant spot diameter darkenings. (Courtesy J. Biol. Phot. Assoc.^^). of the fibrosis to the specific location of the deposited radioactive particles. The autoradiographic technique for Sr^oS04 in lung tissue of mice, administered 1 /iC per animal by inhalation, is mentioned here for comparison with the foregoing meth- ods. Sections from lung tissue fixed in 80% ethyl alcohol were processed in a manner similar to lung tissue containing Ru*"^ par- ticles (7). The film used in this case, how- ever, was 25 n No-Screen x-ray, single coated on 1 X 3 inch slides. These slides had a pro- tective coating. As counting techniques indi- cated a low activity densitj^ for the sections, they were exposed for 8 weeks. The use of X-ray emulsion made it possible to register activity densities that were approaching the lower limits of detection by normal counting methods. The autoradiogram of the lung section containing Sr^", illustrated in Figure 8, shows both a diffuse and particulate response to the radioactive material. The adjacent his- autok\i)I()<;h\phy ■7 ;' ■ v»-lf % ,^ PARTtCfe€- -, REAC; h-6Q0/i."i Fig. 8. Autoradiogram of lung section from chronic Sr'''S04 inhalation experiment. Particulate and diffuse distribution with heavy background grain density of x-ray emulsion visible. (Courtesy J. Biol. Phot. Assoc.^^). tological section is included, as Figure 9, for comparison purposes. A noticeable back- ground of grains, larger than in the NTA and NTB emulsions, is observable. This is due in part to the 25 ^ thickness of the x-ray emulsion and in part to the greater sensitiv- ity. Even greater sensitivity could have been obtained had the film used been of the double coated type. However, the image rendered would have been more diffuse because the emulsion layers are coated on either side and separated by an acetate base. In this case detail has been sacrificed to record low activity density deposition of diffuse and particulate Sr^". The next example is that of a beta emitter, F'\ which has a 14.5 day half -life (8). A gross, fast survey method for the detection of P^- administered to rats by intraperitoneal injection of 2.5 /xc per gram of body weight is outlined. The rats were sacrificed 24 hours following administration of the P^^ ^s Nao- HP^^o^ The emulsion of choice in this case was 25 n, single layer, No-Screen x-ray emulsion on 1 x 3 inch slides. Detail was sacrificed for rough localization of P^^ by using an emulsion with a grain size varing from 3 to 5 /i. The increase in the sensitivity of the emulsion and the thickness insured a positive record- ing of the radioactive sites, with a minimum exposure period. Noticeable disadvantages with this techniciue are the large grain size which makes cellular autoradiography im- possible. The extreme thickness of the emul- sion makes illustration of the recorded data on the film difficult, except by macrographic methods. Figure 10 is an autoradiogram of a rat ovary section from an animal sacrificed 24 hours following intraperitoneal administra- '^W ,»*-e<,y.4-" I'i't^?'. %. I— 500/i-li,, Fig. 9. Hematoxylin and eosin stained histo- logical section adjacent to Fig. 8, for orientation purposes. {Courtesy J . Biol. Phot. Assoc.^^). ALTORADIOGRAPHY OF TISSUE .-">~^jjgg«»>- -*'*rie'i-'«Trs '."^s^K-? •-»»• * # Fig. 10. Autoradiogram of rat ovary 24 hours after administration of 2.5 juc of P'^ pgj. gram of rat. Concentrations of P^^ appear over cellular ele- ments lining the follicular cavity, surrounding the ova and in some corpora lutea. {Courtesy J . Biol. Phot. Assoc.^^). tion of P^2_ ^n increased grain density is noticeable in the area of three developing ova. The autoradiogram may be compared with the adjacent hematoxylin and eosin stained section in Figure 11. In direct contrast with this gross survey method for the localization of P^- is the ex- treme detail and precise localization ob- tained by Guidotti (9). Guidotti mounted paraffin sections, 2 to 4 ;u thick, which were stained, dehydrated, and allowed to air dry. These were given a thin coating of 1 % "Plexi- glas" solution in chloroform. The chloroform was allowed to evaporate completely in a dry atmosphere. An emulsion having a dried thickness of 100 to 150 microns, was pre- pared from Ilford G-5 type emulsion in gel form and glued to the section by means of a 15 % solution of shellac in absolute alcohol. The surface of the emulsion was then cleaned with absolute ethyl alcohol, to remove the impermeable layer. This method allows the fixing and staining of the tissues by means of all of the reagents commonly employed in histology, without any damage to the emulsion. The exposure of the tissue was for 24 hours at 2°C. The emulsions were proc- essed by the Temperature Development Method. Good adhesion and minimum sep- aration between specimen and emulsion are obtained, thus permitting reliable extrapo- lation of the electron tracks from P^-. The Temperature Development Method of Dilworth, Occhialini and Payne, 1948, and Dilworth, Occhialini and Vermaesen, 1950, as modified by Guidotti, is worthy of specific mention. Essentially it consists of soaking the slides in distilled water for 30 to 40 min- utes, beginning at room temperature, and '^ Fig. 11. Stained histological section adjacent to Fig. 10 of rat ovary 24 hours after administra- tion of 2.5 Aic of P'2 per gram of rat. Three develop- ing ova are seen in central portion. (Courtesy J. Biol. Phot. Assoc.'^). AUTOH ADIOGR A1»IIY then lowering the temperature gradually to about 4°C. The cold phase of development is for about 30 minutes. The developer used in this case was Amidol. The cold developer is poured off and the surface of the slides dried with filter paper. Warm slides to 22° to 24°C in a thermostatically controlled tank. The temperature of the warm stage must be chosen by trials. The slides are then placed in a stop bath of 0.2% acetic acid solution for about 30 to 40 minutes at 4°C. Fixation is at 4°C in a 40 % sodium thiosul- phate solution until the emulsion is clear. The emulsion is thoroughly washed at the low temperature and the emulsion allowed to dry slowly. The preparation is then capped with a thin coverslip. One of the reasons for Guidotti's success with P^- autoradiography is the varied tem- perature development allowing all layers of the emulsion to be penetrated by the devel- oper; another is the thinness of the section, which is considerably less than most routine preparations. This allows the site of origin of the beta tracks to be more accurately located. Fig. 12. Autoradiogram of rat ovary and adja- cent tissue illustrating even diffuse distribution of Cs"'. Cs^^'^, a beta and gamma emitter, with a half -life of 30 years, is an example of a water- soluble isotope that presents difficulties re- garding methods for autoradiographic regis- tration. The usual fixing fluids leach Cs''^ from tissues as will any contact with water; there- fore, a mixture of 80% acetone and 20% benzene was used as a fixative (10). This combination was decided upon after ex- perimentation in which the leaching loss was determined by radiochemical analysis. There was a minimum cesium loss when this fixa- tive was used. Tissues were then processed through "Cellosolve" and benzene and blocked in paraffin. Since the sections could not be floated on water for transfer to the emulsion, sections were attached to the No-Screen x-ray emul- sion by warming the slide until the emulsion became tacky, then the sections were at- tached by finger pressure. The slides were then exposed, the paraffin dissolved by im- mersion in xylene. The slides were run down to water from "Cellosolve" and the autora- diograms stained and the slides capped. Figure 12 is an illustration of Cs'^^ deposi- tion in a rat ovary. It will be noted that there is a homogeneous distribution in the ovary and the surrounding tissue. There is no evi- dence of leaching into the emulsion area not covered by the tissue. Figure 13 is the adja- cent section for histological comparison. Greater definition could have been obtained by using an NTB emulsion. Limited Autoradiographic Techniques for Plant Tissue An interesting experiment performed by Hungate et at. (11) involved exposure of plants to fallout from the experimental burn- ing of an irradiated fuel element. The fission products resulting from the burning of the fuel element were carried by the air to plants located downwind from the burning site. Later the plant leaves were exposed to double-coated Type KK indus- 8 alt()hai)io(;raphy of tissue trial x-ray film. The gross autoradiographic result of this exposure of two plant leaves is illustrated in Figure 14. It will be noticed that there is a general darkening over the entire leaf with an increased density at its periphery. In addition, there are small black circular areas of increased grain density that are relatively indistinct. These circular areas of increased grain density suggest that some of the fission products were in the particulate form. In order to substantiate the observation regarding specific deposition of particles, additional 1x3 inch plates coated with a 100 M NTB^ emulsion and a "T" coat were exposed to portions of the plant leaves for 4 days and developed in D-19 developer. These also demonstrated a particle response that was more detailed than that observed with the double-coated x-ray emulsion. The detailed microscopic particle response of fission material on plant leaves is illustrated in Figure 15. In this case the autoradio- 7 ; \$* Fig. 14. Glu^^ auloradiograms of plant leaves following e.xposure by air of the leaves to fission products. An over-all darkening with increased density at the peripher}^ is noticed. Some small circular spot densities can be seen. graphic technique demonstrated the presence of particles on lea^'es that was not detectable by routine counting procedures. Fig. 13. Stained tissue section of ovary and re- lated tissue adjacent to section used for autoradi- ography of Cs"' in Fig. 12. General Discussion Up to this point practical problems involv- ing autoradiography have been discussed. ]\Iany specific applications have been AUTORADIOGRAPHY >^>.-^^ ^'^:;mt PARTICLE REACTION ^-lOO^^ Fig. 15. Microscopic autoradiographic re- sponse of leaf exposed to fission products. Specific spot grain responses indicate that some of the ma- terial was in particulate form. omitted. Among them is the classical work involving plutonium deposition in the bones of dogs administered plutonium. This is the excellent work of Jee (12) in which the results of autoradiographic methods involv- ing hard bone are discussed in detail. Autoradiographic techniques have been applied to electron microscopy. O'Brien and George (13) have examined sectioned yeast cells, previously suspended in a Po^^*^ solu- tion. In this case the specimen grids were coated with the emulsion by touching them to a drop of diluted NTA emulsion. Borasky and Dockum (14) examined Pu^^^ particles contained in an aerosol sample collected on "Formvar" coated grids. The grids bearing the particles were chrome-shadowed and placed on "Lucite" plugs held upright in wells drilled in a metal block. A small wire loop was placed in diluted NTA gel emulsion and retracted, thus forming a very thin layer of emulsion. The loop bearing the emulsion was then lowered over the specimen grid and the loop retrieved by lifting the plug and withdrawing the loop from beneath the plug. The grids were exposed for 4 hours in a light- tight box when high activity specimens were examined. The emulsion-coated grids on the "Lucite" plugs were immersed in D-19 de- veloper for 5 minutes, then washed in dis- tilled water for one minute, after which they were fixed in liquid x-ray fixer for 3 minutes, washed and dried. The individual grains of the alpha tracks could be plainly seen when examined by an RCA EMU-2C electron microscope. The point of origin of the tracks was located by the shadow cast by the Pu^^^ particle. George and Vogt (15), using unshadowed grids, examined plutonium particles collected on millipore filters, and prepared electron micrographs of selected areas of particles before and after autoradiography, thus dif- ferentiating radioactive from non-radioac- tive particles. It is probable that small cubes of tissue of approximately 150 microns could be infil- trated in diluted gel emulsion long enough to penetrate the tissue. These small cubes could be exposed for the desired amount of time, developed by the Temperature Devel- opment Method, after which the cubes could be embedded in methacrylate and sec- tioned by ultra-thin sectioning methods. Selected areas of cells could be studied in relation to the developed grain deposition either by electron microscopy or by oil im- mersion phase microscopy, if mounted on a glass slide (16). This is an avenue that will lead the microscopist into cellular and sub- cellular autoradiography. REFERENCES 1. Johnston, M. E., "A bibliography of bio- logical applications of autoradiography, 1954 through 1957", UCRL-8400, 1958. Johnston, M. E., "A bibliography of biologi- cal applications of autoradiography, 1958 through 1959", UCRL-8901, 1959. 10 SHADOW AUTOR ADIOG K A PHY 2. Boyd, G. A., "Autoradiography in biology and medicine", Academic Press, New York, 1955. 3. DocKUM, N. L., Coleman, E. J., and Vogt, G. S., "Detection of plutonium contami- nation in humans by the autoradiographic method". Stain Technology, 33(3), 137-142 (1958). 4. DocKUM, N. L., AND Case, A. C, "Autoradio- graphic analysis of plutonium deposition in human skin", Acta Radiologica, 50, 559-64 (1958). 5. Marks, S., Dockum, N. L., and Bustad, L. K., "Histopathology of th3'roid gland of sheep in prolonged administration of 1"^", Am. J. Path., 33, 219-250 (1957). 6. Dockum, N. L., and Healy, J. W., "Spot diameter method of quantitative auto- radiography of ruthenium^o^ particles in lung tissue". Stain Technology, 32(5), 209- 213 (1957). 7. Unpublished data, W. J. Bair. 8. Vogt, G. and Kawin, B., "Localization of radioelements in rat ovary". Document HW-53500, p. 120-123 (Unclassified), 1957. 9. GuiDOTTi, G. AND Levi Setti, R., "Auto- radiography of tracks from beta particle emitters in tissues". Stain Technology, 31, 57-65 (1956). 10. Unpublished data, N. L. Dockum. 11. Hungate, F. p., Uhler, R. L., Cline, J. F. AND Stewart, J. D., "Decontamination of plants exposed to a simulated reactor burn", Document HW-63173 (Unclassified), 1959. 12. Jee, Webster S. S., Arnold, J. S., Mical, R., Lowe, M., Bird, B. and Twente, J. A. "The sequence of histopathologic bone changes in bones containing plutonium", p. 148-189. Univ. of Utah Radiobiology Lab- oratory Annual Progress Report, COO-218, 1959. 13. O'Brien, R. T., and George, L. A. II, "Prep- aration of autoradiograms for electron mi- croscopy". Nature, 183, 1461-1462 (1959). 14. Unpublished data, Borasky, R. and Dockum, N. L. 15. George, L. A. II, and Vogt, G. S., "Electron microscopy of autoradiographed radioactive particles", Nature, 184, 1474-1475 (1959). 16. Dockum, N. L., Vogt, G. S., and Coleman, E. J., "Applications of autoradiography in biological research", J. Biol. Phot. A,s.soc.,27, 1-18 (1959). Norman L. Dockum SHADOW AUTORADIOGRAPHY Shadow autoradiography is a technique that enables one to differentiate between radioactive and non-radioactive particles. The source of the particles may be suspen- sions or aerosols. The technique is described below. Droplets of particle suspensions are placed on clean glass microscope slides, spread and allowed to dry. Aerosol particles may be collected directly on the glass substrate by gravity or by impaction methods. If the par- ticles from aerosols are collected on mem- brane filters, the membrane filter is dis- solved in a suitable volume of acetone and (a) (b) Fig. 1. Alpha emitters. Photomicrographs of chrome-shadowed particles before (a) and after (b) autoradiography. Alpha-emitting radioactive particles are those surrounded by star clusters of alpha tracks. (Courtesy N. L. Dockum). 11 AUTOR ADIOGKAPIIY (a) [b) Fig. 2. Beta emitters. Photomicrographs of chrome-shadowed particles before (a) and after (b) autoradiography. Beta-emitting particles are covered by spots of dense granules. (Courtesy of N. L. Dockum and R. Borasky, Nucleonics, 15, 110 1957). W ■'^ % ^ ££^ '♦•- i "* K Wl^ ■■* • ' ^ ^ . .^^•%' >5* (b) Fig. 3. Photomicrographs of superimposed negatives of areas before and after autoradiography. (a) Alpha-emitting particles, (b) Beta-emitting particles. 12 SHADOW ALTORADIOGRAPIIY aliquot s of the acetone suspension treated in the same manner as described above. The glass slide containing the particles is shadowed with chromium at an acute angle (e.g. 30°) (1). The shadowed slide is next examined in the optical microscope and areas of interest are recorded by stage mi- crometer readings and by photomicrography. Following the microscopic examination, a section of 5 micron nuclear track emulsion stripping film (e.g. NTA for alpha-emitting particles and NTB for beta-emitting par- ticles) is floated over the sample and allowed to dry. The latter step is performed in a dark room using a series I Safelight filter. The slide bearing the particles and emulsion is placed in a light-tight box with a desiccant for a suitable standardized time for exposure at 3°C. The exposed slides are developed in D-19 developer for 5 minutes, washed, fixed, dehydrated and cleared, and a coverslip mounted over the emulsion. The slide is re- examined in the microscope and previously selected areas are photographed. Alpha activity is manifested by star clus- ters of alpha tracks (cf. Fig. la and lb). Beta activity is manifested by spots of dense granules (cf. Fig. 2a and 2b). Beta activity may be determined quantitatively by spot- diameter autoradiography (2). The method for differentiating radioactive from non-radioactive particles is as follows. A sheet of tracing paper is placed over the photomicrographs or negatives of selected areas of the shadowed slide before auto- radiography. Particle positions are noted with small circles. The tracing is next placed over the photomicrograph or negatives of the slide after autoradiography and activity centers noted with a check mark or cross. The tracing now has small circles represent- ing non-radioactive particles and crosses or check marks denoting radioactive particles. Siriking results are obtained by superimpos- ing the negatives of the areas before and after autoradiography (cf. Figure 3). The physical characteristics of the particles are determined from the photomicrographs of selected areas on the shadowed slide before autoradiography. REFERENCES 1. Dempster, W. T., and Williams, R. C, Anat. Record, 96, 27 (1946). 2. DocKTJM, N. L. AND Healy, T. W., Stain Tech- nologij, 32, 209 (1957). 3. DocKUM, N. L. and Borasky, R., Nucleonics, 15, 110 (1957). R. Borasky Chemical microscopy ALKALOIDS AND ALKALOIDAL-TYPE PRECIPI- TATION An attempt is made in the article on Chemical Microcrystal Identifications to keep the discussion of practical work on a sufficiently general basis so that any analyti- cal chemist concerned with identification may be able to see some application to his own work. If a consistent effort is made to use and develop this kind of chemical iden- tification, it is usually best to work from tests that are already known and satisfac- tory in a restricted field, and to extend the coverage first to chemically related com- pounds, having the same or modified re- agents, and by stages, systematically try to cover entirely different groups of compounds with other reagents and new crystal-produc- ing reactions. 13 CHEMICAL MICROSCOPY The traditional tests on the well-known alkaloids are made on aqueous solutions. About 80 substances, mostly natural alka- loids, or of natural origin although syntheti- cally modified, have been studied repeatedly by different investigators for their reactions in such tests. Table 1 summarizes the data for 16 of these with the 11 reagents which give most of their major aqueous micro- crystal tests. These reagents are equally applicable to numerous other alkaloids. It is luifortunate that publication of the isolation of a new plant alkaloid is almost never accompanied by some individual or highly characteristic analytical test by which subsequent investi- Table 1. Some Important Microcrystal Tests FOR ^ ^LKA LOIDS i) ■3 o to K- 1 s 1— i u u i-i PQ d < c o o K u < c 5 en u r2 < u u 8 Cli U W ■a o fa other Caffeine (c) C c C — — — HAuBr4 Ephedrine a C — a (ac) a — C — — — HaPtIs rgt Nicotine c c ac c C c C c c — — HoPtBre Aconitine ac ac a a a a a — a C ■ — HMn04 ; HCIO4 Scopolamine c C a C C c (a) (ac) ac — C I-KI, 1:35 Hyoscyamine C C a c C c c a c (ac) c I-KI, 1:35; Br-HBr Atropine C c a c ac C (c) (a) c ac c I-KI, 1:50; Br-HBr Morphine C(l) C C(2) ac a a (c) (a) a(c) c — HCl-HAuBr4 (HCl soln) Codeine C c(a) C a a a c a a(c) (e) — Hgl2-NaCN & Nal (3) Diacetyl-mor- a ac ac ac a(c) ac C C(4) c C a HAuBr4 in (2 + 3) phine (Her- H2SO4 oin) Meperidine a ac a ac ac c ac C C a c Pbl2 in K acetate soln Procaine ac a(c) a C a C c(a) c c(a) (a) (c) H.PtBre ; Br-HBr Cocaine a a a C C C a C(5) c c c HaPtBre ; Pblo in K acetate soln Quinine a ac a ac a(c) c ac a a a(c) C Herapathite test; Na2HP04 ; HgCU & HCl (6) Strychnine c a(c) C C C C C c C c C CrOs ; numerous others Narcotine a a a a a a a a a C a K2Cr04 ; K acetate soln a, amorphous precipitate, not crystallizing c, crystals form, either directly or from an amorphous precipitate a(c), amorphous precipitate, crystallization poor or uncertain c(a), normally crj^stals but precipitate may remain amorphous ac, discrepant reports, usually because of variations in reagents or because crystallization, while fairly certain, is quite slow (a), (c), or (ac), amorphous precipitate or crystals may not form in most tests because of lack of sensitivity — , no result C, major crystal tests 1-6, Figures. 14 ALKALOIDS AND ALKALOIDAL-TYPE PRECIPITATION gators might distinguish it. When, in some study, alkaloids are isolated from plant ma- terial on a small scale, and are not among those commercially available, one usually finds that, even though the plarit from which they came is known, and the alkaloids found are distinguished by analytical reactions, it is still impossible to tell which one if any corresponds to some previous^ studied al- kaloid of that plant, without going through most of the original procedure. This necessi- tates a large quantity of material, obtaining Fig. 1. A well-known crystal test: morphine with iodine-KI (aqueous solutions). ""% ''^. "t fh ^* % ^^' Fig. .3. A less-known alkaloidal crystal test: codeine with Hglo-NaCN and Nal (aqueous solu- tions). Fig. 4. A well-known crystal test: diacetylmor- phine (heroin) with H2PtCl6 (aqueous solutions). : ■ >ti I Fig. 2. A well-known crystal test: morphine with K2Cdl4 (aqueous solutions) . melting-points, and analyzing for elementary composition, except in the rare event that a sample of the original isolate can be obtained for comparison. This kind of gap between research and application is very common. Other Compounds with Aqueous Re- agents. Alkaloids do not constitute a dis- tinct chemical group. Their precipitation with the "general alkaloidal reagents" de- pends upon their being compounds of basic 15 CHEMICAL MICROSCOPY Fig. 5. A well-known crystal test: cocaine with H2PtCl6 (aqueous solutions). Y^^--M^ "W Fig. 6. A less-known alkaloidal crystal test: (2) quinine with mercuric chloride and HCl (aque- ous solutions). nitrogen, usually of a fairly high degree of complexity. There are now numerous drugs that are not considered alkaloids, but which have the same characteristics: they are nitro- gen-bases, soluble in dilute acids, precipi- tated by exactly the same general reagents, and give characteristic crystals with certain of them (different ones, depending on the compound), just as the alkaloids do. Among these are the synthetic narcotics, local anes- thetics, antihistamines, antimalarials, atro- pine-like drugs and some kinds of tranqui- lizers. In some cases, whole classes of these drugs are recent introductions, and for all of them, the number in use has expanded greatly in recent years. E. G. C. Clarke has given microcrystal and color tests for 101 alkaloids (including those synthetically modified), and for 15G of these other drugs (to the date of writing), with the same reagents. Other classes of drugs which are amine bases are the sympathomimetics and central stimulants. Some of these also give good tests with the traditional aqueous reagents. Others, relatively simple in structure and often having hydroxyl groups, may be too water-soluble in their compounds for good tests in this way, and precipitation in phos- phoric-acid solution is used. Many different acidic and anionic re- agents are available for basic compounds and it would seem that it should be equally possible to use basic reagents or cations to provide microcrystal tests for acidic com- pounds, at least those of some complexity. As a matter of fact, tests have been pub- lished for various acids using salts of silver, lead, copper, nickel, mercury, zinc, and pal- ladium, usually in a solution made slightly basic with ammonia, pyridine, triethanol- amine, or the like. Such tests are useful but as yet most of them hardly seem in the same class with the better alkaloidal tests, either for proving identity or in sensitivity, and much study to develop good tests is needed. Aqueous Precipitation of the Free Substance. Many alkaloids are precipi- tated from acid solution by making the solution basic, or even, in the case of weak bases, by merely reducing the acid strength, for example with potassium acetate. Charac- teristic crystals are often produced and the tests also give chemical information merely by the fact of precipitation. The strength or weakness of the base is shown to some degree by the strength of the basic reagent required for sensitive precipitation. K2Cr04 may pre- 16 ALKALOIDS AND ALKALOIDAL-TYPE PRF.CIPITA TION cipitate the chromate of a strong base but its really sensitive tests are those in which it acts as a basic group reagent for weak com- plex alkaloids. Substances soluble in both strong aqueous acids and bases may be insoluble at some intermediate point. The method is obviously general and acidic substances may be pre- cipitated from alkaline solution by ordinary acids. This is one way of obtaining crystal tests for barbitm-ates. While using various ways of separating the free substance for microcrystal tests, the chemical meaning of the reactions should always be noted. Tests are also made on an acidic substance (which may be in fine particles, or amor- phous, to start with) by dissolving it in NaOH solution, adding HCl in slight excess, and allowing the drop to evaporate to dry- ness. The reagents leave merely grains of NaCl, which are isotropic, and therefore in- visible with crossed nicols; the acidic sub- stance set free may crystallize, and may often be birefringent and easily seen, and then its refractive indices and other optical properties can be determined. Halogen Reagents for Nitrogen Com- pounds in Aqueous Solutions. Bromine in water, or in HBr or NaBr solution, and iodine in KI or HI solution, are the reagents commonly used; iodine may also be used in HBr or NaBr solution. These are very gen- eral precipitants not only for the alkaloids but in general for amine compounds of any complexity. In a somewhat different type of reaction their use extends even to compounds in which the nitrogen no longer has appre- ciable basic properties. The alkaloid-type reaction takes place in acid solution, or at least the alkaloid is com- bined with acid in a salt and itself acts as a cation. When an amine-derivative is pre- dominantly acidic and is precipitated as an iodine compound in acid solution, it may not be clear whether the reaction should be attributed to residual basic qualities, or not. The related but anionic reaction is most clearly seen as a distinct type in neutral or slightly basic solution; occurs best with a high concentration of iodine. This type of reaction, which has not been explored nearly enough occurs with some barbiturates and with other compounds without appreciable basic properties, and also with some com- pounds that can react either as bases or acids, including the alkaloids caffeine, theo- bromine, theophylline, and colchicine, which give both the alkaloidal-type and the ani- onic-type of iodine precipitation. They all have acidic C=0 groups, / Oxygen Acids. These may be divided into two kinds, complex and simple. The complex oxygen acids, phosphoritungstic, phosphorimolybdic, arsenimolybdic, sihco- tungstic, etc., are very general, precipitating from aqueous solution all the alkaloids and related compounds, and, generally speaking, potassium and the heavier alkali metals, ammonium and the simple amines, as w^ell. As a rule they give crystals only with the simpler bases; most of their alkaloidal pre- cipitates are amorphous. Phosphorimolybdic acid is, however, very useful for comparative precipitation. Instead of the absolute con- centration of reactive substance, sensitivi- ties may be reported in terms of the relative sensitivity compared with phosphorimolyb- dic acid. This is also important in classifying alkaloids by precipitation. The simpler oxygen acids used as alkaloi- dal precipitants are perchloric, chromic, permanganic, etc., used either as free acid or alkali salt. Permanganate especially fre- quently reacts as an oxidizing reagent and is itself reduced, but it gives some excellent crystals with compounds not readily oxi- dized. Unlike the complex oxygen acids these compounds are not very general or sensitive as reagents, and so are chiefly use- ful with fairly complex bases which are easily precipitated. Chlorochromic acid, HCrOsCl, is more sensitive and general than CrOs , 17 CHEMICAL MICROSCOPY but still gives similar precipitation and crj^s- tals. Anionic Complexes of Central INIetals. These reagents are commonly called "gold chloride", "platinum chloride", etc., and their precipitates are called "double salts", but actually the metals are in anionic com- plexes and the precipitating agents are really chlorauric acid, HAuCU , chloroplatinic acid, H2PtCl6 , etc. Thus the reagents are given by those metals a7id anions which will form anionic complexes of this particular type, and the group might be subdivided either according to the metals or the simple anions concerned. The metals are central to the long periods of the periodic table; the anions are halogens (Cl~, Br~, I~) and "pseudohalogens", such as CN~, SCN~, NO2-, N3-, etc. The iodide reagents of bismuth and plati- num are colored and so sensitive and general in aqueous solution that they are commonly used for spraying to bring out the spots of alkaloids or related compounds in paper chromatography . The following table gives the reagents of greatest value for microcrystals, designated as R; other definite reagents within the scope of this table are indicated by a small r. Chloride Bromide Iodide Cyanide Gold (+3) R R — R Platinum (+4) R R R R Palladium (+2) R r Mercury (+2) R r R r Cadmium (+2) r r R Bismuth (+3) r r R A "gold iodide" reagent is in use, made from HI and chlorauric acid, but the HAUI4 immediately decomposes, and the chief effects seem to be due to iodine-HI, rather than any gold compound, with a few results due to an aurous iodide complex, a gold (-f-1) reagent. The following might also be mentioned: chlorides (chloro-acids) of Fe, Zn, Sn (stan- nous and stannic), UO2 (uranyl), and VO (vanadyl), which are most effective with a high content of HCl to form the chloro-acid ; iodides of Pb, Sn (+2), Zn, and Ag, with alkali iodide; the complex cyanides of Fe, both ferro- and ferri-, and also nitroprusside; and complex thiocyanates of Ft, Hg, Sn (-f2), Cd, Zn, Co (4-2), Ni (+2), Mn (+2), and Cr (-F3). Most precipitates with thiocyanate reagents do not crystallize very readily. A remarkable exception is reinecke salt, NH4Cr(NH3)2(SCN)4 , which is very general and yet gives many crystals. Double complexes are also possible; e.g., mercuric cyani-iodide or mercuric chloro- iodide, made by dissolving the insoluble Hgl2 in cyanid-? solution or in strong HCl, respectively. On the whole this group is the most useful of all fo^^ niicrocr^^stal tests, and by using other strong acid; besides HCl the use of these reagents may be extended over all compounds of basic nitrogen. Simple Halides. These are useful with relatively complex bases, particularly iodide and thiocyanate, used as alkali salts. Organic Reagents. The best reagent of this group is picric acid, and several others are highly nitrated compounds, e.g., styph- nic and trinitrobenzoic acids. A few other kinds are known, e.g., sodium alizarin sul- fonate. Sodium tetraphenyl-boron is a new type of reagent (in some ways alhed to the complex oxygen acids), recently introduced, which is very useful for crystalline precipi- tates with free amines in volatility tests (without acidification), and quite sensitive to the lower amines in this way. Reagents in Strong Acids. The study of strong acids as media for the precipitation of nitrogen-bases began with hydrochloric acid, which, because many of the precipitat- ing agents are halides, often has special effects. Some reagents were already known which require a high content of HCl either to prevent precipitation of basic chloride (e.g., SbCls) or to form the chloro-acid 18 ALKALOIDS AND ALKALOIDAL-TYPE PRECIPITATION (e.g., FeCls). These were more or less assimi- soluble in H3PO4 tliaii in water, and far lated to the ordinary aqueous reagents; but more soluble in acetic acid than in water, some 20 or 25 % of concentrated HCl in the The reagents with acetic acid are therefore reagent will profoundly affect the precipi- less general than others, and find their espe- tation and crystal-forming 'properties of cial application with highly complex com- HAuCls , for example. Next, reagents made pounds, very easy to precipitate, particularly with concentrated HCl were tried; then in- those that yield only amorphous precipitates stead of using only aqueous solutions of the in the conventional tests. Instead of having substance tested, concentrated HCl was also to study new precipitating agents which are used for this purpose The value of other less and less general in effect, we can simply acids was later discovered. It became evi- try using acetic acid in the test-drop, as a dent that instead of dissolving the substance medium for new, useful tests. tested in quite a number of different acids, The effect of phosphoric acid in increasing it would be simpler to apply the reagents, the range of the tests is even more impor- in the various acids, directly to a Httle tant. There are in fact three effects. One is of the solid substance. This is now the the effect of a nonvolatile liquid medium in method used, but precipitati^on from solution allowing a test to stand as long as may be could be used. The precipitating agents are desired for crystallization to occur. H3PO4 the same ones that have long been used in has no very strong tendency to absorb water, aqueous solution to precipitate alkaloids. nor, when it is used already a little diluted, The acids proved to have very different to dry out. It has no side effects correspond- effects. The ones chiefly used are diluted or ing to sulfonation, or to the withdrawal of syrupy (85-88 %) H3PO4 , diluted H2SO4 , water from the molecule of a dissolved sub- concentrated HCl, and (2 + 1) acetic acid, stance, both often given by H2SO4 even when Concentrated H2SO4 would react with too it is somewhat diluted. Another effect is many of the substances to be tested, and that of any strong mineral acid in suppress- would decompose nearly all precipitating ing acidic qualities of amphoteric substances agents, but may be used up to (1 + 1) and enhancing their basic qualities Third strength with chloride reagents, as strong as and especially important is the particular (2 + 3) with bromides, and up to (1 + 3) effect of phosphoric acid in increasing the with iodides. Acetic acid may be used up to insolubility of precipitates in it, as already the glacial strength except that it then tends mentioned. to creep and spread all over the slide, and The most useful of all the crystal-produc- for this reason (2 + 1) is usually the maxi- i"g compounds are the chloro- and bromo- mum strength employed. ^^^^^ «^ ^'^^^ ^^^.^ platmum, the iodide re- ^-r -J. ■ r J ^T_ i XT- • -^ i- agents of platinum and bismuth, and Now it IS found that the precipitation- . ^,. ^,^ ^ . ,. ^^^ , ,' , „» , . , ,,!•«. .- 1-, . 1 TT o/-i lodme-KI or lodme-Hl. Among other ad- eft ect is not greatly different in diluted H 2^04 , ,, 11 , 1 J ^1 , ^^_, ^ .... vantages these are all colored, and crvstals or m concentrated HCl from that in plain ^ ji.i u j-ij-^- "^-uj ^ formed by them can be readily distinguished water, only a little greater m diluted H2SO4 ^^.^^^ ^^.^.^^^^ crystalline material, or from and a little less m concentrated HCl; and ^lystals, e.g., phosphates, formed simply by these acids are chiefly used, as media, to ob- ^^le acid used. They are compatible, with tain certain crystallization effects not ob- proper formulas, with all the acids men- tainable in plain water. The most surprising tioned, and extend the use of microcrystal effects are obtained with phosphoric and tests for identification to all compounds of acetic acids. The compounds formed by the basic nitrogen. Phosphoric acid reagents in precipitants with bases are immensely less particular are used for relatively simple, and 19 CHEMICAL MICROSCOPY feebl}^ basic and partly acidic compounds, and those of high solubiHty in water. Inorganic Precipitation with Re- agents for Basic Nitrogen. Potassium and ammonium have similar solubilities for their salts, and so give similar precipitation reac- tions; and, in a more general view, the al- kaloids and other compounds of basic ni- trogen react much like the heavier alkali metals. Cesium (ion) is precipitated by most of the "alkaloidal" reagents and rubidium > \ N^sr-^ n Fig. 7. Sodium bromaurate crystals. Na2S0 (solid) with HAuBr4 in H3PO4 . Fig. 8. Magnesium bromauraurate crystals MgCl.-eHsO (solid) with HAuBr4 in H3PO4. (Pho- tographed with red-sensitive plate.) by perhaps a third of them, from aqueous solution. The most general of such reagents often precipitate ammonium and potassium; and conversely, a reagent used to precipitate potassium is worth trying as a general pre- cipitant of nitrogenous bases. Sodium ion is not precipitated from aque- ous solution by the "alkaloidal" reagents, but in a medium of syrupy phosphoric acid the relationship of the lighter elements be- comes apparent. Not only sodium, (Fig. 7) but even lithium, magnesium (Fig. 8), zinc, and cadmium, become more or less subject to such precipitation. Even the hydrogen compound verges on insolubility, and for example bromauric acid itself, HAuBr4 , will precipitate from H3PO4 solution with drying. These inorganic reactions have been studied very little, except the formation of bromau- rates. Charles C. Fulton CHEMICAL MICROCRYSTAL IDENTIFICATIONS The microscope should be used throughout chemistry; it is surely an obvious instrument (and not a new one) simply for taking a closer and better look at small things. Chem- ists should turn to it as naturally as to test- tubes or the bunsen burner, whenever it will help. It has applications requiring knowledge and study, too, particularly the polarizing microscope, which is by far the most valu- able. Its especial use, discussed here, is in a basic branch of analytical chemistry— the making of identifications. This is a subsid- iary science in its own right, not well ex- plained by the vague general term "qualita- tive analysis". It is, of course, a part of qualitative analysis, as distinguished from quantitative, but it is a special part, in which we ought to use properties, tests, and reactions especially characteristic, or even specific for individual substances. Physical properties, although obtained by some gen- eral method, are often used when they can be accurately measured in such a way as to 20 CHEMICAL MICROCRYSTAL IDENTIFICATIONS distinguish an individual compound even The science of making chemical identifica- from those others that are closely related, tions is important in forensic chemistry (cf. Likewise, a chemical reaction may be general Forensic Microscopy, p. 338), especially in law for a whole class of compounds, and yet we enforcement, because many legal cases in- may be able to distinguish particular results, volve questions of identity rather than For example, a color reaction of phenols may quantity, as regards narcotics, poisons, po- give different colors with different phenols, tent drugs, adulterants, contaminants, sub- The value of the microscope is to see dif- stitutes, and so on, and such cases require ferent results given by different substances exacting proof of the identity of the sub- even in the same general precipitation reac- stances concerned. The uses are more gen- tion. eral, however, for questions of chemical Chemical tests are mainly divided into identity which can be answered in similar two kinds — color tests and precipitation ways may arise in any field involving chemi- tests. Now the ordinary spectrophotometer cal substances. Usually too little attention (cf. Absorption Spectroscopy — ^Visible and is paid to the possibilities of microscopic Ultraviolet, in "Encyclopedia of Spectros- chemistry, which may provide tests for very copy") may be regarded as giving not only a minute quantities, or such firm assurance of means of studying what might be called the identity as to be the preferred method, when "colorimetric" properties of a substance it- it can be used by anyone acquainted with self, but also, when a reagent is used, an it, regardless of the amount of substance enormous extension of the color tests of available for tests. chemistry. Not only are comparisons and Microcrystal tests are exceptionally good colorimetric readings made better, in the for court purposes, because they are as sim- visible range, but also they are extended pie and direct as tests can be. In many cases, deep into the ultraviolet, where nothing just looking at the crystals under the micro- at all could be seen with the unaided eye. scope enables the analyst to make the identi- In a somewhat analogous way, the micro- fication, and the things looked for are the scope enormously extends not merely the characteristics of an actual, visible com- study of the form and crystal properties of pound of the substance to be identified. It a substance itself, but also the ordinary seems to be pretty generally realized that precipitation tests of chemistry, without simple, direct tests are the best for forensic changing their essential character as chemi- purposes. What sometimes seems to be over- cal reactions. Not only are details of the looked or not realized as it should be, is that precipitate, when it is crystalline, seen bet- they are also best for chemical purposes, ter, and down to a smaller size, but with the In these microcrystal tests the polarizing polarizing microscope the observation of microscope should by all means be used, and crystal characteristics is extended to things the relation to optical crystallography is, of of an entirely different nature from those course, close. Now optical crystallography, that can be seen with the unaided eye by for which, in the field of mineralogy, the ordinary light. Thus microscopic identifica- polarizing microscope was developed, is of- tions by crystal forms and properties are at ten an ideal means of identifjdng the crystal least potentially capable of extension from species, for any substance that is already substances which are already crystalline to present in microscopically sizable crystals all substances that give precipitation reac- or fragments of crystals, and in fair pro- tions in ordinary chemistry, or indeed to portion in the material examined; and de- those that give any kind of crystal-forming termination of refractive indices plays a reaction. large part. On the other hand, microscopic 21 CHEMICAL iMICROSCOPY chemical identifications, which are the sub- ject here, usually depend upon the forma- tion of crystals by the action of a chemical reagent, which may even pick out the sub- stance in a complex mixture for the analyst. Refractive indices are very seldom used, but optical-cry stallographic properties which can be observed directly, without separating the crystals from the solution in which they are formed, are used, and therefore the polarizing microscope is needed. The two procedures go well together, but they are distinct. Figure 1 illustrates crystalline de- posits examined between crossed nicols; the compound is aminotriazole. Figure lb repre- sents only 0.4 microgram in the deposit. Other types of crystallization of a sub- stance without a reagent can be seen in the photographs of r/-, (//-, and (d + dl)-airi- phetamine hydrochloride and of NH4CI (see pages ()7-69). It should be noted that optical crystallog- raphy for substances already crystalline focuses upon the exact crystal species pres- ent, determined by such things as the acid with which a basic substance is combined, and even the proportion of molecules of water of crystallization, and sometimes the particular one of two or several polymorphic forms of the same substance, for these things often make major differences in the refrac- tive indices and all the other crystal proper- FiG. 1. Aminotriazole as seen with the polarizing microscope, between crossed nicols: (a) deposit from evaporated aqueous solution; (b) deposit of 0.4 microgram. 22 CHEMICAL MICKOCRYSTAL IDENTIFICATIOxNS ties. In the microscopic chemical tests, on the other hand, one can test directly for a basic drug (e.g., morphine, or amphetamine) without necessarily being concerned with whether it is originally present as the sulfate or hydrochloride or in some other combina- tion, and similarly for other kinds of sub- stances. Indeed, when a substance must be sepa- rated from a complex mixture, whether by extraction, chromatography, or some other means, its original state of combination is lost, and to use optical crystallography, or for that matter, to get a melting-point or x-ray diffraction pattern, it must be obtained in some kind of crystalline form. The great- est sensitivity, simplicity, and directness for any test requiring crystals exist if the crys- tals are obtained as some very insoluble com- pound which still crj^stallizes quite readily and in forms that are recognizable by direct inspection and observation. This is precisely the idea of the chemical microcrystal tests. This subject is, therefore, a branch of chemistry. It involves the use of the micro- scope in making identifications by means of chemical reactions, especially precipitation reactions, yielding crystals. The optical crystallography used is limited and usually need not be of a specialized kind. In fact the chemistry used is not very unusual either, but naturally the reagents and reactions are selected as those especially useful in con- junction with the microscope. The science is to some extent a blend of chemistry and microscopy, but analysts should particularly keep in mind the chemical meaning of the tests, and development of the field will pro- ceed along chemical lines. It is strange indeed that it has been so generally neglected by chemists, over a period of a hundred years. Advantages and Disadvantages. The outstanding advantages of this kind of iden- tification tests are: (1) The high assurance with which micro- crystal identifications can be made. (2) The direct character of most of these tests. (3) Their simplicity, convenience, and speed. (4) Selectivity, in the sense of noninter- ference, or no great interference, by most impurities. (5) High sensitivity, not necessarily in degree of dilution, but especially, with the aid of the microscope, in respect to amount of the substance for which the test is made. In addition to various objections that are often made but have no very substantial ba^is, the main disadvantages are: (1) The results are not readily classifiable. (2) Applicability of coverage limited in organic chemistry. (3) The general neglect. When an identification is wanted for a chemical substance the identification should be specific. However, it is seldom reached by a single specific test, but usually as the only conclusion possible from two or more — com- monly several — group tests or characteristic tests and properties. Often most of the chemical tests used are rather general, and are then supplemented with some physical measurement, often made on a derivative, to show differences between closely related compounds within the ascertained group, or provide final confirmation. The combined characteristics then become specific. Much misunderstanding of the microcrys- tal tests seems to result from a failure to appreciate their chemical character. In a chemical microcrystal test we know, fij'st of all, what reagent we used; and we know, or ought to know, what its chemical effects may be. We take a close look through the micro- scope at the crystals formed. Often we can recognize them immediately by their visible characteristics, assuming we have seen them before, and always remembering that we know what reagent was used. The results are far more definite for indi- vidual compounds than in most chemical tests, l)ut the procedure of identification is 23 CHEMICAL MICROSCOPY essentially the same, and these tests can well be used along with others, especially color tests for the spot-plate or designed for mi- nute residues, or "spot tests" as developed by Feigl. However, the microcrystal tests are usually so highly characteristic, that two or three of them, even just using the ordinary microscope, and making comparisons, as necessary, with a known sample, will often make an identification certain, without nec- essarily having or obtaining any other type of information. With the polarizing microscope the results are especially definite since a number of fur- ther observations can be made, even without removing the crystals from the solution in which they form. They do not have to be separated, washed, dried, recrystallized, and so forth, as for melting-point determinations and most other types of further tests, includ- ing refractive indices. All of the following are simply a matter of direct observation: not only the shape of the crystals, their grouping, color, size, and so on, but also bire- fringence, whether high, medium, low, or absent; whether extinction is inclined, and if so, the angle of extinction; in the case of colored crystals whether dichroism is pres- ent, and if so, the two different colors shown, and if the crystals are regularly elongated, the direction of dichroism (sign of absorp- tion), and in the case of colorless elongate crystals, the sign of elongation. All these things can be observed without having to treat or prepare the crystals in any way, once they form in a solution. Microcrystal tests are especially useful for distinguishing among closely related reactive compounds. Usually suitable tests can be found even to distinguish an isomer from the corresponding racemate, by completely different crystals. In fact, such tests have been used to identify (/-amphetamine even in the presence of d/-amphet amine, and vice versa. E. G. C. Clarke has shown how micro- crystal tests can sometimes be used to dis- tinguish microgram amounts of a d-isomer from the Z-isomer, and without using an op- tically active reagent, which is another possi- bility. The question of identifying a d- or Z-isomer is important in some cases now be- cause the ^isomer of certain new synthetics is restricted as a narcotic while the d-isomer is also offered commercially as a different drug not subject to narcotic restrictions. The distinction is based on the fact that the race- mate will give crystals in certain cases where a separate isomer will not. Besides knowing and having a suitable reagent, the analyst needs one known isomer, let us say the d-. If he mixes some of this with the suspected substance and it is 1-, then he has the race- mate and can get the appropriate crystals, but if the suspected substance is d- and he mixes d- with it, then of course he still can- not get crystals due to the racemate com- pound. If he has both known isomers for a cross-check, this kind of test can be com- pletely certain. The question about microcrystal tests so often asked, "what will they do on mix- tures?"— reveals fundamental misunder- standing. Most identification tests require that a substance to be identified be fairly pure. This is perhaps especially true of most physical and physicochemical tests, because they are so very general — melting-point tests, for example. A strictly chemical test or reaction, on the other hand, is bound to be selective in some degree, and will pick out a substance in the presence of all kinds of impurities ex- cept those that react with the same reagent, and occasionally even in the presence of compounds that react, but in a different way. A microcrystal test, where not merely the fact of a reaction, but crystals of a particular kind are looked for, cannot in general be expected to work on a complex of closely related, reactive compounds, although some tests that are remarkably searching, in this sense, are known. The application is often 24 CHEMICAL MICKOCRYSTAL IDENTIFICATIONS to a residue that has been separated as satis- may be put at about .000-4 microgram of factorily as possible from a mixture. atropine. "Selective" is a term used in two different In general, however, the writer agrees with senses: a reagent may demonstrate that a Chamot and Mason: "It is perhaps unfor- compound belongs to a particular group, or tunate that so much stress has been laid it may actually pick out a reactive com- upon the sensitivity of microscopical crystal pound among others that are unreactive tests, both because their advantages of ease or that do not react as strongly or in the same and directness are thereby obscured, and way. Thus one kind of identity test may because the impression is given that they require an absolutely pure substance and be are appropriate only in case extremely mi- specific, or characteristic, or merely selective nute amounts of material or low concentra- (in the first sense); while another type of test tions are to be dealt with." Sensitivities are may not require great pmity, and therefore important — -sometimes very important, even when specific or characteristic for a com- absolutely essential in such a science as pound may also be quite selective in the sec- toxicology — -but it does seem that when a ond sense, as well. The latter is very true of microcrystal test is described in any detail, the chemical microciystal tests. sensitivity is likely to be the one point These tests will work on mixtures with seized on, in abstracts for example, to the nonreactive material, with little or no inter- exclusion of everything else. This degrades ference, and can generally be applied directly the microscopical crystal tests to the level to medicinal tablets, for example, in which of mere detection tests, which are often sen- a drug is mixed or diluted with material that sitive but very general, and worthless for is inert in the chemical reaction concerned, identification whereas the purpose of the The reagent itself seeks out the particles or microcrystal tests is identification, whatever molecules of the substance that interests us, the amount of substance available. Their ad- among particles and molecules of all sorts vantages are not only in ease and directness, of other (nonreactive) substances. In the as mentioned, but in providing highly char- same way, these tests may be just the kind acteristic and even specific identification needed for extracts or isolates when it is not tests, not mere detection. To the extent that possible to get extremely high purity without sensitivity statements obscure this, they risk of losing the little bit of material that should be played down; at the same time it is all that is available. should be realized that the better micro- When procedures on a very small scale are crystal tests are exceedingly sensitive, and combined with the use of the microscope, as thus quite adequate for identification on al- is done by Clarke, a quite common sensitiv- most any minute scale, ity for a highly characteristic crystal test The statement is rather frequently en- will be 2 or 3 hundredths of a microgram, countered, that in general color tests are Atropine gives tiny characteristic crystals more sensitive than precipitation (or micro- with the right variety of iodine-KI reagent crystal) tests. This is both incorrect and down to an aqueous dilution of 1 to 1,000,- fallacious. Some compounds give excellent 000. Using a fair-sized ordinary drop of 0.04 color tests and few and poor precipitation ml the crystals can therefore be obtained tests, or none; others are just the reverse, with 0.04 microgram; by using a microdrop Some compounds, e.g., morphine, are re- of 0.1 microliter (.0001 ml), the extreme markable for both kinds, and in such cases sensitivity is .0001 microgram; or, as the they are hkely to be of the same order of sen- crystals are hard to find at the extreme limit, sitii-ity, as a consultation of published state- the working sensitivity, using microdrops, ments in regard to morphine tests will show. 25 CHEMICAL MICROSCOPY The sensitivities can be great for both, but oped. They are destructive aUke to the speed, they are usually due to different reactive simplicity, ease, and directness of the micro- radicals or atoms, sometimes both kinds crystal tests, in most cases without any con- present in the same molecule, as with mor- current gain, since the observation of char- phine, and sometimes only the one or the act eristic crystals was already adequate to other present for any known and useful tests, the pmpose of identification. Much of the It would be almost as sensible for a chemist literature on the use of picric, styphnic, and to cut off one hand, as to refuse to use one or picrolonic acids, reinecke salt, etc., is of this the other kind, according to their merits in type, particular cases. A related erroneous idea sometimes en- The preeminent radical for color tests is countered is that special measures should be probably the phenolic hydroxyl. Even more taken to get the crystals as well-formed crys- preeminent for precipitation and micro- tallographically as possible — whereas in fact, crystal tests is the basic or partly basic ni- crystals ordinarily forming in characteristic trogen atom, no matter whether it is "pri- distortions have far greater value. In most mary", "secondary", or "tertiary", or cases even procedures such as warming or whether it belongs to a straight chain of adding alcohol are of value only when a atoms, or is attached to a benzene ring, or is particular compound is suspected that needs itself part of a ring. The chief count eractant special measures to induce crystallization to the precipitating effect of basic nitrogen with a much-used reagent, and then a special is acidic oxygen, which shows why color reagent is usually better if one can be found tests may occasionally "oppose" crystal that will readily give crystals withovit such tests, i.e., the change of a molecule to in- measures. elude the preeminent radical for color tests Crystals given by a particular organic may weaken and lessen the sensitivity of compound may be specific, that is, of a kind precipitation tests, especially for relatively given by no other substance. Sometime ob- simple compounds. Paradoxically, however, jection is made to calling such tests specific, oxygen itself may be basic, in certain cases, but on grounds that would prevent any tests and then forms oxonium salts, which are sub- ever being called specific: that not every ject to precipitation tests. possible substance has been tried, or even While an identification may well be based that some new compound might be discov- on several results obtained separately by ered or synthesized sometime, that might be different disciplines, it is unfortunate that confused with the old one. To the writer it numerous researchers consider the form and seems more reasonable to call a test specific obvious characteristics of crystals only as if it is specific, so far as we know or can rea- dressing for procedures resulting in micro- sonably foresee, for our knowledge never can melting points, refractive indices, or other be infinite. numerical determinations. That is, they On the other hand, we should exercise make the tests over in the usual pattern of considerable caution in designating a test as organic confirmatory tests: formation of a specific. The crystals are, in general, direct derivative, isolation of the derivative and compounds of the substances for which the often its further purification (e.g., by recrys- tests are made, and therefore in a particular tallization), and physical measurement of case they are more or less different from some characteristic of the isolated, purified those given by any other substance, just as derivative. Such procedures have been pro- each substance is, in itself, finally unique, if posed even in the alkaloidal field, where the tests could be carried far enough to take in microcrystal tests are already best devel- all its physical and chemical properties. 26 CHEMICAL MICROCRYSTAL IDENTIFICATIONS However, the question is whether the crys- tals in a particular case are sufficiently differ- ent from most others, or from all others, to be distinguished from them by informed direct observation. The writer is satisfied to call most of these crystal tests, even most of the very good ones, characteristic rather than specific; not only when we know of some definite resemblance but also whenever the crystals are some general type that might occur with some other compound, even if we do not yet know what it would be. However, as an example of a specific test the intensely pleochroic crystals of hera- pathite, as a test for quinine, may be cited. These crystals of quinine iodosulfate (Fig. 2), produced by a suitable reagent, are rec- ognizable at once and the test is so sensitive to cjuinine, and so insensitive to interference, that the writer has used it as a direct test for quinine in the diluted, adulterated, impure heroin mixtures of the illicit drug traffic. Of course, one almost never depends wholly upon any single test by itself, even if it is specific and the intense fluorescence of qui- nine sulfate solution in ultraviolet light, which is characteristic but not specific, is also easily observed. Other cinchona alkaloids give highly char- iA *^% ./«■■ ^|i' ^'•i*'^ . ^-^v^ ■y^-^^ y Fig. 2. Crystals of quinine iodosulfate (hera- pathite) (with polarized light). lodine-KI reagent C-3 or a similar reagent. Fig. 3. Crystals produced by quinidine with lodine-KI reagent C-3 or a similar reagent. wn Fig. 4. Crystals produced b.y Cinchonidine with lodine-KI reagent C-3 or a similar reagent. acteristic, wholly different, iodosulfate crys- tals, with the same reagent used for quinine (Figs 3, 4, 5). A strange objection is often made to crys- tal tests as, supposedly, a reason for not us- ing them: that one substance with one re- agent may give two or more different kinds of crystals. This undoubtedly happens, but, in the first place, it also happens with all other tests that depend on crystal form and properties, including melting points. As a difficulty it is very unlikely to be too serious 27 CHEMICAL MICROSCOPY Fig. 5. Crystals produced by Cinchonine with lodine-KI reagent C-3 or a similar reagent. with microcrystal tests, because the crystals are not some that come to the analyst al- ready formed mider unknown conditions, but are formed under test conditions, and comparisons should be made with a known sample under as nearly as possible the same conditions. Secondly, even if one crystal type was un- expectedly changed into another which was unknown to the analyst, this would pre- sumably result in a failure to make the identification, but not in an erroneous identi- fication. There is probably no way of testing known that may not sometimes fail to give a recognizable result because of some kind of interference with the usual test conditions. On the other hand, if both kinds of crystals are known, even if the reason for one chang- ing into the other is not, the test is still good. Finally, the objection is especially strange because in most cases, having two or more types of test-crystals is a positive advantage. If we can learn the reason for the change — and this is usually not very hard to do, and is known in many cases — then we have two tests instead of one, which together often provide enough evidence for identification in themselves. Or, the two kinds of crystals may occur in the same test, in which case both types together characterize the sub- stance, and the single test is all the more likely to be specific or very highly character- istic. One crystal type is occasionally known to be changed to another by temperature or by stirring. However, the most common cause is a different concentration of the sub- stance, especially relative to the precipitat- ing agent in the test -drop. Particular advan- tage is taken of this in the new tests made by addition of a strongly acid reagent to a very little of the dry substance tested. A cover-glass is applied and this prevents the tested substance from diffusing equally through the test-drop. Thus the precipita- tion and crystallization take place at differ- ent concentrations in each single test even with only a minute amount of the substance. Therefore when a substance gives two or more completely different types of crystals at different concentrations, they are nearly always obtained in each single test, so that such tests are generally highly characteristic and may comparatively often be of specific rank. For example, in one such test with an iodine-KI reagent in HCI-H3PO4 , morphine gives four kinds of crystals (Fig. 6a), de- pending upon concentration: black needles, Fig. 6a. Crystals given by extracted Morphine with Iodine-KI reagent M-2. 28 CHEMICAL MICROCRYSTAL IDENTIFICATIONS brown threads in rosettes, tiny rectangular orange-brown blades, and comparatively large brown to red square-cut and jagged birefringent plates. All four -kinds can be obtained with only a few micrograms of morphine (in a small spot) ; the black needles are sensitive, with no special effort to reduce the scale, to 0.1 microgram. The test is ex- traordinarily resistant to interference, that is, the morphine need not have a high degree of purity to yield good crystals; in fact, the black needles and red plates have been ob- tained on a little dried poppy-juice (Papaver somniferum) without first separating the morphine from the other opium alkaloids or even purifying the alkaloids as such (Fig. 6b). That specific tests are available for some substances is only one factor in deciding what tests to use in given circumstances. Even a very general reaction may be useful to prove the absence of a compound. For general identification work, reagents will be preferred which give many different charac- teristic results, rather than "tailor-made" reagents for a few specific results, particu- larly in using them before one has much idea what is present. Other important factors are sensitivity, and ease of obtaining crystals even when the substance is not quite pure. Some specific tests rank high in these ways also, but they do not necessarily go to- gether. Regardless of other factors, there are always cases where what we want, when we can have it, is a test that, when successful, will prove the identity of the substance then and there, beyond doubt. This is possible with some microcrystal tests. The very diverse nature of microcrystals is at the same time the strength and weakness of the method. The results are highly char- acteristic or specific, but they are hard to classify. Largely, this is due to the fact that the tests are not measurements; the results are not even directly numerical at all, and so they lack the automatic classification that numbers give. It is not that the tests are Fig. 6b. Crystals given by Morphine in dried juice from Popaver somniferum — ^no purification of the morphine, not even any separation of the al- kaloids— with lodine-KI reagent M-2. good for only a few particular results: plenty of good results are obtainable. The difficulty is to have a way to look them up if they are not already familiar. Possible solutions are being sought, now that the problem is becoming acute. Even if anyone wanted still to limit the scope of the tests to some 50 familiar alkaloids, it is not possible even in the drug field, for there is now a host of new drugs that react in the same way (by precipitation and crystal- forming reactions) with the same traditional reagents. Several different solutions of the classification problem are possible in prin- ciple, at least such that punched-card sorting would quickly lead one from results on an unknown back to the proper "known" if it had been previously studied and classified. However, at present, and probably for some time, the analyst has to rely largely on chemical classification, involving method of isolation, color reactions, precipitation re- actions regarded chemically, and any other clues. Chemical evidence may reduce the number of compounds that have to be con- sidered to something manageable, and then microcrystal tests will show exactly what the substance is, if "knowns" for compari- 29 CHEMICAL MICROSCOPY sons are available. This goes well beyond about this state of affairs is that it should using the tests as merely "confirmatory" of be a challenge and a stimulus to an analyst something already known or suspected, but to promote application of the microscope to even that is one of their useful functions. any problems of identification that occur in The second real disadvantage listed above his own work, for these tests is the limited field of applica- It may be pointed out that these disad- bilitj^ or lack of coverage outside a limited vantages, serious as they are in some cases, field. It has already been pointed out that in no way interfere with the use of micro- the applicability is in fact not nearly so lim- crystal tests by an analyst acquainted with ited as generally supposed, but the disad- them, in: (a) confirmatory tests, where ten- vantage might be stated more accurately tative identification has been made, and that tests have as yet been well worked out comparison with a known sample is possible; only for inorganic ions and the familiar al- (b) in proving the presence of a particular kaloids. The same reagents and procedures compovmd (here, if the compound sought as for alkaloids can be used for a great many turns out not to be present, the test may do other drugs, e.g., antihistamines, and with more than most to show what is present); some changes they can be extended over all (c) in deciding, among a few alternatives, to derivatives of basic nitrogen, and even fur- which one an unknown corresponds: this is ther, but certainly other types of organic often possible by microscopic test-compari- precipitation have not been studied nearly sons with the known alternatives even if enough, and in fact with the many new drugs nothing has been previously published or the coverage is inadequate even for those studied on these particular compounds; (d) giving tests with the traditional reagents. in distinguishing between closely related Rather than any danger of making erro- compounds (including distinction of an neous identifications, as seems to be feared isomer from the racemate, or even of the by many (perhaps rightly in the case of the d- from the ^form, as previously explained); totally inexperienced), the bane of the ex- (e) in a larger and larger field of tests as the perienced analyst is the finding of an occa- experience of the analyst grows, sional compound, in the line of work, which yields beautiful crystals, sometimes with practical norK several different reagents, but which, never- As textbooks are nearly non-existent, theless, he still cannot identify. This is due some advice on practical work will be at- much more to lack of study of the com- tempted. In the limited space it still must pounds than to the difficulty of looking up be rather general. those already studied, because of lack of First, accustom oneself to the micro- classification. The individual chemist or a scope by using it. Much analytical work will single laboratory cannot possibly keep abreast be helped enormously by preliminary micro- of developments. scopic observation of the material to be Both of the disadvantages just discussed examined, either by reflected light (if it is are related to general neglect, which also opaque) or by transmitted light. This may, has other featm'es. The analyst must often for example, show at the outset whether the train himself; he usually cannot obtain material is a mixture or appears to be all one much instruction, or even textbooks, in this substance. field. At the present time the writer does Then, bring the chemical work down to not even know of a good textbook for micro- the level of convenience for microscopic ob- crystal tests for alkaloids that is now in servation. A precipitation test can be ob- print. The only good thing that can be said served just as well with a drop of solution 30 CHEMICAL MICKOCKYSTAL IDENTIFICATIONS on a slide as with several ml in a test-tube, before using the microscope at all. Color tests on the spot-plate are a useful adjunct. Accustom oneself to working on some such scale, which will probably be found more convenient than the "usual" one, anyway. Keep a polarizing microscope at hand, and use it. If a polarizing microscope cannot be used, polarizing attachments for the ordi- nary microscope are helpful, but are at best a poor makeshift. Observation of the micro- crystal tests is almost always by transmitted light, although it would be possible to use reflected light in many cases. "Low Power" of 80-100 X is usually sufficient for the tests, at least for all the initial observations. If the crystals are very small, a higher power is then used. Make a practice of looking at the result of a precipitation test microscopically, whether the textbook mentions crystals or not. For example, the iodoform test for ethyl alcohol is made quite definite for the product iodoform (although in any case it is not specific for ethanol) by microscopic observa- tion, although textbooks often fail to men- tion this and may only refer to the odor. Pictures that can be consulted are helpful in dealing with a test not yet familiar, and when time permits or the case is important enough the analyst should take them of his own results, as a reminder to himself and information to others. There is a real art in taking photomicro- graphs of known crystals so that they will most plainly show the truly characteristic forms that should be looked for. Pictures for reference, however, valuable as they are, are no substitute for at least a little actual ex- perience, and no substitute, either, for actual comparison of the results given by an un- known with those given by a known com- pound. Most of the familiar tests, for organic as well as inorganic substances, are made with aqueous solutions. Usually the ordinary flat microscope slide is quite satisfactory. Com- monly, a cover-glass is not used, unless or until examination by high power is wanted. The test is observed while evaporation takes place. This time can be prolonged by setting a petri-dish over the slide when it is not under immediate examination. Cavity slides are less convenient for focusing, but the evaporation is more gradual, and can be stopped by simply laying a slide over the one on which the test is made. Cavity slides are also more convenient for mixing three or four ordinary drops, in more complex tests. The reagent drop may be added directly to the drop tested, or placed beside it on a plain slide and the two drops allowed to flow together. Reagent is usually added in about the same size drop as the solution tested. The actual concentration of substance tested is therefore reduced by about half in the test-drop, but it is customary to report the sensitivity or optimum concentration, etc., in terms of the solution tested, before it is mixed with reagent. A full drop of water (let fall from a fairly wide opening) is about 0.05 ml. Use smaller rather than larger "drops", by touching the pipet to the slide and letting just a little flow out, by dropping from a narrow orifice, or by using a rod to transfer a drop. In most tests with no especial need to conserve ma- terial ordinary 1-ml pipet s may be used for handling both the solutions to be tested and the reagents. Not only crystals, but also amorphous precipitation, or absence of precipitation, are observed. Amorphous precipitates may be finely divided or curdy, or in drops, the latter often a prelude to the formation of large crystals. Crystals may form directly, or from an amorphous precipitate. As evaporation is usually not controlled, the time during which a test is observed is then only that required for the drop to dry up. Test-crystals often form around the edge, especially with dilute solutions, as the drop dries. The reagent it- self may crystallize out when evaporation has gone far enough, and a reagent-blank 31 CHEMICAL MICROSCOPY should always be run in comparison with any mm diameter) are used to make these micro- test, until the analyst is thoroughly familiar drops and add reagent (see Clarke), with crystals or any other effect that may To prevent too rapid evaporation a micro- arise from the reagent alone. test is made in a hanging drop. It may be Test solutions can be made up of weighed put on an ordinary slide and inverted over quantities in a few ml or less of water, or a cavity slide. This procedure is also used made right on the microscope slide. The for an ordinary small test-drop (say 0.01 ml), analyst soon learns to judge the amount of simply as a means of preventing or greatly substance needed, say about 0.2 mg if there slowing evaporation and the slide can be is no need to conserve the substance, giving reinverted whenever rapid evaporation is de- with an ordinary drop of about 0.04 ml a sired. As an alternative to sealing-in, to be concentration of about 1:200, which is very quite sure of preventing evaporation while good for many microcrystal tests, usually the test stands, a drop of water, or better, better than more concentrated. Most of the a large "blank" test-drop, which will give tests are best at concentrations between virtually the same humidity as the hanging 1 : 100 and 1 : 5000 (0.4 mg to 8 7 in 0.04 ml) ; drop, can be put in the depression of the the very good ones still succeed at greater cavity slide. The enclosure technique also dilutions, requiring only a few micrograms slows the evaporation of other substances even in an ordinary drop, and the tendency besides water, e.g., iodine from solutions is always to study and use more and more which have only a low content of iodide to sensitive tests. hold it in solution. The average analyst needs micro tests far With a standard "thin" slide (thickness more than "micro" procedures and equip- not over 1 mm) the usual second power ment, aside from the microscope itself. It is magnification (objective 20 or 21 X, or 8 often more convenient to carry out an ex- mm) can be used through the slide. If ob- traction, for example, with ordinary equip- servation under still higher power is wanted, ment, and volumes of, say, 10 ml of solution of course it is necessary to mix the test-drop and 5 or 10 ml of solvent at a time, for a few on a cover-glass. In Clarke's technique the shake-outs, even if the amount of material cover-glass is sealed in with 25 % gum-arabic is quite small and the final residue micro- solution to prevent its slipping and ensure scopic. Often, in fact, the analyst has to only extremely slow evaporation. The seal process quite large amounts of material to may be made not quite complete at fii'st if get an extremely minute residue. Chemistry it is desired to allow a little evaporation has reached a stage where more and more (until precipitation occurs) before finally sensitivity is wanted in tests, and they are sealing. Square 25 mm cover glasses may be almost never too sensitive if satisfactory in most convenient, other respects. In new tests for basic substances in a me- A further long step downward can be dium of strong acid, a very small amount of taken when necessary, by using small drop- solid material is put on a slide (finely pow- lets. If all the substance available will make dered or in a thin deposit in one spot), a only one small "macro" droplet of solution drop of reagent placed near it, and then a of, say, 0.01 ml, it will still be possible to cover-glass, ordinarily of 18-mm size (round try up to about 100 microcrystal tests, using or square), is added so that the reagent microdrops of as little as 0.1 (mm)^ or 0.1 flows over the solid. Or, the drop of reagent microliter, each containing only 1 to 0.02 is put on the cover-glass, which is then in- microgram of substance, to be within the verted over the substance. The reagent con- best range for most tests. Thin solid rods (1 sists of some precipitating compound in solu- 32 CHEMICAL MICROCRYSTAL IDENTIFICATIONS tion in a strong acid, most often syrupy H3PO4 , (2 + 3) H2SO4 , concentrated HCl, or (2 + 1) acetic acid — acids which give re- agents of very different effects even with the same precipitating compound. With a medium of strong phosphoric or sulfuric acid the drop cannot dry up, but its moisture content may gradually change on standing (generally increasing, sometimes considerably), depending on the humidity or dryness of the air. These tests are usually observed during only about two hours, so that ordinarily no special precautions are needed. Small cover-glasses of 12-mm diameter are obtainable. A test made under this size can be inverted over a cavity slide, a procedm'e which will very nearly prevent any evapora- tion of a volatile acid or water, or changes due to humidity or dryness of the air. This is generally sufficient control for at least 24 hours without sealing-in. A further refine- ment is to put a drop of diluted sulfuric acid of a certain strength in the cavity of the lower slide, to regulate the humidity. H2SO4 (45 -j- 55) will correspond pretty closely to a reagent of syrupy H3PO4 ; H2SO4 (1 -f 5) will provide controlled humidity; H2SO4 (55 + 45) provides gentle drying for H3PO4 reagents. Reagents in strong acids may also be added to aqueous solutions, or applied to solids in a form somewhat diluted with wa- ter, without a cover-glass. Then formation of a precipitate and crystallization may occur promptly, or only as the drop evaporates to a higher acid strength, particularly in us- ing phosphoric acid. Sometimes drying in a desiccator may be useful, e.g., with HAuCU in H3PO4 , or a controlled humidity to allow drying just to a certain degree, or to prevent complete drying when the air is very dry, e.g., with the reagent HsBile in (1 -f 7) H2SO4 . A hanging drop of reagent is also used in tests for volatile substances. This is more or less familiar in chemical and microcrystal tests for ammonia from ammonium salts, and as a test for urea when it is decomposed to ammonia by urease. Microcrystals tests in the hanging drop for the readily volatile d- and c?/-amphet amine were perhaps first used by Griebel, with aqueous reagents. The ma- terial (e.g., a small amount of scrapings from a tablet containing an amphetamine salt), or its solution, is treated with dilute alkali on a cavity slide, and a hanging drop is in- verted over it. There are now at least a dozen such sympathomimetic drugs regu- larly on the market, that are readily volatile, to be tested in this way, and others that are less readily volatile. It is certainly not yet realized how far such tests can be extended, how many sub- stances usually considered nonvolatile are in fact slightly volatile with water vapor even at room temperature and can be detected in a hanging drop with an exposure of, say, two hours, for example, I- and c?/-ephedrine, and phenmetrazine. Gentle heat from above, e.g., from a desk lamp, may be used to shorten the time required, and may improve results, e.g., in the vaporization of pentylenetetrazol, but a longer exposure at room temperature is as good or better for most compounds. On the acid side benzoic and salicylic acids may be mentioned as sufficiently vola- tile with water vapor for vaporization tests. A hanging drop of an aqueous solution of reagent, above an aqueous solution on a cavity slide, does not dry up or expand un- less considerable hygroscopic material is present in one solution, and a long exposiu'e may be used. For basic vapors, a reagent in diluted H3PO4 , (1 + 4) to (1 + 2), may be exposed for some two hours or more, with some increase in the size of the hanging drop due to absorption of water. The slide with the hanging drop is then reinverted to allow evaporation, even put in the desiccator if necessary, and tests can be obtained for a great many basic substances even if only slightly volatile. A convenient purification as well as a test 33 CHEMICAL MICROSCOPY for a volatile base is to catch the vapor in a substances, many such tests undoubledly hang;ing drop of dilute HCl (say 1 % by will be found. However, the most familiar volume of the concentrated acid), over an microcrystal tests at present are those for alkaline drop on a cavity slide. Then the alkaloids and related substances, and related slide with the hanging drop is reinverted and tests for other derivatives of basic nitrogen, the drop allowed to dry up, leaving the hy- drochloride of the volatile base. This solid Criteria for Selecting the Best Identi- hydrochloride may then be examined micro- iicatioii lests scopically, if it crystallizes, or it may be (1) Highly characteristic crystals (which tested with HAuBr4 in H3PO4 , or some other may be so by reason of two or more types reagent of high acid content; or with this occurring in the same test -drop, or for other purification it may be simply dissolved in reasons besides form); (2) Easily recogniz- water for some test. able crystals; (3) Crystals forming very Some substances decompose in alkaline readily; (4) Crystals of the same type or solution and give off a volatile base. For types over a wide range of concentration or example, the drugs levarterenol, epinephrine, with widely different amounts of material and isoproterenol, all diphenols and of simi- (when direct addition is used); (5) Crystals lar uses, give off ammonia, methylamine, reasonably stable; (6) Crystals not essen- and isoproterenol give off ammonia, methyl- tially changed by the presence of nonreactive amine, and isopropylamine, respectively, and material (e.g., tablet excipients); (7) Crys- can be readily distinguished in this way, with tallization not prevented nor badly altered a hanging drop of sodium tetraphenylboron when impurities of other nitrogenous bases solution. The assm'ance of the identity of are present in quite minor amounts (ex- the substance given off that is afforded by perience with actual uses of the test affords microcrystals raises these tests much above the best information on this point); (8) High the usual chemical decomposition tests. sensitivity; (9) Colored crystals are prefer- Still, a decomposition test is usually not able (i.e., use of a colored precipitating com- the best, if a direct test is available. In show- pound); (10) The reagent by itself (blank ing contamination by rodent urine, instead test) should not give any possibly confusing of the indirect test for urea by decomposition crystals; (11) A permanent reagent is prefer- with urease and detection of the ammonia able; (12) A slow-evaporating or non-drying evolved (a ubiquitous substance, and usu- reagent is preferable; (13) A reagent of gen- ally detected merely by its alkalinity, al- eral usefulness for microcrystals is preferable, though microcrystals can be used here, too) or alternatively (for some purposes) one that a direct microcrystal test with xanthydrol does not precipitate at all with most other can be used. In (2 -f 1) acetic acid this is alkaloids and amines; (14) Comparisons, amazingly prompt and sensitive for an or- both in general, and with the most closely ganic reaction, and results in crystals of a related compounds available, should estab- urea compound, dixanthylurea, which are lish the meaning and valvie of the test for observed directly, with obvious forensic and the particular substance. other advantages for the test. This is an example of the scattered microcrystal tests Future Lses which are possible and depend on varied It seems likel^M hat for the near future the organic reactions. usefulness of microcrystal tests will remain If analysts, generally, would try to find chiefly in the drug field, as regards the sub- microcrystal tests whenever this type would stances tested, and in general in regulatory be the most advantageous way of identifying and law-enforcement work. Here, the highly 34 CHEMICAL MICKOCRYSTAL IDENTIFICATIONS individual character of the tests is much more an advantage than a defect. Here, highly characteristic and specific tests, the results of which can be photographed for possible introduction in court or in other proceedings, are welcomed. Here, it is a great advantage that the tests are simple and di- rect, and distinctions made on the basis of obvious, visible characteristics. The identi- fications clearly distinguish between closely related compounds, even between an isomer and the racemate, and by certain tests, be- tween the d- and Z-isomers; and this is just what is needed in the drug field and in forensic chemistry and related work. In these applications the microcrystal tests are already used, and so too are color tests, of a kind with which the crystal tests form a natural partnership. The classification of the microcrystal tests, to enhance and extend their usefulness, re- mains a problem. Compounds are onty too likely to be met with occasionally, to which the tests are applicable but which still can- not be identified. This of course is due not merely to lack of classification, but is, at present, primarily because the tests for so many new drugs and other compounds have not been studied at all. However, this is a related aspect of the problem, for if a satis- factory system of classification is once worked out, the necessary experiments to find where all the compounds likely to be met with fit into the scheme are more likely to be made. WTiether a classification of crystal results should be primarily chemical, or morphologi- cal with the different reagents, or based simply on the fact of crystallization of each sub- stance with certain reagents out of a stand- ard list, there can be no reasonable doubt but that the science of microcrystal tests should be integrated with analytical chem- istry, not pushed aside as a "specialty" nor even regarded as being more microscopy than chemistry. The idea of using the microscope to make tests better for identification by simply look- ing at the different crystals formed in the usual precipitation reactions and other crys- tal-forming reactions of chemistry seems so sensible and also so obvious that it is hard to explain how it can be overlooked, as it usually seems to be. The idea develops nor- mally in three stages: (1) Microscopic observation of chemical tests: The microscope is used to distinguish the crystals occurring in chemical precipita- tion tests and other reactions, and thus to particularize them further; and the tests which are found especially characteristic are noted. (2) Microscopic study of microcrystal tests: The tests thus found of especial value be- come in themselves the basis for identifica- tion, and the best reagents thus found are tried systematically on other, related com- pounds. In this stage, appreciation of the chemical nature of the tests sometimes al- most disappears. (3) Chemical extension of microcrystal tests: Chemical reagents and reactions are reconsidered and studied with the objective of improving microcrystal tests and extend- ing them to new groups of compounds. As the tests come to cover a wider area in or- ganic chemistry than simply "alkaloids", their chemical meaning in various cases be- comes important. The chemist needs the microscope, not only for primarily observational use, as in looking at the material he is dealing with, and as a convenient aid in all kinds of chemi- cal procedures and research, for example, whenever he has a deposit or residue and merely wants to know whether it is crystal- line or not. . . . He needs the polarizing mi- croscope, not only as a means of identifying crystals of chemical substances by refractive indices and other properties shown by opti- cal crystallography. . . . Most of all, in his own capacity of chemist, he needs it, as this article has tried to show, in analytical iden- tification chemistry, as the prime means of 35 CHEMICAL MICKOSCOPY making chemical tests more definite. The science of microcrystal tests is ah-eady useful now, especially in the drug field as well as in inorganic chemistry, but it needs to be fully integrated with analytical chemistry, and extended over all organic compounds that can give characteristic, identifying crystals in chemical reactions. REFERENCES 1. Stephenson, Charles H., "Some Micro- chemical Tests for Alkaloids; including Chemical Tests of the Alkaloids Used," by C. E. Parker, Philadelphia and London, J. B. Lippincott Co., 1921. 2. Kley, p. D. C. (a) Behrens-Kley, "Mikro- chemische Analyse," 1st part, (as 4th ed. of "Anleitung zur Mikrochemischen Analyse," by H. Behrens), Leipzig, Leopold Voss, 1921. (Inorganic), (b) Behrens-Kley, "Or- ganische Mikrochemische Analyse" (as 2nd ed. of "Anleitung zur Mikrochemischen Analyse der Wichtigsten Organischen Ver- bindungen," by H. Behrens), Leipzig, Leo- pold Voss, 1922. 3. Amelink, F., "Schema zur mikrochemischen Identifikation von Alkaloiden; ubersetzt von Marga Laur." (In German). Amster- dam, N.V.D.B. Centen's Uitgevers Maat- schapij, 1934. 4. Geilman, W., "Bilder zur qualitativen Mikro- analyse anorganischer Stoffe," Leipzig, Leopold Voss, 1934. Republished by Author- ity of the Alien Property Custodian, by J. W. Edwards, lithoprinted by Edwards Brothers, Ann Arbor, Michigan, 1944. 5. Rosenthaler, L., "To.xikologische Mikro- chemie," Berlin, Gebriider Borntraeger, 1935. Republished by Authority of the Alien Property Custodian, by J. W. Edwards, lithoprinted by Edwards Brothers, Ann Arbor, Michigan, 1946. 6. Wagenaar, M., a series of articles on micro- chemical tests for particular alkaloids and related substances, (in Dutch), Pharma- ceutisch Weekblad voor Nederland, 64-67 (1927-30); 70 (1933); 72-73 (1935-36). 7. Herzog, Alois, "Mikroskopische Bilder fiir den Chemiker," Zeiss Nachrichten, 2 Folge, Heft 5, 149-81 (1938), and Heft 6, 183-215. (English summary at end of the volume.) 8. Whitmore, W. p., and Wood, C. A., "Chemi- cal Microscopy of Some Toxicologically Im- portant Alkaloids," Mikrochemie , 27, 249- 334 (1939); "Scheme for the Microchemical Separation of Some Toxicologically Impor- tant Alkaloids," Mikrochemie, 28, 1-13 (1940). (Both in English.) 9. Chamot, Emile Monnin, and Mason, Clyde Walter, "Handbook of Chemical Micros- copy, Volume II. Chemical Methods and Inorganic Qualitative Analysis," New York, John Wiley & Sons, 1940. 10. DucLOux, Enrique Herrero, "Notas Micro- quimicas sobre 'Doping'," Buenos Aires, Peuser Lda., 1943. 11. Fulton, Charles C. (1) Reagents: "The Pre- cipitating Agents for Alkaloids," Ain. J. Pharmacy (April, 1931); "New Precipitating Agents for Alkaloids and Amines," Am. J. Pharm., (Feb. & Apr., 1940); "The Relation of Alkaloidal Chemistry to Inorganic, and the Use of Bromauric Acid as a Reagent for Inorganic Microcrystal Tests," J. Am. Pharm. Assn., Scientific Edition, 31, 177 (1942). (2) Classification by aqueous pre- cipitation: "Alkaloids and Their Reagents," Am. J. Pharm. (May, 1939); for earlier, see /. Association Official Agricultural Chemists, 13, 491 (1930) . (3) Particular substances and development of new-type reagents: Atro- pine, ihid., 12, 312 (1929); Cocaine and Pro- caine, Ajn. J. Pharm. (July & Aug., 1933); Heroin (with G. D. Williams), A7n. J. Pharm. (Sept., 1933); Morphine, /. Lab. Clinical Medicine (March, 1938) ; Cinchona alkaloids, Ind. Eng. Chem., Anal. Edition, 13, 848 (1941) ; Morphine, Heroin, Dilaudid, Cocaine, in American Journal of Police Science, incorporated in Journal of Criminal Law and Criminology, 32, 358-65 (1941) ; Methadone, Narceine, Thebaine, in United Nations document (mimeograph) E/CN.7/ 117, 14 April 1948, (also includes photomi- crographs of the natural narcotine crystals of opium— see "Origin of Opium"), and Codeine in Add. 1 to this document, 22 Sept. 1948; Colchicine, Jour. AOAC, 41, 756 (1958). 12. Hartshorne, N. H., and Stuart, A., "Crys- tals and the Polarizing Microscope," London, Edward A. Arnold & Co., (1943), 2nd ed. 1950. (Optical Crystallography.) 13. Farmilo, Charles G., Levi, Leo; Oestrei- CHER, P. M. L.; AND Ross, R. G. "Micro- crystal and Colour Tests for the New Syn- thetic Narcotics," Bulletin on Narcotics, 4, No. 4, 16-42 (1952). 14. Clarke, E. G. C, and Williams, Margaret, "Microidentification of the Opium Alka- loids," Bulletin on Narcotics, 7, No. 3-4, 36 FORMS OF MICROCRYSTALS 33-42 (1955); "Microchemical Tests for the Identification of Alkaloids," /. Pharmacy and Pharmacology, 7, 255-62 (1955). 15. Clarke, E. G. C. (a) "Microchemical identi- fication of Sugars as Osazones", J. Phys- iology, 135, 28-9 P, from "Proceedings of the Physiological Society," December 1956. (b) In /. Pharmacy and Pharmacology, "Microchemical Identification": Local Ana- esthetics, 8, 202-6 (1956); Some Less com- mon Alkaloids, 9, 187-92 (1957); Antihista- mines, 9, 752-8 (1957); Antimalarials, 10, 194-6 (1958). Atropine-Like Drugs, 11, 629-36 (1959). "Microchemical Differentia- tion between Optical Isomers of N-Methyl- morphinan Analgesics," 10, 642-4 (1958). (c) "Microchemical identification of some modern analgesics," Bulletin on Narcotics, 11, No. 1, 27-44 (1959). Charles C. Fulton FORMS OF MICROCRYSTALS Descriptions and classifications of micro- chemical crystal forms have scarcely any relation to orthodox crystallography. What is needed is not a deduction of the crystal system — even if that were generally possible, which it is not — but simply a straightforward way of describing and classifying the forms just as they are seen as a result of the chemi- cal test. In the comprehensive study and classifi- cation of microcrystals, along with /orm goes size, and the latter is measurable. In the fol- lowing classification of forms, however, the concern is not with the actual measured size, but only with considering, in relation to form, how many different measurements, of value to a description or classification, might be made on a particular type of crystal. Crystals exist, of com'se, in three dimen- sions, but may extend significantly only in one direction, length, if they are fine needles, —or only in two, length and breadth, if they are plates or blades of negligible thick- ness. In the latter case, moreover, they may lie flat, especially if under a coverslip; so that such crystals, certainly as we see them, extend simply in two dimensions. The number of dimen.sions subject to use- ful measurement will not (in general) be greater than the number of dimensions of ex- tension, but may be smaller. In the case of a regular hexagonal plate, for example, only one measurement is needed to complete the description (so far as form and size go) — no matter whether we make it as the length of a side, or as a radius, or as the diameter from one vertex to the opposite one. The writer uses the term "grain" for crys- tals which resemble the familiar grains of sand, salt, sugar, etc., as they appear under moderate magnification. The three dimen- sions of "extension" are essentially equal, so that only one measurement is needed (the diameter) to complete a description in a particular case. These dimensional concepts give six classes; but there are two distinct kinds of crystals extending in three dimensions, each requiring two different measurements. They may best be taken as two separate classes (see table). It is also convenient to separate elongate or directional plates from blades. Together with a zero class for crystals of no significant dimensions (appearing as mere specks with the magnification usually used), this arrangement therefore provides nine classes in all (0-8) for the simple crystal forms. Distortions, contortions, and skele- tons, as well as aggregates, should be referred to the corresponding sunple forms. Skeletonized crystals are here classed with the forms from which they are derived (which may occur along with them), rather than with the simple forms they may finally resemble in their parts. For example, some crystals vary from square plates to crosses. The flat, thin cross should still be assigned to the same class as regular plates although it might be described as having arms of blades. Skeletonized forms can often be distin- guished from aggregates by the birefring- ence. If a complex form extinguishes and brightens as a whole, it usually is basically 37 CHEMICAL MICROSCOPY Table of Crystal Forms 0-0 Specks; precipitute often aniorphous-lookiiiji with l)irefriiigent specks seen Avith crossed nicols. 1-1 Needles (fine) distortions — hairs aggregates — fans, tufts, sheaves, bundles, rosettes, dendrites of needles; burs; wigs, spirals, balls of hairs 2-1 Plates essentially regular (triangles, squares, hexagons, octagons, disks) unformed or vague — precrystalline disks, smudge rosettes distortions — irregular but not definitely elongate plates, flakes skeletons— crosses, stars (formation all one crystal) 2-2 Elongate or Directional Plates (diamonds, rhomboids, elongate hexagons, etc.) skeletons — X's, flat nail shapes, flat combs, etc. twins — hourglass shapes, books, etc. aggregates — rosettes of plates; stepped or segmented plates; spears (built of dia- monds) 2-2 Blades (thin flat forms with length considerably greater than breadth) distortions — irregular blades; splinters, ribbons skeletons — serrate and feathered blades aggregates — fans, sheaves, rosettes, dendrites of blades 3-1 Grains (cubes, polyhedra, pyramids, gems, kernels, sharply angular grains) unformed or shapeless — globes, globulites, spherulites; nuggets aggregates — granular clumps; sphero-rosettes of kernels; chains of grains 3-2 Rods, prisms; spindles, thick coarse needles,— (unimmeasurable cross-section) distortions — sticks, wires aggregates — fans, sheaves, rosettes, clusters, dendrites of rods, etc. 3-2 Regular Tablets (regular plates with thickness) distortions — tablets irregular but not definitely elongate skeletons — regular forms with thickness (e.g., crosses); regularly branching iso- tropic dendritic skeletons 3-3 Elongate Tablets and Bars with edges — chisels; thick ribbed blades skeletons — thick combs, ladders, etc.; dendritic anisotropic skeletons aggregates — clusters of bars; irregular multiformed dendrites a single crystal, but if different parts brighten and extinguish independently, the whole must be composed of different simpler crys- tals growing from each other or from the same point. The above table summarizes the sug- gested classification, including numerous kinds of distortions, skeletons, and aggre- gates. This is not a final answer to the prob- lem, but shows that the many extremely di- verse kinds of crystals can be grouped in a small number of classes, using primarily the two simple concepts of extension and meas- urement. Charles C. Fulton HISTORY The microscope was invented between 1590 and 1609, and thus was available dur- ing the very beginning of scientific chem- istry. Scientists of the 1700's used it to look at all sorts of small things, including natural crystals and such things of chemical nature. Some trvie chemists, as soon as there were any who deserved this name, used it, in an observational way, in their chemical re- searches. True "chemical microscopy", although not then known by this name, began at least as early as 1742. In that year "The Micro- scope Made Easy", by Henry Baker, Fellow 38 HISTORY of the Royal Society and Member of the Society of Antiquaries of London, was pub- lished. Baker stated the following, under the chapter heading, '"Of Salts in. Mineral Wa- ters": "The Microscope may be of great Service to determine by ocular Examination, what kind of Salts our medicinal Springs are charged with, whence to form a Judgment in what Cases their Waters may be drank to Advantage." Five kinds of "fossile salts" were then enumerated. Marggraf, in 1747, first published his re- search, which proved the presence of a true sugar in the beet, and thus laid the founda- tion of the German sugar industry. Despite his early date, he called himself and may truly be called a chemist (then "chymist"), and he used the microscope as a matter of course, as an aid in this research, remarking that little white crystals, looking like those of sugar, could be seen sprinkled over dried slices of the root, as a preliminary indication of the presence of sugar. Another book by Baker, "Employment for the Microscope", was published in 1753. This work had two parts, of wiiich Part I, con- tainmg 232 pages out of 442 for the book, had the title: "An Examination of Salts and Saline Substances, their Amazing Configura- tions and Crystals, as formed undex* the Eye of the Observer." His method was to pre- pare a saturated solution of a salt and then observe the formation of the crystals as a drop, spread out on a slide, began to dry up. There were 55 short chapters on dif- ferent kinds of salts, and 9 plates out of 17 for the book illustrated their crystals. Not much was yet knowai about chemical com- position, and Baker did not use reagents in his procedure. He was primarily a micros- copist and such chemistry as he had was about as crude as it could be ; but it may be pointed out that he preceded Raspail, often cited as having been "the first" or "earliest" in microchemistry, microscopical chemistry, or chemical microscopy, by 80 years; also he preceded Emich, who still is often consid- ered a pioneer, by over 150 years. Chemistry became fully scientific in the late 1700's and in the early 1800's micro- scopic chemistry was still basically an obser- vational science. Chemists examined rocks, sections of plants, any and all kinds of things under the microscope, as a means of learning something about their chemical constitution. By this time, they tried reagents on the things seen, to observe w^hich parts or par- ticles reacted and how. The reagents might be of any kind and in fact color reactions were perhaps most used. Crystals, certainly, were observed, as they had been by Baker and Marggraf, but they were still usually natural crystals, or crystals of common chemicals, or of substances isolated by some chemical process; as yet they w'ere very sel- dom crystals formed in tests intended to produce them for diagnostic purposes. This microscopical chemistry, which was the original "microchemistry", was the chemis- try of substances observed through the mi- croscope. Eminent chemists of this transitional time who used microscopic methods in some of their chemical researches included WoUaston in England and Dobereiner in Germany. Other chemists also used the microscope on occasion, chiefly to look at their material to be worked on; and they included some mi- croscopically observed data in their writings. Thus Hehvig notes that Hiinefeld had in- cluded mention of the crystal forms of the alkaloids then known in a book of 1823. Most emphatic of the early chemists in recommending the microscope for the uses just described, and indeed throughout chem- istry, was F. V. Raspail, whose "Xouveau Systeme de Chimie Organique, Fonde sur des Methodes Nouvelles d 'Observation", w^as first published in Paris in 1833. Organic chemistry then and for some time still was the chemistry of living things, their constit- uents and products; Raspail may be called a biochemist in modern terms. 39 CHEMICAL MICROSCOPY At about this time a new means of ob- not the first, as stated by the "Dictionary serving crystals microscopically was devel- of National Biography", for Talbot had al- oped, that is, the polarizing microscope, ready taken the first photomicrographs of Double refraction had been observed by history as early as 1835, and reported on his Bartholinus as long ago as 1069, and about work in January 1839. Both men used the 1678 was partially explained by Huygens, solar microscope, on which Reade made im- who later in the century also made further provements. Talbot specifically called atten- observations bearing on polarization. The tion to the value of the photography for definitive discovery of polarized light is recording the appearance of microscopic, credited to Mains, over a hundred years chemical crystallizations, later, in 1808. By 1811 it was completely At least as early as 1838 toxicologists were explained in terms of light theory by Fres- concerned with how they might identify al- nel and by Arago. The optical crystallog- kaloids, for obviously no drastic treatment raphy of minerals was developed rapidly by could be used as in the separation of a min- Malus, Biot, and others, especially Sir David eral poison, and then there was the problem Brewster from 1811 to 1819 and later. of distinguishing these poisons from harm- Brewster seldom referred to magnifying less, possibly prevalent, natural bases. Al- aids but used them as convenient; e.g., by ready some suggestion had been made that mounting a plano-convex lens on a plate of microscopic distinctions might be possible, tourmaline or agate as an analyzer, and even, but in 1838 there was as yet little or no idea on occasion, directing polarized light through of using a reagent to produce crystals which a compound microscope. Sources of the could be seen microscopically to be different polarized light were reflection (Mains' origi- for different substances, nal observation), or oblique transmission This idea, however, gradually appeared through a set of plates of glass, as well as during the 1840's and 1850's. It was present, light transmitted through tourmaline, which somewhat rudimentally, in some work of has extreme dichroism. A black mirror could Thomas Anderson of Edinburgh, who de- also be used to "analyze" by reflection; and scribed test-forms of some free alkaloids and also, as early as 1819, Brewster had discov- their thiocyanates in 1848. A specific crystal ered how to extinguish one of the images of test appeared in 1852-53, with Herapath's calcareous spar, nearly perfecting the ana- discovery of the remarkable quinine iodo- lyzer, and anticipating, in a way, Nicol's sulfate, and his own recommendation that invention. the crystals could be formed in a drop of Nicol invented his remarkable prism in reagent as a test for quinine. Both of these 1828. In 1834 (W.) H. F. Talbot made the were included in "The Micrographic Die- polarizing microscope a definite instrument, tionary", by J. W. Griffith and Arthur Hen- and immediately applied it to chemical sub- frey, first edition 1856. In 1859 Taylor, in jects. This w^as the same Talbot who became the second edition of his toxicology, "On one of the founders of photography, and of Poisons", gave seven different microcrystal modern archeology. Apparently chemists in tests for strychnine, and even made refer- general were not alert to utilize the new in- ence to the "polarizing properties" as help- strument, and only in recent years have ing to characterize the crystals, as well as they begun to borrow it back from the min- giving some other microcrystal tests, includ- eralogists. ing one for cyanide using crystals given by J. B. Reade, another chemist, microsco- HCN vapor in a hanging drop of AgNOs ; pist,and photographic discoverer, took photo- but the microcrystal tests were not system- micrographs in 1839. However, they were atically developed. 40 HISTORY The new science of microcrystal identifica- tion tests came of age in the 1860's. Perhaps it may be said that the idea had crystalhzed with the issuance of the prospectus of Worm- ley's book in the United States in 1861. This pubhcation was delayed by the Civil War, and actually the honor of the first book of this science, at least the first findable by the writer, goes to Helwig, whose "Das Mikros- kop in der Toxikologie" was published in Germany in 1864 and 1865. Chamot and Mason have also mentioned a similar book by Erhard, "Die giftige Alkaloide u. d. Aus- mittelung auf Mikroskopischen Wege", 1866. Wormley's great book, "Microchemistry of Poisons", was fuially published in 1867 (1869). He gave attention to sensitivities and established the science on a firm basis. His book is a landmark; the tests are still good but, of course, the number of com- pounds covered has now become quite inade- quate. A second edition was issued, without very much change, in 1885. In those days, most poisons were inorganic or else natural plant alkaloids, and crystal tests were developed for both kinds. Later, such tests were extended over the whole field of inorganic chemistry, but their ex- tension to other organic compounds besides the alkaloids has been \'ery slow and meager. Inorganic tests were further developed by Haushofer in "Mikroskopische Reaktionen" (1885), and by Klement and Renard in "Reactions Microchimiques" (Brussels, 1886). Behrens' outstanding work, with some publication as early as 1882, culminated near the end of the century in "Anleitung zur Mikrochemischen Analyse", 1895-97. He brought the inorganic part to a high level, and made a strong effort to develop the science throughout the organic field, with many examples of reactions and descriptions of crystals suitable for microscopic identifi- cation tests. In spite of this, the majority of chemists even today think of such tests as suitable only for alkaloids, if they use them at all. A new edition of Behrens' work, by Kley, in 1921-22, represented the fourth edition for the inorganic part, the second for the organic part. The inorganic field has been developed further, and admirably, by Chamot and Mason, especially in volume II of the second edition of their "Handbook of Chemical Mi- croscopy" (1940). The work on alkaloids was carried on in excellent fashion by Stephenson, "Some Microchemical Tests for Alkaloids", 1921, and Amelink, "Schema zur mikrochemischen Identifikation von Alkaloiden", 1934. Ame- link gave attention to just a few compounds besides alkaloids, reacting with the same reagents. "Toxikologische Mikroanalyse", by Ro- senthaler, 1935, included tests for various inorganic and other organic compounds as well as alkaloids and their modern analogs, and is very valuable although it often seems not too well based on the best preceding work. (It was reissued in the United States in 1946, and is still "in print".) While Behrens was especially influential, "microchemistry" usually meant the use of the microscope in making chemical identifi- cations, especially by means of reactions pro- ducing characteristic crystals. However, in the last 40 years, to most chemists the term "microchemistry" has come to mean merely small scale chemistry. Meanwhile the oldest type of microchemistry, i.e., the chemistry of things observed through the microscope, also survived and was further developed by botanist-chemists, Tunmann and Molisch in particular; it now includes considerable use of crystal-forming reagents, but is not lim- ited to them. Only the beginning of optical crystallog- raphy has been noted above, and no attempt has been made to trace the development of micro-sublimation, or micro-melting-points and the modern fusion microscopj'. 41 CHEMICAL MICROSCOPY Numerous studies giving microcrystal tests for various alkaloids and occasionally other compounds can be found in the periodi- cal literature of the last century and this one. In 1932 and 1940 the present writer reviewed the field of "alkaloidal" reagents used for such tests, and began the work of extending them to cover all compounds of basic ni- trogen. In spite of the solid body of work on mi- crocrystal tests cited above, and a consider- able total of scattered work, largely on al- kaloids, the general impression of past history, and certainly the present situation, is one of neglect. In most fields, the authors who have contributed to the development of microscopic chemistry are surprisingly out- numbered by those others who do not recog- nize any application of the microscope to that field at all. This is most astonishing in toxicology, for here the science of crystal tests began. Hel- wig thought it strange that many toxicolo- gists seemed unaware of the obvious advan- tages of the use of the microscope in their science, but this is still true nearly 100 years later, as published works show. The neglect is at least ec^ually great elsewhere. In most colleges and universities, courses in ciuali- tative analysis or analytical chemistry, whether inorganic or organic, usually do not even mention use of the microscope; if a query about it is made, it is passed off as a "specialty". How has this neglect come about? Un- doubtedly one factor is simply that the field underwent some development very early. It has often happened that a subject developed earliest is not developed best, for later re- searchers hesitate to work in a field already partly tilled. A most important factor, however, has been the strangely long period through which chemistry has passed, during which only quantitative applications were considered to have any value at all. The microscope is not primarily a quantitative tool; this is one limitation on the tests. It has even happened that reagents already well-known and useful as alkaloidal precipitants for microcrystal tests were renamed "Wagner's reagent" and "Mayer's reagent", for example, for alleged quantitative uses almost devoid of value. For three-quarters of a century or more it was the use of the microscope in chemistry, in one way or another, either purely obser- vational, or a necessary aid in tests with crystal-producing reagents, that was com- monly known as "microchemistry". Then, this name was appropriated for what is merely chemistry on a small scale, often not even on a "micro" scale at all; this new "mi- crochemistry" had quantitative aspects, and by about the 1920's it was enthusiasically received. Emich himself spoke of the micro- scope as indispensable, and described nu- merous microcrystal tests, largely derived from Behrens, in his "Lehrbuch der Mikro- chemie" (1911), endorsing also the use of optical crystallography; but since then the microscope seems to have been lost in the shuffle. The use of microcrystal tests in the or- ganic field never has disappeared among narcotic chemists, law-enforcement chemists in general, drug analysts, and others. At the present time E. G. C. Clarke in England is doing very effective work in extending the coverage of microcrystal tests, once again in toxicology. The flood of new drugs in recent years has vastly increased the demands and difficulties for reliable identification tests. The need of extended development of micro- crystal tests was never greater than it is to- day. Charles C. Fulton MIXED FUSION ANALYSIS Microscopic mixed fusion analysis consists of the microscopical observation of the melt- ing and solidification behavior of mixtures of two or more fusible substances. Two main methods of mixture preparation are em- 42 MIXKD FUSION ANALYSIS ployed: the components may be thoroughly mixed physically or a contact preparation may be made. Melting and solidification phenomena are usually observed on prepara- tions contained between a microscope slide and cover-glass; however, it is sometimes desirable to contain the mixture in a sealed micro-capillary tube. Equipment. The equipment necessary for microscopic mixed fusion analysis is usu- ally quite simple. Any microscope with magnification up to about 200 X and capable of accommodating a hotstage is satisfactory. Some method for achieving polarized light is desirable; although this may be as simple as two pieces of "Polaroid" mounted to serve as a polarizer and an analyzer. For observa- tions at various temperatures, a good hot- stage is mandatory. Hot stages covering the range 30-350°C are commercially available as is a coldstage capable of functioning above — 100°C. In addition, there are published in the literature designs for both hot and cold stages applicable at more extremes of tem- perature. Temperatures are measured either with a thermometer or a thermocouple. The commercial stages employ thermometers. Temperatures and heating rates are usually controlled with a variable resistor or a variable transformer. It is imperative that the temperature measuring devices be accurately calibrated with known melting point standards at a heating rate nearly identical with that at which an unknown is to be measured. Usually, a standard heating rate (such as 3°C/min) is selected and all calibration and melting point determinations are made at this heating rate. It is very helpful to have a good voltmeter in parallel with the hotstage heating element so that heating rates may be approximately ad- justed to the desired value at any given temperature from predetermined stage char- acteristics. A hot bar is a very useful adjunct to the hotstage microscope, especially for contact preparations. These devices are commer- cially available or may be constructed in the laboratory. However, it is possible to employ an alcohol lamp, a micro burner, or even a soldering iron to achieve the same ends. A micro sul)limator for purification of com- pounds is also a useful item for microscopic mixed fusion analysis. Contact Preparation Methods. Con- tact preparation methods rapidh^ yield in- formation concerning the nature of the phase diagram of a two-component system. A more restricted definition of microscopic mixed fusion analysis would include only contact preparation methods. A contact preparation is made in the following fashion. A small amount of the higher melting component (A) is melted between a microscope slide and cover-glass so that, on solidification, approxi- mately one half of the area of the cover-glass contains crystalline material. A small amount of the lower melting component (B) is then placed adjacent to the crystals of component (A) and in contact with the cover-glass. When the slide is warmed so that component (B) melts, the melt flows under the cover-glass and dissolves a portion of component (A). The entire preparation is then allowed to solidify. It is freciuently de- sirable, in order to insure adequate mixing, to melt back the preparation so that most of the higher melting component is melted and then to allow the entire preparation to resolidify. A concentration gradient, ranging from pure A on one side to pure B on the other side, exists in such a preparation. All inter- mediate compositions exist in some area of the zone of mixing. ^Microscopical observa- tion of the mixing zone dm'ing the cooling process yields information concerning the nature of the phase diagram between A and B. If the system is simple eutectic, both components will crystallize rather rapidly vmtil the crystal fronts reach the mixing zone. Crystal growth will then proceed at a greatly reduced rate. When the eutectic tem- perature is reached, fine grained crystals of 43 CHEMICAL MICROSCOPY eutectic composition will crystallize rapidly without appreciable change in velocity as throughout the mixing zone. The appearance the preparation is cooled. The resultant solid of a eutectic zone is quite characteristic and crystals will appear homogeneous through easily recognized once the observer is famil- the preparation and the polarization colors iar with the phenomenon. will be uniform. On heating, the entire prep- Molecular addition compound formation aration will melt at the same temperature as is readily observed in that a third solid the two starting components. Small amounts phase may be seen to crystallize in the mix- of impurities modify the behavior only ing zone of the two-component system. One slightly. If the two components are not or two eutectic zones are also observed, de- identical, there will normally be a marked pending on whether the addition compound depression of the melting point in the mixing melts incongruently or congruently. The zone. various types of solid solution are also read- Frequently, if a contact preparation be- ily detected on solidification of the melt, tween an unknown and an easily supercooled With solid solution, both component A material such as thymol is allowed to digest and component B crystals are seen to grow on the hotbar, the unknown will develop well through the mixing zone with a change in defined crystal faces in the mixing zone. It growth velocity but without the appearance is then possible to measure accurately such of an area of eutectic composition or a mo- quantities as profile angles, extinction an- lecular addition compound. If there is partial gles, dichroisom, and refractive index rela- miscibility of liquid A and liquid B, the two tive to the melt for identification purposes, liquid phases may be seen in the mixing It is also sometimes possible, with the proper zone. choice of second component, to nucleate un- Observations made during heating of a stable polymorphic forms of a given com- contact preparation serve to confij-m the pound. This enables the investigation of more rapidly obtained conclusions drawn polymorphic forms difficultly available or from observations made on cooling. In addi- unattainable by other means, tion, it is possible to measure the various An identification scheme based upon eu- melting points such as the eutectic tempera- tectic melting points has been published by tures, molecular compound melting points, the Koflers. Unknown compounds are sub- maxima or minima in solid solution systems, divided into various restricted temperature and the melting points of the two starting ranges and two reagents are used for each components. The accuracy of such measure- temperature range. The eutectic temperature ments is dependent on the accuracy of between the unknown and the two reagents, calibration of the hotstage and the care with together with the refractive index of the melt which the proper heating rate is maintained, determined by the Koflers glass powder Aside from the general determination of method serves to identify the unknown com- the type of phase diagram between two com- pound. Over 1200 compounds have been so ponents and the determination of the signifi- catalogued and either the contact prepara- cant temperatures, there are other applica- tion method or the method of mixtures may tions of the contact preparation method, be used. In principle, this method may be Identity or non-identity of an unknown com- applied to any fusible compound provided pound and a suspect compound may be rap- suitable reagents are chosen, idly established both by observation on A different method of identification based cooling and by observation on heating. If upon molecular addition compound forma- the two are identical, the growing crystal tion and applicable to aromatic compounds front will pass through the mixing zone has been published by Laskowski, Grabar, 44 MIXED FUSION ANALYSIS and McCrone. A contact preparation be- tween the reagent (2,4,7-trinitrofluorenone) and the unknown estabhshes if the unknown is of the class which forms addition com- pounds with the reagent. If an addition com- pound forms, the preparation is observed microscopically on heating. Identification is achieved on the basis of melting point of the unknown, the molecular addition compound, the eutectic between the reagent and the addition compound, and the eutectic be- tween the addition compound and the un- known. Besides these four melting points the color of the addition compound is also ob- served. In several systems the addition com- pound was found to exist in two polymorphic forms, hence additional melting points are available for characterization. This method is rapid and is applicable to liquids as well as solids. It is also applicable if the addition compound melts incongruently. The above identification scheme has been applied to alcohols by Laskowski and Adams. Although the alcohols do not gener- ally form addition compounds, they may be converted to 2,4,6-trinitrobenzoates. The 2,4,6-trinitrobenzene grouping leads to ad- dition compound formation with a variety of aromatic substances. Contact prepara- tions w^ere made between various trinitro- benzoate esters and both naphthalene and phenanthrene as reagents. Four significant temperatures (three if the addition com- pound melts incongruently) were obtained with each ester and each reagent. Satisfac- tory discrimination was achieved among all all of the alcohols investigated. The proce- dure of reacting a functional group with a suitable reagent so that the resultant deriva- tive forms addition compounds with other reagents offers promise for wide applicability of this method of identification. Method of Mixtures. Mixtures of known composition may be prepared by weighing the components together and grinding until a uniform composition is achieved. Since only small amounts of material are required for determination of a melting point with a hotstage microscope, it is possible to deter- mine temperature-composition diagrams rapidly with a minimum expenditure of ma- terials. The points of initial and final melt- ing are easily determined microscopically. In addition it is frequently possible to ob- serve polymorphic transitions in one or both of the starting components. In such mixtures, it is also possible to determine microscopically the effect of com- position and temperature on crystal growth velocity, nucleation rate, rate of nucleation and growth of unstable polymorphic forms, and the effect of added components on crys- tal habit. The method of mixtures is appli- cable to any number of components. SELECTED REFERENCES The literature on microscopic mixed fusion analysis as well as the various experimental tech- niques involved is covered extensively in the books by Kofler (1) and McCrone (2) and the re- view by Cecchini (3). Specific applications of mixed fusion analysis include the identification of aromatic compounds (4), the investigation of molecular compound formation (5), (6), isomor- phic relations between organic compounds of sulfur and selenium (7), identification of alcohols (8), identification of fibers (9), effect of composi- tion on crystal habit (10), and studies on crystal growth velocity (11). Although this list of references is not intended to represent an exhaustive survey of the literature on microscopic mixed fusion analysis, it does serve to orient the reader in the general area. 1. Kofler, L., and Kofler, A., "Thermo- Mikro-Methoden zur Kennzeichnung Or- ganischer Stoffe und Stoffgemisch," Wag- ner, Innsbruck, 1954. 2. McCrone, W. C. Jr., "Fusion Methods in Chemical Microscopy," Interscience Pub- lishers, New York, 1957. 3. Cecchini, M. A., Selecta Chimico, 16, 95 (1957). 4. Laskowski, D. E., Grabar, D. G., and McCrone, W. C. Jr., Anal. Chem.. 25, 1400 (1953). 5. FtJRST, H., AND Praeger, K., Chem. Tech., 11, 653 (1958). 6. Laskowski, D. E., and McCrone, W. C. Jr., Anal. Chem.. 26, 1497 (1954). 45 CHEMICAL MICROSCOPY 7. Cecchini, M. a., and Giesbrecht, E., J . Org. Chem., 21, 1217 (1956). 8. Laskowski, D. E., and Adams, O. W., Anal. Chem., 31, 148 (1959). 9. Grabar, D. G., and Haessly, R., Anal. Chem.., 28, 1580 (1956). 10. Arceneax, C. J., Anal. Chem., 27, 970 (1955). 11. Gilpin, V., J. Am. Chem. Soc, 70,208 (1948). Donald E. Laskowski NITROGEN-BONDED RADICALS: IDENTIFICATION Most of the really excellent microcrystal tests known for organic compounds are due to basic nitrogen, but the crystals are so affected and modified by other elements and radicals, even quite separated from the ni- trogen atom, and by the whole structure, that they identify the molecule of a specific substance as a whole. That they can do so is, in fact, their great value. However, some- times it may be as well, and more conven- ient, to distinguish closely related com- pounds by precisely the points in which they differ. The difference may be in the radicals on a basic nitrogen atom, which volatilize while still attached to it in a deamination re- action. For example, the three USP drugs levarterenol, epinephrine, and isoproterenol, all diphenols and of similar uses, decompose spontaneously in alkaline solution to give ofT respectively, ammonia, methylamine, and isopropylamine, which can be distinguished readily by microcrystal tests in a hanging drop. Still more important are cases where it is desired to learn the identity of radicals on the nitrogen atom as a step in analysis, either in identifying an unknown as some- thing already known, or in formulating the structure of a compound whose precise con- stitution is not known. In either case there are of course chemical tests, such as those given by Feigl, for distinguishing primary, secondary, and tertiary amines; but the microcrystal tests are more definite and show exactly what radicals are attached to the nitrogen atom, provided the deamination reaction can be obtained. With many com- pounds, stable to alkali alone, alkaline oxida- tion with permanganate will cause deamina- tion, the nitrogen carrying with it any simple carbon-hydrogen (non-oxygenated) radical attached to it, as it comes off. For example, hydroxyamphet amine and methoxamine yield ammonia in this way, while phenylephrine and ephedrine yield methylamine (ephedrine volatilizes slowly unchanged, if not oxidized). Hordenine in the same way yields dimethylamine, cho- line yields trimethylamine, and N-ethyl- ephedrine yields methylethylamine. A test of an antibiotic, the exact structure of which was not known, showed that the amine com- ing off, either spontaneously from alkaline solution or more rapidly with alkaline oxida- tion, was unmistakably dimethylamine, showing that the nitrogen atom bore two methyl groups in the original compound. Hanging-drop tests above an alkaline so- lution apply also, of course, to basic com- pounds that are themselves volatile. Also there are cases where an ester is soon hy- drolyzed by alkali, and if basic nitrogen is present in the part that supplies the alcohol rather than the acid of the ester, the simpler basic compound resulting will very probably be volatile, as occurs with procaine and other synthetic anesthetics. Such decomposition compounds can be detected and identified by microcrystals in the hanging drop. However, the tests given below are suggested specifi- cally for the simplest compounds resulting from deamination. The same tests are of course directly applicable to any minute amounts of the lower amines, whether formed by decomposition of a larger mole- cule or already individual compounds. The tests may be useful in several fields, e.g., botanical chemistry, as well as food and drug work. Procedure. Stir a very little of the sub- stance into a drop of 5% NaOH in the de- pression of a cavity slide. Set a plain slide 46 NITROGEN-BONDED RADICALS: IDENTIFICATION on the cavity slide and put on it, over the fringencc, dichroism if it occurs (as with center of the cavity, a little droplet of sodium diethylamine hioniaurate), etc., as well as tetraphenylboron solution (1:20 in water); form. To the case of the red and brownish- then invert this slide. Tetraphenylboron is yellow hexafi;onal crystals with dimethyl- extremely sensitive to ammonia and the amine (as well as the similar crystals, lower amines, and if one of them is evolved, brownish -yellow only, with methylethyl- characteristic crystals form rapidly in the amine) it is surprisingly easy to get a good hanging drop. Examine them with a polariz- interference figure (a uniaxial cross), and ing microscope with the slide in place, or determine the sign of the crystal (positive), transferred over an empty cavity to prevent Working sensitivities, except for reagent further formation. In case of a mere trace of 3, are of the order of hundredths of a micro- ammonia, compare with a blank test using gram of the evolved amine captured in the the same reagents. Ammonia is so common hanging drop. Reagent 3, although less sen- from various causes, including contamina- sitive than was desired, can be vised for tion, and the tests so sensitive, that often highly characteristic crystals with as little its crystals are not very distinctive, as those as 4 or 5 micrograms of ethylamine. The of the lower amines are. alternative given for ethylamine, reagent 4, Usually a result appears within a few min- has the desired sensitivity, but has disadvan- utes, if at all. If there is no result in an hour tages due to the reagent itself, and because or so, add a drop of 1 % KMn04 solution and the crystals of the particular test are less eas- invert over the cavity a fresh droplet of ily distinguishable from others than is the tetraphenylboron solution. Or better, if case with the other recommended tests. The there is no shortage of material, run the test tests with reagent 5 can be obtained in the with oxidation at the same time as the one presence with NH4CI. (See table), with alkali alone. The oxidation test may Compare results closely with the crystals give a different result even when there is given by knowns, until ciuite familiar with some evolution of a volatile base. Examine them. the crystals wdth the slide still inverted, over Selection of a "best test" for ethylamine an empty cavity. gave the most difficulty. Its test with If either technique is effective for crystals IIAuBr4 in 2H-iP04- IHBr was passed over at apparently due to a lower amine, prepare the time the table was drawn up, because another test. Use a hanging drop of simply good crystals form only at the periphery of 1 % of concentrated HCl in water. After an precipitation. However, they are very char- exposure which may be judged from the acteristic, divided hexagons, quite different previous reaction, or up to about an hour, from the crystals of methylamine (or any reinvert the slide and allow the drop to dry others so far seen) with this reagent. HAuBr4 up (it may be put in a desiccator if hygro- in 2H3P04-1 (Acetic acid) also gives good re- scopic). Examine the residue with a polariz- suits. These tests may be compared with the ing microscope and then test with one of the tw^o that had been selected for the table, reagents suggested below, according to the Various examples of microcrystals of com- identification or indication of the tetraphen- pounds are illustrated in Fig. 1 for NH? , and ylboron test. Fig. 2 for methyl-, ethyl-, dimethyl- and tri- Put a droplet of reagent on a small cover- methylamines. glass, then invert it upon the residue of hy- Test for Aninioiiia with Formalde- drochloride. Examine the crystals, using a hyde. There should be no difficulty in iden- polarizing microscope, with magnifications tifying ammonia with certainty in the fore- of about 100 X and 200 X, observing bire- going procedure, noting the small crystals 47 CHEMICAL .MICROSCOPY Recommended Tests Structure Base given off Recommended reagent Reagent No. H / — N \ H Ammonia HAuBr4 in H3PO4 , (20) 1 H / — N \ CH3 Methylamine HAuBr4 in 2H3P04-lHBr, (24) 2 H / — N \ C2H5 Ethyl amine HAuBr4 in 9H3P04-2H20-15HBr, (26) or HaPtle in diluted H2SO4 , UOO) 3 4 CH3 / — N \ CH3 Dimethylamine lodine-KI reagent B-1 5 CH3 / — N \ C2H5 Methylethylamine lodine-KI reagent B-1 5 CHs +/ — N— CH3 \ CH3 Trimethylamine lodine-KI reagent B-1 5 H / — N CH3 \ / CH \ CH3 Isopropylamine HAuBr4 in H3PO4 , (20) 1 C2H5 / — N \ C2H5 Diethylamine HAuBr4 in H3PO4 , (20) 1 48 NITROGEN-BO.NDKU KADICALS: IDENTIFICATION with tetraphenylboron, the characteristic isotropic crystalUzation of NH4CI, and the characteristic crystals with HAiiBr4 in H3PO1 . However, a test especially for am- % 3:^ '••'is ^•%:#^ 4 Fig. la. Ammonium chloride isotropic crystal- lization (from water or dilute HCl). 66X. monia is desirable, as it is the commonest product. Ammonia condenses very readily with form- aldehyde to form methenamine (hexa- ^,-:^ > X-, Fig. Ic. NH4CI with HAuBr4 in H3PO4 , (20) crystals at periphery of precipitation, with a slightly moist reagent. lOOX. Fig. 2a. Methylamine (from epinei)liriue in Fig. lb. NH4CI deposit with HAuBr4 in volatility test with Na tetraphenylboron (1:20 in H3PO4 , (20). 135X. water), in hanging drop. lOOX. 49 CHEMICAL MICROSCOPY k>'' I -»» * , . \ \ ' i %* . A I '^ Fig. 2b. Ethylamine HCl with 1.3 HAuBr4 in 9H3P04-2H20 15HBr, (24), 135X. ^ ;>.^,- / Fig. 2c. Ethylamine HCl with 1.8 HoPtle in diluted H2SO4 , (100). lOOX. methylenetetramine), (CH2)6N4 . The low organic amines cannot give such a highly- condensed product, and under the conditions specified here will hardly give more than trace reactions nevertheless they will inter- FiG. 2d. Dimethylamine HCl with iodine-KI reagent B-1. Yellow-brown diamonds and hexa- gons, fairly large at periphery of crystallization. Similar crystals are given by methylethylamine HCl, but only dimethylamine gives red hexagons forming among and from the little crystals (2e) . « A Fig. 2e. Dimethylamine HCl with iodine-KI (5:80) in H3PO4 (2:1) (iodine-KI reagent B-1). (direct addition) lOOX. Crystal red plates, form- ing from and among little brownish yellow plates or flakes. 50 OBSERVING IMICKOCRYSTALS ^**t 1. ^ t^^^ i"^ ^.^* ^" ^ ^ V, ■+4» ^ If V'V^ ## # Fig. 2f. Trimethylaniine HCl with iodine-KI (5:80) in H3PO4 (2:1) (iodine-KI reagent B-1). (direct addition), crystals black, opaque. lOOX. fere more or less with the ammonia reaction. The recommendation of the test, therefore, is still for cases where ammonia is the only product coming off, or at least the chief one. Proceed as previously described, but in- stead of fixing the evolved ammonia with dilute HCl, use a hanging drop of neutral, 1% formaldehyde solution (0.1 ml of the usual concentrated formaldehyde, 37 % by weight, diluted to 4 ml with water). After exposure above the alkaline solution, rein- vert the hanging drop and allow it to dry. Methenamine gives branching isotropic crystallization, frequently with tripartite forms. Examine for a trace soon after drying, as it has a slight volatility. In a blank test there is at most a very slight deposit of tri- oxymethylene left from the formaldehyde solution itself, which usually does not show anything definite microscopically, and does not react in the following crystal tests. To confirm methenamine, redissolve the deposit in a little droplet of water, add a droplet of iodine-KI solution (1:1 g in 100 ml), and preferably rein vert o\'er a cavit}^ containing a little crystal of iodine — ^this prevents evaporation of iodine from the test- drop, and the crystals can be examined at relative leisure. Characteristic birofringent crystals form at once. Methenamine can also be confirmed with reagent 3 above, HAuBrj in 9H.3P04-2H.>0- 15HBr. Put a droplet of the reagent on a small cover-glass and invert on the dry resi- due. Allow a short time for the formation of characteristic crystals with a very small amount . Of the lower amines, methylamine gives the most noticeable results, including minute microcrystals, but it could not possibly be confused with ammonia if more than a trace of the latter is concerned, or if the character- istic crystals are obtained and observed. The formaldehyde test for ammonia, al- though confirmed by the usual sort of basic- nitrogen microcrystal tests, in the formation of methenamine illustrates the use of ciuite a different reaction for microscopical chem- istry. Charles C. Fulton OBSERVING MICROCRYSTALS To make full use of microcrystals, the chemist-analyst must learn to see and under- stand all the characteristics that may possi- bly distinguish them. In studjdng a particu- lar test, not every detail descriptive of the crystals needs to be permanently recorded, but everything that can readily be observed should be observed, in deciding what is worth writing down for a formal description, and what should be the points of compari- son to establish identity between known and unknown. Too often those who try such tests content themselves with a very superficial observation of the crystal forms alone. It is surprising how much can be done with such observations, often using mereh^ the ordi- nary microscope; it is not nearly as bad as failing to use the microscope at all, but still 51 CHEMICAL MICROSCOPY a gross neglect of potentialities. The polariz- interference colors mainly of a higher order, ing microscope is necessary, and this means Even with crystals having a deep color of a good instrument, not just "polarizing at- their own the birefringence may be expressed tachments", which are no more than a make- as dim, moderate, or bright. With crystals shift. that are colorless or only lightly colored in A magnification of about 100 X is ordi- themselves the place of an interference color narily used, and about 200 X when a higher in the first order can be specified as precisely power seems advisable or necessary. as the uniformity of the crystals warrants; Looking at crystals through the micro- the order of higher colors can be determined scope, the chief characteristics of form and with the quartz wedge, aggregation immediately catch the eye. (See Note, on more than one crystal, whether "Forms of Microcrystals", p. 37.) Size, not extinction is parallel to a principal side (or measured but rather loosely appreciated rela- to the general direction of an irregular crys- tive to other crystals commonly seen, is also tal), or bisects a principal angle, or is oblique, obvious. Further things to observe will now The possibility of measuring an important be suggested. angle, at least approximately, by using the Note how many dimensions of the crystals revolving stage, should not be forgotten, may be worth measuring. Note whether the This applies both to angles of form and to crystals are fairly uniform, or show diversity the angle of extinction, within one type, or whether there are two Unless the angle of extinction is close to or more distinct kinds. Look over the whole 45°, if the crystals are elongate, or direc- precipitate, if there are many crystals, to tional at all, the sign of elongation is impor- note how some forms develop from others, tant, and finding it usually requires no more Thus disks or shapeless plates when more than pushing in the red plate and observing perfectly formed may be hexagons, or per- the result. With high-order interference col- haps octagons. Squares and hexagons often ors it can usually be determined with the skeletonize into 4- and 6-armed stars; ob- quartz wedge. All crystals can be put into longs and rhomboids into X-shapes. That four groups by the sign of elongation: -1-, the diamond shape is related to the hexagon — , =t, and indeterminate, is frequently evident; also diamonds may Birefringence often shows whether a com- extend into daggers and spears. Often very plex form is all one crystal or an aggregate irregular forms can be related to a fairly of several. X-shaped crystals are usually simple form. Regarding details of descrip- skeletons of a rhomboid or an oblong, along tion, note particularly the ends of blades, the diagonals. The direction of the acute an- rods, prisms, bars: whether square, slanting, gle shows the elongation of the parent form, pointed, incised, ragged. A number of points regardless of distortions of the arms. (Fig L) in regard to form have to be noted in connec- Colored crystals are nearly always due to tion with birefringence. a colored reagent, and to this extent the Usually the next step is to look at the color does not distinguish the substance crystals with crossed nicols. First, note the tested, but the color is more significant the degree of birefringence. It will vary, of course, more it differs from the usual color produced with the thickness of individual crystals, but in crystals by the particular reagent. The very often the crystals of a precipitate will color between crossed nicols is often signifi- appear uniformly of much the same color, a cant. Some crystals are even known which do weak gray, or gray-white, or bright white to not extinguish completely but turn a differ- yellow, in the first order, or, on the other ent color at "extinction" positions, hand, nearly all may show different brilliant Dichroism is common in crystals with 52 OBSERVING MICROCRYSTALS some colored reagents, giving not merely the fact of dichroism to distinguish certain sub- stances, but two different colors to be ob- served and specified. Some crystals (mostly with iodine reagents) are pleochroic, with three extreme colors, and as seen on the slide usually shoAV quite a variety of colors with polarized light, changing with rotation of the stage. Pleochroism is the general term (in- cluding dichroism), but as there are usually only two quite different colors, and moreover an individual crystal as it lies on the slide can only show two extreme colors with rota- tion of the stage, the writer prefers to use the term dichroism when it is applicable. Dichroism can usually be noted merely by rotating the stage, but when it is feeble it may be necessary to test the extinction posi- tions, which show the extreme colors, to ob- serve it. A crystal is turned to extinction position between crossed nicols, then ob- served using only the polarizer or only the analyzer, then observed again in the next extinction position, at right angles. The change in color may be slight, or great; di- chroic microcrystals are known with iodine reagents which change from slight yellowish to black, with bromauric acid which change from pale yellowish (sometimes appearing colorless) to deep bright red, with iodopla- tinic acid which change from pink to dark blue, or from green to purplish red. Note not only the two different colors, but also their orientation. The crystal is said to have a positive sign of absorption when the darker color is "lengthwise", negative when it is "crosswise". The sign of absorp- tion, which also depends on elongation, nearly always agrees with the usual sign of elongation, when both can be observed. Dichroic crystals often show a peculiar quality of Color in ordinary light, and also show the deep dichroic color where they overlap at right angles. Pleochroic (tri- chroic) crystals may be recognizable even in ordinary light. Observe the relation of various forms to Fig. 1. dZ-Amphetamine with HAuCli in (1 + 2) H3PO4 , applied directly to tablet material. Crossed nicols. 66X . The X-crystals show negative elongation. birefringence. Hexagons may or may not show birefringence. When they do not show it, lying flat, there are often interspersed rods that do. Crystals appearing square or four- parted may be isotropic, or only some of them may show birefringence, in this case usually not very strong even when present (interference figures possible). All may show definite birefringence — even quite high — and may extinguish parallel to the crosshairs in some cases, or diagonally with crystals of a different kind. Thus crystals of the same form may be of quite different types when birefringence is also considered. Examine for interference figures — this requires at least a 20 or 21 X objective — when sizable, transparent cr.ystals show only low or no birefringence, especially if in the latter case the crystals show birefringence when tilted, or some of them are more or less birefringent (tilt of the axes in the crystal), or are accompanied by other crys- tals (possibly of the same crystal system, but a different elongation) which are quite 53 CHEMICAL MICROSCOPY birefriiigent. Interference figures, while im- portant in optical crystallography, have been very Httle used in niicroerystal tests. In most cases it would be a waste of time to look for them, as they could not be found. On the other hand as soon as the analyst learns to recognize the kinds of crystals on which it may be possible to find them, they become an added diagnostic characteristic of value, and can often be obtained, even using the 20 or 21 X objective, quite clearly on surprisingly small crystals (e.g., down to about 25 M diameter, in the case mentioned in the previous article). Moreover, the sign of the crystal can then usually be obtained, as distinguished from the sign of elongation, which may or may not be the same, or the crystals may not be elongate. Some kinds of crystals will give only indistinct figures, and for test purposes it is not worth trying to see them. A good example, but one seldom followed, was set by the Behrens-Kley text in stating actual sizes of microcrystals. No doubt merely saying small, medium, or large is often sufficient, and of course in a compari- son one can instantly see whether the con- trol crystals are of the same order of size as the crystals obtained with the sample. Moreover, one must often allow for change of size with various factors; dilution or stir- ring, for example, may cause diminution in size. However, if an ocular micrometer (kept in an extra ocular) has once been cali- brated it certainly does not take long to use it, and the size of fairly uniform crystals under controlled conditions (i.e., in the test as usually made) can be a valuable and measured characteristic. Also the ratio of length to breadth of oblongs, for example, is then measured rather than estimated, and similar proportions for other shapes. These refinements may be reserved for important tests, but ought not to be completely neg- lected. Refractive phenomena have been seldom noted in these tests. Measurement of refrac- tive indices can be used in a few cases with- out invoking special procedures of filtration, purification, recrystallization, etc.; e.g., when a bircfringent acidic substance is pre- cipitated from dilute NaOH solution with HCl, and the drop then allowed to dry up. However, even this involves more than sim- ply observing the crystals in the test-drop in which they form, the topic here. Observation with an ordinary microscope which supplements the polarizing micro- scope is often useful, chiefly as a matter of convenience when it is troublesome to change objectives on the polarizing micro- scope because they are the type sliding on, while those on the ordinary microscope are on a turntable. Sometimes one wants to see what the crystals look like in ordinary light; the appearance of pleochroic or highly di- chroic crystals is sometimes noteworthy. Darkfield observation may also be used. Many kinds of crystals show up remarkably well, but there is not the distinction between crystals and non-crystals usually obtained with polarized light and crossed nicols. The role of darkfield therefore cannot be more than secondary. In fact, no vital distinctions (not seen otherwise) have been observed in this way. At present it is not particularly recommended, since to have a darkfield ready for immediate use still another micro- scope might be reciuired. With a phase micro- scope both ordinary light and darkfield effects may be observed, but phase micros- copy has not proved of value in these crystal examinations. There is one other mode of observation of great advantage in a few cases, namely, use of incident light. This cannot supplant transmitted light, which is far more useful, but can supplement it. A light should be set up beside the microscope, which will throw a beam down on the stage; it is then no trouble at all to shut off the transmitted light and turn on the incident light. Usually the results are negative; that is, the crystals can 54 OPIUM, ORIGIN OF Fig. 2. Dihydromorphinone with HoPtBre in strong H2SO4 , applied directly. These crystals are opaque; photographed by reflected light. be seen much better by transmitted light and show no special phenomenon with inci- dent light. However, there is added diagnos- tic value in crystals that show up well (espe- cially on a dark background) or in an interesting way with incident hght. (Fig. 2.) Certainly not all these means of obser\'ing microcrystals will be used in routine tests, because then the analyst will be told or know from experience what to look for. However, he should know them all, and would do well to try them on new tests and in examining unknown crystals. Charles C. Fulton OPIUM, ORIGIN OF One of the best tests for determining the origin of seized opium is simply examination of a smear between crossed nicols of the polarizing microscope, using a magnification of about 80-lOOX (and higher when de- sired). Certain kinds of opium, notably Indian (Fig. la) and Iranian, are full of well- formed, highly birefringent, rod crystals, while other kinds, notably Turkish (Fig. lb) and Yugoslav, contain much less crystalline material, and that mostly in the form of small shapeless or roundish particles. Often not e\'en a single well-formed rod can be found in a smear of Turkish opium, while Indian opium contains a multitude of rod crystals. 8till other kinds of opium, as Afghan (Fig. Ic), are intermediate between these types. A small amount of the opium is treated with a drop of water and spread out on a slide. The water is primarily just a dispersing agent and the crystals can be seen floating in it. However, the examination is best made after the smear has dried up. With a strong light, the brown amorphous material is fairly transparent in a thin layer when dry. Alkali solution can be used to dissolve most of this other material, leaving the crystals, but generally this is not necessary. The usual sort of examination of a crude drug with the ordinary microscope will dis- close some of the large crystals in Indian or Iranian opium; in fact this distinction from Turkish opium was mentioned by The National Standard Dispensatory (Hare and others) in 1905, and Levine, in examining samples of seized opium for the U. S. Nar- cotics Bureau in 1945, used the crystal rods as one origin test, along with some others, for distinguishing Indian opium. However, the crystals are not at all easy to see in any number with the ordinary mi- croscope, or using only a polarizer, but spring brilliantly into view when the nicols are crossed. The present writer began using the polarizing test in 1947, also for the U. S. Narcotics- Bureau, and show^ed the crystal- line material to be narcotine, in work on methods for determining the origin of seized opium, which was later continued at United Nations, and which finally resulted in the U. N. Narcotics Laboratory now at Geneva. The number and form of the crystals de- pend partly on the content of narcotine and partly on the physicochemical reaction of the other constituents. This one test, of ao CHEMICAL MICROSCOPY (a) (b) (c) Fig. 1. Opium, (a) Indian; (b) Turkish; (c) Afghanistan. course, is not sufficient by itself to prove a particular origin, but in conjunction with various chemical assays and ash analysis it is very useful. After years of study, it is still the easiest origin test to make, and at the same time one of the best. Charles C. Fulton PURPOSE A fundamental error, which seems to be prevalent in modern "microchemistry", is the idea that the microscope has value in chemistry only as an adjunct to procedures on a small scale. Actually this is the least of its uses; and the early chemists, say from 56 QUINOLINE AS A REAGENT QUINOLINE AS A REAGENT 200 to 100 years ago, who are cited for the istry were chiefly observational: noting the beginnings of "microchemistry", were not, constituents in material to be analyzed, and primarily, groping for small-scale procedures making identifications by microscopic ap- when they turned to the microscope. They pearances existing in the material. The last had a much better appreciation of its real use was later extended to identification of value. mineral crystals by their properties in polar- The uses of the microscope in chemistry ized light, and now is gradually being applied comprise at least the following: in chemical science. About a hundred years (1) Taking a closer and better look at ago, identification by means of chemical things, and observing their minute charac- microcrystal tests was developed in some teristics, regardless of the amount of mate- aspects, chiefly inorganic and alkaloidal, but rial available. the possibilities here were grossly neglected (2) In particular, observing characteris- w^hile chemistry "went quantitative", and tics that cannot be seen at all with the un- still await anything like adequate develop- aided eye: for a century and a quarter now ment. this has meant not merely characteristics too small to be seen by the unaided eye, but Charles C. Fulton also those revealed by polarized light and an analyzer, plus compensators and the other fittings of the polarizing microscope. This use also does not depend on whether Quinoline is a useful reagent in chemical much or little material is available. microscopy for detection of a number of cat- (3) Making identifications by microscopic ions and as a group reagent for the elements observations and microcrystal tests, for named below. Pure quinoline produces char- which the microscope is essential, again re- acteristic crystals with the solid chloride of gardless of the amount of material available, any single one of the following: divalent (4) Using the microscope as an adjunct to cobalt, copper, iron, manganese, mercury, procedures on a small scale. nickel, cadmium, calcium, and zinc, and These uses are not mutually exclusive or monovalent copper. When more than one of even distinct; they overlap greatly. There is these chlorides are present, mixed crystal no intention of saying here that the early formation may occur, so additional tests are chemists never thought of the fourth of the needed for confirmation, but the mixed crys- above uses: of course they did, but they had tals formed are still a good indication of primarily in mind the other uses, which which cations are most probably present, modern chemical science seems to have for- When a cation which normally forms a col- gotten, and which most modern chemists ored product with quinoline is involved in disregard or overlook. When the early chem- mixed crystal formation with a cation which ists turned to small-scale procedures it was, normally forms a colorless product, the re- at least as often as not, to adapt them to sultant crystal generally shows the shape of microscopic observation, rather than the the colorless specie and the color of the col- reverse, to adapt the microscope to small- ored specie. WTien two cations which nor- scale chemistry. mally form colored products are involved, All the uses apply both to the materials they may form off -colored crystals; heating to be analyzed or studied, and to the results will usually develop characteristics which re- of chemical reactions, particularly the crys- semble one of the components, tals resulting from chemical precipitations. A drop of sample solution which has been The earliest uses of the microscope in chem- converted to chlorides is evaporated on a 57 CHEMICAL MICROSCOPY slide ; as the slide is removed from the source of heat, a coverglass from which a drop of quinoline is han<>;ing is placed on the solids and when the slide has cooled the prepara- tion is examined under a microscope. If larger, more euhedral crystals are desired, the slide may be gently warmed and then allowed to cool, but the cupric chloride- quinoline compound, if present, should be identified prior to this because its color is destroyed by heating. If the sample is a dry solid, it may be mounted in quinoline to establish the presence of the above-named cations as chlorides. The characteristic crystals formed by quinoline with the metal chlorides are de- scribed below, grouped by color. Yellow crystals indicate cuprous copper or ferrous iron. The copper compound appears as needles or rhomb-shaped plates and tablets; the iron compound appears as pleo- chroic (pale yellow to yellow) rectangular plates. Blue crystals indicate cobalt, nickel, or cupric copper. The cobalt compound forms pleochroic (light blue to blue) rhomb-shaped or rectangular plates and tablets; the nickel compound forms pleochroic (violet to blue) blue-violet plates and tablets; the copper compound forms pleochroic (green to blue to blue-violet) crystals shaped like elongated hexagons or footballs. Colorless, needle-like crystals indicate cal- cium or cadmium; these two are not readily differentiated from each other. Colorless, well-defined crystals indicate manganese, mercury, or zinc. The manganese compound forms small elongated rectangu- lar plates; the mercury compound forms pseudo-hexagonal plates; the zinc compound forms rhomb-shaped plates and tablets. The chlorides of the less commonly en- countered elements indium and thallium (thallous) also react with quinoline to yield colorless crystals. The indium compound forms diamond -shaped and rectangular plates; the thallium compound forms hexag- onal, rhomb-shaped, and rectangular plates. Primarily, (juinoline serves as a group re- agent; a positive test indicates that one (or more) of the above-named metal-chlorides is present and a negative test indicates ab- sence of an apprecial)le ([uantity of any of them; under favorable conditions qviinoline offers a specific test for these cations. When the sample is a solid, ciuinoline may distin- guish these metal chlorides from their oxides, or sulfates, or free metals, and it simultane- ously distinguishes the valence state in the cases of copper chlorides or iron chlorides. J. M. MUTCHLER REAGENTS FOR MICROCRYSTAL IDENTIFICATIONS The reagents given here are primarily precipitant s of organic compounds contain- ing basic nitrogen. They include, but go far beyond, traditional alkaloidal reagents. A few give good inorganic tests for certain ions; a few extend to nitrogenous compounds that are almost completely acidic, such as the barbiturates. Each of the outstanding precipitating compounds, HAuBr4 for example, makes a number of quite different reagents, by the use of different solvent media, particularly: (1) Syrupy H3PO4 ; diluted H3PO4 , (2) Diluted H2SO4 (up to (1 + 1) for chlorides, (2 + 3) for bromides, (1 -f 3) for iodides), (3) Water; and aqueous solutions only slightly acid, or made acid only by the pre- cipitating compound itself; sometimes even neutral, slightly basic, or alkaline, (4) Concentrated HBr (40%); diluted HBr, (5) Concentrated HCl (38%); diluted HCl, (6) Acetic acid (usually diluted at least (2 -|- 1) simply to prevent its spreading all over the slide). These are given above in the order of de- 58 REAGENTS FOR MICROCRYSTAL IDENTIFICATIONS dining effectiveness for precipitation. Crys- tallizing effectiveness varies with the kind of substances tested, and is greatest when a resultant compound is neither too soluble nor excessively insoluble. When insolu- bility is very high, the precipitate is likely to come down amorphous, and crystalliza- tion may not be obtainable. However, the very best tests combine ease of obtaining crystals with a high sensitivity which usu- ally means great insolubility of the product formed. All these media have specific effects dis- tinct from the general solubility effect. Combinations of media are used to get the best possible results in some cases. The reagents may be added to aqueous solutions of the substances tested (the tra- ditional way), usually without a cover glass; or, particularly those with high concentra- tion of acid, direct to a very little of the solid substance, usually with a cover glass added. In the list below, the major reagents of the most widely applicable precipitating com- pounds are given first, then a selection of others, more or less in the order of declining precipitating power, so far as this can be reconciled with a certain listing of related reagents together and related compounds near each other. The list as a whole will be found to bear but little resemblance to lists of traditional ''alkaloidal reagents", al- though the best of the traditional reagents are included. Iodine-Iodide Reagents The precipitating compound, iodine-HI or iodine-KI, or presumably the anion Is", has the widest range of any. However, the con- ditions for precipitation and suitable crys- tals vary greatly depending on the type of substance, and more types of reagents are used than for any other precipitating com- pound. Only those of greatest established value are given here. Inorganic applications certainly exist, at least with reagents made with H3PO4 , but have not been studied by anyone, to the writer's knowledge. With Phosphoric Acid. (1) lodine-HI- H,PO, reagciU. Iodine 0.08 g, HI (57%) 0.5 ml, syrupy H3PO4 (85-88%) 3.5 ml. May be kept in a small rubber-bulb dropping bottle. Remake when it has lost considerable iodine strength. Very wide range of pre- cipitation, but precipitates of many sub- stances remain amorphous or in drops. Used for crystals with various sympathomimetics, aminoacetic acid, etc. (direct addition). (2) Iodine-KI reagent B-1. Mix 2 ml iodine-KI solution (5 g iodine and 80 g KI in water to make 100 ml) with 4 ml syrupy H3PO4. Pour off from any KI that crystal- lizes out. The mixed reagent keeps quite well in an 8-ml rubber-bulb dropping bottle. Used for barbiturates and similar compounds (added to the slightly alkaline aqueous solu- tion); also (added directly to the hydro- chloride) for some of the simplest amines, etc. (3) Iodine-KI reagent M-2. Mix 2 ml I-KI solution (5:30 g in 100 ml) with 3 ml coned HCl and 3 ml syrupy H3PO4 . Used espe- cially for morphine (direct addition). With Acetic Acid. (4) Iodine-KI reagent C-3. Mix 0.4 ml iodine-KI solution (10:10, dissolved in about 15 ml water, then diluted to 100 ml), 2.0 ml glacial acetic acid, 3.2 ml water, 0.4 ml (1 -f 3) H2SO4 . Remake when it loses strength. Used especially for ciuinine, and for other cinchona alkaloids, which form iodosulfate crystals; also for codeine, etc. (direct addition). Wide range of useful crystallization with complex compounds. (5) Iodine-KI reagent 0-1. Mix 4 ml aqueous I-KI reagent No. 2 (1 : 1.75 g in 100 ml) with 2 ml glacial acetic acid. Originally labeled "0-1" because of crystals with a number of the opium alkaloids and their synthetic modifications. Aqueous. (6) Concentrated aqueous iodine reagents. Three of these have already been mentioned when used as stock solutions. They may also be used as reagents added to 59 CHEMICAL MICROSCOPY aqueous solutions, when a high concentra- tion of iodine is wanted, (a) I-KI, 10:10 (g in 100 ml). This is saturated with iodine; often the most useful ratio for crystals, although iodine evaporates from it comparatively readily. Also the more dilute solution (1:1) is chiefly used for complex compounds (such as alkaloids). (b) I-KI, 5 : 14. Used for atropine in trace ; also, added to bicarbonate solution and with added KCl (or other similar cation), for caffeine, theobromine, etc. (c) I-KI, 10:50. Used for colchicine (neu- tral or bicarbonate solution) ; etc. (7) A series of aqueous ^'alkaloidal" iodine reagents. The precise ratio of iodine to iodide has usually been ignored as an important factor, and is sometimes not even stated. However, it is vital. Using 1 g iodine to make 100 ml solution, the following amounts of KI give distinctly different reagents, of declining sensitivity (of precipitation) in some cases, and different crystals in many cases: 1 g; 1.75 g; 2.75 g; 5 g; 10 g; 20 g; 35 g; 50 g. Dissolve the iodine and KI in no more water than necessary and dilute to 100 ml after complete solution. Used for in- numerable precipitations and microcrystals of alkaloids and many related compounds (generally by addition to aqueous neutral or slightly acid solutions). Bromauric Acid Reagents The bromauric acid reagents as a group, and HAuBr4 in H3PO4 , and in concen- trated HCl, particularly, are probably the most useful of all know^n reagents for micro- crystal tests with compounds of basic ni- trogen. 1 g of the commercial "gold chloride" (HAuCU -31120) converts to about 1.3 g HAuBr4 ; dilutions of acid with water are given in terms of volume of concentrated acid -1- volume of water, e.g., (2 + 3)H2S04 ; final volume (from 1 g of the starting ma- terial) in parentheses at the end: these features may be used in specifying a par- ticular reagent precisely, especially in places where the full formula is not immediately in view. (8) HAuBri in H.POa . HAuCl4-3H20 crystals 1 g, HBr (40%) 1.5 ml, H2O 1 ml, syrupy H3PO4 (85-88%) 17.5 ml. Virtually shows the limits of the basic quality of the N atom. Used for all sorts of simple N bases, feeble ones, and those partly acidic; for certain inorganic cations; for oxonium com- pounds; also for sympathomimetic drugs, etc. (direct addition). (9) HAuBri in {3 -\- l)HzPOi , {1 -j- 3)- HzPOa , {2 + 3)H2SOi , water, cone. HCl, (1 + 3)HCl, or {2 -\- 1) acetic acid, etc. HAuCl4-3H20 1 g, HBr (40%) 1.5 ml; one of the media, e.g., (2 -f- 3)H2S04, to make 20 to 30 ml solution. In this concentration especially used for addition to aqueous solu- tions; for direct addition to residues of an alkaloid or similar compound a greater dilu- tion, to (45), or (60), may often be prefer- able. Considering both aqueous and direct uses, HAuBr4 in cone. HCl is perhaps the best single reagent known for alkaloid-type compounds. 1.3 HAuBr4 in (1 -{■ 3)HC1, (45), is used especially for caffeine, theo- bromine, etc. (direct addition). (10) 1.3 HAuBvi in 2HzP0vl{2 -f 3) HiSOi, {90). Dilute 1 part of 1.3 HAuBr4 in (2 + 3)H2S04 , (30)— preceding formula— with 2 parts by volume of syrupy H3PO4 . Used especially for the fully basic relatives of amphetamine, etc. (direct addition to crushed tablet material containing a salt of the base). (11) HAuBrA in H^PO, and HBr. (a) 1.3 HAuBr4 in 2H3P04- lHBr,(24). HAuCl4-3H20 1 g, HBr,(40%) 8 ml, H3PO4 16 ml. Used for methylamine hydrochloride especially; also for ammonium salts, iso- propylamine and diethylamine hydrochlo- rides; etc. (b) 1.3 HAuBr4 in 9H3P04-2H20- 15HBr, (26). HAuCl4-3H20 1 g, HBr(40%) 15 ml, H2O 2 ml, H3PO4 9 ml. Used for ethylamine and methylamine hydrochlorides, etc. 60 REAGENTS FOR MICROCRYSTAL IDENTIFICATIONS Chloraiiric Acid (Gold Chloride) Rea- gents (12) HAuCh in H^PO, , (1 + 2)HzP0i , (1 + l)H2S0i , water, cone. HCl, (1 + 3)HCl, or {2 -{- 1) acetic acid, etc. 1 g HAuCU -31120 in 20 ml (usually) of the appropriate sol- vent; further dilution (60) with H3PO4 , (1 + 1)H2S04 , coned HCl, or (2 + 1) acetic acid may be used for tests of direct addi- tion with easily precipitated compounds. In the traditional procedure, that is, simply in water and applied to aqueous solu- tions, HAuCU is probably the best single reagent known for alkaloid-type compounds. HAuCU in (1 -^ 2)H3P04 is especially used in hanging drop tests for volatile bases and in direct addition to solids (salts of simple bases, etc.) — as well as for addition to aque- ous solutions. No cover-glass is used and water is then allowed to evaporate from the test-drop if necessary for precipitation and crystallization. Bismuth and Platinum Iodide Reagents HsBile and H2Ptl6 are very general precipi- tants, so much so that they are the two pre- cipitating agents most com.monly used — though only in aqueous solution — for devel- oping spots of alkaloid-type compounds in paper chromatography. (At this time the writer does not yet know of any paper-chro- matographer who has used phosphoric acid, or even diluted sulfuric, to increase the gen- erality and sensitivity of these reagents.) Bismuth Iodide Reagents An aqueous (strongly acid) HsBile reagent is a traditional one from the last century; but as commonly made, it soon decomposes in part; then some effects are still due to the HsBile , but others to iodine-KI. Some users age their reagent for a time to obtain consistent results. Amelink even added io- dine crystals to have a mixed reagent from the start. The reagents given here, however, depend upon HsBile alone. For crystal purposes, phosphoric acid has been difficult to use because colored crystals are likely to form due to the reagent itself. The use of Bils and HI to make up the rea- gent has not as yet solved any important problems either, so the formulas given here simply employ a concentrated bismuth nitrate solution. The reagents have a bright orange color and are to be remade when they darken appreciably. This occurs very soon with (1 -f 7)H2S04 , and too rapidly for convenience even with aqueous reagent, if no preservative is used. Sodium hypophos- phite is here introduced as a preservative, which makes the reagents comparatively long-lasting. Cone. Bi(N03)3 soln: Dissolve 50 g bis- muth subnitrate in 70 ml (1 -f 1)HX03 and dilute to 100 ml with w^ater. (13) H^Bile in {1 + 7)H2SOi or in water, etc. (a) HgBile in (1 -f 7)H2S04 . KI 1.25 g, H2O 2.0 ml, (1 + 3)H2S04 2.5 ml, cone. Bi(N03)3 soln 0.5 ml, Na hypophosphite 0.05 g. Mix. Used especially for hanging- drop tests and direct-addition tests for sym- pathomimetics, etc. (b) HsBile (aqueous). Omit the diluted H2SO4 , using simply 4.5 ml water. Used direct for theophylline and related com- pounds; also for addition to aqueous solu- tions (traditional). (0.04 g hypophosphite is enough.) (c) Double strength HsBile (aqueous) may sometimes be preferred. KI 2.25 g, Na hypophosphite 0.05 g, HoO 4.0 ml, cone. Bi(N03)3 soln 1 ml. Mix. Platinic Iodide Reagents 1 g H2PtCl6-6H20, with iodide, makes about 1.8 g H2Ptl6 . (14) HiPth with H^POa , (A and B). Formulas for small dropping bottles. (A) 1.8 IloPtle in (2H + 1)H3P04 , (250), with minimum Nal. Dissolve 0.04 g Nal in 0.5 ml H2O; mix with 1.8 ml H3PO4 , and add 0.2 ml of 1:20 aqueous platinic chloride 61 CHEMICAL MICROSCOPY solution (1 g H2PtCl6-()H.,0 in 20 nil water). Keeps well for only about a week to a month at most; can be made up on half the above small scale. (B) 1.8 HsPtle in (4 + l)H;iP().. , (250), high Nal. Dissolve 0.5 g Nal in 0.:] ml water, add 2.0 ml H3PO4 and 0.2 ml of 1 :20 aqueous platinic chloride solution. Let stand about a day to develop good differentiation from (A). Keeps much better than the pre- ceding, (A); in a few cases a rather old reagent even gives the best results for cer- tain crystals. Both of the above are among indispensable reagents for sympathomimetics, central stimulants, and other compounds which are of simple structure or partly acidic, not too easily precipitated. (15) HiPth in diluted H2SO4 , (100). H.PtCle soln (1:20 in water) 0.8 ml, (1 -f 3) H2SO4 3.2 ml, Nal 0.46 g. Preferably let stand at least 2 or 3 hours (or overnight) before use. Thereafter it slowly deteriorates, but can be used for a long time. The crystals it gives change to some extent with the age of the reagent. Also, the reagent itself de- posits colored crystals as it partially dries, first around the edge of the cover-glass, where the solution is not covered. Used for ethylamine hydrochloride; uses not much explored. (16) H-iPth {acid aqueous). HoPtCle soln (1:20 in water) 4 ml, HCl 1 ml, Nal 1.25 g. If the HCl is omitted, the alkalinity of com- mercial Nal may cause precipitation. Only a little acid would be needed to overcome this, but Amelink indicates generally better re- sults in an acid than in a "neutral" test- drop, anyway. Platinic Bromide Reagents (17) H.PtBr^ in H.POi , (/ + 3)H,P0i , (^ -f 3)HoS0a , water, HBr, {1 -+- 3)HCl, or (2 + 1) acetic acid, etc. H2PtCl6-6H20 1 g [makes about 1.3 g H^PtBrg]; HBr(40%) 2.5 ml, appropriate solvent to make 20 ml (usually). The aqueous reagent (which can be made with NaBr instead of HBr) is excellent although strangely neglected in the past. New uses (hanging drop and direct, for sympathomimetics, etc.) especially concern H.PtBrg in (1 + 3)H:iP()4 . The reagent with syrupy H3PO4 develops some precipitation in the bottle, but the clear supernatant solu- tion can be used (rather than adding enough water to prevent precipitation). (18) HoPtBrCl^ in cone. HCl. If the above formula is used with concentrated HCl for the solvent, even with twice as much HBr (not exceeding the molecular ratio of 1 HBr to 5 HCl), there is evidence — from micro- crystals — that the precipitating compound is HoPtBrCU . The four other possible Br-Cl combinations form in proportioned mixtures of the strong acids. They can be used by direct addition, for the particular effects; as originally worked out on morphine they were diluted to (60) with the strong acids and some water (up to one-fifth) for best results. Much dilution with water tends to give HoPtBre , regardless of more HCl than HBr being present. Platinum and Palladium Chlorides (19) HoPtCh in H,POi , (/ + 3)H,P0, , (1 -f 1)H2S0a, water, (1 + 3)HCl, etc. 1 g H.PtCle-eH.O in 20 ml of the solvent. The aqueous reagent is of course the tra- ditional one and very valuable. The reagent with (1 + 3)H3P04 is particularly used in the hanging drop for d- and rfZ-amphetamine. (20) HoPdCU in H,POi , or water, etc. PdCl2-2H20 1 g; cone. HCl 0.9 ml (in H3PO4) or 0.8 ml (hi water, etc.); H3PO4 , or water, etc. to make 20 ml. Aqueous or partly aqueous reagent may be made wdth NaCl 0.75 g, instead of the HCl (forming NasPdCb). Tetraphenvlboron and Reinecke Salt (21) Na tctraphcnijlboron, 1 g in 20 ml water (not acidified). Used especially as a hanging drop; remarkable for its sensitivity 62 REAGENTS FOR MICROCRYSTAL IDENTIFICATIONS and crystals with ammonia and the lower amines. (22) Reinecke salt, NHiCr{NHs)2{SCN), . A fresh, approximately saturated, aqueous solution is used. Stir a little of the com- pound into about 0.5 ml water at room tem- perature, to provide a few drops for use. Properties of the reagent begin to change within a few hours. Rosenthaler says that it is useless when the ferric salt test for thio- cyanate ion can be obtained (red color). Some very interesting crystals with lower amines have been obtained with a still- effective reagent aged for a number of hours — e.g., overnight — but the writer does not know how to control these changes or sta- bilize any such intermediate stage. Gradu- ally, in two or three days, effectiveness is completely lost. Bromine-Bromide Reagents Although various forms of bromine- bromide reagents should be possible and valuable (as with iodine-iodide), the diffi- culty of keeping any reagent without the bromine evaporating, and its disagreeable character, have so far prevented anything but a limited aqueous use. (23) Br in HBr solution; Br in NaBr solution. HBr(40%) 10 ml, water 90 ml; or NaBr 5 g, water 100 ml. Saturate with bromine. Used for barbiturates, etc., as well as for basic compounds. Other Reagents for Aqueous Tests The following are added to aqueous solu- tions (usually of a salt of the base tested), unless otherwise stated. Complex Oxygen Acids. (24) Phosphori- timgstic acid. Obtainable commercially. 10 g in 100 ml water. Sihcotungstic acid is used similarly. (25) Phosphorimolyhdic acid with HNO?, . 10 g of the commercial phosphomolybdic acid in 90 ml water and 10 ml coned HNO3 . Phosphorimolyhdic acid has especial value as a standard for sensitivity determinations. Formerly the reagent for this purpose was made with only a few drops HNO3 ; but there is probably no sufficient reason to maintain a separate formula for it; the form with 10% HNO3 may be better for crystals as well as being used foi- the follow- ing reagent. (26) Phosphorimolyhdic acid with H^POi . A fresh solution of the commercial phospho- molybdic acid is decolorized immediately by H3PO4 , but the preceding solution with HNO3 , after standing at least 6 weeks, is more stable. To 4 ml of the stabilized yellow phosphorimolyhdic acid add 0.6 ml syrupy H3PO4 and mix. The solution should remain yellow for use (it will keep for a few days). Less sensitive but gives crystals (e.g., with narceine and atropine) which are unobtain- able with the usual phosphorimolyhdic r.cid solution. Platinum and ^lercury Thiocyanates. (27) H^PtiSCN), . H2PtCl6'-6H20 I'g, water 20 ml, NaSCN 0.95 g. This hardly has out- standing importance for microcrystals so far, but probably is the best of the "normal" thiocyanates (i.e., aside from Reinecke salt); and the writer takes this opportunity of correcting a former statement that onh- about a third as much NaSCN need be used : the NaSCN must be sufficient to replace all six CI atoms, or the reagent will precipitate on standing. (28) K2Hg{SCN)i. Dissolve 3 g KSCN in 100 ml water and saturate with Hg(SCN)2 (5 or 6 g recjuired). This is an outstanding reagent for inorganic microcrystals of a num- ber of cations, especially Zn, Cd, Cu, Co, and Au. It precipitates alkaloidal-type com- pounds but its value for this is minor. Mercuric Iodide Reagents. (29) KoHgli. Dissolve 2 g KI in 100 ml water and saturate with Hgl2 (nearly 3 g Hgl2 required). Only occasional value for crystals but much used for general tests of alkaloid-type precipita- tion. (30) Hglo and HCl. Dilute 27 ml cone. HCl to 100 ml with water (makes an actual 6a CHEMICAL MICROSCOPY 10% HCl), then saturate with Hgl2 (does not take much). Used both for aqueous solu- tions and direct addition to the solid, for heroin, etc. (31) Hgh'NaCN and Nal. Dissolve 0.5 g good NaCN in 100 ml water and saturate with Hgl2 (use 4J^^ g). (This is a reagent in itself, and more sensitive, but the reagent with excess Nal is recommended more highly for crystals.) Filter the saturated solution and add 8 g Nal per 100 ml. Crys- tals with codeine, etc. Cadmium and Lead Iodides. (32) K^Cdh . Cdl2 5 g, KI 4.5 g, water 100 ml. (33) Phi 2 in K acetate solution. Dissolve 4 g lead acetate (3H2O) and 30 g potassium acetate in water to make 100 ml, and add glacial acetic acid drop wise just to faint acidity to methyl red (reacts brown instead of yellow); then add 4.5 g KI. Many crystals with both the above. Mercuric Chloride Reagents. (34a) Simple HgCl2 (5%) is commonly used but usually added to solutions containing dilute HCl, or the hydrochloride of the base, so that actually more or less of the following chloro-acid is present. (b) HHgCls . HgCla 5 g, coned HCl 1 ml, water 99 ml. (c) NaHgCls . HgCl2 5 g, NaCl 0.75 g, water 100 ml. (d) HgCl2 & HCl. HgCl2 5 g, cone. HCl 15 ml, H2O 85 ml. This is less sensitive; useful especially for quinine. Ferric Chloride Reagents. (35a) FeCls in HCl; HsFeCle . FeCU-GHzO 10 g, cone. HCl 100 ml. For addition to aqueous solu- tions (cocaine, methadone, etc.). (b) HsFeCle for direct addition to solids: FeCl3-6H20 10 g, coned HCl 17.5 ml, water to make 100 ml. Organic Reagents. (36) Picric acid is outstanding. (a) A saturated aqueous solution (about 1K%). Many crystals, (b) A 0.2% aqueous solution. Picric acid crystals (10%) water) 0.2 g, water 100 ml. Cinchonine is an example of a base yielding crystals far more readily with this dilute reagent than with the satu- rated. This weak solution may also be used for direct additions. (c) Half-saturated Sodium Picrate. Pre- cipitate sodium picrate from saturated picric acid solution with concentrated sodium acetate. Filter, then prepare a saturated solution of sodium picrate. Filter this from excess crystals and dilute with an equal vol- ume of water. Crystals with bases are often obtained more readily than with the satu- rated acid. (37) Other nitro-organic reagents. Styphnic acid (trinitroresorcin) and Trinitrobenzoic acid are used in saturated solutions. Simple Oxygen Acids. (38) CrOz and HCrOzCl. (a) 5 % CrOs . The chloroacid or its salt (following formulas) is more sensitive and better. (b) HCrOsCl. Stock solution of 20 g CrOa in water to make 100 ml (this may be used as a concentrated CrOs reagent in itself). Reagent: 1 ml of foregoing stock solution plus 2 ml water and 1 ml coned HCl. Will keep for some time; slowly darkens. (c) NaCrOsCl. Add 1 g NaCl to 4 ml of the 5 % CrOs . Keeps well. (39) Perchloric acid. 5 % solution of HCIO4 or NaC104 . (40) Permanganic acid oxidizes most alka- loids and other bases, but the stable crystal- line permanganates are quite distinctive. Unfortunately reducing impurities can easily spoil a test. (a) KMn04 2 g in 100 ml water, with a few drops of syrupy H3PO4 . (b) HMn04 in dilute H2SO4 , for direct addition. Use a stock solution of 1 % KMn04 . Reagent: 2 ml stock solution, 1 ml (1 -f 3) H2SO4 . Make up fresh for use. Used for cocaine, meperidine, methadone, etc. Basic Reagents Precipitation of the free base by addition of the reagent to a neutral or slightly acid solution of the salt of an alkaloid, etc. 64 SYMPATHOAll.METICS AND CENTRAL STIMULANTS (41a) 5% NaOH. Sometimes a concen- trated solution is used. (b) 5 % Na3P04 . The alkalinity is about the same as for 5% Na2C03 , -which is more often used, but effervesces when added to an acid solution. (c) 5% K2Cr04 . Precipitation of a chromate of a strong base is possible, but the principal use and value is as a basic group reagent for the weak alkaloids that are quite insoluble in water. It must, of course, be added to solutions that are only slightly acid, or, if strongly acid, the effect is that of K2Cr207 (similar to CrOs). (d) Concentrated K acetate, 30 g in water to make 100 ml. For precipitation and crystals with very weak insoluble bases. Cyanides (42) Gold Cyanide Reagent. Dissolve 1 g HAuCU-SHaO crystals in 20 ml water. Add 0.5 g NaCN a little at a time; if there is im- mediate precipitation add just enough NaCN to redissolve; otherwise 0.5 g should be the right amount. Makes a colorless solution, which sould not turn red litmus blue (if it does, acidif}^ with acetic acid). (43) Platinum Cyanide Reagent. Dissolve 1 g H2PtCl6-6H20 m 18 ml water and add 1.5 g NaCN. Solution may warm up and be- come rather brown; if not, warm a little on the water bath until it just begins to darken, then cool. Cautiously acidify with 2 ml (1 -t- 3) H2SO4 . There is usually a small amount of brown precipitate, apparently due to the reaction going a little too far, in part ; this is removed by filtering either before or after acidifying. The solution is brown but not a dark brown. (44) HiFeiCN)^ with H^POi . A stock so- lution is kept of 10 g K4Fe(CN)6-3H20 in water to make 100 ml. Reagent: Mix 3.5 ml of the stock solution with 0.25 ml syrupy H3PO4 . Has to be made fresh as it will not keep. Simple Halides and Pseudohalides. (45) 5 % solutions of KI, NaSCN, NaNOz . They are also used in concentrated solutions. Charles C. Fulton SYMPATHOMIMETICS AND CENTRAL STIMULANTS Most of these are relatively simple com- pounds, compared with alkaloids (q.v.), and are not so readily precipitated. Some which have alcoholic and phenolic hydroxyl groups are not precipitated at all by the usual aqueous reagents, which are described in the article "Alkaloids and Alkaloidal-type Pre- cipitation". The following 14 tests are se lected for tabulation in Table 1. Five rea- gents are applied directly to a very little of the solid, with addition of a cover-glass : (1) 1.3 HAuBr4 in H3PO4 , (20) (2) 1.3 HAuBr4 in 2H3P04-1(2 + 3)- H2SO4 , (90) (3) 1.8 H2Ptl6 in (2H + 1)H3P04 , (250), with minimum Nal (4) 1.8 H2Ptl6 in (4 -f 1)H3P04 , (250), high Nal (5) Iodine-HI-H3P04 reagent. Three reagents, which contain more water than the five above, are used in two tests each. They are applied to the solid without a cover-glass, and let stand for partial evap- oration; they are also used as the reagent of a hanging drop, in volatility tests above an alkaline solution on a cavity slide. In the latter case the test-slide is in general rein- verted after exposure to the vapor (of up to two hours), and evaporation to a higher content of the non- volatile acid then occurs. (6, 7) HsBile reagent in (1 -f 7)HC2S04; 6 direct application, (7), volatility test (8, 9) HAuCU in (1 + 2)H3P04 , (20) (10, 11) 1.3 HoPtBre in (1 + 3)H3P04 , (20) The two following reagents are also used for volatility tests (the first permitting rein- version, but the second only as a hanging drop) : (12) HaPtCle in (1 + 3)H3P04 , (20) 65 CHEMICAL MICROSCOPY to H Z < H CD a.S < S H to (D PS o ^ -u CO H OJ trl W o H C J "^ «i m « fJ C) ^ IZ T-H H J m ■< H « "5 + >yil 0-3 a m TO . _ ■(-3 to -» I— I CJ o CO + .2 o o to ^^ . d. K%"0 .^ -w -^ ^ 2.2 g ^ 'D (u '•+J '-t^ '-i^ '•+J ' c3 c3 Cw c3 'o'o'o'o > > > > -< ^ t^O O o 0^0^^ O O OOU O-^oO 0J2 0000 OS-O 0-T3 Q o 0) c a, ID tp c3 a S >> X O a) rt O O =3 -►^ a c c tj G o "^ 2-" G"^ =; _o a^H "t: c cS O o o rt • O o N o3 (-• +i fi ^^ (—1 o 0^ r- "T"* • rH c - >> £3 c a 0 OJ 0 G G tami etam meth ?^:J G tp a 0 c s s <» -G a S Methamphe -Methamph he nyl propyl -►J a ;_ T3 -i-i .. s up IV Amph 5 Mephen Propylh c 3 ^ "^o '^ Ph 2^ -TS -e-^P-. H 66 SYMPATHOMIMETICS AND CKNTKAL STIMULANTS (13) Sodium letraphenylboron in water decomposition) in a lianging drop of dilute (1-20) HCl (about 1% by volume of the eoncen- (14) Volatility tests are also made by trated acid) above a drop of alkaline solution catching the volatile amine -(or amine of on a cavity slide, then evaporating the HCl \ m V Ik f / N '-I •> r- * \ >N^ ••. \i ^ t '*«fe » \ 1 4 \ Fig. la. Hydroxvamphetamine with HAuBrj Vt^ o i t? u i • -^u i o tta n • u u^-. /on^" J ■ • .1 /^ w l^ I ^^^- 2a. 1-Ephednne with 1.3 HAiiBrj in in H3r(J4 (20), red spindles (hrst formed), brown ott pr^ i /o i oa tt cfi mew CENTRAL STIMULANTS Fig. 5c. Nylidrin, 1.3 HAuBr4 in 2H3P04-1 (2+ 3) H,804 (90), on standing, with humiditv. lOOX. amphoteric compounds (even when mainly acidic and only feebly basic) in which the strong H3PO4 brings out the basic character; this also occurs with bromauric acid in H3PO4 . Iodine reagents are indispensable but not more so than bromauric acid; the iodine precipitates more often fail to crystal- lize, and different reagent-formulas are needed for them. In general, the precipitates are not neces- sarily crystalline. With complex and defi- nitely basic compounds, such as the alkaloids and antihistamines, the precipitates with HAuBr4 in II3PO4 , even when it is added to an aqueous solution, are far too insoluble for the best results. Such precipitates are usually amorphous or too minutely crystalline to have any value for microscopic crystal tests. On the other hand, compounds that are rela- tively simple but too water-soluble or too feebly basic, or both, to yield precipitates from an aqueous solution with reagents dis- solved in water, will generally yield beautiful crystals with bromauric acid in a medium of phosphoric acid, although occasionally only drops are formed. The crystals show great differences from one compound to another, not only in their forms, but also in colors, birefringence, and dichroism. The reagent is added directly to the dry substance to be tested. Crystals of the bromaurate com- pound are easily distinguished from undis- solved material, or any other crystals that may form, by their color. The reagent also has a general use in de- termining whether any compound capable of this precipitation is present. A little powder from a tablet, for example, may be scattered thinly on a slide, a drop of the reagent and a cover-glass applied, and bromaurate crystals or precipitation looked for under the micro- scope immediately and after standing. In this direct addition there is some danger that a compound with too great bromaurate insolubility may not show up, because the insoluble precipitate may completely cover the surface of the material and prevent further solution. Therefore, some less sensi- tive reagents should also be tried, or the test tried on an ac^ueous or dilute acid solution of the substance, before concluding that no compound of basic nitrogen is present. Precipitates may be due to other kinds of basic substances. These include: (a) Inorganic: all the alkali metals, and magnesium and zinc, in particular, in the form of thcMr salts, as well as ammonium and hvdroxvlnniine. 71 ELECTRON MICROSCOPY (b) Compounds of basic oxygen: i.e., com- pounds of a certain complexity and capable of forming oxonium salts also react (for example coumarin). These are rare by com- parison Avith the compounds of basic nitro- gen, but in a general test it should be remem- bered that some organic bases do exist which do not contain nitrogen. The reagent: Gold chloride crystals (HAuCU-SHaO) 1 g (makes about 1.3 g HAuBr4); HBr (40%) 1.5 ml; H2O 1.0 ml; syrupy (85-88%) H3PO4 to make 20 ml. This may be named in short, HAuBr4 in H3PO4 ; or in more detail as 1.3 HAuBr in syrupy H3PO4, (20). Excellent crystals for identification tests may be obtained with aminoacetic acid, betaine, glutamic acid, urea, acetamide, etc., as well as with many sympathomimetic drugs and other substances. Charles C. Fulton Electron microscopy AEROSOLS CONTAINING RADIOACTIVE PARTICLES The electron microscope is an ideal tool for the analysis of aerosol particles. This per- tains especially to particles with submicronic dimensions down to 0.002 micron (2 X 10"^ cm). Electronoscopic observations provide data on size, shape, aggregation tendencies and population density of the particles. In addition it may be possible to obtain electron diffraction data as a means for chemical identification (1). Correlation of electrono- scopic observations with data obtained by other analytical methods makes possible a complete description of a particular aerosol. The purpose of this article is to describe the electron microscopic appearance of particles from aerosols containing Sr^'^S04 , Ru^'^'Oa , or Pu23902 . Methods Particles from the various aerosols were collected directly on "Formvar" coated elec- tron microscope supporting grids or on mem- brane filters. In the latter case the particles had to be transferred to coated grids before observation in the electron microscope was possible. This was accomplished by a modi- fied Kalmus technique (2). Each aerosol was represented by an appropriate number of specimen grids (minimum of six grids per aerosol) to provide a statistically valid sam- ple of particles. All specimen grids used were preinspected for cleanliness, or pre-shadowed to delineate contaminations. Screens representing the various aerosols were surveyed in the electron microscope and appropriate fields were photographed at magnifications of 2,000X, 6,500X, and/or 10,000 X. The electron micrographs illus- trating this report are photographic enlarge- ments. Size distribution data were obtained by measuring several hundred particles chosen at random on the prints representing the different samples. The particles were all measured in the same direction. Observations and Discussion Description of the Particles. Strontium Sulfate. The particles obtained from aerosols containing Sr^''S04 are characteristically in the form of needles. Tj^pical examples are illustrated in Fig. 1. For the most part the needles occur in clusters. The dark cuboidal or spherical material noted in these micro- graphs may be undissolved membrane filter 72 AEROSOLS CONTAINING RADIOACTIVE PARTICLES Fig. 1. Particles from an aerosol containing Sr*'S04 . 1. Particles transferred from membrane filter to a "Formvar" -coated screen. 80OOX. The broad gray band in the lower left corner is a common contaminant in this type of preparation. 2, 3 and 4. Selected fields illustrating particles collected directlj^ on "Formvar"-coated screens. 25,000X. from which the particles were transferred, or they may be small particles of Pluronics, a dispersing agent present in the aerosol gener- ation suspension. The individual needles are rarely longer than 1 micron, or wider than 0.05 micron. Ruthenium Dioxide. Particles obtained from aerosols containing Ru^''^02 are illus- trated in Fig. 2. The particles are character- istically three-dimensional chain aggregates, each aggregate consisting of many roughly spheroid unit particles. Ruthenium dioxide 73 ELECTHON MICKOSCOl'Y Fig. 2. Particles from an aerosol containing Ru'''^02 . 1. Particles transferred from membrane filter to a Formvar coated screen. 6,375X. The aggregate particles are slightly fused because of exposure to the electron beam in the electron microscope. 2. Three-dimensional chain aggregate particle. 14,450X. This preparation was shadowed with palladium. Negative print. Note the spheroid nature of the particles making up the aggregate. 3. and 4. The same field before and after pro- longed exposure to the electron beam. 6,375X. The arrow in figure 3 points to a large ag- gregate particle. The arrow in figure 4 points to the same particle after it was exposed to the electron beam for 60 seconds. Such behavior is common for ruthenium dioxide par- ticles. The aggregate particle is too large to dissipate the heat evolved liy the impact of the electron beam on the specimen, and the melting point is low enough to cause fusion of the small particles into the resultant sphere. 74 AEKOSOLS CONTAINING RADIOACTIVE PARTICLES Fig. 3 The large particle (a) in Fig. 3 is about 0.65 /x across and about 0.65 ^ high d/m = 0.38 if particle is Pu()2 with a S.G. 11.44. Particle (b) is appro.xi- mately 0.05 m- The aggregate particle (c) consists of a countless number of grains 0.05 fj. and less. (About 4500 X.) Fk,. 1 In Figure 4, the large particle (a) appears to be made up of several cubes and a brick shaped particle. The volume of this particle can be ap- proximated very roughh'. Particles like these are frequently observed. This particle illustrates the discrepancy between actual or real volume and calculated volume based on a single linear meas- urement. The actual volume is roughly guessed to be about 7 >x^. The calculated volume using the dimension shown is about 46 m'- If this particle is sensitive to the electron beam, that is, when exposed to high beam intensity the ag- aggregates meh and fuse to form a single sphere. Large aggregates are more sensitive to the electron V)cam than small aggregates. Plutonium Dioxide. Particles from Pu-'^02 containing aerosols are characteristically- cubic or brick shaped. These are illustrated in Figs. 3, 4, 5, 6, and 7. Unhke Sr^^SO^ and Ru^oeQo which occur predominantly as ag- gregate particles, Pu-^^Oo occurs predomi- nantly as individual non-aggregated parti- FiG. 5 Figure 5 shows a particle that appears to be almost a perfect cube approximately 0.4 /x on a side. Volume is therefore about 0.064 ju'. Particles of this shape are commonly found in these sam- ples. If this particle is PuOs (S.G. 11.44) the dis- integration rate is about 0.09 d/m. The halo sur- rounding the particle is probably the remains of moisture that were associated with the particle. The round globule just outside the halo is an arti- fact— namel}' a bubble of carbon produced in shadow casting. (About 15,000X.) Fig. 4. — Continued represents a PuOo particle (S.G. = 11.44) the actual disintegration rate is somewhere near 10 d/m, whereas on a calculated basis the activity density would lu' aljout 62 d/m. This micrograph also illustrates the variation in jiarticle size en- countered in these samples. Compare particle (a) 3.55 M across with particles labelled (b) 0.05 ix across. (About 4500 X.) 75 ELECTRON MICROSCOPY luu. 6 Figure 6 shows another characteristic shape (brick-shaped) of particles found in this sample. (About 15,000 X.) Fig. 7 In Figure 7 are shown particles that are ob- served occasionally. This particle is surrounded by a halo indicating moisture was associated with this particular particle. The particle is an aggregate of different sized particles standing one on an- other. The variation in width of the particle is manifested by the difference in width of the shadow. The widest portion of this particle is about 0.4 n and its height is about 1.75 y.. (About 15,000X.) cles. Aggregate particles of Pu-^^O-) rarely consist of more than 5 or 6 cubes. The particles shown in Figures 3, 4, 5, 6, and 7 were collected directly on tungsten ox- ide and carbon preshadowed specimen screens. The specimens were shadowed again with chromium at a 30° angle before exam- ination in the electron microscope. Double- shadowed particles therefore would indicate contaminants on the specimen grid prior to exposure to the aerosol. Physical Data. Pertinent information on the physical characteristics of the particles from each of the aerosols is presented in Table 1 and Figure 8. The data and observations presented above are very general in that they were based on measurements of 100 particles rep- resenting a single aerosol. However, analysis of other aerosols containing Sr^''S04 , Ru^"®- O2 , or Pu-^^02 gave similar results. Al- Table 1. Physical Characteristics of Particles from Aerosols Containing Sr^oSO* , RU106O2 , OR Pu^^Oa Aerosol Contains Size Range (m) Mean Size ±S.D. Per Cent 0.5 m or less Remarks Sr'<'S04 0.05 0.38 74 Particles - 1.30 ± 0.22 are nee- dles or needle clusters. RuiosQa <0.05 0.36 75 Particles - 1.30 ± 0.29 are three dimen- sional chain ag- gregates of small spheriod type par- ticles. PU"902 0.05 0.20 99 Particles - 0.60 ± 0.09 are cubic or brick- shaped 76 i I BLOOD [Z~\ PARTICLES FROM AEROSOL WITH Sr^SO^ ^^ PARTICLES FROM AEROSOL WITH Ru'^^Oj {■ PARTICLES FROM AEROSOL WITH Pu^'^O, 1 050 060 070 SIZE IN MICRONS Fig. 8. Particle size distribution though the size range and mean size for the particles from the different aerosols are similar, there are striking differences in shape and aggregation tendencies. REFERENCES 1. KuMAi.MoTOi, "Encyclopedia of Microscopj^," 1961. 2. BORASKY, R., AND Mastel, B., AEC R + D Report, No. H.W. 46722 General Electric Co., Richland, Washington, 1956. 3. Fitzgerald, J. J. and Detwiler, C. G., KAPL-1088, General Electric Co., Schenec- tady, New York, 1954. R. BORASKY BLOOD* This method is designed to involve the least possible technical manipulation of the blood sample both before and after fixation. The sample was centrifuged to concentrate the buffy coat, thereby obtaining minimal contamination by erythrocytes. No antico- agulant or any other foreign substance was added to the blood prior to fixation. About 6-7cc of blood was obtained by venipuncture either (a) by withdrawal with a lOcc syringe fitted with a 20-gauge needle * (Excerpt of fixation, preparation, obtaining of, and microscopy of specimens taken from "Elec- tron Microscopic Atlas of Normal and Leukemic Human Blood") and transference to a lOcc Lusteroid centri- fuge tube (International), precooled to 5-10° C; or (b) by needle drip directly into the tube. The syringe had been previously silicon-coated with Dow-Corning 200, 2 per- cent in ecu , by immersion and baking for 3^-1 hour at 450-550° C. The needle had been coated with 10 percent aqueous Armour Monocote [tris-(2-hydroxyethyldodecyl)- NH4CI] by immersion, draining, and air drying. The ice-cooled sample was then centrifuged at 1500 rpm for 15 minutes at 0° C (relative centrifugal force — 265; Inter- national model PR-2, refrigerated, angle head centrifuge). The buffj- coat was aspi- rated with a silicon-coated pipette and trans- ferred to a glass tube containing 5cc of 1 percent Veronal-buffered (pH — 7.4) OSO4 at 5-10° C. It was fixed for }^-l hour, usually the former. Between the successive steps (3-^-1 horn-) of fixation, dehydration, and methacrylate infiltration, the specimen was centrifuged for 1-lH minutes at 1500 rpm (relative centrifugal force — 385; Claj^-Adams Safeguard centrifuge) in glass tubes (alcohol dissolves Lusteroid!). After each centrifuga- tion the supernatant fluid was decanted, the next fluid added, and the tube manually agitated to produce a suspension. The last methacrylate suspension (6 parts n-butyl, 1 part methyl) was permitted to settle by gravity in 00 gelatin capsules for 3'^-l hour 77 Eosinophil Fig. 1. (a) Cell of normal V)lood It Lymphocyte Fig. 1. (h) Cell of normal blood 78 BLOOD ^:^«r Neutrophil Fig. 1. (c) Cell of normal blood Monocjte I Fig. 2. Cell of normal blood 0 >.- %> 79 ELECTRON MICROSCOPY V " *-vt fit -■• ■Hi .-i^il^^ M- «« J5 Neutrophilic Promyelocyte Fig. 3. (a) Cell of granulocytic leukemia V ^ E:!^^ aitj* Neutrophilic metamyelocyte Fig. 3. (b) Cell of granulocytic leukemia Neutrophilic myelocyte Fig. 3. (c) Cell of granulocytic leukemia to avoid close packing. This also eliminated bubble formation during polymerization, which was performed overnight at 47° C with dry heat in an oven. Sections were cut on a Porter-Blum ultramicrotome using a glass knife and were mounted on copper grids covered by Formvar or carbon mem- branes. Three RCA electron microscopes were used for viewing and photography — an EMU-2, an EML-IB, and an EMU-3. The micrographs were taken on 2 by 10 inch or 3M by 4 inch Kodak lantern-slide medium plates and were printed by projection en- largement. Neither the negatives nor prints were retouched. James A. Freeman BOTANICAL APPLICATIONS Introduction. The high resolution ob- tainable with the electron microscope per- mits it to be used to great advantage for a mmiber of different problems in botanical 80 BOTANICAL APPLICATIONS Fig. 1. Section through part of a pollen wall (acetolyzed and chlorinated) of Rhododendron ponticum, XHflOO. (Afzelius, by courtesy of Grana Palynologica) research. The two main techniques employed have been thin sections and surface repUcas, the former being particularly valuable in plant cytology where internal detail is of interest, and the latter being mainly applica- ble to taxonomic and morphological studies. In addition it is possible to obtain informa- tion in a number of particular cases using direct examination. The various results ob- tained in different fields of botany will be briefly described here. Paly no logy. Pollen morphology has been fairly widely studied in the electron micro- scope. The results may be of interest in fields other than botany, for example, in studying the history of post-glacial flora by means of pollen analysis. In addition it may be of in- terest to those working on such problems as hay fever and asthma. Two specimen preparation techniques, namely sectioning and replicas, provide dif- ferent pictures of the sporoderm. Thin sec- tions provide a great deal of information concerning its stratification, but in order to obtain a complete picture of the sporoderm of a given pollen grain it is desirable to have a knowledge of the surface topography in addition to sub-surface stratification. Earlier work on pollen grains was confined to the study of thin sections in the electron micro- scope (1, 2, 3). It is difficult to prepare sections of pollen cell walls because they are extremely hard. However, advances in the technique of ultra- microtomy have permitted a considerable amount of information to be obtained in this way (4). The stratification of the sporoderm Fig. 2. Shadowed carbon replica of the surface of a fresh pollen grain of Rhododendron ponticum, X7000. is an extremely complex subject and cannot be discussed in detail here. An interesting study of both sections and replicas of pollen grains has been carried out by Miihlethaler (5), who interprets elec- tron micrographs of sections and carbon replicas (6) in terms of existing terminology on pollen morphology. An interesting com- parison between the electron micrographs obtained by different authors of the same type of pollen grain is shown in Figures 1 and 2. Figure 1 shows a section through part of an acetolyzed and chlorinated pollen wall of Rhododendron ponticum (4), taken by Afzelius. This can be compared directly with the carbon replica of Rhododendron ponticum taken by the author and shown in Figure 2. It can be seen that the section shows indica- tions of a surface structure which is similar in character to that revealed clearly in the replica. The potentialities of the electron micro- 81 ELECTRON MICROSCOPY scope compared with the Ught microscope in the study of pollen grains have been discussed by Bradley (7). For example, when studying pores, their outline can generally just be distinguished in the light microscope. With the electron microscope, however, the entire morphology is clearly resolved. A comparison between the pores of Plantago media and Plantago lanceolata is shown here. The pore of P. media (Figure 3) has a ragged outline and contains a number of irregularly scat- tered large protrusions; that of I\ lanceolata is circular and completely different in form. This difference can just be detected in the light microscope, but the true structures cannot be resolved. Surface replicas have indicated that the effect of the acetolyzation process on the sub-microscopic structure of the pollen grain surface is negligible. It might be expected that the use of powerful reagents such as those employed in acetolyzation would pro- duce artefacts in the sporoderm. This is not the case with pollen grains studied in the light microscope and there appears to be no Fig. 3. Shadowed carbon replica of a fresh pol- len grain of Plantago media, X9000. (Courtesy of the New Phytologist) Fig. 4. Shadowed carbon replica of a fresh pol- len grain of Plantago lanceolata, X9000. (Courtesy of the New Phytologist) noticeable effect at electron microscope levels of resolution. An important problem in the study of pol- len grains is the distinction between ap- parently identical grains of different species. If a separation of these species could be ob- tained it would be of considerable value in quaternary research. Preliminary electron microscope studies of Cannabis and Humu- lus, Coryhis and Myrica have not produced the distinction hoped for. In the former case Cannabis and Humulus appeared identical in the electron microscope with regard to their surface structures. However some slight but definite structural distinction was found between Coi'ylvs and Myrica grains. Rowley (8) has used both sectioning and surface replicas in an exhaustive study of the pollen wall in eleven species in the Com- melinaceae. No basic differences were found in the structural elements making up the mature pollen wall; morphological ^•aria- tion at light microscope level was due to variations in the arrangement of these elements. Rowley also studied the develop- 82 BOTANICAL APPLICATIONS ment of the pollen grain of Tradescantia pahidosa. It is of considerable interest that he found that the basic form of exine sculp- turing orginated very early in development. The intine was not recognizable until much later. The entire structure of the sporoderm, both internal and external, can be full}^ elucidated by the judicious employment of replicas and thin sections. Moss Spores. The spores of mosses and similar plants present a similar problem in replication and sectioning to pollen grains. Afzelius, Erdtman and Sjostrand (3) have studied the fine structure of the outer part of the spore wall of Lycopodium davatum using thin sections, the results indicating that the spore wall is divided into two layers, the outer being laminated and the inner granulated. Moss spores have not been studied ex- tensively, the only example being by Bradley (9), who shows electron micrographs of the surface structure of spores of Atrichum undu- latum and Dicranella heteromalla. It seems that the value of studying such specimens in the electron microscope is somewhat limited. Fungi. The direct examination of fungus spores of a number of different species was carried out by Gregory and Nixon (10). The spore structure is of interest in studies of asthma as is the case with pollen grains. Direct electron micrographs only provide a silhouette of the spores and little can be seen of their surface structure; the use of replicas, however, shows the surface structure clearly as in Figure 5. An interesting application of the electron microscope using both surface replicas and sections has been carried out independently by two authors on the division of Saccharo- myces cerevisiae (11, 12). The morphology of the different types of yeast bud scars and the mechanism of the division process was studied by replicas in the case of Bradley (12) and sections in the case of Agar and Douglas (11), both authors independently reaching similar conclusions. Algae. A group of algae which has been studied extensively in the electron micro- scope is that comprising the diatoms. So much work has been carried out that it is impossible to include more than a brief ref- erence. Much of the work was done in the earh^ days of electron microscopj^ and sev- eral important contributions were made by MuUer and Pasewaldt (13), Kolbe and Golz (14), Hustedt (15), and Hendey, Cushing and Ripley (16). The electron microscope is continually being used, generally as a tax- onomic aid in studies of new species or popu- lations. The light microscope is fully adequate for distinguishing the diatom genera, and also for separating the great majority of species. However, the electron microscope permits a much fuller examination of the submicro- scopic details of the silica valve and thus enables a far better understanding of the meaning of its finer visible features to be achie^'ed. In addition, the hope is that it will provide useful information about the de- velopment of this fine structure. The electron microscope has also been used in the study of one family and three genera of algae belonging to the Chryso- FiG. 5. Shadowed carbon replica of a spore of the fungus Russula mairei, X9000. 83 ELECTRON AllCHOSCOI'Y Fig. 6. Shadowed carbon replica of the calcite scales of a species of the family Coccolithophorida- ceae, X 10,000. phyccae. The first of these to be described here, CoccoHthophoridaceae, is of wide interest since the unicellular organisms form minute calcite scales. These scales form sedimentary chalk deposits. They are deposited on the ocean bed after the cells have died and the protoplasm has disintegrated. There is a very large number of species in the family and classification is a difficult matter when the light microscope is used because of the lim- ited resolution available. However, the elec- tron microscope provides the increased reso- lution recjuired for a full taxonomic study and as a result much useful information has been obtained which is of particular value to both botanists and palaeontologists. As in many other cases, electron micros- copy of the scales (coccoliths) requires a detailed terminology which is provided by Halldal and Markali (17). These authors give a comprehensive survey of a large num- ber of species of the genus using direct examination in the electron microscope. Though much information can be gained by studying coccoliths directly, the results are much clearer if carbon replicas are prepared (18). Figure fi shows a carbon replica of the calcite scales and the full surface topography is clearly shown. The remaining two genera studied in de- tail in ihe electron microscope are Synura and MaUomonas. These are closely related to the CoccoHthophoridaceae , bvit the latter are marine organisms whereas the former are fresh-water organisms. Synura and Mal- lomonas are also unicellular and covered with scales, but these arc generally much smaller and are composed of silica. The genus Synura contains only a few species. These have been studied in detail in the electron microscope by Manton (19), Fott (20), Petersen and Hansen (21) and Harris and Bradley (22, 24). The latter two authors concentrated on their taxonomy us- ing the electron microscope, but Manton studied internal morphology and employed thin sections. The genus MaUomonas contains a much larger number of species, many of which have only been discovered recently. The elec- tron microscope has been of great value as a taxonomic aid, since the classification of the genus depends almost entirely on the struc- ture of the minute silica scales covering the organisms. Harris and Bradley (23, 24), also Harris (25), have studied the scales directly and used carbon replicas to show up their fine structure. Asmund used direct examina- tion and has studied the occurrence of Mal- lomonas species in Danish ponds (26). Although the cells of many MaUomonas species disintegrate when dried, most of them become sufficiently rigid after careful fixa- tion to be studied complete in the electron microscope. Figure 7 shows a direct electron micrograph of a scale of a species of Synura and Figure 8 shows a carbon replica of a complete MaUomonas. A small genus, ChrysosphaereUa, about which relatively little is known, has also been studied in the electron microscope. It is rather similar to Synura and only consists of two or three species. The cells are also coated with silica scales and have long spines at- 84 BO r A\ IC AL APPLICATIONS Fig. 7. Direct electron micrograph of scales of Synura eichinulata; the structure at the base of the spine is internal thickening, X 13,500. t ached to them. ChrysosphaereUa is rela- tively uncommon compared with Mallo- monas and Synura. Petersen and Hansen (27) have studied some organisms associated with the surface of the water as opposed to those types of phyto-plankton which are free-swimming. By means of a special technique they were able to study single cells as they were situ- ated on the water surface before drying. Marine algae have been studied in detail by Parke, Manton and Clarke (28, 29) who used direct examination and sections. These authors are concerned mainly with the micro-anatomy and taxonomy of Chryso- chromulina. Their descriptions are extremely detailed and informative. Manton and Clarke (30) have also given a detailed and interesting description of the spermatozoid of Fucus serratus. This inter- esting organism is shown in Figure 9. Man- ton and Clarke ascertained by comparing UV micrographs and electron micrographs that the body shrinks but does not alter its shape in the electron microscope. The func- tion of the fine hairs on either side of the front flagellum is not known. The proboscis, which is highly mobile in the living state, is a fimnel-shaped membrane surrounding the front flagellum and attached to the body at the base. When dry, the organ is flattened and thirteen characteristic concentric thick- enings can be seen. The study of the flagella is particularly interesting since the morphol- ogy can be compared with other flagella and cilia. Electron micrographs of the disinte- grated front flagellum show it to be com- posed of eleven strands; in the rear flagel- lum there are only nine. It is interesting to note that fern cilia bear a close numerical relationship (31) yet there is no phyletic or Fig. 8. Shadowed carbon replica of a complete cell of Mallomonas coronata, X9000. (Courtesy of Research) 85 ELECTRON :\IICKOSCOPY Fig. 9. Direct electron micrograph of a sha- dowed spermatozoid of Fucus serratus; the struc- ture of the proboscis is particularly well shown, X 18,750. {After Manton and Clarke, courtesy of the Annals of Botany) structural relationship with broAvn algae. Eleven strands can also be found in ani- mals, for example, Paramoecium (32), the sperms of domestic fowl and fish. The prevalence of the number eleven suggests some fundamental property in the geometric relations of fibres. Bacteriology. The electron microscope has been used extensively in the study of bacteria and related organisms. The de- velopment of the thin sectioning technique has permitted bacteria to be sectioned and internal structiu'es to be examined in great detail. In general the study of the surfaces of bacteria using replica techniques is not particularly rewarding, but in a few cases it has been possible to obtain useful informa- tion in this way. The study of bacterial cytology is a complicated and controversial subject which cannot be described in any detail here. Much of the controversy arises from interpretations of electron micrographs of thin sections of bacteria. The appearance of sections of bacteria using different types of embedding and staining techniques is always at variance. The recent development of the use of epoxy resins for embedding specimens prior to sectioning (33) has indi- cated that some of the previous work using methacrylate embedding materials is sus- pect. There is no doubt that the wealth of information now available on this subject is becoming much more co-ordinated. The use of surface replicas in bacteriology has generally been connected with taxonomic studies such as in the case of the genus Bacillus (34). Here it was shown that the surface sculpturing of spores was different in different species, so that once again the electron microscope proved to be a valuable taxonomic aid. Plant Cytology. Plant cytology now covers a wide field. The electron micro- scope has been used in the study of cell walls, mitochondria, chromosomes and other cyto- plasmic inclusions. The specimen techniques required are variable, much of the work being carried out using thin sections, but direct examination and replicas of cell walls have provided much information on their structure. Detailed general reviews of the electron microscopy of the plant cell have been given by Miihlethaler (35) and Buvat (36) and some of the more important findings are described here. The Cytoplasm. The structure of the cytoplasm varies according to the fixing agent and may appear as granular or in the form of a fine network. It seems probable that the reticulate structure does not corre- spond to the living state. Studies of the distribution of the various albumens and nucleic acids in the cytoplasm have been attempted by forming heavy metal complexes to act as specific electron stains (37), Strugger (38) combined OSO4 fixation with uranyl-acetate treatment and 86 BOTANICAL APPLICATIONS was able to resolve sub-microscopic filament- on their development (39, 40). The various like elements. stages in the development of the barley Plastids. These have received a great deal chloroplast are shown in Figure 10. Firstly of attention, much work being carried out (1) a proplastid develops in the leaf meri- FiG. 10. A diagrammatic representation of the barley chloroplast (see text). (After Diter von Wettstein, courtesy of Hereditas) 87 ELECTRON MK.HOSCOPY Fig. 11. Section through a proplastid from Be- gonia, X25,000. (By K. Muhlethaler) Fig. 12. Section through a further developed proplastid from Begonia, X 25,000. {By K. Muhle- thaler) Fig. 13. Section through a chloroplast of Elo- dea canadensis, X 16,000. (By K. Muhlethaler) Stem. This shows more or less the same struc- lurc as the mitochondria. Next (2) the size iiicroasos and the plastid center, containing a tul)ular structure, is formed. Starch grains now appear (3), and then the material for the formation of the layer structure pro- trudes radially from the center (4). The starch breaks down (5) and the lamellae l)egin to form. They multiply by thickening and splitting until the continuous lamellar structure of the chloroplast (G) is formed by the fusion of short lengths. The resulting plastid is traversed by continuous double lamellae (7). Finally further splitting forms the grana of the fully differentiated chloro- plast (8). It is interesting to compare Figure 10 with the electron micrographs of Miih- lethaler showing plastid development in Begonia and the chloroplast of Elodea canadensis (Figures 11-13). The proplastid from Begonia (Figure 11) is generally similar to the drawing of the barley proplastid (Figure 10) and the further development of the Begonia plastid (Figure 12) is like stage (4) of the description. The fully-differ- entiated chloroplast of Elodea canadensis (Figure 13) is also generally similar to that of barley. From this it may be inferred that the general pattern of chloroplast develop- ment is similar in different species. It is not possible to study the structure of the chloroplast in detail here. The structure is generally similar in all plants, but there is much variation in the dimensions and spacings of the lamellae. The chloroplast is the site of photosynthesis, the chlorophyll being concentrated in the grana. The grana are distributed in the stroma which forms the body of the chloroplast. The Nucleus. Sections through the nu- cleus show chromosomes in the despiralized condition and with good resolution a fine granular structure can be detected in the chromosomes and nuclear cytoplasm. How- ever, the electron microscope has added little to our knowledge of chromosomes. I 88 BOTANICAL APPLICATIONS Mitochondria. The electron microscope shows that the mitochondria are funda- mentally different morphologically from other cell particles. As in the case of animal mitochondria, those of plant cells show a double-membrane system. Within the mito- chondrion are complex fine structures, the cristae mitochondriales, which consist of in- FiG. 14. Shadowed portion of primary cell wall Valonia, X5500. (After Steward and Miihlethaler, courtesy of the Annals of Botany) vaginations of the mitochondrial membrane reaching from one wall to the other. These structures are highly variable from cell to cell and between different species of plants. It is now known with reasonable certainty that the mitochondrion is the site of respira- tory activity of the plant. The Cell Wall. Many studies of the pri- mary cell wall have been obtained by macer- ating the cells and reducing them to a very thin film. These thin films are in the form of sub-microscopic strands of cellulose which in turn can be split into even finer threads by means of such techniques as ultrasonic irra- diation. The micro-fibrils obtained in this way are elementary in nature because they correspond to the crystalline micellar strands which can be detected by means of x-ray diffraction. It can be seen in Figure 14 that the micro-fibrils of the primary cell wall are interwoven to form a very dense network. In the secondary cell wall Avhich consists entirely of micro-fibrils of cellulose, the strands tend to lie parallel instead of in the form of a network as shown in Figure 15. In successive lamellae the orientation of the parallel fibrils is shifted. Fig. 15. Shadowed portion of secondary cell wall of Valonia, X11,0()0. {After Steward and Miihlethaler, courtesy of the Annals of Botany) 89 ELECTRON MICROSCOPY Fig. 16. Pit membrane in a simple pit in the radicle of maize, X 15,000. (After Milhlethaler, courtesy of Die Naturwissenschaften) It is possible to study changes in the pri- mary cell wall structure during cell division. The manner in which the cellulose micro- fibrils are deposited can be examined. The secondary cell wall contains characteristic perforations known as pits. These pits have been studied extensively in the electron microscope and whereas optical evidence suggested that the pit membrane, which stretches across the perforation, acts as an impassable barrier, the electron microscope indicates that it is porous in nature (Figure 16). Cell wall growth has been studied in detail in the onion root tip by Scott et al. (40). Conclusion. The electron microscope is clearly very valuable in botanical research. It is extremely useful as a taxonomic aid, and much morphological information has been obtained in the field of plant cytology. It remains for these results to be correlated with biochemical investigations. Acknoivledgments. The author would like to thank the following for material and advice: Professor and Mrs. T. M. Harris, University of Reading; Dr. K. Miihlethaler, Eidgenossische Technische Hochschule, Zurich; and Dr. B. E. Juniper, University of Oxford; also Dr. T. E. Allibone, F.R.S., Director of the Research Lab- oratory, for permission to publish this article. REFERENCES 1. Fernandez-Moran, H. and Dahl, A. O., Science (Lancaster, Pa.) 116, 465 (1952). 2. MiJHLETHALER, K., Mikroskopie, 8, 103 (1953). 3. Afzelius, B. M., Erdtman, G., and Sjos- TRAND, F. S., Sv. Bat. Tidskr., 48, 155 (1954). 4. Afzelius, B. M., Grana Palynologica, 1, 22 (1956). 5. MtJHLETHALER, K., Platita, 46, 1 (1955). 6. Bradley, D. E., Brit. J. Appl. Phys., 5, 96 (1954). 7. Bradley, D. E., New Phytol., 57, 226 (1958). 8. Rowley, J. R., Grana Palynologica, 2, 3 (1959). 9. Bradley, D. E., Mikroskopie, 13, 180 (1958). 10. Gregory, P. H. and Nixon, H. L., Trans. Brit. Mycol. Soc. 33, 359 (1950). 11. Agar, H. D. and Douglas, H. C., J. Bac- teriol., 70, 427 (1955). 12. Bradley, D. E., J. Roy. Micros. Soc, 75, 254 (1956). 13. MtJLLER, H. O. AND Pasewaldt, C. W. a., Natunviss., 30 (1942). 14. KoLBE, R. W. AND GoLZ, E., Ber. dtsch. hot. Ges., 61, 91 (1943). 15. HusTEDT, F., Arch, hydrobiol. Plankt., 41, 315 (1945). 16. Hendey, N. I., Gushing, D. H. and Ripley, G. W., /. Roy. Micros. Soc, 74, 22 (1954). 17. Halldal, p. and Markali, J., Avhandlinger Utgitt av Det Norske Videnskeps-Akademi, Oslo. Mat-Natruv. Klasse, No. 1 (1955). 18. Bradley, D. E., J. Appl. Phys., 27, 1399 (1956). 90 CELL ULTHVSniUCTLRE I\ MAMMALS 19. Manton, I., Proc. Leeds Phil. Soc, 6, 30G (1955). 20. FoTT, B. AND LuDviK, J., Prcslia, 29, 5 (1957). 21. Petersen, J. B. and Hansen, J. B., Biol. Medd. Dan. Vid. Selsk., 23, 1 (1956). 22. Harris, K. and Bradley, D. E., Discovery, 17, 329 (1956). 23. Harris, K. and Bradley, D. E., J. Roy. Micros. Soc, 76, 37 (1957). 24. Harris, K. and Bradley, D. E., ./. gen. Microbiol., 18, 71 (1958). 25. Harris, K., J. gen. Microbiol., 19, 55 (1958). 26. Asmund, B., Dansk. Bat. Arkiv., 18, 7 (1959). 27. Petersen, J. B. and Hansen, J. B., Saertyk. Af. Bot. Tid., 54,93 (1958). 28. Parke, M., Manton, I. and Clarke, B., /. Mar. Biol. Assn. U.K., 35, 387 (1956). 29. Parke, M., Manton, I. and Clarke, B., /. Mar. Biol. Assn. U.K., 37, 209 (1958). 30. Manton, I. and Clarke, B., Ann. Bot., 15, 461 (1951). 31. Manton, I. and Clarke, B., /. Exp. Bot., ii, 125 (1951). 32. Jackus, M. a. and Hall, C. E., Biol. Bull., 91, 141 (1946). 33. Glauert, a. M., Nature, 178, 803 (1956). 34. Bradley, D. E. and Franklin, J. G., J. Bac- terial., 76, 618 (1958). 35. Muhlethaler, K., Naturwiss., 44, 204 (1957). 36. Buvat, R., Ann. des Sci. Nat. Bot., 11th Series, 19, 121 (1958). 37. Lamb, W. G. P., Stuart-Webb, J., Bell, J. L. G., BovEY, R., and Danielli, J. F., Exp. Cell Res., 4, 159 (1953). 38. Strugger, S., Naturwiss., 43, 357 (1956). 39. Muhlethaler, K., Private communication. 40. DiTER von Wettstein, Hereditas, 43, 303 (1957). 41. Scott, F. M., Hanmer, K. C, Baker, E. and Bowler, E., Amer. J. Bot., 43, 313 (1956). D. E. Bradley CELL ULTRASTRUCTURE IN MAMMALS The term "ultrastructure" refers to the fact that the cell structures accounted for here have been analyzed with the aid of the electron microscope. The finer details of cell structures cannot be resolved with the light or phase contrast microscopes and, there- fore, have been considered as being beyond or ultra this level of resolution. Methods The techniques applied in modern electron microscopy are found elsewhere (electron microscopj^: specimen preparations). It should be emphasized, however, that thin sectioning (section thickness about lOOA) is required to permit the best resolution. Several specially designed microtomes are commercially available such as the Porter- Blum (U. S. A.), the Sjostrand (LKB, Sweden), the Sitte (Reichert, Austria), the Moran (Leitz, Germany), the Hanstra (Philips, Holland). The microtome preferred by the author is the LKB Ultrotome (Swe- den) which is the most recent and most superior in design. The preservation of the tissue has been thoroughly worked out by Dr. Palade (U. S. A.) and Dr. Sjostrand (Sweden). It is believed that any analysis of the cell ultra- structure today can be made without the risk of describing artifacts because so much is known about how various factors may influ- ence the cell structures: the tonicity and acidity of the fixative, the temperature, post mortem changes, etc. The contrast of the electron microscopic picture (electron micrograph) can be en- hanced by applying special staining tech- niques, either during the fixation or dehy- dration periods or after sectioning. The fixative itself gives, however, cjuite sufficient contrast for most studies. The most com- monly used is osmium tetroxide (often called osmic acid, although it is not an acid). After fixation, the specimen is dehydrated in graded alcohols and subsequently embedded in liquid plastics (usually a mixture of butyl and methyl methacrylates). The monomers are polymerized and the embedded tissue block is then ready to be sectioned. Cell Shape The general shape of cells varies from tis- sue to tissue (Fig. 1). This has been ac- curately analyzed with the light microscope and very little has been added to previous 91 ELECTROIV MICROSCOPY Fig. 1. Plasma cell in loose connective tissue of the human epididymis with most of the features present which usually are found in a mammalian cell: nucleus (N), nucleohis (Ne), Golgi zone (G), mitochondria (M), ergastoplasm (E). The cell is freely suspended in between the connective fibers and displays a number of surface extensions (S). Magnification 10,500X descriptions by applying electron micros- ruled out, as for example in the epidermis copy. However, the relationship between and in the heart muscle. Only in rapidly cells has been elucidated clearly. The old dividing cells, as for instance in the testis, conception of cells being connected by inter- can one find two or more cells interconnected cellular bridges to a syncytium has been at a certain stage of their differentiation. 92 CELL ULTKASTRUCTURE L\ MAMMALS Nucleus ranged in a layer outside the proteins. Also As a rule, there is only one nucleus in each ^ mosaic arrangement of tlic lipid and pro- cell. The striated muscle cell is an exception ^^"^ molecules has been suggested. The ap- and may contain several nuclei. A double- P'^'"^'"^ uncertainty is explained by the fact contoured membrane surrounds the nucleo- ^^'^^ ^'^^y ^^^^^^ ^^ known about what struc- plasm with a total thickness of about 250A ^^"'^ ^^ stained most intensely with osmium (Fig. 3). Discontinuities have been demon- tetroxido the proteins or the lipids, strated in the nuclear membrane, reminiscent ^*'*^^ Surface. The plasma membrane on of pores. There does not seem to be a free ^^^ ^^'^® surface of cells shows four types of communication between the nucleoplasm differentiation— microvillus, brush border and the cell cytoplasm, however, because extension, stereocilium, and cilimu. the "pores" appear to be plugged by a dense Microvilli. The microvillus is essentially a substance of unknown nature. The structure ^^^/'^ ^"^^ ^'^"^ projection of the cytoplasm of the nucleoplasm or chromatin is finely '^^'hich is covered by the plasma membrane, granulated. The granules have a diameter of ^^^ micro\'illus does not contain any pecu- about 250A and are clustered in a zone ^^^^' structures but a slightly dense ground adjacent to the nuclear membrane but can, substance. The free surface of most epithelial in addition, be seen distributed evenly cells does display a varying number of micro- throughout the nucleoplasm. The nucleolus ^^^ ^^'^^^ ^ g^eat variation in length and represents a heavy aggregation of these thickness (Fig. 4). Supposedly, the microvilli granules (Fig. 1). Very little has been done ^^^^ resorptive functions and certainly so far on the ultrastructure of the chromo- contribute to the increase of the cell surface, somes. Brush border extensions. The brush border extensions are longer than the microvilli, Centriole mostly thicker and occur in greater abun- The cell center or the centriole (centro- dance. They all are of the same length and .some) is located in the neighborhood of the are found on the surface of the intestinal nucleus, mostly within the Golgi zone of the epithelial cells (Fig. 5) and on the proximal cell. The centriole is a round or slightly convoluted tubule cells of the nephron (cross elongated body with size and structure es- reference: kidney ultrastructure). Similar sentially similar to the basal body of the structures (Fig. 8) are also seen in the ef- cilium (cross reference: ciliated epithelia) ferent ducts of the testis (cross reference: which represents a dense cortex and a lighter ciliated epithelia ultrastructure). Although core. The cortex is composed of nine paired essentially representing extensions of the filaments and a matrix which embeds the apical cell cytoplasm, the brush border ex- filaments. The core is structureless (Fig. 2). tensions do contain some fine and dense _, ^w , striations oriented longitudinally, sometimes Flasnia Membrane t. a- i ^^i "^i ^i extendnig down nito the apical cytoplasm The plasma membrane is the outermost below the level of their bases. Histochemical limit of the cell. It has an average thickness tests seem to prove that the enzyme alkaline of 70-100 A and stands out as a single dense phosphatase is associated with the brush line in the electron micrograph. It is com- border extensions. posed of lipid and protein molecules, their Stereocilia. Stereocilia are longer and mutual arrangement remaining unknown. It narrower than the brush border extensions, has been assumed that the dense line seen in but the lumibcr of stereocilia on each cell is the electron micrograph may represent the about identical with that of the brush border protein layer. The lipids would then be ar- extensions. Each stereocilium contains three 93 ELECTRON MiCHOSCOl'V W^' lA ms M 1 '» 4P ^r^*^ Gr Go \ •; .» CS' ur II * »n^.. Fig. 2. Detail of a plasma cell (human epididymis). The nucleus (N) is enveloped bj' a triple-layered membrane. The Golgi Zone (Go) has both lamellar and vesicular components. In addition, dense large granules (Gr) are seen, each surromided by a single membrane. In the center of the Golgi area is the centriole (CS) cut at an angle through its fibrillar components. Above the mitochondrion (M) is the ergastoplasm (granular endoplasmic reticulum) E, with its membrane bound flat cisternae and at- tached RNA granules. At ms the cisternae are more spherical. This shape is predom- inant after cell fractionation and centrifugation. These rounded structures then cor- respond to the microsome -fraction of the cell. The cell border is seen at P with some collagenous fibers in the interstitial space. Magnification 33,000X 94 CELL ULTKASTKUCTIRE L\ MAMMALS Fig. 3. Detail of a proximal tubular cell of the mouse kidney. A number of struc- tures are seen which can be encountered in most mammalian cells. The nucleus (N) with its triple-layered envelop. Two mitochondria are present, one sectioned longi- tudinally (MI), the other cross cut (M2). In addition to several microbodies (m), Golgi membranes and vesicles (Go) a dense large body (D), may be seen. Magnifi- cation 67, OOOX to five longitudinal fine filaments. In man, mediate stage between the ordinary brush the stereocilia are found in the duct of the border extensions and the ciha. epididymis (Fig. 6). Their function is not Cilia. The ciha are extremely long ex- clear because they do not have any con- tensions of the apical cell cytoplasm. They tractibility. They seem to form an inter- are coarser than the stereocilia and display 95 ELECTRON IVIICROSCOPY \ ..J^ l.u w I Fig. 4. Typical arrangement of microvilli (V) on the surface of an epithelial cell in the distal tubule of the mouse kidney. The microvilli are slender, short processes with seemingly poor rigidity, usually widely spaced. Mitochondria (M) are seen to- gether with a few small vesicles in the apical cytoplasm. Magnification 15,000X L 1,0 M Fig. 5. Brush border extensions of the mouse intestinal mucosa. Brush borders are closely packed, rather rigid surface processes. They display a higher density than the microvilli. Some mitochondria (M) and several absorbed lipid droplets (L) are seen. Magnification 28,000X 96 CELL ULTKASriU C: Tl KE L\ >L\MMALS 1% "^^ Fig. 6. Stereucilia on the surhice of epithelial ceils in the human ductus epididy- mis. The stereocilia are long and slender and lack basal bodies. They do not seem to have any mobility, but a few basal rootlets can be resolved. Magnification 10,800X a peculiar inner structure of fine longitudinal are all joined in the tip of the cilium and in filaments which are arranged in nine pairs the basal body below the cell surface (cross peripherally and two single in the center reference: ciliated epithelia ultrastructure). (Fig. 7). The filaments are contractile and Tubular invaginations. The ireeceWsuriace Fig. 7. Cilia on the surface of the cells of the rat trachea. Cilia are shorter and coarser than stereocilia and display distinct fine inner structures which terminate in the basal body (B). Magnification 31,500X 97 ELECTRON MICROSCOPY displays small tubular invaginations which membrane close to the free surface of all penetrate about half a micron into the cell epithelial cells (Fig. 8). In a sense, they are (Fig. 8). They are found mostly in cells with reminiscent of the coopering bands of a a high rate of fluid uptake. It has been as- barrel, although located inside the barrel, sumed that they represent another means by The terminal bar of one cell is always located which fluid and substances can be taken in opposite a similar strvicture in a neighbor- by a cell without first penetrating the plasma ing cell. Furthermore, the intervening space membrane. Their activity has been com- between the cells is filled with a denser struc- pared with the uptake of water by an ameba ture and is smaller than usually recorded in (pinocytosis). The word "micropinocytosis" other places, undoubtedly indicating the (or membrane flow) has been suggested, firmness by which the cells are attached, indicating that once the fluid or substance Desmosomes. The desmosomes are in a has been taken in by the microtubules, it section having almost the same appearance can be entirely surrounded by the tubular as the terminal bars. However, they repre- membrane, forming a vesicle which can be sent only local points of attachment and are transported elsewhere in the cell (cross actuafly paired button-like structures with reference: kidney ultrastructure, ciliated one ''button" attached to the intracellular epithelia ultrastructure). aspect of the plasma membrane of either Cell Border. The plasma membrane cell (Fig. 8). In addition, the intercellular which faces a neighboring cell is called cell space between the two "buttons" is larger border. than in other places and is occupied by a Lateral inter digitations. This portion of the substance of high density which frequently plasma membrane frequently displays small contains even denser structures of lamellated lateral projections of the cell cytoplasm nature. The desmosomes are most abundant which penetrate into the cell in juxtaposi- in the cells of the epidermis (Fig. 14) but are tion (Fig. 14). In some instances, the pro- also found in other epithelial cells. Struc- jections become quite numerous, as is the tures of identical appearance are also demon- case between the cells of the ciliary epithe- strated in striated muscle and in the specific hum of the eye or in the intestine. It was first tissue of the heart (the impulse conducting believed that the lateral interdigitations system). Besides their adhesive function in helped in maintaining cell cohesion; how- these tissues, they probably also serve as ever, it has been suggested recently that they points of less resistance across which the rather are the result of a certain compression impulse of contraction can travel with much or expansion during different functional higher speed than elsewhere, stages of the cells, much like what happens Basal Surface. The basal plasma mem- to the bellows of an accordion. The varia- brane which faces the basement membrane tion of volume does not occur in the cell is from time to time elaborately infolded, itself to any large extent, but rather to the Epithelial cells. This is particularly the intercellular space, which is expanded by case in the cells of the nephron, of the ciliary fluid from time to time. Cell cohesion is, on epithelium of the eye, and of the choroid the other hand, definitely accomplished by plexus of the ventricles of the brain. The two peculiar structures which are closely infoldings sometimes reach to the depth of related to the cell border — the terminal bars half the cell and may also be seen inter- and the desmosomes. digitating laterally with each other. Un- Terminal bars. The terminal bars are ring- doubtedly, they serve to increase the basal like reinforcements of the cell border at- surface of the plasma membrane upon which tached to the inner aspect of the plasma enzymatic activity can more readily occur. 98 CELL LLTKASTRUCTURE L\ MAMMALS Jk. .Xj Fig. 8. Detail of epithelial cells of the human ductus efferens (connection between the testis and the epididymis). The surface of the cell's is provided with brush border extensions (B) extending into the lumen (L) of the duct. Tubular invaginations (T) of the surface membrane descend into the upper part of the cells. The cell boundaries (CB) are held together by terminal bars (Tb) close to the surface, and by desmosomes (De) at lower levels. Except for mitochondria (M), these cells display large dense granules (D) as well as extremely osmiophilic bodies (P) , some of which may represent pigments, others lipid granules. A large vacuole (Va) is seen in the center. Magnifi- cation 26,000X 99 ELECTRON MICROSCOPY Nerve cells. The extremely elongated cyto- plasmic extension of a nerve cell is called the axon. The axon is covered by a number of Schwann cells along its course. It is the cytoplasm of the Schwann cell which wraps itself around the axon to form the myelin sheath. The axon and the myelin sheath to- gether represent the nerve fiber. The main component of the myelin sheath is the plasma membrane of the Schwann cell which builds up the myelin sheath in a varying number of layers. It is, therefore, justifiable to consider the myelin sheath as being a specialized infolding of the plasma membrane of the Schwann cell. Basement IVIembrane The basement membrane forms the struc- ture upon wliich most cells rest. It is a structureless layer with a thickness varying between 400 and 1000 A (Fig. 13). Accord- ing to histochemical tests, it does contam mucopolysaccharides but so far no peculiar ultrastructure has been detected in its homogeneous layer. The old conception of the basement membrane being a layer with a thickness of several microns in some tissues has been ruled out with the aid of the elec- tron microscope. The thick basement mem- branes seen in the light microscope contain, in addition to the just described homogene- ous layer, a number of fibrillar structures most of which are reticular fibers (Fig. 10) with some additional collagenous ones. It is not quite clear what kind of cell is responsible for the formation of the homogeneous base- ment membrane seen in the electron micro- scope. In some instances, it is surely laid down by fibroblasts and it should, therefore, be looked upon as being part of the connec- tive tissue. However, sometimes no fibro- blasts can be detected in adult tissue in connection with the basement membrane and it is, therefore, beheved that other cells may have the ability of laying down this struc- ture during their differentiation. Cell Organelles In the cell is recorded a number of or- ganelles some of which are large and have a definite form— rmioc/iondna, microhodies, large granules, and pigments. Other or- ganelles have more flexible form— Go/gi apparatus, vesicles, and ergastoplasm (rough surfaced endoplasmic reticulum). Within the homogeneous ground substance of the cytoplasm, as it was looked upon by means of light microscopy, structures have been detected with the electron microscope which are of rather small diameters, but which definitely should be listed among the cell organelles— /?A^A-6franw?es and glycogen gran- ^Jl PS Mitochondria. The mitochondria are dis- crete bodies within the cell. They may vary in number, size, and shape, but their ultra- structure is remarkably unchanged from cell to cell and from tissue to tissue. In the fight microscope, they can be selectively stained by Janus Green B, but even so, it is sometimes difficult to distinguish them from other granular structures in the cell. The mitochondria are surrounded by a double-contoured membrane, the thickness of which is in the neighborhood of 180 A. The matrix of each mitochondrion has a higher density than the surrounding cyto- plasm. The matrix itself is homogeneous or slightly granulated. It is traversed by a vary- ing number of double-contoured membranes (or cristae) which mostly are arranged parallel to each other. The inner mito- chondrial membranes (or plates) with a thickness of roughly 150A are always con- nected with the mitochondrial capsule for some distance, but there is no open connec- tion between the cytoplasm of the cell and the mitochondrial matrix (Fig. 3). Extremely electron-dense spherical bodies of different sizes are sometimes seen embedded in the mitochondrial matrix between the mem- branes. The mitochondria are the carriers of 100 CELL ULTKASnaCTLRE IN MAMMALS most cellular enzymes and one believes that the membranous struct m'es of their interior represent either the enzymes proper or the surfaces upon which the enzymes and the metabolites interact. Microbodies. The microbodies are spheri- cal and usually smaller than the mitochon- dria. They are surrounded by a single mem- brane, about 50A thick. The matrix has the same appearance as that of the mitochondria but lacks the inner membranes of the latter (Fig. 3). They have been demonstrated so far only in kidney, liver, cortical adrenal cells, and in the delta cells of the endocrine portion of the pancreas. They may represent the precursors of mitochondria although this has not been convincingly proved. It is still a mystery how mitochondria develop. Some people believe they can arise de novo from the ground substance of the cytoplasm, where possibly the microbodies would con- stitute an intermediate stage. Others con- sider a splitting or budding of already exist- ing mitochondria to be more likely, as is known to occur during cell mitosis. Large Granules. The large granules en- countered in cells of various tissues are usually very dense; this has led most in- vestigators to believe that they contain lipids. They have, therefore, often been called cell lipid granules. However, the den- sity varies from time to time and it is diffi- cult to predict if this is due to a certain func- tional variation or if it mainly reflects a difference in structures. Most granules in cells represent a secretory product and as such will be dealt with below (Secretion). The re- maining large granules are of two types - — spherical granules and granules with ir- regular outlines. The large spherical granules are surrounded by a distinct single mem- brane and display a medium dense structure- less matrix (Fig. 3). It is most likely that they represent end products of substances taken in by the cells by means of a micro- pinocytotic activity through the tubular invaginations of the plasma membrane. By a certain metabolic process, the engulfed fluid, and therein dissolved substances, become concentrated and now appear as dense granules. This is surely the case with macrophages and has also been dem- onstrated in connection with uptake of proteins by the proximal convoluted cells of the nephron and of the intestinal cells. The granules with irregular outlines most likely represent lipid granules which the cell handles as part of its metabolism (Fig. 8). Structural evidence is at hand for a certain interaction between the lipid granules and the mitochondria, as for instance in liver cells. The intense blacken- ing of lipid granules by osmium tetroxide indicates that these granules contain un- saturated sulfhydryl groups. Saturated fats do not take stain with osmic acid; this can be clearly demonstrated in fat cells where the fat globules are convincingly stained with Sudan III for light microscopy but in the electron microscope show up as un- stained vacuoles with a bordering thin membrane. Pigments. The pigments represent an- other type of spherical granule which can be encountered in a number of cells. Similar to lipid granules, they stain intensely with osmium tetroxide (Fig. 8). The pigments of the retina are surrounded by a single mem- brane. Their matrix is homogeneous. The pigments encountered in the cells of the epidermis (often called melanin granules) do not have a limiting membrane but unveil an abundance of small pigment micelles each of which has a diameter of about 75A. The pig- ments of the epidermis are supposedly formed within special cells, the melanocytes, and migrate into the basal cells of the epidermis. Golgi ApiJaratus. The Golgi apparatus is located near the nucleus mostly surrounding its one pole like a halo. It consists of a sys- tem of paired membranes, small vesicles and 101 ELECTRON MICKOSCOI'Y granules. The membranes have a smooth surface and enclose clear spaces. The num- ber and length of the membranes vary, presumablj'^ because of different functional stages of the Golgi apparatus. There are mostly several pairs of membranes arranged in parallel form with each other and one can occasionally see that the clear space which each pair of membranes encloses is distended to a vacuole of varying size (Fig. 3). The small vesicles are quite numerous and bordered by a smooth membrane. Fre- quently, the clear centers of the small vesicles become condensed, thus obtaining the appearance of small granules (Fig. 2). The diameter of the small vesicles and the granules ranges between 200A and lOOOA. The appearance of the Golgi apparatus as a whole varies greatly from cell to cell and from tissue to tissue. It is best developed in secretory cells where its function presimiably is involved in the secretory process. Evi- dently the secretion products are prefab- ricated elsewhere in the cell, but the end products appear as small secretory granules within the Golgi zone. Here they enlarge and migrate eventually to the upper part of the cell. In non-secretory cells, the Golgi ap- paratus presumably plays an important role in the metabolism of the cell, either by offering its membranes as surfaces for en- zymatic activity or by being utilized as a system of channels for the intracellular flow of metabolites and fluid. Small Vesicles. Vesicles of the same order of magnitude as those found within the Golgi zone may be traced elsewhere in the cell. Their origin is unknown but they could possibly be derived from the Golgi ap- paratus. In some instances, as in non-ciliated cells of the ciliated epithelia of the bronchi and bronchioles (cross reference: ciliated epithelia ultrastructure), in the distal con- voluted tubule cells of the kidney, and in Fig. 9. Surface area of a dark, intercalated cell of the collecting tubvile of the mouse kidney. The cytoplasm is pervaded by abundant microvesicles, one in connec- tion with the surface (at 9). Tiny microvilli (Vi) extend into the tubular lumen (Lu). In between the vesicles, which all are bound by a smooth membrane, are abundant granules (arrows) with the size of about 150A, probably corresponding to granules which in other cells have been demonstrated to contain RNA. Magnification 57,000X 102 CELL ULTRASTRLCTURE L\ MAMMALS the parietal cells of the gastric mucosa, they appear in great abundance in the luminal portion of the cell. They seem to migrate towards the cell surface and fuse with the surface plasma membrane (Fig. 9). It may be assumed that they participate in a certain secretory process, involving the transport of large amounts of fluid to the cell surface. Similar vesicles are a prominent feature of the capillary wall, where many such struc- tures are found thi'oughout the endothe- lial cytoplasm as well as in connection with both the surface and the basal plasma membrane (Fig. 10). The theory has been forwarded that they represent the structural evidence of fluid being transported across the capillary wall. In nerve endings and in connection with synapses, similar vesicles have been demonstrated to contain acetyl- choline. The presence of synaptic vesicles has suggested that these might be involved in either the formation of the chemical mediator or in its transmission across the synapse. Ergastoplasm (rough-surfaced endo- plasmic reticulum). In the light micro- scope, areas with basophilic structures have been called ergastoplasm. The electron microscopical analysis of such areas reveals highly complicated systems of paired mem- branes, each membrane having a thickness of about 40A (Fig. 1). Each pair of membranes surrounds a light space called cisterna. It has been demonstrated in nerve cells that the cisternaeof the Nissl body communicate with each other. The cytoplasmic aspect of each membrane is studded with numerous small granules with a thickness of 150A (Fig. 2). These granules have been separated from the membranes by cell fractionation and subse- quent differential centrifugation. Enzymatic tests prove that they contain ribonucleic acids Fig. 10. Part of a lymph capillary in the human epididymis. The cytoplasm of the capillary endothelium (L) contains a large number of microvesicles. One mitochon- drion (M) is seen. At the top is the capillary lumen (Lu) and at the bottom part of a fibroblast (F) as well as several cross sectioned collagen fibrils (C). The lymph capil- laries lack the usual type of basement membrane. Instead, a network of fine fibrils (Re) probably of reticular nature, establish the immediate base upon which the endo- thelial cells rest. Magnification 34,oOOX 103 ELECTRON MICROSCOPY and the granules or particles have, therefore, been called RNA-particles. The number of membranes varies from tissue to tissue; In cells engaged in heavy protein synthesis, such as the exocrine cells of the pancreas, the ergastoplasm dominates the cell with its complicated pattern of membranes and cisternae throughout the entire cell ex- cept in the Golgi area. Evidently the mem- branes and cisternae produce the precursors of the zymogen granules which, at a later stage appear within the Golgi apparatus. In nerve cells, the Nissl bodies have the same ultrastructure, participating in the produc- tion of proteins needed within the cell itself. Again in other cells the ergastoplasm may be seen as scattered short pairs of membranes with the RNA-particles attached. The terminology used in connection with the ergastoplasm is somewhat confusing. As mentioned, the term "ergastoplasm" refers to basophilic areas which can be seen in the light microscope. The term "endoplasmic reticulum," introduced by Porter and Kail- man (1952), originally referred to a structure presumably vesicular or tubular w^hich was observed in whole cells in tissue cultures. Extended studies demonstrated that the endoplasmic reticulum corresponds to the ergastoplasm. When later structures like infoldings of the plasma membrane, pino- cytosis vesicles, and the membranes of the Golgi apparatus were sufficiently analyzed, it was found that some of these structures could be in direct continuity with the er- gastoplasm. It was, therefore, suggested that all membranes wdthin the cell, whether smooth or rough surfaced, may represent diverse differentiations of one single mem- branous system. Therefore, the ergasto- plasm wdth its RNA-dotted membranes is usually referred to as the rough-surfaced endoplasmic reticulum, whereas the Golgi membranes, cytoplasmic vesicles and in- folded or invaginated portions of the plasma membrane are called smooth-surfaced endo- plasmic reticulum. For a more detailed dis- cussion of this problem, consult Haguenau (1958) International Review of Cytology, VII. Another example of a purely smooth- surfaced endoplasmic reticulum is foimd in striated muscle cells, here called sarcoplasmic reticulum.Ii is an elaborate network of smooth tubules around the mj^ofibrils with expan- sions of the system specifically localized in relation to the Z-band. It has been suggested that this system might function in the inward spread of the excitation impulse to contract. Microsonies. When using differential cen- trifugation the membranes and cisternae of the ergastoplasm break up to form small spheres which can be isolated at a certain cent rifugat ion speed. They then represent the microsome fraction (Fig. 2). RNA-particles. In most cells, the cyto- plasm displays an abundance of small granules or particles with a diameter of about 150A (Fig. 9). They appear single or in clusters of 3-5 and are freely dispersed throughout the cytoplasm. They are iden- tical in size and shape to the particles which are attached to the membranes of the ergastoplasm. It has been clearly demon- strated that either type of granules contains ribonucleoproteins and is, therefore, called either RNA- or RNP-particles. Glycogen Granules. Particles have been demonstrated in the cytoplasm of the stri- ated muscle and in the heart muscle which have a diameter ranging between 150A and 300A. Thus, they are somewhat larger than the ribonucleoprotein particles with which they can easily be confused. The larger particles are more variable in their size and shape and have less sharply defined margins. Histochemical tests seem to prove that they contain glycogen and it is, therefore, likely that they represent a particulate form of glycogen. One has not been able to demon- strate similar granules in mammalian liver cells. The glycogen-rich areas of the liver cell cytoplasm usually show a diffuse, cot- ton-wool texture of low density when using solely osmium tetroxide as a tissue stain. 104 CELL ULTKASTRUCTLRE IN MAMMALS However, in appl^dng an additional stain of heavy metal to the thin sections, as for in- stance phosphomolybdic acid, the large glj^cogen particles also become visible in liver cells. Fibrillar Structures Fibrillar structures can easily be resolved with the electron microscope. The}^ may be located within the cell cytoplasm as is the case with myofilaments, tonofilaments, and the neurofibrils, or they are found at the outer surface of the cells or in the interstitial space, here recognized as collagen, reticular fibers and elastin. Intracellular. Striated muscle filaments. The most evident myofilaments are found in the striated skeleton and heart muscle cell (Fig. 11). Here they occupy the main portion of the cell oriented longitudinally, and they represent the contractile elements of the muscle cell. The myofilaments extend along the whole length of a sarcomere which is the structural, repeated unit of the muscle fiber. The thickness of the individual myo- filament varies within different areas of the myofibril, with its smallest diameter related to the area of the I-band and the H-band, and its largest diameter within the Z-band, S-band, and M-band. In a stretched muscle fiber, presumably corresponding to the re- laxed position, the mean diameter of the myofilament is about lOOA in the H-band, whereas the same value for the M-band is in the neighborhood of 150A. During muscle contraction, the diameter of the individual myofilament increases about three times as compared to its stretched diameter. It has been proposed that each myofilament in turn is composed of three subunits. Each subunit consists of rows of rodlets measuring 20A in length. The subunits are assumed to represent cables of supercoiled a-helices of protein molecules. The main constituents of the myofilaments are the proteins actin and myosin. As yet, it has not been convincingly proved what part of the myofilament represents the actin and what represents the myosin. It has been sug- gested that actin is represented by one set of filaments and myosin by another set. There is also a num})er of theories about how the contraction occurs from a structural point of view. Further extensive investiga- tions are needed until a definite solution is arrived at regarding the striated muscle. Smooth muscle filaments. In the smooth muscle cell, the contractile elements are not as easily demonstrated as in the striated one (Fig. 12). This is particularly true in the smooth muscle cells of the blood vessels. Although it is well known that these cells do contract, the cytoplasm is remarkably devoid of any fibrillar structures. However, in the smooth muscle cells of the small bronchi and bronchioles of the lung as well as in those of the intestinal wall, a longitud- inal striation can more readily be seen. The diameter of the filaments here average 150A. The length of each filament is more difficult to determine and it seems that it does not extend for a length of more than half a micron. The difference in ultrastructure be- tween the striated and the smooth muscle filaments seems to reflect the difference in their function, since it is well known that the smooth muscle cell has a much slower rate of contraction than the striated one. Tonofilaments. In the cells of the epidermis an elaborate system of tonofibrils crisscrosses the cytoplasm. The tonofibrils are well dis- tinguished in the light microscope. Their ultrastructure is characterized by an abun- dance of .small tonofilaments, oriented longi- tudinally to the axis of the tonofil^ril (Fig. 13). Each tonofilament has a thickness of about 190A and an approximative length of 0.5 micron with seemingly tapered ends. The filaments have a light core and an electron dense wall, the latter with a diameter of about 70A. It has not as yet been possible to determine whether the tonofilaments are spindleshaped or slightly twisted around 105 Fig. 11. Myofibril in the specific tissue (im- pulse conducting system) of the steer heart. The longitudinallj' arranged myofilaments are seen with their various thickenings related to particu- larly the Z and M bands. Mitochondria (M) and microvesicles (Ve) are abundant throughout the sarcoplasm. Magnification X 49 ,000 Fig. 12. Longitudinal section through the con- tractile region of the sarcoplasm of a smooth mus- cle cell of the mouse intestine. Fibrillar units may be distinquished (arrows) and a certain longitud- inal arrangement is vaguelj' indicated. In the con- tractile region is always present a certain number of dense, ovoid structures (0), typical for the smooth muscle sarcoplasm. At the cell boundary (CB) are aggregations of microvesicles (Ve). The mitochondria (M) are clustered in the central por- tion of the cell. Magnification X47,000 CELL ULTUASTKL'CTURE IN MAMMALS 'aSESii.. Fig. 13. Detail of a basal cell of the human epi- dermis. The cytoplasm is run through by abundant tonofilaments (Tf) most of which are sectioned longitudinally. They display a hollow structure (arrows) and originate from dense areas (X) of the plasma membrane which faces the basement mem- brane (BM). Magnification 83, 000 X each other m the formation of the tonofibrils. The tonofibrils originate from and terminate at the desmosomes which are buttonhke structures associated with the plasma mem- Cl ^H^mSi^ ^i^-*' a2ju Fig. 14. Cell l)()iiiul:uu'.-> ui luu adjacent cells (Cl and C2) of the basal part of the human epider- mis. The cells are highly interdigitated and the mutual attachment established through desmo- somes (De) in association with which the in- tercellular space is wider than elsewhere. The tonofilaments (Tf) terminate (or originate) at the desmosomes. Magnification 90,000X brane (Fig. 14). Histochemical and bio- chemical tests prove that the main com- ponent of the tonofilaments is keratin, a tough noncontractile, in young tissue quite 107 ELECTRON MICROSCOPY elastic, structure, which gives the cells of the organ and is probably dependent on the age epidermis a certain elasticity and firmness, and the function of the fibril. Each collagen Neurofibrils. In the axoplasm of nerve cells fibril is in turn composed of small protofibrils still another fibrillar cytoplasmic structure or tropocollagcn units with a diameter of loA can be found. The neurotibrils of classical and a length of 2G00A. The tropocollagcn histology represent aggregates of axon fila- units are tied up in a staggered fashion to ments large enough to be resolved in the form the collagen fil)ril. The bands of the light microscope. Deposition of heavy met- collagen fibril represent discontinuities in the als, as with the histological silver and gold staggered arrangement of the protofibrils technics, favors the detection of the very and stand out in the electron micrograph thin fibrous structures. In the early days of because heavy metals have a greater affinity electron microscopy, these structures were for this irregularity. The formation of the mistaken for tubules, hence the first name to immature collagen seems to occur at the sur- be coined was neurotubules. Presently, there face of the fibroblasts, the main cell type of seem to be two kinds of fibrils in the axo- all connective tissues. Small fibrils without plasm — the protoneurofibrils and the neuro- periodicity appear at the surface of the filaments. The proto7ieuro fibrils appear as fibroblast. These unit fibrils organize out of smooth threads with a thickness of about 60 or polymerize from material present at the to 80A. Their length ranges between 0.5 to cell surface, and from here the fibrils, in 1 micron. They have an irregular course and many cases already in bundles, are shed into are seen to branch and interconnect. The the intercellular space. The fibrils will first neurofilaments have a thickness of about appear with an axial periodicity of 210A, but 150A with indefinite length. They are some- as they increase in size they also change into times of a double-edged appearance. Their a 640A periodicity. The subsequent fibril surface is smooth and they do not seem to growth apparently occurs by accretion of constitute any part of the endoplasmic materials from the general environment of reticulum. The function of either type of the intercellular spaces and not by fusion of fibrillar structure is unknown. smaller fibrils as believed earlier. Subsequent Extracellular. Collagen. The main fine layers of collagenous material are deposited component of the connective tissue is the upon the core, represented by the tropocoi- collagen fibers. The width of the collagen lagen unit fibril. The origin of the tropocol- fiber is about one micron and its length is lagen fibril is not settled, but considering the indefinite. Each fiber is in turn built up of presence of abundant rough-surfaced endo- numerous collagen fibrils which also may be plasmic reticulum in the cytoplasm of the seen single or in groups of two or more fibrils fibroblast which presumably is involved in scattered in the interstitial tissue (Fig. 15). the protein synthesis, it seems justifiable to Within each group of collagen fibrils the suggest the following mechanism as being fibrils are usually parallel with each other, the most likely regarding the role of the The collagen fibril has a thickness which fibroblast in the formation of the collagen varies between 400 and 2400A depending on fibrils. From the cisternae of the rough-sur- the age (Fig. 10) ; its length is indefinite. The faced endoplasmic reticulum of the fibroblast fibril has an axial periodicity of light and the monomeric form of the tropocoUagen dense bands with a length of each period of unit fibril is discharged to the environment of approximately 640A. In the light and dense the cell and is here ciuickly induced to segments can be seen several smaller bands polymerize by enzymes resident in tem- of varying density and thickness. The length plates at the cell surface or in the unit fibrils and number of bands varies from organ to themselves. 108 CELL ULTRASTRUCTURE IN MAMMALS Fig. 15. Connective tissue of the tracheal submucosa of the rat showing elastic fibers (E), collagen fibrils (C), fibroblasts (F) and reticular fibrils (Re). Magnification 24, 000 X Reticular fibers. In the light microscope fine fibers are found in loose connective tissue and in the interstitial substance of cartilage which can be elect ively impregnated with silver and they are, therefore, often called argyrophil fibers (Fig. 15). Their submicro- scopic structure is like that of collagenous fibers. They have a thickness of about 200 A and have the characteristic cross striations of collagen. In growing tissue it has been demonstrated that the bundles of fibers increase in thickness and finally lose the ability to be impregnated with silver. The reticular fibers found in the interstitial sub- 109 ELECTRON ^IICROSCOPY stance of cartilage do not always show the cross striatioiis (Fig. 10), and it may, there- fore, be assnmed that not all reticular fibers can be transformed into collagenous ones. Elastin. Elastic fibers are found in loose connective tissue (Fig. 15). They have a thickness of about one micron and seem to have an indefinite length. They are not fibrillar but have a homogeneous appearance in the light microscope. Elastic fibers at most show only a weak positive birefrin- gence, but become strongl}^ birefringent on stretching. This is caused by an orientation of the submicroscopic components in the direction of the fiber axis. These ultrastruc- tural components are difficult to demonstrate in osmium-fixed specimens, but it has been possible to distinguish thin filaments with a thickness of about 70A at the periphery of the elastic fiber. Hence, it seems that the elastic fiber has two main components. The dominating structure is a non-fibrous dense cement substance, presumably an albumi- noid which embeds the less apparent elastic filaments. Crystals There are only a few examples of where crystals may be found in normal mammalian tissues. Intracellular. Intracellularly located are the so called crystalloids of the interstitial (Sertoli) cells of the human testis. These crystals are probably of protein natvn-e. The crystals can be seen in the light microscope. They are usually elongated structures with rounded or pointed ends. Each cr3^stalloid body is made up of numerous dense granules with a diameter of about 150A. They are spaced about 190A apart along two axes which are approximately at right angles to each other. This pattern is thought to represent the arrangement of macromole- cules in the lattice of a protein crystal. The function of the crj^stals and the reason for their presence in the Sertoh cells of the testis is unknown. It has been suggested that these cells have an endocrine glandular function and, therefore, the crystals may be involved in the production of the hormone. Extracellular. Extracellularly located crystals are encountered in the bone tissue where they make up the strong and resistant component of the skeletal system. The crys- tals have a width of about 35A. They are scat- tered in the extracellular matrix of the bone tissue but they are also lined up within the collagen fibrils with their long axes oriented parallel to the long axis of the collagen fibril. Selected-area electron-diffraction of these structures has revealed that they are crys- tals of apatite, more specifically hydroxy- apatite [Caio(P04)6(OH)2]. Similar crystals appear under pathologic conditions in areas undergoing calcification like in aging car- tilage, and they also form the main com- ponent of concretions precipitated through- out the urinary system (Fig. 16). Secretion As already mentioned, most of the large granules encountered in secretory cells are associated with secretory processes and have, therefore, been called secretory granules. Although formed within the cell and from the beginning being a part of the cytoplasm, they become discharged and perform their action outside the cell territory. There are two types of secretory cells, namely the exocrine and the endocrine cells. Exocrine Secretion. The exocrine secre- tory granules appear within the Golgi zone, but are evidently preformed in association with the ergastoplasmic sacs (or the cis- ternae of the rough-surfaced endoplasmic reticulum). The fii'st indication of secretory granules within the Golgi zone is a condensa- tion of the Golgi vacuoles or a swelling of the Golgi vesicles and small granules. Sei'ous secretion. In secretory cells with a serous production, the indi^-idual secretory granules are usually surrounded by a single membrane and there is no indication of a coalescence of granules. The granules mi- 110 CELL I LTR ASTRUCTl RE IN >IA>L>IALS Fig. 16. Hydroxyapatite crystals, located in a concretion, experimentally pro- duced in the tubular lumen of the proximal convolution of the rat kidney by injecting subcutaneoush" large doses of parathyroid hormone. (From Engfeldt, et al., 1958). Magnification 16o,000X grate toward the luminal part of the cell Mucous secretion. In secretory cells with a where, in some instances, it can be seen that production of mucin, the secretory granules the limiting membrane of the granule fuses are not bound by a membrane. Frequenth', with the surface plasma membrane and the fusion of several mucous granules occurs, granule empties its content into the lumen, and the discharge of mucin does not occur 111 ELECTRON MICROSCOPY until the whole luminal part of the cell (the goblet) is filled by a muhittide of mucous granules which at all limes are discharged into the lumen. It is believed that either type of exocrine cell has the ability to re- generate new secretory granules. Accord- ing to earlier theories, the cell would disin- tegrate after the secretory cycle is completed. Endocrine Secretion. The endocrine granules are usually smaller than their exocrine relatives. They are also more elec- tron-dense and are mostly surrounded by a single membrane. Few high resolution stud- ies have so far been performed on endocrine glands, but judging from the data available, the formation of the endocrine granules is similar to what is known about the exocrine. A varying amount of rough-surfaced endo- plasmic reticulum (ergastoplasm) as well as an abundance of RNA-particles seem to contribute the necessary prerequisites for the production of the early stages of endo- crine granules. In the beta-cells of the endo- crine portion of the pancreas, it has been quite convincingly demonstrated that the granules first appear in the Golgi zone. The matter of bringing these products in contact with the capillaries of the gland is still not fully explained but evidence is at hand for a migration of the endocrine granules toward that part of the cell which faces the capillary. In some instances, it has also been possible to demonstrate that the membrane of the granule fuses with the plasma membrane, thus giving the dense granule the oppor- tunity to be discharged into the small extra- cellular space existing between the plasma membrane and the basement membrane which surrounds the endocrine cells. From here, the content of the granule may quite easily diffuse into the interstitial space and from there into the capillary. Further studies are, however, needed to prove that this occurs in relation to all endocrine organs. Ground Substance of the Cytoplasm In light microscopy, the term "ground substance" referred to that part of the cyto- plasm which was not organized as mito- chondria, Golgi apparatus, specialized cell inclusions like zymogen granules, pigments, or structures such as myofibrils. With the introduction of the ultrastructural era, it was demonstrated that the ground substance does contain particular well-defined struc- tures like cytoplasmic membranes of various kinds, vesicles, RNA-particles, osmiophilic large granules, and various fibrillar struc- tures. These ultrastructures may, of course, still be regarded as being part of the "ground substance" of the cytoplasm. However, it may also be justifiable today to call that W'hich is as yet not defined as discrete struc- tural entities as representing the ground sub- stance awaiting new techniciues to be de- veloped before this portion of the cytoplasm can be described. In most electron micro- graphs, a certain homogeneous background characterized by a certain electron density can always be recorded. This "background" represents the ground substance in our present concept of the cytoplasm. It is con- ceivable that this background contains a great variety of salts and ions as well as carbohydrates, proteins and fats w^hich available staining techniques and resolving power of the electron microscope fail to bring out. Until these are further developed, let us consider the homogeneous background of our electron micrographs as being the true ground substance. In doing so, we may create the necessary challenge to explore the un- known structural world of the atoms of the cytoplasm. REFERENCES General Reviews Selby, C. C, "Microscopy. II. Electron micro- scopy: a review," Cancer Research, 13, 753 (1953). Sjostrand, F. S., "Electron microscopy of cells and tissues," in "Physical Techniques in Biological Research," 3, 241 (1956). Eds. G. Oster and A. W. Pollister, Academic Press, Inc., New York. Sjostrand, F. S., "The ultrastrvicture of cells as revealed by the electron microscope", Int. Rev. Cytol., 5, 455 (1956). 112 CELL ULTRASTRrCTURE IN MAMMALS Oberling, C, "The structure of the cytoplasm," Int. Rev. Cytol., 8, 1 (1959). Miller, F., "Orthologie und Pathologie der Zelle im elektronenmikroskopischen Bild," "Vehr. Deutsch. Ges. Pathologie," p. 261. Gustaf Fischer Verlag, Stuttgart, 1959. Selby, C. C, "Electron microscopy: techniques and applications in cytology," in "Analytical Cytology," p. 273, Ed. R. C. Mellors, Mc- Graw-Hill Book Company, Inc., New York, 1959. "The cell," Vol. 2: Cell Constituents, Eds. J. Brachet and A. E. Mirsky, Academic Press, Inc., New York and London, 1960. Basement Membrane Weiss, P. and Ferris, W., "The basement la- mella of amphibian skin", /. Biophys. Bio- chem. Cytol., 2, 275 (1956) Suppl. VAN BrEEMEN, v. L., ReGER, J. F., AND CoOPER, W. G., "Observations on the basement mem- branes in rat kidney," /. Biophys. Biochem.. Cytol., 2, 283 (1956)' Suppl. Cent Hole Yamada, E., "The fine structure of centriole in some animal cells", Proc. 1st. Reg. Conf. in Asia and Oceanic, p. 247, Tokyo, 1956. Bernhard, W. AND DE Harven, E., "Sur la presence dans certaines cellules de Mammi- feres d'un organite de nature probablement centriolaire", C. r. Acad. Sci. Paris, 242, 288 (1956). de Harven, E. and Bernhard, W., "Etude au microscope ^lectronique de I'ultrastructure du centriole chez les vertebras" Z. Zellf., 45, 378 (1956). Cell Surface Fawcett, D. W., "Structural specializations of the cell surface," in "Frontiers in Cytology," p. 19, Ed. S. L. Palay, Yale University Press, New Haven, 1958. Rhodin, J. AND Dalhamn, T., "Electron micro- scopy of the tracheal ciliated mucosa in rat," Z. Zellf., 44, 345 (1956). Rhodin, J., "Anatomy of kidney tubules" Int. Rev. Cytol., 7, 485 "(1958). Weiss, P., "Cell contact," Int. Rev. Cytol., 7, 391 (1958). Cilia Fawcett, D. W. and Porter, K. R., "A study of the fine structure of ciliated epithelia," J. Morph., 94, 221 (1954). Rhodin, J. and Dalhamn, T., "Electron micros- copy of the tracheal ciliated mucosa in rat," Z. Zellf., 44, 345 (1956). Rhodin, J., "Ciliated epithelia". Int. Rev. Cytol., 10 (1962). Collagen Gross, J., "The behavior of collagen units as a model in morphogenesis," J. Biophys. Bio- chem. Cytol., 2, 261 (1956) Suppl. Crystals (extracellular) Robinson, R. A. and Watson, M. L., "Crystal- collagen relationships in bone as observed in the electron microscope," Ann. N. Y. Acad. Sci., 60, 596 (1955). Knese, K.-H. and Knoop, A.M., "Elektronenop- tische Untersuchungen iiber die periostale Osteogenese," Z. Zellf., 48, 455 (1958). Glimcher, M. J., "Molecular biology of mineral- ized tissues with particular reference to bone," Rev. Modern Physics, 31, 359 (1959). Crystals (intracellular) Fawcett, D.W. and Burgos, M. H., "Observations on the cytomorphosis of the germinal and in- terstitial cells of the human testis," in Vol. 2, "Ageing in Transient Tissues, Ciba Foun- dation CoUoquia on Ageing," p. 86, Eds. G. E. W. Wolstenholme and E. C. P. Millar, J. & A. Churchill, Ltd., London, 1956. Elastin Hall, D. A., "The fibrous components of con- nective tissue with special reference to the elastic fiber," Int. Rev. Cytol., 8, 212 (1959). Endocrine Secretion Ferreira, D., "L'ultrastructure des cellules du pancreas endocrine chez I'embryon et le rat nouveau-ne," /. Ultrastructure Research, 1, 14 (1957). MuNGER, B. L., "A light and electron microscopic study of cellular differentiation in the pan- creatic islets of the mouse," Am. J. Anat., 103, 275 (1958). Lacy, P. E., "Electron microscopic and fluorescent antibody studies on islets of Langerhans," Exp. Cell Research, 7, 296 (1959) Suppl. Ekholm, R. and Sjostrand, F. S., "The ultrastruc- tural organization of the mouse thyroid gland," /. Ultrastructure Research, 1, 178 (1957). Herman, L., "An electron microscope study of the salamander thyroid during normal stimu- lation," J. Biophys. Biochem. Cytol., 7, 143 (1960). 113 ELECTRON MICROSCOPY Endoplasmic Reticuhim Palade, G. E., "Tlie piidoplasmic reticulum" J. Biophys. Biochem. Cytol., 2, 85 (1956) Suppl. Haguenau, F., "The ergastoplasm: its history, ultrastructiire, and biochemistry," Int. Rev. Cytol., 7, 425 (1958). Exocrine Secretion Sjostrand, F. S. and Hanzon, V., "Membrane structures of cytoplasm and mitochondria in exocrine cells of mouse pancreas as revealed by high resolution electron microscopy," Exp. Cell Research, 7, 393 (1954). Rhodin, J. and Dalhamn, T., "Electron micro- scopy of the tracheal ciliated mucosa in rat," Z. Zellf., 44, 345 (1956). Palay, S. L., "The morphology of secretion," in "Frontiers in Cytology," p. 305, Ed. S. L. Palay, Yale University Press, New Haven, 1958. Ekholm, R. and Edlund, Y., "Ultrastructure of the human exocrine pancreas, J. Ultrastruc- ture Research, 2, 453 (1959). Glycogen Bernhard, W. and Rouiller, C, "Close topo- graphical relationship between mitochondria and ergastoplasm of liver cells in a definite phase of cellular activity," /. Biophys. Bio- chem. Cytol., 2, 73 (1956), Suppl. Fawcett, D. W. and Selby, C. C, "Observations on the fine structure of the turtle atrium," J. Biophys. Biochem. Cytol., 4, 63 (1958). Watson, M. L., "Staining of tissue sections for electron microscopy with heavy metals," /. Biophys. Biochem. Cytol., 4, 475 (1958). Golgi Apparatus Dalton, a. J., "A study of the Golgi material of hepatic and intestinal epithelial cells with the electron microscope, Z. Zellf., 36, 522 (1952). Sjostrand, F. S. and Hanzon, V., "Ultrastruc- ture of Golgi apparatus of exocrine cells of mouse pancreas," Exp. Cell Research, 7, 415 (1954). Gatenby, J. Bronte, "The Golgi apparatus," R. Micr. Soc, 74, 134 (1955). Haguenau, F. and Bernhard, W., "L'appareil de Golgi dans les cellules normales et canc^reuses de vertebras," Arch, d'anatomie micr. et morph. exp., 44, 27 (1955). Dalton, A. J. and Felix, M., "A comparative study of the Golgi complex," J. Biophys. Biochem. Cytol., 2, 79 (1956) Suppl. Large Granules Rhodin, J. and Dalhamn, T., "Electron micros- copy of the tracheal ciliated mucosa in rat," Z. Zellf., 44, 345 (1956). NiLSSON, O., "Ultrastructure of mouse uterine surface epithelium under different estrogenic influences," /. Ultrastructure Research, 1, 375 (1958). Zelander, T., "Ultrastructure of mouse adrenal cortex," ./. Ultrastructure Research, 2, (1959) Suppl . Ladman, a. J. and Young, W. C, "An electron microscopic study of the ductuli efferentes and rete testis of the guinea pig," J. Biophys. Biochem. Cytol., 4, 219 (1958). Microbodies Rhodin, J., "Correlation of ultrastructural organi- zation and function in normal and experi- mentallj' changed proximal convoluted tubule cells of the mouse kidney," Thesis, Karolinska Institutet, Stockholm, 1954. Rouiller, C. and Bernhard, W., "Microbodies and the problem of mitochondrial regenera- tion in liver cells," /. Biophys. Biochem. Cytol., 2, 355 (1956) Suppl. Engfeldt, B., Gardell, S., Hellstrom, J., iVEMARK, B., Rhodin, J. and Strand, J. "Effect of experimental hyperparathyroidism on renal function and structure," Acta Endo- crinologica, 29, 15 (1958). Microsomes Palade, G. E. and Siekevitz, P., "Pancreatic microsomes," /. Biophys. Biochem. Cytol., 2, 671 (1956). Siekevitz, P. and Palade, G. E., "A cytochem- ical study of the pancreas of the guinea pig," /. Biophijs. Biochem. Cytol., 4, 309 (1958). Mitochondria Palade, G. E., "The fine structure of mitochon- dria," Anat. Rec, 114, 427 (1952). Sjostrand, F. S., "Electron microscopy of mito- chondria and cytoplasmic double mem- branes," Nature, 171, 30 (1953). Steffen, K., "Chondriosomen und Mikrosomen (Spharosomen)," "Encj^clopedia of Plant Physiology," p. 574, Ed. W. Ruhland, Springer-Verlag, Berlin, 1955. Siekevitz, P. and Watson, M., "Cytochemical studies of mitochondria," J. Biophys. Bio- chem. Cytol., 2, 639 (1956). Palade, G. E., "Electron microscopy of mito- chondria and other cytoplasmic structures," in "Enzymes: Units of Biological Structure 114 CELL I LTR AS TRUCTl RE L\ MAMMALS and Function," p. 185, Ed. O. H. Gabler, Academic Press, Inc., New York, 1956. Nerve Cells Palay, S. L. and Palade, G. E., "The fine struc- ture of neurons," /. Biophys. "Biochem. Cytol., 1, 69 (1955). Fernandez-Mo ran, H. and Brown, R., "The submicroscopic organization and function of nerve cells," Exp. Cell Research, 5, (1958) Suppl. Neurofibrils EsTABLE, C, Acosta-Ferreira, W., and Sotelo, J. R., "An electron microscope study of the regenerating nerve fibers," Z. Zellf., 46, 387 (1957). ScHMiTT, F. O., "Molecular organization of the nerve fiber," Rev. Modern Physics, 31, 455 (1959). Nucleus Afzelius, B. a., "The ultrastructure of the nuclear membrane of the sea urchin oocyte as studied with the electron microscope," Exp. Cell Research, 8, 147 (1955). Watson, M. L., "The nuclear envelope. Its struc- ture and relation to cj'toplasmic membranes," J. Biophys. Biochem. Cytol., 1, 257 (1955). Haguenau, F. and Bernhard, W., "Particulari- tes structurales de la membrane nucleaire," Bulletin du Cancer, 42, 537 (1955). Watson, M. L., "Further observations on the nuclear envelope of the animal cell," /. Bio- phys. Biochem. Cytol., 6, 147 (1959). Nucleolus Bernhard, W., Bauer, A., Gropp, A., Hague- nau, F., and Oberling, C, "L'ultrastructure du nucleole de cellules normales et cancereuses," Exp. Cell Research, 9, 88 (1955). Pigments Falk, S. and Rhodin, J., "Mechanism of pig- ment migration," "Proc. Stockholm Confer- ence on electron microscopy," p. 213, 1956, Eds., F. S. Sjostrand and J. Rhodin, Almquist and Wiksell, Stockholm, 1957. Wellings, S. R. and Siegel, B. V., "Role of Golgi apparatus in the formation of melanin granules in human malignant melanoma," /. Ultrastructure Research, 3, 147 (1959). Wellings, S. R. and Siegel, B. V., "Electron microscopy of human malignant melanoma," J. Nat. Cancer Inst., 24, 437 (1960). Charles, A. and Ingram, J. T., "Electron mi- croscope observations of the melanocyte of the human epidermis," J . Biophys. Biochem. Cytol., 6, 41 (1959). Plasma Membrane Robertson, J. D., "The molecular biology of cell membranes," in "Molecular Biology," p. 87, Ed. D. Nachmansohn, Academic Press, Inc., New York and London, 1960. Reticular Fibers Jackson, S. F., "The morphogenesis of avian ten- don," Proc. Royal Soc, B, 144, 556 (1956). Wassermann, F. and Kubota, L., "Observations on fibrillogenesis on the connective tissue of the chick embryo with the aid of silver im- pregnation," J. Biophys. Biochem. Cytol., 2, 67 (1956). Porter, K. R. and Pappas, G. D., "Collagen for- mation by fibroblasts of the chick embryo dermis," /. Biophys. Biochem. Cytol., 5, 153 (1959). Zelander, T., "Ultrastructure of articular carti- lage," Z. Zellf., 49, 720 (1959). RN A -Particles Palade, G. E., "A small particulate component of the cytoplasm," J. Biophys. Biochem. Cy- tol., 1,59 (1955). Palade, G. E., "Microsomes and ribonucleopro- tein particles," in "Microsomal Particles and Protein Synthesis," p. 36, Ed. R. B. Roberts, Pergamon Press, New York, 1958. Small Vesicles Rhodin, J. "Anatomy of kidney tubules," Int. Rev. Cytol, 7, 485 (1958). Hally, a. D., "The fine structure of the gastric parietal cell in the mouse," J. Anat., 93, 217 (1959). DE Robertis, E., "Submicroscopic morphology of the synapse," Int. Rev. Cytol., 8, 61 (1959). Fawcett, D. W., "The fine structure of capil- laries, arterioles and small arteries," in "The Microcirculation," p. 1, Eds. S. R. M. Rey- nolds and B. W. Zweifach, The University of Illinois Press, Urbana, 1959. Vial, J. D. and Orrego, H., "Electron micro- scope observations on the fine structure of parietal cells," J. Biophys. Biochem. Cytol., 7, 367 (1960). Smooth Muscle Caesar, R., Edwards, G. A., and Ruska, H., "Architecture and nerve supply of mamma- lian smooth muscle tissue," J. Biophys. Bio- chem. Cytol., 3, 867 (1957). Thaemert, J. C, "Intercellular bridges as proto- 115 ELECTRON MICROSCOPY plasmic anastomoses between smooth muscle cells," J. Biophys. Biochem. Cytol., 6, 67 (1959). Stereocilia Rhodin, J., "Ciliated epithelia," Int. Rev. Cytol., 10 (1962). Striated Muscle Hodge, A. J., "Fibrous proteins of muscle," Rev. Modern Physics, 31, 409 (1959). "Structure and function of muscle," Vol. 1, Struc- ture, Ed. G. H. Bourne, Academic Press, Inc., New York, 1960. Tonofilaments Selby, C. C, "An electron microscope study of the epidermis of mammalian skin in thin sec- tions," J. Biophys. Biochem. Cytol., 1, 429 (1955). Odland, G. F., "The fine structure of the inter- relationship of cells in the human epidermis," /. Biophys. Biochem. Cytol., 4, 529 (1958). Setala, K., Merenmies, L., Stjernvall, L., and Nyholm, M., "Mechanism of experimental tumorigenesis. IV. Ultrastructure of inter- follicular epidermis of normal adult mouse," /. Nat. Cancer Inst., 24, 329 (1960). Johannes A. G. Rhodin CILIATED EPITHELIA ULTRASTRUCTURE Ciliated epithelia are so called because many of the cells which form the epithelial layer are provided with a great number of small motile structures on their surface — the cilia. The ciliated epithelia are found in con- nection with organs and tissues where a sur- face has to be kept clean and moist, as in the respiratory tract (nose, trachea, bronchi), or in small ducts where a certain propulsion of the content of the duct is facilitated and aided by the beating of the cilia as in the male reproductive tract (efferent ducts of testis) and in the female reproductive tract (oviduct). Function The ciliated epithelium in mammals is characterized by several cell types (Fig. 8) the most prominent of which is the cili- ated cell. The other cells have been classi- fied as tion-ciliated cells which comprise secretory cells of two types {serous and mucous or goblet cells), the brush cells and the hasal cells. The functions of the different cells de- pend on the location of each cell. The goblet cell is predominant in all ciliated epithelia and keeps the surface moist by discharging continuously a more or less viscous mucin. The secretion product of the serous cells either dilutes the mucin and/or adds en- zymes (activators) to the content of the ducts. The brush cells, presumably young cells which have migrated from the basal portions of the epitheliimi, are primarily the precursors of the ciliated cells. They may probably also be transformed into mucous cells as well as serous cells. The appear- ance of the brush cells of the efferent ducts of the testis indicates that their function here is mainly a secretory one. However, certain structural features speak for the fact that they also have absorptive functions. The ciliated cells with their abimdant sur- face extensions, the cilia, participate in keeping the mucus blanket moving. The ciliary beat is rather complicated. It is com- posed of a rapid forward stroke and a slow backward stroke. The cilium is rigid and erected during the forward stroke, whereas it folds itself beneath the mucus in a limb and yielding movement during the backward stroke. The activity of the cilia is syn- chronized in local areas and moveinents of the cilia over larger areas have been com- pared with the appearance of a rye field when the wind blows across it. The basal cells are presumably the precursors of all the cells of the ciliated epitheUa. The mitotic activity among these cells is high and they migrate to the upper portions of the epi- thelial layer in order to replace ciliated or serous cells when they are sloughed off and lost in the moving mucus blanket. 116 CILIATED EPITIIELIA ULTRASTRUCTURE T. ..: ■ MU vt f ^/:^^r-; ".\vf^:.- ^'f^^_._ • -A/i ^ 'i'../'.-.- BM 'iff^T!11fffi Fig. 1. Longitudinal section of the pseudostratified columnar epithelium of the human trachea composed of ciliated cells (C), mucous cells (G), brush cells (BC), and basal cells (B). At the top is the tracheal lumen with the mucous blanket (MU); at the bottom, the basement membrane (BM) with its abundant reticular and collag- enous fibrils. Only the basal cells rest on the basement membrane. Intercellular spaces (I) are evident, and a wandering cell, presumably a lymphocj-te (L), is located in the space between the basal cells. Magnification 1,800X 117 ELECTRON iMICROSCOPY Structure are attached lo the nienihraiiesof the rough- Goblet cell. The goblet cell cytoplasm is surfaced endoi)lasinic reticulum (ergasto- characterized by a high content of ribonu- plasm). Precursors of the mucin are evi- cleic acid (RNA) particles, some of which dently formed within the cisternae which are Fig. 2. Detail ol I'ig. 1 shuwing twu muc-ous cells (G) and one ciliated cell (C). The mitochondria (m) are abundant in the upper part of the ciliated cell in which the Golgi apparatus (go) is also seen. The cilia (cl) emerge from basal bodies at the top of the cell into the tracheal lumen together with a nmnber of microvilli (mv) beneath which a dense structure, the terminal web (tw), is resolved. The cells are attracted by the terminal bar (tb). The mucous cells (G) are filled with mucin granules which are about to be discharged at the cell surface. Magnification 6,000X 118 CILIATED EPITHELIA ULTRASTRUCTURE bound ]iy these membranes. The mucous state when the upper part of the mucous cell granules appear, howe\-er, within the smooth is transformed into a goblet filled by numer- membranes of the Golgi apparatus where a ous large mucous granules. A certain fusion heavy accumulation of various sized mucous of mucous granules occiu's intracellularly granules can be identified previous to the before they are discharged at the surface of 1 ^'U 0.2J1 #' •TmiM ifUfr .*' Lff. ^ Fig. 3. Each cilium emerges into the lumen from a basal body. Rootlets (r) and a lateral fibrillar projection (arrow) secure the ciliary filaments in the luminal part of the cell cytoplasm. The lateral ciliarj' filaments are continuous from the ciliarj^ body to the tip of the cilium. The central filaments cannot be demonstrated within the basal body. Magnification 39,000X Fig. 4. Cross sectioned cilia display a ring of nine paired peripheral filaments and two central single filaments. The plasma membrane which shows up as a .single dense line is well preserved in all but two cilia. Magnification 100,000X 119 ELECTRON MICROSCOPY the epithelium. The nucleus of the goblet cell is pushed to the base of the cell because of the heavy accumulation of mucous gran- ules. After the discharge of the mucin, the cell resumes its resting columnar shape and a new production cycle of mucin starts again (Figs. 1,2). Serous cell. The structure of the serous cells is quite variable, depending on what organ the cells are located in. In the trachea only few serous cells have been recognized, but in the fine bronchi and in the hronchioles they become quite abundant. Their cyto- plasm has an abundance of RNA-particles, presumably related to the ability of the cells to secrete enzymes. In addition, a large number of cytoplasmic vesicles indicate that the cells are turning out fluid or substances dissolved and transported within the mem- brane-bound vesicles. It is believed that these cells play an important role in the mechanism involved in pulmonary edema. In the oviduct of some mammals (man not included), the serous cells display a large Fig. 5. A schematic representation of the fine structure of the mammalian cilium. (After Rhodin and Dalhamn, 1956). mmiber of secretion granules. They appear first within the Golgi zone and migrate from here to the cell surface where they are dis- charged without previous fusion with one another. The granules seem to have some nutritional or enzymatic relation with the ovum when it passes along the oviduct to the uterus. The surface of the serous cells is characterized by a number of microvilli, the size of which is smaller than either the cilia of neighboring cells or the brush border ex- tensions seen on the cells of the intestine or the proximal convolution of the nephron in the kidney. The nucleus is located in the center of the cell and is not dislocated during the secretory cycle, the latter being struc- turally less obvious than what is noticed in the mucous cells. Basal cell. The basal cells are always lo- cated near the basement membrane upon which most of the cells of the ciliated epi- thelia rest. The basal cell is round or slightly elongated and does not reach the surface of the epithelium. Its cytoplasm contains small fibrils of unknown function. Eventually, the cell migrates to the upper part of the epi- thelium where it gains contact with the sur- face and starts to differentiate into a cell type which is ready to replace any of the two kinds of cells that dominates the epi- thelium, the ciliated and the mucous cell (Fig. 6). Brush cell. It is probable that the brush cell represents a basal cell which has recently moved to the surface and started to differen- tiate into a ciliated cell. This is quite obvious in the trachea, where several features of the brush cell are identical with the basal cell. The surface of the brush cell is covered with a large number of brush border-like exten- sions. In man, a dense plate is found at a distance of about half a micron beneath the surface of the brush cell, presumably the site of differentiation of the ciliary basal bodies. Another typical feature of the brush cell is the clustering of small mitochondria beneath this plate in the brush cell of the 120 CILIATED EPITHELIA L LTRASTRUCTURE n f'r # n% ^ :i. ■" **«^, nil ^ nu « \ rm ^^ I ■* % * A*^ • IN •^. B ;\ % K ■\. V ,♦*. I . #««(» - (C « BM Fig. 6. Detail of Fig. 1 showing four basal cells (B). The nuclei (nu) are large and prominent features of these cells with several darker bodies, the nucleoli, the latter a possible indication of mitotic activity. The mitochondria (m) are mostly aggregated beneath the nucleus. The basal cells rest on the basement membrane (BM). The in- tercellular spaces (I) are sometimes the site of lymphocytes (L). Magnification 5,600X human trachea. In the efferent ducts of the that they are obviously secretory cells. The human testis, the brush cells display an siu'face structures are predominantly of elaborate rough-surfaced endoplasmic reticu- brush border type, but resemble more lum, a multitude of freely dispersed RNA- those found in the kidney than those particles, and a large Golgi zone, indicating of the intestine. IVIoreover, a highly de- 121 ELECTRON MICROSCOPY Iji mv Fig. 7. Detail of Fig. 1 showing the top part of a brush cell with its numerous and slender microvilli (mv). Three cilia (ci) are seen, possibly indicating that the brush cell is being transformed into a ciliated cell. Mitochondria (m) are more abundant than in ordinary ciliated cells, and particularly evident is the large number of quite small mitochondria and small vesicular structures in between. The arrow points to a dense structure with a light center which is reminiscent of a cross sectioned basal body. A developing cilium? The dense line beneath the microvilli, the terminal web (tw), is present only in relation to the microvilli. Magnification 17,500X 122 COLLOIDS, LYOPHOBIC veloped system of surface invaginations, reminiscent of tubules, speaks for the fact that structural evidence for micropinocytosis is present. It is, therefore, concluded that the brush cells of the respiratory tract and those of the male reproductive tract are func- tionally different, although the structure of their surfaces is almost identical (Figs. 2, 7). Ciliated cell. The ciliated cells have a cytoplasm which contains only a small amount of RNA-particles and endoplasmic reticulum. These cell organelles are probably used only for maintaining the restricted amount of protein that is synthesized for metabolic processes within the cell. The number of cilia per cell is large; in the rat trachea, it amounts to between 250 and 300. The cilium is covered by the plasma mem- brane and its interior represents an exten- sion of the cytoplasm of the cell. Within this cytoplasm is a number of distinct fibrils oriented longitudinally. They originate from the basal body which is located intracellu- lar ly below the level of the cell surface. There are two central single filaments and nine pe- ripheral double ones. They all join in the tip of the cilium. The central ones divide and split within the basal body and wrap around its central core. The peripheral filaments ex- tend below the basal body and terminate at various levels in the upper part of the ciliated cell as ciliary rootlets. Beneath the basal bodies are clusters of mitochondria, the carriers of enzymes in all cells. It is believed that the filaments of the cilium are contrac- tile. The motion center is represented by the basal body, and the energy recjuired for the contraction is derived from the nearby mitochondria. It has been demonstrated that the ciliary beat can occur as long as the con- tact between the basal body and the cilium proper is maintained. Physical damage, which involves a break between the cilium and the basal body will, therefore, stop the ciliary beat. However, it has also been shown that toxic gases as well as cigarette smoke at a certain concentration stop the Fig. 8. A schematic representation of the col- umnar ciliated epithelium of the rat trachea. Con- trary to that of man, this epithelium lacks an evident layer of basal cells (BC) although some may be seen in between the bases of the ciliated (C) and the non-ciliated cells. Among the non-cili- ated cells are found mucous cells (C) in various states of mucous secretion, and brush cells (BRC). The basement membrane is thinner in rat than in man. (After Rhodin and Dalhamn, 1956). ciliary activity. Furthermore, infections like influenza damage the ciliated cells so dras- tically that eventually these cells die and become sloughed off. They are then replaced by basal cells which develop into brush cells and from there into cihated cells (Figs. 3, 4, 5). REFERENCES Fawcett, D. W., and Porter, K. R., "A study of the fine structure of ciliated epithelia," /. Morph., 94, 221 (1954). Rhodin, J. and Dalhamn, T., "Electron micros- copj- of the tracheal ciliated mucosa in rat," Z. Zellf., 44, 345 (1956). Rhodin, J., "Ciliated epithelia," Int. Rev. CytoL, 10 (1962). Johannes A. G. Rhodtn COLLOIDS, LYOPHOBIC The Colloidal State The colloidal state is essentially that state in which matter exists with at least one dimension in the size range 10~^ to 10~^ cm. In this state can be included large molecules 123 ELECTRON >IICROSCOPY such as proteins, enzymes, viruses and high 5G00 A) the maximum resohition obtainable polymers, which ha^•e molecular weights in is of the order of 2000 A, while with ultra- the region of 1,000 to several million and violet light resolution of the order of 800 A exist in molecular solution, and smokes, may be obtained; such resolution is totally mists, gels, etc.. In this size range, many inadequate for the examination of most col- inorganic materials can be prepared as two- loidal dispersions. However, the wavelength phase systems of small particles in liquid of an electron beam produced at an accelerat- media, and are spoken of as colloidal dis- ing potential of 80 kV is 0.043 A and there- persions or sols. The latter groups are usually fore theoretically resolution of the order of termed hjophobic colloids, and it is with this one A unit should be possible. Final resolu- class of colloidal material that this article is tion, however, is limited by the difhculty of mainly concerned. correcting the lens aberrations and with The most important properties of a lyo- earlier microscopes the resolving power was phobic colloidal dispersion are: only of the order of 30 to 50 A. With many (a) the size and shape of the particles, modern instruments the resoh^ing power is and whether the system is dispersed or floe- of the order of 5 A and hence it is possible to culated, resolve objects of the order of atomic di- (b) the chemical structure of the particles, mensions. (c) the nature and structure of the sur- £„_g Experimental Techniques (d) the mode of nucleation and growth of Preparation of Specimen Supports. the particles, and the possible production of The essential criteria for a supporting mem- monodisperse sols, brane are that it should be rigid enough to (e) the electrical charge on the surface withstand manipulation, remain stable in (electrical double layer), and its relation to the electron beam during examination and stability. have a thickness of the order of 100 A. Many Moreover, in phenomena such as nucleation, types of materials have been suggested as growth, coagulation and aging of sols the supporting membranes (1) and probably the dynamic aspects of the system have to be membranes most commonly used are pre- considered and a knowledge of changes in pared from "Formvar" or nitrocellulose. The the system with time is required. The elec- "Formvar" or nitrocellulose membrane is tron microscope may be employed to obtain formed on a dish of distilled water and then answers, or some of the answers, to all these picked up on the surface of a copper mesh questions with the exception of (e). It must grid (2). These films, however, are not suffi- be remembered, however, that as specimens ciently stable for high -resolution work; un- are subjected to a high vacuum (ca. 10~^ mm less they are carefully prepared, they have a mercury) in the electron microscope they tendency to drift under the influence of the cannot be investigated in their natural liciuid electron beam. Greater stability can be environment. The exposure of the specimen achieved by evaporating a thin layer of car- to a beam of high energj^ electrons also means bon on to the plastic film, but care must be that suitable precautions must be observed taken not to increase the support thickness to prevent heating, with subsequent sublima- beyond the limits required for high resolu- tion or decomposition of the specimen; tion. examination of large specimens may be pre- One of the most stable supporting mem- cluded by this factor. branes can be obtained by evaporating ear- In the case of an optical microscope, when bon onto carefully cleaned glass slides or viewing by white light (average wavelength freshly cleaved mica. On immersing the slide 124 COIJ.OIDS, LYOPIIOBIC in water, the carbon film floats on the sur- rectly, concentration can be achieved by- face and can be transferred to the grid. If means of electrodecantation (3). the colloidal dispersions are spread on the Deposition of Sols on Supporting glass slide and allowed to dry before applying Membranes. The simplest method of trans- the carbon, the particles remain embedded in ferring a sol sample to the supporting mem- the film on removal and can be used for direct brane is by the use of a fine loop of platinum viewing. Although carbon forms a very wire. The wire can readily be cleaned by stable film it is not suitable for the examina- flaming before transference of the specimen, tion of all materials. Electrostatic effects are An alternative method, often useful for often encountered which make it difficult to quantitative measurements, is to spray the obtain drop adhesion and solutions contain- sol on to the supporting membrane by means ing surface-active agents usually disrupt the of a nebulizer. Freeze-drying of the sample film; in these cases nitrocellulose-carbon films is often useful and this can be carried out are the most useful with the liquid applied simply by placing the grids on a copper block to the nitrocellulose side. immersed in a freezing mixture; the aqueous Silicon monoxide has also been used to drops then freeze immediately on making obtain a stable supporting membrane, with contact with the supporting membrane, little structure (2). Specimen Contrast. In order to obtain Preparation of Colloidal Dispersions, an image of the particle in the electron mi- Where colloidal dispersions are produced by croscope, electrons must be scattered out of the interaction of two ionic reagents, the the field so that they do not reach the pho- sol produced usually contains considerable tographic plate. The amount of scatter gen- quantities of extraneous electrolyte. This, if erally depends upon the atomic number and left in the sol, tends to crystallize on the sup- the density of the specimen, and it is for this porting membrane and the resultant crystals reason that materials composed mainly of may be confused with the colloidal particles carbon, hydrogen, nitrogen and oxygen, i.e., during examination. Even if actual crystal- of the same composition as a nitrocellulose lization does not occur poor backgrounds supporting membrane, are difficult to ob- may result, or flocculation of the sol may serve. In the latter case the specimen con- occur due to the high concentration of elec- trast can usually be increased by staining or trolyte reached during evaporation . Removal by shadowing with a heavy metal vapor such of any extraneous electrolytes is therefore ^s that of chromium, gold or uranium in a advisable either by dialysis or electrodialysis; ^^Sh vacuum. care must be taken, however, not to remove ^^^'"^^ ^^ ^^^ low penetrating power of stabilizing potential-determining ions. Small electrons it is necessary to use very thin samples may be rapidly dialysed against specimens if mtenor details are to be ob- distilled water in cellophane dialysis sacs; served (see later). With crystalline specimens f i.i-i- 1 r-i- considerations other than random scatter lor electrodialysis a number of simple pieces , ... c , , , , ., 1 1 . V have to be taken into account (2). ot apparatus have been described which can „ , . /• r- i, . , , « . , , ,., , /..N T., , Kesolution ot Colloidal Particles. In be readily constructed (3). Electrolyte may j- ^ • .• r n -j i ^- i . , , , ^ ^ -^ -^ direct examination of colloidal particles, also be removed by passage of the sol ^^^jy ^^e two-dimensional aspects of the through a suitable ion-exchange resin pro- panicle are seen. In this connection it was vided that precautions are taken to avoid realized by von Borries and Kausche (4) that contamination of the sol by particles or crystalline colloidal particles, which should complex ions from the resin. be bounded by plane faces intersecting in WTiere sols are too dilute to be used di- geometrical lines, should be revealed as well 125 ELECTRON MICROSCOPY ..^..H. Fig. 1. (top) Diagram illustrating resolution of the shape of a colloidal particle. The shape of the particle can only be recognized if 6 > 5. (bottom) Comparison of the size of particles of different shape required for resolution of that shape. (After Borries and Kausche). defined shapes in microscopes of infinite re- solving power. Such a condition cannot be reahzed in practice, however, and the efi^ect of finite resolving power has to be considered, von Borries and Kausche (4) supposed that the geometrical boinidaries of the object appeared in the image as boimdaries of finite width, within which the intensity fell con- tinuously from that of the particle to that of the background. The effective width of the boundary was considered to be twice the resolving power, 5, of the microscope and the physical boundary of the particle at a point midway between the particle and back- ground intensities. As a consequence of the finite resolving power the image boundary at the intersection of two straight lines is rounded, and it was assumed that the curva- ture was such that the arc of a circle was tangential to the intersecting edges at a dis- tance 5 from the point of intersection ob- tained by geometrical construction (see Fig. 1). Thus recognition of shape is only possible if the length of the straight portions of the edges, b, is greater or equal to 8. li h < 8 the particle would appear circular and in fact many early workers found that small colloidal particles appeared circular, an effect often due to lack of resolving power. ( )n this basis it is clear that the exact shape of a particle having less than six corners should be easier to recognize and therefore ti'iangu- lar particles should be recognizable as such at much smaller dimensions than the hex- agonal type. Conversely, octagonal plates must be seven to ten times larger than the triangular particles in order not to appear circular. The sizes of particles, relative to a triangle, required for the resolution of a defi- nite shape are illustrated schematically in Fig. 1. In order to test the resolution of an elec- tron microscope it is useful to have a suitable test object, and it has been found (5) that silver and silver iodide sols can be prepared which contain particles having sizes ap- proaching the limits of present-day micro- scopes. Particles having dimensions of the order of 5 A can be clearly resolved from the background (Fig. 2) ; the fact that these par- ticles could be reproduced on separate photo- graphic plates clearly established their identity as colloidal particles. The limita- tions in resolving power appear to be mainly governed by chromatic errors (i.e., stabiliza- tion of high tension) and lens aberrations. The precise measurement of particle sizes be- low 10 A becomes difficult due to phase con- trast effects at this level of resolution. A suitable test object for high resolution work is also found in the case of metal phthalo- cyanines; for example, the metal-bearing «> « ''.•'«,'*« ** * * ♦*'* » ♦-1 * . • '. ' '/,♦.•.■",', •• " *-'♦ • .'• • . ,' ; . .*. *• " • . •;■.•■'*...•**. ,*...-. -,•«•♦■.« « . ■.'•,•?.' "", "* *. ♦"••■• *• * * ' ■■ ■ '■' .*•. • * •'• . ' l1<1^i^ *■■ • • *;' • '♦ Fig. 2. Electron micrograph of silver iodide sol of small particle size. 126 COLLOIDS, LYOPHOBIC planes of platinum phlhalocj-aninc can be resolved in the electron microscope and are found to be 11.97 A apart (6). It is also pos- sible to check resolution by means of Fresnel fringes which are visible when' the objective lens is slightly off -focus (7, 8). The use of image intensifiers with the electron micro- scope may well prove useful in the future in the field of high-resolution work. Shadow Casting. Normally, only a two- dimensional aspect of the colloidal particle on the grid can be obtained. This is insuf- ficient for many purposes particularly if the 3-dimensional shape of the particle is re- quired. In order to enhance contrast and obtain an approximate idea of the vertical height and shape of the particle, shadowing with a heavy-metal vapor such as that of gold, platinum, chromium or uranium may be employed. The metal is evaporated onto the sample in a high vacuum at a suitable angle; usually a special evaporator unit is needed to meet the strongest requirements of high resolution work (1). From the known angle of shadowing and the length of the shadows, a three-dimensional picture of the particle shape can be built up. These shapes can be checked by constructing models and shadowing with a beam of light. IMore reli- able information can be obtained if the ma- terial is shadowed in two directions at right Ofn Fig. 3. Colloidal silver iodide particles sha- dowed with uranium at 50°, a) and c) in one direc- tion, b) and d) in two directions at right angles. Fig. 4. Diagram of particle models based upon shadowed micrographs of the type illustrated in Fig. 3. angles, but it is essential that the shadowing be very light; overdeposition of metal com- pletely obliterates the first shadow. In Fig. 3 micrographs of specimens shadowed with uraniimi at 50° in one direction only and with uranium at 50° in two directions at right angles are shown. From these micro- graphs particles were found to have shapes such as those shown diagrammatic ally in Fig. 4. Replication. One of the most useful tech- niques for determining the exact shape of particles and also their surface structure is replication; this technique is, moreover, in- valuable for the study of specimens which under normal conditions are decomposed by the action of the electron beam. The tech- nique is carried out by depositing samples of the sols either on clean glass slides or freshly cleaved mica and allowing them to dry. A thin film of carbon is then evaporated on to the slide; normally it is advantageous at this stage to shadow very slightly with a heavy metal vapor such as chromium. The carbon film is then removed from the slide using as a hquid substrate a solvent for the embedded particles. When this solvent is a concentrated salt solution it is advisable to follow by wash- ing with more dilute solutions of the salt and then to give a final rinse in distilled^ wa- ter. For the best results from the replication technique it is essential that the films em- ployed should be extremely thin {ca. 100-300 A). 127 ELECTRON MICROSCOPY Fig. 5. Carbon replicas of silver iodide parti- cles shadowed with chromium at 50°. Results obtained by the replica technique with subsequent shadowing are illustrated in Fig. 5. Distribution Curves. It is possible from electron micrographs of a field containing a large number of particles to determine not only their shapes but also the frequency distribution of diameters, or appropriate di- mension. Once these factors are established the surface area of a sample may be obtained, and from a knowledge of the physical density, the weight of the particles. Thus a distribu- tion curve can be made by plotting the fre- quency of appearance of a certain diameter against that diameter. A typical example is given in Fig. 6. Strictly, a number of photo- graphic plates should be taken of different fields and several thousand particles meas- ured in order to obtain a truly representative curve. In practice, however, counts are usu- ally made on 300 to 400 particles; for reason- able representation it is essential to take several micrographs of different parts of the field. Such a curve enables information to be obtained on the degree of polydispersity of the system with respect to size and shape and can be of great assistance in following rate processes such as nucleation, particle growth and coagulation. The shape of the size distribution curve depends on the rates of the nucleation and growth processes (see later). In general there is good agreement be- Iween the size of particles determined by electron microscopy and those determined by other methods. Determination of Absolute Particle Number per unit Volume. Two procedures can be employed for this determination. In the first the sol is sprayed on to the grid in the form of fine droplets using a high-pres- sure nebuUzer (9, 10). Under favorable con- ditions the drops are clearly visible and assuming the diameter of the dried drop to be the same as that in the spray, the drop volume can be estimated. A typical drop formed by spraying a polystyrene latex sus- pension is shown in Fig. 7. Thence from a count of the number of particles contained in the drop the number of particles per unit volume of the original sol can be calculated. In the second method it is essential for accurate results that a monodisperse sol should be used. Electron microscopy can be used to determine the diameter, or in the case of non-spherical particles the appropri- ate dimension, and the volume of a particle 40 30 % PARTICLES 20 lO - ! j— - ! I 1 i ! r^ ■ ! 1 1 — ■ i i i ! j 1 II; ill! ! — ! — 1 ' i '— ! ! !—l ! 1 1 1 . 1 300 600 900 I200 I5CXD PARTICLE DIAMETER IN A° Fig. 6. Particle size distribution curve for a sol of silver iodide. 128 COLLOIDS, LYOPIIOBIC calculated. Then if the particle density is known the weight per particle can be calcu- lated and if a subsidiary determination is made of the weight concentration of the sol used, the number of particles per unit volume can be calculated. Both methods have been used extensixely, but in some cases care must be exercised in applying these methods since it is not easy to determine drop diameters accurately, and the second method may yield spurious re- sults, either because of particle shrinkage (beam intensity too high), or because the particles do not possess a well-defined geo- metrical shape. A comparison has been made (11) between values for particle numbers per unit volume determined by electron microscopy, and those determined by other methods, such as, direct ultramicroscopic counting using a flow method, turbidity measurements and counts using a haemocytometer cell. The results of the comparison which was carried out using polystyrene latex particles are given in Table 1. In general it was found that the dry- weight method yielded results very close to those determined by other methods, whereas the spray method tended to give rather high results in terms of absolute numbers. A Table 1. A Comparison Between Particle Numbers per ml Determined by Electron Microscopy and by other Methods Polystyrene Latex Particles Method 0.216 m Diameter 1.029/1 Diameter Electron Microscopy, spray method Electron Microscopy, dr}^ weight Particle counter (flow method) Turbidity measure- ments Haemocytometer 5.60 X 10'^ 1.93 X 10" 1.90 X 10'3 2.53 X W 1.76 X lO'i 2.01 X W 1.26 X W 3.40 X 1011 method for the direct application of micro- drops of reproducible known volume is clearly desirable. Processes Involved and Destruction in Sol Formation Fig. 7. "Droplet" of polj'styrene latex parti- cles obtained by spraying with a Vaponefrin nebu- lizer. A variety of methods exist for the prepara- tion of colloidal dispersions of different tj^pes (3) and of these perhaps the most commonly used, and the most studied, is the mixing of two ionic solutions. A typical example is the formation of a sol of silver bromide by mixing silver nitrate and potassium bromide at concentrations sufficiently high to exceed the solubility product ; for a stable sol the stabi- lizing ions Ag+ or Br~ have to be present in certain proportions. In most cases of low- .solubility inorganic materials stable sols can be formed provided that a stabilizing ion is present . The nuclei originally formed in the solu- tion, which are usually very small crystals, grow to form the larger sol particles which are usually known as primary particles. This phenomenon is known as ageing, and is usu- ally explained as the growth of extremely small particles to form larger ones either by regular addition to the lattice, i.e., smaller particles going into solution so that the larger ones can grow at their expense, or by a process of ordering of the disordered lattice 129 ELECTRON M l( .IU)S( .( )I»Y of the particles originally formed. Sols of primary particles are often stable for long periods of time, but addition of electrolyte beyond a certain concentration, or the addi- tion of strongly adsorbed organic ions, causes the particles to clump together (coagula- tion); this process may, or may not, be ac- companied by recrystallization to form even larger particles, depending on the system. The process of sol formation and destruc- tion of the sol either by coagulation or recrystallization can be represented sche- matically in the following manner: Ions Nuclei growth (ageing) Large crystals (ageing)/^ J Primary particles \ Coagula Most of the stages represented in this scheme can be investigated by electron microscopy and will therefore be considered separately. Nucleation. Nucleation can be defined as the formation of a discrete particle of a new phase in a previously homogeneous solu- tion. Nuclear or amicronic sols contain par- ticles which cannot be resolved as discrete entities in the ultramicroscope. They can be prepared from many materials, e.g., gold, silver, silver iodide etc.. Electron micro- scopic examination of these sols shows par- ticles down to 10 A or less (see Fig. 2) which can be clearly resolved. The resolution of these particles constitutes one of the highest resolutions so far achieved with the electron microscope. From the point of view of col- loid chemistry this illustrates the small size range in which colloidal particles can exist and it is of considerable interest that the regular shapes of many of the particles ap- pear to be maintained down to the limits of resolution. Thermodynamically, the smaller particles would be expected to have a larger solubility than the larger ones and thus would be expected to go into solution as ions and deposit on the larger particles to increase their size. Charge would be expected to in- fluence this process (12), but in view of the large value of the free energy of most solid- liquid interfaces it is doubtful whether this does in fact play a significant role. Several theories have been proposed for the mechanism of nuclei formation in dilute solu- tion, of which the impurity, organizer and fluctuation mechanisms appear to have re- ceived the most attention. The impurity theory is based on the idea that nuclei are introduced into the system as foreign bodies, e.g., dust particles; it has been found, for example, that in the preparation of colloidal gold different sols are obtained according to the state of the glass vessel used. However, it was concluded by Turkevich, Stevenson and Hillier (13), who prepared gold sols un- der many different conditions, that impuri- ties were not a variable in their investigation. These authors proposed the organizer mecha- nism to account for the formation of nuclei in gold sols. Their suggestion was that the nucleating agent, e.g., hydroxylamine, grad- ually built up a complex between the gold ions, chemically binding a large number of gold ions and reducing agent molecules into large macromolecules. It was suggested that the latter underwent a molecular rearrange-i ment to give metallic gold and oxidation products of the reducing agent. Some sup- port was lent to this hypothesis by the na- ture of the reducing agents, but there are clearly many conditions under which such a mechanism cannot apply. The fluctuation theory of nucleation is probably that most widely accepted. It is based on the hypothesis that the formation of a nucleus occurs only when a statistical fluctuation of the ionic (atomic or molecu- lar) concentration brings a sufficiently large number of ions together to form a particle of thermod3niamically stable size. This theory has been shown to apply to the forma- tion of colloidal sulfur (14). Studies of Nucleation bv Electron 130 COLLOIDS, LYOPHOBIC Particles N No.of Portlcles Per Unit Volume. Particle Fig. 8. a) Particle size distribution curve for sol and b) nucleation curve obtained there- from. (After Turkevich, Stevenson and Hillier). Microscopy. Two methods can be used to examine nucleation by electron microscopy (13) firstly, quenching the nucleation process at different times and directly examining the samples and secondly examination of the particle size distribution curves of the formed sol. In the first method the reaction can be quenched at suitable time intervals either by addition of a suitable reagent to stop the reaction, or by a large dilution to slow the reaction down by several orders of magni- tude. Thus from a direct examination of the samples obtained at different times, particle size distribution curves can be obtained for each sample, and a nucleation curve con- structed. ,The second method, which is less tediotis than the first, was suggested by Turkevich, Stevenson and Hillier (13) and consists of determining the nucleation curve from the particle size distribution curve of the com- pleted sol. The method is based upon the assumption that the principal cause of spread in particle size of the sol is the spread in time in nuclei formation. Thus particles formed in the early stages of nucleation commence to grow immediately, while nuclei formed later have smaller sizes corresponding to a shorter growing time. In the final sol therefore the particles first formed have the larger size and the size distribution curve can be con- sidered as a distorted image of the nticleation curve. Thus if a particle size distribution curve of the type shown in Fig. 8a is con- sidered, particles of diameter Dx may be chosen, which correspond to the diameter at- tained by these particles at a time tx on the nticleation curve (Fig. 8b). Thus at a time tx there are Nt^ particles per unit volume, a number which can be expressed as the frac- tion of the number of particles eventually formed at infinite time, i.e., N{t). The total number of particles formed up to time tx is represented by the area shaded in Fig. 8a or the number of particles with diameter greater than D{tx) is given by. NiO = niD) dD 0(tx) Tiu'kevich, Stevenson and Hillier (13) found that the growth of gold nuclei was given by an eciuation of the form / = (a — log D)/b, where a was a function of the rate constant and of the time at which an arbi- trarily selected reference particle formed; h was proportional to the rate constant. From independent observations of a and h, nuclea- tion curves w^ere constructed from particle size distribution curves. The Growth Process. The formation of a sol involves two processes, formation of nuclei and growth of nuclei. If the formation of nuclei is slow and growth rapid the sol consists of a small number of large particles; if nuclei formation is rapid and growth slow the result is a large number of small parti- cles. When both processes are slow, a broad distribtition of sizes is obtained. In nuclea- 131 ELECTRON MICROSCOPY tioii the number of particles formed as a f miction of time is studied, whereas in the case of growth the rate of increase in the size of the particle with time is the important factor; strictly, in order to study growth the number of nuclei should be kept constant. One method of maintaining this condition in practice is to add nuclei to a slightly super- saturated solution of the growing species. The growth process which appears to have been investigated in most detail by electron microscopy is that of colloidal gold. Turke- vich, Stevenson and Hillier (13) took ad- vantage of the fact that in a slightly acid solution of chlorauric acid and hydroxyla- mine hydrochloride, in a very clean closed vessel, colloidal gold was not produced until a sufficient number of nuclei were intro- duced. Thus when the growth medium was inoculated with nuclei, the chlorauric acid was reduced by the hydroxylamine and the metallic gold was deposited only on the nuclei; hence the nuclei increased in size but not in number. The mean diameter of the resulting particles, Dg , was shown to be given by D, = Dn , 3/M„ -f Mc, M„ where Dn was equal to the mean diameter of the nuclei and Mn and Mce were the respec- tive masses of the metallic gold in the nuclei and ionic gold in the growth medium. From an examination of the particle size distribution curves obtained from the ki- netics of citrate reduction determined chem- ically, it was found that the growth law was of the form dt = kD where k was a constant dependent on tem- perature and reagent concentration but not on particle size. The growth of gold particles in monodis- perse gold sols produced by the action of sodium citrate on chlorauric acid has also been studied by Takiyama (15) using elec- tron microscopy. The size of a particle at a time t (minutes) was expressed by the mole number of one particle, x (mole), as calcu- lated from the mean particle diameter D by the relation, X = 4/37r(D/2)'p/M where p and M are the density and molecular weight of gold, respectively. It was found that the rate of growth was expressed by the equation, — = A;x2/3(x„ — x) , at where x and Xoa are the mole numbers of the particles at a time t and after completion of growth; k was a rate constant. It was found that the growth process was autocatalytic with respect to the surface of the gold parti- cles. The Ageing Process. A lyophobic sol is never stable in the thermodynamic sense and is always proceeding in the direction which involves a decrease in the surface free energy of the solid-liquid interface. Thus there is always present a tendency for the total surface area of the sol to decrease, until a pseudo-stable equilibrium is reached; at this stage the sol consists of dispersed pri- mary particles. The process of change from the initial sol, which may consist of a large number of small particles, to the final "sta- ble" sol which may consist of a small number of large particles is termed ageing; the dis- solving of smaller particles accompanied by growth of larger ones is sometimes termed Ostwald ripening (Fig. 9). The rate of ageing may be slow or fast, according to the condi- tions and the material employed. In the case of barium sulfate, for example, the ageing process even at room temperature is rapid and large crystals are formed ; it is difficult to prepare a very stable finely dispersed sol. On the other hand, in the case of silver iodide, sols of finely dispersed particles can be pre- pared which are stable for several years. The 132 COLLOIDS, LYOPHOBIC "^^:^^ -o ^^r-^ ""c^.. -. ' "-c"-"" . r* f*. X> ^ " ' ■ > V<' ;> '. "^^ VHI ' • r ■ ( , ^ ^. V ' f""' ' _^- , ' '">o ■G> • , ^ ■; • i -; Lf^u ■ »^ ^. ^ a i ^ J ^ ^^8 ABC Fig. 9. Sequence of electron micrographs of carbon replicas showing Ostwald ripening of silver iodobromide emulsion crystals in a solution containing gelatin and ammonium bromide at 50°C. a) immediately b) 5 minutes, and c) 20 minutes after mixing. (By courtesy Messrs Ilford Ltd.) ageing process is very important industrially, principally in the production of photographic emulsions, where the nuclear sols produced in the presence of gelatin are allowed to "Ost- wald ripen" before coating onto plates or film base. The ageing process in poh^disperse systems is usually considered to involve the smaller particles going into solution and the larger particles growing at their expense. An alter- native explanation is that coagulation of the small particles occurs followed by recrystal- lization of the coagula to form regular particles. Most evidence would appear to favor the former mechanism but in some cases mosaic crystals have been found (see for example Fig. 17a) which would tend to favor the latter. It is possible that in prac- tice both mechanisms occiu' with the former usually being the predominant one. Electron microscopy forms a suitable method for the examination of the ageing process since both the average size of the particles and the number present per unit volume of sol may be evaluated at a given time. Moreover, from the shape of the par- ticles formed during the ageing process, it is possible to tell whether growth of the parti- cles occurs preferentially in certain direc- tions. A detailed study of the ageing of silver bromide sols has been carried out by Kolthoff and his collaborators (16). An interesting example of the ageing proc- ess is found in the case of vanadium pentox- ide sols. The sol particles formed in the initial sol, e. g., in a Biltz sol (17) have been found to be small needles several hundred ang- stroms long and 140 A thick. On ageing these are transformed into fibrous crystals several microns in length. In detailed studies on the ageing process (18, 19) it was found that the large numbers of small needle-like particles were redistributed to give smaller nmnbers of large filaments; growth was attributed to recrystallization of the fibrils. Formation of Monodisperse Sols. Inti- mately connected with the study of nu- cleation and growth is the problem of pro- ducing monodisperse sols. The latter may be defined as sols in w^hich all the particles con- tained therein have exactly the same size and shape. The conditions for the preparation of monodisperse sols, which are verj^ important from the viewpoint of colloid chemistry, have been investigated in detail by LaMer and his collaborators (20), and may be il- lustrated by consideration of Fig. 10. Thus if a slow chemical reaction occurs which con- tinuously generates molecules of a disperse phase, the concentration of these molecules increases steadily, passes the point of satura- 133 ELECTRON MICKOSCOPY Concentration o( Otspcrtc PhQtc in Solution. Fig. 10. Schematic diagram illustrating the mode of production of monodisperse sols. (After La Mer). tion A and exceeds at B the level at which the rate of niicleation becomes appreciable. When the rate of production of molecules is slow, however, the sudden appearance of nuclei relieves the supersaturation so rapidly and effectively that the region of nucleation (II) is restricted in time and no nuclei are formed after the initial outburst. Hence the nuclei produced grow uniformly by a dif- fusion controlled process (region III) and a sol of monodisperse particles is obtained. If the initial solutions used are not very dilute, then the rate of production of mole- cules becomes so rapid that their concentra- tion in solution continually exceeds the saturation concentration (Co) and continuous creation of nuclei in addition to growth oc- curs. Thus a polydisperse sol is formed since the size of any particle depends upon the stage at which it was formed. A luimber of monodisperse sols have been prepared and investigated by electron mi- croscopy and other methods. The formation and properties of monodisperse sulfur sols were investigated by LaMer and collabora- tors (20) apparently without detailed exami- nation by electron microscopy. The mono- disperse sols most widely investigated by electron microscopy are undoubtedly those of polystyrene latex (21, 22) and sized samples of these particles are now widely used for the magnification calibration of electron micro- scopes. Watillon, Grunderbeeck and Hautecer (23) by reducing selenium oxide with hydrazine in the presence of amicronic gold particles produced monodisperse sols of selenium. Ex- amination by electron microscopy showed the particles to be almost perfectly spherical with a deviation from perfect sphericity of about ±2%; the spherical shape was con- firmed by shadowing experiments (see Fig. 11). Selected area micro-diffraction showed the particles to be essentially amorphous. ^Monodisperse barium sulfate sols have been prepared by Takiyama (24) by decom- posing the barium-EDTA complex with hy- drogen peroxide in the presence of am- monium sulfate. The particles were shown by electron microscopy to be spindle 6 Fig. 11. Monodisperse sols (a) electron micrograph of selenium sol particles {bij courtesy of Dr. A. Watillon), (b) carbon replica of silver bromide sol particles shadowed with chromium at 60°, (c) carbon replica of silver iodide sol particles shadowed with chromium at 60°. 134 COLLOIDS, LYOPIIOBIC shaped, and the mole number x per particle was calculated from the equation X = ^7r(a/2)(6/2)2p/M where a and h were the mean length of the long and short axes, respectively, and p and M were the density and molecular weight of barium sulfate. The size of the particles was found to increase with the concentration of the reagents used and the relation between the mole number of the particle and the concentration of the reagents (C) was given by the expression xC" = K where a and K were constants. ]\Ionodisperse gold sols have also been prepared by Takiyama (15) using the reduc- tion of chlorauric acid by sodium citrate. The sol particles thus prepared were found to be almost spherical, the average diameter being 172 A with a standard deviation of 13 A. Silver bromide and silver iodide sols have been prepared in a monodisperse form by Ottewill and Woodbridge (25). The silver bromide sols were prepared by slow cooling of a hot solution of silver bromide, which was slightly supersaturated at room temperature, and the silver iodide sols by dilution of the potassium iodide complex with water. The silver bromide particles were found to be cubes and the silver iodide particles rhom- boids; typical micrographs of both types of sol particles are given in Fig. 11. Effect of Additives on Sol Formation. It is well known that molecules of dyes or surface-active agents often show a preference for adsorption onto particular crystal faces (26) and thus can exert profound influences on crystal shape (27). Support for the idea of preferential adsorption on certain crystal faces has been found in electron microscope studies on the coagulation of sols. For exam- ple, in the coagulation of hexagonal plates of silver iodide by dodecylpyridinium ions (28) it was found that the plates were joined by , .^. a . ^a^u. ;:C ,»Jb Fig. 12. Electron micrographs illustrating the injfluence of additives on the formation of silver iodide sol particles, a) sol formed in the presence of 2.2 X 10"'' M mercaptotriazole, b) sol formed in the presence of 4 X 10" M dodecylpyridinium io- dide, c) sol formed in the presence of 7.74 X 10~^ M dodecylpyridinium iodide. edge to edge adhesion suggesting a preference for adsorption of the ion on the 0110, 1010, 1100, Olio, lOlO and IlOO faces. The influence of different media on the shape of sol particles can conveniently be investigated by electron microscopy. In Fig. 12a are shown particles of silver iodide formed in the presence of 2.2 X 10~^ M mercaptotriazole; comparison with Fig. 12b shows that the crystal form has been altered from predominantly flat plates or tetrahedral particles to rod-like particles. It is advisable when such changes are noticed to carry out micro-diffraction experiments on the parti- cles to determine whether any of the additive has been incorporated into the crystal. High concentrations of surface-active agents, particularly those above the critical micelle concentration, often have a consider- able influence on the sol formation process (28). In the case of silver iodide particles, it appears that a twofold adsorbed layer of sur- face-active agent ions is formed on the particles at the nucleus stage; the particles formed are finely dispersed with many of diameter less than 25 A (Fig. 12c). Owing to the strongly adsorbed layer the ageing proc- ess appears to be retarded and a protected 135 ELECTRON MICROSCOPY nuclear sol is obtained. Further growth is coagula obtained can be determined from slow and the sol shows a fairly narrow parti- micrographs. These are found to vary con- cle size distribution. siderably according to the system and coagu- An important additive to the silver halide lating agent employed, typical examples sols formed for coating photographic plates being edge to edge adhesion of flat plates, is gelatin. An electron microscope examina- irregular clumps, chains and massive "lace- tion of the growth of silver bromide particles like" aggregates of particles embedded in in gelatin has been carried out by Ammann- additive (see Fig. 13). Brass (29), and the effect of a number of other polymers of high molecular weight by ^he Structure of Sol Particles Perry (30). In the latter case the nature of Micro -diffraction Examination. With the polymer was found to have a profound an electron beam, as with an x-ray beam, a effect on the growth process and on the regular crystal lattice acts as a diffraction morphology of the crystals obtained. grating and gives rise to a diffraction pat- Studies on Flocculation. The floccula- tern. Thus the electron microscope may be tion of sols is usually considered to include used as a diffraction camera. In earlier both the process of coagulation, i.e., the ac- machines a different specimen holder was tual adhesion of the particles, and the subse- often used for diffraction but in many mod- quent processes of recrystallization etc. ern machines the specimen is left in the same Coagulation can often occur during the dry- position and hence diffraction and micros- ing down of specimens for electron micro- copy can be carried out consecutively on scopic examination and as such is of purely the same specimen. Moreover, with the elec- nuisance value. Direct electron microscopic tron optical arrangements available it is studies, however, do provide a useful method possible to select a specimen area of diameter of obtaining information on coagulation down to 2,000 A and to obtain a diffraction processes provided precautions are taken to pattern from this region alone. It is also pos- eliminate artefacts occurring during the sible to modify instruments so that diffrac- drying down process. tion patterns are produced under the same Electron microscope studies on the coagu- illuminating conditions as used for micros- lation of silver iodide sols have been carried copy (32). Hence by this means, single col- out by Mirnik, Strohal, Wrischer and Tezak loidal particles can be isolated and diffraction (31) and by Home, Matijevic, Ottewill and patterns obtained directly from them; in the Home (28). The former authors studied the case of thin plates with a cross-sectional dis- formation and coagulation of the sol as a tance greater than 2000 A it is possible to function of time and the latter authors the isolate particular regions for examination, effect of dodecylpyridinium iodide, at vari- Thus with colloidal particles which are single ous concentrations, on the sol formation crystals a symmetrical spot pattern is ob- process. tained (see Fig. 14). If the inclination of the So far electron microscopy does not appear single-crystal specimen to the electron beam to have been employed to carry out quanti- is changed, the diffraction pattern alters ac- tative studies on the kinetics of coagulation, cording to the amount and direction of the but it does form a useful method of confirm- inclination. This effect has been studied in ing whether effects observed in quantitative detail bySuitoandUyeda (33) using lamellar measurements by other methods, e.g., spec- gold crystals. For the detailed interpretation trophotometry, are really to be attributed to and analysis of micro-diffraction patterns coagulation or to other effects such as re- the relationship between the reciprocal lat- crystallization. Moreover, the form of the tice of the crystal and the Ewald sphere must 136 COLLOIDS, LYOPHOBIC K* i Fig. 13. Electron micrographs of varioiLS types of flocculation, a) side to side ad- hesion of hexagonal plates, b) formation of chains, c) formation of clumps, d) "lace- like" aggregates of particles embedded in added surface active agent. be considered, since the section of the former by the latter can be regarded, ap- proximately, as the electron diffraction pat- tern obtained. Apart from the direct use of diffraction patterns to obtain the structm-e of the par- ticles, it is very useful in the colloid chemical field to use the "x-ray" structure of the bulk material, if known, to identify the colloidal particle, or confirm its identity, and to de- termine the orientation of the particle; moreover, the indices of the crystal faces ex- posed can be obtained from the disposition of the spots. Diffraction analysis can also be used to give an indication of the imperfec- tions present in the particle. The main limitation to the use of this technique on crystalline colloidal particles is the thick- ness of the particles since for more than a certain thickness of the particle, which varies according to the nature of the material, too much of the beam is scattered, or absorbed, for a distinct pattern to be obtained. In Fig. lob is shown a selected area micro- diffraction pattern from a thin colloidal par- ticle of silver iodide (thickness ca. 100-200 A) and in Fig. 15a a selected area micrograph of the portion of the particle from which the diffraction pattern was obtained. The pat- tern is that expected for a hexagonal crystal of silver iodide of a = 4.59 A and c = 7.49 A resting on the OOOi plane. The clarity of the diffraction pattern demonstrates clearly the almost perfect crystalline nature of colloidal particles of this type. If instead of a single particle, a field is taken containing a number of small particles then a ring pattern is obtained (see Fig. 14b). Only those planes which satisfy the Bragg equation contribute to the pattern and thus a series of discrete rings are obtained; if only a small number of particles are present the rings are broken up into spots. A typical ring pattern, obtained from a group of silver iodide particles of particle size 300-400 A, is shown in Fig. IG. The angular breadth of the diffraction line depends upon the diame- ter of the particle D and the wavelength of the incident radiation X, the quantity \/D usually being termed the Scherrer breadth. Thus theoretically an estimate of particle size can be obtained from the breadth of the diffraction lines (2, 34). However, other facts 137 ELECTRON MICKOSCOl'Y Electron Beam Electron /O o -3020 Beam \/ B ^-^ 20fO 21 10 1 1 So OOOJ loTo (a) (b) Fig. 14. Schematic representation of micro-diffraction, a) diffraction pattern ob- tained from a single crystal particle, b) ring pattern obtained from a collection of colloidal particles. Fig. 15. Micro-diffraction pattern from a col- loidal particle of silver iodide, a) selected area of particle, b) micro-diffraction pattern from this se- lected area. such as finite aperture of illumination, etc. play a part in line broadening. Dark Field Image Analysis. A useful complement to micro-diffraction experiments is dark-field image analysis. This method was developed by Mollenstedt and others (35, 36? 37) and has principally been applied to lamellar crystals. The technique consists of isolating a single spot on the diffraction pattern by means of an aperture in the plane of the objective diaphragm, from which the part of the electron beam, focused as the spot, passes through the small aperture; the latter is then moved aside from the normal position on the axis into the position for dark field. Thus a dark-field image can be obtained which corresponds to the portion of the crys- tal containing the lattice planes from which diffraction was actually taking place. In this manner each spot of a single crystal diffrac- tion pattern can be studied individually and 138 COLLOIDS, LYOPHOBIC a set of dark-field images obtained which correspond to the sections of the crystal con- taining the diffracting lattice planes. Com- parison of the dark-field images with the bright -field images reveals that each black line in the latter becomes a bright line in the former, and thus the Miller indices of the lattice planes giving rise to the bright lines in dark field can be identified. An in- teresting study using this technique has been carried out by Suito and Uyeda (33) on lamellar colloidal gold crystals, in which they were able to identify the crystal planes giv- ing rise to the striped patterns often observed on thin gold crystals (see Fig. 18). Studies of Internal Structure in Thin Colloidal Particles. Many colloidal parti- cles, when studied by direct electron micros- copy, appear completely opaque, and there- fore any interior details which might be expected to be visible, such as grain bound- aries from mosaic growth, dislocation lines, etc., are completely obscured. Dislocation lines usually appear dark on micrographs, an increase of contrast which is thought to be due to the increased Bragg reflection from the strained region around the dislocation line. Moreover, studies on thin metal foils of aluminum, stainless steel, etc., have shown that such imperfections are readily visible in the electron microscope (38, 39). Very few direct observations have been recorded on the presence of these defects in colloidal par- ticles, although such defects may play an Fig. 16. Micro-diffraction pattern from a group of colloidal silver iodide particles, a) selected area of particles, b) micro-diffraction pattern. important role in the stability of colloidal systems. In fact, observation of these defects in colloidal particles does not appear to be an easy problem and may well mean that the particles are very close to perfect crystals. For observations of interior structure it would appear necessary for the thickness of the particles to be of the order of onlj^ a few hundred angstrom units. Thin crystals may often be prepared by the controlled ageing of nuclear sols; this method is particularly successful in the case of silver iodide and some detailed studies on thin hexagonal plates of this material have been carried out by Home and Ottewill (40, 41). Some crystals were found to exhibit a mosaic appearance, sections of different con- trast, which would correspond to different crystal orientations, being clearly visible (Fig. 17a); extinction contours were also clearly visible. In other crystals dark bands were observed (Fig. 17b) which appeared to be due to the presence of dislocations or stacking faults; these were often seen to migrate across the crystal under the influ- ence of the beam. In the case of thin lamellar particles, such as those of gold, striped or spotted patterns are often obser\-ed; a typical example is shown in Fig. 18. These patterns are thought to arise from diffraction effects due to local curvature of the crystal, that is the sub- strate with which the crystals are in close contact undergoes local curvature in the form of a valley; the crystals follow this curvature and have a common axis of bend- ing along the \'alley. The thin strips which accompany the central one can be considered to arise from reflections which correspond to the subsidiary maxima which surround the main diffraction spots (33). The crystal planes giving rise to these diffraction effects can be identified by dark-field image analy- sis. The displacement of the subsidiary maxima from the main spot, which precisely satisfies the Bragg condition, is closely re- lated to the thickness of the crystalline 139 ELECTKO\ MICROSCOPY Ol M. I ^-H Fig. 18. Patterns formed on lamellar gold crys- tals due to diffraction effects. Fig. 17. Electron micrographs of thin colloidal particles of silver iodide showing the presence of imperfections, a) mosaic crystal, b) dislocation lines and hexagonal dislocation loop. ness of such crj\stals (33). Reasonable agree- ment was obtained between the thickness of particles obtained by this method and that suggested by shadowing experiments. Examination of the Surface Structure of Colloidal Particles by Replication. Closely related to the problem of internal structure of particles is the surface structure, since the history of the mode of growth of a particle is often recorded in its surface, in the form of growth spirals, kink sites, twin plane grooves, etc. Such information can usually best be obtained by an examination of very thin replicas of the particles shad- owed very lightly with a heavy metal ^'apor (42). However, in the case of very thin crys- tals, composed mainly of elements of low atomic weight, it is often possible to detect details of surface structure by direct micros- copy or by lightly shadowing the specimen with a heavy metal vapor. Typical exam- ples obtained with a metal detergent crystal, barium tetradecyl sulfate, are shown in Fig. 19. The spiral terraces of the plate-like crys- tals are clearly visible, the lengths of the shadows indicating the depth of the steps to be of the order of 50 A, i.e. bimolecular, and indicate that growth probably occurs by a screw dislocation mechanism. Similar effects have been observed with single crystals of the paraffin n-nonatriacontane and stearic acid (43). In the latter case the crystals were I ^ t Fig. 19. Electron micrographs of barium tetra- decyl sulphate crystals showing surface structure, a) shadowed with chromium at 50°, b) direct micrograph. particle. This fact has been utilized as a method for the determination of the thick- 140 COLLOIDS, LYOPHOBIC replicated with silicon monoxide and the rep- licas shadowed with palladium vapor. Dis- location centers were clearly visible, indicat- ing that growth had occurred by a screw dis- location mechanism ; the spiral step heights were found to be of the order of 45 A (see Fig. 20). An interesting example of the use of elec- tron microscopy in elucidating the growth mechanism of crystals occurs in the case of silver bromide. It was suggested by Berri- man and Herz (44) that the reason for tabu- lar growth in silver bromide (sodium chlo- ride-type structure) was the occurrence of twinning on the 111 plane; some confirma- tion was found for this hypothesis in that Laue photographs taken by them exhibited six-fold symmetry. Additional support for the mechanism was obtained by Hamilton and Brady (45) in an electron microscope ex- amination of shadowed carbon replicas of tabular silver bromide crystals. The replicas were examined at an angle of 45° to the in- cident beam when the convex and concave intersections at the twin planes were visible on the edges of the crystal. Examples of crystal replicas showing the presence of twin planes are given in Fig. 21. The fact that in the immediate vicinity of a crystal imperfection the lattice has a higher chemical potential than in the more perfect parts means that when the crystal is placed in a solvent preferential attack occurs at this point. Thus, provided the reaction is stopped before extensive solution of the crystal oc- curs, etch pits are formed at the sites of preferential attack. In the case of colloidal particles this technique appears to have been employed primarily on silver halide crystals in an attempt to detect dislocation sites. For example, Hamilton, Brady and Hamm (46) carried out chemical etching-studies on large grains of silver bromide and sih^er bromoio- dide emulsions. Etching was carried out by immersion of the particles, for a hmited time, in a silver halide solvent. The particles were then replicated with carbon, and the replicas Fig. 20. Silicon monoxide replica of a stearic acid crystal, shadowed with palladium, showing growth from a single dislocation and bimolecular steps. {By courtesy of Dr. I. M. Dawson) Fig. 21. Carbon replicas of colloidal particles showing evidence of twinning, a) silver iodide, b) silver bromide. Arrows indicate the position of twin planes. examined after shadowing with platinum- palladium at a 5 to 1 angle. The concentra- tion of chemical etch pits and the geometry of the pits were found to be dependent on the solvent used; potassium bromide gave octa- hedral pits and sodium sulfite and potassium cyanide dodecahedral pits. A general increase in etching on certain faces was found after intentionally straining the grains, but the effect was-not sufficiently strong to establish a one-to-one relationship between etch pits and dislocations. The etching experiments did not provide any conclusive e\'idence that normal grains of silver bromide were poly- crystalline in nature. A micrograph of a carbon replica of a sil- ver bromide crystal etched with potassium cyanide is shown in Fig. 22. 141 ELECTRON MICROSCOPY Fig. 22. Carbon replica of a silver bromide par- ticle etched with potassium cyanide. (By courtesy of Dr. J. F. Hamilton) Stability of Colloidal Particles in the Electron Beam. One of the difficulties en- countered in electron microscopy is that the amount of energy which is transferred from the beam to electrons in the specimen can often be in excess of the chemical binding energies. Hence, precautions must be taken against possible decomposition of the speci- men in the beam. Stability depends on whether, after excitation of electronic energy levels, the substance reverts to its original structure or to a new configuration. Thermal stability is also important, since consider- able temperature changes of the specimen can occur; temperatures of the order of 100°C or more can easily be obtained. In many cases decomposition of the specimen in the beam is rapid, making direct examina- tion impossible; For example, both silver chloride and silver bromide rapidly decom- pose to yield a mass of metallic silver (47, 48); thus replica techniques are normally used for examination of these crystals (49). The interaction of the specimen with the beam, however, may frequently be helpful in studying the decomposition of crystals. It was shown by Sawkill (50) that single crystals of silver azide could be decomposed in the beam of the electron microscope and that the decomposition could be followed in (l(;1ail by examining micro-diffraction pat- terns and electron micrographs taken at various stages. In a similar type of study Goodman (51) has examined the dehydration of single crystals of magnesium hydroxide to magnesium oxide imder the influence of the electron beam. Electron microscope observa- tions have also proved useful in studies on the motion of electrons and holes in photo- graphic emulsion grains (52) and in studies on latent image formation (53). A useful technique for examining dynamic changes in crystals of colloidal dimensions is to employ a cinecamera to record the image obtained on the fluorescent screen of the microscope. The first observations of this type were carried out by von Ardenne (54) using a camera fitted into the microscope, and later by Preuss and Watson (50) using an external cinecamera. In recent studies on the movement of dis- locations in thin metal films (38, 39) and dynamic changes in silver iodide particles (40, 41) cine recordings were made of the phenomena taking place. In this work the fluorescent viewing screen was tilted at a suitable angle and recordings were made by photographing directly through one of the observation windows at microscope magnifi- cations of 40,000 X or 80,000 X. A Kodak cine special camera was used and modified to take a 1 in. f/0.95 Angenieux lens at a work- ing distance of 15-20 cm; a speed of 16 frames per second was generally used. The studies on silver iodide were made directly on colloidal particles. Two types of effects were noticed under the influence of the beam — mobile changes of contrast within the particles and filament growth from the particles. The first effect was obtained with hexagonal plates of silver iodide. Changes of contrast were observed under the in- fluence of the electron beam which were highly mobile and migrated within the par- ticle boundaries at rates dependent on the beam intensity. The nature of these changes 142 COLLOIDS, LYOPIIOBIC and the speed with which they occurred are illustrated in Fig. 23. Particles are shown which have undergone considerable changes within the period of two frames of cine film (He sec). The electron-transparent regions moved very rapidly inside the crystal with- out effecting any change in the external shape, and continued to do so indefinitely under constant beam conditions. With certain types of silver iodide parti- cles, mainly the tetrahedral variety, fila- r i I- lOOO A' H Fig. 23. Sequence from a cine film showing a silver iodide particle undergoing changes of con- trast. ^_OJj^ Fig. 24. Filament growth from a single tetra- hedral particle of silver iodide. ments appeared to be pushed outwards from the interior as the intensity of the electron beam was increased. Fig. 24 illustrates typi- cal filament growth from the interior of a particle. In many cases the filaments ap- peared to be ribbons with widths as low as 30 A; these remained, however, quite rigid and were able to push holes in the support- ing membrane. Some filaments also show^ed well-defined contrasting bands (see Fig. 24). Filament growth was also observed in some rather coagulated regions which received strong electron irradiation. A sequence from a cine film of filament growth is given in Fig, 25. Many other dynamic processes should be amenable to investigation by electron mi- croscopy using this type of technique. General Morphology of Colloidal Par- ticles. The number of substances which can be prepared in the colloidal state is very 143 ELECTROiN MICROSCOPY Fig. 25. Sequence from a cine film showing the growth of filaments from irradiated silver iodide particles. large. Furthermore, the size, shape and struc- ture of particles vary considerably from material to material. No attempt has been made in this article to cover all the work carried out on colloidal particles nor to con- sider in detail the question of general mor- phology. This subject has been reviewed elsewhere (56) and it was concluded that the morphological forms in which colloidal par- ticles are formed could be subdivided into amorphous small particles, small regular forms (spheres, cubes, hexagons, octahedra, etc.), fibers and plates. Most of these forms have been described in this article but the main emphasis has been laid upon the de- scription of techniques which enable any colloidal particle to be examined in the elec- tron microscope. Acknowledgments. It is a pleasure to record my thanks to Mr. R. W. Home for his continuous en- thusiastic collaboration in much of the work described in this article. I should also like to express my thanks to Drs. I. M. Dawson, G. F. Hamilton and A. Watillon for generously supply- ing the micrographs acknowledged in the text. REFERENCES 1. CossLETT, V. E. AND HoRNE, R. W., Vacuum, 5, 109 (1955). 2. Hall, C. E., "Introduction to Electron Mi- croscopy," McGraw Hill Book Co., New York, 1953. 3. Kruyt, H. R., "Colloid Science," Vol. I, Elsevier, Amsterdam, 1952. 4. VON BORRIES, B. AND Kausche, G. A., Kol- loid-Z., 90, 132 (1940). 5. Ottewill, R. H., and Horne, R. W., Kol- loid-Z., 149, 122 (1956). 6. Menter, J. W., Proc. Roy. Soc, A236, 119 (1956). 7. Hillier, J. and Ramberg, E. G., /. Appl. Phys., 18, 48 (1947). 8. Haine, M. E. and Mulvey, T., J . Sci. Instr., 31, 326 (1954). 9. Williams, R. C, and Backus, R. C, /. Am. Chem. Soc, 71, 4052 (1949); J. Appl. Phys., 21, 11 (1950). 10. G^ROVLD, C.B.., J. Appl.PhTjs., 21, 183 (1950). 11. Ottewill, R. H. and Wilkins, D. J., J . Colloid Sci., 15, — (1960); in press. 12. Knapp, L. F., Trans. Faraday Soc, 17, 457 (1922). 13. Turkevich, J., Stevenson, P. C, and Hil- lier, J., Disc. Faraday Soc, 11, 55 (1951). 14. LaMer, V. K. AND Kenton, A. S., /. Colloid Sci., 2,257 (1947). 15. Takiyama, K., Brdl. Chem. Soc. Japan, 31, 944 (1958). 16. Kolthoff, I. M. AND Bowers, R. C, J. Am. Chem. Soc, 76, 1503 (1954). 17. Biltz, W., Ber., 37, 1095 (1904). 18. Takiyama, K., Bull. Chem. Soc. Japan, 31, 369, 555 (1958). 19. Kerker, M., Jones, G. L., Reed, J. B., Yang, N. P., and Schoenberg, M. D., J. Phys. Chem., 58, 1147 (1954). 20. LaMer, V. K., Ind. Eng. Chem., 44, 1270 (1952). 21. Harkins, W. D., /. Am. Chem. Soc, 69, 1436 (1947). 22. Bradford, E. B., Vanderhoff, J. W., and Alfrey, T., /. Colloid Sci., 11, 135 (1956). 23. Watillon, A., van Grunderbeeck, F., and Hautecler, M., Bull. Soc. Chim. Belg., 67, 5 (1958). 24. Takiyama, K., B^lll. Chem. Soc. Japan, 31, 950 (1958). 25. Ottew^ill, R. H., and Woodbridge, R. F., in press. 26. Buckley, H. E., "Crystal Growth," 458, Wiley, New York (1952). 144 CRYSTAL LATTICE RESOLUTION 27. Rehbinder, p., Disc. Faraday Soc, 18, 151 (1954). 28. HoRNE, R. W., Matuevic, E., Ottewill, R. H., AND Weymouth, J. W., Kolloid-Z., 161, 50 (1958). 29. Ammann-Brass, H., Chimia, 10, 173 (1956). 30. Perry, E. J., J. Colloid Sci., 14, 27 (1959). 31. MiRNiK, M., Stromal, P., Wri.scher, M., and Tezak, B., Kolloid-Z., 160, 146 (1958). 32. Reicke, W. D., "Proc. First Regional Euro- pean Conference on Electron Microscopy," 98, Stockholm, 1956. 33. SuiTO, E. and Uyeda, N., "Proc. Interna- tional Conference on Electron Microscopy," 223, London, 1954. 34. Heidenreich, R. D., Phys. Rev., 62, 291 (1942). 35. Mollenstedt, G., Optik, 10, 72 (1953). 36. Rang, O., Z. Phys., 136, 465, 547 (1953). 37. Ito, K. and Ito, T., J. Electronmicroscopy (Japan), 1, 18 (1953). 38. HiRscH, P. B., HoRNE, R. W., and Whelan, M. J. Phil. Mag., 1, 677 (1956). 39. Whelan, M. J., Hirsch, P. B., Hornb, R. W. AND BoLLMANN, W., Proc. Roy. Soc, A240, 524 (1957). 40. Horne, R. W. and Ottewill, R. H., /. Phot. Sci., 6, 39 (1958). 41. Horne, R. W. Ottewill, R. H., "Fourth International Conference on Electron Mi- croscopy," Berlin, Springer Verlag, 1, 140 (1960). 42. Bradley, D. E., Brit. J. Appl. Phys., 10, 198 (1959). 43. Anderson, N. G. and Dawson, I. M., Proc. Roy. Soc, A218, 255 (1953). 44. Berriman, R. W. and Herz, R. H., Nature, 180, 293 (1957). 45. Hamilton, J. F., and Brady, L. E., /. Appl. Phys., 29, 994 (1958). 46. Hamilton, J. F., Brady, L. E. and Hamm, F. A., /. Appl. Phys., 29, 800 (1958). 47. Levenson, G. I. P. AND Tabor, J. H., Sci. and Indust. Phot., 23, 295 (1952). 48. Klein, E., Mitt. Forsch. Agfa, 10 (1955). 49. Hamm, F. A., and Comer, J. J., /. Appl. Phys., 24, 1495 (1953). 50. Sawkill, J., Proc Roy. Soc, A229, 135 (1955). 51. Goodman, J. F., Proc Roy. Soc, A247, 346 (1958). 52. Hamilton, J. F., Hamm, F. A., and Brady, L. E., J. Appl. Phijs., 27, 874 (1956). 53. Hoerlin, H. and Hamm, F. A., J. Appl. Phys., 24, 1514 (1953). 54. von Ardenne, M., Z. Phys., 120, 397 (1943); Kolloid-Z., 108, 195 (1944). 55. Preuss, L. E. and Watson, J. H. L., /. Appl. Phys., 21, 902 (1950). 56. TuRKEvicH, J. AND HiLLiER, J., Atiol. Chcm., 21, 475 (1949). R. H. Otteavill CRYSTAL LATTICE RESOLUTION The ultimate goal of electron micro.scopy is, of course, the resolution of atoms in any structure, and this possibility has been ana- lyzed theoretically by several of the eminent workers. If atoms or molecules are regularly arranged in a crystal lattice there is a much stronger chance of resolved image formation than in the case of two isolated atoms be- cause there are definite phase relationships between electrons scattered from neighbor- ing noncoherent atoms. The resolving power of the best electron microscopes produced in the world is limited first by diffraction error and spherical error to about 2.8 A, and chro- matic error and astigmatism increase this to at least 7 A. At this value it should be possi- ble to observe crystal lattices in crystals with fairly large lattice parameters of the order of 10 A or greater. Great success prior to 1956 in the investi- gation of macromolecular crystals of viruses and proteins by the replica technique had been achieved by Wyckoff and associates at the National Institutes of Health (3). The surface of a needle-shaped crystal of the jack bean protein concanavallin, with molecule weight 42,000, among the smallest thus far replicated, reveals a rectangular net of about 62 X 87 A of particles (30-40 A in diameter) which are not in contact. The most likely next step downward in dimensions would be for crystals of organic molecules of inter- mediate molecular weights, sufficiently thin and properly oriented for direct transmis- sion micrographs. Menter (1) was the first to produce elec- tron micrographs of the lattice planes in crystals. He was particularly fortunate in his choice of metal phthalocyanins, especially 145 ELECTRON MICROSCOPY Fig. 1. A packing drawing of the nickel phtha- locyanine structure viewed along the bo-axis. The metallic atom is black, the nitrogen atoms are line-shaded. (R.W. G.Wyckoff,'' Crystal Structures'^) the copper and platinum derivations of phthalocyanin, a blue dye with a flat ring structine (Fig. 1) and the metal atom in the center. The crystal structure of platinum phthalocyanin may be considered to com- prise fairly widely spaced planes of heavy metal atoms (c^2oi = 11.94 A) embedded in a matrix of the light elements nitrogen, car- bon and hydrogen. The thin ribbon habit of the crystal is such that when supported on a specimen grid the 20l planes are almost parallel to the electron beam. The electron micrograph at a magnification of 1,500,000 (Fig. 2) shows the clearly resolved planes spaced 11.94 A apart. Similarly the copper phthalocyanin shows a spacing of 9.8 A. These parallel lines are the image of the projection of the 20l planes seen edge on. Imperfections are seen sometimes in the form of edge dislocations (Fig. 2) caused by in- complete planes. It has been possible also to resolve the 111 planes of the inorganic crys- tal sodium faujasite, a silicate with the com- position 2 Al2O3-CaO-Na2O10SiO2-20H.,O, with a spacing of 14.37 A. The mechanism of image formation de- pends upon the fact that the ihiii crystals form a cross grating diffraction spectrum. With a 50-mi(*ron objective aperture the spectra contributing to the image from the {)hthalocyanin crystals are (20l), (402), (201) and (402). These spectra recombine with the zero order beam in the image plane and form an image of the crystal grating in accordance with the simple Abbe theory of image formation by a lens. Calculations by Menter suggest that it should be possible to resolve planes with spacings considerably smaller than 10 A providing the divergence of the illuminating beam is made sufficiently small. Figs. 2. (a). Electron micrograph of part of crystal of platinum phthalocyanine showing image of lattice planes 12 A apart. Magnification 1,500,- OOOX. (b). Part of crystal of platinum phthalocy- anine showing edge dislocation. The dislocation line is perpendicular to the plane of the paper. Magnification as before, (c). Sketch copied from (b) showing exact position of extra plane of mole- cules. (Menter) 146 ELECTRON OPTICS If this is accomplished in practice there an extra dark band caused by osmium at- will be new possibilities for studying crystal tached to the double bonds, structures and detecting imperfections. The The RIAS workers have compared sepa- images are poor Fourier projections and fail rate micrographs of multilayers made from to reveal detail obtained by x-ray analysis; Cis, C22; and C36 acids salts. Periodicities of but they do reveal the exact location of im- the bands show ratios of 16 to 22 to 36. With perfections, thus permitting the direct study longer chain barium soaps, the lighter bands of the behavior of dislocations in solids under increase in width while the dark bands (me- a variety of physical conditions. tal ions) remain constant. The width of the Very recently, as reported in the July 4, dark bands is greater than the actual thick- 1960 Chem. Erig. News, it has been demon- ness (2 to 3 A) of the metal ion double strated that electron microscopy can show sheath, but resolution of the electron micro- the actual structure of soap multilayers at scope goes down only to 8 to 15 A. the molecular level. Two workers at the The band image of the micrographs repre- Research Institute for Advanced Stud\^ in sents actual ''multilayer architecture" and Baltimore, Md., have made micrographs that not merely interference fringes. For one picture highly regular sequences of Hght and thing, structures of 40, 80, or 120 double dark bands. And the spacing of these bands layers of soap molecules (made of saturated conforms with the length of fatty acid chains fatty acids) show exactly 40, 80, or 120 that make up the soaps, they find. bands, as the case may be. Moreover, such This combination of electron microscopy details as line dislocations in the micrographs and film techniques is the first observational are noted as in the case of the metal phthalo- method to confirm the double-layer spacing cyanins. of multilayers. Dr. Hans J. Trurnit told the references 34th National Colloid Symposium sponsored , ,^ ^ „. ,,„ , ,;r. t^ 1 1 4 ^01 T^- • • r ^11 • 1 ^1 • 1- Menter, J. W., "Electron Microscopy, Pro- by the ACS Division of Colloid Chemistry, ceedings of Stockholm Conference, Sept. at Lehigh University. The double spacing 1956," Academic Press, N. Y., 1957, p. 88. (previously derived from optical and x-ray 2. Neider, R., Ibid., p. 93. methods) results from a head-to-head, tail- ^- Wyckoff, R. W. G., Koninkl. Nederl. Akad. to-tail arrangement when the layers are laid W^tenschappen, Amsterdam, Proc, B59, 449 down on the surface. Dr. Trurnit and his associate, George George L. Clark Schidlovsky, build the multilayer soap films on methacrylic ester slides. Then they expose DISLOCATIONS IN METALS. See TRANS- sample strips of the slides to osmmm tetrox- MISSION ELECTRON MICROSCOPY OF ide (contrast inducer and fixing agent), im- METALS-DISLOCATIONS AND PRECIPITA- bed them in the polymerizing plastic, and tION d 291 slice them into thin sections (about 500 A.). They could not use free fatty acids because of ELECTRON OPTICS: ELECTRON GUN AND their solubiUty in the plastic bed. ELECTROMAGNETIC AND ELECTROSTATIC Dark bands in the electron micrographs lcinoco correspond to sheaths of metal ions in the The electron optical system of the electron soap layers, when the metal used has a high microscope comprises (a) an electron gun in enough atomic number to give electron scat- which electrons emitted from the tip of a hot tering. Magnesium ions do not register, but tungsten hairpin-shaped filament are ac- calcium, barium, and other such ions do celerated to produce a narrow conical di- show up. Also, unsaturated fatty acids show vergent beam of electron illumination to ir- 147 ELECTRON MICROSCOPY radiate the object; (b) a condenser lens to concentrate the beam onto the object and in- corporate an aperture system to control the angular aperture of the irradiating beam; (c) an objective lens and objective aperture followed by one or two projector lenses to form an image of the object on a fluorescent screen or photographic plate magnified usu- ally between 1,000 and 200,000 times. The condenser lens may be a double one to allow the reduction of the area of illumination to minimize heat dissipation at the object. The objective lens must have the minimum spher- ical and chromatic aberration and astigma- tism consistent with meeting the geometric requirement of allowing adequate space for the object holder and objective aperture, and their manipulation. The projector lenses must be designed to give the widest range of magnification possible without image dis- tortion. The Electron Gun The electron gun used in the electron microscope utilizes a tungsten wire hairpin cathode, an apertured grid and an anode. An example of a typical electrode system is shown in Figure 1. The anode is at ground potential and the cathode at the full 50-100 kv accelerating potential. The grid is oper- ated with a negative bias of a few hundred volts with respect to cathode. To obtain even barely adequate final im- age intensity at magnifications of 50 to 100,- 000 required for high resolution working, the requirements for gun performance are strin- gent. Since the illumination beam angle is limited to give optimum resolving power, the main gun requirement is to give the maxi- mum possible current density per unit solid angle of emergent beam. Langmuir (1937) has shown that a theoretical limit to the cur- rent density per unit solid angle (or "bright- ness") is imposed by the spread in the emis- sion velocities of the electrons from the cathode. The limiting value of brightness (iS) is given by: (3 = pco/TrkT a) where pc is the emission current densitj^, cpo the accelerating voltage, k = Boltzman's constant (8.6 X 10"^ e.V./°K) and T the absolute temperature. It can readily be shown that the current density obtainable in a focused image of the virtual gun source is independent of the mag- nification of the focusing system where the beam angle is fixed. Haine and Einstein (1952) have shown that the electron micro- scope gun will give this theoretical bright- ness, for a wide range of geometrical con- figurations of the electrodes, provided op- timum bias conditions are maintained. A certain choice of geometrical configurations will give a narrower angle of divergence of the beam and hence conserve total current. Thus, increase of the cathode shield diame- ter and reduction of the height of the cathode behind the shield both reduce beam angle and therefore total current. Under such con- ^^ ATHODE SHIELD 7^ / / NODE (a) FLAT SHIELD Cb") RE-ENTRANT SHELD Fig. 1. The typical geometry of an electron gun. 148 ELECTRON OPTICS ditions the optimum bias potential is in- creased but the brightness and hence the image intensity is unchanged. Haine, Einstein and Borcherds (1958) have discussed the use of automatic bias, i.e., the generation of bias potential by the po- tential drop across a resistance connected in series with the high voltage supply. Apart from the simplicity of this arrangement, the negative feedback action of the gun mutual conductance and series resistor gives a high degree of stabilizing action to the beam cur- rent. Further restrictions are imposed on the choice of geometrical configuration to en- sure that the bias potential is maintained at the optimum value. To obtain adequate image intensity for very high resolution working (3 to 10 A), the electron gun cathode must be operated at an emission density of 1 to 3 amp/cm-. This emission current density requires an operat- ing temperature at which the tungsten cath- ode life, as limited by evaporation, is only a few tens of hours (Bloomer 1957). Electron Lenses The requirements of the electron micro- scope are met with magnetic or electrostatic lenses which provide a "bell-shaped" magnetic field or electrostatic potential distributions along the axis. The magnetic lens comprises a solenoidal excitation coil wound on an axi- ally symmetric kon circuit with a gap in the core and an axial hole bored to allow the passage of the electrons. Basically the lenses comprise parallel pole pieces spaced S cm apart and an axial hole diameter D cm with an excitation NI ampere-turns applied be- tween the pole pieces (Figure 2). Little or no advantage is gained by changing the shape of the pole pieces from this simple geometry. The electrostatic lens comprises basically three parallel plate electrodes with axial holes. The outer plates are at ground poten- tial and the central one at the negative po- tential of the electron gun cathode. The elec- trodes are usually shaped to minimize the surface electric field strength to avoid flash- overs, but the lens theory applied to the O CWDI 002 0-03 0-04 005 0-0 6 Vr/CNI)' Fig. 2. Curves showing the focal properties of magnetic lenses as functions of the excitation param- eter. Vr/{NI)\ 149 ELECTRON MICROSCOPY simple sliape gives results of adequate ac- curacy for practical purposes (Archard 1954). The relations between the geometric fac- tors, applied excitation and the focal proper- ties and aberrations have been calculated and measured by many different workers (e.g., Glaser, 1941, Ramberg, 1942, Lenz 1950, Liebmann andGrad, 1951). The results are now well established and it is more rele- vant to describe these results than the meth- ods for deriving them. The results are more complete for the magnetic than for the elec- trostatic lens, but those for the latter are adequate to show its inferiority. All the rele- vant properties are given in a series of uni- versal curves derived from data calculated by Liebmann and Grad (1951). The first order focal properties of impor- tance are the focal length (Jo , for objective and/i , for projector) and focal distance (zq). These parameters can be expressed in the simple universal curves of Figure 2 in terms of the ratios /o/(>S -f D), fi/(S + D), Zo/ (S + D) and the excitation parameter Vr/ (Niy where eVr is the relativistically cor- rected electron energy, and NI the effective ampere-turn excitation of the lens. By effec- tive is meant the ampere-turn excitation drop across the lens gap. The dotted line in the figure represents the thin lens approxima- tion given by / = 25 VriS + D)/{Niy (2) It is seen that the projector focal length passes through a minimum. This is of im- portance in that it determines the maximum magnification which can be obtained with a given stage length. Of paramount importance is the spherical aberration of the objective lens which sets the ultimate limit of resolving power. The spherical aberration is defined by the con- stant Cs where Cs a^ gives the radial error in position of a ray leaving the object at angle a with the axis, the distance being measured in object space. The value of Cs/f is given within an error of ± 10 % for values of the ratio S/D between 0.2 and 2 by the full curve of Figure 3. The 16 Cs/f 14 12 lO ■' " r T~ / \ / 1 y ^ tA / ^^^^0'^'^'^ ^ THIN L Cs/f = ENS APP 5f/Cs■^ ROX 0-2 0-4 0-6 08 lO 1-4 1-6 1-8 f/(Si-D) Fig. 3. Variation of the ratio CJj with// (5 -h Dy. 20 150 ELECTRON OPTICS 4000 (nO . - •■-^*~*^^- 1 .^^ 6000(ni) \^^-- r'"~""..---' — 8ooo(ni) 8000(ni) Fig. 4. The resolving power as a function of the pole piece spacing, magnetic field and excitation for 100 KeV electron energies. dotted curve shows an approximation which becomes accurate for weak lenses and is given by: CJf = 5[//(S + D)Y (S) The resolving power of the microscope is theoretically limited by spherical aberration (requiring a minimum aperture angle) and diffraction (requiring maximum aperture angle) to a minimum value (d) at an opti- mum aperture angle a, given by: d = 0.43 C'/^XS/" a = 1.4(X/C.)i/* (4) (5) Figure 4 shows the variation of resolving power under optimum aperture conditions as a function of pole piece spacing (S), mag- netic field strength (Hp) and excitation (A^7) for 100 Kev electron energy. It is seen that for a given field strength a flat optimum oc- curs around a particular value of spacing. A further parameter of importance is the chromatic constant (C<.) which gives the variation of focal length with small changes in the high voltage or lens current Sf = CciSV/V - 257/7) (6) 8f must clearly be kept within the depth 151 ELECTRON MICROSCOPY lo Cc/f 0-9 0-8 0-7 0 6 ..^l I i j ^ OOI 002 003 004 005 006 O07 008 O 09 OIO \2 Vr/(NI)^ Fig. 5. The variation of the ratio of the chromatic constant (Cc) to the focal length with the excita- tion parameter Vr/(Niy. of focus of the instrument. The value of Cc/f is given by the curve of Figure 5 as a function of the excitation parameter. To ensui'e no limitation in resolving power or contrast, the variation in focal length (5/) due to ripple on the electron accelerating voltage or lens current supplies must be kept below one quarter the depth of focus of the instrument: 5/ < 0.7 dVX This requires a voltage and current sta- bility meeting the requirement: dV/V - 25/// < 0.7 dyxCc For a lens near the minimum focal length (Cc '^ 0.4 cm) and for a resolving power near the optimum ('~2-4 A), the required stabilities are in the region of 2 or 3 parts in a million (Haine, 1960). The magnetic lens has the peculiar property of rotating the image with respect to the object. Astigmatism arises in objective lenses as a result of very small departures from axial symmetry of the pole piece bores or faces. Only the elliptical component of asymmetry is of significance. To an adequate approxima- tion the astigmatic distance between the tangential and sagittal foci (za) is given by the expression: Za = 1005(2 + 3 S/D)Vr/iNiy (7) where 5 is the departure from symmetry. For the astigmatism not to limit the re- solving power, Za should be small compared with the depth of focus. To achieve the nec- essary symmetry tolerance, which may be a few millionths of an inch for optimum re- solving power, represents an almost impossi- ble mechanical engineering task. Fortun- ately, it is possible to correct a small degree of residual astigmatism by the inclusion of a weak cylindrical lens of variable power and orientation. The progress of correction has been discussed in detail by Haine and IMul- vey (1954). Design Considerations The number of stages of magnification in- cluded in the microscope is dependent on the maximum magnification and the range of magnification required. A fairly definite op- timum of three stages can be deduced, giv- ing two stages following the objective. The various factors affecting the choice of stage length and the focal length of the various lenses include the maximum and range of magnification, the minimum being limited by image distortion. The practical design of a lens for given pole piece geometry must ensure an adequate magnetic circuit, ade- quate heat dissipation from the excitation coil and a design which can be manufactured within very close symmetry tolerances. The magnetic design is discussed by Mulvey 152 40 004 OOI ELECTRON OPTICS 0 0025 Vr /(Nl)^ Niy/vv Fig. 6. The image rotation angle in radius plotted as a function of NI/\/V\- and Vr/iNI)^ (on top scale). (1953). Permissible thermal loading of the coil varies from about 800 ampere-turns per square centimeter of cross-sectional area of coil for a lens without water coohng, to 1200 ampere-turns per sq cm for a carefully de- signed water-cooled coil with no interleaving paper. The values are fairly independent of the wire size used, which may be chosen to give a coil impedance most suitable for the current supplies. It was at one time thought that the great precision of symmetry required in the objec- tive lens required the pole pieces to be manu- factured separately from the main iron shroud and fitted precisely within its bore. That this is not so, and in fact leads to un- necessary complication, has been shown by Haine (1954). Although some manufacturers still utilize separate pole pieces, the tendency is toward the simpler lens made from two pieces of iron. Electrostatic Lenses The data for electrostatic lenses cannot be expressed by such simple means as for mag- netic lenses. The variation of focal length and spherical aberration with the dimensions for three aperture unipotential lenses of spac- ing (*S), central electrode aperture diameter D and thickness T are given in Figures 7 and 8. The objective cannot be immersed in the electrostatic field and it will be seen that the spherical aberration is about an order of ten greater than for the magnetic lens. 153 ELECTRON MICKOSCOPY 4-5 4-0 \ V , s ^ ^ — 1 — '/// \ \ y /// v/ 3b 3-0 2-5 2-0 \\ \\\ ri y/ // \ \ /i-o/ / /''V \\\ /// l'5 l-O ■s. 0 -5 O 1 T/D O 0'2 0-4 0-6 0-8 I-O 1-2 1-4 1-6 IB Fig. 7. The focal properties of an electrostatic lens as a function of its geometry. 90 80 70 Cs/S 60! SO 40- 30 20- lO V s o S / i D \\\ ' U- / T '/ ' // We 1 y-o \ I / \\\ ^!;s ^0\^ 1 — • 1 1 T/O 1-2 1-4 O 02 0-4 0-6 0-8 I-O Fig. 8. The spherical aberration of an electrostatic lens as a function of its geometry. 154 HISTORY OF ELECTRON OPTICS REFERENCES Archard, G. a., "Proc. Internat. Conference on Electron Microscopy," London (1954). Pub- lished by the Royal Microscopical Society (1954). Bloomer, R. N., Brit. J. Appl. Phys., 8, 83 (1957). Glaser, W., Z. Physik, 117, 285 (1941). Haine, M. E., "Proc. Internat. Conference on Electron Microscopy," London (1954). Pub- lished by the Royal Microscopical Society (1954). Haine, M. E., "The Electron Microscope" (E. and F. N. Spon Ltd., London). Haine, M. E. and Einstein, P. A., Brit. J. Appl. Phys., 3, 40 (1952). Haine, M. E., Einstein, P. A., and Borcherds, P. H., Brit. J. Appl. Phijs., 9, 482 (1958). Haine, M. E. and Mulvey, T., /. Sci. Instr., 31, 326 (1954). Langmuir, D. B., Proc. I.R.E., 25, 977 (1937). Levt, F., Z. Angew. Phys., 2, 448 (1950). Liebmann, G. and Grad, M. E., Proc. Phys. Soc (London), 64B, 956 (1951). MuLVET, T., Proc. Phys. Soc (London), 66B, 441 (1953). Ramberg, E. G., J. Appl. Phys., 13, 582 (1942). M. E. Haine FIBERS (TEXTILES). See GENERAL MICRO- SCOPY, p. 343. HISTORY OF ELECTRON OPTICS The history of electron optics starts more or less with the discovery of cathode rays. The earliest experiment of Pliicker in 1859 already showed a rectilinear propagation of these rays. Ten years later, Hittorff dis- covered that cathode rays can be deflected by a magnetic field, and that an axially sym- metrical magnetic field will concentrate such rays. Another ten years elapsed before Crookes demonstrated a better proof of rec- tilinear propagation. The first attempt to calculate the trajectory of charged particles in fields of forces originated with Riecke in 1881. These were rather incomplete, and the first quantitative theoretical and experi- mental studies on the deflection of an elec- tron beam had to wait until 1897 when J. J. Thomson determined the ratio of the charge to mass of the elementary particles carrying the elementary quantity of electricity and confirmed, in a manner of speech, the beauti- ful calculations performed earlier the same year by H. A. Lorentz. The first intentional use of a solenoid for concentration of cathode rays was done by Wiechert in 1899. He used a relatively long solenoid; a short coil for concentration was not used until 1903 by H. A. Ryan. In the same year, the first electrostatic concentration of an electron beam was performed by Weh- nelt. Very extensive calculations of trajec- tories of charged particles in field.s of forces were carried out in 1907 by C. Stoermer. The basic equations of electron optics are im- plicitly contained in the work of Stoermer, although this had not been recognized for many years. The haphazard use of electron optical ele- ments and of calculations was followed by a more deliberate one in the middle twenties of this century. De Broglie's thesis in 1924 on the wave nature of the electron became the foundation of physical electron optics. The formal foundation of geometrical electron optics was set down in 1926 by Busch. Busch actually considered a magnetic coil as an optical lens, and derived the lens equation for it, although he failed to use the coil as an imaging element. The combination of these two discoveries stimulated thinking in two directions. His- torically, the first was the development of ap- plied physical electron optics in the form of electron diffraction. The momentous dis- covery of diffracted electron beams by Davis- son and Germer m 1926-27 was soon fol- lowed by the similar experiment of G. P. Thomson. On the other hand, geometrical electron optical methods were first applied around 1929, when cathode ray oscillographs were built on the basis of electron optical elements. These attempts were pursued simultaneously by Briiche iri Germany and by Zworykin in the United States. The next step in the development of geo- metrical electron optics was the study of ax- ially symmetrical fields as lens elements. 155 ELECTUON MICKOSCOPY These studies started probably around 1929. ber with two-way motion of the specimen At least, this is the date of a patent applica- and an air lock, as well as a photographic tion on electrostatic lenses by I^oll. The chamber for internal photography of the fii-st publications on the study of electro- electron micrographs. This photographic static and magnetic lens elements date from chamber was also provided with an air lock. 1931 when Davisson and Calbick published The first electron micrographs of biological a note on the focal length of an electrostatic objects in 1934 were made on material im- lens and Knoll and Ruska studied the be- pregnated with osmium salts because of the havior of magnetic lenses. Further studies of then prevalent notion that the electron electrostatic lenses were published by Briiche beam would destroy any biological material, and collaborators. Within a year, it was recognized that this Experimental electron microscopy had its impregnation was not absolutely necessary beginning in 1931. The formal beginnings are and that by reducing the total exposure time marked by the patent application of Riiden- and beam intensity one can achieve biologi- berg and the publication of a paper by Knoll cal pictures without destroying the material, and Ruska; however, the subject goes back This was possible because of the introduction at least three or four years earlier to discus- of the internal photography described above, sions m physics colloquia in the Berlin area During 1934, Ruska had also demon- as seen from a publication by Gabor. Gabor's strated that the electron microscope had a name should be mentioned here in another resolution somewhat greater than that of the respect too. The studies of Knoll and Ruska light microscope. These observations, com- on magnetic electron lenses were greatly bined with the theoretical prediction that the facilitated by the fact that an ironclad mag- electron microscope is able to reach a re- netic lens had been developed in 1926-27 by solving power exceeding that of the light Gabor. During 1931-32, the first instru- microscope by a factor of several hundred, ments which merit the name of electron led to increasing efforts toward an improve- microscopes had been built. The fii'st of these ment of the resolving powder of the electron probably is one built by Knoll and Ruska for microscope. the observation of emitting objects, which From the beginning, it was understood utilized magnetic lenses. A second was the that a good part of the mechanism of the instrument built by Briiche and Johannsen, image formation is due to scattering of the with electrostatic lenses, to study emission electrons within the transmission-type ob- phenomena. Before the end of 1932, Mar- jects. Both Ruska and Marton referred to a ton had a simple instrument using a magnetic scattering mechanism in their 1934 papers, lens to investigate emission phenomena, and In 1936, Marton made the first quantitative later transmission phenomena. During 1933, attempt to explain the image formation on two instruments were built. One was Knoll the assumption of multiple scattering by the and Ruska's revised and improved transmis- object. The same approach was used a couple sion instrument, using two magnetic stages of years later by von Ardenne. Very soon, and having a limiting magnification of however, it became obvious that the average 12,000. IVIarton's second instrument also specimen of electron microscopy is much too used magnetic lenses for transmission ob- thin for multiple scattering. Consequently, servation, but it was simpler and its limiting ]\Iarton and Schiff in 1941 studied image for- magnification was of the order of 2,000. The mation assuming single scattering. In the first biological observations were carried out postwar years, single scattering calculations with Marton's two-stage instrument in 1934. were further improved by von Borries, Lenz Marton built in 1934 an improved two-stage and others, instrument, incorporating a specimen cham- In the meantime, efforts were under way 1.56 HISTORY OF ELECTRON OPTICS to improve the performance of the electron two-stage magnification. Around 1941, Hil- microscopes. These were twofold — ^one was lier attempted to use a compound projection building new microscopes of which a few lens. Three-stage magnification was intro- merit brief mention: The- instrument by duced in 1942-43 by Marton. This idea was Scott and iVIcMillen (1936) for the emission- taken up, probably independently, late in type observation of biological subjects. This 1944 by Ruska and von Borries in a patent was the first compound emission microscope application, and in 1944-46 by Le Poole. Le in America. Second, the instrument of Mar- Poole also introduced selected area diffrac- tin, Whelpton, and Barnum in England tion as an adjunct to electron microscopy, (1937) which sparked the British work in this using a three-stage arrangement, field; and the third, the new instrument of The postwar years have seen the rapid Ruska and von Borries (1938). This last in- development of commercially available strument paved the way to the fii'st commer- models in different countries. An important cial model, completed in 1939. There were contribution to the present-day development also advances in applications. In 1937, Mar- was recognition of the role of lens astigma- ton had taken the first bacteriological pic- tism by Hillier and Ramberg in 1946. Recog- tures. These were followed in 1938-39 by nition of this defect led, in the ensuing years, improved bacteriological pictures by Ruska to the development of different types of and collaborators. The first virus pictures electrostatic or magnetic stigmators and were obtained by Helmut Ruska, brother of produced the present high resolving power Ernst Ruska, in 1940. of modern instruments. The existence of the In 1938-39, some papers also appeared high resolution transmission type micro- by von Ardenne analyzing some of the de- scope stimulated the development of several fects of the microscope. The most important related instruments. First of these was the so- single item probably was an analysis of the called scanning microscope proposed by von alignment defects on the basis of which von Ardenne in 1938, followed by further im- Ardenne built a very complicated and highly provements by Zworykin, Hillier, and Sny- successful instrument establishing a tem- der in 1942. In 1939, the electron shadow porary world record of 30 A for resolving microscope was proposed by Boersch. In the power. Coincident with that development same year, 1939, von Ardenne proposed the was development also of the first electro- X-ray shadow microscope. The practical de- static microscope by Mahl and Boersch, as velopment of this instrument had to wait well as the development of the fu'st Canadian until the 1950's when Cosslett and Nixon microscope by Burton, Prebus and Hillier. built the first working instrument. Another This was followed by the development of the related instrument is the microprobe ana- first American commercial model — -the RCA lyzer of Castaing first described in 1951. Ion Type A was completed by Marton in 1940, microscopes are another group of derived followed by RCA Type B in 1941 by Zwory- instruments. The first attempt was made kin, Hillier, and Vance. These two last- with lithium ions in 1947 by Boersch, In named instruments were the first ones to use 1949, Magnan and Chanson published the highly regulated electronic power supplies, first description of a proton microscope. The German work relied on the stability of The 1932 attempts of Knoll and Ruska, as storage batteries to achieve the required well as Briiche and Johannsen, in emission constancy of the magnetic lenses. microscopy has been mentioned earlier. The Japanese work on electron microscopy resolving power of these emission micro- started in 1939 partly by Higashi and Tani, scopes was quite poor, and no progress was also by Tadano. made until Recknagel in 1941 developed a All the above-described instruments used theory of immersion objectives. In 1942, 157 ELECTKO.N MlCKOSCOl'Y Mecklenburg and Malil produced much bet- ter emission micrographs with thermionic electrons. Secondary electrons for producing emis- sion microscopy had been hrst used in the middle thirties. A crude image had been shown in 1933 by Zworykin, and a year or two later Knoll had shown better results. The best results to date are those of Mollcn- stedtin 1953, who replaced the primary elec- trons with ions for the excitation of second- ary electrons. Photo emission as a source of emission microscopy had been first used by Briiche and by Pohl in 1933 and 1934. The method has been further developed in the hands of Grivet and Septier in 1956. The most spectacular emission microscope is not a true microscope in the optical sense. Field emission microscopy started with a two-dimensional model of Johnson and Shockley in 1930. Very soon thereafter, in 1935-37, Mueller developed the point emis- sion microscope. This was followed in 1951 In' the invention of a field ion microscope with a best achieved resolving power of about 2.7 A. A method derived from dark field micros- copy is the so-called schlieren observation of electrostatic and magnetic fields and other perturbations of the optical medium. One of the first observations of this kind was pub- lished by Boersch m 1937, when he demon- strated the dark field image of a vapor stream. In 1943-44, von Ardenne deliber- atel\^ produced schlieren conditions. Due to the war, this paper was not known in the United States, and it was independently dis- covered by Marton in 1946-47 who with Simpson and Lachenbruch applied this method for quantitative evaluation of mag- netic fields. As this review is hmited to the electron microscopical aspects of electron optics, a very brief enumeration of other de^•elop- ments of electron optics may be sufficient. The oscilloscope tube development was started in 1894 by A. Hess and in 1897 by F. Braun. These early tubes used gaseous discharge with cold cathodes. The hot cathode in the cathode ray tube was introduced by Wehnelt in 1903. The first proposals for the use of cathode ray tubes as image transmitting ele- ments were made in 1900-07 by Dieckmann and Glage, as well as by Rosing. Storage of the image now used in television pickup tubes was first announced in 1908 by Camp- bell-Swinton. Modern development is linked to the names of Zworykin between 1925-33, Round in 1926, Farnsworth in 1927, and Henroteau in 1929. Mass spectrographs are another interest- ing chapter of electron optics. They were first conceived by J. J. Thomson in 1897; 180° magnetic deflection was fu'st used by Classen in 1908. Ten years later, Dempster used magnetic focusing; and in 1919, Aston invented ^-elocity focusing. Modern mass spectroscopy (developed in 1932-34) is linked to the names of Herzog and Mattauch w^ho developed the electron optics of mass spec- troscopy devices, as well as Bainbridge, Bar- ber, and Stephens. The development of electron optical theory is linked to three names essentially: Stoermer we mentioned earlier, but the for- mal theory of axially symmetrical devices was not developed until the years 1933-36. Foremost are the names of Glaser and Scher- zer: the first used an essentially Fermat- Hamiltonian approach, whereas the latter preferred the trajectory method. The first linking of geometrical electron optical theory to wave mechanics was done by Glaser in 1943. We now present a short description of the development of physical electron optics. Electron difTraction instrumentation for the observation of crystallographic structures has been greatly improved by the addition of a magnetic lens to the diffraction instru- mentation by Lebedeff, in 1931. Modern diffractographs took advantage of the high resolving power of combined electron optical 158 IIVIAGE FORMATION MECHANISM and diffraction elements as manifested by the instrument developed by Cowley and Rees in 1952. Boersch has contributed greatly to the understanding of the physical optics of elec- tron microscopical phenomena. First he had shown, in 1936, the correlation between dif- fraction diagram and the electron optical image. Recognition of this fact was essen- tially responsible for the development of all modern combined electron microscopical and diffraction instrumentation. In 1941, he showed the correlation between image prop- erties of crystalline objects and Bragg reflec- tion. Then, in 1943, he discovered the exist- ence of Fresnel diffraction in electron microscopical images. This discovery led in 1946 to the observation of phase contrasts by Hillier and Ramberg and to the method called "microscopy by reconstructed wave- fronts" invented by Gabor in 1948. Electron interferometry started with a proposal by Marton in 1952. Following this proposal, Marton, Simpson, and Suddeth built a three-crystal, wide beam interferome- ter in 1953. About the same time, accidental interferences had been observed in electron microscope images by Rang. The use of Fres- nel biprisms for electron interferometry was introduced in 1954 by Mollenstedt and Diiker. L. Maeton IMAGE FORMATION MECHANISM Two criteria are commonl}- used for judg- ing an image formed by an optical system: one is resolution, and the second is contrast. Some of the factors governing both are com- mon, and for this reason the ensuing discus- sion will consider both. The essential mecha- nism, responsible for both qualities, is scattering of electrons in the broadest phys- ical sense. It is customary, however, to desig- nate the coherent effects of scattering from an aggregate of atoms by the term "diffraction," reserving thus the expression "scattering" to the manifestations of localized electron- specimen interaction at the atomic or quasi- atomic level. These two effects constitute the major contributions to resolution and con- trast in the image. Two minor contributions are called "refraction" and "absorption." Refraction is that process in which the major change brought about by the scattering is limited to a change in the phase of the wave function, while in absorption there is a large change in the amplitude of the wave func- tion due to the scattering. In this discussion of image formation, we will consider only those effects which arise from the interaction of the electrons with the object and those modifications of these effects which are produced by the proper- ties and defects of the optical system. How- ever, the general effect of the optical proper- ties of the system on image formation will not be treated. Diffraction is due to the wave nature of the electron which produces deviations from the simple geometrical model of energy propagation. These deviations are manifested by the appearance of dark and bright bands, the Fresnel diffraction fringes. The least re- solved distance of an optical instrument (most commonly called resolving power) can be taken to be the distance at which the first diffraction maximum from a point-like object coincides with the zeroth diffraction maximum of the next object. Using this so- called Ra3^1eigh criterion in the Abbe rela- tion, we obtain 5 = 0.61X n sin a (1) where 8 is the least resolved distance, X is the wave length of the electron (X = h/p, where h is the Planck constant and p is the momentum of the electron), and a is the aperture angle of the optical system, n is the refractive index of the electron optical me- dium wliich for many applications can be as- sumed to be unity. The numerical constant 0.61 changes if another criterion is substi- 159 ELECTRON MICROSCOPY tutod for the Rayloish criterion in the defini- makes an angle 0 (satisfied by equation (S)) tion of the least resolved distance. Equation with a lattice plane of the crystal, essentially (1) can be derived either from the theory of specular reflection may occur from this lat- Fraunhofer diffraction or from the quantum tice plane. If the optical system of the elec- mechanical uncertainty relation. Both rela- tron microscope has a wide enough aperture tions give qualitatively the same results to collect both the primary beam and the within a small numerical factor. reflected (diffracted) beam, and the objective Contrast is also governed to some extent is properly focused so that the two beams by the diffraction phenomena at the edge of are brought to the same focus, the intensity an object. Using the theory of diffraction, distribution in the image of the crystal will the fringe distribution at the edges of an appear to be w^hat may be called "normal." opaque or semi-opaque object can be calcu- If the objective aperture, however, is small lated ; and such calculations have been found with respect to the Bragg angle, such that to be in reasonably close agreement with the reflected beam will be intercepted (most experiment, provided that in the case of of the intensity being in the reflected beam), semi-opaque objects a refractive index of the the directly transmitted part of the image material, corresponding to an internal po- will appear unusually dark. A further mani- tential of 10-15 ev, is assumed. The experi- festation for this phenomenon is when a wide ments show that fringes appear in out-of- aperture system is slightly defocused and focus images and that the fringe spacing both the dark and bright images appear side depends upon the degree of defocusing of the by side. This variation in intensity, due to image. The apparent contrast at the edge of Bragg reflection, contributes also to appear- an image is optimum in the slightly defocused ances of the so-called extinction contours, image. For perfect focusing, the contrast is These extinction contours appear in slightly found to be lower than for the out-of-focus bent crystals where, accidentally, part of the condition. crystal happens to be at the proper angle to The intensity distribution in Fresnel dif- show Bragg reflection. Under action of the fraction fringes can be interpreted as inter- electron beam, thin crystals may warp, show- ference phenomena. Accidentally occurring ing a displacement of the extinction contours interference fringes can be observed also in while under observation, selected specimen areas illustrating again the Related also to Bragg reflection is the ob- role played by interference phenomena in the servation of crystal defects, such as disloca- contrast in the image plane. tions, stacking faults, etc. Due to the exist- In the case of crystalline specimens, dif- ence of these crystal defects, certain lattice fraction from the lattice planes can produce planes show different orientations from the an important modification of the intensity surrounding lattice planes and produce in distribution in the image plane. Electrons this manner a marked contrast in the final which are scattered from a crystal lattice image. will show diffraction maxima according to While the above considerations are re- Bragg's equation stricted to Fresnel diffraction and to crystal- „ , . line diffraction, Gabor has demonstrated that dmraction phenomena may be impor- where d is the distance between the lattice tant in noncrystalline specimens too. He pro- planes, 0 is the angle between the direction posed the use of a coherent electron beam for of the incident beam and the diffracting forming a diffraction image, called hologram, planes, and n is an integer. of the specimen. In principle, such a diffrac- If an electron beam is directed so that it tion image should contain all information 160 IMAGE FORMATION MECHANISM about the specimen except the phase. Through the use of proper optical equipment, the image can be reconstructed from such a hologram in a process called "microscopy by- reconstructed wave fronts." Although prac- tical difficulties until now made it impossible to reach high resolutions by this method, its existence amply shows the importance of dif- fraction in the image forming process. At extremely high resolution, crystal lat- tices and other periodic structures can be ob- served provided that the aperture of the sys- tem allows the zero order spectrum to pass together with at least one of the first-order diffraction spectra. This is in agreement with the Abbe theory of image formation in light optics. Moire patterns are due to the com- bination of a doubly diffracted beam, orig- inating on overlapping crystals, with the in- cident beam. The intensity at any point of the image is governed by the number of electrons which have been scattered within the solid angle formed by the aperture of the objective lens. This number can be WTitten as for the elastically scattered electrons is given by where '^'" - ^ (z - fy- (5) '^j t^^^c ^'l^f *^ >'»^^ L^--" ,d^ ^mI *-T-^iii«* Fig. 3. Cross section of a mouse kidney proximal convolution (top) and a distal convolution (bot- tom) with surrounding capillaries (C). In both tubules the nuclei (N) and mitochondria (M) are evi- dent. The lumen is usually closed in the proximal convolution but open (Lu) in the distal. Numerous brush border extensions (B) line the cells in the proximal convolution but only few microvilli (Vi) in the distal convolution. Vacuoles (V) containing autofluorescent material are prominent. Occasionally, a cilium (Ci) is encountered in the distal convolution. Magnification 2,900X. cupying the space that normally would be mal convolution and the abundance of brush called the lumen of the tubule. Evidence is at border extensions seems to facilitate this ac- hand showing that about 85 per cent of the tivity of the cells by largely increasing their urine reabsorption takes place in the proxi- absorptive surfaces. Therefore, the w'ay the 168 KIDNEY ULTRASTRUCTURE urine seems to be passed along the proximal convolution is probably more like the slow filtration through a finely porous sieve than like the rapid drainage of a sink with a wide open plumbing system. The "pores" of this particular sieve are then represented by the narrow slits between the millions of brush border extensions. It can be i^ecorded in high resolution micrographs that the plasma membrane which covers the extensions has a thickness of about 50A. However closely packed, the extensions do not get in closer contact than lOOA apart. The intervening space is occupied by a substance with less density than the plasma membrane. It has been suggested that this substance represents an additional layer of possibly lipid mole- cules with a depth of 50 A, identical with the intervening layers between adjacent cell bor- ders. When the urine passes along the proxi- mal convolution, it expands slightly the two lipid layers of adjacent brush border exten- sions, thus creating the narrow slits men- tioned above. By this mechanism is estab- lished a much closer contact between the urine and the surface membrane of the brush border extensions than would be the case with a wide open lumen. Some substances, as for instance proteins, which cannot be absorbed through the sur- face membrane, are taken up by the tubular invaginations. These structm'es are located between the bases of the brush border ex- tensions and represent minute tubules which are in open connection with the tubular lumen (Fig. 8). The tubular invaginations are ex- panded to vacuolar profiles when fluid and substances are taken in. A certain condensa- tion of the vacuolar content is noted con- comitantly with a breaking-off of the con- nection between the expanded tubular invagination and the lumen of the nephron. The condensed vacuole can then be found anywhere in the cell and as the condensation of its contents proceeds, the vacuole is trans- formed into what has been called a large granule in electron microscopy. This whole mechanism of taking in fluid-dissolved sub- stances has been called micropinocytosis or membrane flow. The mechanism by which substances other than protein are taken up by the cells of the proximal convolution is structurally not clear. It involves substances and ions such as glucose, sodium, potassium, phosphate, sulfate, amino acids, urea, and creatinine to mention some of them. The reabsorption of water is a passive process, but most of the substances listed involve an active process which requires energy. The enzymes recjuired for these active processes are supplied by the multitude of mitochondria present in the cells of the proximal convolution (Fig. 4). The ultrastructure of the mitochondria has been thoroughly investigated (cross reference: cell ultrastructure) and it is believed that the efficiency of the mitochondrial work is in- creased by the number of mitochondria and the presence of structurally intact internal membranes. Organelles have been recorded in these cells which presumably constitute mitochondrial precursors. They have been called microbodies, and represent small spher- ical bodies without internal membranes and with a single membranous capsule as com- pared with the double-contoured mitochon- drial outer membrane. Structures have been found in the basal portion of the cells of the proximal convolu- tion which probably facilitate the flow of re- absorbed fluid and substances through the cell body. They represent infolding s of the basal plasma membrane which extend to a varying degree into the cell (Fig. 4). The mitochondria have a close relationship with these infoldings and it is tempting to as- sume that this facilitates a certain interac- tion between the enzymes carried by the mitochondria and the infolded plasma mem- brane (Fig. 8). Once the reabsorbed fluid and substances have obtained contact with the basal plasma membrane, they may penetrate it by means of an enzymatic activation (Fig. 9). And when the extracelhilar space is 169 Fig. 4. Basal part of a proximal convoluted tubule cell of the mouse kidnej'. The plasma membrane which faces the basement membrane (BM) is highlj^ infolded, but at the point where it turns (arrow) another portion of the cytoplasm is always interposed. The basal cell cytoplasm is thus divided into coarse lamellae which contain mitochondria (M) and abundant RXA particles (R). Magnification 32,000X. Fig. 5. Basal part of a collecting tubule cell of the mouse kidney. As in the proximal convolution, the basal plasma membrane is infolded, but as a rule, the turning point (arrow) can be seen without anj' interposed cytoplasmic lamella. The cytoplasmic lamellae are here thinner than in the proximal con- volution and not broad enough to house mitochondria (M). Some RNA-granules (R) are located in the lamellae. The basement membrane (BM) is quite thin. Magnification X46,000. 170 BM Fig. 6. Two cells of the thick, a.- ivstiiig on the basement membrane (BM) and joined by a terminal bar (TB). The cell boundary is indicated by the dotted line. The nucleus (N) is located in the apical part of the cell with the Golgi zone (Go) close to it. Extending into the tubular lumen (Lu) are numerous microvilli (Vi). The cytoplasm is finely granulated because of the presence of abundant RNA particles. The autofluorescent vacuoles (V) are less numerous than in the proximal convolution. The mitochondria (Ml) are long and densely packed. Sometimes, their shape is more like an ice cream bar than that of a sausage, which can be seen when they are sectioned at an angle (M2). The basal cell membrane is deeply infolded and the turning points clearly seen at the arrows show that cytoplasmic lamellae, also containing mitochondria, are always interposed in an interdigitat- ing fashion. Magnification 11,000X. 171 Fig. 7. Survey of the papilla of the rat kidney showing mostly cross sectioned collecting tubules (D), thin segments of Henle's loop (T) and capillaries (C). The cells of the collecting tubule are cuboidal with few mitochondria, easilj^ recognized cell boundaries, and tinj- microvilli on the surface. The cells of the thin segment are fairly squamous in sections showing scalloped appearance due to the frequently inter- digitating, narrow cell extensions. The endothelium of the capillaries is extremely thin. In the electron micrograph, there seems to be no chance of mistaking a thin segment for a capillary because of the dif- ference in thickness of the lining cells, but also because of the obvious staining of the blood plasma in the capillaries. The interstitial cell (X) is unidentified. Magnification 2,700X. 172 KIDNEY ULTRASTRUCTURE reached, the access to the surrouncUng capil- laries is easily achieved. The Golgi zone of these cells is quite re- stricted and it is believed that it is mainly- involved in secretory processes, particularly in the synthesis of protein used either by the cell itself or for extracellular purposes. This hypothesis is supported by the fact that rough-surfaced endoplasmic reticulum is ab- sent in these cells, but the cytoplasm is filled by numerous RNA-particles which presuma- bly are the structural evidence of cytoplasmic proteins. The Loop of Henle. When the nephron leaves its convoluted portion in the cortex of the kidney, it assumes a straight course down into the medulla and papilla. The first part of this portion is called the straight de- scending loop of Henle. Once down in the papilla, it bends back again and the turning part is called the thin segment of HenWs loop. PROXIMAL CONVOLUTED TUBULE Fig. 8. Schematic representation of the basic structures of the proximal tubule cells of the mouse kidney: nucleus (N), mitochondria (M), microbodies (m), Golgi apparatus (Go), auto- fluorescent vacuoles (V), large dense granules (D), ribonucleoprotein particles (R), brush border extensions (B), tubular invaginations (Ti), plasma membrane (PM) with infoldings, terminal bars (TB), and basement membrane (BM). (After Rhodin, 1954). cell borders bcsame^f rnembrane secfioma " cytoplasmic laimllat cytoplasmic lamtliaz frcm eel Is pulled auaq terminal bars rnifochondria nucleus cell borderi ■ THIN SEGMENT OF HENLE'S LOOP Fig. 9. Three-dimensional reconstruction of cells of the proximal convolution and thin seg- ment of the mouse kidney. The reconstruction is not based on serial sections. (After Rhodin, 1958). From there on, it approaches again the cortex and the neighborhood of the glomerulus. This portion is called the ascending (thick) limh of Henle's loop. It should be stressed that in most nephrons, the thin segment comprises a considerable part of the descending loop, the hairpin turn itself, and for some distance, the ascending loop. Straight Descending Loop. The cells of the straight descending loop of Henle have a light cytoplasm due to a certain scarcity in mitochondria and RNA-particles. The sur- face of the cells shows scattered brush border extensions, but these are shorter and coarser than those found in the proximal convoluted portion of the nephron (Fig. 10). The lumen of the tubule is frequently open, presumably because of the relatively few brush border extensions. Tubular invaginations of the surface plasma membrane are few and the basal infoldings of the plasma membrane, so numerous in the proximal convoluted part, are shallow and few, often completely absent. 173 ELECTRON :\lICKOSCOPY ^^^M^^ 6 Fig. 10. Diagram showing the essential fea- tures of the colls lining different parts of a tj'^pical cortical nephron in mammalian kidney as seen with the electron microscope. 1) Collecting tubule (arched or proximal part) : dark intercalated cell; 2) Collecting tubule (arched or proximal part): light cell; 3) Proximal convo- luted tubule: proximal part; 4) Distal convoluted tubule: intercalated part (Schaltstiiek); 5) Proxi- mal convoluted tubule: terminal portion; 6) Distal convoluted tubule: thick (ascending) limb; 7) Thin segment of Henle's loop. (After Rhodin, 1958) These phenomena seem to indicate that less resorptive activity is performed by the straight descending loop as compared with the convoluted proximal part of the nephron when judging from a structural point of view. Thin Segment. The epithelium of the thin segment of Henle's loop is of a squamous type with extremely flattened cytoplasm (Fig. 7). The cell surface shows only scattered short microvilli and tubular invaginations are not recorded. The mitochondria are scarce and exceedingly small. The cytoplasm is light due to a small number of RNA-par- ticles. Basal infoldings are present only be- neath the nucleus where the cytoplasm is of greater thickness than elsewhere. The at- tenuated part of the cell shows some quite characteristic features of this portion of the nephron. It displays a large number of cyto- plasmic extensions which rest on the base- ment membrane. These extensions resume the shape of the arms of a starfish and they interdigitate frequently with similar exten- sions of neighboring cells (Fig. 9). This pat- tern is reminiscent of the way the epithelial cells of the glomerular capillaries are ar- ranged. However, in the case of the thin seg- ment, the interdigitated cell processes al- ways show a terminal bar close to the surface which presumably secures the firm attach- ment of individual cell processes to one an- other. It can, therefore, be assumed that there do not exist any "slit-pores" between the cells of the thin segment as was indi- cated to be the circumstances regarding the glomerular capillary epithelial cells. In the papilla and deeper portions of the medulla, the descending and ascending parts of the thin segment are closely opposed to each other as well as to the capillaries (vasa recta) and the collecting ducts. It has been suggested that the close juxtaposition would facilitate the mutual exchange of fluid, sub- stances and energy because of the close re- semblance of this system to the basic princi- ple of a counter current system with streams moving in opposite directions. The ultra- structure of the thin segment seems to sup- port this hypothesis. The thin walls of its loop are evidently ideal to serve this ex- change of fluid and substances. Ascending (Thick) Limh. The ascending limb of Henle's loop is slightly wider than the other portions of the nephron and has, therefore, been called the thick limb. The cells have a cuboidal shape and a free open lumen is always to be found. The extreme multitude of mitochondria makes this sec- tion of the nephron stand out more clearly than the others, both in light and electron microscopy. Not only are the mitochondria numerous, but they also have a coarse and elongated shape as compared to the spherical form of the other parts of the nephron (Fig. 6). The mitochondria are densely packed and arranged with their long axes perpendicular 174 KIDNEY ULTR VSTRUCTURE to the basement membrane of the tubule, convokition is less coiled than the proximal They extend from the basement membrane one. The cells become small and a wide lumen to a level of about one micron from the cell opens up (Fig. 3). The surface of the cells surface. The mitochondrial fine structure is usually has longer microvilli than is the case of the same appearance as elsewhere in the in the thick limb, but the length varies from nephron with the exception that the inner cell to cell. The mitochondria decrease no- membranes (or plates) are more frequently ticeably in number and size, as does the seen arranged parallel with the long axis of number of RXA-particles, features which all the mitochondrion. contribute to a light appearance of the cell The free surface of the cells displays a cytoplasm. A few shallow and narrow in- number of small and scattered microvilli. The foldings of the basal plasma membrane can luminal portion of the cells, more or less de- be recorded, but they are too narrow to ad- void of mitochondria, is pervaded by abun- mit any mitochondria to be enclosed in the dant microvesicles bounded by a smooth small cytoplasmic strands they give rise to single membrane. Some of the vesicles are (Fig. 5). A fair number of small vesicles is connected with the surface plasma mem- still to be found in the luminal part of most brane. The Golgi apparatus of the cells oc- of the cells, but cells can be recorded which cupies a restricted area around the upper part are completely devoid of these structures, of the nucleus. Its fine structure is identical When the distal convolution approaches with the one recorded in the proximal con- the neighborhood of the vascular pole of the volution. The basal plasma membrane is glomerulus, it becomes attached to the angle deeply infolded between the elongated mito- between the afferent and efferent arterioles, chondria leaving but a narrow strip of cyto- This part of the distal convolution has been plasm in between. The RNA-particles are called the macula densa because it can be numerous and scattered among the small seen that the cells of the distal convolution vesicles in the apical cytoplasm. here assume a more columnar shape than It is obvious that the work performed by elsewhere, thus causing the nuclei to stand the thick limb requires a big supply of en- out very clearly in close juxtaposition (Fig. zymes judging from the great number of 1). Apart from their extreme columnar shape, mitochondria present in the cells. Sodium is the cells of the macula densa do not show any taken up by these cells by an active process peculiarities as far as their fine structure is which also leads to a simultaneous retention concerned which would discriminate them of water. Furthermore, formation of ammonia from other cells of the distal convolution, and the acidification of the urine seem to take Cortical Collecting Duct. This part of place in this portion of the nephron. It may the nephron immediately follows the distal be reasonable to assume that the numerous convolution and serves only as a short con- microvesicles are structural evidence for one nection between this part and the collecting or both of these processes. Similar micro- tubule. It is characterized by two cell types, vesicles are a prominent feature of the pari- the first' identical with the one occurring in etal cells of the gastric mucosa. These cells the distal convolution, the second, with the supposedly are responsible for the production appearance of the cells found in the collect- of the hydrochloric acid of the stomach. ing tubule. As the cortical collecting duct is Distal Convolution. The thick ascend- gradually transformed into the collecting ing limb of Henle's loop is gradually trans- tubule, the cell type reminiscent of that of formed into a convoluted position, the dis- the distal tubule disappears. Because these tal convolution, when the nephron again cells seem to be interposed between the main reaches the cortex of the kidney. The distal cell type, they have been called intercalated 175 ELECTRON ^IKKOSCOPY cells. Tlie intor('alat(Hl coll has a large luunbcr of sphorical initDchoiulria wliich arc clustered mostly above the nucleus, thus giving the cell a dark appearance which has lead some investigators to call it a dark cell (Fig. 10). This seems justifiable when considering that the main cell type has few mitochondria and a light cyt(»i)lasni. This cell is therefore called a light cell. There are other structurally important differences between the two cell types. The luminal surface of the dark cell usuall}^ has quite abundant microvilli, whereas the light cell has few or none. The vesicles of the luminal part of the dark cell cytoplasm and the RNA-granules are numer- ous as against a scarcity of these structures in the light cell. The similarity between the dark intercalated cells of the cortical collect- ing duct and the cells of the distal convolu- tion strongly suggests that the former ac- tually represents a certain variety of cells of the distal convolution which are distributed along the cortical collecting duct. Collecting Tubule. The collecting tubule begins in the outer cortex and runs in a straight course toward the medulla, each nephron being connected to a collecting tu- bule by the cortical collecting duct. The con- vergence of successive orders of collecting tubules in progressively deeper layers gives rise to vessels of mcreasing caliber until fi- nally the papillary ducts are reached. These ducts discharge their contents into the renal pelvis. It is believed that very few processes re- lated to the composition of the urine occur in the collecting tubule. The final concentra- tion of the urine seems, however, to be estab- lished here, possibly mediated by a certain reabsorption of sodium chloride. It has also been suggested that the permeability of the cells of the collecting tubules may be in- fluenced by the action of the antidiuretic hormone (ADH) of the pituitary. Structurally, the cells of the collecting tubule are rather poor (Fig. 7). The cells are of a cuboidal type with a very light cyto- plasm containing a small amount of IINA- particles and a few small and widely scat- tered spherical mitochondria. Large granules of the lipid type occur frequently, together with a small Golgi complex. The luminal sur- face displays a varying number of very short microvilli and the basal plasma membrane shows remnants of extremely shallow infold- ings. The cell borders are straight and the cells are held together by distinct terminal bars close to the surface. The nuclei are fairly large, occupying a good portion of the cell body. REFERENCES Sjostrand, F. S. and Rhodin, J., "The ultra- structure of the proximal convoluted tubules of the mouse kidney as revealed by high reso- lution electron microscopy," Exper. Cell Re- search, 4, 426 (1953). Hall, B. V., "Studies of the normal glomerular structure by electron microscopy," Proc. A7inual Con}. Nephrotic Syndrome, 5th Conf., p. 1, 1953. Rhodin, J., "Correlation of ultrastructural or- ganization and function in normal and experi- mentally changed proximal convoluted tubule cells of the mouse kidney," Thesis, Stockholm 1954, Karolinska Institutet. Pease, D. C, "Fine structures of the kidney seen by electron microscopy," /. Histochem. Cyto- chem., 3, 295 (1955). HalL; B. v., "Further studies of the normal struc- ture of the renal glomerulus," Proc. Annual Conf. Nephrotic Syndrome, 6th Conf., 1955. Rhodin, J., "Electron microscopy of the glomeru- lar capillary wall," Exper. Cell Research, 8, 572 (1955). Pease ; D. C, "Electron microscopy of the vascu- lar bed of the kidney cortex," Anat. Rec, 121, 701 (1955). Pease, D. C, "Electron microscopy of the tubular cells of the kidney cortex," Anat. Rec, 121, 723 (1955). RusKA, H., Moore, D. H., and Weinstock, J., "The base of the proximal convoluted tubule cells of rat kidney," /. Biophys. Biochem. Cytol., 3, 249 (1957). Rhodin, J., "Anatomy of kidney tubules," Int. Rev. Cytol., 7, i85 (1958). Rhodin, J., "Ergebnisse der elektronenmikros- kopische Erforschung von Struktur und Funktion der Zelle," Verhandl. deutsch., Gesellsch. Path. 41st. Tagung, 18, 274 (1958). 176 LEAF SURFACES Rhodix, J., "Electron microscop}' of the kidney," Am. J. Med., 24, 661 (1958). Johannes A. G. Rhodin LEAF SURFACES The electron microscope and its associated techniques have now reached a stage in de- velopment where they can be used by biolo- gists as tools of research, not simply as instruments for making interesting new mor- phological discoveries or for confirming in- formation gained from other sources. Our investigations, using the new techniques, have demonstrated the existence of a fine structure on the surfaces of many plants. Al- though some such structure beyond the reso- lution of the light microscope had been in- ferred, because plant surfaces vary widely in their ability to repel water droplets, its morphology, diversity, and the problems of its development had never been suspected. Prior to this work several attempts had been made to examine the surfaces of animal and plant cuticle with the electron micro- scope. Holdgate, Menter and Seal (1954) used reflection electron microscopy to study insect cuticle. This technique, although suc- cessful in demonstrating changes in the insect cuticle, suffered from the disadvantages as- sociated with reflection electron microscopy; the difficulty of interpretation is due to dis- tortion and a restricted magnification only a little above that of the light microscope. Most of the work on plant cuticle has used some form of replica technique. The pioneer work by Mueller, Carr and Loomis (1954) and Schieferstein and Loomis (1956) used liquid polyvinyl alcohol as the first stage of a two-stage replica method. However, it was necessary in their technique to wet the surface of the leaf with wetting agents to make the liquid plastic adhere. The liquids used to wet the surface ought to affect both the behavior and fine structure of those sur- faces, and their results appear to confirm this. A technique which does not involve wet- ting the leaf's surface has therefore been de- vised. It is basically the single-stage carbon method of Bradley (1954) in which the speci- men is coated with a film of carbon, and everything but the carbon is subsequently dissolved away. The technique is as follows with leaf specimens (Juniper and Bradley, 1958). The leaf to be examined is fixed to a glass slide with cellulose adhesive tape, and the slide is then placed in the evaporating chamber along with a porcelain/oil marker to gauge the thickness of carbon deposited. The chamber, which in the work described was an Edwards Coating Unit Model 12 EA, is then evacuated. While pumping is in prog- ress very little gas would appear to be given off by the leaf and a level of vacuum (10~' mm Hg) sufficient to allow the evaporation of carbon is easily obtained. Carbon is evapo- rated by passing a heavy alternating current through the points of two carbon rods lightly pressed together. The points of the rods are 15 cm above the specimen to be coated. A film of carbon 15-20 m/x thick is deposited on the leaf surface and the leaf is then removed from the vacuum. In spite of the level of \-acuum reached in the chamber, the leaf does not suffer any superficial distortion due to the escape of gas provided that the pumping time is kept short. The pumping time for most leaves is about 9 minutes, but succulent leaves may take up to 11 minutes. The subsequent stages of the technique are described with reference to Fig. 1. The carbon film is backed succes- sively with thin layers of "Formvar" and "Bedacryl" 122x, allowing each in turn to dry completely (Fig. lb). "Formvar" and "Bedacryl" 122x are quick-setting liquid plastics, flexible when dry and extremely soluble in certain solvents. Then the com- bined film of carbon and plastic (Fig. lb) is backed with cellulose adhesive tape and the leaf is peeled away from this composite film (Fig. Ic). The film is then immersed in ace- tone (Fig. Id). This removes the "Bedacryl" 177 KLK( ; i l« )N M l( :k( >SC<)PY from l.ctwvfii I he cellulose tape and the "Formvar," but does not affect the latter, wiiich is insoluble in acetone, and keeps the carbon fihn fiat and intact in the solvent. Specinu'ii ^lids are inserted into the space made by the removal of the "Bedacryl" (Fig. Id). Athene, New 200, 3.05 mm grids have been used in this work. The "Formvar" Carbon Rodi I Carbon Carbon _^,,„, Leaf Layer l-oy«r ^'^"l* (0) (b) («) Solvent for Bcdocryl I 1 i Solvent lor Formvar , Chromic acid 1 / / In ^J'^/^^J — ■ ■ •! • • >»v ■ 1 — fr -^\ fr^ "^\ — Support Grid Replica on Grid Replica on Grid (<") («) M Fig. 1. Stages in the carbon replica technique. A'Jf'*^ *.ei>l Fig. 2. Adaxial leaf surface of Chrj^santhemum segetum. and the carbon film is then lifted on the grid from the acetone bath and dried. The "Form- var" is finally washed away in a chlorofoi'm bath (Fig. le). A final .stage (Fig. If), which is rarely necessary, is to clean the film in a bath of chromic acid. This is only necessary where field material is examined and dirt picked up from the leaf might contaminate the microscope. Finally the replica is shad- dowed with gold/palladium through the grid bars at a fairly high angle, i.e., 1:1 or 2:1. Although originally intended as a high- resolution shadow-casting technique, the method of Bradley (1959) for the simultane- ous evaporation of carbon and platinum has been used to make self-shadowed carbon replicas. The carbon/platinum film is de- posited from 45° or 33° and the later stages of the technique are as described above. Very successful results with this adapted tech- nique have been achieved. The compound replicas are tougher, the shadows more clearly defined and the tedium of two stages of making a replica and shadowing are avoided. The fact that one can, apparently at the same time, shadow and form a com- plete replica of a specimen would appear to be a contradiction. The explanation is that evaporating carbon, but not platinum, does not travel in a straight line. The carbon replica is in a sense a negati^'e of the original leaf surface and the design of the electron microscope reverses original densi- ties in the image. The practice has therefore been adopted of using an extra negative stage in printing from the original electron micro- scope plates. This is contrary to certain conventions in electron microscopy, but it results in dark shadows and light protrusions from the plant surfaces — ^a result which is aesthetically more satisfying and easier to interpret. The contact angle of a drop of liquid on a surface is a measure of the wettability of that surface. The contact angle for water of a smooth surface of wax is about 100°. Many plant surfaces have contact angles for water 178 LEAF SURFACES above 140° and some micro-roughness was inferred to be responsible for these high con- tact angles. Two examples of the micro- roughness revealed by this technique are shown in Figs. 2 and 4. Projections above the level of the cuticle of many different forms have been discovered. They vary in height from 0.25 /x to 4 /x and are assumed to be mainly of wax. We believe that it is this fine structure which is responsible for the bloom in many species, for contact angles above 100°, and (most important to the agronomist) for the lack of adhesion of water droplets. All species which are found to have this fine Svax blanket' structure on their surfaces have initial contact angles for those surfaces above 140°. A number of problems connected with the retention of liquids on plant surfaces and thought to be associated with properties of the cuticle have been investigated with this technique. Fogg (1947) showed that different parts of the same plant differed in their abili- ties to shed water droplets. The electron microscope shows that the fine structm'e of the surface of a plant may vary with its position on a plant. Figures 3 and 4 show, respectively, the abaxial and adaxial surfaces of the same leaf of a pea plant. Differences were thought to occur in the fine structure on the surfaces of plants with variations in the light intensity (Dorschner and Buchholtz, 1956). Figures 4 and 5 show the adaxial surfaces of pea leaves grown at 5000 and 1500 ft-candles, respectively; as the light intensity falling on the leaf is pro- gressively reduced, the wax projections de- veloped on those surfaces are smaller, less dense and more angular in appearance. This agrees with the observation that the sur- faces of plants grown under reduced light conditions are more delicate and more sus- ceptible to mechanical damage. Differences in susceptibility to mechanical damage are apparent between different spe- cies of plants grown at the same light in- tensity. The leaf surface of Oxalis corniculata Fig. 3. Abaxial leaf surface of Pisum sativum var. 'Alaska'. Fig. 4. Adaxial leaf surface of Pisum sativum var. 'Alaska'. is practically impervious to mechanical dam- age and the leaves usually remain unwettable until senescence; Hyacinthus orientalis, on the other hand, is readily damaged bj'' ordi- nary aerial weathering, and the leaves are wettable soon after they have emerged. Very little is known about the comparative chem- istry of leaf waxes which must contribute to hardness. 179 ELECTRON MICKOSCOI'V Fig. 5. Adaxial leaf surface of Pisum satium var. 'Alaska'. Grown at 1500 ft. candles light in- tensity. Fig. 6. Adaxial leaf surface of Pisum sativum var. 'Alaska'. Immediately after transfer from darkness to light. Treating the soil in which peas are germi- nating with 2,2-dichlorpropionic acid ("Dal- apon") affects the susceptibihty of the pea plants subsequently sprayed with certain herbicides (Dewey, Gregory and Pfeiffer, 1956). It was thought that the "Dalapon" interfered with the wax formation of the cuticle and so enhanced the retention and penetnition of the lu'ri)icides. The electron microscope shows that there is a progressive reduction in the complexity, size, and density of the wax projections on the pea leaf sur- faces W'ith increasing concentration of "Dalapon" in the soil. Increases in the con- centration of "Dalapon" in the soil beyond 0.32 lb per acre finally result in a smooth cuticle with no w^ax projections visible at all (Juniper, 1959). The initial development of the fine struc- ture of a leaf is shown by the electron micro- scope to take place before the earliest stage at which replicas can be made. The stages of the development of the fine structure of a pea leaf surface have therefore been followed by growing the pea plant to an advanced stage in darkness (during which time no pro- jections are developed) and then transferring it to light. Replicas are then taken of corre- sponding leaves at 24 hour intervals. The projections from the surface develop very rapidly. Figure 6 is of a pea leaf surface im- mediately after transfer. The appearance of Fig. 4 is achieved after about five days. New techniques and the electron micro- scope have revealed the existence of a fine structure on the surfaces of many plants which had never previously been suspected. Oiu' investigations have confirmed some of the observations made by other workers using different techniques in the same field. It has revealed morphological differences re- sulting from en\'ironmental changes which were not previously suspected because the techniciues for detecting such changes in the fine structure were not sufficiently refined. The results are interesting both to the de- velopmental morphologist and to the agrono- mist because the view that everything external to the epidermal cells is static and immutable can, with present evidence, no longer be held. REFERENCES Bradley, D. E., "Evaporated carbon films for use in electron microscopy," Brit. J. Appl. Phys., 5, 65 (1954). Bradley, D. E., "High-resolution shadow-casting 180 METALS BY TRANSMISSION technique for the electron microscope using the simultaneous evaporation of platinum and carbon," Brit. J. Appl. Phys., 10, 198 (1959). Dewey, O. R., Gregory, P., and Pfeiffer, R,. K., "Factors affecting the susceptibility of peas to selective dinitroherbicides," Brit. Weed Control Conf. Proc, 1, 313 (1956). DoRSCHNER, K. P. AND BucHHOLTz, K. P., "Ef- fect of wetting ability of herbicides on alfalfa stands," Agron. J., 48, 59 (1956). Fogg, G. E., "Quantitative studies on the wetting of leaves by water," Proc. Roy. Soc, B134, 503 (1948). HOLDGATE, M. W., MeNTER, J. W., AND SeAL, M., "A study of an insect's surface by reflection electron microscopy and diffraction," Proc. Int. Conf. Electron Microscopy, 129, 555 (1954). Juniper, B. E. and Bradley, D. E., "The carbon replica technique in the study of the ultra- structure of leaf surfaces," /. Ultrastructure Res., 2, 16 (1958). Juniper, B. E., "The effect of the pre-emergent treatment of peas with trichloracetic acid on the sub-microscopic structure of the leaf surface," NewPhyt., 58, 1 (1959). Mueller, L. E., Carr, P. H., and Loomis, W. E., "The submicroscopic structure of plant sur- faces," Am. J. Bot., 41, 593 (1954). Schieferstein, R. H. and Loomis, W. E., "Wax deposits on leaf surfaces," Plant Physiol., 31, 240 (1956). B. E. Juniper METALS BY TRANSMISSION* Although replica techniques have been used with great success for the examination of metals, these methods have their limita- tions. With the development, therefore, of instruments having high resolving power and fitted with a double condenser lens that en- ables high beam intensities to be obtained on the specimen, a new approach was made to the problem of preparing metals in a form suitable for examination by transmission in the electron microscope. During the past few years a number of suitable techniques for preparing metal foils with a thickness of 100-2000 A have been devised. These tech- * Abridged from paper in Jour. Inst. Metals, 87, 385 (1959), by permission niques, which will be reviewed in this paper, can be classified into three principal groups: (a) Deposition. (6) Deformation. (c) Dissolution. The deposition methods involve the pro- duction of metal foils by condensation of the metal vapor in vacuo, the precipitation or elect rodeposit ion of thin crystals from aque- ous solution, and the casting of foils from the liquid state. These techniques are of Hm- ited use only, since the results obtained from foils prepared in this way are seldom typical of bulk material. The deformation methods used to pre- pare thin foils comprise beating and machin- ing in diamond-bladed microtomes. The former is very limited in its application, but the latter shows interesting possibihties, par- ticularly for the examination of multiphase alloys. The dissolution techniques are perhaps the most generally applicable to the preparation of foils of pure metals and alloys. The meth- ods adopted have involved simple chemical etching, ionic bombardment, and electro- polishing. The latter is the most successful and has been widely used, since it is thought that the structures obtained from foils pre- pared in this way are typical of those to be expected in bulk material. Deposition Methods Vacuum Evaporation. It has been known for some time that thin films of metal, transparent to electrons, can be pro- duced by evaporating a metal in vacuo and condensing the vapor onto a substrate. The evaporated film may then be stripped from the substrate, or the substrate dis- solved away. If the vapor is deposited onto a crystalline substrate at high temperatures, the metal film exhibits "epitaxy," an effect discussed by Pashley (1). In this way it is possible to grow single-crystal fihns of cer- tain metals. The structure of evaporated fihns is not on the whole typical of bulk material, and 181 ELECTRON IVIICKOSCOPY these films arc useful only in affording infor- nuitioii about the properties of thin films as such, although they may give indications of changes which might occur in bulk materials. l^ajioratcd films of iron, subsequently nitridcd, austcnitized, and quenched to mar- tensite, have been examined by Pitsch (2) by transmission electron microscopy. The martensitic product and the mechanism pro- posed for its formation differ markedly from those in bulk material. This is due to the absence of a constraining lattice on either side of the thin film. Takahashi and his co-workers (3-6) have prepared evaporated foils of aluminium- copper alloy by simultaneous evaporation of the two metals. These have been examined by transmission electron microscopy and electron diffraction in the as-deposited con- dition and after heating in the microscope. Again, the structure of these evaporated alloys differs from that of the bulk alloy. Deposition from Solution. Large thin single-crystal flakes of gold have been pre- pared by Suito and Uyeda (7) by reduction of a dilute auric chloride solution. The thin o crystals (100-200 A thick and several mi- crons across) showed trigonal and hexagonal habit and grew parallel to (111). The value of such thin crystals in giving information about bulk properties is limited. Electrotleposition. Two methods have been used to produce thin elect rod eposits for examination in the electron microscope. The first, developed by Weisenberger (8), consists in using thin carbon supporting films as conducting electrodes onto which the metal is electrodeposited. This method can be used to in\'estigate the nucleation of electrodeposits, but beyond this the tech- nique is severely Umited. The second method, first used by Weil and Read (9), consists in electroplating nickel onto copper or zinc and stripping the thin electroplate from the cathode. This techniciue was appHed by Reimer (10) to the study of the epitaxial growth of electrode- posits. Apart from the investigation of the structure of electrodeposits, this technique has little application. Thinning Molten Drops by Surface Tension. Takahashi and Kazato (11) have developed a technique for the preparation of thin metal foils from molten material. Foils are made by dipping an elliptical loop (major axis = 1 cm., minor axis = 0.3 cm.) of wire into molten metal and withdrawing it at the rate of 2 cm. /sec. This must be done in an inert atmosphere, when dealing with metals that oxidize readily in air, but otherwise the technique should be applicable to most met- als and alloys. Alloys of tin and lead, alumin- ium and silver, and aluminium with tin or copper have been investigated by Takahashi and his co-workers (6, 11, 12). It was found that the microstructure of the thin parts of the foil was not typical of bulk material, but that the structure of the thicker portions resembled the microstructure of the bulk alloy. This shows that the transition from thin foil to bulk behavior occurs in the o thickness range of a few thousand Ang- stroms. However, other properties and other alloys may show a different lower limit of thickness for truly bulk behavior, and it is dangerous, therefore, to extend results ob- tained from such foils to bulk material. Deformation jVIethods Thickness Reduction by Mechanical Work. Some very ductile materials such as gold, silver, and platinum can be beaten into foil '^lOOO A thick. Such foils are so heavily distorted as a result of the deformation that they are virtually useless for pro\'iding data on the properties of the metal. Beaten gold foils were examined in the electron mi- croscope by Hirsch, Kelly, and Menter (13), but the examination yielded only limited in- formation. Thin Sections Cut from Bulk Mate- rial with an Ultramicrotome. An ultra- microtome utilizing thermal expansion for advancing the specimen and the magneto- 182 METALS BY TRANSMISSION striction of a nickel rod for withdrawal dur- in multiphase alloys. Even with pure metals ing the return stroke, was developed by a difficulty is the formation of an etching Haanstra (14) for cutting thin biological substructure (17), but this can be avoided by specimens 50-100 A thick could be cut from proper choice of etching reagents and condi- a specially shaped specimen by a diamond tions. knife, but the resulting structure was heavily Ion Bombardment. A method of thin- deformed, ning metals from the bulk state by bom- Reimer (15) used a diamond knife and an barding a thin disc (1-2 ^u thick) of the mate- ult ramie rot ome developed by H. Fernadez- rial with a beam of ions has been developed Moran to cut thin metal foils of Al, Ni, Cu, by Castaing (18, 19). In using ions to re- Au, Fe, Pd, Pt, and Ag, but in this case, too, move metal atoms from a specimen, the en- the foils were very heavily deformed. In spite ergy of the ions must be adjusted to give of the severe deformation introduced, thin random atom removal. Too high an energy sections can be readily prepared in this Avay, of the ions will result in heating and possible and, in the case of alloys containing dis- damage to the structure, while too low an persed phases, the results are not hkely to be energy will give an etching effect. Therefore, very different from those found by examin- critical adjustment of the accelerating po- ing elect ropolished foils. This technique holds tential (3000 V in the case of aluminium) is great promise as a rapid method of obtaining necessary. Attempts to extend this tech- foils, nique to stainless steel (21) and a-brass (18) have been unsuccessful. Dissolution Methods A difficulty in ionic thinning is the very low speed of metal removal — ^10 n in 24 hr. Chemical Etching. To produce a thin Thus, although with great care the method foil that is representative of bulk material, is capable of producing clean, uniform foils it is necessary to remove metal from a large from bulk material, the technique is compli- section without destroying or modifying the cated, tedious, and difficult to control in the structure of the material. One method of final stages. doing this is by chemical etching of a sheet Electropolishing. Heidenreich (20), us- of material 1 /j, or more thick. Hirsch, Kelly, ing an electropolishing technique, produced and Menter (13) examined gold by etching the first thin metal foils from bulk material the beaten foil in a dilute solution of potas- in a form suitable for transmission electron slum cyanide. The foils produced were very microscopy. His techniciue consisted in elec- uneven in thickness, but the heavily cold- tropolishing discs of aluminium and alumin- worked structure of the foil before etching ium-copper alloy, 3 mm. in dia., cut from w^as preserv^ed. Dislocation movements and rolled sheet 125 m thick. The electropolishing arrangements in aluminium were studied by produced holes at the center of the specimen Hirsch, Home, and Whelan (16), on thin and the areas near these holes were thin foils, made from sheet 0.5 /x thick by etching enough to be transparent to electrons. The in dilute hydrofluoric acid. The foils showed foils were often dirty and thin areas were not large uniform areas which were transparent obtained from every specimen, but when to electrons, and there was no evidence of success was achieved the results were very substructure due to etching. encouraging. Man}^ workers have since used This technique produces good results, but electropolishing techniques of different types its application is limited to pure metals or to prepare thin foils of a wide variety of single-phase alloys since it is difficult to pre- metals. These techniques are based mainly vent preferential attack of one of the phases on experiment, with very httle theoretical 183 KI.K( IKON MKHOSCOPY understanding of the conditions necessary to produce a uniform thin foil. Tiie main cUfhculty in i)r()ducing a thin foil is that of attaining random removal of metal from the surface of the specimen. In order to achieve this during electropolishing, the current density at all points on the speci- men must be the same. Since it is difficult to measure the thickness of the metal in the region 200-2000 A, the criterion that the specimen is thin enough is usually the ap- pearance of one or more holes. Thin regions occur near the edges of the holes, but as soon as a hole is formed the current-density dis- tribution is disturbed, with the result that the edges of the holes become rounded and the thin regions lost. Wlien a sheet-metal specimen is placed vertically as the anode in an elect ropolish- ing solution, with a flat vertical cathode, it is found that the current density is maximum at the edges of the specimen and at the "waterUne" of the solution. In addition, elect ropolishing solutions can be divided into two classes: (i) Solutions forming a viscous layer which flows down under gravity. (ii) Solutions which evolve bubbles of gas at the anode. Solutions of the first type will give en- hanced attack at the top of the specimen, since the viscous layer is flowing away from that region and is increasing in thickness as it flows down the specimen. Solutions of the second type will form a blanket of gas bub- bles which is thin at the bottom of the speci- men and thick at the top; as a result metal is removed more quickly from the bottom of the specimen. A horizontal anode eliminates the effect of gravity on the viscous layer and on the bubbles, but now the conditions are different on the two sides of the specimen. Hence, in this case it wiU be difficult to at- tain the correct polishing conditions on both sides of the specimen, especially if the cur- rent-density range for polishing is small for the electrolyte. To eliminate the effect of preferential pol- ishing at the edges of the specimen, these may be coated with a non-conducting lac- fiucr. The result of this is preferential polish- ing at the edge of the lacquer coating, leav- ing the center of the specimen thicker than the outside. To overcome this difficulty, BoUmann (21, 22) used pointed cathodes mounted close to the center of a metal disc (^^2 cm. in dia. and 25-200 /x thick). The specimen was elec- tropolished until a hole appeared in the center, when the electrodes were moved ^^1-2 cm. away and polishing was continued. Perforation then occurred at the edge of the lacquer coating and continued toward the central hole. The specimen was then removed and washed, and foils were cut from the regions near the junction of the two holes. To avoid removing the specimen from the solution to adjust the position of the cath- odes, a modification of the electrode assem- bly has been designed in which there are two sets of cathodes, a set of pointed cath- odes close to the specimen and a set of flat cathodes far away from the specimen. Either set may be connected to the negative of the power supply, so that the same sequence of events can be followed merely by switching from one set to the other. The Bollmann technique is hmited to polishing solutions with a low throwing power and may produce wedge-shaped foils, but otherwise the method is simple and ver- satile. The preferential polishing of a vertical specimen due to the formation of a heavy viscous layer has been utilized in the "win- dow" technique developed by Nicholson, Thomas, and Nutting (23) for use with alu- minium alloys, and later by Tomlinson (24) for making thin foils of Mg, Ni, Al, Cu, Fe, and Co. This consists of lacquering the edges of a thin sheet (25-200 ^l thick) and mount- ing it vertically with a vertical cathode in an electropolishing solution. Perforation occurs at the top edge of the lacquer coating and 184 METALS BY TRANSMISSION advances downward. When perforation has was carried out immediately in methyl alco- extended about half-way down, the specimen hoi kept below — 20°C to prevent attack by is removed and washed, and foils are cut from the anodic layer before it was completely re- the edge of the hole. Tomlinson states that mo^'ed. the foils are more uniform in thickness over Another techniciue employing the viscous larger regions if the current is switched on polishing layer has been developed by Bran- and off rapidly near the end of the polishing, don and Nutting (26) for use with iron. A No explanation of the mechanism of this sheet of material 50-100 n thick was electro- phenomenon has been given. polished down to ^^20 ju in the usual way. Better foils are obtained if the specimen is If perforation occurred during this polish- removed as soon as perforation begins at the ing, its progress was stopped by coating with top, and the perforated edge coated with lacquer. When the thickness of the specimen lacquer. The specimen is then in\-erted, re- had been sufficiently reduced, one surface placed, and polished until perforation occurs was painted with non-conducting lacquer to at the other end of the specimen. This can give a "figure-of-eight" outline. The speci- be repeated until there is a narrow band at men was further electropolished and perfora- the center of the specimen from which the tion occurred at the top of the specimen foils are cut. and under the "arms" of the "eight," leav- The advantages of both the "window" ing the portion between the "arms" suitable and the Bollmann methods have been incor- for thin foils. porated in a low-temperature polishing tech- To a\'oid the preferential rounding off of nique for copper alloys, developed by Swann the edges of holes, Mirand and Saulnier (27) and Nutting (25). The "window" technique have backed the specimen with a piece of is used first, the specimen being inverted the same material; thus, when a hole ap- after each successive perforation, until a pears, attack continues on the backing metal, band about half the width of the original The polishing was continued until fragments specimen is left. The Bollmann techniciue is dropped from the specimen into the electro- then used to thin the center of this band lyte, where they were collected for examina- until perforation occurs. The final polishing tion. This technique works well for titanium is done with the "window" technique. alloys, but the difficulty in extending it to In producing thin foils of copper alloys, other metals lies in finding an appropriate there are two main difficulties. First, the so- polishing solution. Most electropolishing so- lution used polishes very rapidly at room lutions will etch the metal, unless a current temperature and this gives little control over is applied, and therefore etch the fragments the final stages of thinning; secondly, the as soon as they fall from the specimen. Elec- anodic layer is so reactive that even imme- trolyte may also seep between the specimen diate washing after removal of the specimen and the backing piece, thus etching the back from the solution is not enough to prevent of the specimen. However, when suitable etching and oxidizing of the specimen. To conditions have been established the tech- slow down the rate of attack and give the nique of Mirand and Saulnier provides a required degree of control over polishing, the rapid and easy method of preparing thin temperature of the bath was kept below foils. — 20°C. Just before the specimen was re- The production of exact geometrical moved, sufficient liquid nitrogen was poured shapes and the preservation of certain pro- onto the top of the solution to form a layer, files durhig electropolishing have been investi- through which the specimen was withdrawn, gated by Michel (28). This work gives an without switching off the current. Washing indication of the methods to be adopted if 185 ELECTKOiN MlCKOSCOl'Y uniform thinning is to be attained. By plot- ting the equipotentials, Fisher (29) has found the apjM-opriato shape of the anode, placed at a gi\(Mi distance from a pointed cathode, necessary to gi\'e uniform current density over the anode surface. The anode must be pear-shaped at its periphery, with a flat central portion. Such a configuration could be achieved by placing suitable washers on either side of a flat sheet. The specimen perforates almost simultaneously at a nvnn- ber of points in the flat area. Polishing must be stopped as soon as the first holes appear in order to maintain a uniform current-den- sity distribution and to stop rounding of the edges. This technique, which seems the most promising of those reviewed above, has so far been applied to stainless steel, iron, gold, and iron-cobalt. It can obviously be ex- tended to other metals and alloys. Electrolytic Jet Machining. The tech- niques for obtaining thin foils from bulk ma- terial described above are limited to mate- rials that can be obtained in the form of thin sheets <200 ^ thick. The other criticism which can be made of these methods is that thermal and mechanical treatments carried out on sheets --^100 m thick may give results which are not tj^pical of bulk material. To overcome this difficulty and to extend the existing thin-foil techniques to hard mate- rials such as alloy steels, which cannot con- veniently be made into sheets of the required thickness, a technique of removing metal uniformly and rapidly from a specimen ~1 mm. thick has been developed by Kelly and Nutting (30). This consists in machining the specimen electrolytically with a jet of acid. A glass jet ~1 mm. in dia., connected to a copper-tube cathode, is moved horizontally backward and forward at constant velocity 30 times a minute, while the specimen, mounted perpendicular to the jet at a dis- tance of 1-2 mm., is moved vertically up and down once every 6 min., also at constant velocity. The cams imparting these motions have a 1-in. throw, so that a square area on the specimen of 1-in. side is covered by the jet. The accuracy of the machining is better than 2% of the metal removed. Using a hydrochloric acid electrolyte at 2 amp. and 50 V on a steel specimen, the rate of metal removal is -^250 ju/amp./hr. Since it is pos- sible to eliminate the effects of mechanical cutting and grinding by electromachining, without destroying or modifying the struc- ture of the interior of a metal, thin foils can be prepared from as large a specimen as re- quired. In this way foils of 1 % carbon steel and 20 % nickel, 0.8 % carbon steel were pro- duced from material 0.75 mm. thick, elec- tromachined to 75 /x and then electropolished by the Bollmann technique. REFERENCES 1. Pashley, D. W., Advances in Physics, 5, 173 (1956). 2. PiTSCH, W., Phil, Mag., 4, 577 (1959); see also ./. Inst. Metals, 87, 444 (1958-59). 3. Trillat, J. J. AND Takahashi, N., Compt. rend., 235, 1306 (1952). 4. Takahashi, N. and Trillat, J. J., ibid., 237, 1246 (1953). 5. Takahashi, N. and Mihama, K., Acta Met., 5, 159 (1957). 6. Takahashi, N. and Ashinuma, K., /. Inst. Metals, 87, 19 (1958-59). 7. SuiTO, E. AND Uyeda, N., "Proceedings of the Third International Conference on Electron Microscopy," p. 223 (London, 1954). 8. Weisenberger, E., Z. wiss. Mikroskop., 62, 163 (1955). 9. Weil, R. and Read, H. J., J. Appl. Physics, 21, 1068 (1950). 10. Reimer, L., Z. Metallkunde, 47, 631 (1956). 11. Takahashi, N. and Kazato, K., Compt. rend., 243, 1408 (1956). 12. Takahashi, N. and Ashinuma, K., ibid., 246, 3430 (1958). 13. HiRSCH, P. B., Kelly, A. and Menter, J. W., "Proceedings of the Third International Conference on Electron Microscopy," p. 231 (London, 1954). 14. Haanstra, H. B., Philivs Tech. Rev., 17, 178 (1955). 15. Reimer, L., T. Metallkunde, 50, 37 (1959). 16. HiRscH, p. B., HoRNE, R. W. and Whelan, M. J., Phil. Mag., 1, 677 (1956). 17. Phillips, R. and Welsh, N. C, ibid., 3, 801 (1958). 186 MINERALS 18. Castaing, R., "Proceedings of the Third Inter- national Conference on Electron Micros- copy," p. 379 (London, 1954). 19. Castaing, R., Rev. Met., 52, 669 (1955). 20. Heidenreich, R. D., J. Appl. Physics, 20, 993 (1949) . 21. Bollmann, W., "Electron Microscopy: Pro- ceedings of the Stockholm Conference, 1956", p. 316. 1957, (Almqvist and Wiksell), Stockholm. 22. Bollmaxn, W., Phys. Rev., 103, 1588 (1956). 23. Nicholson, R. B., Thomas, G. and Nutting, J., Brit. J. Appl. Physics, 9, 25 (1958). 24. Tomlinson, H. M.,Phil. Mag., 3, 867 (1958). 25. SwANN, p. R. AND Nutting, J., private com- munication. 26. Brandon, D.G. AND Nutting, J., fin7. J. Appl. Physics, 10,255 (1959). 27. MiRAND, P. AND Saulnier, A., Compt. rend., 246, 1688 (1958). 28. Michel, P., Sheet Metal Ind., 26, 2175 (1949). 29. Fisher, R. M., private communication. 30. Kelly, P. M. and Nutting, J., J. Iron Steel Inst., 192,246 (19-59). 31. Nicholson, R. B., Ph.D. Thesis, Cambridge University'. 32. SiLCOX, J. AND Hirsch, p. B., Phil. Mag., 4, 72 (1959). 33. Irving, B. A., Institute of Phj^sics Exeter Conference, 1959. 34. Kerridge, J. F., Johnson, A. A. and Mat- thews, H. I., Nature, 184, 356 (1959). 35. Saulnier, A. and Mirand, P., Compt. rend., 247, 2351 (1958). 36. Wilsdorf, H. G. F., Cinquina, L. and Vak- KER, C. J. Proc. Int. Conf. Electron Micr. (in press) (1958). P. M. Kelly J. Nutting MICROTOMY. See GENERAL MICROSCOPY, p. 343. MINERALS The sizes, shapes and surface features of minerals and mineral particles provide es- sential and sometimes otherwise unobtain- able information about such important sub- jects as atomic structure, the nature of origin and growth, and the reaction of minerals to physical and chemical forces impo.sed either by nature or man. Textural relationships of minerals with other materials of their own or a different kind are often keys to the nature of bonding forces and to the probable behavior of the aggregate — be it rock, soil, ceramic body, paint film, or drilling mud — under the various conditions to which it is subjected. Considered on the })asis of the size range of mineral particles and their surface fea- tures, one-third of the mineral kingdom was "invisible" prior to the advent of electron microscopy. The light microscope, penetrat- ing to the tenths-of-a-micron region, leaves unseen all objects measured in tens and hun- dreds of Angstrom units; and although some of these objects are reproduced in larger min- eral particles, many ha^-e no counterpart in the world of the light microscope or unaided eye. Such, for example, are the detailed fea- tures of tubular crystals of halloysite clay, seen in replica in Figure 1, and of chrysotile asbestos, shown in ultra-thin section in Fig- ure 2. Until 1948 when the tubular form of halloysite crystals was first revealed by the electron microscope and the structure ex- plained (1), it was not known that cylindri- cal crystals with curved planes of atoms existed as stable structures in the mineral kingdom. In this and many other cases the electron microscope removed the barriers standing in the way of direct observation of one of the most fertile parts of the mineral kingdom, that lying beyond the realm of the light microscope and bordered on the other side by the land of measm-ement of the atoms themselves. In mineralogy, therefore, as in other fields, the electron microscope has two functions : to serve as a supplement to the eye and the light microscope by providing fur- ther data on features visible at lower mag- nification; and to reveal to the eye the details of objects indigenous to the 10 to 10,000 A region. The ability of the microscope to perform these functions has depended on the de\'elop- ment of technique. Early mineralogical ap- plications such as studies of mineral dusts, clays, and pigments involved the morphol- 187 ELECTKO.N >llC;H()SCOPY Fig. 1. Halloysite clay, Wendover, Utah. Platinum-carbon replica of a fracture surface showing tubular and lath-shaped crystals. The scale in all illustrations is one micron except where otherwise indicated. (All micrographs not other- wise expedited were taken by Joseph J. Comer, electron microscopist. College of Mineral Indus- tries, The Pennsylvania State University.) X 84 ,300. ogy of fine particles that could be, or already were, dispersed for direct observation. Thus, as soon as the instrument became available it began to fill the long standing needs of the mineralogist concerned with the shape, size, concentration, interaction, behavior, pro- duction, separation, and use of particles measured in microns. The effect in areas of research such as silicosis and clay mineralogy was immediate and dramatic; and although new technicjues have greatly extended the area of applicability of electron microscopy, many present day problems can be soh'ed by the simple technique of directly viewing par- ticles that are dispersed upon a suitable sub- strate. Two techniques provide additional infor- mation lo the scientist interested in the na- ture of particles in dispersed systems : shadow casting and freeze drying. The former soon became standard procedure in electron mi- croscope laboratories because it produced additional contrast and provided a means of measuring particle thickness. The freeze-dry technicjue is invaluable when the morphology and interaction of particles in suspension is important. In Figure 3 the convolutions of the Wyoming bentonite flakes suspended in water were "captured" when the em- bedding water droplet was suddenly frozen. Had the particles been allowed to settle and the water removed by evaporation the resulting shape of the films would bear no relationship to that which, to a large extent, controls the behavior of clay-water bodies. In this illustration note the effect of Fig. 2. Chrysotile asbestos, Transvaal. Ultra- thin section showing circular cross sections of tubular crystals. The scale is 0.1 micron. X94,000. {Courtesy Robert V. Rice, Electron Microscope Laboratory , Mellon Institute) 188 MINERALS shadow-casting in giving depth to the field and reveahng the size, shape and texture of the smaller particles that make up the twisted flakes. The de^'elopment of replfca methods ex- tended the applicability of the electron mi- croscope to particles of any size and to the study of surfaces, thus removing many of the limitations of the instrument for mineralogi- cal and petrographic work. As a result, elec- tron microscopy began to play a more signifi- cant role in studies of crystal growth and solution, chemical reactivity and physical wear, high temperature reactions, texture of natural and artificial mineral aggregates, and applications of minerals in many industrial uses. Early replica work, particularly using plastic films, was restricted to relatively smooth, impermeable and non-porous sur- faces. Later a very important forward step was the successful apphcation of the plati- num-carbon replica technique to clays (2, 3) and other porous substances, thereby extend- ing the method to all mineral materials, in- cluding rocks and soils. Thus, Figure 1 is a platinum-carbon replica of a fractured sur- face of the halloysite clay. The technique is especially suitable for minerals of this type where sensitivity to hydration state may re- sult in severe morphological changes if prep- aration procedures become too involved and time consuming. Thus, replicas of fractured surfaces may indicate a very different parti- cle size distribution than that revealed when particles have been dispersed and allowed to settle on a substrate. Thin sectioning for the electron micro- scope is a technique that, until recently, has been of limited use to mineralogists because of the hardness and brittleness of the mate- rials involved. However, improvements in diamond knives have permitted an impor- tant "breakthrough" in this area (4) and it is to be expected that sectioning of minerals and mineral aggregates will provide impor- tant new data in the future. Figure 2 demon- strates how effectively application of this h -I Fig. 3. Wyoming beutonite, Upton, Wyoming. Freeze-dry preparation of montmorillonite clay from aqueous suspension. X 20 ,200. technique to a bundle of chrysotile asbestos fibers supports previous data as to the tubu- lar morphology and the packing arrange- ment of the individual crystals as seen in cross section (5). Fine-grained Minerals Mineral particles that belong in the size range appropriate to effective electron micro- scope study of morphological features fall into two categories. In one are those that are simply small scale representatives of min- erals which also occur in the form of coarser 189 ELECTKON ^IICKOSCOPY crystals iiiul fra»;moiils easily studied at lower nuignificatiou. Many of the particles in mineral dusts, paints and pigments, fine- grained rocks, ceramic bodies and other syn- thetic aggregates fall in this group. Thus, Figure 4 pictures quartz crystals that have grown in pai-allc^l to near-parallel orientation in a minute fissure in chert, a high-silica rock. Although, for minerals of this group, the moi-phological characteristics and behavior of the finest particles can frequently be pre- dicted from larger specimens of the same material, the electron microscope still pro- vides vital information as to size limitations and distribution of the fine particles; as to surface features, textural relationships, nu- cleation and growth phenomena; and as to chemical and physical stability under vari- ous conditions. The second group, smaller in number, con- sists of the "dwarfs" of the mineral kingdom which, because of restrictions imposed by their crystal chemistry, rarely attain suffi- cient size to be studied effectively with the light microscope. It was at this group that the electron microscope was first "aimed" because it was here that the resolution of the instrument was most needed to provide basic information of the type long known for coarse-grained minerals. The clay minerals were among the first investigated (6, 7, 8, 9) and have been more thoroughly subjected to electron microscopic study than any other mineral family (see, for example, 10, 11, 12, 13, 14). Not only do these minerals fall in just the right size range but, because of the large number of commercial applications of clay and clay products, detailed high magni- fication study has been industrially as well as scientifically important. The morphological complexity of the minerals is revealed by a comparison of Figures 1 and 5, replicas of two Fig. 4. Quartz crystals in chert, Caddo Gap, Arkansas. Particles on most fracture surfaces in this rock do not show the crystal forms exhibited here. X 12,100. 190 MINERALS members of the kaolin group, halloysite and kaolinite, respectively, which have the same chemical composition but crystallize as tubes or laths in the former case (15), and hexago- nal plates in the latter. A similar situation exists in the serpentine minerals where tubes of chrysotile (Figure 2) contrast sharply with laths and flakes of antigorite (16, 17, 18, 19). Such striking morphological differences within single mineral groups reflect differ- ences in environment and growth conditions, and obviously must be taken into account in any application which depends on the physi- cal attributes of the materials. Figure 3, referred to previously, provides further evidence of the large morphological variation among clay minerals. Of all the clays, those of the so-called montmorillonite group demonstrate the greatest morphologi- cal variation with change of hydration state (20, 21, 22). Precise measurement shows that individual sheets approach unit cell thick- o ness of 10 to 30 A yet have a real extent of the order of hundreds of sc^uare microns. Such characteristics are basic to the applica- bility of bentonite in oil-well drilling-mud and as filler in many industrial materials, as well as in many areas of colloid chemistr}^ agron- omy and mineralogy where ion exchange and surface behavior of such materials are of interest. Synthetic minerals occur in both groups discussed above but deserve particular men- tion because of the important role played by the electron microscope in their study. In the field of geochemistry, particularly, where much emphasis is placed on high tem- perature, high pressure phase equilibrium studies, conditions are frec^uently not con- ducive to the formation of large crystals and the products may be visible only at high magnification. Such is the case for synthetic chrysotile (Figure 6) produced from Mg-Si gels under hydrothermal conditions at a tem- perature of 450°C and a pressure of 40,000 psi (23). The tubular character of the fibers is apparent, as is also the "cone-in-cone" Fig. 5. Kaolinite-coated paper. The character- istic platy form and hexagonal outline and cleav- age of the clay particles are apparent in this platinum-carbon replica of a sheet of high-gloss paper. X 15,000. growth feature only rarely observed in natu- ral material. Minerals are now being synthesized under a wide variety of conditions, many for use in numerous industrial applications, and the electron microscope has an important role in the research involved with the development, production and control of the product. Fig- ure 7 shows an aggregate of crystals of a synthetic zeolite, the mineral group so im- portant in water softening. Whether the minerals be synthetic or nat- 191 ELECTRON MICKOSCOrY Fig. 6. Synthetic chrysotile asl:»estos. The tubuhir crystals have internal diameters of 50-lOOA but vary widely in outer diameter. The "cone-in-cone" habit is common only in synthetic material. X02,000. ural the electron microscope has been inval- uable in studying their applicability to and usefulness in many industrial applications. The replica showing the kaolinite flakes in Figure 5, for example, was made of a sheet of paper such as that you are now reading, the role of the clay particles being to provide the high gloss necessary for good reproduction of illustrations. To obtain maximum reflectiv- ity the perfection and orientation of individ- ual crystals is of utmost importance, and platinum-carbon replicas such as that shown have provided vital information to the paper coating trade (24). Similarly, fundamental work has been done on subjects such as Ti02 , iron oxide and other mineral pigments in paint and enamel (25, 26, 27, 28), crystals in glass (29, 30), diamond powders (31), and calcium-silicate hydrates in cement (32, 33, 34), just to mention a few. Along "biomineralogical" lines much re- search has been conducted on such subjects as mineral dust (35, 36) and its concentra- tion in lung tissue (37, 38) ; and the structure and chemical resistance of apatite tooth enamel (39, 40). Diatoms with their intricate structures have long been a fa\'orite subject of microscopists (41, 42), and high magnifi- cation studies of their structures and mode of occurrence in diatomaceous earth are of academic and industrial importance. Studies of coal surfaces have revealed structures, such as those shown in Figure 8, which are presumably of botanical origin. Surface Features of Minerals and Min- eral Aggregates The development of replica techniques opened the way to the study of mineral structures which either were not well repre- sented or could not be observed on or in mineral fragments. Study of fracture sur- faces of crystals or fragments of single min- erals provides evidence about such features as cleavage, fracture patterns, and inclusions (43, 44, 45, 46) ; whereas in compomids, ag- gregate textures, morphology and relative hardness of mineral components, and the amount and role of cementing material can be evaluated. External surfaces yield data on features produced during growth or by subsequent physical or chemical action under conditions imposed by nature or man. This area of study penetrates many branches of science and thousands of electron micro- graphs of mineral materials appear through- out the scientific literature of the miner- alogist, chemist, physicist, soil scientist, ceramist and engineer. Surface changes on catalysts (47), the abrasion of diamond (48), oxidation of sulfides (49), dislocations in crystals (50, 51), grain growth at high tem- peratures (52), nucleation of ice crystals (53), epitaxial relationships (54), gliding, twin- ning, scratches: all these and many more 192 MINERALS Fig. 7. Synthetic zeolite A. The crystals are bounded by cube and dodecahedron faces and show good penetration twins. X 16,500. (Courtesy of D. W. Breck, Linde Company) Fig. 8. Subbituminous coal, Gilette, Wyoming. The platinum-carbon replica was obtained from a collodion replica stripped from the fractured coal surface. The identity of the various objects is un- known. X11,000. (Courtesy of R. R. Dutcher, Coal Petrography Laboratory , College of Mineral Indus- tries, The Pennsylvania State University) 193 ELKCIKON AIKKOSCOPY Fig. 9. Synthetic sylvite (KCl). The beauti- fully revealed cubic cleavage contrasts strikingly with curved surfaces. X 10,300. Fig. 10. Quartz fracture surface. Collodion- carbon replica, Pd-shadowed. Evidence of crystal- lographic control of the fracture pattern is prom- inent in the fine as well as the coarse fracture lines. X5,700. {Courtesy of A. van Valkenburg and Elisabeth Mitchell, Constitution and Microstructure Section, Mineral Products Division, National Bureau of Standards) subjects have received much attention from electron microscopists. Figures 9-14, illus- trating some typical surface features, can only suggest the nature and extent of this large and rapidly expanding area of mineral- ogical study. Figures 9 and 10 contrast the cubic cleav- age of sylvite (KCl) with the more irregular but apparently crystallographically-con- trolled fracture parallel to the X-cut of a o quartz crystal. Figure 11 shows 50 A growth steps observed at very high magnification on a crystal of the zeolite pictured in Figure 7. In studies such as this the electron micro- scope has played an important role in help- ing to delineate the type of growth step pictured here from that produced in spiral growth. Etching techniques have long been im- portant in the investigation of structure, crystal growth and chemical stability (55) and the electron microscope has greatly ex- tended the apphcability and utility of this Fig. 11. Synthetic zeolite A. Platinum -carbon replica showing 50 A growth steps on a cube face. X59,000. (Courtesy ofD. W. Breck, Linde Company} 194 MINERALS method of investigation (56, 57). The vastly different results produced by etching syn- thetic NaCl with ethyl alcohol as compared to water are illustrated in Figure 12 (58). An entirely different type of surface reaction is shown in Figure 13 which pictures crystals of mullite which have formed from a kaolinite clay surface (see Figure 5) held at a tempera- ture of 1130°C for 20 hours. The arrange- ment of the mullite needles results from the pseudo-hexagonal symmetry of the clay min- eral. Similar application of electron micros- copy to the study of minerals in various ceramic materials has been extensive (see 59 for references). The type of result that may be obtained from a study of fine-grained aggregates is illustrated in Figures 1, 4 and 14. Whereas the first tAvo show textural and crystallo- graphic features of particles in monomineral- lic specimens, the pyrite in Figure 14 occurs in a fine-grained shale containing many min- eral varieties. The area pictured here is part of a pyrite nodule made up of an assemblage of pyritohedrons in parallel orientation. Other examples of electron microscopy in petrographic studies involve studies of such subjects as slate (60), ore textures (61, 62), meteorites (63), and glacial till (64) as well as cherts and fine-grained limestones. Despite the extensive research done with the electron microscope in the study of minerals and rocks (for general articles see 65, 66, 67, 68, 69), the potential of the instru- ment is far from realized. One reason has been the tendency to regard the areas cov- ered by electron and light microscopy as separate rather than overlapping fields. With the increased application of the replica tech- nique supplemented by devices designed to Fig. 12a 195 ELECTRON MICKOSCOPY Fig. 12b Fig. 12. Synthetic halite (NaCl). A, Chromium-shadowed formvar replica of a crystal face etched with water. X15,300. B, the same after etching with ethyl alcohol. X36,500. (Courtesy of A. Oberlin, Labora- toire de Mineralogie, Faculte des Sciences, Paris) Fig. 13. MuUite crytals grown from a kaolinite surface at 1320°C. The hexagonal arrangement re- sults from the structure of the parent material. X 14,000. 196 MINERALS 1^ g 4^*-^ %gf *^^ Fig. 14. Pyrite nodule, Chattanooga shale. Fracture took place around rather than through individual crystals revealing their pvritohedral habit. X 16,700. Fig. 15. Step fracture of calcite crystal in Solenhofen limestone. X28,000. (Courtesy of H. P. Studer, Shell Development Company) facilitate study of a selected area of the sub- ject with both light and electrons, the bound- aries between the two disciplines are dis- appearing and a more integrated picture is being obtained of the entire mineral king- dom. With the further development of ultra- thin sectioning of hard and brittle substances it is probable that no morphological feature of minerals and mineral compounds will long remain "invisible." REFERENCES 1. Bates, T. F., Hildebrand, F. A. and Swine- ford, A., "Morphologj' and structure of endellite and halloysite," Geol. Soc. Am. Bull, 59, 1310 (1948); ^TO.MweraL, 35, 463- 485(1950). 2. Comer, J. J. and Turley, J. W., "Replica studies of bulk clays," /. Appl. Phys., 26, 346-350 (1955). 3. Bates, T. F. and Comer, J. J., "Electron mi- croscopy of clay surfaces," Natl. Acad. Sci.-Natl. Res. Council Publ. 395, 1-25 (1955) . 4. Pfefferkorn, G., Themann, H., and Urban, H., "Anwendungen der mikrotomschnitt- technik auf elektronenmikroskopische mine- raluntersuchungen," Proc. Stockholm Conf. on Electron Microscopy, 333-334 (1956). 5. Maser, M., Rice, R. V., and Klug, H. P., "Chrysotile morphology," Am. Mineral. 45, 680-688 (1960). 6. Eitel, W. H., Muller, H. 0. and Radczew- SKi, O. E., "Examination of clay minerals with the ultramicroscope," Ber. Deut. Keram. Ges., 20, 165-180 (1938). 7. Ardenne, M. von, Endell, K., and Hoff- mann, E., "Examination of the finest frac- tions of bentonite and cement with the universal microscope," Ber. Deut. Keram. Ges., 21, 209-227 (1940). 8. Humbert, R. P. and Shaw, B. T., "Studies of clay particles with the electron micro- scope," Soil Sci., 52, 481-487 (1941). 9. Endell, K., "Significance of the electron microscope for identifying the microstruc- ture of clays," Tonind. Ztg.,Q5, 69-72 (1941). 10. PiNSKER, Z. G., "Electrographic and elec- tronomicroscopic study of clay minerals," Trudy Biogeokhim. Lab., Akad. Nauk SSSR, 10, 116-141 (1954). 11. Bates, T. F., "Electron microscopy as a method of identifying clays," Proc. 1st Natl. Conf. on Clays and Clay Minerals, 197 ELECTRON MICKOSCOPY Calif. Dopt. Xal. Res., Div. Mines Bull. 169, 130-150 (1955). 12. T.\GGART, M. S., Mulligan, W. O., and Studer, H. p., "Electron micrographic studios of clays," Nail. Acad. Sci.—Natl. Res. Council Publ. 395, 31-64 (1955). 13. Neuwirth, E., "The determination of clay minerals with the electron microscope," Tschcnnaks mineralog. u. petrog. mitt., 5, 347-3G1 (1956). 14. SuiTO, EiJi, "Electron microscopy of clay minerals," Nendokagaku no Shimpo, 1, 166- 180 (1959). 15. Bates, T. F. and Comer, J. J., "Further ob- servations on the morphology of chrysotile and halloysite," "Clays and Clay Min- erals," 237-248, Pergamon Press, New York, 1959. 16. Bates, T. F., Sand, L. B., and Mink, J. F., "Tubular crj'stals of chrysotile asbestos," Science, 111, 512-513 (1950). 17. Noll, W. and Kircher, H., "Zur morphologic des chrysotilasbest," Naturwiss., 37, 540-1 (1950). 18. ZussMAN, J., Brindley, G. W., and Comer, J. J., "Electron diffraction studies of serpen- tine minerals," Am. Mineral., 42, 133-153 (1957). 19. Kalousek, G. L. and Muttart, L. E., "Stud- ies on the chrysotile and antigorite compo- nents of serpentine," Am. Mineral., 42, 1-22 (1957). 20. Corbet, H. C. and Wolffes, J., "The use of a freeze-drying technique in the investiga- tion of sodium-montmorillonite by electron microscopy," "Proc. Stockholm Conf. on Electron Microscopy," 334-335 (1956). 21. Mering, J., Oberlin, A., and Villiere, J., "Etude par electrodeposition de la mor- phologic des montmorillonites. Effet des cation calcium," Bull. Soc. franc, mineral, et crist., 79, 515-522 (1956). 22. HoFMANN, U., Fahn, R., and Weiss, A., "Thixotropie bei kaolinit und innerkristal- line quellung bei montmorillonit. Wirkung der austauschfahigen kationen der flussig- keit und der elektrolj^te einer wassrigen losung," Kolloid-Z., 151, 97-115 (1957). 23. Roy, D. M. and Roy, R., "An experimental study of the formation and properties of synthetic serpentines and related layer silicate minerals," Am. Mineral., 39, 957- 975 (1954). 24. Comer, J. J., Stetson, H. W., and Lyons, S. C, "A new replica technique for making electron micrographs of surfaces of paper sheets," Tech. Assn. Pulp Paper Ind., 38, 620 (1955). 25. Schmieder, F., "Electron microscopic investi- gation of the relation between covering quality and grain size of pigments," Kolloid- Z., 95, 29-33 (1941). 26. Honda, T. and Mukai, F., "Electron micro- scopic observation of pigment (hematite) particles in a porcelain glaze," Rept. Osaka Pref. Ind. Research Inst., 3, 56-57 (1950). 27. Cooper, A. C, "The electron microscope examination of pigment powders and their dispersions in oil media," 1st Int. Cong. Electron Micros. Paris, 1950, Proc. 461-463, Paris Ed. Rev. Optique. 28. Smith, N. D. P., "The electron microscope in the study of paints," Trans. Inst. Metal Finishing, Advance Copy, No. 12, 10 pp. (1955). 29. Bates, T. F. and Black, M. V., "Electron microscope investigation of opal glass," Glass hid., 29, 487 (1948). 30. Fadeeva, V. S. AND Yakovleva, M. E., "Investigation of opaque zirconium glazes with the electron microscope," Doklady Akad. Nauk SSSR, 81, 667-669 (1951). 31. Meldau, R. and Harsewinkel, "Electron microscopy investigations of diamond pow- ders; Electron-optical study of types I and II diamonds," Industrial Diam. Rev., 11, 7-8 (1951). 32. Radczewski, O. E., Muller, H. O., and EiTEL, W., "The hydration of tricalcium aluminate," Naturiviss., 27, 837-838 (1939). 33. Sliepcevich, C. M., Gildart, L., and Katz, D. M., "Crystals from Portland cement hydration; an electron microscope study," Ind. Eng. Chem., 35, 1178-1187 (1943). 34. Shekhter, a. B., Ser-Serbina, N. N., and Rebinder, p. a., "Investigation with the electron microscope of the effect of a surface admixture on the crystallization of hydrated minerals of cement clinker," Doklady Akad. Nauk., SSSR, 89, 129-132 (1953). 35. Cartwright, J. and Skidmore, J. W., "The measurement of size and concentration of airborne dusts with the electron micro- scope," Ministry Fuel and Power (G. Brit.) Safety in Mines Research Estab. Research Rept. No. 79, 31 p., 1953. 36. Talbot, J. H., "Identification of minerals present in mine dusts by electron diffraction and electron microscopy," Proc. Stockholm Conf. on Electron Microscopy, 353-355, 1956. 37. Pfefferkorn, G., "Vergleichende unter- suchungen uber den mineralinhalt von 198 MINERALS silikoselungen," Arch. Hyg. u. Bakteriol., 135, 14-22 (1951). 38. RiMSKY, A., Oberlin, A., and Ceccaldi, P. F., "Le diagnostic cristallographique des silicoses pulmonaires," Bull. Soc. Franc. Mineral, et Crist., 78, 418-424 (1955). 39. Scott, D. B. and Wyckoff, R. W. G., "Stud- ies of tooth surface structure by optical and electron microscopy," /. Ain. Dent. Assn., 39, 275-282 (1949). 40. Gray, J. A., Schweizer, H. C., Rosevear, F. B., AND Broge, R. W., "Electron micro- scopic observations of the differences in the effects of stannous fluoride and sodium fluoride on dental enamel," /. Dental Res., 37, 638-648 (1958). 41. Krause, F., "Electron-optical pictures of diatoms with the magnetic electron micro- scope," Zeits.f. Physik, 102, 417-422 (1936). 42. Mahl, H., "Photographs of diatoms with the supermicroscope," Naturwiss., 27, 417 (1939). 43. Zapffe, C. a. and Worden, C. O., "Fractog- raphy as a technique in crystal chemistry," Acta Cryst., 2, 377-382 (1949). 44. Pfefferkorn, G. and Westerman, H., "Electron microscope studj' of the deforma- tion of calcite," Neues Jahrb. Mineral., Monatsh, 1951, 97-103 (1952). 45. Seal, M. A., "Reflection electron microscope study of diamond cleavage surfaces," Proc. Stockholm Conf. on Electron Microscopy, 341-343, 1956. 46. Wolff, G. A. and Broder, J. D., "Micro- cleavage, bonding character and surface structure in materials with tetrahedral coordination," Acta. Cryst., 12, 313-323 (1959). 47. Shekhter, a. B. and Tretyakov, I. I., "Electron-microscopic study of the surface changes of massive catalysts at work," Bull. Acad. Sci. USSR Div. Chem. Sci., 397-402 (1953). 48. Seal, M., "The abrasion of diamond," Proc. Royal Soc. A, 248, 379-393 (1958). 49. Watanabe, M., "Submicroscopic topography of oxidized surface of stibnite and etch surface of sphalerite," J . Phys. Japan, 12, 874-882 (1957). 50. Indenbom, V. L., "Dislocations in crystals," Soviet Physics. Crystallography, 3, 112-132 (1957). 51. Hashimoto, H. and Uyeda, R., "Detection of dislocation by the moire patterns in electron micrographs," Acta Cryst. 10, 143 (1957). 52. Stuijts, a. L. and Haanstra, H. B., "Grain growth in a ceramic material, studied by means of electron microscopy," Trabajos reuion Intern. Reactividad Solidos, 30 Ma- drid, 1, 671-701, 1956. 53. Maruyama, H., "Electron microscopy study of the ice crystal nuclei," Papers Meterolo. and Geophys. (Tokyo), 7, 251-266 (1956). 54. HocART, R. AND Oberlin, a., "Epita.xie des microcristaux d'or sur le graphite. Etude au microscope electronique et par diffrac- tion electronique," Compt. Rend. Acad. Sci.. 239, 1228-1230 (1954). 55. Honess, a. p., "The nature, origin, and inter- pretation of the etch figures on crystals," 164 pp. Wiley and Sons, New York, 1927. 56. Pfefferkorn, G., "Electronmicroscopic in- vestigation of calcite and its true crystal structure," Optik 7, Sonderheft 2, 208-216 (1950). 57. Stanley, R. C., "Etch pits on calcite cleavage faces," Nature, 183, 1548 (1959). 58. HocART, R., Roult, G., and Oberlin, A., "Decroissements par strates d'une face (001) de chlorure de sodium naturel et artificiel. Observations au microscope elec- tronique," Compt. Rend. Acad. Sci. 240, 2245-2247 (1955). 59. Comer, J. J., "The electron microscope in the study of minerals and ceramics," Am. Soc. Testing Materials, Spec. Tech. Publ. 257, 94-120 (1959). 60. Bates, T. F., "Investigation of the micaceous minerals in slate," Am. Mineral., 32, 625- 636 (1947). 61. Syromyatnikov, F. V. and Filimonov, A. F., "Study of ore structure with the help of the electron microscope," Izvest. Akad. Nauk. SSSR., Ser. Geol., 5, 135-140 (1953). 62. Ohmachi, H. and Hagamara, T., "On the studies of structure and texture of the manganiferous iron ore by electron micros- copy," Kagaku {Tokyo), 25, 637-638 (1955). 63. Orsini, p. G. and Maggi, N., "Electron microscopy of an iron meteorite," Gazz. Chim. Hal., 88, 482-486 (1958). 64. Droste, J. B., White, G. W., and Vatter, A. E., "Electron microscopy of the till matrix," J. Sediment. Petrol., 28, 345-350 (1958). 65. O'Daniel, H., "jSIineralogical research with the electron microscope," 1939-1946, Min- eralogy, 1-6 (1948). 66. Bates, T. F., "The electron microscope ap- plied to geological re.search," Neiv York Acad. Sci. Trans. Ser. II, 11, 100-108 (1949). 67. Oberlin, A., "Application of the electron 199 ELECTRON iMICROSCOPY microscope iu the study of crystallized media," Bvll. Soc. Franc. Mineral, et Crist. 77, 833-839 (1954). 68. DwoRNiK, E. AND Ross, M., "Application of electron microscope to niineralogic studies," Am. Mineral. 40, 261-274 (1955). 69. Fahn, R., "Applications of the electron micro- scope for the investigation of rocks and minerals," Tonind. Ztg. u. Keram. Rund- schau, 80, 171-180 (1956). Thomas F. Bates PAINT SURFACE REPLICA TECHNIQUES Since it is impractical to view paint sur- faces directly because of difficulties in elec- tron transmission it was necessary to develop indirect methods. These procedures classi- fied as replicathig methods are based upon a reproduction of the surface. The reproduc- tion may be in the form of a negative or positive replica. In the negative replica the heights and hollows are reversed with re- spect to the original surface. The positive replica reciuires an intermediate replica from which the final replica is made and the heights and hollows are in the same relative order as on the original surface. As a rule both methods employ metal shadow casting to enhance the contrast of the replica and to emphasize the dimension normal to the sur- face. Both negative and positive replication techniques may be employed in the study of paint surfaces. However, the replica tech- nique employed is governed strictly by the nature of the surface; that is, the solvent system used in replication should not attack the paint surface and introduce artifacts. A detailed description of various replica tech- niques which may be used in the examina- tion of paint surfaces will be presented as well as examples illustrating paint films in var- ious stages of degradation. Xo attempt will be made to discuss the mechanism of paint film formation since this paper will be con- fined strictly to techniques. Replicating Systems The organic nature of most paint sur- faces presents a problem in replication since the solvents used may attack the organic phase and cause some uncertainty in the rep- lication process. However, this problem may be circumvented by using water-soluble in- termediate plastic replicas or by carefully selecting replicating media the solvent sys- tems of which are known not to be miscible with the paint surface under study. Perhaps the simplest method of replica- ting paint films is by making unbacked negative replicas; this method will be dis- cussed first. Negative Replica (Unbacked). In the unbacked negative replication technique the surface is reproduced in such a manner that the replica represents a surface in which the heights and depressions are reversed with re- spect to the original surface. This method employs only one replicating medium. A schematic representation of the steps used in the preparation of an unbacked negative replica of a paint surface is given in Figure 1 . A plastic solution from a dropper is flowed over a small portion of the paint surface and the excess solution drained off by drying the film at room temperature on the paint panel in a vertical position. Some typical plastic solutions are given below: (1) 1-2% parlodion or collodion in amyl acetate. (2) Freshly prepared 1 % solution of Form- var in ethylene dichloride or dioxane. (3) 1-2 % ethyl cellulose in ethylene di- chloride. One should remember that the chemical nature of the sample should be such that it is not attacked by the replicating solution. To remove the replica from the paint sur- face, moisture from the breath is condensed onto the surface and then removed with Scotch tape; first, however, several specimen screens about %" apart are placed on the replica. If there is some tendency for the replica to wrinkle on the screen after strip- 200 PAINT SURFACE REPLICA TECHNIQUES ping, this difficulty maj' be eliminated by immersing the screen in dilute solution of an adhesive before placing the screen on the replica. The tape is carefully pulle'd away and the plastic replica of the surface removed with it. To remove film-coated screen from the Scotch tape, a dissecting needle which has been sharpened to a fine point is used to cut the film around the edge of screen. Fine- tipped tweezers are then used to remove the screen from the tape. Sticking of the screen to the tape may be eliminated by placing a small circular disk of paper (}^'' diameter) between the tape and screen. The replica is then shadowcast by coating the surface with a suitable metal, but more about the shadow- casting process later. Negative Replica (Backed). An alter- nate procedure for preparing a negative rep- lica of a paint surface is to back the negative replica with a heavy coating of plastic, as shown schematically in Figure 2. This pro- cedure is used in cases where stripping an unsupported film is difficult, as in rough paint surfaces, and the resultant replica would be distorted or torn b}^ stresses ap- plied in stripping. Plastic # 1 is applied to the surface with a dropper and allowed to drain vertically. When dry, plastic # 2, in a heavy concentration which is not miscible with plastic ^1, is applied similarly and dried. Some typical plastic combinations which have been applied successfully to the study of selected paint surfaces are given below. (A) 1-2% solution of 'Tormvar" in ethyl- ene dichloride followed with 2-3% solution of "Parlodion" in amyl acetate. (B) 1-2% solution of "Parlodion" in amyl acetate followed with 4-5% of polyvinyl alcohol in water. The double ffim is then stripped and brought in contact with an appropriate sol- vent to remove the backing film. In the case of (A) amyl acetate would be used to remove the "Parlodion" while in (B) warm water ^tc^- Fig. 1. Negative replica technique (unbacked). ^LC^- Fig. 2. Negative replica technique (backed). Fig. 3. Intermediate shadowed negative rep- lica. would be used to solubilize the poly\dnyl al- cohol. In both cases, plastic # 1 should be uppermost and once solution is complete the replica is mounted on grids (200 mesh stain- less steel wire disks) brought from below and shadowcast with an appropriate metal. Intermediate Shadowed Negative Replica. Another variation of the negative replica techniciue requires the use of an in- termediate as shown in Figure 3. In this case, a thick plastic is applied to the paint surface, dry stripped with Scotch tape, in- verted and shadowed obliquely with a suit- able metal and then coated at normal inci- o __ o dence with 200 A of silica or 100 A of carbon in the vacuum evaporator. The plastic rep- lica is removed by washing in a suitable 201 ELECTRON MICROSCOPY solvent. The silica or carbon replicas are picked up on specimen screens and studied as usual. The use of carbon and silica will be covered in greater detail in the section on positive techniques. Preshadowed repli- cas afford the obvious advantage of provid- ing better detail by eliminating the inter- ference of the replica structure itself. Shadow Casting Shadowmg of paint replicas with metal is needed in order to increase the contrast since the replica alone scatters electrons more or less uniformly over the entire area and con- trast in electron microscopy is dependent mainly on the difference in electron scatter- ing among area increments of the specimen. After shadowing, areas in which the depos- ited metal is the densest are the most opaque to the electron beam. This process of shadowcasting requires the use of a vacuum evaporator which consists essentially of a bell jar evacuated by a dif- fusion pump backed with a forepump. The mechanics of evaporation can be best illus- trated with the aid of Figure 4. Here, the replica preparation at B is placed below and to one side of the tungsten filament C, which can be charged with a given weight of metal and heated electrically to the temperature required for ^-aporization of the metal. If a sufficiently good vacuum exists in the bell jar at the time of evaporation (10~^ mm. of Hg) the metal atoms will travel in straight hues in all directions from the filament and some of them will be deposited on the sam- S4MPLE Fig. 4. High vacuum evaporation. pie. If a metal is used that does not ingrate after deposit, its thickness will be greatest on those aspects of the preparation that face the oncoming atoms, and there will be re- gions of the substrate lying behind high de- tail that will be so shielded as to receive no metal. These clear regions will appear as if they were shadows when viewed under the electron microscope. The varying opacity that results from the varying thickness of the metal creates the impression that one is see- ing the preparation in three dimensions il- luminated by light having the direction of the oncoming shadowing atoms. Although in principle the process of vac- uum evaporation is simple there are certain conditions which must be fulfilled if satis- factory preparations of paint films are to re- sult. A suitable shadowing material must be used and the evaporation must take place under vacuum. It would be expected that sharp shadows could not be obtained unless the mean free atomic path in the vacuum evaporator bell jar exceeds the distance from the heater to the sample, that is, unless the evaporated atoms can follow a collision-free path. Experience has shown that the vacuum should be considerably better than this mini- mum. Seemingly, this good vacuum is needed partly for adec^uate degassing of the surface of the preparation before evaporation and partly to take care of some of the gas evolved during evaporation. The best material to use for shadowing will depend both on the type of preparation and the kind of observation to be made. A shadowing material must not have any per- ceptible structure of its own and must not migrate over the face of the paint replica af- ter deposition. In addition, it must absorb or scatter so strongly that the needed thickness does not appreciably distort the shape or de- tail of the sample. Chromium was the first metal to be suc- cessfully used for metal shadowcasting and it continues to be used for much work. How- ever, when the thickness of the chromium 202 PAINT SURFACE REPLICA TECHNIQUES layer (40A) is objectionable, either platinum o or palladium (4-8A) may be substituted. The metal is deposited at an angle of from tan~i 1:1 or 1:3 depending on the inherent roughness of the surface. The lower angle is preferable for fine detail (high gloss paint surfaces) whereas the higher angle is used for coarse structures (weathered paint surfaces). Positive Replica Technique A schematic representation of the steps used in the preparation of the positive replica is given in Figure 5. In this technique (3) an intermediate thick replica is reciuired from which a thinner second replica can be stripped. Careful control of the solution con- centration of the intermediate replicating medium is highly critical since too low a concentration results in stripping difficulties while too high a concentration results in ex- cessive shrinkage of the medium at the con- tact surface of the sample resulting in the development of wrinkles. For paint surfaces a 5 % aqueous solution of methylcellulose (Methocel, Dow Chemi- cal Co.) or poly\'inyl alcohol (DuPont) in distilled water is employed as the inter- mediate replicas. The solution is applied uni- formly with the aid of a dropper maintaining the paint surface in a horizontal position. When dry, the film is detached slightly with a sharp razor and then carefully pulled off with tweezers. The stripped film is then at- tached with Scotch tape to a microscope slide and placed in the vacuum evaporator for coating with carbon or silica at normal incidence. i/e* oi«. cmeoN kods ^ ^ IIIV£«1T VERTICIL 0EP0S1TI0M :>( C»R90N 0« SILICI CHROHIU SHtDONE / tt.r.:M C3 Fig. 5. Positive replica technique. Fig. 6. Carbon evaporation apparatus. Carbon Evaporation Procedure The method of carbon evaporation consists of passing 60-cycle alternating current of approximate^ 20 to 50 amperes through y^" diameter graphite rods (National Carbon, pure spectrographic grade) in a vertical posi- tion above the specimen with one rod ta- pered to a fine point. The carbon rods are held in close contact by a spring as showai in Figure 6 so they do not separate during the evaporation process. Intense local heating occurs at the regions of contact. A satisfactory carbon thickness is main- tained by visually observing the color (light brown) developed on a piece of white porce- lain containing a small drop of oil (Apiezon B) (1, 2). The condensed carbon is clearly visible on the porcelain but not on the oil drop which shows up in sharp contrast. When the evaporation is completed the carbon plastic -coated slide is removed; cut into 0.3 cm squares with a sharp razor and placed on the surface of a Petri dish contain- ing distilled water until the methycellulose is completely dissolved. From two to three hours are required for complete solution of the methylcellulose, with frequent replace- ment of distilled water. Specimen screens are placed on top of the floating carbon sections and remo^'ed with forceps. When dry, the carbon-coated screens are placed in the evap- orator and shadowcast with a suitable metal. 203 ELECTRON MICROSCOPY Silira Kvaporation Procedure In the process of silica evaporation (4, 5) the intermediate repUca of the paint film is placed about 7 cm below a conical tungsten (O.o mm) wire basket containing about a 1 mgm chip of ciuartz. The evaporator is evac- uated to about 10-* mm of Hg and a current of 20 to 30 amps passed through the basket for 20 to 30 seconds. If some difficulty is experienced in evaporating the (quartz) sil- ica, a colloidal suspension of colloidal silica (Ludox) may be substituted and micro- pipetted on a comparable tungsten filament, dried and e^'aporated. Another alternative is to use SiO which evaporates more readily than (quartz and does not attack the tung- sten. As in the carbon technique, the plastic- silica combination is immersed in water to solubilize the plastic. Once solution is com- plete the silica replica is picked up on speci- men screens, shadowcast and examined as normally. Sometimes if the silica films are extremely thin, special lighting conditions may have to be employed to enhance the visibility. Photof;ra|>hic Procedure An RCA-type EMU electron microscope equipped with self -biased electron gun and binocular viewer was employed. A platinum objective aperture with an opening of 50 microns was used to increase the contrast in the final image. Negatives were taken at an electronic magnification of 1500 X and en- larged optically to 8000 X. Kodak Medium Plates were used and developed in Kodak Versatol Developer. Electron micrographs of the paint surfaces prepared by the negative replica technique were processed to the posi- tive print stage, while micrographs prepared by the positive replica technique were proc- essed to the negative print stage. This was done to make heights and depressions more readily appreciated. 1 r If Ik .» 1 m M > « ' V •• • * • • * . • % «^ l-«».^ 4 « fk- 'PI 4 * « A **, "♦ « V > . * ' ■ <* V m .■ (a) Cellular pattern (X 8000) (b) Pigment flocculation near air interface (X 8000) Fig. 7. Negative replicas of paint surfaces; a. Unbacked technique; b. Backed technique. 204 PAINT SURFACE REPLICA TECHNIQUES (a) Enamel (X 8000) (b) Flat paint (X 8000) Fig. 8. Positive replicas of two different paint surfaces. (a) Before weathering (X 8000) (b) After weathering (X 8000) Fig. 9. Positive replicas of paint surfaces. 205 ELECTRON ^MICROSCOPY Applications PATHOLOGY: KIDNEY A typical example of the utilization of the The method for renal biopsy was intro- unbacked negative replica technicjue for the diiced by Iversen et at. (1951) and further examination of a paint surface is illustrated developed by Kark and Muerhcke (1954). in Figure 7a. The surface in this case is It affords an excellent opportunity to collect basically smooth, illustrating a typical cellu- material for electron microscopy from living lar pattern. Loose material attached to the patients with different stages of kidney dis- replica during stripping can be observed in ease. As a consequence of this the electron the background. The micrograph of the paint microscope is used in many institutes all over surface illustrated in Figure 7b was prepared the world where human renal pathology is by the backed negative replica technique, studied, and several reports have been pub- The roughness in this case appears to be lished concerning this subject.* caused primarily by pigment flocculates near A large number of electron microscopic in- the air interface. vestigations of experimentally produced re- The positive replica technique using a wa- nal lesions in animals have also been per- ter-soluble intermediate with either carbon formed. It must be emphasized, however, or silica as the final replica offers more ver- that experimental lesions are not always satility than the simple negative technique identical to human diseases. Our investiga- and for this reason is used quite extensively tion on the normal anatomy of the human in the study of paint surfaces. kidney has also convinced us that there are Tj'pical examples of the utilization of this important morphological differences between technique for the examination of paint sur- the kidney of man and that of animals. Thus faces are illustrated in Figures 8a, 8b, 9a and the diameter of the basement membrane of 9b. Figure 8a shows an enamel with a smooth the glomerular capillaries has been calculated o glossy surface while Figure 8b shows a flat at 800-1200 AU in different animals and o pamt with a roughened surface. Figure 9a at about 3500 AU in adult man (Bergstrand demonstrates an automative enamel before and Bucht, 1958). The "physiological" varia- weathering, while Figure 9b illustrates the tions due to differences in function of the enamel after outdoor exposure to test the re- tubular epithelial cells seem to be much sistance of the coating to weathering. greater in man than in animals which are inbred and kept under uniform environ- Acknowlegments.TYie author is indebted to Mr- ^^^^^^^ conditions (Bergstrand and Ericsson, A. Jbi. Jacobson for helpful discussions and to the -,^rr^^ m^ i ^- • i i-i National Lead Company for supporting this pro- ^^^^^^ ^^us observations on animal kidneys gram. should be carefully and critically evaluated when used in discussions of human path- REFERENCES ology. L Bradley, D. E., J. Appl. Phys., 5, 65-6 (1954). The word "pathological" must also be used 2. Bradley, D. E., J. Appl. Phys., 27, 1399-1412 with great care. The electron microscope re- ^^^^^^- veals cell organelles such as vacuoles, mem- 3. Lasko. W. R., Anal. Chem 29 784-786 (1957) . Cranes and mitochondria which are probably 4. Twiss, S. B., Teague, D. M., and Weeks, ,. ,, , • xi • . . n i W. L., Off. Dig 28 93 (1956) continually changing their structure, parallel 5. BoBALEK, E. G., Lebras, L. R., Powell, A. S., AND VON Fischer, W., Ind. Eng. Chem. 46, * When not stated otherwise, all works dealing 572 (1954). with human pathology and referred to in this paper have been performed on material obtained W. R. Lasko with this method. 206 PATHOLOGY to changes of function, also in the heakhy e^ aZ., Feldman), since the techniques used for subject. It is very difficult and many times demonstration of the localization of antibod- impossible to ascertain which are physiologi- ies does not permit an absolutely certain cal variations and pathological phenom- separation of the pedicels and the thin cover- ena. The latter word is used in'this paper to ing of endothelium from the basement mem- describe morphological changes which differ brane. Thus the reaction may take place in with reasonable certainty, quantitatively or any of these structures. Considering the qualitatively, from what is seen in healthy chemical nature of the basement membrane, subjects. as far as it is known, and the experimental evidence referred to above, the basement Glomerulonephritis membrane is the most probable locahzation, Experimental. Electron microscopic however, studies of glomerular changes in nephrotoxic Renal lesions in the acute generalized serum nephritis in rats and mice have been Schwartzman reaction were studied in rab- published by a number of authors (Simer, bits by Bohle et at. (1958) and in rats by 1954; Pielei aZ., 1955; Reid, 1956; Miller and Pappas et al. (1958). The animals were Bohle, 1957; Sakaguchi et al., 1957; Vernier treated twice with 24 hours interlude with an et al., 1958b; Miller et al., 1958; and Bohle intravenous injection of hpopolj^saccharides et al., 1959). The earliest lesions observed from Escherichia coli. Light microscopy one hour after the serum injection was showed abundant thrombi in the glomerular a diffuse thickening of the basement mem- capillary lumina eight hours after the last in- brane and a slight swelhng of the capillary jection. Electron microscopy demonstrated endothelial and epithelial cells. According to vacuolization of the endothelial cell cyto- Bohle et al., an increased swelling and vacuo- plasm and fibrils with a periodicity of 371.5 lization of the epithelial cells takes place dur- AU, probably representing fibrin in the ing the first 24 hours, accompanied by a thrombi. gradual and slow decrease of the diameters of Experimental glomerular lesions have also the basement membrane to normal values, been produced by Bencosme et al. (1959) in Between 24 and 72 hours the epithelial-cell rats treated with uranyl nitrate. They ob- changes subside and instead a marked swell- served a fusion of the foot processes and a ing of the endothelial cells takes place. These vacuolation of the epithelial cell cytoplasm cells increase in size to such an extent that with an accumulation of dark bodies cor- the capillary lumina may be completely ob- responding to hyaline droplets observed in literated. The latter changes have been ob- the light microscope. A foreign substance was serv'ed by all authors and are commonly precipitated between the capillary basement regarded as typical for experimental glom- membranes and large cells inside them, which erulonephritis. the authors regard as intercapillary cells. In Similar changes have been observed in rab- the newly formed masses collagen fibers were bits (Feldman, 1959) and in rats (Sitte, 1959) formed and the authors therefore consider after single intravenous injections of bovine that the intercapillary cells are not of endo- serum. thelial origin. It is possible that the initial swelling of the Human Glomerulonephritis and Lu- basement membrane is produced by the reac- pus Erythematosus. A case of subacute tion between antigen and antibody which, hemorrhagic glomerulonephritis was de- according to many authors, takes place in this scribed by Bergstrand and Bucht (1956). A site. This cannot be proved, however (Bohle number of large vacuoles were found in the 207 ELEC ri{o\ M I ( .H( >S( ;< >I'Y cndotholial cell cytoplasm. In the capillary nicnihraiie and accumulation.s of a newly luniina tiiere were numerous rounded bodies formed material were most striking in the bordered by a sinfj;le membrane and contain- subacute and chronic cases. The large masses ing small cell organelles, mostly vacuoles, of this material seen in two cases of Lupus The authors concluded that these bodies Erythematosus correspond to the wire loops were parts of the cell cytoplasm ejected from obsen-ed with the light microscope, the endothelial cell which was damaged by Cases of subacute and chronic glomeru- the inflammation. Later investigations have lonephritis with similar observations have shown that this may also be observed in also l)een reported by Spiro (1958, 1959). healthy humans and animals, but in the au- Judging from these reports glomerulone- thors' opinion the process was markedly in- phritis is primarily a disease of the vascular creased in the case of glomerulonephritis, endothelium in the glomeruli with a swelling There were no definite changes in the base- and vacuolization of the cytoplasm of these ment membrane or in the epithelial cells. cells followed by a precipitation of a foreign Farquhar da/. (1957a) have described two material in the basement membrane and cases of acute, two of subacute and three of probably also in the endothelial cell cyto- chronic glomerulonephritis. They observed plasm. The basement membrane is thick- in the acute stage of the disease a prolifera- ened and subseciuently the epithelial cells tion and swelling of the endothelial cells and are also destroyed along with the whole glo- later of the epithelial cells in the glomeruli, merulus, as is well known from light micro- The epithelial foot process organization ap- scopic studies. From the following sections it peared normal in contrast to the observa- will be seen that similar changes also occur tions in nephrosis (see below). The basement in other renal diseases, membranes were thickened and large masses of a basement membrane-like material was Morphological Changes Associated with accumulated probably inside endothelial the Nephrotic Syndrome cells. They concluded that this material was Experimental Nephrosis. The glomeru- produced by the endothelial cells. In the later lar lesions in aminonucleoside nephrosis in stages of the disease the capillaries were de- rats have been studied by Feldman and stroyed and replaced by masses of the base- Fisher (1959), Harkin and Recant (1959), ment membrane-like substance. Collagen and Vernier et al. (1958, 1959). Daily subcu- fibrils were not observed. taneous injections of small amounts of this Similar alterations were seen by the same substance evoked proteinuria in seven days, authors in three cases of Lupus Erythema- Simultaneously a swelling and vacuolization tosus. The thickening of the basement mem- of the capillary epithelial cells were observed, brane was more pronounced than in glomer- The foot process structure was dissolved and ulonephritis. Nodular accumulations of a the outer surface of the basement membrane basement membrane-like substance were also covered by a nearly continuous layer of epi- observed in the endothelial cells. Collagen thelial cell cytoplasm. After prolonged treat- fibrils were not found. These observations ment there was also a diffuse thickening of were confirmed by Putois (1959). the basement membrane and a precipitation In a subsequent paper by Vernier et al. of a foreign material on the endothelial side (1958a) 11 cases of glomerulonephritis and of the membrane (Vernier e^ ai.). There were four cases of Lupus Erythematosus in chil- no changes of the endothelial cells. In the dren were described. Proliferative changes in tubular epithelium a .swelling of the mito- the endothelium dominated in the acute chondria and a number of large vacuoles in cases whereas the thickening of the basement the cytoplasm were observed. 208 PATHOLOGY Similar changes were produced in rats with low). The basal folds of the cell membranes rabbit anti-kidney serum (Ehrich and Piel, had almost disappeared. Similar changes, but 1953) and in rabbits by intravenous injec- less prominent, were also present in the loops tion of saccharated iron oxide (Ellis, 1959). of Henle and the distal convoluted tubules. A nephrotic syndrome has occasionally Most authors consider the swelling of the developed in children during treatment with epithelial cells with loss of foot process or- trimethadione "Tridione". Electron micro- ganization to be the first observable lesion in scopic investigation of biopsy material has these diseases. Experimental observations shown glomerular lesions very similar to indicate that there is a causal relation be- those observed in aminonucleoside nephrosis tween proteinuria and epithelial cell changes. (Gribetz et al., 1959). Several authors have discussed the possibil- Lipide Nephrosis and Familial Nephro- ity that the primary lesion (swelling) is in the sis. A great number of reports on electron basement membrane with an increased per- microscopic studies of renal changes in pa- meability and leakage of proteins. The pres- tients with lipide nephrosis or familial ne- ence of proteins in considerable amounts in phrosis has been published recently (Ver- the filtrate would give rise to the changes of nier et al., 1956, 1958, 1959; Farquhar et al., the epithelial cells. (Vernier et al., 1958a, b; 1957a, b, c; Piel and Williams, 1957; Folli et Movat and McGregor, 1959a, b; Sitte, 1959.) al., 1957, 1958, 1959; Dalgaard, 1958a; Spiro, It has been shown by several authors that 1958, 1959a, b; Fiaschi et al., 1959; Putois, the epithelial cell changes may be reversible. 1959). The glomerular changes were very They disappear in cases which respond favor- similar to those observed in experimental ably to hormone treatment simultaneously nephrosis. They were first localized to the with the proteinuria. (Piel and WilUams, capillary epithelial cells. The organization of 1957; Dalgaard, 1958a, b; Folli, 1958; Pu- the foot processes was destroyed. The vari- tois, 1959; McDonald et al., 1959; Vernier ous epithelial cell processes on the outside of et al., 1958a, 1959b.) the capillary wall were fused so that the wall Membranous Glomerulonephritis. In was covered by a practically continuous layer cases with a long standing nephrotic syn- of epithelial cells. These cells also showed an drome and light microscope changes increased number of large vacuoles and can- described as chronic membranous glomeru- aliculi belonging to the endoplasmic reticu- lonephritis there is a more marked thicken- lum. These findings were obsei'ved in all ing of the glomerular capillary basement cases regardless of the clinical phase of the membranes (PoWak etal., 1958; Fiaschi e^ a/., disease. In more severe cases of long duration 1959). There is also a swelling and prolifer- irregular thickenings w^ere noted in the base- ation of the endothelial cells comparable to ment membranes alternating with empty- what is seen in acute glomerulonephritis. In appearing spaces. The endothelial cells were some cases ("mixed type") the epithelial also swollen but contrary to what was ob- cells may also be destroyed or increased in served in glomerulonephritis, this occurred number as in proliferative glomerulonephri- only in late stages of the disease. tis. In the epithelial cells of the proximal con- Experimentally Produced Renal Am- voluted tubules there was a focal loss of the yloidosis. Miller and Bohle (1956) have brush border. Many large vacuoles were ob- produced amyloidosis in mice with sodium- served in the apical part of the cells. The caseinate and ACTH. They were able to mitochondria were fragmented or swollen demonstrate a diffuse thickening and nodular with loss of the inner structure as observed protrusions of the glomerular capillary base- during increased protein resorption (see be- ment membranes. The nodules bulged into 209 ELECTRON MICROSCOPY Fig. 1. "Pure" Nephrosis. Very little changes in the endothelial cells (End). The epithelial cell is swollen with several vacuoles and partial destruction of the foot processes. the epithelial cells, i.e., toward the cavity of Bowman's capsule. Within the protrusions a fine spongy or felt-like structure with inter- woven filaments was observed. The thick- ness of the filaments was calculated as 30-40 o AU. There were no pores in the altered base- ment membranes. Earlier observations with the Ught microscope indicated that amyloid is deposited between the basement mem- brane and the endotheUum. The authors therefore concluded that the changes in the basement membranes w^ere probably not due to amyloid but to an increased leakage of protein through the capillary walls. Human Amyloidosis. A case of renal amyloidosis secondary to chronic osteomye- litis was reported by Geer et al. (1958). The glomerular capillary basement membranes were uniformly thickened with a diameter of approximately 0.5 micron. Large masses of material with the same structureless appear- ance as the basement membrane but with less density were observed in the basement membranes in the "basilar" portions of the 210 PATHOLOGY capillaries. They corresponded to the amy- loid masses which were observed in the light microscope. The endothelial cells were vacou- lated and the number of perforations of the cytoplasm (pores) was decreased. There was also a proliferation of the endothelial cells. The epithelial cells showed no damage other than a focal loss of organization of foot proc- esses similar to what has been observed in lipide nephrosis. There were small nodular thickenings of the basement membranes in the epithelial cells from the proximal and distal tubules. Similar material as described above was ob- served in the interstitial tissue, but it could not be demonstrated with certainty that this was amyloid. The brush border of the proxi- mal tubular epithelium was flattened with a decreased number of cytoplasmic processes and the basal infoldings of the cell mem- branes were also diminished in number. In the cytoplasm of the tubular epithelium were droplets of fat and large opaque granules. Further observations on renal amyloidosis in man have been published by Spiro (1959a). This author has treated his sections with PTA (phosphotungstic acid) for 24 hours after the routine preparation. In these sections dense and thickened areas were found in the glomerular capillary basement membranes alternating with apparently emptj^ spaces with a diameter of several hundred AU. The author considered the lat- ter to be large ''pores" in the basement membranes and responsible for the abundant proteinuria in his patients. We have not been able to confirm these observations in our laboratory; we do not think that it is possi- ble to exclude that they are artifacts due to the prolonged treatment with PTA. Spiro also obsei-ved destruction of the epitheUal cells and fine filaments in the amyloid masses, similar to those described by Miller and Bohle. In severe cases larger fibrils with a striation similar to that of collagen were observed. Bergstrand and Bucht (1959a) have stud- •jri--**- ^^'^■^^::. c, 4' 1, :j Fig. 2. Nephrotic Syndrome. "Mixed type" Nephrosis. Severe swelling and vacuolization of the endothelial cell (End). Slight focal thickening of the capillary basement membranes. Also marked swelling of the epithelial cell (Ep) with partial loss of foot process structure (Fp). I^J^,^ Fig. 3. Renal amyloidosis. Several capillary lumina (Cap) with a thin covering of endothelium. The amyloid masses are localized to the capillary basement membranes and to the space between them where remnants of destructed epithelial cells (Ep) are seen. 211 ELECTRON .^IICKDSCOPY Fig. 4. Renal amjdoidosis. Detail of amyloid precipitate inside an endothelial cell. A system of parallel fibrils are seen in the amyloid. ~;4 ■ ''^: -t Ci/?i Fig. 5. Diabetic glomerulosclerosis. Diffuse thickening of the capillary basement membrane (Bm) and focal accumulation of foreign material between the basement membrane and the endo- thelial cell (End). ied seven patients with renal amyloidosis. Clinical examination showed massive pro- teinuria, decreased glomerular filtration and a low filtration fraction. Electron microscopy of renal biopsies showed changes similar to those described by Geer et al. In early cases there was swelling and vacuolization of the endothelial cells. The basement membrane was thickened and folded but without visible structure. In more severe or prolonged cases a foreign substance accumulated between the endothelial cells and the basement mem- ])ranes. The epithelial cells were also altered and finally completely desquamated and re- placed by amyloid masses. Similar observa- tions have been described by Movat (1959), Putois (1959) and Meriel et al. (1960). In severe cases very fine filaments were seen in the newly formed substance in the basement membranes and above all in the desquamated epithelial cells. Diabetic Glomerulosclerosis. The glo- merular lesions in diabetic glomerulosclero- sis have been described by Irvine et al. (1956); Bergstrand and Bucht (1957, 1959); Dalgaard (1958b); Cossel et al. (1959); Farquhar et al. (1959) ; and Hartman (1959). The histologic alterations as viewed with the electron microscope were very similar to those of renal amyloidosis. The foreign "hyaline" substance was accumulated be- tween the endothelial cells and the glomeru- lar capillary basement membranes which were uniformly thickened. These changes correspond to the "diffuse" type of glomeru- losclerosis as observed by light microscopy. The hyaline masses also bulged into the vacuolated endothelial cell cytoplasm. The endothelial cells were partly destroyed by the large masses which obliterated the capillary lumina. These changes correspond to the "nodular" lesions, observed through the light microscope. In the hyaline material close to the endothelial cell membrane very fine filaments could be demonstrated. There was little damage to the epithelial cells in early stages of the disease, and these changes, which included a loss of foot process organi- zation, were regarded by the authors as sec- ondary to the lesions in the basement mem- brane. Conclusions. The diseases which are as- sociated with the nephrotic syndrome are by many authors regarded as metabolic disor- 212 PATHOLOGY ders of the connective tissue ground sub- stance and related structures. In the diseases described here, the first and most important changes are considered to be located in the capillary basement membranes. According to Gersh and Catchpole (1949) the latter are condensations of the ground substance. The changes of the basement membranes probably include a precipitation of a foreign protein-containing substance, such as amy- loid, and perhaps also a metabolic change, such as a depolymerization, of the muco- polysaccharides pre-existing in the basement membrane. It may be presumed that a change in permeability is evoked simultane- ously with the thickening of the basement membrane which may explain the apparently controversial observations of a decrease in filtration fraction and proteinuria in the same patient (Bergstrand and Bucht, 19o9b; Sitte, 1959). Important differences exist, however, between the diseases associated wdth a ne- phrotic syndrome. In lipide nephrosis and fa- milial nephrosis the epithelial cells are se- verely changed whereas little or no damage is seen in the endothelium. In renal amyloi- dosis both kinds of cells are damaged and in glomerulosis the lesions are mainly located in the endothelial cells. Furthermore, in the cases of lipide and familial nephrosis, the glomerular fi.ltration seems to have been nor- mal. In amyloidosis and glomerulosis a de- crease in filtration fraction was observed in most cases. It is interesting to note that in glomerulo- nephritis the main changes are also located in the capillary basement membranes and the endothelial cells. No definite conclusions concerning eventual relationships between the latter disease and those described above should be drawn from this fact at the pres- ent moment, however. Shock Kidney Electron microscopic investigations of the renal changes in acute anuria (post-opera- tive shock) have been published by Dalgaard 05// Fig. 6. Diabetic glomerulosclerosis. Capillary- basement membrane without apparent structure but about three times thicker than normal and with a thin band (B) of foreign material close to the endothelial cell membrane (End). (1958a; 1959) and Putois (1959). There were no glomerular lesions but severe damage to the epithelial cells in the proximal and distal convoluted tubules. The brush border w^as partially or completely destroyed and in the cytoplasm large vacuoles, severely damaged mitochondria and "macrobodies" were ob- served. The basal infoldings of the cell mem- branes w^ere decreased in height or com- pletely absent. In many cells these changes had developed into complete necrosis with destruction of most cell organelles. No re- generation phenomena were observed. Experimental Tubular Changes Engfelt et al. (1957) have studied the in- fluence of parathyroid hormone on the kid- neys in rats. Electron microscopy, which has been reported in more detail by Rhodin (1958), showed an accumulation of very 213 EIKCTKON MICKOSCOPY dense l^odies in the basal parts of the epi- droplets are formed in the apical parts of the thelial cells of the proximal convoluted tu- cells, where there are no mitochondria, prob- bulos when very small amounts of the hor- ably from the vacuoles or from enzyme-pro- mone were given. Wlien larger doses were ducing granules (cytosomes). An occasional used, extremely dense small granules were fusion between a mitochondrion and a drop- foiuid in the basement membranes of the let may take place if they are located close tubular epithelium. The authors presumed to each other. A primary accumulation of that these granules were precipitations of protein inside the mitochondria is not prob- calcium salts. In the apical parts of the epi- able. thelial cells of the proximal convoluted tu- Policard et at. (1957) produced lesions in bules large bodies were formed, which had the epithelial cells of the proximal convo- a xery low density and were PAS -positive, luted tubules in rats by injecting a colloidal The.y increased in size and occasionally filled solution of 5 per cent sodium silicate in the the larger part of the cells. In the basal parts peritoneum. The brush border was partly of the cells the number of mitochondria was destroyed and in cells with more severe dam- decreased, probably because several of them age the apical parts of the cells were cut off had developed into the above-mentioned and ejected into the tubular lumen. The electron-dense bodies. The tubular lumina basal infoldings of the cell membrane disap- were filled by the same PAS-positive sub- peared. The mitochondria were large and stance as is seen in the tubular epithelium, pale with vacuoles or a severely changed The formation of concrements probably inner structure. started by the precipitation of calcium salts Rouiller and Modtjabai (1958) perfused a in this substance. Electron microscopy of the sodium-free 5 per cent solution of glucose concrements showed a structure quite similar through the peritoneal cavity of rabbits. The to that of apatite crystals. Similar observa- cells of the proximal convoluted tubules tions have been made by this author in showed a marked increase in size and were biops3^ specimens from patients with hyper- very pale in the light microscope, as is ob- parathyroidism and renal calcifications. served in human subjects which have been Rhodin (1954) first studied the effect of treated with large intravenous infusions of, intraperitoneal injections of egg white on the for instance, dextran. Electron microscopy proximal tubular epithelium in mice. There showed an extremely low contrast in the cy- were no changes in the appearance of the toplasm of some cells of the proximal convo- brush border. In the apical parts of the cells luted tubules and very large and numerous there was increased vacuolization. Large vacuoles inside or between the cells. The api- opaque bodies were found in the cytoplasm cal part of the cells were partly destroyed corresponding to the hyaline droplets seen and ejected into the tubular lumen. The by light microscopy. Inside these bodies were mitochondria w^ere few^ and small. According remnants of double membranes. The author to the authors, these changes were due to an concluded that these bodies developed from increased content of water in the tubular the mitochondria during resorption of pro- cells. In some mitochondria were accumula- tein as previously described by Oliver (1948) tions of small dark granules, which were re- and others. garded by authors as evoked by the resorp- These observations were confirmed by tion of glucose. Similar changes were induced Miller and Sitte (1955) and by Gansler and in mice by Yolac (1959) with sucrose. Rouiller (1956). According to Miller (1959), The tubular lesions in potassium-depleted the association between the hyaline droplets rats were studied by Tauxe et al. (1957), and and the mitochondria is only secondary. The Muehrcke and Bonting (1958). There were 214 PATHOLOGY no changes in the proximal tubules. In the Electron microscopic investigations of the collecting ducts the mitochondria were en- excretion of radioactive mercury compounds larged or destroyed. Ten days after potas- through the kidneys of rats have been made slum had been administered 4^he mitochon- by Bergstrand et al. (1959). Mercury was dria had regained a normal appearance. An demonstrated in the proximal tubular cells as increased number of macrobodies and lipoid accumulations of very small and dark parti- droplets was observ^ed in the tubular epi- cles inside large bodies similar to the "sider- thelium of animals with cholin deficiency osomes" described by Richter. The presence (Ash worth, 1959a, b). of mercury in the cytoplasm could not be The effect on the tubular epithelium of demonstrated by the electron microscope, heavy metals has been studied by several After homogenization of the kidneys and authors, Dempsey and Wislocki (1955) stud- centrifugation at high speed, a very high ac- ied the site of accumulation of silver in the tivity could be demonstrated in the micro- tissues by giving a solution of silver nitrate some fraction which contained only the in the drinking water to mice, rats and guinea smallest cell organelles, such as RNA-gran- pigs for 6 to 12 months. The silver was ules, but no traces of mitochondria or "sider- mainly located in the basement membranes osomes." Thus it is very probable that of the glomerular capillaries and the tubular mercury was also present in the cytoplasm as epithelium. In the epithelial cells of the prox- was demonstrated with iron by Richter. The imal tubules silver could be demonstrated amounts of mercury were low and there was inside mitochondria which had retained their no destruction of the tubular cells, original structure and were easily identifi- Bencosme et al. (1959) have studied the able. The metal appeared as very dense and effect of uranyl nitrate on the kidneys of o small particles w^ith a size of about 20-30 AU rats with the electron microscope. Extensive or as larger aggregates with a diameter of damage to the epithelial cells was observed one or several microns. Similar observations in the proximal tubules. The intercellular on rats have been made by van Breemen et al. spaces were widened, forming large cisternes (1956) and by Olcott and Richter (1958). probably containing resorbed tubular fluid. Richter (1957) gave repeated injections of No deposits of the metal were observed in the hemoglobin intraperitoneally to rats. Hemo- epithelial cells. The author concludes that siderin could be demonstrated in the proxi- the passage of fluid through the tubular walls mal tubular epithelium. Electron microscopy takes place mainly through the intercellular revealed large opaque bodies in the cells spaces, containing numerous very dense particles with a mean diameter of 55 AU. Similar par- Conclusions on the Morphological Basis tides could be demonstrated in the cyto- "^ Glomerular Filtration plasm without apparent connections to any Electron microscopy of normal glomeruli cell organelles. The author named the large from man and animals has failed to reveal bodies "siderosomes" and considered them the structures in the capillary walls which to be derivatives of mitochondria. The small are necessary for the filtration process. Sev- particles corresponded in size to the iron eral attempts have been made to elucidate miscelles of purified ferritin, and the author the problem through a comparison of clinical concluded that ferritin is probably a com- data from patients with impaired renal ponent of hemosiderin. Similar iron-contain- function to the corresponding morphological ing bodies have been described under the changes of the capillary wall as revealed by name of cytosomes in other organs and in the electron microscope, macrophages (Lindner, 1958; Miller, 1959). Hall (1957) has concluded from studies on 215 ELECTRON MICKOSCOI'Y normal kidneys that the sHt-pores between the foot processes are the filtration pores postulated by Pappenhcimer (1955). It could therefore be expected that in cases where the foot process organization was impaired, filtra- tion should also be low. As pointed out by Farquhar et al. (1957a, c), Vernier et al. (1958a) and Rhodin (1959), this is not the case in familial nephrosis or lipide nephrosis. On the contrary, a decreased renal function was observed in cases with subacute glomer- ulonephritis where the foot processes were intact. The problem has been discussed by Berg- strand (1959) on the basis of investigations in amyloidosis and diabetic glomerulosis. According to this author, changes in filtration rate cannot with certainty be correlated to changes of the capillary wall. A severe reduc- tion in filtration fraction, on the other hand, is most probably a sign of an increased re- sistance to filtration through the capillary walls. This was observ^ed in most of his cases and he concluded from the electron micro- scopical observations that the most probable cause of the decrease in filtration fraction was a thickening of the basement membrane, perhaps associated with a change of its ultrastructure, which could not be demon- strated with certainty with the present tech- niques. Bergstrand presumes that the filtra- tion pores in the normal capillary wall are localized in the basement membrane. The same opinion has been expressed by Sitte (1959). Other attempts to solve the problem have been made by studying the passage of very small particles through the glomerular capil- lary wall. Farquhar and Palade used ferritin molecules, Latta and Maunsbach thorotrast, and Vernier silver-labeled protein molecules. These studies were performed both in normal and nephrotic animals. The authors observed a very rapid passage of particles with a diam- eter less than 75 AU, whereas larger particles, abo\'e all thorotrast, were retained in the bloodstream. No changes of structure could be observed in the basement membrane when particles were passing through. No definite conclusions could thus be drawn about the morphological Imsis of glomerular ultrafiltra- tion. REFERENCES 1. AsHWOBTH, C. T., "Ultrastructural Aspects of Renal Patho-Physiology," Tex. Rep. Biol. Med., 17, 60-72 (1959a). 2. AsHWORTH, C. T. AND Grollman, A., "Elec- tron Microscopy in Experimental Hyperten- sion," A.M.A. Arch. Path., 68, 148-153 (1959b). 3. Bencosme, S. a., Stone, R. S., Latta, H., AND Madden, S. C. "Acute Reactions with Collagen Production in Renal Glomeruli of Rats as Studied Electron Microscopically," y. UUrastr. Res., 3: 171-185 (1959). 4. Bergstrand, A., "The Morphological Basis for the Filtration Process in the Glomeruli," Acta Chir. Suppl. 245, 336-342 (1959). 5. Bergstrand, A. and Bucht, H., "Electron Microscope Investigation on Biopsy Mate- rial frum Patients with Renal Diseases: A Case of Subacute Glomerulonephritis," Proc. Stockholm Conference on Electron Microscopy, 256-258, 1956. 6. Bergstrand, A. and Bucht, H., "Electron Microscopic Investigations on the Glomeru- lar Lesions in Diabetes Mellitus (Diabetic Glomerulosclerosis)," Lab. Invest., 6, 293- 300 (1957). 7. Bergstrand, A. and Bucht, H., "Anatomy of the Glomerulus as Observed in Biopsy Material from Young and Healthy Human Subjects," Z. Zellforsch., 48, 51-73 (1958). 8. Bergstrand, A. and Bucht, H., "Electron Microscopic Observations on Renal Amyloi- dosis," (1959a) (to be published). 9. Bergstrand, A. and Bucht, H., "The Glo- merular Lesions of Diabetes Mellitus and Their Electron-Microscope Appearances," J. Path. Bad., 77, 231-242 (1959b). 10. Bergstrand, A., Friberg, L., Mendel, L., and Odeblad, E., "Localization of Subcu- taneously Administered Radioactive Mer- cury in the Rat Kidney," (Abstract) y. Ultrastruct. Res., 3: 238, 1959. 11. Bergstrand, A. and Ericsson, J., "Electron Microscopic Investigation on the Anatomy of the Kidney Tubules in Young and Healthy Human Subjects," (1959) (To be published). 12. Bohle, A., Sitte, H., and Miller, F., "Elek- 216 PATHOLOGY tronenmikroskopische Untersuchungen am Glomerulum des Kaninchens beim gene- ralisierten Schwartzmann - Phanomen , ' ' Verh. deut. Ges. Path., 41, 326-332 (1958). 13. BoHLE, A., Miller, F., Sitte, H., .a.nd Yolac, A., "Fruhveranderiuigen bei der Masugi- Nephritis der Ratte. Immunopathologie. I. Int. SA'mposium," Basel, Seeligsberg, 1958 70-81 (1959). 14. V. Breemen, V. L., Reger, J. F., and Cooper, W. C, "Observations on the Basement Membranes in Rat Kidney," /. Biophys. Biochem. CijtoL, 2, suppl. 283-286 (1956). 15. COSSEL, L., LiSEWSKI, G., AND MOHNIKE, G., "Elektronenmikroskopische und klinische Untersuchungen bei diabetischer Glomeru- losklerose," Klin. Wschr., 37, 1005-1018 (1959). 16. Dalgaard, C. Z., "Electron Microscopic Investigation of Material from Renal Biopsj- in Two Patients with so-called Genuine Lipoid Nephrosis," Ugeskr. Laeg., 120, 1358-1363 (1958a). 17. Dalgaard, C. Z., "Electron Microscopic In- vestigations on Renal Biopsies from Pa- tients with Renal Diseases. Verh. IV Int. Kongress Elektronenmikroskopie," Berlin, (1958b) In print. 18. Dalgaard, C. Z. and Pedersen, K. J., "Renal Tubular Degeneration. Electron Microscopy in Ischaemic Anuria," Lancet 11, 484-488 (1959). 19. Dempsey, E. W. and Wislocki, G. E., "Use of Silver Nitrate as a Vital Stain, and its Distribution in Several Mammalian Tissues as Studied with the Electron Microscope," /. Biophys. Biochem. Cytol., 1, 111-118 (1955). 20. Ehrich, W. and Piel, C. F., "Morphologic Differentiation of Nephritis in the Rat and the Therapeutic Effects of Anticoagulants and Proteolytic Enzymes," Proc. 5th Ann. Conference Nephrotic Syndrome, p. 117, Nat. Nephrosis Foundation Inc., New York (1953). 21. Ellis, J. T., "Glomerular Lesions in Rabbits with Experimentally Produced Proteinuria as Disclosed bj- Electron Microscopy," (Ab- stract.) A7n. J. Path., 34, 559-560 (1959) 22. Engfelt, B., Gardell, S., Hellstrom, J., IVEMARK, B., RhODIN, J., AND StRANDH, J., "Effect of Experimental Hyperparathj-- roidism on Renal Function and Structure," Acta Endocrinol., 29, 15-26 (1958). 23. Farquhar, M. G., Vernier, R. L., and Good, R. A., "An Electron Microscope Stud}' of the Glomerulus in Nephrosis, Glomerulone- phritis, and Lupus Erythematosus," /. Exper. Med., 106, 649-660 (1957a). 24. Farquhar, M. G., Vernier, R. L., and Good, R. A., "Studies on Familial Nephrosis II: Glomerular Changes Observed with the Electron Microscope," Am. J. Path., 33, 791-805 (1957b). 25. Farquhar, M. G., Vernier, R. L., and Good, R. A., "Application of Electron Microscopy in Pathology: Study of Renal Biopsy Tis- sues," Schweiz. Med. Wschr., 87, 501-510 (1957c). 26. Farquhar, M. G., Hopper Jr., J., and Moon, H. D., "Diabetic Glomerulosclerosis: Elec- tron and Light Microscopic Studies," Aju. J. Path., 35, 721-754 (1959). 26a. Farquhar, M. G., and Palade, G. E., "Seg- regation of Ferritin in Glomerular Protein Absorption Droplets," J. Biophys. Biochem. C?/ioL, 7: 297-304 (1960). 27. Feldman, J. D., "Electron Microscopy of Serum Sickness Nephritis," J. Exptl. Med., 108, 957-962 (1959). 28. Feldman, J. D. and Fisher, E. R., "R nal Lesions in Aminonucleoside Nephrosis as Revealed bj' Electron ^licroscopy," Lab. Invest., 8,371-385 (1959). 29. FiASCHi, E., Amdres, G., Giocomelli, F., and Naccarato, R., "Renal Histopathology in the Para-Nephritic Nephrotic Syn- drome," Sc. Med. Ital., 7, 639-742 (1959). 30. FoLLi, G., Pollak, V. E., Reid, R. T. W., PiRANi, C. L., AND Kark, R. M., Electron Microscopic Studies of Renal Biopsies Taken from Nephrotic Patients before and after Diuresis. (Abstract.) J. Lab. Clin. Med. 50, 813 (1957). 31. FoLLi, G., Pollak, V. E., Reid, R. T. W., PiRANi, C. L., AND Kark, R. M., "Electron Microscopic Studies of Reversible Glomeru- lar Lesions in the Adult Nephrotic Syn- drome," Ann. Int. Med., 49, 775-795 (1958). 32. FoLLi, G., "Studi di microscopia elettronica nelle nefrosi," Atti. Soc. Lombarda Sci. Med. Biol., 13,200-207 (1958). 33. FoLLi, G. AND Onida, L., "Biopsj' and Elec- tron Microscopy of the Kidnej'," Sc. Med. Ital., 8, 19-58 (1959). 34. Gansler, H. and Rouiller, C, "Modifica- tions physiologiques et pathologiques du chondriome," Schweiz. Z. Path. Bad., 19, 217-243 (1956). 35. Geer, J. CStrong, J.P.,McGiLL Jr., H. C, and Muslow, I., "Electron Microscopic Observations on the Localization of Amy- loid in the Kidney in Secondary Amj'loi- dosis," Lab. Invest., 7, 554-565 (1958). 217 ELECTHO-N >l l( HOSCOl'Y 36. Gersh, I., AND Catchpole, H. R., "Organiza- tion of Ground Substance and Basement IMpmbrane and Its Significance in Tissue Injury, Disease and Growth," Am. J. Anal., 85, 457-507 (1949). 37. Gribetz, a., Mantner, W., and Kohn, J. L., "Trimethadione (Tridione) Nephrosis: Re- nal Biopsjf and Electron Microscopic Stud- ies," (Abstract.) A.M.A. J. Dis. Child., 98, 76, (1959) 38. Harkin, J. C. AND Recant, L., "Earliest Lesion in Aminonucleoside Nephrosis; an Electronmicroscopic Study,"(Abstract.) Am. J. Path., 34, 559 (1959) 39. Hartman, J. F., "Elektronenmikroskopie der menschlichen Niere bei Diabetes Mellitus im Vergleich mit Nephritis und Nephrose," Zentrbl. Allgemeine Pathologic und Path- ologische Anatomie, 98, 313 (1958) Abstr. 40. Hall, V., "The Protoplasmic Basis of Glo- merular Ultrafiltration," A7n. Heart J., 54, 1-9 (1957). 41. Irvine, E., Rinehart, J. F., Mortimore, G. E., and Hopper Jr., J., "The Ultrastruc- ture of the Renal Glomerulus in Intercapil- lary Glomerulosclerosis," Am. J. Path., 32, 647-648 (1956) Abstract. 42. IvERSEN, P., AND Brun, C., "Aspiration Biopsy of the Kidney," Am. J. Med., 11, 324-330 (1951). 43. Kark, R. M. and Muerhcke, R. C., "Biopsy of Kidney in Prone Position," Lancet, 1, 1047-1049 (1954). 43a. Latta, H., and Maunsbach, A., Rep. I Int. Congress Nephrol. Evian 1-4 Sept. (1960). In print. 44. Lindner, E., "Der elektronenmikrosko- pischen nachweis von Eisen im Gewebe, Ergebn. Allgemeine Pathologic und Patholog- ische Anatomie, 46-91 (1958). 45. McDonald, M. K., Lambie, A. T., and Rob- son, J. S., "Resolution of Glomerular Le- sions in the Nephrotic Syndrome treated with Cortisone: Electron Microscopic Stud- ies in an Adult Case." 45a. Meriel, P. Moreau, G., Sue, Y. M., and PuTois, Y. Rep. I Int. Congr. of Nephrol. Evian, 1960. 46. ^IiLLER, F., "Orthologie und Pathologie der Zelle im elektronenmikroskopischen Bild," Verh. deut. Ges. Path., 42, 261-332 (1959). 47. Miller, F. and SiTTE, H., "Elektronenmikros- kopische Untersuchungen an Mausenieren nach intraperitonealen Eiweissgaben," Verh. deut. Ges. Path., 39, 183-190 (1955). 48. Miller, F. and Bohle, A., "Comparative Investigation bj' Means of the Light and Electron Microscope of the Glomerular Capillary Basement Membrane in Experi- mental Renal Amyloidosis of the Mouse, Klin. Wschr., 34, 1204-1210 (1956a). 49. Miller, F. and Bohle, A., "Electron Micros- copy of the Glomerular Basement Mem- brane in Experimental Amyloidosis in the Mouse," Proc. Stockholm Conference Elec- tron Microscopy, 254-256 (1956b). 50. Miller, F. and Bohle, A., "Elektronen- mikroskopische Untersuchungen am Glome- rulum bei der Masugi-Nephritis der Ratte," Virchows Arch., 330, 483-497 (1957). 51. Miller, F., Sitte, H., and Bohle, A., "Elekt- ronenmikroskopische Befunde am Glome- rulum bei der Masugi-Nephritis der Ratte," Ver. deut. Ges. Path., 41, 333-335 (1958). 52. MovAT, H. Z., "The Fine Structure of the Glomerulus in Amyloidosis," (Abstract.) Am. J. Path., 35, 708 (1959). 53. MovAT, H. Z. and McGregor, D. D., "Fine Structure of the Glomerulus in Membra- nous Glomerulonephritis (Lipide Nephrosis) in Adults," A771. J. Clin. Path., 32, 109-127 (1959a). 54. Movat, H. Z. and McGregor, D. D., "Fine Structure of the Glomerulus in Membranous Glomerulonephritis (Lipoide Nephrosis)," (Abstract.) Am. J. Path., 35, 670, (1959b) 55. Muehrcke, R. C. and Bonting, S. L., "Elec- tronmicroscopic and Ultrabiochemical Stud- ies of the Potassium-depleted Kidney," Clin. Res., 6, 413-414 (1958). 56. Olcott, C. T. and Richter, G. W., "Experi- mental Argj'rosis. VI. Electronmicroscopic Study of Ingested Silver in the Kidnej- of +hp Rat." Lab. Invest., 7, 103-109 (1958). 57. Oliver, J., "Structure of the Metabolic Proc- ess in the Nephron," J. Mt. Sinai Hosp., 15, 175-222 (1948). 58. Pappas, G. E., Ross, M. H., and Lewis, T., "Studies on the Generalized Schwartzmann Reaction. VIII. The Appearance by Electron Microscopy of Intravascular Fibrinoid in the Glomerular Capillaries during the Re- action," J. Exptl. Med., 107, 333-340 (1958). 59. Pappenheimer, J. R., "tjber die Permeabili- tat der Glomerulummembranen in der Niere," Klin. Wchschr., 33, 362-365 (1955). 60. Piel, C. F., Dong, L., Modern, F. W. S., Goodman, J. R., and Moore, R., "The Glomerulus in Experimental Renal Disease in Rats as Observed by Light and Electron Microscopy," /. Exper. Med., 102, 573-580 (1955). 218 PATHOLOGY 61. PiEL, C. F. AND Williams, G. F., "Long Con- tinued Adrenal Hormone Therapj' in Child- hood Nephrosis," J. Am. Med. Women's Assoc, 12, 273-279 (1957). 62. PoLiCARD, A., Collet, A., and Pregermain, S., "Sur quelques points de la cytopatho- logie des nephrites toxiques," Presse Med., 65, 168^1688 (1957). 63. Pollak, V. E., PiRANi, C. L., Muehrcke, R. C, AND Kark, R. M., "Asymptomatic Persistent Proteinuria: Studies by Renal Biopsy," Guy's Hosp. Rep., 107, 353-372 (1958). 64. PuTOis, J., "Structure et ultrastructure re- nale. Aspects normaux et pathologiques," Librairie Arnette, Paris, 1959. 65. Reger, J. F., Neustein, H. B., and Hutt, M., "Observations on the Fine Structure of Nephron Units in Various Diseases," (Ab- stract.) Anat. Rec, 130, 460 (1958) 66. Reid, R. T. W., "Electron Microscopy of Glomeruli in Nephrotoxic Serum Nephri- tis," Australian J. Exp. Biol. Med. Sci., 34, 14.3-150 (1956). 67. Rhodin, J., "Correlation of Ultrastructural Organization and Function in Normal and Experimentally Changed Proximal Con- voluted Tubule Cells of the Mouse Kidney," Thesis, Stockholm, 1954. 68. Rhodin, J., "Ergebnisse der elektronenmikro- skopischen Erforschung von Struktur und Funktion der Zelle," Verh. deut. Ges. Path., 41,274-284 (1958). 69. Rhodin, J., "Microscopic electronique du rein," Bruxelles-Med., 39, 409-426 (1959). 70. RiCHTER, G. W., "A Study of Hemosiderosis with the Aid of Electron Microscopy," /. Exptl. Med., 106, 203-218 (1957). 71. RouiLLER, G. AND MoDTJABAi, A., "La ucph- rose experimentale du lapin. Comparaison entre la microscopic optique et electronique. I. Les modifications des cellules a bordure striee," Ann. Anat. Path., 3, 223-250 (1958). 72. Sakaguchi, H., Suzuki, Y., and Yamaguchi, T., "Electron Microscopic Study of Masugi Nephritis," Acta Path. Japan, 7, 53-66 (1957). 73. SiMER, p. H., "Electron Microscopic Studies of the Glomerulus in Nephrotic Mice of the NH Strain," Anat. Rec, 118, 409 (1954). 74. SiTTE, H., "Changes in the Glomeruli of Rat Kidneys after Administration of Foreign Proteins and a Hypothetical Explanation of the Glomerular Filtration," Verh. deut. Ges. Path., 42, 225-32 (1959). 75. Si'iRO, D., "Electron Microscopic Studies on Human Renal Biopsies. The Structural Basis of Proteinuria. Verh. IV.," Int. Kongress Elektronenmikroskopie, Berlin, 1958. In print. 76. Spiro, D., "The Structural Basis of Pro- teinuria," Am. J. Path., 35, 47-74 (1959a). 77. Spiro, D., "Further Studies on the Ultra Structure of the Kidney in the Nephrotic Syndrome," (Abstract.) Am. J. Path., 35, 716 (1959b). 78. Tauxe, W. N., Wakim, K. G., and Baggen- STOss, A. H., "Renal lesions in experimental deficiency of potassium," Atner. J. Clin. Path., 28,221-232 (1957). 79. Vernier, R. L., Farquhar, M. G., Brunson, J. G., AND Good, R. A., "Studies on Familial Nephrosis," (Abstract.) J. Clin. Invest., 35, 741 (1956) 79a. Vernier, R. L., Rept. I Int. Congress Nephrol. Evian 1-4 Sept. (1960). In print. 80. Vernier, R. L., Farquhar, M. G., Brunson, J. G., and Good, R. A., "Chronic Renal Disease in Children," A.M. A. J. Dis Chil- dren, 96, 306-343 (1958a). 81. Vernier, R., Papermaster, B., Arhelger, R., AND Mecklenburg, P., "Electron Microscopy of Experimental Renal Dis- ease," A.M.A. J. Dis. Child. 96, 597-598 (1958b). 82. Vernier, R. L., Papermaster, B. W., and Good, R. A., "Light and Electron Micro- scopic Pathology of Experimental Amino- nucleoside Nephrosis: Relations to Ne- phrosis in Man," Fed. Proc, 17, 463 (1958c) Abstract. 83. Vernier, R. L. and Good, R. A., "Renal Biopsy in Children," Pediatrics, 22, 1033- 1034 (1958d). 84. Vernier, R. L., Papermaster, B. W., and Good, R. A., "Aminonucleoside Nephrosis I. Electron Microscopic Studj- of the Renal Lesions in Rats," /. Exptl. Med., 109, 115- 126 (1959a). 85. Vernier, R. L., Worthen, H. G., and Good, R. A., "Electron Microscope in Medical Research," Journ. -Lancet, 79, 423-27 (1959). 86. YoLAC, A. B., "Electron Microscopic Studies Concerning the Morphology of the Epithe- lium of the Proximal Convoluted Tubules of Mouse Kidneys after Injection of Sucrose Solution," Verh. deut. Ges. Path., 42, 235-40 (1959). Anders Bergstrand 219 EI.ECTRON MICROSCOPY PLASTICS. Srr GENERAL MICROSCOPY, p. 390. PULP AND PAPER. See GENERAL MICROS- COPY, p. 394. REFLECTION I In the early days of electron microscopy prior to the de\elopment of replica methods, most specimens were so thick as to be elec- tron-opaque and only their silhouettes could be examined. To enable direct surface studies to be made, Ruska (1) in 1933 attempted to image electrons scattered from a metal sur- face. The specimen was irradiated with an electron beam which was at 90° to the axis of the imaging system, but the resolution was limited in this first experiment to about 20 to 30 microns by the chromatic aberra- tion of the lens and the very large bandwidth of the energy spectrum of the scattered elec- trons. Further work by Ruska and Miiller, (2) using glancing angles of incidence, and viewing at 90°, yielded a resolution of 5,000 A. Von Borries (3) in 1940 suggested the use of glancing angles of both illumination and viewing, thus reducing the energy loss in the scattered beam and giving a resolution of o about 250 A. Such low angles (the total deviation of the beam being about 8°) make for extremely difficult interpretation of the Fig. 1. Schematic diagram of reflection elec- tron microscope. image and moreover only the smoothest surfaces can be examined. It was for these reasons that with the advent of repUca meth- ods, interest in reflection electron microscopy diminished. There was a revival of interest in the years 1951 to 1953 when it was realized that in certain circumstances the reflection method of Von Borries could have distinct advan- tages over the now well established replica method and this revival took place almost simultaneously in several schools. Kushnir, Biberman and Levkin (4), Cosslett and Jones (5), and Fert and Saporte (6) worked with microscopes designed and built for reflection work while Menter (7), and Haine and Hirst (8) made adaptations for the Metropolitan Vickers microscope. The characteristics of the image in this type of work are very different from those pertaining to other forms of microscopy and can be considered under the following head- ings: Viewpoint. The highly oblique viewpoint is a most unusual one; Von Borries stated that the image has the appearance of a road illuminated by the headlights of an ap- proaching car. The most important feature of this obhque viewpoint is the marked fore- shortening effect in the image. Distances along the line of sight (direction y, Fig. 1) are foreshortened with a viewing angle of for example 5° by a factor of 12:1. Thus a circle in the plane of the specimen appears as an ellipse of this eccentricity; this led Emerton (9) to suggest that the scale on reflection electron micrographs should be indicated by ellipses, the diameter of which could be quoted. It also follows that linear features on the specimen appear to be ori- ented more nearly perpendicular to the line of sight (direction x). A line at an angle of 45° to the line of sight appears in a micro- graph (foreshortened 12: 1) to be at an angle of 85°. This viewpoint does nevertheless have certain advantages. The form of the surface 220 REFLECTION 1 in the direction perpendicular to the plane of the specimen (direction z) is often brought out with a clarity which is not possible by other microscopical methods and the reflec- tion image is very sensitive to changes in this direction. Contrast. This is high and arises largely from the difference between the intensity from illuminated areas and from areas "in shadow" behind raised features. In addition, there is a tone range within the illuminated areas due to the local variation in angle be- tween the surface and the axis of the imaging system. The contrast is thus extremely easy for the eye to interpret and this accounts for the very pleasing appearance of reflection micrographs. Resolution. Electrons are scattered in all directions, evenly illuminating the objective aperture the size of which therefore governs directly the confusion in the image due to aberrations. Of these, chromatic aberration is the overriding effect due to the large energy spread in the scattered beam. The radius of the disc of confusion is given by d = aCc8v/v (1) where a is the semi-angular aperture of the accepted beam Cc is the chromatic aberration con- stant of the objective lens f)v is the half width of the energy spread of accepted electrons V is the accelerating voltage d can be reduced by reducing Cc , al- though this appears to be limited by the present design of lenses to approximately the focal length, or by reducing the objective aperture. There is a limitation to the latter course due to the increasing effect of diffrac- tion, but this is in fact never a problem be- cause a reduction in a decreases the intensity of the image. In practice the lowest value of a is chosen which gives an image just suffi- ciently bright to be focused and recorded at the necessary magnification. Fig. 2. Cleavage fracture on mica. (After Menter).© , -H © 2 = 14° Magnification 5,100 X. Typical values are as follows: a = 1.5 X 10-^ Cc =1.0 cm Sv/v = 0.0015 o This gives a resolution of about 200 A, and this cannot be appreciably improved on. Equation (1) gives the resolution on the axis of the system. Abaxial points generally have an inferior resolution because of the field chromatic aberrations. Page (10) has shown the importance of these and has given a practicable method for their correction in a triple lens electron microscope. Depth of fieltl. This is given by D = 2d /a {=2CcSv/v) (2) The typical values above give a depth of field of about 30 microns and this ensures that any part of the specimen is entirely in 221 ELECTKON ^IICHOSCOPY '^««H*.' Fig. 3. Scratch on brass surface (after Page). © i + ©2 = 26". Magnification 1,200X. The advantages and disadvantages of the Von Borries reflection method may be sum- marized as follows. Adva7itages (1) Direct observation of the specimen (2) Large depth of field (3) Contrast easily interpreted (4) Information available in the dimen- sion perpendicular to the specimen plane. Disadvantages (1) Extremely foreshortened image (2) Unsuit ability for rough surfaces (3) Resolution limited to 200 A (4) Damage to some specimens by high intensity electron beam In spite of the disadvantages, the reflec- tion method has been of value in the fields of metallurgj' and crystallography, in the :i.v<^'". Fig. 4. Etched germanium surface. (After Halliday and Newman). ©1 and Q 2 adjusted to select electrons diffracted from (-400) planes. Mag- nification 66,000X. focus in a direction perpendicular to its plane. 999 REFLECTION 11 examination of fibers both natural and syn- thetic and in the study of wear. jMore recent work has enabled some of the above disadvantages to be overcome. Brad- ley (11) has suggested the use of sohd metal replicas for those specimens which suffer beam damage. Halliday and Newman (12) have photographed etched germanium sur- faces at high resolution by imaging diffracted electrons which have a lower energy spread than scattered electrons. The greatest ad- vance has, however, been in a return to much larger angles of beam deviation though not to the extreme value of 90° used b\^ Ruska. Fert, Marty and Saporte (13), Page (14) and Ito, Ito, Aotsu and ]\liyamae (15) have all used angles of deviation in the region of 25 to 30°. The considerably reduced fore- shortening obtained at the higher angles of viewing gives to the image a strikingly three- dimensional appearance which can be highly informative. It would seem though that no further significant advance in the reflection technique will be made until lens designers can find a practical method for the correc- tion of axial chromatic aberration. If such a correction became feasible we might have an instrument capable of direct and rapid observation of surfaces at a resolution com- parable to that attained at present by rep- lica methods. REFERENCES 1. Ruska, E. Z., Phys., 83, 492 (1933). 2. Ruska, E. and MtiuLER, H. O., Z. Phys., 116, 366 (1940). 3. vox BoRRiES, B., Z. Phys., 116, 370 (1940). 4. Ku.sHxiR, Yu. M., Biberman, L. M., and Levkin, N. p., Bull. Acad. Sci., URSS Per. Phys., 15, 306 (1951). 5. Cosslett, V. E. AND Jones, D., J. Sci. In- strum., 32, 86 (1955). 6. Fert, C. and Saporte, R., C. R. Acad. Sci., Paris, 235, 1490 (1952). 7. Menter, J. W., /. Inst. Metals, 81, 163 (1952). 8. Haine, M. E. and Hirst, W., Brit. J. Appl. Phys., 4, 239 (1953). 9. Emerton, H. W., Research (London), 7, S56 (1954). 10. Page, D. H., Brit. J. Appl. Phys., 9, 268 (1958). 11. Bradley, D. E., Brit. J. Appl. Phys., 6, 191 (1955). 12. Halliday, J. S. and Newman, R. C, "Fourth International Conference on Electron Microscopy," Berlin, September 1958. 13. Fert, C, Marty, B., and Saporte, R., C. R. Acad. Sci., Paris, 240, 1975 (1955). 14. Page, D. H., Bn7. J. Appl. Phys., 9,60 (1958). 15. Ito, K., Ito, T., Aotsu, T., and Miyamae, T., /. Electron Microscopy, 5, (1957). D. H. Page REFLECTION II In electron microscopy, as in fight micros- copy, it is possible to examine both trans- lucent and opaque objects — ^the former with transmitted and the latter with reflected illumination. Most of the electron micro- scopes built have been of the transmission type since higher resolutions are attainable with this arrangement. Only the thinnest films are translucent to electrons and so it has been common practice to use replica methods when information about the surface structure of solid specimens was wanted. However, considerable progress has been made with techniques for the direct examina- tion of the surfaces of solid objects. Those used at present include scanning, emission, and reflection electron microscopy. Each gives a different type of information and this article will deal with one of them only — reflection electron microscopy. In the reflection electron microscope the electron gun is tilted with respect to the axis of the objective and projector lenses. The specimen is also tilted with respect to this axis but at a smaller angle to it. The arrange- ment is shown in Figure 1. The beam of electrons from the electron gun strikes the specimen at an angle di and the electrons are there scattered in all directions. Those scat- tered at angle 62 to the surface in the direc- tion of the objective aperture pass through it and subsequently through objective and 223 ELECTKON IMICKOSCOI'Y BtiM moil iixomoii (na OBJECTIVE AND PBOJECTOB i£Ha;s Fig. 1. Schematic diagram of a reflection elec- tron microscope. projector lenses to form an image of the surface. Factors governing the choice of di and $2 . The first reflection electron microscope built by Ruska in 1933 had di -\- 62 = 90°. The image was faint and the resolution only about 5000 A. Von Borries (1940) had better results using glancing angles of illumination and viewing {di and 62 about 4° each). The potentialities of the technique were not at that time fully explored, but after the war a number of workers developed the method further. These included Kushnir, Biberman, and Levkin (1951), Menter (1952), Fert and Saporte (1952), Haine and Hirst (1953), Cosslett and Jones (1955). In all this work the von Borries arrangement of glancing angles of incidence and viewing was used — values of Oi of about 1° and 62 about 8° were common. Images of moderately good resolu- tion and intensity were obtained, but since the surfaces were viewed obliquely they ap- peared foreshortened. In such images the magnifications are different in different direc- tions and it is usual to call the maximum magnification (in the direction perpendicular to the plane of incidence) Wj. and the mini- mum magnification (in the foreshortened direction parallel to the plane of incidence) mi| . Then m|| = m^ sin 62 and with an angle 62 of about 8° the foreshortening factor (mj./w||) is about 7; hence there is a con- siderable distortion of the image. This makes it diflEicult to interpret the pictures — never- theless, much work has been done with this type of arrangement with useful results. The electrons scattered at the specimen change velocity by different amounts, which can be quite large. Consequently, chromatic aberration in the objective lens becomes the factor limiting the resolution of the reflection electron microscope. The best resolution so o far obtained is about 300 A and this is prob- ably somewhere near the limit attainable unless some means of reducing the chromatic aberration (e.g., a velocity filter or achro- matic lens) can be developed. The figure of o 300 A represents the resolution in the direc- tion of maximum magnification. The resolu- tion in the foreshortened direction is worse by the foreshortening factor since a circle of confusion at the image corresponds to an ellipse at the specimen. It would obviously be desirable to reduce the foreshortening and several methods of doing this have been tried. The simplest is to view the final image through a cylindrical glass lens arranged to magnify it in the fore- shortened direction. An alternative is to use a cylindrical electron lens (Fert, 1956) to correct the distortion before the image is formed on the fluorescent screen of the elec- tron microscope. This approach is not very fruitful since the resolutions differ in differ- ent directions in the image. The most promis- ing method is to use larger values of 62 . The main reason for the common use of small values of 62 is that the intensity of the image falls off very rapidly as the total angle of deviation is increased. It is difficult to see to focus the image if 62 is too big. However, by using a suflEiciently powerful electron gun Fert and his colleagues at Toulouse have ob- tained good images with 61 = 2° and 62 = 224 REFLECTION 11 23° or 38° (see Figure 2). The distortion was then by a factor of 2.6 or 1.6 onh^ and the pictures were comparatively easy to inter- pret. One ditHculty encountered was the rapid contamination of the specimen surface by the intense electron beam. Fert overcame this by using ionic bombardment of the surface to remove the contamination con- tinuously or, if desired, to etch the surface. The resolution in the direction of maximum o magnification was again about 300 A and was affected very little by the value of d^. . The resolution in the direction of minimum magnification was worse by a factor of 1/sin di because of the foreshortening. Hence as 02 was increased the resolution in this direc- tion improved markedly. The high quality of the images was at first thought to be rather surprising since it had generally been supposed that the energy losses of the scat- tered electrons would increase as the devia- tion of the electron beam increased, thereby introducing a serious loss in resolution be- cause of the increased chromatic aberration. When the spectrum of energy losses was measured (Fert, Pradal, Saporte, and Simon 1958) it was found that this supposition was incorrect, for the proportion of electrons with high energy losses actually decreased as the scattering angle was increased. It has thus been shown that for best re- sults the largest practicable value of di + d^ should be used. Page (1958a) has taken good pictures with Qi -{- d^ = 26.5° using a modi- fied commercial instrument and without ionic bombardment of the specimen. Un- fortunately the commercial instruments available at present do not allow very large angles of tilt of the electron gun. The choice of angle ^i is important. Since the surface of the specimen is illuminated obliquely any asperity on it will cast a shadow. If di is small this shadow will be long compared with the height of the asper- ity. Vertical features may thus be given very high contrast: suppose for instance that Fig. 2. Pearlitic steel specimen after ionic etching: montage of three reflection electron mi- crographs. © 1 = 2°, © 2 = 23°. (Photograph repro- duced by courtesy of Professor C. Fert.) 01 = 1°; then a step 100 A high would cast o a shadow 5700 A long and at a magnification mj. of 2000 X would have a length in the image of 0.16 mm if 62 = 7° or 0.50 mm if 02 = 25°. This shadowing effect is one of the most valuable features of reflection electron microscopy, and in order to obtain good con- trast it is normal to use an angle di of about 1° or if the surface is very smooth a smaller angle which may be as low as 0.1°. Also it is necessary to use a small value of 9i to obtain good resolution. Fert, Marty, and Saporte (1955) have shown that di must be less than 3 or 4° if the resolution is not to suffer. Thej'' attribute this to the fact that if 61 is small the electrons will not penetrate far into the specimen and there will be less chance of an electron scattered there in the direction of the objective suffering a second collision be- fore leaving the specimen (particularly if 62 is large). Thus the angles 61 and 62 should be chosen as follows: di should be made small (-^1°) and its exact value chosen to give the desired contrast for the particular speci- men being examined, and 62 should be made as large as is practicable. 225 ELECTRON IVI I C KOSCOP Y Fig. 3. Reflection electron micrograph of the surface of a polished sapphire ball (diameter 1.27 cm). Though curved, the surface is well focused. Q J -|-©„ = 11°,© i~l to2° (varying across field), nu = 2500X,m|| + 400 X. Applications There are two main types of specimen for which reflection electron microscopy is a particularly suitable method of surface examination. First, there are very smooth surfaces such as polished metals, glass, cleavage surfaces of some crystals, etc. Fig- ure 3 shows a micrograph of the surface of a polished sapphire ball and is representative of the results which can be obtained from this type of surface. If the angles di and do are known, then the heights of asperities can be calculated from the shadow lengths using simple geometrJ^ The method can give useful quantitative information but there are a number of possible sources of error to be kept in mind. There may be transmission of the electron beam through the edges of asper- ities; the local surface on which the shadow falls may be at an angle to the plane of the specimen surface as a whole; the incident electron beam may not be strictly parallel and may therefore give penumbra; electron- induced contamination can build up on the edges of scratches, etc. and give misleading results. However, comparison of results ob- tained by this and other methods (see e.g., Bailey and Seal 1956) has shown that for o surface roughnesses of 100 A or more the method gives accurate results provided that reasonable care is taken in the operation of the instrument and interpretation of results. Features down to about 30 A in height can be resolved by the shadows they cast, but for such small features the quantitative re- sults are less reliable. The second type of application depends on another effect. This is the high depth of field inherent in the electron microscope be- cause of the very small angular aperture of the objective lens. The depth of field of a microscope objective is ± 5/tan a where 5 is the resolution and a. the semianglar aperture of the objective. For an optical microscope at high magnification 8 would be about 2500 A and a about J-^ radian giving a depth of field of ztVo u. A reflection electron micro- ^ o scope would have 5 about 300 A and a about 3.10"^ radian giving a depth of field of ±10 n. Thus the depth of field of a reflection electron microscope is considerably greater than that of an optical microscope at similar magnifi- cation. Consequently, focused pictures of highly curved objects such as balls, wires, fibers, etc., and of features such as scratches or grooves can be obtained. Examples are shown in Figures 3 and 4: the polished sapphire surface shown in Figure 3 had a radius of curvature of 0.63 cm; Figure 4 shows one groove of a gramophone recording. Although it is possible to examine gross features of this kind there is a restriction: they must either be single isolated features on an otherwise smooth surface or have extent in one direction only. It is not possible to examine rough surfaces of large extent since the shadows cast by asperities would obscure neighboring detail: in such cases the mountain tops onlj^ are illuminated and the valleys left in shadow. With a feature such as a scratch the illumination should be parallel to the scratch since otherwise the edge would probably cast too long a shadow. Figure 5 shows a surface (of lathe-turned brass) which is about as rough as can usefully be examined by this technique. In reflection electron microscopy one examines the specmien directly and there is 226 REFLECTION 11 therefore the possibihty of performing ex- periments on the surface while obsenang it. This has not received the attention it de- serves, but Cosslett and Jones (1955) have buih a reflection electron microscope with a hot -stage and have been able to observe the changes brought about by heating silver and other materials. Most work in reflection electron micros- copy has been done with metal specimens but it is possible to examine other materials. It is necessary for the specimen surface to be electrically conducting since otherwise it be- comes charged and repels the electron beam. If an electrical non-conductor is to be ex- amined, it is usual to render its surface con- ducting by evaporating a layer of metal (usually a few hundred angstroms thickness of silver) onto it. There are also some diffi- culties in examining organic materials since these are readily decomposed by the electron beam: the use of very low beam intensities is often necessary in order to obtain pictures of these materials. These difficulties can be overcome by using the replica technique of Bradley (1955) — a rigid metal replica of the surface to be examined is prepared and forms a robust specimen for the reflection electron microscope. There appears to be no loss in resolution, but the advantage of direct examination of the surface is of course lost. A few references to typical applications of reflection electron microscop}^ follow. Metals with different surface preparations have been studied by Halliday (1955, 1957); diamond surfaces of different types by Seal and Menter (1953) and Seal (1956, 1958a and b); the epicuticular surfaces of insects by Holdgate and Seal (1956); natural and synthetic fibers and the effect of abrasion on them by Chapman and Menter (1954); lubricant layers of graphite and other lamel- lar solids by Deacon and Goodman (1958): cleavage faces of zinc crystals by Moore (1955); surface damage on mica by Bailey and Courtney-Pratt (1955) and on rocksalt by King and Tabor (1954); paper and pulp Fig. 4. Reflection electron micrograph of a single groove of a gramophone record. There is a sine wave modulation of which rather more than one cycle is visible. © i = 2°, © 2 = 9°. nij. = 570 X, mil = 90 X Fig. 5. Reflection electron micrograph of part of a lathe-turned brass surface. The circular grooves left by the cutting tool appear foreshort- ened as portions of ellipses. © 1 = 2°, © 2 = 9°. nij. = 1000 X,mi! = 160 X. fibers by Amboss, Emerton, and Watts (1954) ; the fretting corrosion of mild steel by Halliday and Hirst (1956). Since the main part of this article was written, further papers on the subject have appeared. Halliday and Newman (1958, 19G0) have published the results of an in- vestigation of reflection electron microscopy using diffracted electrons. In these experi- ments crj^stalline specimens were used and the angles of tilt of the electron gun and specimen adjusted so that a diffracted beam passed through the objective aperture. This beam was used to form an image, analogous to the "dark field" image often used with 227 ELECTRON IMICKOSCOI'Y transmission specimens. It was stated that the resokition should be higher than usual since the energy spread of electrons diff- fracted at a Bragg angle would be smaller than that of inelastically scattered electrons. A resolution of 80 A Avas obtained experi- mental! J^ Kushnir and Der-Schwarz (1958) dis- cussed various problems in reflection elec- tron microscopy and described some experi- mental results. They dealt with the spatial intensity distribution and velocity spectrum of electrons scattered at a massive object, the correction of chromatic aberration and geometrical distortion, and the effects of ob- jective aperture displacement. Further use has been made of the possibil- ity of carrying out experiments on a specimen whilst observing it in a reflection electron microscope. Halliday and Rose (1959) de- scribed the direct observation of wear proc- esses in this way. There was also an earlier paper by Takahashi, Takeyama, Ito, Ito, Mihama, and Watanabe (1956) describing a hot stage for reflection electron microscopy and some results obtained from this. Equations A number of equations relevant to the interpretation of reflection electron micro- graphs and the use of the instrument follow. The magnifications in directions parallel to and perpendicular to the plane of incidence are related by The magnification along a line in the image at azimuth ^p is m^' = VI X sin d-i sec f I / ^ .ill ,. , ■ — (4) and a circle in the specimen will be imaged as an ellipse. The height of an asperity h is related to the length of the shadow it casts L (measured in the image) by h = L sin di m± sin (di + Bo) (5) If the validity of this equation is not to be affected by penumbra, a restriction must be imposed on the angular spread of the inci- dent electron beam 2Aai < 5 J. sin^ di h sin $2 (6) Thus good collimation becomes important for small ^i . Other factors affecting the validity of equation 5 have been listed on page 225. The depth of field is approximately rtSx/of, (7) where a is the semi-angular aperture of the objective lens. The chromatic aberration of the objective lens limits the resolution to aCy, (8) m\\ = mx sin do , (1) and the resolutions in these directions by a similar equation 5|| = 5 J. /sin $2 (2) Angular relationships in the image are dis- torted because of the foreshortening. A line in the specimen at azimuth cp (measured to a line in the specimen in the plane of incidence) will be imaged as a line at azimuth cp' where tan HCKOSCOPY Specimen Holder Specimen Fig. 1. Specimen mounted in plastic block and placed in wedge-shaped specimen holder. polished cross section of the coating and tungsten wire. Upon completion of the polishing, the samples were etched in a 10 % nitric + 3 % hydrofluoric acid solution for approximately five minutes. This etch was sufficient to reveal the crystallographic de- tail in the interfacial region. This surface was then replicated by the process pre- viously described. The plastic block containing the sample was then mounted in a "Lucite" wedge, as showTi in Fig. 1. The wedge containing the sample and block was placed in the poly- ethylene Buchner funnel, as shown m Fig. 2. With the hosecock closed, etchant was added until the surface of the liquid had crossed the lower edge of the mounting block and was nearly in contact with the lower edge of the shadowed specimen. Distilled water was then added from the buret until the edge of the etchant surface was properly positioned at the bottom edge of the sample. Prior to this step the edges of the sample had been scored so as to break the carbon film and permit the etchant to attack the sample material. The progress of the etchant was followed by means of a 15 X microscope which could be swung over the sample on a pivot arm. As the etching progressed, water was added from the buret to raise the level of the liquid and hence help pull the replica from the sample and float it upon the etchant surface. When the separation was completed the replica floated freely upon the surface of the etchant. The etchant level was then raised by means of the buret until it was completely free of the mounting block. It was then neces- sary to dilute the etchant to approximately twenty times its volume to raise the pH to a level where the replicas could be picked up on the metal grids. The hosecock shown in Fig. 2 was opened and the stopcock of the buret opened. The flow into and out of the funnel was adjusted so that the liquid level remained essentially constant. The distilled water from the buret was introduced below the surface of the etchant via a poly- ethylene tube. This step greatly reduced the turbulence of the surface and hence reduced Buret Specimen Specimen Holder Buchner Funnel [PolyethylcneJ Hosecock Fig. 2. Apparatus for separating replica from specimen in solutions of low pH. 240 SCANNING the tearing of the repHcas during the diki- tion. When the pH of the solution in the top of the funnel was approximately 5, the rep- lica was placed on a grid. The diluting process could be accelerated by removing the plastic block containing the sample from the etchant after the replica was released. The danger of the resulting turbulence tearing the replica usually pre- cluded the use of this time-saving step. This technique should be applicable to the removal of carbon replicas from any solid material which can be etched from beneath the replica by acidic solutions, particularly those containing hydrofluoric acid. V. J. Texnery, C. G. Bergeron, AND R. BORASKY RESINOGRAPHY. See p. 525. SCANNING In the scanning electron microscope, an electron optical system is used to produce a fine electron spot which, in turn, is caused to move over each point of the specimen. The electron current leaving the specimen is collected and amplified, and the resulting signal is used to modulate a recording device moving in synchronism with the electron spot but with a greater amplitude. If the specimen possesses some property which causes the electron current leaving it to vary from point to point, then a record will be built up which will represent in some way the variation of that property over the area of the specimen which is scanned. The re- solving power of such a microscope is de- termined by the size of the electron spot. A microscope intended for the study of secondary emission phenomena and based upon the above principles was first proposed by Ivnoll (1) in 1935. The first practical scanning microscope was described by von Ardenne (2, 3) in 1938, the main purpose of this instrument being the examination of thick specimens in the transmission mode*. This instrument, which used electrostatic lenses and electromagnetic scanning, pos- sessed no amplifying device; instead, the electron beam, after passing through the specimen, was allowed to fall on a film mov- ing in synchronism with the beam but with a greater velocity. The magnification was given in this case by the velocity ratio of the beam and the moving film. This early instrument possessed several practical dis- advantages among which w^ere very low effi- ciency, necessitating recording times of 30 minutes or more, and no direct means of observing and focusing the picture before recording. von Ardenne suggested that living biologi- cal specimens might be examined in air by first passing the beam through a thin Lenard window, and further that the sur- faces of opaque specimens could be exam- ined by collection and amplification of the secondary electrons to modulate a cathode- ray tube display. A scanning microscope designed specifi- cally for the direct examination of the sur- faces of metallurgical specimens was de- scribed by Zworykin et at. (4, 5) in 1942; it used electrostatic lenses and electro- magnetic scanning of the beam. The speci- men was placed with its surface in a plane perpendicular to the axis of the objective lens, and secondary electrons emitted from the surface were accelerated back through this lens to be detected by a fluorescent screen followed by a photomultiplier. After further amplification, the signal was applied * In thescanning instrument the electron beam does not permit any focusing after its interaction with the specimen; hence, energy losses occurring in a thick specimen observed in transmission are less serious than in the case of the conventional transmission instrument. In practice, however, the scanning method is only advantageous for one particular type of specimen typified by highly scattering particles lying on or near the surface of a substrate of much lower scattering power (see reference 6). 241 ELECTRON MICROSCOPY to a facsimile recorder scanning in syn- gether with direct viewing of the picture chronism wilh the electron beam. Contrast realized the possibility of an instrument in the image was produced partly as a result suitable for practical laboratory application. of variations hi the composition of the sur- The techniques and applications described face and partly because of the surface topog- in the following sections relate largely to raphy. The energy of the electron beam in- this instrument (6) and developments there- cident on the specimen was made low, only from (7, 8, 9). 800 volts, in order to achieve a high sec- Formation of the Electron Spot. The ondary-emission coefficient. Focushig of this function of the electron optical system is to instrument was achieved by observation of focus a fuie electron spot of high intensity the video signal with an oscilloscope; the onto the specimen. Two principal factors high-frequency content of the signal was limit the performance in this respect — aber- used as a criterion for adjustment. rations of the lenses, and limited brightness A disadvantage of this instrument was output from the electron gun. Expressions that no means was provided for observing are derived below which indicate how these the field prior to recording. The use of a factors affect the performance, low-energy beam reduced the intensity of The electron spot which is a reduced im- the spot, thus necessitating rather long re- age of the virtual cross-over in the electron cording times of about 10 minutes. Also, at gun is formed by a two-stage lens system low primary energies, contamination of the (Figure 1). By varying the strength of the specimen surface with organic vapors, al- first lens, the demagnifying power of the ways present in a demountable vacuum sys- system, and thus the size of the spot, may tem, was likely to affect significantl}^ the be controlled over a wide range without secondary-emission coefficient. affecting significantly the position of the A scanning electron microscope, similar spot. Because the effective aperture at which to the above instrument but incorporating the first lens works is very minute, the aber- several new features, was proposed by Oat- rations of this lens do not contribute signifi- ley and McAIuUan in 1948 and constructed cantly to the limitation of the performance, by McjMullan at the Engineering Depart- The spot is focused on the specimen by ment, University of Cambridge. This instru- means of the final or objective lens. The ment used high-energy electrons (10-40 kV) specimen, which may be as large as 1 cm in to reduce the effects of surface contamina- diameter, is placed as close to this lens as tion, oblique scanning of the specimen, direct possible since the focal length and conse- amplification of the emitted electrons with quently the spherical aberration is reduced an electron multiplier suitable for use in a thereby. If ferrous specimens are to be exam- demountable vacuum system, and direct ob- ined, the working space must also be field- seni-ation of the field on a long persistence free. The requirements of a magnetic scan- cathode-ray tube display prior to recording ning objective are best satisfied by an the picture. While oblique scanning resulted asymmetrical pole-piece design (10); typical in a foreshortening and reduction of resolv- values of focal length and spherical aberra- ing power in the vertical direction of the tion coefficient for this type of lens with a image, it allowed a much simpler and more working distance of 0.5 cm are 1 cm and efficient collection arrangement. The fea- 3 cm, respectively. Astigmatism of the ob- tures incorporated in this instrument re- jective caused by pole-piece ellipticity may moved most of the disadvantages of earlier also be a significant fimitation. instruments. In particular, the reduction of Spherical aberration and astigmatism of recording times to 5 minutes (or less) to- the objective and, at the smaller spot sizes, 242 SPECIMEN COLLECTOR PHOTOMULTIPLIER OBJECTIVE LENS BEAM DEFLECTING COILS FIRST LENS GUN "^LIGHT PIPE VIDEO AMPLIFIER (GAMMA tPNTROLl MAGNIFICATION ICONTROLS z tm -JTSTI SCANNING GENERATOR SCANNING VISUAL DISPLAY TUBE RECORDING TUBE Fig. 1. Schematic diagram of an electromagnetically focused and deflected scanning microscope. diffraction at the objective aperture will each cause a point object to be imaged in disks of confusion, the diameters of which are given by the folio whig expressions. Spherical aberration Astigmatism da = Zaa Diffraction d/ = 1.22 X/a The maximum current density, ja , in the Gaussian spot, i.e., in the absence of aber- rations, may be determined from Langmuir's formula (12) which in the form applicable to the electron microscope is (1) (2) (3) eV Ja = J — a^ amp/cm2 (5) where Cs is the coefficient of spherical aber- ration, Za is the axial astigmatism, X is the electron wavelength and a is the semi-angu- lar aperture which the objective subtends at the specimen. In order to estimate the combined effect of spherical aberration and diffraction in the conventional transmission microscope, the disks of confusion may be assumed to con- tain a Gaussian distribution of intensity and their diameters may be added in quadrature (U). This procedure will be applied in the present case to estimate the total effective diameter of the electron spot, i.e. or d2 = do- + dr + da- + rf/2 C,2 (1.22X)2 (4) where do is the Gaussian spot diameter, where e is the electronic charge, V is the ac- celerating voltage, T and j are, respectively, the temperature and emission density of the gun filament and A' is Boltzmann's constant. [Haine and Einstein (13) have verified that the current density predicted by this equa- tion is achieved by the type of gun used in the electron microscope when operated at moderate values of emission density (<2 amp/cm'-). For higher values of emission the current density falls below the predicted value because of space charge effects.] Langmuir's theory predicts that awaj^ from the center of the spot the current den- sity will fall off according to a Gaussian dis- tribution, hence, the diameter of the spot is often defined as being the width of the Gaussian distribution curve at half-height (13). A definition for the effective diameter which appears to be more satisfactory for cathode-ray tubes and for the scanning mi- croscope is that diameter which includes 80 % of the total current (14, 15). 243 ELECTKON MICKOSCOPY Using this definition, the total current, i, in the spot will be i = 0.62 — • Ja amp 4 where the numerical factor 0.62 adjusts for the Gaussian distribution of intensity and applies only to this particular definition of do gi\'en above. „ -n-do^ . eV , i = 5.65 X 10' - rfoW considering the case for the corrected lens, equation 7 becomes 1 C.2 c/2 = ri.l77iV - X 10'" + (1.22X)2 a2 4 and d is a minimum when a = ctopi given by 1/8 ....... 09 (-) (7.92X10.,- + .) / J> \3/8 ( 7.92 X lO^j - + 1 j d„.ia = 1.29C,"4X5" (8) (9) In terms of electron wavelength i = 8.48 X 10-" 4- doW 1 \' (6) Equation 6 may be transposed for do and combined with equation 4 to give: d2 = 1.177fX2 -7 X W + (1.22X)2 3 C 2 4 (7) Thus, for a fixed value of spot current, the contribution of the first term in l/a^ is to reduce the diameter of the spot with increase of a while the contribution of the remaining terms is to increase the diameter with in- crease of OL. For any given current, there will be an optimum aperture setting and a cor- responding optimum setting of do (deter- mined by equation 6) at which the total spot diameter will be a minimum. This mini- mum may be derived from equation 7 as it stands, but the resulting expressions are less cumbersome if two cases are considered, (a) for an objective corrected for astigmatism and (b) for an uncorrected objective. In the latter case, astigmatism will in general pre- dominate and the spherical aberration term may be neglected. However, since correction for astigmatism is relatively easily accom- plished, spherical aberration imposes the more fundamental limitation. Therefore, It may be shown from the foregoing equa- tions that the Gaussian spot diameter, f/o , must be set to approximately y/Ji dmin for the optimum conditions. Expressions equivalent to 8 and 9 may be derived for the case of an uncorrected objec- tive in which astigmatism is predominant. Equation 9 shows clearly the hmitations imposed by spherical aberration, diffraction, and by the electron gun. If only very small spot currents are required, the first term in the brackets may become insignificant, in which case the optical aberrations become the ultimate limitation. For the usual values of filament temperature and emission exist- ing in the electron microscope, this occurs at spot currents of approximately 10"^^ amp. Currents somewhat greater than this are re- quired in the scanning microscope. Limitations Imposed by Fluctviations in the Electron Beam Current. Because of the particulate nature of an electron beam, the illumination of an object in the electron microscope is a random process. A mean number of electrons n, falling on an element of the object, has associated with it a fluctuation of r.m.s. magnitude \/w. The basic signal-to-noise ratio is then nl\Jn = \/^- If the secondary emission ratio at the specimen is not less than unity and no further noise is introduced during interaction with the specimen, a change An in the number of electrons emitted from the surface as the beam is scanned from one 244 SCANNING element to the next will be detectable as a change AB m the brightness of the picture only if the basic signal-to-noise ratio exceeds B/AB by a factor of about, 5 [Rose (16)]. AB/B = C is then the threshold contrast. It may be shown (9, 17) that during in- teraction of the beam with the specimen, the signal-to-noise ratio may be reduced by roughly a factor of 4. Further, possibly only one-third of the emitted electrons are effec- tively collected, thus / n 20 3 - C or n > 1200 (10) Therefore, to detect a contrast change of 5 % between adjacent picture elements requires a minimum of 4.8 X 10^ electrons per ele- ment . If the specimen is scanned in a square raster containing N lines, i.e., A^- elements, in a total time Ts seconds, then each element is covered in a time T^/N^ seconds and the required spot current will be i = ne —- amp t = 1200 T3::r Camp (11) C^Ts (For a 300-line picture, a threshold contrast of 5% and a scanning time of 5 minutes, i = 2.31 X 10-" amp). Substitution of this expression for i in equation 9 gives for the minimum spot diameter dmin = 1.29C,i'^X3'4 1.52 X 10-« (" m c^, f \3/8 -T + lj (12) This basic equation includes all the fac- tors which affect the resolving power of the scanning microscope. In Figure 2, the spot diameter is plotted for three different values of A^ as a function of the recording time for the following typical parameters: Cs = S 120 180 240 300 RECORDING TIME (SECONDS) Fig. 2. Electron .'jpot diameter, dmin , as a function of recording time, Ta , according to equation 12. cm, X = 10-9 cm, (F = 15 kV), c = 57o, T = 2800°K, j = 3.5 amp/cm^ [effective emission density at a filament life of 8-10 hours (13, 18)]. It will be seen that a re- solving power close to 100 A is theoretically obtainable if the field is reduced to 150 lines. Poor focusing will occur at the top and bottom of the picture if the number of lines or elements scanned in the vertical direction of the picture is too great. The more oblique the angle of observation the fewer the num- ber of lines or elements which can be re- solved. At an angle of 45° this will be about 300 lines. The number of elements scanned in the horizontal direction may be much greater, although this will entail either longer recording times or lower resolving power according to equation 12. The Beam Scanning System. The de- flecting system may be electromagnetic or electrostatic. The latter allows a somewhat simpler physical construction and associated circuitry but, at high beam-voltages, the deflecting potentials may become incon- veniently large, the deflection sensitivity be- ing proportional to the beam-voltage, and raster distortion may be introduced. These disadvantages may be overcome by intro- 245 ELECTRON MK.KOSCOPV . . . . I. i.V . / MEAN PLANE . f < < IETIIODS 8. KuMAi, M., "Electron-microscope study of snow-crystal nucli II." Geofisica pura e applicata-Milano, 36, 169-181 (1957). 9. KuRoiWA, D., "Electron-microscope study of atmospheric nuclei." T. Hori (ed.) "Studies on fog," Sapporo, pp. 349-382 (1953). 10. Lodge, J. P., "Analysis of micron-sized par- ticles." Anal. Chem., 26, 1829-1831 (1954). 11. Maruyama, H., "Electron-microscope study of the ice crystal nviclei." Meterology and Geophysics, Tokyo, 7, 251-266 (1956). 12. MUGURUMA, J. AND HiGUCHI, K., "On the etch pits of snow crystals," /. Meteor. Soci. of Japan. Ser. II. 37, 71-75 (1959). 13. Nakaya, U., "The formation of ice crystals." Compendium of Meterologj^, American Metero. Soc, pp. 207-220 (1951). 14. Nakaya, U., "Surface nature of ice crystals," in "Artificial Stimulation of rain," H. Weickman (ed.) Pergamon Press, New York, pp. 386-389, 1955. 15. Nakaya, U. and Kumai, M., "Electron- microscope study of center nuclei of snow crj^stals III. J. Meteor. Soci. Japan, 75th anniversary volume, pp. 49-51, 1957. 16. Ogiwara, S. and Okita, T., "Electron- microscope study of cloud and fog nuclei," Tellus, 4, pp. 233-240 (1952). 17. Schaefer, V. J., "The formation of ice crys- tals in the laboratory and the atmosphere," Chem. Rev. 44, 291-320 (1949). 18. Schaefer, V. J., and Harker, D., "Surface replicas for use in the electron microscope," /. Applied Phys. 13, 427-433 (1942). 19. Schaefer, V. J., "The question of meteoritic dust in the atmosphere," in "Artificial stimulation of rain," H. Weickman (ed.) Pergamon Press, New York, pp. 18-23, 1955. 20. VoNNEGUT, B., "Nucleation of ice formation by Agl particles." Final Rep. No. RG 140, Gen. Elec. Res. Labs., pp. 26-34, 1948. 21. Yamamoto, G. and Ohtake, T., "Electron microscope study of cloud and fog nuclei," Science Report. Tohoku Univ. Ser. 5 Geo- phys., 5, 141-159, 1953. MoTOi Kumai SPECIAL METHODS One of the aims of electron microscopy is to observ'e the arrangement of constituent atoms and molecules of materials. Recent improvement in the electron microscope has made it possible to observe the lattice image of crystals of which spacing is less than 10 A. Miiller, using the field emission type elec- tron microscope, has succeeded in observhig directly the arrangement of atoms such as W, Re, etc. Besides the problem of high reso- lution, various special methods for electron microscopic observation have been tried to develop the fields of its application. In this chapter some of these special methods, which have been studied mainly in Japan, arc in- troduced and their constructions, perform- ances and experimental results are briefly described. They are as follows: (1) Reflection method (2) Specimen cooling method (3) Specimen heating method (4) Gas reaction method Reflection Method In obser\dng solid surfaces by the trans- mission electron microscope, we usually use the replica film which is reprinted from the surface structure of the solid. Then, a method for direct observation of solid surfaces is naturally desired, just as for an optical mi- croscope for metallurgy. There are two meth- ods in the deflecting mechanism of the elec- tron beam in a reflection microscope: (1) makes the electron beam strike the specimen by mechanically inclining the electron gun and condenser lens at any desired angle ; the other (2, 3) makes the electron beam strike the specimen by using two pairs of deflecting coils without inclination of the illuminating system. This article describes the latter. Construction of Reflection Electron Microscope. Fig. 1 shows the principle of the reflection device for JE]\I-5Y universal electron microscope and Fig. 2 the specimen chamber equipped with this device. For changing the incident angle /3, two pairs of deflecting coils are inserted between the condenser lens and the specimen chamber as shown in Fig. 1. The electrons emitted from the electron gun are deflected from the optical axis by the upper pair and again to the speci- men by the lower pair. It is possible to 259 ELK(II{( >N M K l{< >SC()I*Y upper ceiU Cower cotU defUcierf beam ntm.JeHected tea Sf>ecimen. Fig. 1. Principle of reflection method. Fig. 2. JEM-5Y Electron microscope equipped with reflection device. change the angle /3 to any angle between 0° and 30°. The reflection specimen may be tilted to any angle between 0° and 30° from the lens axis, about an axis perpendicular to the plane of incidence, and may be rotated in its own surface plane, i.e., azimuthal plane. It is helpful to change the azimuthal angle for the reflection method, because many observations from several directions are needed to give a perfect interpretation of the image obtained. The specimen-shifting mechanism for the two horizontal directions is the same as that for general use of the transmission type. In the reflection electron microscope, the electrons striking the specimen surface at a small angle arc scattered to the angle Q and then the scattered electrons are imaged. When the angle Q becomes larger, the inten- sity of the image is much lower; when the angle Q is smaller, the distortion of the image is more marked. The relation between the angle Q and the distortion of the image is sho\vn in Fig. 3. In the present device, as an angle Q larger than 20° is used, distortion of the image can be considerably reduced. Because of its large chromatic aberration, however, the resolution of reflection is less than that of transmission. On the other hand this method has two features. The change of the specimen at high temperature can be directly observed by us- ing a specimen-heating device at the same time. It is also possible to obtain the electron diffraction pattern and microscopic image from almost the same area of the specimen. In the reflection method, however, it is diffi- cult to obtain the diffraction pattern of the /2h O* 30" C0° Observation an^ic 10' Fig. 3. Relation between image distortion and observation angle. 260 SPECIAL MK/niODS selected area corresponding completely to the reflection image except in some special cases, because in electron diffraction the in- cident angle of the electron beam to the specimen is less than 1°. Considerations on Resolution. In gen- eral the resolution of the electron micro- scope is determined mainly by spherical and chromatic aberrations, astigmatism and diff- raction. In the reflection electron microscope, chromatic aberration is comparatively great because of the larger amount of inelastic scattering suffered by an electron beam. Therefore, it is considered that chromatic aberration is mainly responsible for limiting the resolution obtainable in the reflection microscope. Here, the order of magnitude of the reso- lution is estimated which is expected in a direction perpendicular to the plane of inci- dence of the electron beam. Neglecting all aberrations except chromatic, we may write: 5x = «•/• AV (1) To achieve higher resolution, it is neces- sary for the objective aperture to be smaller. For example, to obtain a resolution better than 100 A, it is estimated that the ol)jective aperture about 10 n would be required. But by using such a small aperture, it is doubtful whether the intensity of the image would be adequate for precise focusing at high magnifi- cation. Menter has obtained the resolution better than 400 A using a 30 /x aperture in electron microscope with a focal length somewhat longer than that in the above calculation. Application of Reflection Device. Fig. 4 is a reflection image of pearlite structure. (a) is taken at 6 = 8° and (b) at 0 = 30°. It is found that the distortion of the image is reduced by increasing the observation an- gle. Fig. 5 shows the reflection image of mar- tensite structure and Fig. 6 the reflection image of pearlite structure of dra\^'n high carbon steel. They were taken at 30°. Studies by a reflection method have not so far given any further information to that obtained by the replica method, but in future where 8±is the radius of the disc of confusion in a plane perpendicular to the plane of inci- dence due to chromatic aberration, a the semi-angular aperture, / the focal length of the objective, V the accelerating voltage and AV the average energy loss of electrons scat- tered from the specimen. The variations in the high tension supply voltage and excita- tion current in the objective lens being much smaller compared with energy losses due to scattering at the specimen, their effects are neglected in equation (1). According to Kushnir et al., AV is lOOF when the acceler- ating voltage is 80 kV. Thus, AV/V = >^oo. Putting a ^ 3 X 10-^ rad. and f = o mm, we can find: 8± ^ 190 A The resolution in a direction in the plane of incidence will be worse by a factor 1/sin d, i.e., for d = 6°, 5,, ^ 2000 A and for 9 = 30°, 5 1, ^ 400 A. Fig. 4. Reflection image of pearlite (above) e = 8°, (below) e = 30°. 261 ELECTRON ^IICKOSCOI'^ Fig. 5. Reflection image of martensite 0 = 30°. Fig. 6. Reflection image of pearlite of drawn high carbon steel d = 30°. its applications may be developed by attain- ment to its higher resolution and by using a specimen heating or gas reaction method. Specimen Cooling Method In electron microscope the specimen must be irradiated by electron beams in the vacuum. Therefore, in the ordinary method it is impossible to study the substance of high vapor pressure, that of low melting pohit or organic substances easily damaged by electron irradiation. Cooling of the speci- men to a large extent can prevent these ef- fects and; develop the field of its applica- tion. This is the best advantage in specimen cooling methods. Construction and Performance of Specimen-Cooling Device. The specimen cooling device previously reported (5) had many unsatisfactory points in its construc- tion and was not suitable for obtaining high resolution. A new specimen cooling device has since been constructed (Figure 7) (6) for JEM-5Y electron microscope. A is the pouring port of refrigerant, B the refrigerant reservoir with capacity of about 30 cc and E the draught port. The refrigerant reser- voir B, the pouring port A and the draught port E are connected to each other by syl- phon bellows C and D. The draught port E makes it easy to pour the refrigerant into the reservoir. The specimen holder F is at- tached to the center of the round porcelain H, which is set up to the specimen shifter /. The temperature is measured by copper- constantan thermocouple J fixed on the specimen holder F. The heater K is used to control the temperature of the specimen. To set up this device with JEM-5Y, the refrigerant reservoir is kept at the position shown by the dotted line, by rotating the pinion A^ with the lever M. Then the whole 0 P,S,T Q 'Tffjfl t^^L^srr^ Fig. 7. Construction of new type specimen cooling device. 262 SPECIAL MKIHODS Fig. 8. Cu-Phthalocyanine taken at spacing 12.5°A. -50°C, (7). The performance of this device has proved very satisfactory. Contamination of Specimen. One of the principal difficulties of the specimen- cooUng device is the contamination of spec- imen due to materials condensing from resid- ual \'apors. In the conventional vacuum of 5 X 10-* to 1 X 10-^ mm Hg, condensed materials increased remarkably below -80°C to - 100°C and covered the specimen surface within a few minutes. Fig. 9 shows an electron micrograph and diffraction pat- tern of ice condensed on collodion film at — 80°C from residual vapor pressure of 1 X 10-* mm Hg. This proves that the larger part of the residual vapors is usually water vapor. Contamination of another nature has been observed in electron microscopic experi- device is inserted into the round port located at the head of the specimen chamber. When the specimen is to be exchanged, the reser- voir is elevated in the same way. Then the specimen cartridge G is moved in and out independently through another chamber for pre-evacuation. The specimen cartridge can be inserted deep in the field of the objective lens R. In order to cool the specimen, the refrigerant reservoir is lowered to contact the specimen holder F. At the lowest position the reservoir is pushed down to the specimen holder by the release of spring S and the elongation of sylphon bellows C and D. The specimen-shifting mechanism is just the same as that for general use of electron microscope and can control the specimen position as fine as 0.1 /x. When the reservoir was filled with liquid oxygen, the tempera- ture of the specimen holder reached to — 160°C. This temperature could be ob- tained within 15-20 min, using about 150 cc of liquid oxygen. The resolution obtained by using this de- vice is measured by copper phthalocyanine. Fig. 8 shows the lattice image of spacing 12.5 A of this specimen taken at — 50°C. Fig. 9. Electron micrograph (above) and dif- fraction pattern (below) of ice condensed on col- lodion film. 263 ELECTRON AI 1< KOSCOPY Fig. 10. Contamination on collodion film pro- duced by electron irradiation. ments. Fig. 10 shows contamination on collodion films produced by electron irradia- tion for several minutes at liquid nitrogen temperature in conventional vacuum of the order of 1 X 10~^ mm Hg. This micrograph was taken after the films were re warmed to 200°C. It is considered that this contami- nation is caused by electron irradiation, be- cause the amount of deposit depends on dosage of electron irradiation. It is very likely that this contamination is a product of hydrocarbon vapor decomposed by elec- trons, which is observed in ordinary electron microscope. In order to reduce contaminations, we must improve the working vacuum as far as possible. First, the inside walls of all parts of the apparatus are thoroughly cleaned, warmed and baked up to 70° to 150°C in the dryer, before assembly. Photographic plates were pre-evacuated in another vac- uum system. When the vacuum was broken, well-dried air was introduced into the appa- ratus. Then, using a 4-inch oil diff vision pump backed by a phosphorus pentoxide trap and a 150 1/min oil rotary pump, the vacuum of the order of 1 to 2 X 10~^ mm Hg could be obtained after half an hour. The refrigerant was poured into the reservoir without inserting the specimen. At this stage the vacuum of the specimen chamber became about 1 to 2 X 10-^ mm Hg. Finally, after being bombarded by the electron beam for several minutes, the specimen was put into the apparatus. This procedure efficiently reduced con- taminations. Fig. 11 shows a pair of electron micrographs of a collodion film of about 400 A thick; (a) taken before cooling and (b) at liquid nitrogen tem^perature after a con- tinual observation of 45 min. There is no appreciable change in the contrast between the film itself and its holes. This proved that contaminations were completely eliminated. Application of Specimen Cooling De- vice. Study of Mercury (5) : This study con- cerned the following points: (a) The char- FiG. 11. Collodion film about 400°A thick, (left) at room temperature before cooling, (right) after 45 minute continual observation at liquid nitrogen temperature. 264 SPECIAL METHODS acteristic crystal growth "whiskers" sug- gested by Sears (8), and (b) the tem- perature rise of the specimen caused by electron irradiation, which it was hoped could be checked by observing the hquid- solid phase transition. Fig. 12 (a) shows hair-like crystals formed by condensing mercur\^ vapor at about — 100°C. (b) was taken two minutes after (a). The crystal in- dicated by the arrow completed its growth within a fraction of a second. Fig. 13 (a) shows a droplet of liquid mercury. This drop- let was formed at -39° to -38°C by heat- ing the solid state crystal. This droplet of liquid mercury was poligonized through crys- tallization at about — 100°C as shown in (b). Study of Natural Cellulose Fiber (9). The atomic structures of an organic substance like cellulose fiber are so hable to be dam- aged by irradiation of electron beams that the diffraction patterns can be observed for only a very short time. We can avoid this difficulty by using a method of cooling the specimen. Fig. 14 reproduces two pairs of electron micrographs and diffraction pat- terns of Valonia microfibril taken by Boersch-le Poole's method ; (a) and (b) were taken without cooling and (c) and (d) by cooling the specimen to — 40°C. This speci- men is composed of two sheets of microfibril having fiber axes along the directions indi- cated by the arrows. In the electron diffrac- tion patterns of Valonia microfibril cooled to — 100°C taken by Hilhers method, a very large number of reflections having fairly good resolution up to as high as tenth layer line has been observed. By analyzing these patterns, the new structure different from that proposed by JNIeyer and Misch (10) for cellulose I has been obtained. Specimen-Heating Method The specimen heating method makes it possible to observ^e continuously the change of a specimen at high temperature, by at- taching a heating furnace of a specimen to ordinary electron microscope. This method is divided into two parts: (1) in which a heat- ing device with a small electric furnace is used for both reflection and transmission (11, 12) and (2) in which the temperature Fig. 12. Hair-like cry.stal.s of mercury, (above) Hair-like crj-stalso frmed Vjy condensing mercury vapor at about — 100°C. (below) Two minutes after picture above. Fig. 13. Mercury droplet, (above) Liquid state, (below) Solid state, crystallized at about — 100°C. 265 ELECTRON >lICKOSCOPY X X (a) (b) (c) (d) Fig. 14. Electron micrographs and diffraction patterns of Valonia microfibril: (a) and (b) without cooling; (c) and (d) cooled to — 40°C. of metallic thin wires as the specimen can be elevated by passing the electric current directly through them (13). In the latter case, only 0.6 to 6 watts is sufficient to raise the temperature of the specimen from 700°C to 3,000°C and some gas, when necessary for the chemical reaction with the specimen. can be introduced. Here, only the former in which we have experience, is described. Construction and Performance of Specimen-Heating Device. Fig. 15 shows the construction of the transmission type of the new specimen heating device (12) for JEM-5Y electron microscope and Fig. 16 266 SPECIAL METHODS the outside view of it. A is a connecting ring for setting this device on to the specimen shifting mechanism of ordinary type elec- tron microscope. B shows a ring-shaped insulator of porcelain. C, D and E are the outside walls of the furnace, which are made of porcelain. F is a heater of Mo-wire. G is an inside wall of the furnace, made of Mo- sheet. The specimen cartridge, composed of Mo and Porcelain I, can be put into and taken out of the furnace, which is heated beforehand, through another chamber for pre-evacuation. L is used as a guide plate, when the cartridge is inserted in the furnace. The temperature of the specimen is measured by Pt-Pt,Rh thermocouple mounted to G. To av^oid charging due to electron irradia- tion, the outer surface of the porcelain is covered with evaporated carbon. The pole piece of the objective lens is so well pro- tector 0 that it cannot be damaged by heat- ing. Reflection type of the specimen heating device is also made according to the same design with the only exception that it is equipped with a mechanism by which the glancing angle of the specimen is arbitrarily changeable. A battery is employed as power source and the specimen temperature reached up to 1,000°C at about 14 V, 8 amp. The drift of the specimen due to thermal expansion from room temperature to 1 ,000°C was about 0.2 mm. This drift can be easily controlled, as the specimen-shifting mechanism is the same as that of ordinary electron microscope. The difference between the temperature of the specimen and that indicated by a meter was about 50°C at about 600°C, from the test of melting point of Al film. The resolution obtained with this device in operation was determined by measuring the distances of two evaporated particles of Pt-Pd. Fig. 17 shows two serial micrographs taken with the specimen heated at 300°C. o The resolving power is better than 30 A. This proved that such a high resolution is also obtainable at high temperature. Application of Specinieii-IIeatiiig De- vice. One example of its application is shown. Evaporated thin films of Al-Cu (SO- SO) alloy have been used as the specimen Fig. 15. Construction of new specimen heating device, transmission type. '•r>ni''rIICROSCOPY Application of Gas Reaction Device. One oxample of its applicutions of tho old type device is shown (17). Copper mesh is used as a specimen and hydrogen sulfide as reaction gas. Though hydrogen sulfide is a very corrosive gas, the furnace and the surfaces of the pole piece were not remark- ably corroded in the gas atmosphere of 10~^ to 10~' mm Hg at temperatures up to 500°C. At 1 mm Hg, however, the conduction wires of copper and the sylphon bellows which give allowance for the specimen shifting, were covered by corrosion layers after a few hours' operation at about 500°C. Fig. 21 shows some stages of the crystal growth of copper sulfide in an atmosphere of 10-2 mm Hg; (a) was taken at 300°C, (b) at 350°C, 5 min after (a) and (c) at 400°C, 15 min after (b). In (a), a needle crystal below 0.1 /i in breadth, marked by an arrow, has gro\\^i to a finite length within a part of one second. Then, the axial growth became so slow that it could hardly be observed on the screen, whereas a growth perpendicular to the axial direction began. In (c), the crys- tal, marked by an arrow, has grown to 1 ^t in breadth. REFERENCES 1. BoRRiEs, B. VON, Janzen, S., VDI-Zeitsch- rift., 85, 207 (1941). 2. Ito, K., Ito, T., and Watanabe, M., J. Electron Microscopy, Jap., 2, 10 (1954). 3. Ito, K., Ito, T., Aotsu, T., and Miyamae, T., /. Electron Microscopy, Jap., 5, 3 (1957). 4. Menter, J. W., J. Inst. Met., 81, 163 (1952). 5. Honjo, G., Kitamura, N., Shimaoka, K., AND MiHAMA, K., /. Phys. Soc. Jap., 11, 527 (1956). 6. Watanabe, M., Okazaki, I., Honjo, G., AND MiHAMA, K., "Proc. 4th Int. Conf. Elect. Micros.," Berlin (1958). 7. Watanabe, M., Okazaki, I., and Honjo, G., (unpublished). 8. Sears, G. W., Acta Metal., 1, 457 (1953); 3, 361 (1955). 9. Honjo, G. and Watanabe, M., Nature, 181, 326 (1958). 10. Meyer, K. H. and Misch, L., Helv. Chern. Acta, 20, 232 (1937). 11. Ito, K., Ito, T., and Watanabe, M., "Proc. 3rd Int. Conf. Elect. Micros., London," 658 (1954). 12. Okazaki, I., Watanabe, M., and Miiiama, K., "Proc. 4th Int. Conf. Elect. Micros.," Berlin (1958). 13. Hashimoto, H., Tanaka, K., and Yoda, E., J. Electron Microscopy , Jap., 6, 8 (1958). 14. Takahashi, N. and Mihama, K., Acta Met., 5, 159 (1957). 15. Ito, T. and Hiziya, K., /. Electron Micros- copy, Jap., 6, 4 (1958). 16. AsHiNUMA, K. AND Watanabe, M., (un- published). 17. Hiziya, K., Hashimoto, H., Watanabe, ^L, and Mihama, K., "Proc. 4th Int. Conf. Elect. Micros.," Berlin (1958). Masaru Watanabe Kazuo Ito SPECIMEN PREPARATION— SPECIAL TECHNIQUES AT LOS ALAMOS A IVIodified Aluminum Pressing Replica Technique Modifications of an aluminum pressing replica technique for metallic and other suit- able surfaces, first introduced in Germany by Hunger and Seeliger, afford greater yield and wider applicability, which make the method considerably more effective. The necessary pressure within the die is produced by hydrostatic force, transmitted through a medium of hot liquid plastic in place of "cold-pressing" as before. In this way, sen- sitive surfaces can be replicated nondestruc- tively. The method also permits simultane- ous replication of both surfaces of any given specimen and/or similar multiple treatment of several samples in one operation. No addi- tional apparatus is required for application, since the necessary routine equipment, such as metallurgical presses and dies found in most laboratories, may be used without al- terations. The modifications as described therefore extend application of this method to surface studies in powder metallurgy, metal-ceram- ics and related industries. 270 SPECIMEN PREPARATION The technique entails the use of a stand- ard laboratory metallurgical specimen press to produce the necessary pressure to form aluminum foil replicas. The pressure is trans- mitted through a Uquid medium of molten plastic. The loading of the mold is accomplished by the following successive steps : 1. Insert the bottom male die 2. Place a "Lucite" disc on the die 3. Several layers of cut-to-fit filter pa- pers 4. One lead (Pb) foil 5. One thick aluminum (Al) foil 6. One thin aluminum (Al) foil 7. Specimen 8. Filter papers 9. "Lucite" disc 10. Insert upper male die For multiple sample replication, the above steps 2 through 8 may be repeated up to full mold capacity. For two-sided replication of a specimen the loading is repeated in reverse on the opposite side of the specimen. Molding is accomplished by heating the thermoplastic to its melting point, with no pressure applied until this temperature is reached. Pressure is then applied evenly and continuously to maximum, and cooUng commenced at once. The pressure is not re- leased until the mold has cooled to room temperature. The separating filter "circles" prevent contact of the plastic with the specimen and foils. There is, however, circumferential flow along the die edges, hence the finished mold- ing must be separated mechanically. The aluminum foils are then stripped care- fully from the specimen to prevent the for- mation of strain lines in the foil itself. Immediately after removal, the aluminum foils are anodized. The foil side bearing the replica and oxide film is then scribed in squares and stripped as usual by chemical means. The mold components were chosen for the following reasons: 1. Solid discs of thermoplastic instead of molding powders are used in order to form a stable support for the foils and specimens and to alleviate the uneven melting condi- tions inherent in loosely packed porous powders. 2. The lead (Pb) foil is added only as mechanical backing for the multiple alumi- num (Al) foils. This technique as described permits high- speed, high-production rate of specimen replication. Fractured surfaces which prevent strip- ping of routinely applied primary plastic coatings can also be successfully treated in this way. Recent work has shown this method to be effective in preparing "cool" replicas of plu- tonium and its alloys by washing the Al foil replica several times in 10 % IINO3 before anodizing. Fig. 1. Modified Al pressing. Composite mold- ing. (Los Alamos Scientific Laboratory) GMX-1 GT-SITE 271 ELECTRON MICROSCOPY Fig. 2. Modified Al pressing. Fractured, un- poled BaTiOs . 7,000X. (Los Alamos Scientific Laboratory) GMX-1 GT-SITE Examples: Fig. 1: A Composite Molding Fig. 2 : Fractured Sm'face of mipoled BaTiOs , 14,000 X J. H. Bender and E. H. K.\lmus Preparation of Aerosols Fall-out particles from dust clouds, com- monly called aerosols, may be prepared for observation with the electron microscope. Specimens are collected by means of various types of precipitators on membrane filters made from cellulose esters. During the proc- ess of dissolving this material in acetone, particles adhering to the filter surface are transferred to "Formvar"-coated specimen screens. Losses are negligible, residual filter background is eliminated and particles are clearly visualized for measuring. Procedure : 1. Prepare "Formvar-" or carbon-coated specimen screens. 2. Cut membrane filters into small squares to fit the screens. 3. Place three supporting bridges made from coarse and fine brass mesh into a large Petri dish. 4. Load filter-screen combinations into marked compartments on bridge surfaces. 5. Fill Petri dish with acetone up to bridge level. Principle of "liquid film" action for dissolution of the filter material. 6. Make three solvent changes, washing 30 minutes each. 7. Cover dish during and in between steps. 8. After discarding last solvent change, transfer screens onto filter paper for drying over night under vacuum. Examples: Figure 3. Bridge; Figure 4. UO2 aerosol after transfer. E. H. IVALMUS Dispersion of Aerosols Ultrasonic dispersion techniciues for car- bon black and graphite particles, with the equipment available (100 KC to 1000 KC output), left much to be desired in effec- tively breaking up large agglomerates. A simple method of producing artificially airborne particles was devised, for better over-all particle distribution; it involved using a large cardboard box and spraying the inside with "Krylon" to prevent con- tamination by cellulose fibers, sealing the FILTER PAPER SPECIMEN SCREEN " PETRI DISH DISSECTING NEEDLE WATCHMAKERS FORCEPS Fig. 3. Bridge. 272 SPECIMEN PREPARATION ' V Fig. 4. UO2 aerosol after transfer. box, and cutting a small opening in the bot- tom to admit a microscope slide with "Formvar"-coated specimen screens. Procedure for sample preparation : A small amount of the sample is placed in an atomizer, the hand bulb remo\'ed and the atomizer attached to a compressed air hne. The nozzle of the atomizer is inserted into the box opening and a short blast of air applied. Immediately after, the coated screens are placed on the floor of the "cloud chamber" and the door is sealed. Settling times from 20 minutes up to several hours, according to particle size range desired, proved to be suffi- cient. With this techniciue, although the agglom- erates are not completely broken up, par- ticles appear in preferential chain-type ag- glomeration with a good percentage of single layers evident; statistical counting, individ- ual particle shape determinations as well as elect ronmicrography are thus facilitated. Example: Figure 5 carbon black particles 74,000 X. J. H. Bender and J. D. Steely Sorting and Fractography of Particles Particles in the size range of 400 to 325 mesh (37-44 m/i), of interest to powder metal- lurgists, are sometimes coated with a dissimi- lar material to enhance or change the basic physical properties and/or form alloys in the sintering process. Since statistical counting of coated vs. un- coated particles and determination of the coating thickness were not feasible by stand- ard E.M. techniques, the following method was developed : Statistical counting may be accomplished optically by color difference at ^-^1000 X on as-received powders using reflected, polar- ized, full-wave compensated light. Standard samples of the coating and base material are used for reference. Coating thickness may be determined in the following way: 1. Mix proportional amounts of particles and epoxy resin. 2. Cast this mixture in a "pencil-shaped" mold. 3. Cut thin sections from the cured cast- ing with a biological microtome. 4. Direct stripping plastic rephcas can then be taken from the cut end of the cast- ing. 0.1 -" Fig. 5. Carbon black particles. 4O,00OX. (Los Alamos Scientific Laboratory GMX-1 GT-SITE) 273 ELECTKON M l( KC )S( < H»Y 5. Examination of the replicas will show some percentage of particles fractured in a manner that will permit thickness measure- ment of the coating material. 6. The thin sections are used for optical microscopy to correlate E. M. findings. J. H. Bexder and E. H. Kalmus Uncurling Carbon Replicas The carbon repUca technicjue introduced bj^ Bradley is by now well established among elect ronmicroscopists and modifications are probably as numerous as there are micro- scopists using it. A fact familiar to most workers is that sihca and carbon replicas with or without backing have a strong tendency to curl up; straightening out these replicas more often than not becomes extremely difficult and time-consuming, if not impossible. During the course of our experiments we noted that carbon replicas are especially prone to curl if primary plastic coatings are used. It was found, however, that such rolled-up squares will unfold immediately and straighten out by transferring them onto the surface of distilled water. Replicas made in this manner contain few, if any, holes and/or breaks. AV. B. Estill* and E. H. Kalmus STAINING, ELECTRON Electron staining is a technique for en- hancing contrast in the electron images of biological specimens. The most widely used electron stains are aqueous and/or solvent- soluble compounds containing heavy metal compounds. Examples of the most com- monty used electron stains are phosphotung- stic acid, osmium tetraoxide, and uranyl acetate. Some of these compounds, e.g., os- mium tetraoxide, are also used as fixatives. The theoretical basis for the use of heavy metal compounds as electron stains is as fol- * Sandia Corporation, Albuquerque, N. M. lows. Biological materials for all practical purposes consist of the elements hydrogen, carbon, nitrogen and oxygen. These are low atomic number elements, and therefore have low electron density or electron scattering power. Thus, the difference in electron den- sity of a specimen and its plastic supporting film is very slight. Consequently contrast between the electron image of the specimen and the supporting film is low. However, when one treats the biological specimen with a solution of a compound containing a heavy metal element, the heavy metal, or ion con- taining the heavy metal is (1) absorbed on the surface, (2) absorbed into or (3) chemi- cally combined with specific reactive groups of the specimen. In either event elements of high electron density (or electron scattering power) become part of the specimen, thus enhancing the contrast between the electron image of the biological specimen and its sup- porting film. Electron staining reactions may be corre- lated with data obtained by other analytical methods. In the case of collagen an elegant correlation is obtained by comparing electron staining reactions with hydrothermal stabil- ity (shrinkage temperature) data (1). Phos- photungstic acid treatment stains the colla- gen, but has no effect on the hydrothermal stabiUty. From this it may be postulated that the anionic complexes containing tung- sten may combine with certain reactive groups (e.g. — NHs"^ groups of the basic amino acid side chains) but do not introduce cross-links to increase the hydrothermal sta- bility of the collagen. Osmium tetraoxide and uranyl acetate increase the hydrother- mal stability of collagen. Thus it is postu- lated that the collagen reacts with the heavy metal containing ions and in the process cross-links are introduced to increase the hydrothermal stability of the collagen. Lead nitrate has a lyotropic effect upon the collagen. The fibrils are markedly swol- len. Hydrothermal stability is decreased substantially yet the fibrils appear to be 274 STAINING, ELECTRON stained. However, close inspection of the micrographs of lead nitrate-treated colla- gen reveals that there are numerous small particles adsorbed on the surface and depos- ited in the microcrevices of the fibrils. This type of enhancement of contrast may be termed pseudo-electron staining. There is no chemical combination of collagen with metal in pseudo-staining. Figures 1 and 2 compare unstained (i.e., non-treated) and electron stained (phospho- tungstic acid treated) collagen fibrils, re- spectively. Note the greater contrast and enhancement of ultrafine structure in the electron stained material. Figures 3 and 4 compare "negative" elec- tron-stained collagen with electron-stained collagen, respectively. In both instances the stain is a chromium complex. The enhanced contrast of the "negative" electron stain is .j^^^^mmiiiiii BjBIIS?"^^mi ■B *^ 1 ■^ f*^.,."' 1 f ■>» , • Fig. 1. Cowhide collagen fibrils. 31,500X. Fig. 2. Cowhide collagen fibrils. 82,600X. Stained with 0.1% phosphotungstic acid (HaPWioO^- 24H2O). Fig. 3. Cowhide collagen fibrils. 15,400X. Treated with chrome alum (3% KCr (804)2- I2H2O + 10% NaCl) liquor— pH 2.8. Note that the chromium is concentrated along the edges of the fibrils. Fig. 4. Cowhide collagen fibrils. 15,400X. Same preparation as in Figure 3 only after neutraliza- tion with sodium bicarbonate to pH 4.5. Note that the chromium is now incorporated in the fibrils. Fig. 5. Cowhide collagen fibrils. 15,400X. Treated with lead nitrate (Pb(N03)2). The lead particles are distributed evenly on the surface of the fibrils delineating fine periodic structure both parallel and perpendicular to the fibril axis. 275 ELECTRON ^IK.KOSCOPY ^S/* Fig. 6. Cowhide collagen fibrils. 52,500X • Same as Figure 5. due to the fact that the electron-dense chro- mium complex deposited on the plastic film supporting the collagen fibrils, especially along the edges of the fibrils, acts as a dark background for the light fibrils. Figures 5 and 6 show collagen treated with lead nitrate. These are examples of pseudo- electron staining as explained in the text above. REFERENCES 1. BoRASKY, R., J. Am. Leath. Chem. Assoc, 52, 596-610 (1957). 2. Hall, C. E., /. Biochem. Biophys. Cytol. 1, 1 (1955). 3. MuDD, S., AND Anderson, T. F., J. Expt. Med., 76, 103-108 (1942). 4. Porter, K. R., and Kallman, F., Exp. Cell. Res., 4, 127 (1953). 5. Symposium on Electron Staining, J . Roy. Micro. Soc, Series III, 78, Parts 1 and 2 (1959). 6. Watson, M. L., /. Biochem. Biophys. Cytol., 3, 1017 (1957). R. BoRASKY TISSUES (CONNECTIVE), BONES AND TEETH Apart from cancer, most of the non-infec- tious maladies from which the human race suffers are related to changes in the structure or function of the connective or hard tissues. The primary aim of the work which has so far been done on this group of tissues has been to establish a clear picture of their normal appearance to compare with similar preparations of pathological specimens. These tissues, which form the framework of the body, consist mostly of a matrix of long chain polymeric compounds with which their formative cellular elements are associated. In bones and teeth the matrix is made rigid by the deposition of calcium salts. Strictly speaking, the connective and hard tissues are mesodermal in origin, and consist of fibrous proteins embedded in a polysaccha- ride-containing medium. Two ectodermal structures will also be considered in this chapter: tonofibrils, which have as their main function the holding of the epidermis together; and dental enamel, which fits naturally into the group of hard tissues. Collagen and polysaccharide-containing ground substance are universal in the animal kingdom. Leech connective tissue and devel- oping human connective tissue have fibro- blasts and collagen which look identical; while in teeth one has to get as far away from the human as a sea urchin to find calcium salts which are not predominantly hydroxy- apatite. On the other hand, there are minor differences in many closely related species. Thus, some details in the development and calcification of the epiphyseal cartilage of the rat and rabbit, both rodents, show differ- ent appearances. In any study of tissue structure the great- est amount of information is obtained when the electron microscopic appearance is con- sidered in conjunction with findings from other methods. It has not superseded other established techniques, the most important of which is still the light microscope. Some- times it is sufficient for the electron micro- scope to fill in details after the greater part of the work has been done using light micro- scope techniques, involving staining, as in the detailed investigation of the growth of epiphyseal cartilage. At other times, the most important part of the investigation is done using the electron microscope, as in the sorting out of the mechanism of dental caries. 276 TISSLES (CONNECTIVE) However, in all cases mistakes can be avoided when orthodox light microscopy is used as a control. The most obvious example of this has been the work, on elastica, in which many workers who had not used ade- quate controls reported observations on the non-elastic components of the tissues. Other methods which have been used are X-ray and electron diffraction, to identify compo- nents and indicate orientation; injection methods, to trace blood vessels ; microradiog- raphy and direct experiments. The use of light microscopy to prevent mistakes is an example of one of the most important aspects of this type of investiga- tion— the recognition of artefacts. Some are comparatively obvious: for example, the fact that both embedding media and many tissue components are polymers of similar density. The divergence of opinion on whether colla- gen fibrils were solid rods or hollow tubes has been enlightening. Those supporting the first theory left their embedding medium in ; those in favor of the hollow tube theory had removed it. Viewing stereoscopically, which has been found a useful method for detecting a number of artifacts, supports the observa- tions that at least those collagen fibrils which are of large diameter are hollow. In this account, emphasis will be placed on the electron microscopic appearances of the extra-cellular matrices, since a sui'vey of the literature suggests that the present state of knowledge of the cellular components is still very uncertain. One of the few conclu- sions about these cells for which there is a substantial mass of evidence from a number of lines of approach, including the use of labeled precursors (e.g. I. Karpishka, C. P. Leblond and J. Carneiro, Arch, oral Biol. 1, 23, 1959), is that the systems of approxi- mately parallel membranes in the cytoplasm, such as that seen in the cytoplasm of the chondrocyte shown in Fig. 1, are closely concerned with protein formation. First, individual components will be dis- cussed, and then the architecture of the main tissues concerned. In the available space a comprehensive list of references is not fea- sible, so many of those quoted in detail will be important ones likely to be missed in a casual surve}^ of the literature. Except for Figures 5 and 10, photographs are of a white object on a black background. Components Tonofibrils. The surface of the skin is covered with a layer of keratin. Immediately under this are the prickle cells, which, when viewed with a high-power light microscope, appear to be connected by fibrils. When reasonably thick sections are viewed in the electron microscope, so that the slightly wavy fibrils remain in the plane of the sec- tion and are not cut away, it is seen that these tonofibrils do, in fact, pass from one cell to another through intercellular bridges. When viewed stereoscopically, the thickness of the section sho^vn in Fig. 2 is seen to be about equivalent to the depth of three inter- FiG. 1. Section of chondrocyte from primitive cartilage of rabbit. The system of parallel mem- branes is believed to be the site of production of the matrix. Formalin fixed. Embedding medium removed. X6000. 277 ELECTFtON IVIICROSCOPY 9 * ' *'i Fig. 2. Section of human epidermis, showing portion of prickle cell. Tonofibrils pass through the intercellular bridges from one cell to another. Formalin fixed. Embedding medium removed. X 10,000. cellular bridges. What is normally regarded as inadequate fixation has been found useful, since with other cytoplasmic components de- graded and removed, individual fibrils can easily be traced through a series of cells. When entering a cell near the nucleus the fibrils are seen to turn through almost a right angle as soon as they are clear of the cell wall. There has been no suggestion that these fibrils are associated with ground sub- stance, but there is a limited amount of evi- dence suggesting that they might be kerati- nous. Ground Substance. Ground substance is ubiquitous, and may be regarded as the glue that holds all the extra-cellular tissues together. Its characteristic components are polysaccharides. The texture of the ground substance varies from thin tough sheets, as in reticulin (Fig. 3), to a three-dimensional gel, as in articular cartilage. Fig. 4 is a sec- tion of cartilage, necessarily dried out, where the spaces are caused by loss of fluid from the gel. Collagen is never found without ground substance, though the proportion of the two may vary considerably, both be- tween different types of tissue and within any one type. Their association is fundamen- tally that of a mixture and not a chemical compound. The ground substance adheres to the outside of the collagen fibrils with particular tenacity at the bands. In tissue such as the epiphyseal growth cartilage, where the banded structure of the collagen is not very apparent, the 640A spacing is most clearly demonstrated by the ground substance stretching between fibrils (see Fig. 16). Collagen. While the ground substance holds it together, collagen fibers provide the mechanical strength for the framework of the Fig. 3. Fragment of reticulin from human liver. Uranium shadowed. X5000. Specimen prepared by Dr. H. Kramer. 278 I TISSUES (CONNECTIVE) Fig. 4. Section of articular cartilage from rabbit. The thickness of this section is less than the diameter of the thicker collagen fibrils. Two or three are shadowed on their inner surfaces. Em- bedding medium removed. Uranium shadowed. X 10,000. body. In the preceding paragraph it was im- pHed that the ground substance is composed of a group of related polysaccharides. Simi- larly, it is most probable that collagen is made up of a group of related proteins, with the 2.86A spacing of X-ray diffraction pat- terns and a periodicity of the order of (340 A along the fibrils as the common factors. Solubility, shape of fiber and mode of for- mation can vary. The evidence at present available suggests that, like ground sub- stance, collagen is polymerized in the cyto- plasm of cells. It is then precipitated in a fibrillar form either at the surface of the cell, or away from the cell in a medium of ground substance. So far two types of collagen have been dis- tinguished in the human. The form which can be more easily brought into solution oc- curs as very fine fibrils, is most abundant in the infant, and is the predominant form of collagen in tissues such as primitive anlage cartilage and growth cartilage. The second form is the more commonly illustrated one, less soluble, with a tubular appearance of its fibrils and more prominant r)40A spacings. It is the major component of mature tendon. Most tissues such as articular cartilage and osteoid matrix contain a mixture of the two. When specimens of the two types are sepa- rated in a relatively pure form, by making use of the solubility differences, minor dif- ferences are noticed in the wide-angle X-ray diffraction pattern. One type of cell which produces collagen directly is the fibroblast. Various workers have studied it with the electron microscope, using thin sectioning techniques. Fig. 5, taken by Dr. J. A. Chapman and Dr. R. Peach of Manchester University, is a photo- graph of regenerating tendon at 6 weeks. This type of specimen is in^-aluable for show- ing cytoplasmic detail, but is not so well suited for demonstrating the surrounding matrix. Thus, Chapman and Peach do not think that there are collagen fibrils within the cytoplasm, while Fitton-Jackson and others think there may be. Fig. 6 is from a thick section of fibrocartilage containing a cell in which the cytoplasm has disintegrated as a result of delayed fixation. This treat- ment does not affect collagen. Here there is no sign of collagen fibrils crossing the cell wall. At one point this wall is expanded into a very fine net -like structure. This observa- tion is common to fibroblasts from a number 279 ELKCTHON MICKOSCOl'Y of different sites. The collagen fibrils formed by a single fibroblast appear to be tubular and of more or less uniform diameter. Collagen fibrils laid down away from cells Fig. 5. Fibroblast from 6 week regenerating tendon. Fixed with osmium tetroxide and stained with P.T.A. Embedded in araldite. X10,000. Photograph bj^ J. A. Chapman and R. Peach, Manchester University. tend to have a different appearance. Often they are very fine, as in the collagen from epiphy.seal growth cartilage (Fig. 10), and separated from the cells by an even finer mesh of ground substance. A much larger range of fiber diameters is also frequent. In Fig. 14, from a specimen of normal adult ar- ticular cartilage, a collagen fiber of very large diameter is seen in the plane of the section. Elastin. Many papers have been written on the electron micro.scopy of elastin, and much confusing and contradictory informa- tion offered. There is now, however, a meas- ure of agreement. R. W. Cox (Thesis, Oxford 1957) has carefully compared the appearance in the electron microscope with alternate sections stained and prepared for light mi- croscope examination. He has also isolated it chemically. When wet it has rubber-hke elasticity; when dry it is brittle. Elastin is found in those tissues where a quick recovery from applied stresses is necessary, and there is some evidence to suggest that it is pol- ymerized in situ, rather than being precipi- tated in a manner analogous to collagen. Its form and size can be very irregular, and in the electron microscope the safest method of identification is to examine the texture in Fig. 6. Fibroblast from dog intervertebral disc. Delayed fixation with formalin. The cytoplasm is degraded, thus showing that no collagen fibrils have penetrated to cell wall. Part of this wall is ex- panded as compared with the rest. Embedding medium removed. Stereoscopic photograph. 280 TISSUES (CONNECTIVE) Fig. 7. Section of rabbit aorta. Smooth muscle cells and elastic tissue alternating. Note the in- creased density of the elastic tissue at its surface. Osmium tetro.xide fixation. Embedding medium re- moved. No shadowing. X 10,000. sections viewed stereoscopically. In sites such as developing arteries, elastin can be found well away from fibroblasts or any simi- lar cells. For instance, in Fig. 7 elastin is seen on either side of a smooth muscle cell, with no different components in the imme- diate vicinity. D. C. Pease is of the opinion that ground substance is necessary for the formation of elastin, and this view is sup- ported by the fact that in areas such as that sho^^^l in Fig. 8, taken from a section of elastic cartilage, the elastin is confined to one type of surrounding matrix. Fig. 9 is from the loose connective tissue of skin, and here it can be seen that the elastin has tended to engulf some of the surrounding collagen fibrils — a very common occurrence. Calcium Phosphates. The mineral which imparts rigidity to the hard tissues is hydroxyapatite. As far as the characteristics of the individual crystallites are concerned, a great deal of work has produced remark- ably few uneciuivocal results. This is partly because some of the difficulties inherent in the method of examination could not be appreciated until the appearance of such phenomena as dislocations, twin boundaries and total reflection were commonly recog- nized. The crystallites in bone and dentine are very small compared with those in enamel. They are rod-shaped, with diameters of approximately 40-50A, and it is now gen- erally agreed that their average length is of the order of GOOA (T. W. Speckman and W. P. Norris, Science, 126, 753, 1957). The X-ray diffraction results of D. Carlstrom and A. Engstrom are in agreement with these dimensions. Experimentally, enamel crystallites are easier to observe. Even with Fig. 8. Section of luunan epiglottis. Elastin in cartilage matrix. The shadowing metal on the cut surface gives a false impression of density. Osmium tetroxide fixation. Embedding medium removed. Uranium shadowed. X 10,000. 281 ELECTKON iVIICKOSCOPY Fig. 9. Section of human skin. The edges of the elastic tissue have embedded some of the neigh- bouring collagen fibrils. Large spaces are due to the high fluid content of the tissue, Osmium tet- roxide fixation. Embedding medium removed. No shadowing. X 10,000. the low magnification of Fig. 21 they are clearly distinguishable. But they are com- posite crystallites, so that in addition to the apparent detail produced by the physical phenomena already mentioned, there are the real sub-boundaries of a column of bricks. This is demonstrated in photographs such as Fig. 10, taken by R. W. Fearnhead at the London Hospital, in which total re- flection of many of the sub-units can be seen in one photograph but not in the other. The angular difference of these two photographs was 3°. The width of crystalUtes is of the order of 500A (they are usually thinner) and total lengths are upwards of 2000 A. In cartilage and bone the crystallites are first laid down in a haphazard manner bear- ing no relationship to the direction of the collagen fibrils. This has been observed both directly (R. A. Robinson and D. A. Ca- meron, J. Biophys. Biochem. Cytol. Suppl. 2(4), 253, 1956) and indirectly, using other techniques (G. Wallgren, Acta Paediatrica, Suppl. 113). In Oxford, a combination of electron microscopy and electron diffraction has shown that in newly calcified dentine the orientation of the hydroxyapatite crystal- FiG. 10. Crystallites from dental enamel. Total reflection of electrons occurs in different regions in the two photographs, taken at an angle of 3°. X50,000. Photographed by R. W. Fearnhead at the London Hospital. 282 TISSUES (CONNECTIVE) lites is also unrelated to the orientation of the neighboring collagen fibrils. The align- ment observed in mature tissues is appar- ently a secondary rearrangement caused by physical stresses. Tissues Reticulin. This is the tissue that holds the various parts of the body in place. Spread underneath the epidermis is a base- ment membrane. Various structures and organs, and sometimes even individual cells, are surrounded by similar membranes. The thickness varies according to the me- chanical strength required at any given site. The thinnest membranes found in such or- gans as the kidney, spleen and liver are of the order of lOOA, while at the other end of the scale they may be several thousand Angstrom units thick. Several workers con- sider that there are distinct forms of reti- culin. E. L. Benedetti and R. Tiribelli have described two in the glomeruli of kidneys (Arch. Ital. Anat. IstoL, 27, 1954) and A. Bairati and B. Pernis (Boll. Soc. Ital. Biol. Sper. 34(6), 250, 1958) suggest that there is a different form for each tissue with which the reticulin is associated. In sections, reticulin membranes are seen to be sometimes single and sometimes mul- tiple, with the material of an apparently more or less uniform texture. When frag- mented material is treated with alkali, the reticulin is observed to be very resistant chemically, as compared with most other tissue components. Consequently it can be isolated, and individual membranes viewed using transmitted electrons. When this is done, the finer fragments (e.g., those from liver, kidney, spleen, adrenal and lung) which are sufficiently thin to allow electron penetration are seen to contain an almost random two-dimensional network of colla- gen fibrils (H. Kramer and K. Little "Nature and Structure of Collagen", p. 33, 1953). The proportion of collagen to the polysac- charide-containing ground substance varies Fig. 11. Fragment of reticulin from human liver. The positions of some of the collagen bands suggest a non-fibrillar structure. Metal shadowed. X5000. from one site to another. A fragment of reticulin isolated from liver is shown in Fig. 11. Capillary walls consist of a reticulin mem- brane, with cells arranged at intervals on its surface, and occasionally on the inner sur- face. Capillaries may be distinguished from veins or arteries by the fact that cells are not a necessary component part of the wall (that is, after it has been manufactured). There is as yet no definite evidence whether or not capillary walls contain collagen. In Fig. 12 is seen a portion of a capillary wall, the lumen being clearly demonstrated by the presence of the barium sulfate particles used as an injection medium. This injection method is particularly useful when the capillary is passing through loose tissue. The one shown in Fig. 12 is taken from a specimen of the metaphysis of a growing bone. Its wall can be seen to consist of more than one layer. D. C. Pease (./. Histchem. 283 ELECTRON ^IICKOSCOPY Fig. 12. Capillary wall in metaphysis of rabbit tibia. Blood has been replaced by the fine particles of a barium sulphate injection medium. Formalin fixed. X3000. Cijtochem., 3(4), 295, 1955) has studied the glomerular capillaries in kidneys. He found that a fenestrated endothelial sheath was cemented onto the basement membrane of the capillary walls, and in many places found a layer of the cementing material on either side of the membrane. In the sections he examined he found no evidence for a fibrillar structure in the basement membrane. Loose Connective Tissue. It is this tis- sue which is mainly responsible for allowing bodily structures to move relative to one another without friction. It contains nerves and blood vessels; too great a mobility, which might damage these, is prevented by a loose interweaving of bundles of collagen fibrils. These bundles are the collagen fibers observed by histologists using the light mi- croscope. The fibroblasts producing these fibers are commonly observed to form bun- dles of fibrils of uniform diameter. This is seen in Fig. 9. In the foetus and infant the fibroblasts of the loose connective tissue pro- duce fibrils of smaller diameter than those found in mature tissue. The tissue contains much fluid, so that there are many spaces in the dried sections used in the electron microscope. In areas where there are greater stresses, notably the subcutaneous tissues, elastin is also found. Such elastic tissue ap- pears to be laid down after the collagen bundles, and the edges of these are often embedded hi the adjoining elastin (Fig. 9). The elastic "fibers" are of variable size, shape and arrangement. In the Ehlers-Danlos syndrome, in which the skin, and sometimes other tissues, has unnatural mobility, it has been reported (L. H. Jansen, Arch. Belg. derm. syph. 10(3), 251, 1954) that collagen formation is defec- tive, with the fibers less entangled than usual. Unless sufficient elastic tissue is pres- ent to compensate for this, the tissues have a very low tensile strength. Tendon and Ligament. Tendons are a specialized form of compact connective tissue, designed to take a high tensile load. Bimdles of collagen fibrils are tightly inter- woven, with their orientation predominantly parallel to the long axis of the tendon. Un- like the loose connective tissues, it can be seen that a single collagen bundle contains fibrils of different diameters. It has been ob- served that the greater the forces to which a tendon is habitually subjected, the larger and more robust are the individual fibrils. Thus larger fibrils with more marked 640A banding are found in kangaroo tail tendon than in himian Achilles tendon. The latter, in turn, are larger than those in some of the smaller tendons. These visual observations are in accord with the results of solubility experiments on tendons of experimental ani- mals. Solubility tends to vary from one tendon to another. Considerable variation is also observed in the proportion of polysac- charide present in the tendon or ligament. In identically prepared untreated sections, individual fibrils will be clearly distinguish- able in human Achilles tendon, and almost obscured by ground substance in the liga- mentum nuchae. It was in tendon specimens that the tubular structure of at least some collagen fibrils was first established (J. J. Kennedy, Science, 121: 673, 1955). The methods of preparation developed by R. W. G. Wyckoff are particularly well suited for the examination of this type of tissue. 284 TISSUES (CONNECTIVE) Ligaments subjected to lateral stresses are usually found to contain elastin. Thus, while the human ligamentum nuchac has been ob- sei'\'ed to contain predominantly collagen and ground substance, specimens of the ligamentum nuchae from goats, sheep and similar animals with the head held in a for- ward direction are frequently seen to contain more elastin than collagen. In these elastic ligaments, the appearance of the elastin is more genuinely fiber-like than in any other site, excepting the vocal cords. It is also noticeable that, whereas in most tendons collagen fibrils lie practically parallel within their bundles, in elastic ligaments the colla- gen fibrils between the elastic fibers present a wavy appearance — no doubt to allow for a greater extension. Fibrocartilage. Tissues such as the meniscus of the knee and the intervertebral discs are compact connective tissues, with tightly interwoven bundles of collagen pro- duced by fibroblasts. In each bundle, the individual fibrils are approximately parallel, but there is not the same over- all Hnear orientation as is found in tendons. The direc- tions of orientation are related to the forces to which the tissue is subjected. When these are variable, as in the meniscus, the average orientation is very low, whereas in the disc it tends to be parallel to the circumference. Such obsei'vations are best made using X-ray diffraction (W. G. Horton, Thesis, London, 1956). When specimens of more Umited size are examined in the electron microscope the finer details of structure are more apparent. In the bundles there is often seen a distinct tendency for fibrils to be arranged in sheets (Fig. 13), while bundles of fibrils with differ- ent orientation can be glued together by ground substance almost as efficiently as the individual sheets within the bundles. In sites subjected to deformation, the collagenous matrix is usually reinforced by elastic tissue (Fig. 8). The proportion in adult epiglottis is often high. Primitive Cartilage. Cartilage proper — the tissue produced by chondrocytes — is of three main types. Articular cartilage and growth cartilage both develop from a primi- tive cartilage which has cells of the type shown in V\g. 1. Frequently these are in groups of two or four, with adjacent faces flattened against one another. They do not have the prominent cell walls of the fibro- blasts, but have ones which are finer than the nuclear membrane. Most cells or groups of cells are in a capsule appearing, when dried, as a very fine network. The matrix between cells, or groups of cells, in their cap- sules appears as a dried gell containing only very fine fibers. There is no indication that any collagen fibrils are formed at the cell wall, but the appearances are consistent with them having been precipitated in situ. Articular Cartilage. Bones develop from a model, or anlage, of primitive cartilage. In •<2a»«L* >^^'Z: Fig. 13. Section of fibrocartilage from menis- cus of the knee. Embedding medium removed. Metal shadowed. X 10,000. 285 ELECTRON MICROSCOPY its early stages it is not subjected to many stresses. When these do arise — in a joint which is beginning to be used — and when these appHed stresses are in \-ariable direc- tions, articular cartilage makes its appear- ance. This is characterized by less prominent cell capsules and the precipitation of fibers of larger diameter in the hitercellular matrix, such as those shown in Fig. 4 and 14. The greater the load the joint has to bear, the higher the proportion of fibrils of large diam- eter. The cartilage is made up of cells and matrix, but it contains no vessels. The nour- ishment for the cells is probably provided by natural movement causing an intermit- tent pvmiping action which circulates syno- vial fluid through the tissue (J. Trueta). All the observed facts are in accord with this hypothesis (for example, articular cartilage wiped dry and then compressed yields syn- ovial fluid), and it also provides a rational explanation of the appearances of the car- tilage in osteoarthritis. Osteoarthritis occurs whene\'er the various forces acting across a joint become unbal- y^^^'*'*: "i?t » ' jC '^^ . V'f- Fig. 14. Section of noriu:il adult human articu- lar cartilage. Embedding medium removed. Metal shadowed. X 10,000. Fig. 15. Section of adult human articular car- tilage from a case of osteoarthritis. The section has been cut parallel to the surface of the bone. Embedding medium removed. Metal shadowed. X 10,000. anced as a result of injury, unnatural activ- ity, or disease; J. Trueta and his collabo- rators have produced it in experimental animals in an analogous manner, by up- setting local forces acting on the joint. On the microscopic scale the first manifestation is the onset of degradation of the ground sub- stance (an observation which is confirmed by chemical analysis) and a lining up of collagen fibrils normal to the surface of the bone. Fig. 15 shows the appearance of a sec- tion cut parallel to the surface of the bone from a fairly early case where the surface of the cartilage was still almost intact. The amount and direction of orientation is best demonstrated by a combination of X-ray diffraction and electron microscope observa- tions (K. Little, L. H. Pimm and J. Trueta, J. Bone and Joint Snrg., 40B: 123, 1958). In the experimental animals removal of the undesirable forces and tensions was followed by recovery of the cartilage, as judged by both macroscopic and microscopic appear- ances. 286 TISSUES (CONNECTIVE) Growth Cartilage. In regions where forces are predominantly unidirectional the third type of cartilage is found. This growth cartilage is mainly found at the growing ends of long bones. Instead of the cells being in groups of two or four, as in the primitive cartilage, a whole column of cells is seen as a single unit. The cells in the column are flattened, with evidence of frequent division, and the direction of the columns is along the lines of force. Between the columns of cells is collagen, oriented in the same direction (Fig. 16). At the metaphyseal end of the cartilage the cells in the columns expand, to form the hypertrophic cells, which are of the order of 40 jLt across. At the same time that expansion occurs there is normally a change in the ma- trix between the cells. This shows up in the electron microscope as a change of texture, the matrix assuming a more compact appear- ance. It is this altered matrix which can Fig. 16. Section from epiphyseal growth car- tilage of rabbit. The collagen fibrils are parallel to the columns of cells. Formalin fixed. Embedding medium removed. X4000. Fig. 17. Metaphj-seal side of growth cartilage in rabbit. A vessel containing red cells has invaded the last hypertrophic cell space. The vessel wall is not yet complete. The first osteoblasts can be seen between the vessel walls and calcified cartilage. Formalin fixed. Embedding medium removed. X2000. calcify, when the necessary vessels, calciun\ phosphate and vitamins are present. The cells, thus removed from their source of nourishment, die and are replaced by the encroaching vessel, as shown in Fig. 17. Here red cells are within the capsule of a hypertrophic cell, with calcified cartilage on either side. It may be seen that the capillary wall is incomplete — a very common observ'a- tion. (J. Trueta and K. Little, /. Bone and Joint Surg 42B: 367, 1960). Bone. At different stages of its develop- ment bone contains primitive cartilage, ar- ticular cartilage, growth cartilage and calci- fied cartilage; but the greater part of the bone matrix, for most of the life of a man or animal, is the osteoid matrix. The osteoid is laid down by cells which differ in form from fibroblasts or chondrocytes. In Fig. 17 are seen the first osteoblasts, which appear to be on or near the capillary walls, and are sending out processes toward the calcified cartilage. Previously existing calcified tissue appears to be necessary for the complete development of osteoblasts. When mature, these lay do^vn a collagenous matrix. As with fibroblasts, the collagen of bone matrix appears to originate at the cell wall of the osteoblast. The cell processes remain, and lie 287 ELECTRON MICROSCOPY Fig. 18. Sfclion through osteocyte in poorly calcified trabeculum. Processes from the cell can be seen to pass through canaliculi in the as yet un- calcified matrix. Formalin fixed. Embedding me- dium removed. in channels in the matrix — the canaHculi. Observation of sections of areas of active bone formation seem to show that the osteoblasts bury themselves within the matrix they manufacture, whereupon they are referred to as osteocytes. The osteocyte in Fig. 18 has several canaliculi cut in the plane of the section, and processes from the cell can be seen to lie in these. Normally by this stage the matrix would be completely calcified, but the cell chosen for illustration was from an experimental animal with defi- cient calcification, so that only about a quarter of the matrix in the photograph was calcified. The matrix contains collagen fibrils of assorted diameters. R. A. Robinson and D. A. Cameron (/. Biophys. Biochem. Cytol. Suppl. 2(4), 253, 1956) have pointed out that age in itself is not responsible for the character of different types of fibril. They found the fibrils from osteoblasts to have 2-5 times the diameter of those seen in growth cartilage, and to possess obvious periodic structure, both types of collagen having been laid down at about the same time. In the normal formation of bone, the stages appear to be the laying down of colla- gen and ground substance, followed by the deposition of apatite crystaUites in the gelati- nous ground substance. R. Frank (Thesis, Strasbourg, 1957) has examined a case of post-traumatic osteoporosis. X-ray difTraction showed no change in the mineral composi- tion, while in the electron microscope the changes were consistent with the disappear- ance of first the crystallites and ground sub- stance, leaving collagen fibrils prominent, followed by the gradual breakdown of these fibrils. The changes occurred in a patchy and irregular manner. Dentine. As compared with bone, a great deal of time and effort has been expended on the electron microscope examination of den- tine, but the conclusions are by no means unequivocal. Many photographs of carious dentine with bacteria within the tubules have been published, particularly by R. W. G. Wyckoff, D. B. Scott and R. Frank; and M. U. Nylen and D. B. Scott have produced a compilation of photographs of early devel- oping dentine (U. S. Public Health Service Publication No. 613). The odontoblasts differ from chondrocytes and osteoblasts, but have some of the characteristics of each. They are found on the border of the develop- ing dentine, and produce matrix, but unlike the osteoblast they never become completely surrounded by it. They have a single large process which passes through the dentinal tubule, and a number of workers have re- ported small branches from these, which are analogous to the canaliculi in bones. Like chondrocytes, they do not produce bundles of fibers on their surfaces, but the collagen is mainly found between the tubules in a manner which has originally only a small degree of ordering. Fig. 19 shows the arrangement of collagen fibrils round a tubule near the pulp. Vessels penetrate to the odontoblast layer, and cal- cification of the matrix takes place at the point farthest from cells and vessels. This may be compared with calcified cartilage in which calcification is at the end away from the vessels which supply the cartilage with its nourishment. M. U. Nylen and D. B. Scott report a change in the appearance of the ground substance of the matrix which precedes calcification. It has been known for 288 TISSUES (CONNECTIVE) Fig. 19. Section of dentine near pulp. The tubule contents have been removed and the section decal- cified, showing the arrangement of collagen fibrils around the tubule. Embedding medium removed. Stereoscopic photograph. X 10,000. some time (M. Deakins, J. Dent. Res., 27, 429, 1942) that there is a water loss in the matrix immediately before calcification. A changed appearance of the calcified matrix, in response to pressure or caries is often ob- served, and Von Th. Spreter {Schweiz. Med. Woch., 88, 635, 1958) has shown that 12- 13 % dentine by w^eight is a liquor contain- ing adequate amounts of protein, amino acids, sugars and minerals for further pol- ymerization or precipitation. As a reaction to caries, crystallites of the same size as those in enamel have been observ^ed randomly precipitated within the dentine matrix. Near the pulp, the tubules are observed to be completely filled by the odontoblast proc- esses. Towards the enamel, between the matrix proper and the odontoblast process, a 'peritubular zone' has frequently been re- ported. Its nature is as yet unresolved. A good factual description has been given by R. Frank (Arch. Oral Biol, 1, 29, 1959). Dental Enamel. Tooth enamel has a matrix and crystallites which are completely unhke any of the other calcified tissues, doubtless because it is the only calcified ectodermal structure. The matrix consists of two main compounds, which from chemical analysis and X-ray diffraction studies can probably be regarded as members of the keratin group. Some reports indicate that a small proportion of polysaccharide may be present. The prismatic appearance of enamel in sections prepared for the light microscope is due to the arrangement of these two com- pounds. Since the composition of neither is certainly known, they will be referred to here by the names of the workers who first provided an adequate description of each — the less soluble and denser as Pincus' pro- tein, and the less dense and more unstable as Stack's protein. Fig. 20 shows the arrange- ment in a section of developing enamel shortly before the calcification commences. Both proteins form an oriented network, and during calcification apatite crystallites are laid down in pockets of these networks. Cal- cification in Stack's protein occurs shortly before calcification of Pincus' protein at the same level in the enamel. When fully calci- fied, enamel has a very much higher propor- tion of apatite to matrix than is found in any other tissue. There are considerable difficul- ties in cutting thin sections (diamond knives chip easily) and most published photographs of mature enamel are of replicas. When 289 ELECTRON l\IICH()SCOPY lightly etched with acid the Stack's protein dissolves, so that the prismatic structure is enhanced. The chalky enamel of enamel caries (which is known to be the first stage of all carious lesions) can he sectioned more easily, and it is seen that at this stage Stack's protein has gone, and the crystallites in it are being washed away, while Pincus' protein and the crystallites embedded in it remain (K. Little, J. Roy. Micr. Soc, 78, Dec, 1958). Bacteria are not seen in this earliest stage of caries (Fig. 21). This problem of enamel caries pro- FiG. 20. Section of developing human dental enamel, before calcification. Stack's protein, liable to be degraded in dental caries, is the less dense of the two organic components of the matrix. Forma- lin fixation. Embedding medium removed. Ura- nium shadowed. X5000. vides a good example of how, while the ini- tial obsen'ations must be made using elec- tron microscopy, since the important details are too small to resolve with the light mi- croscope, work with the light microscope can then be interpreted by making use of this information, as has been done by A. I. Darling {Brit. Dent. J., 105, 119, 1958). Once it was shown that the initial stage of caries was almost certainly the loss of Stack's protein; further advances are possible using histological methods. For example, if teeth are treated with ethylene diamine, a protein solvent in which apatite is completely in- soluble, sections viewed under polarized light show the translucent appearance typical of the earliest zone of enamel caries. (A. I. Darling and K. V. Mortimer, lADR British Section, 1959). Again, by normal histological methods it can be demonstrated that an im- portant action of fluorine is to modify the formation of enamel matrix (A. I. Darling and A. W. Brooks, lADR British Section, 1959). In normal enamel treated with a for- mic acid solution, sections viewed in the electron microscope show chalky enamel ap- parently identical wdth that formed in natural caries. When fluorized teeth are similarly treated, electron microscope exam- FiG. 21. Section of chalky enamel. Stack's protein has mostly gone, leaving the crystallites which had been embedded in it free to be washed out during processing. The prism walls are intact, with the crystallites still held in place by the organic matrix. Embedding medium removed. Stereoscopic photo- graph. X3000. 290 TRANSMISSION ELECTRON MICROSCOPY OF METALS illation shows that in large areas of the sec- Textures by Means of Thin Sections" (1). tions both components of the matrix remain In this paper Hes the origin of practically- intact. This suggests that fluorine protects all the future developments in the field of teeth from decay by drastically reducing the transmission electron microscopy — the theo- solubility of the organic matrix. retical explanation of the observed effects, the preparation technique and the applica- K. Little ^ions to metal physics, Hke deformation, re- crystallization and precipitation. In Europe TRANSMISSION ELECTRON MICROSCOPY ^Jj^ ^^'^^^ was continued by Castamg (2) on OF METALS- DISLOCATIONS AND ^^l precipitation of CuAl. m Al + 4 % Cu. p___.p._ _._^ Other transmission microscopy was done in Japan by Suito and Uyeda (3), Hashimoto All high resolution electron microscopy is (4), Takahashi et al. (5). made by transmission, but the phrase "trans- Until 1956 essentially the effects due to mission electron microscopy" is used here in perfect crystals and two-phase systems a restricted sense, whereby the electrons (precipitates) were studied, for example, ex- having transmitted a specimen form an im- tinction contours (dark bands arising from age which furnishes information about the Bragg diffraction). A new possibility ap- interior of the material and not merely about peared when it was shown by Hirsch, Home the surface, as in the repHca method. Elec- and Whelan (6) for aluminum, and independ- trons passing through matter can be ab- entty by Bollmann (7) for stainless steel, sorbed or scattered inelastically or elasti- that dislocations could be obsen^ed directly cally. In the latter case the scattering may inside the metal by transmission electron be incoherent or coherent; coherent scatter- microscopy. The theory of dislocations had ing is known as "diffraction". Absorption been extensively developed before. (The and incoherent scattering give "radio- books of Cottrell (8) and Read (9) appeared grams" indicating variations in thickness in 1953). Transmission microscopy furnished and density, but the interesting information direct verification of these theories. In par- justifying a separate article on "transmis- ticular the cinefilms of the group Hirsch, sion electron microscopy" is obtained from Whelan et al. showing the movement of dis- diffraction effects. This means that the speci- locations in aluminum (6) and in stainless mens to be studied have to be crystalline, steel (10) helped to make the dislocation For practical reasons most of the specimens theory appreciated by metallurgists. So studied mitil now have been metallic but transmission electron microscopy today is minerals are in principle not excluded. A an important tool for research in metal phys- further advantage of transmission electron ics. microscopy is that in addition to micros- It is of interest that the first pictures of copy, selected area diffraction can be ap- dislocations by transmission microscopy pUed, which furnishes information about the were pubhshed by Heidenreich in his basic crystal structure and the orientation of the article (1) and that he even mentioned that: specimen. "... a studj'' of the fine details of the con- Certain transmission effects have been ob- tours in sections plastically deformed under served since the beginning of electron mi- controlled conditions may yield important croscopy, but transmission electron micros- information concerning dislocations". But copy in its proper sense started in 1949 with the theory of dislocations at that time was Heidenreich's paper "Electron Microscope not sufficiently developed to give a full inter- and Diffraction Study of Metal Crystal pretation of the pictures. 291 ELECTRON IVIICKOSCOPY The next section will describe the different preparation techniques for obtaining thin specimens. After that, a short introduction to dislocations is given and an enumeration of different metallurgical ai)plications. In this article no work on moir^ patterns (the interference patterns of difTerent superposed crystal layers) will be discussed. We shall refer as far as possible to review papers. The papers of a symposium on thin film techniques are published in the August 1959 issue of the Journal of the Institute of Metals (11). A review of Japanese work on electron metallography including transmission work is given in the booklet "The World through Electron Microscopes (Metals)" (12). Note added in proof: Since this article was sent in, many new papers have appeared, most of which are collected in the Proceedings of the Delft Con- ference 1960 (72). Preparation Techniques To be traversed by 100 keV-electrons a crystalline specimen has to be thinner than about 5000 A, and to get information about the interior, the surface has to be as smooth and clean as possible. When these conditions are fulfilled, the information in the pictures depends only on the internal state of the specimen and its orientation. Specimens can be prepared in two main ways: A. they can be built up directly as a thin foil or B. they can be cut out of a block and thinned. Class A covers specimens produced by vapor condensation in vacuum, e.g., on cold or hot rocksalt or by electrolytic deposition. A method, where a metal ring is dipped into liquid metal to produce a metal skin, anal- ogous to a soap skin, has been developed by Takahashi et al. (13). The evaporation technique has been perfected to a high degree by D. W. Pashley et al. (14). While the specimens of class A are used mostly to study the behavior of the thin films themselves, specimens produced by a method of class B are considered as repre- sentative of the bulk material. Most of the thinning techniques are based on the electro- polishing method introduced by Heidenreich (1). Heidenreich clectropolished a mechan- ically thinned disk in a special holder, pro- tecting the edges, branched as anode against a pointed cathode, first from one side to take away the mechanically damaged surface layer and then from the other side until the first tiny hole broke through. In the sur- roundings of these holes, the specimens were thin enough to be penetrated by the elec- trons. Castaing (2) added to the electro- polishing an ion bombardment treatment with the ion gun built directly into the electron microscope, and was able to control this treatment to obtain the optimum condi- tions. This technique is especially useful for studying precipitation, e.g., CuAU in Al, where the aluminum is preferentially dis- solved by electropolishing, while the pre- cipitates are thinned by the ion bombard- ment. While Hirsch and his co-workers (6) studied beaten Al foils annealed and chem- ically etched in hydrofluoric acid, Bollmann (7) continued on the line traced by Heiden- reich. The modifications to Heidenreich's arrangement were that (a) the electrolytic attack was symmetrical from both sides of a specimen, (b) instead of stopping the attack when the first small holes appeared, two large holes were produced and the attack was stopped when these holes joined. The symmetrical attack has the advantage that the specimen is a symmetry plane of the potential distribution, thus the current is not especially concentrated at the edges of a hole and does not round them off, so that the edges remain sharp. The point (b) has the advantage that the right moment to stop the treatment can be foreseen. The best specimens are found normally at the point where the two holes join, and may be cut out after cleaning in water and methyl alcohol. It is even possible to continue the 292 TRANSMISSIOX ELECTRON MICROSCOPY OF METALS attack after this and obtain other fairly good material, especially the distribution of dis- specimens. Surveys of electrolytic polishing locations, the question arises how far this methods, with detailed information con- situation has been affected by the prepara- cerning the polishing solutions and condi- tion. Our impression is that the effect on the t ions for various metals and alloys, are given distribution of dislocations is small, though by TomUnson (15), and Kelly and Nutting minor local changes, as for example straight- (16). Other variations, such as the "Window ening of dislocations or a weak relaxation Method," the "Figure of Eight Alethod," of piled up groups (Fig. 19) may take place, and the "Uniform Field Method" are also It should be said that the dislocations usually described in reference 16. are pinned at the surface of the foil and that Another technique was introduced by dislocation networks are very stable. An- Mirand and Saulnier (17), who treated other case arises when dislocations start to specimens, first mechanically thinned down move after a certain irradiation or after to 0.04 mm, in a commercial elect ropolishing heating or stressing. Here the situation apparatus for metallographic purposes (Disa- changes and the velocities of movements of Electropol). The specimens were attacked dislocations, as filmed by the group Hirsch, alternately from both sides over a diameter Whelan et al. (6, 10) are expected to be quite of about 5 mm and then pohshed do\\ai until different in thin foils from those in the bulk nothing remained of the original area, except material. for a few small flakes of metal swimming A preparation technique for semicon- in the solution; this was subsequently de- ductors, such as germanium and silicon by canted and the flakes were rinsed several chemical etching has been developed by times and then fished out with a specimen Innng (18). For different crystal faces dif- grid. A certain danger of this method may ferent etchants have been used (e.g. 1 % be that a flake, having remained a short time sodium hypochlorite solution for the (111)- in the polishing solution without current can face of germanium). Also bismuth telluride become etched, but this will depend on the has been chemically etched by Geach and metals and solutions. Phillips (19). There are a lot of possible electropohshing A completely different way of obtaining techniques but normally it takes some time thin metal specimens is microtome cutting to master any one of them. It is not im- with a diamond knife, introduced by Fer- portant which way one originally goes, but nandez-Moran (20). Metals have been cut it is very important to see from a result in in this way by Haanstra (21), Tsuchikura which sense the conditions have to be cor- and Ichige (22) and Reimer (23, 24). This rected. For electropohshing it is essential to method can be useful for obtaining informa- know that pohshing, i.e., the non-specific tion on different phases but the metal is uniform attack, is only achieved at high strongly distorted by the cutting process, current density in a relatively small range. At low current density the attack is metal- Dislocations lographic which means that grain boundaries Definition. In a perfect ideal crystal all etc. are specifically attacked. On the other the atoms are arranged in a periodic lattice hand when the specimen shows a crust on structure. After a plastic deformation of the the surface or a spotty attack, the current crystal this periodicity is disturbed. For density or the temperature of the bath is too reasons of energy content, the disturbances high. of the lattice are concentrated locally in fines As the electropofished specimens are used while the great part of the lattice is periodic to study the internal situation of the bulk as before, except that it is elastically 293 ELECTRON MICKOSCOPY ^^^ i<" Fig. 1. Schematic representations of disloca- tions, (top) in edge-orientation (center) in screw orientation (bottom) changing from edge to screw orientation. strained. Such localized line defects of the plastically deformed lattice are called "dis- locations". Schematic representations of dislocations are shown in Fig. 1. The ma- terial in the core of the dislocation line, where the crystal order is destroyed, is called "bad" material; the crystalline material around the core even if elastically strained is called "good" material. From the geometry of crystal lattices it follows that dislocation lines cannot end inside the material. They have to be closed loops or must end at the surface of the crystal or at grain boundaries in poly crystalline material. In transmission electron microscopy of crystals the contrast is essentially given by diffraction effects which are described by Bragg's law 2d sin d = n\ where d is the spacing of lattice planes, 6 the diffraction angle, X the wavelength of the electrons, and n the order of diffraction (integer). Dislocation lines become visible because the lattice is strained around the line core which means that d varies locally and thus the Bragg condition and consequently the distribution of the electron intensity between the direct and the diffracted beam varies. As the diffracted beam does not contribute to the image, because it is elimi- nated by the aperture diaphragm, the in- tensity of the image is given by the intensity of the direct beam only. A survey of the theory of the contrast due to dislocations and stacking faults etc. is given by Whelan (26). Fig. 2 shows pictures of dislocations as they can he seen by transmission electron microscopy. As dislocations cannot end in- side the material, the ends of the lines lie on the top or the bottom of the foil. A ciuantitative characteristic of a disloca- tion line is the so-called "Burgers vector" which is defined in the following way : First the dislocation line is given an arbitrary direction. Then, around the line a circuit is marked in the sense of a right-hand screw far enough from the core of the dis- location to be always in "good" material. The circuit has to be traced in such a way that it would be closed in a perfect crystal Fig. 2. Dislocations lines in stainless steel, crossing the foil top to bottom and traversing a twin boundary. 294 TRANSMISSION ELECTRON MICROSCOPY OF METALS (as many steps to the right as to the left, etc.). The vector pointing from the end to the beginning of the circuit is the "Burgers vector" b. Examples are shown in Fig. 1. The sign of the Burgers \-cct()r has only a meaning in connection with the direction of the dislocation line. Changhig the one also changes the other (right-hand screw!). Frank has given another definition of the Burgers vector, where a closed Burgers circuit in a dislocated crystal is imaged onto an ideal crystal where the closure failure there is the Burgers vector. The result is essentially the same but the procedure is academically more correct. The Burgers vector in magnitude and Fig. 3. Interaction of dislocations of different orientations but the same Burgers vector in stain- less steel during cold-working. (Whelan^^ Courtesy Royal Society) bj.b^ »^ Fig. 4. Interinlion of dislocations forming a network during cold-working. {Whelan,^^ Courtesy Royal Society) direction is constant along a dislocation line, even when the line changes its direction (Fig. Ic). That part of a dislocation Une perpendicular to its Burgers vector is said to be in an "edge" orientation, that parallel to its Burgers vector in a "screw" orienta- tion. A dislocation line can be curved and follow all intermediate stages between edge and screw orientations. Dislocations can interact and form nodes. The law here is that the sum of the Burgers vectors of all dislocations entering the node is equal to the sum of the Burgers vectors of those emerging from the node (analogous to Kirchoff's law). Complicated network can be formed, of which examples are given in Figs. 3, 4 and 5. A detailed analysis of dis- location interactions in stainless steel has been given by Whelan (29). Movement. Essentially two kinds of movement of dislocations have to be dis- tinguished— -"glide" and "climb"; these are schematically shown in Fig. G and 7. Glide is a movement parallel to the Burgers vector, while climb is perpendicular to it. Glide does not need material transport, as it is the movement of a configuration analogous to a 295 ELECTRON MICROSCOPY Fig. 5. Hexagonal network forming a twist boundary between two (lll)-planes during poly- gonization in aluminium. (Hirsch, Home and Whelan,^ Courtesy Philosophical Magazine) wave, while climb does. Glide is produced by a shear stress, while climb needs hydrostatic pressure or tension. Glide is essentially re- sponsible for cold working, while climb con- tributes to high temperature creep because diffusion is only possible at high tempera- tures. The glide of dislocations is shown in the films of the group Hirsch, Whelan et al. (6, 10) on aluminum and stainless steel (Fig. 8). The specimen is locally heated by the elec- tron beam inside the electron microscope. The thermal stresses produced by this heat- ing perhaps in connection with the surface contamination through oil vapor act so as to move the dislocations. This movement can be slow because of pinning of the disloca- tions at the oxide layer on the surface or it can be so quick that only the slip traces left by the moving dislocations mark their path. Wilsdorf (30) and also Berghezan and Foudreux (31) have studied the movement of dislocations due to externally applied stresses. A moving dislocation cuts the volume into two domains, one of which is displaced (dis- located!) with respect to the other. Which one is displaced and how much can be recognized in the following way. (vX I) = m where v is the velocity vector of the dis- location, / a vector in the direction of the GLIDE W 4^iH~ ^\Tr^ » \U \ / / Fig. 6. Glide movement of a dislocation in edge orientation {Courtesy Schweizer Archiv) CLIMB I 7 t7. D i t 1 Fig. 7. Climb movement of a dislocation in edge orientation {Courtesy Schweizer Archiv) 296 TRANSMISSION ELECTRON MICROSCOPY OF METALS (a) (C) Fig. 8. Sequence ul ilic lilm mi .-^i;(lllle8.s -steel .showing the ghde nioveinent of dislocations. The black dots in the center are marks on the fluorescent screen of the microscope. The object has been displaced during observation. {Whelan, Hirsch, Home and Bollman^" Courtesy Royal Society) line in the defined line direction, m is a vector which points to one of the two do- mains. This indicated domain is displaced with respect to the other by the Burgers vector b when the dislocation passes. This rule can be verified for the cases in Fig. 1. Thus glide steps on the surface of a material are produced by the sum of the displace- ments due to the individual dislocations; these have been studied at the edges of thin fihns by Hirsch et al. (32) (Fig. 9). Forces Between Dislocations. The energy of a dislocation, i.e., the work that has to be done to produce a unit length of a dislocation line is proportional to h- {h = Burgers vector). This shows that a disloca- tion always tends to attain the smallest possible Burgers vector. The energy E of a dislocation with a Burgers vector (2 6o) would be twice the energy of two separate and sufficient far apart dislocations with Burgers vector ho each: (2 6o 2( bo' + bo'). Two parallel dislocation lines, the Burgers vec- tors of which include an angle smaller than Fig. 9.- Glide steps on stainless steel. The speci- men was etched only from one side. {Hirsch, Par- tridge and Segall,'^^ Courtesy Philosophical Maga- zine) 90° repel each other. If this angle is larger than 90° they attract each other. (The 90°- case corresponds to the theorem of Pytha- goras!) A detailed analysis of the stress field betv.-cen dislocations shows that the forces 297 ELECTRON MICROSCOPY can be noncentral ones. Such problems are discussed in the previously mentioned books on dislocation theory, (8, 9, 25) and the repulsion and attraction of dislocations are Slacking Order in Close Packed Structures A B C A ABA Cubic Face Centered . A B CA B C Hexagonal Close Packed ...A B A B A S. . Fig. 10. Stacking order in close packed struc- tures. STACKING SEQUENCES Orientation (001) <110> ■(112> Cubit: Face Centred K C K B K A K C K B K A K C K B K A fW -tt-Jtt mA ABCABCABCAB / \ Cubic Sequence [^ ABC/ orCBAv / H!«S9 Sequence [-j BAB\ or ABA Hexagonal close packed H B H A H B H A H B H A H B H A H B H A ABCABCABCAB Fig. 11. Stacking sequence in the face-centered cubic and hexagonal close-packed structures. H B ^ -O Central Symmetry Fig. 12. Symmetry of nearest neighbour atoms in relation to the stacking energy. directly observable in the films (G, 10) (Fig. 8). Stackiiij^ Faults and Partial Dislocations As is well known, a cubic face-centered lattice can be understood as a close-packed structure of spheres. On a basic layer in position A a second layer can be placed in two different ways, in position B or C (Fig. 10). If the position B is chosen for the sec- ond layer there remain C and A for the third one and so on. A stacking sequence . . . ABCABC . . . builds up a face-centered cubic lattice, a stacking sequence . . . ABAB . . . a hexagonal close-packed lattice. The fact that a given material crystallizes in a certain structure, e.g. face centered cubic, shows that once the first two layers A and B are ready, the work to build up the third layer must be lower for position C than for position A, because if there would be no difference the stacking sequence would be random (Fig. 11). Looking at all connec- tions to nearest neighbors of an atom in the second B layer one sees that these connec- tions (directions of binding forces) are of central symmetry in the f.c.c. structure (ABC sequence) and of mirror .symmetry in the hexagonal c.p. structure (ABA sequence). Fig. 12. Thus in a f.c.c. lattice the ABC (CBA, BCA, etc.) sequence with central symmetry marked by K has a lower energy (needs less work to be built up) than the ABA (CAC. BCB, etc.) sequence marked by H. ' Different kinds of faults can occur in the stacking sequence; here we consider only the f.c.c. case. A single H inserted into all K means a growth fault (twin boundary; Figs. 13a and b). Two consecutive H mean a so-called "deformation stacking fault" (Figs. 14a and b). Thus to a first approximation a deforma- tion stacking fault has twice the energy of a twin boundary (Fig. 12). The Burgers vector of a (total) disloca- tion line normally is the shortest connection 298 TRANSMISSION ELECTRON MICROSCOPY OF METALS between two atoms in the lattice surrounded by atoms in the same symmetry. In a f.c.c. lattice it is of the type (a/2) (110) (connec- tion between the corner to the centre of the face in the unit cell). At the same time it is the connection between two spheres in a close packed plane (Fig. 15). When a dis- location moves between a layer A and a Twins (cub. f.c. ) ABC ABC ABC AB Fig. 13a. layer B, the atoms in the layer B have to jump from the starting position to the neighboring position the Burgers vector h away. This jump can be achieved in two steps Bi ^ C and C — > B2 . C in this case is not a lattice site but the site occupied when a stackhig fault is at this place. Such a splitting of a total dislocation into two "partial dislocations" or "partials" is de- scribed by the equation in Fig. 15 for the Burgers vectors. A partial dislocation is not a dislocation line in the usual sense but the boundary of a stacking fault. As the Burgers vectors of the two partial dislocations include an angle smaller than 90° they repel each other. By separating, the two partial dislocations have to do work to produce a stacking fault in between them. WTien the stacking fault energy is high, the stacking fault ribbon between the two partials becomes narrow and vice versa. The splitting of dislocations into stacking fault f-r^- Fig. 13b. Fig. 13. Twin boundary in stainless steel. (Whelan, Hirsch, Home and Bollman^'> Courtesy Royal Society) 299 ELECTRON :MI(.K<>sr.OPY Stacking Fault (cub. f.c.) ABCABCABCA B (a) ****■» f- -tmt' m^ ^ ^g^dmibHsk (b) Fig. 14. (a) Schematic representation of a stacking favilt. (b) System of .stacking faults in cobalt on all 4 (lll)-planes. These stacking faults stabilize the observed region in the cubic phase at room temperature although it normally would be hexagonal . ribbons can be seen in the film on disloca- tions in stainless steel (10) (Fig. 16). Dislocations glide especially easily on certain planes (the so-called glide planes) which in the case of the cubic face centered lattice are (lll)-planes (close-packed planes perpendicular to the space diagonal of the unit cell). If a dislocation ribbon is narrow, it can easily change from one glide system to the other, when the Burgers vector is parallel to both glide planes. This move- ment is called cross-slip (Fig. 17). It has been observed in aluminum, which is a metal with high stacking fault energy and thus with narrow dislocation ribbons. If the dis- location ribbon is wide, the Burgers vector is composed of the two vectors of the partial dislocations, both lying in the glide plane. To change the glide plane, the ribbon has to be contracted and to dissociate again into two new partial dislocations lying in the new glide plane. In a metal with low stacking fault energy like stainless steel, this is only possible at a serious obstacle as for example a twin boundary (Fig. 18) but another dis- location on the same slip plane is not suf- ficient to introduce cross slip so that series of dislocation originating from the same source become piled up against each other (Fig. 19) and thus form piled up groups. Contrast Due to Stacking Faults A theory explaining the observed fringes from stacking faults has been given by WTielan and Hirsch (33, 34) on the basis of the kinematical and dynamical theory of electron diffraction. These authors showed SPLITTING OF A DISLOCATION LINE INTO PARTIAL DISLOCATIONS b, = "b, + t, |[,oi>i[n2].i[2iil C A B C ^A \ A A / \ / B — C — ■» B / / / / / / / ,'-/-'/ / ',.../.// / V A ->•,> \/' / /•--/-' / A y / / / Fig. 15. Separation of a total dislocation into partials. 300 TRANSMISSION ELECTRON MICROSCOPY OF METALS Fig. 16. Splitting of dislocations in stainless steel into stacking fault ribbons {Whelan, Hirsch, Home and Bollman^", Courtesy Royal Society) that the fringes are shifted by half a period, when two stacking faults are present, and disappear completely when three or a mul- tiple of three stacking faults follow each other (Fig. 20). Condensed Vacancies A theory of Kuhlmann-Wilsdorf (35) con- cerns the condensation of vacancies, when a metal is quenched from a high temperature. These features have been observed by Hirsch ei al. (36). In aluminum the vacancies condense in a close packed plane as small disks, which collapse. Such a collapsed disk would form a stacking fault surrounded by a so-called sessile partial dislocation with a Burgers vector of the type (a/3) (111). As the stacking fault energy of aluminum is high, the disk does not collapse perpendicu- lar to the plane but inclined, adding a ghssile partial dislocation (a/3) (111) + (a/6) (112) = (a/2) (110) thus forming a disloca- tion ring (Fig. 21). As the Burgers vector Fig. 17. Cross slip trace in aluminium (Silcox (in 27) Courtesy Institute of Metals) sticks out of the plane this dislocation ring can only glide on a cylinder parallel to the Burgers vector. Under shear stress, the dis- location ring expands over the cylinder, one part gliding downward, the other upward. In a metal with low stacking fault energy, like gold, the disks of condensed vacancies collapse reallj^ to a stacking fault, but to 301 ELECTRON MICROSCOPY lower the energy even more this extends over In other cases the interactions between all 4 (lll)-planes, thus forming small tetra- vacancies and dislocations produce helical hedra of stacking faults (Silcox and Hirsch dislocations as observed by Thomas and (37)), (Fig. 22). ^ATielan (38) (Fig. 23). ,23^ SS78^ O.SfJ ^^kem'- Fig. 20. Supt. : pu^cu .-lu^i^u.j; :.»...;.-. {Whelan and Hirsch,^- Courtesy Philosophical Magazine) Fig. is. Dislocation lines traversing a twin boundary {Whelan, Hirsch, Home and BoUmann,^" Courtesy Royal Society) Fig. 19. Groups of dislocations piled up against a twin boundary (Whelan, Hirsch, Home and Boll- mann,^^ Courtesy Royal Society) Fig. 21. Dislocation loops surrounding col- lapsed disks of condensed vacancies in quenched aluminum. {Hirsch, Silcox, Smallman and West- inacott,^^ Courtesy Philosophical Magazine^ 302 TRANS-MISSION ELECTRON MICROSCOPY OF METALS Fig. 22. Tetrahedra of stacking faults in quenched gold. (Silcox and Hirsch,^ Courtesy Philosophical Magazine) Some Metallurgical Topics^ Studied by Trans.imssioii Electron Microscopy Transmission electron microscopy is now used to a large extent to get a deeper insight into metallm-gical effects. Here only a short and incomplete hst can be given of work achieved in this field. This kind of research is in full development. Work hardening finds its explanation in the networks of dislocations which hinder the movement of further dislocations. As the movement of dislocations means plastic deformation, a hindrance of this movement means increased resistance against plastic deformation, i.e.. hardening. Metals with high stacking fatilt energj* like almninum work-harden less than those with low stack- ing fault energy like stainless steel, because of the possibUity of cross slip. Herein be- longs a lot of work on dislocation reactions (6, 10, 29). Thomas and Hale (39) investi- gated the relation between surface structure and underh-ing dislocations due to deforma- tion. Quench hardening is explained by the formation of dislocation loops (35, 36) or stacking fault tetrahedra by condensed vacancies (37). The interaction of disloca- tions and vacancy clusters has been dis- cussed by Wilsdorf and Kuhlmann-Wilsdorf (40^. Fatigue has been studied by Segall and Partridge (41). In aluminimi dislocation loops of condensed vacancies and compli- cated shapes of dislocation fines were ob- sen-ed which indicate the importance of vacancies in this process (Fig. 24). In stain- less steel (41) ven*' narrow sfip bands and extrusions and mtrusions were seen. Obser\'ations on recovery and recrystalliza- tion have been made by Fujita and Xishi- jama (42), and Berghezan and Foudreux (43) on alununum by Saulnier and Develaj'' on titanium (44\ b^' Bollmann (45) on nickel (Fig. 25) and by Bailey (46) on silver. These studies have given an insight into the poly gonizat ion and the formation of new cri'stal grains in hea\ily rolled material as well as into recr\'stalfization bv the move- ...«gt '^^ I .^€ Fig. 23. Helical dislocations produced by the condensation of vacancies on dislocation lines. {Thomas and Whelan,^ Courtesy Philosophical Magazine) 303 ELECTRON IVIICKOSCOI'Y Fig. 24. Dislocation arrangements in fatigued aluminum: loops of condensed vacancies and ir- regular dislocation shapes resulting from climb. (Segall and Partridge,'^^ Courtesy Philosophical Magazine) Fig. 25. Crystal grain containing twins grow- ing in incompletely polygonized surroundings in nickel. {Bellman,*^ Courtesy Institute of Metals) ment of old grain boundaries at low de- formation. A study on the formation of precipitates in Al-Ag alloy has been made by Funkano and Ogawa (47). A review on precipitation phenomena has been given by Nicholson, Thomas and Nutting (48), (Fig. 26) who took a cine film showing the formation and dis- solution of precipitates. Eutectic structures have been studied in SnPb by Takahashi and Ashinuma (49) and on pearlite by Nishiyama et al. (50). Studies of Silcox and Hirsch (mentioned in (27)) on radiation damage in copper show the production of condensed vacancy loops and the growth of these loops with higher neutron doses (Fig. 27). Wilsdorf (51) has shown similar effects in neutron irradiated nickel. Anti-phase domains in a CuAu-alloy and in AuCuZn-alloy have been studied by Ogawa et al. (52, 53) and in CuAu by Glossup and Pashley (54). CuAu II shows a super- lattice, with layers of copper and of gold atoms stacked on each other in the (001) direction while the sequence changes every five elementary cells in the (100) direction, so that phase boundaries are regularly spaced every 20 A. The formation of these domains was directly observed whilst the specimen was heated (Fig. 28). This work is reviewed by Pashley and Presland (55). Observations on martensitic transforma- tions in thin films of iron alloys have been published by Nijama and Shimizu (56) and K -li.* ^ >> mm. 4 1 , ^^ ^ .. Fig. 26. Guinier-Preston zones in Al-4%-Cu. 304 TRANSMISSION ELECTKO.N AIICKOSCOPY OF METALS by Pitsch (57). Pitsch suggested a trans- formation mechanism for thin foils and gave evidence for the mechanism in bulk material by a combination of transmission microscopy and diffraction. Thin metal films and their properties have been studied for the case of condensed layers by Takahashi and INIihama (58) for an Al-Cu alloy. The recrystallization was followed, while the object was heated on a hot stage inside the microscope. Twins in condensed layers of silver, gold, copper and nickel (59) as well as electrodeposited nickel (60) have been investigated by Ogawa et at. Reimer made similar observations on electrode- posited metals (61) and condensed silver (62). Studies on the formation of condensed layers (63) and the mechanical properties such as crack propagation within films are reviewed by BassettandPashley (64). Beaten gold (65) and precipitated gold flakes have been studied by Briiche et al. (66). Brame and Evans (67) studied the deformation of thin films on solid substrates. Studies on different alloys are summarized by Saulnier and Mirand (68). Observations on semi-conductor material as germanium. **p. V<:j-r'- ^•>'1--JC "/■ ■; y'. =1;- ' Fig. 27. Loops due to condensed vacancies in neutron irradiated copper. (Silcox (in 27) Courtesy Institute of Metals) Fig. 28. CuAu II in the stage of formation. {Pashley and Presland,^^ Courtesy Institute of Metals) silicon (69, 70) and bismuth telluride (19) have been made by Geach, Phillips and Indng. Surface together with the underlying structure have been investigated for the case of deformation by Thomas and Hale (39) and for studying an etching structure of aluminum by Phillips and Welsh (71). This kind of research is possible by electro- polishing only one side of the specimen, while the other is protected by a varnish. These few examples mRv show that trans- mission electron microscopy is a powerful means for the study of defects and precipi- tates in crystalline sohds, the knowledge of these defects being of great importance for the understanding of the physical behavior of crystallized matter. REFERENCES 1. Heidenreich, R. D., J. Appl. Phys., 20, 993 (1949). 2. Castaing, R., "Proceedings the Third International Conference on Electron Microscopy" (London 1954) p. 379. 3. SuiTo, E. AND Uyeda, N., Proc. of the Japan Academy, 29, 324-330 (1953). 4. Hashimoto, H., /. Phys. Soc. Japan, 9, 150- 161-(1954). 305 ELECTRON IMICROSCOPY 5. Takahasiii, N., Takeyama, T., Ito, K., Ito, T., MlHAMA, K., AND WaTANABE, M., J. Electromnicroscopy, 4, 16-23 (1956). 6. HiRSCH, P. B., HOBNE, R. W., AND Whelan, M. J., Phil. Mag., 1, 677 (1956). 7. BoLLMANN, W., Phys. Rev., 103, 1588-1589 (1956); "Proc. Stockholm Conference on Electron Microscopy," 1956, p. 316-317, Stockholm (1957). 8. CoTTKELL, A. H., "Dislocations and Plastic Flow in Crystals," (Oxford, 1953). 9. Read, W. T., Jr., "Dislocations in Crystals," New York (1953). 10. Whelan, H. J., Hirsch, P. B., Horne, R. W., and Bollmann, W., Proc. Roy. Soc. A240, 524-538 (1957). 11. /. Inst. Metals, 87, 385-458 (Aug., 1959). 12. "The World through Electron Microscopes," Japan Electron Optics Laboratory Co. Ltd., (Tokyo, 1959). 13. Takahashi, N., Ashinuma, K., and Wat- anabe, M., J. Electronmicroscopy , 5, 22-27 (1957). 14. Pashley, D. W., Phil. Mag., 4, 316 (1959). 15. ToMLiNsoN, H. M., Phil. Mag., 3, 867-871 (1958). 16. Kelley, P. H. and Nutting, J., /. Inst. Met., 4, 385-391 (1959). 17. MiRAND, P. and Saulnier, a., Compt. rend. 246, 1688 (1958). 18. Irving, B. A., Brit. J. Physics (in press). 19. Geach, G. a. and Phillips, R., 4. Int. Kon- gress f . Elektronenmikroskopie (1958) Berlin 1960, p. 571-574. 20. Fernandez-Moran, H., J. Biophys. Biochem. Cytol.2 (suppl.),29 (1956). 21. Haanstra, H. B., Philips Techn. Review, 17, 178-183 (1955). 22. TsucHiKXJRA, H. and Ichige, K., Nature, 181, 694 (1958). 23. Reimer, L., Z. Metallkunde, 50, 37 (1959). 24. Reimer, L., Naturwiss., 46, 68 (1959). 25. Friedel, J., "Les Dislocations," Paris, 1956. 26. Hirsch, P. B., Metallurgical Reviews. 27. Hirsch, P. B., /. Inst. Metals, 87 (12), 406-418 (1959). 28. Whelan, M. J., /. Inst. Metals, 87 (12) 392- 405 (1959). 29. Whelan, M. J., Proc. Roy. Soc, A249, 114-137 (1959). 30. Wilsdorf, H., 4. Int. Kongress f. Elektronen- mikroskopie (1958) Berlin 1960, p. 559-562. 31. Berghezan, a. and Foudreux, A., Compt. rend., 248, 1333-1335 (1959). 32. Hirsch, P. B., Partridge, P. G., and Segall, R. L., Phil. Mag., 4, 721 (1959). 33. Whelan, M. J. and Hirsch, P. B., Phil. Mag., 2, 1121, (1957). 34. Whelan, M. J. and Hirsch, P. B., Phil. Mag., 2, 1303 (1957). 35. KuHLMANN-WiLSDORF, D., Phil. Mag., 3, 125- 139 (1958). 36. Hirsch, P. B., Silcox, J., Smallman, R. E., AND Westmacott, K. H., Phil. Mag., 3, 897 (1958). 37. Silcox, J. and Hirsch, P. B., Phil. Mag., 4, 72 (1959). 38. Thomas, G. and Whelan, M. J., Phil. Mag., 4, 511 (1959). 39. Thomas, K. and Hale, K. F., Phil. Mag., 4, 531 (1959). 40. Wilsdorf, H. G. F. and Kuhlmann-Wils- DORF, D., Phys. Rev. Letters, 3, 170-172 (1959). 41. Segall, R. L. and Partridge, P. G., Phil. Mag., 4, 912-919 (1959). 42. FujiTA, H. and Nishiyama, Z., Metal Physics (Japan) 3, 108 (1957) (in Japanese) (to be published in English). 43. Berghezan, A. and Foudreux, A., Compt. rend., 247, 1194-1196 (1958). 44. Saulnier A., and Develay R., "Symposium de metallurgie speciale Saelay" (1957). 45. Bollmann, W., J. Inst. Metals, 87, (12) 439- 443 (1959). 46. Bailey, J., Thesis, Cambridge University, 1959. 47. Funkano, Y. and Ogawa, S., Acta Cryst., 9, 917 (1956). 48. Nicholson, R. B., Thomas, G., and Nutting, J., J. Inst. Metals, 87, (12), 429-438 (1958). 49. Takahashi, N., and Ashinuma, K., J . Inst. Metals, 87, 19-23 (1958-59). 50. Nishiyama, Z., Kore'eda, A., and Shimizu, K., J . Electronmicroscopy, 7, 41-47 (1959). 51. Wilsdorf, H. G. F., Phys. Rev. Letters, 3, 172- 173 (1959). 52. Ogawa, S., Watanabe, D., Watanabe, H., AND KoMODA, T., J. Phys. Soc. Japan, 14, 936-941 (1959). 53. Ogawa, S., Watanabe, D., Watanabe, H., AND KoMODA, T., Acta crystal., 11, 872-875 (1958). 54. Glossop, a. B. AND Hashley, D. W., Proc. Royal Soc. A250, 132-146 (1959). 55. Pashley, D, W. and Presland, A. E. B., /. Inst. Metals, 87 (12), 419-428 (1959). 56. Nishiyama, Z. and Shimizu, K., Acta Met., 7, 432 (1959). 57. Pitsch, W., J. Inst. Metals, 87 (12), 444-448 (1959). 58. Takahashi, N. and Mihama, K., Acta Met. 5, 159 (1957). 306 UNSOLVKD rUOBLEMS 59. Ogawa, S., Watanabe, D., and Fujita, E., /. Phys. Soc. Japan, 10, 429 (1955). 60. Ogawa, S., Mizuno, J., Watanabe, D., and Fujita, E., J. Phys. Soc. Japan, 12, 999 (1957). 61. Reimer, L., Z. Metallkunde, 47, 631 (1956). 62. Reimer, L., Optik, 16, 30 (1959). 63. Fashley, D. W., Phil. Mag., 4, 324 (1959). 64. Bassett, G. a. and Pashley, D. W., J. Inst. Metals, 87 (12) 449-458 (1959). 65. Bruche, E. and Schulze, K. J., Z. Physik, 153, 571-591 (1959). 66. Bruche, E. and Demny, J., Z. Naturfor- schung, 14a, 351-354 (1959). 67. Brame, D. R. and Evans, T., Phil. Mag., 3, 971-986 (1958). 68. Saulnier, a. and Mirand, P., Revue de I'Ahiminium, 266, 687-697 (1959). 69. Geach, G. a., Irving, B. A., and Phillips, R., Research, 10, Oct., 1957. 70. Phillips, R., /. Inst. Metals (July 1958) p. 72 (Bull. Vol. 4). 71. Phillips, R. and Welsh, N. C, Phil. Mag., 3, 801 (1958). 72. The Proceedings of the European Regional Conference on Electron Microscopy Delft 1960 (Delft 1961). W. BOLLMANN UNSOLVED PROBLEMS The following comments are germane only to applications of the electron micro- scope and related microscopy in non-bio- logical areas. In general, it appears that the electron microscope has been used in the biological sciences more as a research tool than as an analytical servdce instrument. Quantitative Particle Size Distribu- tion Data. This is in principle a very im- portant application and yet there are es- sentially no accurate studies of this kind in the available literature. Size and shape as- says of paint pigments, magnetic powders, semiconductor powders, etc., are of funda- mental importance particularly where the peak of the size distribution curve is of the order of several hundred angstroms. The color or tone of duplicating media is often related to the Mie theory for particle scat- tering which in turn uses the sixth power of the particle (sphere) radius. Also the sur- face area and surface to volume ratio are important to properties of semiconductors such as photo- and dark conductivity. Ultrathin Sections. Multilayer coat- ings such as magnetic tapes and photosensi- tive copy papers exhibit properties which are dependent on the distribution and orientation of particulate solids within a binder. Thin sections normal and parallel to the plane of the coating are needed, but to date little progress has been made in pre- paring sections suitably thin for electron microscopy. Unlike the biological tissue sections, the difference in hardness of the two or more components in the sample militate against the preparation of sections of the order of 500 angstroms thick. How- ever more effort is needed in this area. Somewhat related to this problem is the inability of x-ray microscopy to provide useful information in such studies of coated samples. Overlapping of particles makes interpretation difficult, or smallness of size with respect to the resolving power of the instrument makes interpretation impossible. For the most part the significant applica- tions of the x-ray microscope have been con- fined to metallic and biological materials. Replicas. Again, it is very difficult to pre- pare surface replicas of certain coatings such as magnetic tape and photosensitive papers. The inertness and insolubility of the binder impairs isolation of the replica by dissolu- tion. H3^drophyllic stripping layers have not proved to be successful. In many cases, despite the inertness of the binder, the effects of heat and pressure, solvents, or the radia- tion and vacuum attendant to evaporation techniques make the isolation of good ciuality replicas difficult, or they may vitiate the results. Nondestructive studies are often obviated by the inadequacy of the stripping layers. A few remarks are in ordoi- in regard to the responsibility of industrial management to the microscopist. First, where feasible the 307 ELECTRON ^IICKOSCOPY management should establish two distinctly tack of the surface by reactive compounds or different groups: a service group and an en- corrosive media and the subsequent rubbing tirely separate exploratory research group, or breaking away of the reaction products. The two groups carry out radically different The various wear processes may occur singly, functions which need no further defining. It simultaneously or in sequence. Progress in is unwise to mix the two. wear prevention can be made only after a Second, it is imperative that the person- real understanding of the basic mechanisms nel in these groups be different. Specifically, has been achieved. Toward this end micro- those people in the service group should be scopes and ancillary eciuipment are probably adept at preparing specimens, maintaining the most useful and widely used tools avail- and operating the instrument, cognizant of able to research workers and investigators the limitations of their techniques, and cer- (1). tainly conversant with a broad spectrum of The topography, chemical and physical people with whom they have to deal. In nature of surfaces can influence the wear short, they should be internal consultants, behavior of many materials, thus the starting On the other hand, those in the basic re- point of any investigation should be a study search group, having their own instruments, of the surfaces involved. Many techniques should be primarily chemists, physicists, are available to examine the topography of biologists, metallurgists, etc. They should be surfaces but optical methods allow the speci- left alone to "get lost" in their particular men to be investigated without destroying long-range programs. Communication be- it or subjecting it to strain or wear. Stylus tween the two groups should be encouraged methods provide an immediate numerical in the interest of mutual assistance. characterization of a surface, but may be As to the cost of duplicating equipment it subject to the uncertainty of the extent of may be said — and not facetiously — that damage done to soft materials by the stylus only money is needed to purchase the itself. The profile microscope (Tolansky (2)), instruments. Intelligent, productive people like stylus methods, gives a single line profile cannot be so easily obtained, whatever the and it may be necessary to observe many available funds may be. In short, equipment different recordings in order to form a com- is a comparatively inexpensive investment. prehensive conception of the surface. Ob- Franklin a. Hamm serving an enlarged image of the surface from above by normal microscopic examina- tion is useful, but permits no conclusion WEAR AND LUBRICATION ,. ' ^ru ■ . ^■ regardmg roughness, the mterierence micro- Wear may be defined as the displacement scope on the other hand can be used to re- or removal of material from a surface ; it has veal and evaluate minute surface irregulari- many forms, its nature and magnitude being ties. In the image these appear as deviations dependent on a variety of factors. The more in the bands or fringes which represent con- significant of these are speed, load, surface tour lines connecting all points of the same condition, metallurgical structure, the pres- level (Fig. 1). The difference in level between ence of abrasive matter and corrosive en- different fringes corresponds to half a wave- vironment. The wear process can be classified length of the light used. The magnification generally into two main types — mechanical is irrelevant and the geometric shape of the and chemical. Mechanical wear involves test piece plays only a subordinate role, processes which may be associated with In wear studies, suitable selection of the friction, abrasion, fatigue and vibration, method of specimen illumination can extend whereas chemical wear is caused by an at- the versatility and range of the modern 308 WEAR AND LUBRICATION (a) Normal illumination. Fig. 1. Surface of a worn gear tooth. (After Scott') 50X (b) Interference micrograph showing rippled na- ture of the surface. microscope. A particular surface feature may be emphasized by a specific form of illumina- tion. Some surface irregularities may be re- vealed more easily by dark field or oblique illumination, the irregularities which diffract the Hght appearing light on a dark back- ground. Phase-contrast microscopes may be used to advantage to examine wear damage and surface deformation as phase contrast is very sensitive to changes in surface levels and can reveal minute slip lines and fine surface irregularities hardly visible or quite invisible under normal illiunination (Fig. 2). The use of polarized light can supplement information obtained with phase-contrast and interference techniques. In order to examine microscopically se- lected surface areas of large, unwieldy speci- mens or parts of machines without disman- tling the machine or seriously interrupting its use, replica techniques may be used to advantage (1). Replicas may be rendered more highly reflecting by coating in vacuo with a metal such as aluminum. To study both surface and sub-surface changes brought about by rubbing, sliding, rolling or corrosive action it is usual to cut a cross-section for microscopic examination ; if the specimen is sectioned at an acute angle to the plane of the surface, a further mag- nification factor is obtained of dimensions perpendicular to the original surface. Taper- sectioning (Moore (3)), carried out by grind- ing at a predetermined small angle to the plane of the surface following suitable plating and mounting of the specimen, is useful in studying changes of surface profile and sub- surface metallographic changes (1, 4, 5). To those engaged in a detailed study of wear and the changes caused by rubbing con- tact, knowledge of the physical properties of individual crystals, aggregates and surface films, etc., is essential, and micro-hardness testing technique (Taylor (6)) may be used in conjunction with microscopic and metal- lographic examination. It is often important to determine the extent of work hardening on a surface, the depth and hardness of surface films, as well as to elucidate metallographic changes produced by the wear process (Fig. 3). The success of microscopic and metal- lographic investigations of surface and sub- surface changes depends largely on the 309 ELECTKON MICROSCOPY r :^ ♦^.^^-^ ..---^" \ '"'I % %. %, v^ (a) Normal illumination. (b) Phase-contrast showing nature of the surface Fig. 2. Stylus trace on an anodized aluminium surface. (After Scotti) lOOX edges of specimens and surface films from damage during preparation for microscopic and metallographic examination, and speci- mens which have a non-conducting surface may be successfully plated if a thin film of metal is first deposited in a high vacuum. For ease of manipulation specimens may be mounted in a suitable plastic. Although much information can be derived from the use of optical microscopy, as the initiation of surface damage may be expected to occur on a sub-microscopic scale a more detailed surface examination is sometimes required and the electron microscope can be used to follow paths of exploration to higher degrees of resolution (9). Examina- tion may be made directly by reflection (Halliday (10)) or by preparing surface replicas which are examined in transmission. The first method is limited to the use of specimens sufficiently small to be accommo- dated in the microscope, while the latter method although occasionally subject to artefacts has the advantage that the replica may be taken from worn surfaces in situ (Figs. 4 and 5). The transmission electron microscope has successfully been used for studying the initiation of wear and fretting corrosion (4, 9). Fig. 3. Taper section of a worn surface showing the surface contour, sub-surface metallographic changes and indentations of a micro-hardness sur- vey with V.P.N, superimposed. Etched Nital. (After Scotti) H = 140X, V = 1400X careful use of metallurgical techniques of sec- tioning, polishing and etching. It is essential that the methods used do not change, obliter- ate, or destroy in any way the evidence being sought. Besides the conventional methods, spark erosion machining (Nosov and Bykov (7)) offers many advantages for harder materials and specific applications while electropolishing (Jacquet (8)) is use- ful. Electroplating is used to protect the 310 WEAK AND LUBRICATION Fig. 4. Electron micrograph of a positive car- bon replica of a finely ground steel surface (1212 micro-inches C. L. A.) 5000X Fig. 5. Direct reflection electron micrograph showing a scratch on a mechanically polished sur- face finished with finest abrasive paper. 2500X Debris from worn surfaces and abrasive particles which have been separated from the lubricant by centrifuging (9) may be conveniently examined in the electron micro- scopes by standard methods (Fig. 6). Such examination may reveal the size, shape and possibly the abrasive natiu'e of the debris while micro-diffraction may be used for identification. Both optical and electron microscopy can be used together with metal- lographic methods to elucidate changes effected by deformation and strain. Elegant extraction replica techniques (11, 12) are available which in conjunction with electron diffraction can be used to determine pre- cipitation segregation while the recent micro- probe anal.ysis technique (13) can be applied to elucidate solution segregation especially on a micro-scale. In prevention of wear, the principal task of a lubricant which maj^ be liciuid, solid or gaseous is to enable one solid surface to move over another wilh low friction and with a minimum of damage. This object can be achieved if the lubricant film is thick enough to keep the surfaces apart and hydrodynamic conditions prevail. However for these condi- tions to be realized loads must generally be low and sliding speeds high. If such ideal conditions cannot be maintained other forms of lubrication such as boundary or mixed film may exist. Where this is so the surfaces come into contact and damage in the form of wear occurs (14, 15). Useful information concerning the be- havior of lubricating oils can be obtained by microscopic investigation of the insoluble carbon with which they become contami- nated in practice (Matthews et al. (16)). Electron microscopy can also be used to study the additives which are incorporated in modern lubricating oils to impart or re- inforce some desirable property. The elec- FiG. (i. Electron micrograph of debris from fretting corrosion damage of a steel surface. Shad- owed at 450° with gold-palladium. lO.OOOX (After Scott and Scott^) 311 ELECTRON ^MICROSC.OI'Y Fig. 7. Electron micrograph of particles in a used (10,000 miles) heavy duty detergent crank- case oil. GOOOX >-' ^»>^ ^ v^rJ -/^ ■^ *" •*II* Vw »>« k f^ ^ .f *. P'iG. 8. Reversed print of an electron micro- graph of a shadowed "Formvar" replica of a worn steel surface lubricated with a mineral oil contain- ing an E. P. additive. 13,500X Iron microscope has been successfully used to study the solubility and dispersivity of de- tergent additives (17, 18) and also, by em- ploying special specimen techniques, to compare contaminants in used plain and detergent oils (19) (Fig. 7). The so-called extreme pressure (E.P.) additives are principally lead or other heavy metal soaps and organic compounds of sulfur, phosphorus and chlorine used in various combinations to prevent destruc- tive metal-to-metal contact between rela- tively moving surfaces. E.P. additives react chemically with the metal of the surface to form a svu'face film which has a lower shear strength than the metal itself, and acts as a boundary lubricant. The electron microscope has provided experimental evidence of this (4) (Fig. 8). _ Conventional lubricating greases contain soap fibers or other particles as thickening agents. As the structure of the thickening agents largely determines the physical char- acteristics of the bulk grease it is often im- portant to examine them microscopically. The optical microscope using polarized light and the phenomenon of birefringence has been used to show the orientation of fibers in grease due to shear (20) as fibers are crystals in the true sense of the word. Knowing how they orientate helps to elucidate grease struct lu-e. As the finer fibers of lubricating greases are beyond the resolving power of optical microscopy the electron microscope is the only tool at present available for studying the changes which soap fibers undergo when a grease is subjected to mechanical work, heating, etc. (21). The successful application of electron microscopy to the study of grease structure depends upon a satisfactory method of specimen preparation which can give reproducible results free from artefacts. The oil phase must be removed and the method of removal must not disturb the fiber arrangement. Solvent techniques (20, 22), bulk freezing and slicing (23) and vacuum evaporation (21) have been suc- cessfully applied to achieve this (Fig. 9). Solid lubricants are used where other lubricants are unsuitable due to tempera- ture limitations or inaccessibility. The mechanism of lubrication is similar for all lubricants, a thin film supporting the sur- faces is sheared during relative movement. Solid lubricants usually have a layer lattice structure and present planes on which shear may easily occur but which are resistant to compression at right angles to the plane. Their success depends on the size and shape 312 WEAR AND LUBRICATION (a) Unused grease. (b) Fine fibres in a mechanically worked grease. Fig. 9. Electron micrographs of fibers in lime-base greases. Oil phase removed by vacuum evapora- tion 20,000X Fig. 10. Transmission electron micrograph showing size and layer nature of graphite solid lubricant. 7500X of the dry powders and the method of bond- ing on to the surface. Control of all these fac- tors can be assisted by microscopic and elec- tron microscopic examination (Fig. 10). REFERENCES 1. Scott, D., (Instn mech. Engrs) "Proc. Con- ference on Lubrication and Wear" Paper 57, 670-673, 1957 (London). 2. ToLANSKY, S., Nature, 169, 445 (1952). 3. Moore, A. J. W., Proc. Roy. Soc. 195, 231 (1948). 4. Milne, A. A., Scott, D., and Macdonald D., (Instn mech. Engrs) "Proc. Conference on Lubrication and Wear" Paper 97, 735- 741 (London) (1957). 5. Welsh, N. C, ibid., 77, 701-706 (1957). 6. Taylor, E. W., /. Inst. Met., 74, Part 10, 493 (1948). 7. Nosov, A. V. AND Bykov, D. V., "Working of Metals by Electro-sparking" (translation from Russian, H. M. Stationery Ofiice, London), 1956. 8. Jacquet, p. A., Inst. Met. Metall. Rev., 1, Part 2, 157 (1956). 9. Scott, D. and Scott, H. M., (Instn mech. Engrs) "Proc. Conference on Lubrication and Wear" Paper 14, 609-612 (1957) (Lon- don). 10. Halliday, J. S., (Instn mech. Engrs) "Proc. Conference on Lubrication and Wear Paper 40, 647-651 (1957) (London) . 11. Fisher, R. M., "A.S.T.M. Symposium on Techniques for Electron Microscopy," 49 (1953). 12. Smith, E. and Nutting, J., Brit. J. Appl. Phys., 7, 214 (1956). 13. Melford, D. a. and Duncumb, P., Metal- lurgia, 57, No. 341, 159-161 (1958). 14. BowDEN, F. p. and Tabor, D., "The Friction and Lubrication of Solids," Clarendon Press, Oxford, 1950. 15. Ming Feng, I., Lubric. Engg., 10, 34 (1954). 16. Matthews, J. B., et al., J. Inst. Petrol., 39, 429 (1953). 313 ELKCTKON MKHOSCOPY 17. BRorr.HTKN, J. L., el aJ., Sci. Lubric, 4, No. 12, 23 (1952). 18. McBrian, R., A.S.M.E., Paper 52-A40 (1952). 19. B.\RWELL, F. T., Grunberg, L., and Scott, D., (Instn mech. Engrs, London) "Proc. Auto Div." No. 5, p. 153 (1955). 20. Brown, J. A., Hudson, C. N. and Loring, L. D., Inst. Spokesm. nat. Lubric. Gr. Inst. Vol. XV, No. 11, p. 8, 1952. 21. Milne, A. A., Scott, D. and Scott, H. M., (Instn. mech. Engrs) (London) "Proc. Conference on Lubrication and Wear," Paper 45, 450-153 (1957). 22. Renshaw, T. a., Ind. Eng. Chem., 47, (4), 834 (1955). 23. Vold, M. J. and Vold, R. D., J. Inst. Petrol., 38, No. 339, 155 (1952). D. Scott WILSKA LOW-VOLTAGE MICROSCOPE The first successful experiments with early electron microscopes showed that almost all the objects that were known from light microscopy appeared more or less like sil- houettes. Even bacteria were too thick to be seen through. With higher accelerating voltage the electrons gained more penetrat- ing power. At about 200 kV the bacteria became fairly translucent but because of the difficulties of shielding and regulating volt- ages of this range, the usual laboratory elec- tron microscopes have an anode potential between 50-100 kV only. Such a potential gives a good penetrability for objects of a thickness from a few hundred to a few thousand Angstrom units. The resolving power of these instruments may be as high as 10 A when measured using ob- jects with heavy atoms. Most biological objects consist of atoms of relatively light weight, and of these practically no image is formed even if their size is several times this optimum resolution hmit. The use of heavy- atom stains or relief shadowing in vacuo increases the contrasts to some degree; yet most of the structures in the 10-50 A region have remained unseen. The need of more contrast became clear about ten years ago. Electrons of 50 kV gave no hope of revealing structures that one wanted to see. For this reason experi- ments were begun with slower electron speeds, down to 25. 10, and even 5 kV. On the fluorescent screen it was easy to see how the contrasts were dramatically improved. Holes that were hardly observable on the transparent collodion membrane at 50 kV were surrounded by a pitch-black area caused by the same membrane at 5 and also at 10 kV. The images were too unstable to be photographed with sharpness, however. Low-voltage electron beams are considerably influenced by disturbances of magnetic stray fields as well as impurities on the walls of the column and at the edges of aperture holes. These impurities become charged by the electrons and affect the beam in an ir- regular manner. Similar charges on the photographic emulsion itself may affect a low-voltage image much more than a high- voltage one. For these reasons it was necessary to build an entirely new instrument to meet the extra demands of low-voltage work. The present electron microscope is the fifth in the series of trial models of "home-made" low-voltage electron microscopes. All its predecessors and the main parts of this instrument have been built in Finland. The construction of the present electron microscope in the United States is now practically completed, too. The instrument is probably the smallest of its kind, the col- umn measuring only 13 inches from the cathode to the fluorescent screen. It has four electromagnetic lenses: condenser, ob- jective, intermediate lens, and projector. Withhi the objective lens circuit there is an electromagnetic stigmator consisting of eight minute coils. The alignment is purely electro- magnetic. Both tungsten and oxide-coated platinum cathodes can be used. To protect against damage by positive ion bombard- ment, there is an ion trap between the anode and the cathode. To prevent negative 314 WILSKA LOW-VOLTAGE MICROSCOPE charges from accumulating on the photo- graphic emulsion during the exposure, there is a tiny generator of positive ions in the image chamber. The change of specimen is both simple and rapid. The specimen cap is inserted into the hole on a rod-like support. Before entering the high vacuum the specimen passes through a fore-vacuum zone. The little amount of air in the specimen cap escapes into the fore \'acuum, and the high vacvmm remains practically intact. For this reason there is no waiting when changing the speci- mens. The microscope can be operated in any position between vertical and horizontal. Tilting does not disturb the image on the screen. Visual observation of the image can be made more accurate by using a magnify- ing lens or an auxiliary light microscope, the latter of which can be put in place in 2-3 seconds. The use of a special fine-grain viewing screen in the electron microscope makes optical magnification of 50-100 times practical. At low voltages there is very Httle of the ''splashing" of scattered and second- ary electrons that would seriously limit the possibilities of similar high optical magnifi- cation at 50-100 kV. For the same reason a high photographic enlargement of low volt- age electronic images is made possible. The 35 mm film used is non-perforated, and about 160 pictures can be made without re- loading the camera. Although ultimate refinements of the instrument are still being made, numerous kinds of specimens have been examined with it, using 18 kV as an accelerating po- tential. The main difference between images obtained with this voltage and the usual 50- 100 kV is the strength of contrasts. Many structures that are too thin or too weak when viewed with the usual voltages are much better visualized with this lower voltage. On the other hand, many specimens that are ideal for 50 to 100 kV are quite too thick for our instrument, a circumstance applying to supporting membranes as well. Fortunately, it is easy to make membranes that cast very Httle shadow even at 18 kV. It is somewhat more difficult to make sec- tions thin enough with the present ultra- microtomes. At the present, the resolving power of the microscope is about 25 A which is quite good considering the greater wavelength asso- ciated with the lower velocity of the elec- trons used. To quote Dr. Frank N. Low: "Many biologically significant macromolecules such as those of hemoglobin and pepsin possess a size well within the resolving power of current electron microscopes, but have never been visualized because of contrast problems. A low voltage electron microscope is able to produce images of such macromolecules without the previous use of heavy-metal shadowing or any structural alteration of the specimen such as staining. This should open up a vast field of research in the visualization of organic substances at the macromolecular level of structural organization." As encouraging as results have been so far, the real era of low-voltage electron microscopy will first come when its supe- riority in contrasts can be combined with a resolving power of only a few Angstrom units. This will enable observation of the elementary shapes of most of the biological structures, e.g., polypeptide chains. There is no hope of doing this without first correcting at least a part of the tremendous spherical aberration inherent in all existing electron lenses. Alvar p. Wilsk.\ 315 Electron mirror microscopy In electron mirror microscopy the elec- trons do not penetrate the specimen as they do in conventional electron transmission microscopy nor do they originate from the specimen as they do in the different cate- gories of electron emission microscopy. In electron mirror microscopy the electrons are not reflected by the material of the speci- men, as in conventional electron reflection microscopy, but are reflected without any scattering at the equipotentials in front of the specimen. The characteristic feature of an electron mirror microscope is an electron optical mir- ror playing a dual role. The mirror serves as an electron optical element and simultane- ously constitutes the specimen. In general, the mirror-specimen remains untouched by the image-forming electrons because the specimen, being an electron optical mirror, is C G L N S B, B EO M H B, B2 EO M Fig. 1. Schematics of two design versions of electron mirror microscopes. biased slightly negative with respect to the source of the electrons, that is with respect to the cathode. The electrons approaching the mirror-specimen are slowed down to zero axial velocity shortly before reaching it and then reverse their direction of motion. Any irregularity on the mirror-specimen capable of deflecting the approaching and receding electrons will modify the spatial density distribution of the returning electron beam. This returning electron beam, there- fore, carries back information about the dis- tribution of the irregularities it encountered during the time it was close to the mirror- specimen. Electron mirror microscopy uti- lizes the fact that such information can be presented on a phosphorescent screen as a magnified, visually observable image of the distribution of the electron-deflecting irregu- larities on the mirror-specimen. The specific design characteristics dis- tinguishing an electron mirror microscope from other types of electron microscopes stem from the introduction of an electron optical mirror and its dual function. The reversal of the electron beam on the equi- potentials in front of the mirror specimen poses the problem of either magnetically separating the incoming, illuminating elec- tron beam from the outgoing image-carry- ing beam or tolerating the inconveniences of a simpler straight design in which the axes of both beams remain identical. Figure 1 shows simplified schematics of these two possible design versions of elec- tron mirror microscopes. The one version (la) uses the straight design (1) in which the two beams are not separated. The incoming illuminating electron beam Bi originating from a cathode C in an electron gun G and coUimated by an electron lens L is shot through a small neck N located in a hole in the center of the viewing screen S. After traversing the electron optical system EO 316 ELECTRON ^IIRROK MICROSCOPY the electron beam is reflected on the eqiii- potentials ui front of the specimen M which is generally biased a few tenths of a volt negative with respect to the cathode C and acts therefore as an electron optical mirror. Every electron-deflecting irregularity on the mirror-specimen whether caused by the topographical relief structure or by differ- ences in contact potentials, in surface charges, in electrical conductivities or in magnetization, influences the low velocity electron beam. The electrons returning through the electron optics EO, which may consist of electric or magnetic lenses, project a magnified pictorial representation, a kind of shadow-schlieren image, of the irregulari- ties on the specimen M onto the viewing screen S. The other design version (2) of an elec- tron mirror microscope is shown in Figure lb. Here the incoming illuminating electron beam Bi is separated for part of its path from the returning image-carrying beam B2 by a magnetic field H normal to the plane of drawing. (Separation of the two beams by simply deviating from normal "incidence" of the electron beam on the mirror-specimen would not be feasible because the mirror- specimen necessarily constitutes a part of the electron optical system, in the vicinity of which axial symmetry must be retained in order to avoid intolerable aberrations.) The auxiliary magnetic field H and the resulting knee-shaped design necessarily complicate the design and construction but result in two basic advantages. First, it leaves the viewing screen unobstructed and, second, it permits the introduction of separated lenses to act on the electron beam before en- tering and after leaving the auxiliary mag- netic field H. This is advantageous because these separated lenses can be optimized in- dependently for their different tasks. The one lens can be designed for optimal per- formance of the incoming, illuminating beam whereas the other lens can be operated for optimal image formation and magnification. In the case of the less complicated straight design, the entire electron optics acts on both the incoming electron beam as well as on the returning, image-carrying beam so that the electron optics cannot be of optimal design for either piu'pose but must be a compromise. The capabilities of electron mirror micros- copy are numerous and varied. In general, every electron-deflecting field caused by an irregularity on the mirror-specimen can be depicted because it is the electron deflection which forms the image contrast. Electron mirror microscopy permits, therefore, visual observation of the distribution of such purely electrical properties as contact potentials (1), surface charges, space charges (2) and elec- trical conductivities (1, 3). It also proved to be a feasible method for the visual obsen^a- tion of magnetic patterns (4), particularly of magnetic domain patterns. Electron mirror microscopy is also capable of depicting the surface relief structure (2, 5) on a specimen because this structure is retained in the structure of the equipotentials immediately in front of the mirror-specimen. Figure 2 illustrates this latter possibility. It shows an electron mirror micrograph of a cleaved surface of a LiF single crystal (5) onto which a thin Pt film had been evap- FiG. 2. I'^lcctron mirror micrograph of a cleaved surface of a LiF single crystal. 317 ELECTRON MIHKOK MICKOSCOPY oratod to make Ihc suiiace conductive. De- piction of the topography of a specimen is certainly not a unique feature of electron mirror microscopy and, in many cases, other microscopic methods are available which are either more convenient, such as optical mi- croscop}^, or have considerably higher re- solving power, such as conventional electron transmission or reflection microscopy. In spe- cial cases, however, electron mirror micros- copy might be utilized quite advantageously for the observation of the surface reUef structure on a specimen because the image- forming electrons do not penetrate, nor even impinge on the specimen and thus no speci- men heating problems exist. Furthermore, no recourse to replica techniques is neces- sary; the resolving power, although not ap- proaching that of electron transmission mi- croscopy, is somewhat better than that of light microscopy, and, perhaps most im- portant, the elevation differences or step heights necessary to form sufficient image contrast are extremely small. The step height of a single barium stearate monolayer (24.4 A) can, for instance, be ob- served with such pronounced contrast that Fig. 3. Noise-like and wave-like electric charge movements on amorphous selenium film. Single frame of a motion picture taken from the screen of the electron mirror microscope. Magnification ap- pro.x. 23 X considerably smaller step heights should cer- tainly be observable. A genuine and some- what imique feature of electron mirror mi- croscopy is, however, its capability of depict- ing purely electrical and magnetic patterns. Besides depicting contact potential distribu- tions or surface charge distributions, it can, for example, also depict the electrical con- ductivity distribution of layers of very low conductivity above conductive substrates. This can be achieved by letting a few elec- trons from the tail end of the Maxwellian dis- tribution impinge on the mirror-specimen, using the majority of the electrons, however, for image-forming purposes. It is a particu- larly valuable property of electron mirror microscopy that it can, in such cases, depict without any time delay electrical charge movements which occur on such layers if their conductivity is extremely low. Figure 3 shows as an example a single frame of a 16-mm motion picture taken di- rectly from the screen of an electron mirror microscope. (The dark shadow extending from the center of the picture to the right is caused by the electron gun protruding from the center of the \'iewing screen.) The speci- men was an amorphous selenium film de- posited on a germanium film substrate. The observations which can be made if one ap- plies a small positive bias potential to the substrate film of such a specimen demon- strate quite well that electron mirror micros- copy has some unique capabilities not to be found in any other type of microscopy. One of the many phenomena which can be observed on such an amorphous selenium film is briefly described here as an example. The small positive bias potential on the sub- strate of the mirror specimen lets some elec- trons impinge on the selenium, thus creating a voltage drop across the thickness of the selenium film. This voltage drop in turn causes a slightly negative surface potential equilibrium so that most of the electrons are still reflected by the equipotentials in front of the selenium film. A genuine electron mir- 318 ELECTRON MIKKOK MICROSCOPY ror type image prevails, therefore, despite pictures" of thin films are obtained by pass- positive applied bias potentials. This is the ing an electric current across the specimen, case around the electrical center because a thus creating in the direction of the current voltage drop occurs across, the thickness of flow, on the specimen and in front of it, a the selenium film caused by a small percent- potential pattern related to the conductivity age of the electrons impinging there. It is also pattern. A high potential gradient in the di- the case outside this area because there the rection of the current flow will correspond to normal component of the velocity of the elec- a low conductivity and a low potential gradi- tron is not sufficient to cause electron im- ent to a high conductivity. Where the gradi- pingcment on the mirror-specimen. The inner ent of the potential changes, i.e., where the area, increasing in diameter with increasing area conductivity changes, areas of convex positive bias potential, is filled with random, or concave cuivature of the equipotentials noiselike fluctuations which are more pro- on the specimen and in front of it will be nounced in the center of that area than near transformed by the particular kind of image the border (see Fig. 3). formation of electron mirror microscopy A single frame from the motion picture into dark or bright areas on the viewing cannot, of course, adequately portray the screen. Areas of lower conductivity will movements observable on the screen. For therefore appear as areas with dark borders example, one has to imagine the granular on the one side and bright border areas on structure close to the center area in constant the opposite side. Because of this configura- random, noise-like fluctuations. At a certain tion of dark and bright areas, conductivity positive applied bias potential a pronounced pictures appear as elevated areas, corre- wave-like motion begins which originates sponding to areas of higher conductivity, il- most often in parts of the white-rimmed luminated by a light source from the side border (see upper part of Fig. 3), and moves where the negative voltage is applied. By in a general direction toward the center of taking two consecutive electron mirror the area. Long before reaching this center, micrographs distinguished solely by the the intensity of the waves diminishes and amount of current passing through the speci- they disappear in random fluctuations. The men, it is even possible in certain cases to velocity of the wave-hke motion described obtain pairs of stereomicrographs which are could be changed from practically zero to a three-dimensional pictorial representations few centimeters per second on the screen, of the electrical conductivity distribution in corresponding, at an average magnification thin films (3). With electrical current ap- of about 50 X, to velocities from zero to plied, potential gradient distributions across about one millimeter per second on the speci- p-n junctions (6) or low angle grain bound- men. At present, neither the random fluctua- aries or dislocation lines in semiconductors tions nor the waves moving toward the elec- can also be obser\'ed by electron mirror mi- trical center are fully understood. There is scropy. Figure 4 shows as an example an no doubt, however, that these phenomena electron mirror stereomicrograph pair of the are of electrical origin. That means electron electrical potential gradient across a low mirror microscopy reveals here for direct angle grain boundarj^ in germanium, visual observation the movements of elec- Finally, there is the broad field of magne- trical charges. tism which may profit considerably from the Electron mirror microscopy is also capa- potentialities inherent in electron mirror mi- ble of depicting the conductivity distribution scropy. Whenever the task arises to observe in thin films of semiconductors deposited on visually the distribution of magnetic field insulating substrates. Such "conductivity intensity above a surface, electron mirror 319 ELECTRON >I1HH()K MICROSCOPY Fig. 4. Electron mirror stereomicrograph pair of the electrical potential gradient across a low angle grain boundary in germanium. Fig. 5. Magnetic pattern recorded on magnetic tape as revealed by electron mirror microscopy. The double track pattern was recorded with an audio frequency of 1000 cps, the upper track at a tape speed of l}-i, in. /sec and the lower one at a speed of 3% in. /sec. microscopy can, in general, be used for this purpose. One can, for instance, observe the magnetic pattern recorded on magnetic computer tapes or the magnetic pattern recorded by conventional audio frequency recording (7). Figure 5 shows as an example a magnetic pattern recorded on magnetic tape as revealed by electron mirror micros- copy. The double track pattern was recorded on a 3'4-inch wide tape with an audio fre- quency of 1000 cps, the upper trace at a tape speed of 73^2 inches/sec, the lower one at a speed of 3^^ inches/sec. For the observation of magnetic domain patterns electron mirror microscopy is par- ticularly well suited. Electron mirror micros- copy reveals magnetic domain patterns in materials with a uniaxial direction of easy magnetization (4) as well as in materials with several directions of easy magnetization, ex- hibiting basically flux closure domain con- figurations (8). Figure 6 shows as an example of the first type of material an electron mirror micrograph of a magnetic domain pattern on barium ferrite at a magnification of about 1500 X. Electron mirror microscopy is inherently suited for the instantaneous depiction of domain patterns in motion. A few micro- graphs cannot, of course, adequately portray the quite impressive view of domains in mo- tion as viewed directly on the microscope screen. Motion pictures taken from the screen portray the appearance of domain pat- terns in motion quite well but are still slightly inferior to direct viewing, particularly in cases where domain walls at very low applied magnetic fields are to be observed. Direct Fig. 6. Electron mirror micrograph of magnetic domain pattern on barium ferrite. Magnification approx. 1500X 320 ELECTRON MIRROR MICROSCOPY visual obsen^ation of the magnetic stray fields on grain boundaries in magnetic mate- rials is also made possible by electron mirror microscopy (9). This possibility might be of significance for the still controversial subject of grain size effects on the magnetic proper- ties of such important magnetic materials as silicon-iron and in studies concerned with the basic physics of magnetism in such narrow regions of distorted order. The wide field of thin magnetic films which is becoming more and more important might also profit from the capabilities of electron mirror microscopy. For instance, magnetic domain patterns and their movements with applied magnetic fields can be observed in films of the "Permalloy" type by electron mirror microscopy. Another application in the area of thin magnetic films is the depic- tion of magnetic patterns recorded on ferro- magnetic thin films such as AInBi films. For example, Figure 7 shows an electron mirror micrograph of a sine wave south pole trace in north pole surroundings recorded on MnBi film with an electron beam (10) by utilizing the method of Curie point writing (11). Image and contrast forming in electron mirror microscopy is not of the Gaussian dioptrics type, but is based on a point-to-di- rection correlation rather than on a point-to- point correlation. Optically speaking, elec- tron mirror microscopy resembles the shadow-schlieren method (12) in which the specimen is also used as a mirror and wherein inhomogeneities in the equipotentials in front of the mirror represent the electron op- tical "schlieren." Formation of the image contrast therefore requires the deflection of the electron pencils by the object points which are to be depicted against their back- ground. Contrast formation is thus based on a kind of deflection modulation of the elec- tron paths by that component of the electric field which is parallel to the mirror-specimen. The force causing the deflection is propor- tional to and in the direction of the electric field. This force is independent of electron •^ ** V. i i 1 Fig. 7. South pole trace recorded onto MnBi film by writing with an electron beam as revealed bj^ electron mirror microscopj'. Magnification ap- prox. 60X velocity and thus independent of the direc- tion of the velocity. The force exerted by the electric field providing the deflection retains, therefore, the same direction for the elec- trons when they approach the mirror-speci- men as when they recede from it. In the elec- trical case, and also, of course, in the case of the depiction of a surface relief structure, the resulting contrast-formmg deflection is therefore the sum of both deflections. If one desires, however, to use the in- homogeneities of a magnetic field in front of a specimen as electron optical schlieren to obtain a pictorial representation of the dis- tribution of magnetic fields on a specimen, the case is quite different. The force on an electron caused by a magnetic field (Lorentz force) is more complex and velocity-depend- ent. There is in the magnetic case, therefore, a deflection canceling trend, because for the electron's velocity component normal to the plane of the mirror the direction of the force on the electrons approaching the mirror- specimen is opposite to the direction of the force exerted on the receding electrons. Yet there remains the possibihty of utiUzing the radial component of electron velocity, small as this may be, for image and contrast forma- tion in the magnetic case. This radial ve- locity does not reverse its sign, and thus the 321 ELECTRON MIKHOK >IICROSCOPY Lorcntz force, which stems from the radial in the center, while patterns of other origin component of the electron's velocity, has the will remain the same or may in many cases same direction at a given object area for ap- become even more contrasty and detailed, proaching as well as receding electrons. This A second criterion for distinguishing mag- radial component will be very small most of netic patterns from those of electrical or re- the time, but immediately in front of the lief origin is most convenient experimentally, mirror-specimen it will generally })ecome It requires the shifting of the pattern in predomhiant, making possible its utilization question through the electrical center to the for image and contrast forming in the mag- opposite side of the viewing screen, an opera- netic case. tion which must result, if the pattern is mag- Image contrast formation in the magnetic netic, in a reversal of the brightness of the case is therefore more complex than in the border lines, i.e., dark border lines must be- other cases; yet this difference provides the come bright and the bright must become criteria which permit the discrimination of dark. magnetic patterns from those of other origin, Radial extensions of magnetic pattern ele- a possibility which is of great value for the ments are depicted as more contrasty than practical applications of this novel research those which extend azimuthally. This is in tool. In the magnetic case of electron mirror itself a third criterion if the pattern contains microscopy image contrast-forming deflec- preferred directions azimuthally as w-ell as tion is caused by that component of the radially. If this is not the case, it can still be Lorentz force which stems from the radial used as a third criterion by rotating a pat- component of electron velocity and that tern in such a way as to let the pattern con- component of the magnetic field which is figuration extend preferentially in the radial normal to the plane of the mirror-specimen, direction first and then in the azimuthal di- The sensitivity of electron mirror micros- rection. In the first position the pattern, if copy for magnetic fields is zero at the electri- it is magnetic, will be much more contrasty cal center of the mirror-specimen. (The elec- than in the second position. If one lets elec- trical center is defined as the location on the trons impinge on the electrical center of the specimen plane where the radial velocity of specimen, the image content is lost within the electrons becomes zero or, in practical the area of impinging electrons. The electrons terms, where the electrons first impinge if reflected there, either elastically or as see- the negative bias of the specimen is de- ondary emission, are scattered, thus having creased.) With increasing distance from this lost their predetermined definite direction, electrical center, the magnetic sensitivity This fact then makes the proper kind of increases. image formation of electron mirror micros- To obtain high magnetic sensitivity, a copy impossible, as it is based on a point-to- shift of the electrical center to the border of direction correlation rather than a point-to- the viewing area or even outside is therefore point correlation. advantageous. To utilize this advantage one The shape of the spot representing the must, of course, decrease the negative bias of area w^here the electrons impinge will gener- the specimen to permit sufficiently deep elec- ally be circular in the nonmagnetic cases of tron penetration into the magnetic fields electron mirror microscopy. If in the mag- under observation. Shifting of a pattern into netic case, however, magnetic field compo- the electrical center, therefore, is one possible nents are present which are parallel to the criterionfor distinguishing magnetic patterns plane of the mirror-specimen, the shape will from others. If the pattern under considera- be deformed and the spot representing the tion is magnetic in origin, it must disappear impinging electrons will become displaced. 322 ELECTRON MIKHOK MICROSCOPY One can, therefore, use as a fourth criterion details fall into tiie plane of the screen. Hence, for the presence of magnetic fields in front "focusing" is accomplished by applying proper of the specimen the displacements and shape bias to the mirror rather than by adjusting deformations of the area corresponding to the voltages of the electron lenses. The lenses the area of impinging electrons. In practice, mainly provide the desired magnification and this check can be executed most conveniently are not used as focushig elements in the by lateral movements of the specimen. In the meaning of the Gaussian dioptrics. This re- presence of magnetic field components paral- quires, of course, that the electrical as well lei to the plane of the specimen the spot rep- as geometrical roughness of the specimen be resenting the impinging electrons will move rather small and uniform within the viewing not only against the reference coordinate sys- area. This is a disadvantage in this kind of tem of the specimen (as it does in the non- electron microscopy. The rather high field magnetic case) but also with respect to the strength in front of the specimen is another coordinate system of the viewing screen. Si- disadvantage. multaneously, the shape of the spot represent- There exists another mechanism for image ing the area of impinging electrons will be contrast formation. Differences in surface deformed, depending upon the shape of the potential across the specimen can result in magnetic field. an area-dependent variation in the intensity Because of the peculiar kind of image for- of the reflected beam. More electrons from mation, care must be taken in the interpre- the velocity spectrum of the electron beam tation of electron mirror micrographs. The will impinge on an area which is more posi- potential reUef in front of the mirror is the five than its surroundings. This ''intensity actual image-forming medium. Optically modulation" type of image contrast forma- speaking, it represents a medium of changing tion is in most cases, however, masked to a refractive index with inhomogeneities in considerable extent by the normal "deflec- front of an irregular electron-reflecting sur- tion modulation" which is also always pres- face. This irregular reflecting surface in con- ent and is in general more sensitive, nection with the inhomogeneous refracting Electron mirror microscopy has a differ- medium forms the image. The image is in ent goal from optical microscopy or con- fact the electron density distribution of the ventional electron microscopy. ^Maximal re- caustic of these two inhomogeneities in the solving power, a major goal in conventional plane of the screen. The amount of the small categories of microscopy, is not the major negative bias of the mirror against the cath- objective in electron mirror microscopy; ode determines how far the electrons pene- rather, it makes accessible to direct visual trate into the potential relief in front of the observation the distribution of those physical mirror specimen and which potential surface properties which are not normally depicted will become the reflecting one. This potential by other microscopic methods. Instead of relief rapidly loses its detail and its refracting revealing the distribution of absorption, re- influence on the electron beam with increas- flection or mass density scattering with high ing distance from the mirror. The closer the resolution, electron mirror microscopy strives electrons are permitted to penetrate toward for the obser^'ation of electrical and magnetic the mirror-specimen the more detailed the patterns even if the resolving power is more image formed by the potential relief will be. limited. Nevertheless, the resolution of the One is, of course, inclined to bias the mirror few presently existing rather crude labora- to obtain the most detailed image, i.e., in tory models of electron mirror microscopes is such a way that the focal points of those about 1000 A, which is slightly better than focusing irregularities which represent the fine the resolution of conventional optical mi- 323 ELECTRON MIHKOK AIICKOSCOPY croscopes. One can reasonably expect that the resolution will be improved to about 100 A. The factor finally limiting the resolving power is the rather long de Broglie wave- length of the rather slow electrons at the location where the information is picked up in the form of the image-formhig deflections. Theoretically (13), the smallest resolvable distance 8 should be d[A] = 3.3 • W volts"! cm J in which E stands for the electrical field strength on the surface of the mirror-speci- men. Such a reasonable field strength as 40,000 \'olts/cm at the viewed area of the specimen leads to the abo\'e mentioned figure for the obtainable resolving power. Electron mirror microscopy has a number of advantageous properties and it also has some inherent disadvantages. The require- ment of a very smooth and uniform speci- men surface and the fact that there will al- ways be inherently a comparatively high field strength in front of the specimen have already been mentioned. The peculiar kind of image formation can sometimes result in a rather complicated correlation between the object and the object's electron mirror mi- croscopical image, a correlation which at first sight appears in some cases as not easily in- terpretable. Fortunately, however, there is an additional parameter available, namely, the bias potential on the mirror-specimen. By varying this bias potential and observ- ing the corresponding changes in the electron mirror microscopical images a correct inter- pretation of the images becomes compara- tively easy after some experience has been gained in this respect. Electron mirror microscopy may appear at first thought as a unique research tool for the broad field of surface physics which is becoming increasingly important. In many respects, of course, it can be utilized for that purpose; in many cases, however, its applica- bility is severel}^ restricLed by the vacuum obtainable in present day electron mirror microscopes. The design of a normal elec- tron mirror microscope neither permits out- baking nor sealing off; the obtainable vac- lumi will therefore not be considerably better than 10~^ mm Hg. This in turn does not permit a clean surface in the meaning of present-day surface physics. Difficult as it may be, it should not be impossible, howe\'er, to design simplified electron mirror micro- scopes for special applications in which the extremely high vacua required in modern surface physics could be maintained. Despite its shortcomings electron mirror microscopy promises to become a valuable research tool extending microscopic obser\^a- tions into areas which are not easily accessi- ble by other means. What makes it even more interesting and versatile is the fact that rather often one can observe with an electron mirror microscope not only static electric and magnetic patterns but actually obsen-e some dynamic behavior, such as strangely moving and changing electric charge pat- terns or magnetic domains set in motion by mechanical strain or applied magnetic fields or magnetic stray fields growing out of grain boundaries. Although the basic features of electron mirror optics have been known for many years (14), its utiUzation for a new type of electron microscopy is rather recent. But even with only a few electron mirror microscopes now in operation, the flow of in- formation from them is considerable and often exceeds the observers' capabilities to digest it. REFERENCES 1. Mayer, Ludwig, /. Appl. Phys., 26, 1228 (1955). 2. Bartz, G., Weissenberg, G., and Wiskott, D., "Proc. Third Interntl. Conf. Electr. Microscopy," London, 1954, p. 345; Lud- wig Mayer, /. Appl. Phys., 24, 105 (1953). See also ref. 14. 3. Mayer, Ludwig, J. Appl. Phys. ,28, 259 (1957). 4. Mayer, Ludwig, /. Appl. Phys., 28, 975 (1957); G. V. Spivak, et al., Dokl. Akad. Nauk SSSR, 105, 965 (1955). 324 FIELD EMISSION MICROSCOPY 5. Bartz, G., Weissenberg, G., and Wiskott, D., "Radex Rundschau," Heft %, p. 163 (1956). 6. Bartz, G., AND Weissenberg, G., .Va^wrmss., 44, 229 (1957). 7. Mayer, Ludwig, J. AppL Phys., 29, 658 (1958). 8. Mayer, Ludwig, J. Appl. Phys. SuppL, 30, 252S (1959). 9. Mayer, Ludwig, J. Appl. Phys., 30, 1101 (1959). 10. Mayer, Ludwig, /. Appl. Phys., 29, 1454 (1958). 11. Mayer, Ludwig, J. Appl. Phys., 29, 1003 (1958). 12. "Encyclopedia of Physics" (Springer Verlag, Berlin, 1956), Vol. 34, pp. 565 and 585; H. Hannes, Oplik, 13, 34 (1956). 13. Wiskott, Det.mar, Optik, 13, 403 (1956); 13, 481 (1956). 14. Henneberg, W., Z. tech. Phys., 16, 621 (1935) A. Recknagel, Z. Physik., 104, 381 (1937) G. Hottenroth, Z. Physik., 103, 460 (1936) R. Orthuber, Z. angew. Phys., 1, 79 (1948). Ludwig Mayer Field emission microscopy The basic purpose of microscopy is to pro- duce a detailed enlarged image of the speci- men. The final objective must be to reveal the arrangement of the individual atoms as the finest building stones of matter. Of all presently known microscopic devices the field emission microscope has come nearest to that goal. In the form of the field electron microscope it is able to show the presence and behavior of monoatomic layers on speci- men surfaces, and in the form of the recent field ion microscope it has made visible the individual atoms that constitute the surface of a solid. The field electron microscope was invented by Miiller (1) in 1936. He had observed that at high temperatures surface migration on a clean metal point tends to produce a per- fectly rounded nearly hemispherical tip which is smooth almost down to the lattice steps of atomic dimensions. Such a tip is placed as a cathode opposite a fluorescent screen on anode potential. If the apphed voltage is high enough to produce a field of about 40 million volts per centimeter at the cathode, electrons are emitted which travel on essentially radial trajectories toward the screen. There they display an electron image of the emitter tip at a magnification approxi- EVAPORATOR ANODE \ Oy SCREEN TO PUMP Fig. 1. Schematic diagram of field electron mi- croscope, and of highly enlarged electron emitting tip section at upper left. mately equal to the ratio of tip-screen dis- tance and tip radius. Tj^pical data are 2000 A for the tip radius, 10 cm for the tip screen distance, a voltage of 4000 volts, the image current in the microampere range, and the magnification 500,000 X. The observation is done visually or by photography of the screen. The image brightness is sufficient for taking motion pictures. Theory of Field Emission The mechanism of field emission is under- stood by using the concepts of quantum me- 325 FIKl.D E\IISSIO> MICHOSCOPY chanies. Electrons inside a metal are moving tails to variation in work function. Some- iii a potential trough, from which they can times, however, it is evident that small crys- escape by adding thermal energy (thermi- tallites or asperities appear in the image onic emission), by energy transfer from pho- because of the local field enhancement. Field Emission Patterns from Clean Surfaces tons (photoelectric effect) or by penetrating the potential barrier at the surface (tunnel effect) when the external field is so high that the barrier width is narrowed down to be There is hardly another experiment that comparable with the electron wavelength shows as clearly as field electron microscopy inside the metal. According to the Fowler- the difficulty of obtaining and investigating Nordheim theory of field emission the current really clean metal surfaces. The number of density J (amps/cm-) as a function of field gas molecules striking a square centimeter of strength F (volts/cm) and work function 0 surface in a conventional high vacuum of (e-volts) at the surface can be described by 10~^ mm amounts to 4 X 10^^ per sec, and P2 as the sticking probability on a clean metal / = 1.55 X 10-6 — e-(« 86X10 *' )/F, (1) gui-face is near unity, it takes only a few seconds to build up a monolayer of adsorp- If in a more elaborate theory the electronic tion at that pressure. Therefore, studies of unage force is taken into account, the expo- clean surfaces or of adsorption films under nent will be reduced by a factor slightly reproducible conditions require ultra high smaller than unity dependent upon ■\/F/- vacuum techniques. In field electron micros- This reduces the field strength necessary for copy the criterion for a clean surface is that a given current density by some 10 to 20%. the emission pattern should be independent As field emission is essentially independent of the annealing temperature except for small of temperature, the behavior of the emitter changes in the relative size of the dark surface can be observed in a wide tempera- areas. ture range from liquid helium temperature Only a few crystallographic planes of low to white heat, which is quite valuable for the Miller index appear on a clean surface as study of adsorption. dark regions, indicating a higher work func- The resolution of the field emission micro- tion than the rest of the surface. Probe meas- scope is limited by the random lateral veloc- urements of the current density in various ity component of the emitted electrons, and regions revealed large differences in work also by diffraction of the electron waves. At function. On tungsten 4>\n was found to be a tip radius of 1000 A the theoretical resolu- 4.31 e-volt, and (^on as high as 5.99 e-volts. tion is 20 A, and at 3000 A radius, 40 A. This Observ^ations of field electron microscope is in agreement with practical experience. At patterns of other metals strongly suggest the small protrusions at the tip surface the reso- general occurrence of such previously not ex- lution may be better due to local field dis- pected large work function differences be- tortions. The image details or the contrast tween different crystal planes. On body-cen- are the result of local variations in current tered cubic crystals, such as W, Mo, Ta, density, which according to Eq. (1) is deter- aFe, the highest work functions are on Oil, mined by the local field F and the work func- followed by 112 and 100 planes. Face-cen- tion ^i'^/F can be the short de Broglie wavelength associated derived from measurements of current den- wdth the relatively heavy ion. For the first sity, which makes the interpretation not un- time in the history of microscopy the indi- ambiguous. Rate determinations for finding vidual atoms as they constitute the lattice activation energies may be influenced by the of metals have become visible. The theoret- electric field present during the obsen''ation, ical resolution is about 1.5 A, and experi- although the field effect is usually .small and mentally the individual atoms on the 111 may sometimes be eliminated by observing plane of silicon, with a triangular spacing of with microsecond pulses (4) rather than 2.8 A have been resolved visually. Adjacent with d.c. tungsten or platinum atoms, 2.74 and 2.77 A apart, have been photographed. The image 1 he I' leld Ion Microscope intensity is low, corresponding to an ion A most promising new research tool is the current of 10~^ amp on a 4-inch image screen, field ion microscope (1, 5, 6), which is a so that a typical exposure time is one min- modification of the field emission microscope ute. in that it uses positive ions rather than elec- Field ion microscopy is apphcable onl}- to 329 FIELD KMISSION MICROSCOPY the more refractory metals. At a sufficiently high field atoms of all metals will be "field- evaporated" in the form of positive ions, even at a temperature close to absolute zero. In order to observe a steady surface the evaporation field of the tip material must be above the field necessary for the ioniza- tion of the image producing gas. Helium, requiring 450 MV/cm for ioniza- tion and gi\'ing the best images, can be used for tips of C, W, Re, Ta, Mo, Nb, Ir, Pt, Rhj and with restrictions also for tips of Zr. Pd, Ni, Fe, Co, Si, and of alloys of these metals. Gases with lower ionization poten- tials require less field strength, and fairly good images have been obtained with neon, hydrogen, nitrogen, oxygen, and argon, al- though the resolution is inferior and diffi- culties are encountered with adsorption and possibly field induced chemical reactions of these gases. The luiique feature of operating the field ion microscope with heUum is that, once the tip surface has been cleaned from all con- taminations by field evaporating a few sur- face layers of the metal, not one contamina- tion atom can reach the tip surface as long as the high field is maintained. All gases have a lower ionization potential than helium, so that all contamination molecules will be ion- FiG. 4. Ion microscope picture of a platinum crystal with nearly perfect lattice, with 001 in the cen- ter, and the four 111 planes in the four corners. Many of the about 1000 high index net planes are re- solved in individual atoms. 330 FIELD EMISSION MICHOSCOPY ized way above the surface, and carried away by the field. This makes it possible to design a field ion microscope with verj^ modest vac- uum requirements, providing, for instance, greased joints for easy tip specimen replace- ment. In spite of the resulting poor vacuum conditions (10~^ mm residual pressure), a specimen surface stays atomically clean for any desired length of time, that is without even the adsorption of onlj^ one contaminat- ing atom. The best results are obtained when the microscope is cooled with liquid h\'dro- gen or Hquid hehum, although quite useful images can be made with liquid nitrogen cooling. The most promising application of the field ion microscope is the direct obsen^ation of individual imperfections in metal crystals. The regular arrangement of the atoms in the faultless crystal can more accurately be de- termined by the classical x-ray diffraction methods. The field ion microscope, however, shows directl^y the individual imperfections such as single vacancies, interstitials and dislocations, the structure of lattice steps and of grain bomidaries. Although only the surface is shown in the image, controlled field e\-aporation can be used to remove one surface la3^er of the specimen after another, so that with this "sectioning technique" the interior of the tip crystal becomes acces- sible. The high field exerts a large electro- static stress FVStt at the surface, which amounts to approximately 1000 kg/mm^ when helium ions are used. Nevertheless, all the metals mentioned above can stand this stress in spite of their much lower bulk strength. By pulsing the field the fluctuating stress can be used for fatigue experiments in situ. As no heating is necessary for removing surface contaminations by field evaporation, it is possible to study metal structures that are subject to changes at elevated tempera- tures, such as cold work effects, quenched-in Fig. 5. Ion image of a section of a platinum crystal with many defects. Each fine bright dot represents an individual atom. defects, precipitation in alloy's, or damage by irradiation with high energy particles. The impact of one indi\idual a particle can be seen directlV; and the disarrangement of the metal atoms in the displacement spike can be investigated in detail. It can be assumed that in the future the field ion microscope will increasingly contribute to our knowledge of the atomic structure of solids. REFERENCES 1. Good, R. H., Jr., and Miller, E. W., "En- cyclopedia of Physics" (2nd ed.), Springer Verlag, Vol. 21, p. 17t>-231 (195C). 2. GoMER, R. "Advances in Catalysis,'" Academic Press, Vol. 7, p. 93-134 (1955). 3. Becker, J. A., "Solid State Physics, "Academic Press, Vol. 7, p. 379-424 (1958). 4. Dyke, W. P. .\nd Dolan, W. W., "Advances in Electronics," Academic Press, Vol. 8, p. 90^185 (1956). 5. ^It'LLER, E. W., "Fourth International Con- gress on Electron Microscopj," Berlin, 1958, Springer Verlag, p. 820-835 (1960). 6. MtJLLER, E. W., "Advances in Electronics," Academic Press. Vol. 13. p. 83-179 (1960). Erwix W. jNIuller 331 Fluorescence microscopy* The examination of fluorescing prepara- tions differs fundamentally from the general method of microscopy in which the light transmitted or reflected by the preparation is observed. In fluorescence microscopy the prepara- tion becomes self-luminous while the radia- tion exciting the luminosity does not con- tribute to the image formation but is eliminated by barrier filters. The fluorescing part of the preparation appears bright, usu- ally colored, against a dark background. For excitation, light is used of shorter wave- length than that emitted by the preparation. Thus blue and green fluorescence can only be excited by ultraviolet (UV), while yellow and red fluorescence may also be excited by intense blue-violet (BV). Since only a small part of the incident radiant energy is converted into fluorescent Ught, it is necessary in fluorescence micros- copy to employ the most intense sources of light available. These however, besides the excitation radiations of short w^avelength, also emit light of greater wavelength which would completely flood the relatively weak fluorescences. Therefore two kinds of filters are a part of every fluorescence microscope: (1) ex- citation filters, which transmit in the illumi- nating beam only the excitation radiation of the total radiation emitted by the light source; (2) barrier filters, which bar the fur- ther passage of excitation radiation in the imaging beam. Just as the inherent color of a substance is due to its transmission or reflection of the nonabsorbed Ught falling upon it, so likewise a primary fluorescence, where it occurs, is mainly a function of the chemical constitu- tion. On this basis, guided by the presence or absence of fluorescence of definite quality, * From a Carl Zeiss brochure, by permission. conclusions can be drawn concerning basic chemical composition (fluorescence analysis). Besides the inherent color, affinity for certain stains may be characteristic. The use of stains in histology is a familiar application of topochemical staining technique. Sim- ilarly the method can also be extended to fluorescence and the affinity of so-called fluorchromes for specific substances. Fluor- chromes do not necessarily have a pro- nounced color — in fact some of them are practically colorless. But they brightly fluoresce in a characteristic color when ex- posed to appropriate exciting radiation. Among others, the following are suitable: basic fluorchromes berberine sulfate, auramine, euchrysin, acridine orange, coriphosphin, rivanol, trypaflavin neutral fluorchromes (lipophil) rhodamine B acid fluorchromes fluorescein, thiazine red, sulforhoda- mine, primuline Various tissues or cellular constituents show a specific affinity for the fluorchromes. A peculiar phenomenon, occurring more fre- quently with fluorchromes than with or- dinary stains, is the so-called metachroma- sia. By this is meant that different parts of a specimen "stain" differently with the same fluorchrome, both quantitatively and qual- itatively, dependent on their chemical con- stitution. But the application is not restricted to histology. Fluorchromes are specially suit- able above all in physiology for tracing the transportation and distribution of metabo- lites (absorption, imbibition, secretion, ex- cretion, transportation) and their storage. In general these examinations are carried out in reflected light, since usually the organs cannot be transilluminated. Organs and organelles plainly show fluores- 332 FLLORESCEXCE MICROSCOPY cence after treatment with greatly diluted fluorchrome solutions in contrast to staining with vital stains, which are effective only at relatively high concentrations. Hence it is possible to fluorchrome any organ in full function without damaging it. In addition to basic research, fluorescence microscopy is being employed more and more for routine chnical diagnostic purposes in the following : hematology for leucocyte counts, for differentiating leucocytes, and for counting thrombo- cytes; bacteriology for streak cultures, differentiation of acid-resistant bacilli (t.b. and leprosy) from nonacid-resistant bacilli, for the demonstration of spirochaetes; parasitology for the demonstration of trypanosomes and other flagellates, plasmodia, and other sporozoa. By means of fluorchroming it is also pos- sible to demonstrate the elementary bodies of the so-called large virus types. For this purpose even small outfits are suitable, equipped solely for fluo-violet (BV) excita- tion, since all these fluorchromed objects have their fluorescence in the long-wave region. For characterizing the fluorescent sub- stances one uses the spectral distribution of the fluorescent phenomena. This is done by means of a pupillary spectroscope attached to the tube of the microscope, using the UV filter combination as excitation filter. Since the spectroscope furnishes an average color value of the observed field of view, an iris diaphragm is mounted in the eyepiece of the pupillary spectroscope for eliminating all structures which do not fluoresce in the same color. Above all, the pupillary spectro- scope in connection with the various excita- tion filter combinations permits adapting the choice of barrier filters to the spectral emis- sion of the fluorescing substance and thus to utilize fully the filter intermediate tube. Naturally the fluorescence outfit is also adapted to the observation and evaluation of virtually structureless fluorescent mate- rial, i.e., for observing the fluorescence of licjuids and (cell) suspensions. For this pur- pose one places on the microscope stage a small shallow cell which after filling is closed with a cover glass. Preferably an objective of low magnification is used. The fluorescence spectrum can be observed through the pupillary spectroscope. The great advantage of this micro method of observing the fluo- rescent spectrum is the possibility of carry- ing out the examination of very small sam- ples. (See also Fluorescence, Fluorophotometry in the "Encyclopedia of Spectroscopy." REFERENCES Richards, O. W. : (Latest advances with extended bibliography) Medical Physics, Vol. Ill, p. 375. Year Book Publ. Inc., Chicago, 111., 1960. Lehmann, H., "Lumineszenzanalyse mittels der UV-Filterlampe." Verh. Dtsch. Physik. Ges. 13, 1101-1104 (1911). Reichert, K., "Das Fluoreszenzmikroskop," Phys. Z., 12, 1010-1011 (1911). Lehmann, H., "Das Luminiszenz-Mikroskop, seine Grundalgen und seine Anwendungen," Z. wiss. Mikroskopie, 30, 417-470 (1913). Dhere, Ch., "Nachweis der biologisch wichtigen Korper durch Fluoreszenz und Fluoreszenz- spektren in: Abderhalden, "Handbuch der biologischen Arbeitsmethodeu," Abt. II, Teil 3, Halfte 1, 3097-3306. Verlag Urban and Schwarzenberg, Berlin-Vienne, 1934. Haitinger, M., "Die Methoden der Fliioreszenz- mikroskopie," Ibid., 3307-3337. Haitinger, M., "Fluoreszensmikroskopie. Ihre Anwendung in der Histologie und Chemie," Akademische Verlagsgesellschaft, Leipzig, 1938; De Ment, J. "Fluorescent Chemicals and their Application," New York, 1942. Meyer, H. and Seitz, E. C, "Ultraviolette Strahlen. Ihre Erzeugung, Messung und An- wendung in INIedizin, Biologic und Technik," Verlag Walter de Gruyter and Co., Berlin, 1942. Fonda, G. R. and Seitz, F., "Preparation and chai'acteristics of solid luminescent mate- rials," John Wiley and Sons, New York, 1948. 333 FLYING SPOT MICROSCOPY Danckwardt, p. W. and Eisenbrand, J., "Lu- mineszenz-Analj^se im filtrierten ultraviolet- ten Licht." 5. Edition Akademische Verlags- gesellschaft Geest and Portig. K. F. Leipzig, 1949. Bandow, F., "Lumineszenz, Ergebnisse und An- wendiing in Physik, Chemie und Biologic," Wiss. Verlagsgesellschaft, Stuttgart, 1950. FoRSTER, Th., "Fluoreszenz organischer Verbin- dungen," Gottingen, 1951. Pringsheim, p. and Vogel, M., "Luminesconce of liquids and solids and its practical applica- tions," Interscience Publishers, Inc. New York, N. Y., 1942. G. L. Clark Flying spot microscopy In 1912, Tschachotin (1, 2) devised an ultraviolet microbeam for the purpose of ir- radiating selected areas of living protoplasm. This ultraviolet microbeam was brought to focus by the use of refracting lenses and was rather crudely positioned on the specimen. In 1954, Uretz, Bloom and Zirkle (3) re- ported a much improved method for ultra- violet microbeam irradiation of living protoplasm. In 1956 Montgomery, Roberts and Bonner (4) annoimced the development of the ultraviolet flying-spot television micro- scope and a new tool for the study of ultra- violet irradiation effects came into being. In brief outline, the ultraviolet flying spot television microscope utilizes as a light source an ultraviolet-emitting flying-spot scanner tvibe. A 250-line raster may be traced upon the face of this tube at variable sweep speeds from one sweep every 10 sec- onds to one sweep every l/20th of a second. For intermittent irradiation studies, the raster may be switched on and off for any integral nimiber of frames following any predetermined number of sweeps. The image of the raster of the ultraviolet emitting cathode-ray scanner tube is minified by re- flecting it in reverse through the optical components of a suitable microscope. The unabsorbed energy transmitted by the speci- men is allowed to strike the photocathode surface of an ultraviolet sensitive photo- multiplier tube. The resulting current gener- ated in the photomultiplier tube is suitably amplified and vised to modulate a monitor tube locked in synchrony with the ultra- violet flying-spot scanner tube. In this way, a black and white ultraviolet absorption image of the living specimen is traced out on the monitor tube. Since the spectral charac- teristics of the ultraviolet emitting cathode- ray scanner tube are centered at 2600 A, the black and white image of the specimen on the monitor tube represents, in the main, a nucleoprotein absorption image of the living protoplasm. By appropriate photographic techniques these images may be recorded by time-lapse motion picture photography. A block diagram of this system may be seen in Figure 1. Modifications of this basic ecjuipment have provided a technique for spot irradia- tion of a specimen and this has been re- ported by Montgomery and Bonner (5). In brief this system functions by feeding the horizontal and vertical sweeps to variable delay, pulse-width generators. The pulses thus generated are compared in a coincidence circuit and at the time when both occur, and output from the coincidence circuit modu- lates the scanner tube grid, causing an in- tensified spot to appear on the scanner tube. This spot may be moved about on the raster 334 FLYING SPOT MICROSCOPY DELAYED PULSe GENERATOR X VERTICAL SAWTOOTH GENERATOr? TIMING PULSE GENERATOR HORIZONTAL SAWTOOTH GENERATOR INTERVAL GATE INTERVAL COUNTER 1 DELAYED PULSE GENERATOR VERTICAL SAWTOOTH AMPLIFIER HORIZONTAL SAWTOOTH AMPLIFIER EXPOSURE GATE EXPOSURE COUNTER COINCIDENCE GATE CAMERA PULL-DOWN PULSE AMPLIFIER FIRST SURFACE MIRROR 1/ wv U.V. SCANNER U8E REFLECTING I MICROSCOPE MONITOR TUBE 16 MM cine' CAMERA Fig. 1. Variable sweep -speed UV flying-spot TV microscope by means of the delaj^ circuit and by vary- ing the pulse \vidths, the spot size may be varied down to one picture element. Figure 3 shows the operation of these circuits. The drawings in this Figure lettered A and B show the individual effect of the pulses and their combined effect is showm on the face of the scanner tube. A recent modification of the equipment now allows simultaneous viewing of the specimen in visible and ultraviolet Hght. A visible light tube has been mounted on the equipment at 90° to the ultraviolet tube. A first-surface mirror and beam-spUtter ar- rangement allow the two outputs to be com- bined optically. To view a specimen simultaneously in both wavelengths of hght, opposite halves of each raster are masked off and the remaining output is brought into optical registration. The first half of the horizontal scan, in the focal plane, is then ultraviolet and the second half is visible light, and the two scan the same area of the specimen. The monitor tube then displays the two images in a side-by- side presentation. This arrangement is shown in Figure 4. With only the ultraviolet tube available, positioning the intensified spot required that the background intensit}' be brought up to a level which would allow viewing of the entire field. This would require increasing the video gain to a level that would cause overloading of the A'ideo amplifier in the intensified area Fig. 2. Arrangement of equipment. 335 FLYING SPOT MICROSCOPY HORIZONTAL SAWTOOTH VOLTAGE COMPARATOR /^^ V VARIABLE WIDTH PULSE GENERATOR VERTICAL SAWTOOTH 1-^ COINCIDENCE CIRCUIT VOLTAGE COMPARATOR VARIABLE WIDTH PULSE GENERATOR .JUIJI- ^U-V EMITTING SCANNER TUBE Fig. 3. Spot-irradiation technique. COINCIDENCE CIRCUIT < uu^ c COMPOSITE IN SPECIMEN PLANE BEAM SPLITTER I <> REFLECTING MICROSCOPE DISPLAYED IMAGE ON MONITOR TUBE Fig. 4. Double scanner tube arrangement. and a subsequent loss of information. There was also the possibiHty that the background radiation would cause biological damage to the cells in the field which were to be used as controls. These difficulties were overcome by apply - to the visible light tube a pulse of opposite polarity to that applied to the ultraviolet tube. The two rasters were brought into optical registration. This allowed the ultra- violet background to be at zero level and the background to be filled in with visible light 336 FLYING SPOT MICKOSCOPY JUUl COMPOSITE IN SPECIMEN PLANE h,"^ I?' •^I»l.l, I Fig. 5. Composite ultraviolet visible light system. of the same intensity as the spot. The video obtained uUraviolet absorption image of the system could then be adjusted to the proper remainder of the cell. level to view the entire field and the danger of ultraviolet damage to the control area was removed. This arrangement is shown in Figure 5. By reversing the pulse polarities, the con- verse situation may be achieved. The back- ground is in ultraviolet and a portion of the cell is protected and simultaneously viewed in visible light. Time-lapse motion pictures demonstrating the application of the above briefly described techniques to living cell systems have been taken to demonstrate the following experi- ments : (1) Contmuous, or intermittent, intense differential ultraviolet irradiation of the nucleolus of a Hvmg cell with simultaneously obtained ultraviolet absorption images of the remainder of the cell (Figure 6). (2) Continuous, or intermittent, intense ultraviolet irradiation of the nucleolus of a living cell with a simultaneously obtained visible Ught image of the remainder of the cell. (3) Continuous, or intermittent, intense differential ultraviolet irradiation of the nucleus of a living cell with a simultaneously Fig. 6. Ultraviolet absorption image of a living HeLa cell. The brightened spot centered in the nucleolus represents a one micron square micro- beam of ultraviolet irradiation. 337 FOKKNSIC MICROSCOPY (4) Continuous, or intermittent, intense ultraviolet irradiation of the nucleus of a living cell with a simultaneously obtained visible light image of the remainder of the cell. (5) Conthiuous, or intermittent, intense ultraviolet irradiation of the cytoplasm of a living cell during which time the nucleus of the cell is entirely excluded from ultraviolet irradiation. (6) Continuous, intermittent, intense ul- traviolet irradiation of the entire living cell with a simultaneously obtained ultraviolet absorption image or a visible light image of the cell. (7) Simultaneous presentation of the same image side-by-side on the same moni- tor, as seen when illuminated by visible light-dark field and ultraviolet dark field. (8) Simultaneous side-by-side presenta- tion of the same specimen as seen when il- luminated by ultraviolet light and visible light. REFERENCES 1. TscHAciiOTiN, S.: "Die Mikroskopische Strah- lenstic'hmethode, eine Z elloperationsme- thodo," Biul. Zentralbl. 32, 623-630, (1912). 2. TscHACifOTiN, S., in "Haiidbuch der biologi- scheii Arbeitsmethoden" edited by E. Ab- derhalden, Berlin, Urban & Schwarzenberg, 1938, p. 877. 3. Uretz, R. B., Bloom, W., and Zirkler, R. E.: "Irradiation of Parts of Individual Cells. II," Science, 120, 197-199 (1954). 4. Montgomery, P. O'B., Roberts, F., and Bonner, W.: "The Filing Spot Monochro- matic Ultraviolet Television Microscope," Nature, 117, 1172 (1956). 5. Montgomery, P. O'B. and Bonner, W.: "A New Technique for Ultraviolet Microbeam Irradiation of Living Cells," Arch. Path., 66, 418-421 (1958). P. O'B. Montgomery, Wm. A. Bonner, and L. L. Hundley Forensic microscopy Forensic microscopy can be defined as the use of microscopic techniques for the examin- ation, study, and evaluation of minute speci- mens or minute structures related in some way to a legal investigation. This legal ap- plication of microscopy is not solely the province of criminal investigations, but may also involve civil litigations. The field of forensic microscopy, as well as the spectrum of specimens examined, is almost too vast to be encompassed by a book, much less by an article. However, with a few exceptions, the methods used by the criminalistician do not differ from standard procedures used by other laboratory investigators. Therefore, as in other areas of microscopy, standard refer- ence works are used as they apply to specific problems. An enumeration of some of the micro- scopic techniques used and the specimens examined will give an insight into the prob- lems arising from efforts to solve crimes. Petrographic and crystallographic proce- dures are used in identifying narcotics, pharmaceuticals, and other toxic compounds in connection with illicit drug traffic and poisonings. By determining ph3^sical con- stants with nondestructive microscopic methods, the maximum information can be extracted from small samples, while often preserving them for confirmatory tests and subsequent court presentation. Inorganic compounds and mixtures must also be stud- ied by petrography; these specimens lying chiefly in the fields of building material, paint, safe insulation, and soil. A micro- 338 FORENSIC MICKOSCOPY scopic study after initial sorting from extra- or identification of the tool is made by com- neous debris is often a prelude to spectro- paring test marks made in a suitable manner graphic or x-ray diffraction analysis. on lead, wax, or other similar material with In some forensic examinations the micro- the original mark, or an accurate reproduc- scope plays an ancillary role; in others it tion, from the crime scene. In making this plays a major part in the examination. It is comparison, such variables as the hardness significant that through the microscope the of the material, speed of the operation, and possible source of a specimen of hair on a the angle of the work surface to the test murder weapon can be determined, blood- piece must be controlled and altered sys- stains found on the clothing of a suspect can tematically until the right combination pro- be typed, seeds transported from a crime duces what is termed, a "match", scene can be identified, dust contaminating Each tool-mark will possess certain defi- the clothing of a burglary suspect can be nite characteristics which will identify the evaluated, and many more possible clues to type of work operation performed by the the perpetrator can be studied. The impor- tool, that is to say, the mark may be the tant contribution of microscopy to the medi- result of straight compression or it may be colegal determination of the cause of death the result of some sUding or cutting action through the histological study of various by the tool. The gross dimensions and gen- tissues removed at autopsy is well estab- eral appearance of the tool mark will suggest Ushed. the type of tool that is involved and these Since most of the microscopic methods would be called class characteristics. Dis- employed in the forensic field are either tributed throughout the tool mark will be standard methods, or adaptations of stand- fine striations of a more or less microscopic ard methods, they need not be discussed in nature which will represent the individual detail here. One examination procedure not characteristics of the tool-mark which will be commonly found in analytical laboratories, the basis upon which an identification is yet frequently used in criminalistics, is com- ultimately made. parative micrography. By the use of the The requirements of a "match" are met comparison microscope, tools are related to when the gross dimensions and appearance marks left at the scene of a crime, bullets and the fine striations or individual charac- and cartridges to firearms. The first of these teristics are sufficiently ahke to permit an applications is called tool-mark comparison, experienced technician to render an opinion and the second is referred to as firearms that the same tool made both test and evi- identification. Like many other areas of dence marks. Although the identification is laboratory endeavor, these fields are part art, based upon fundamental principles of proba- part skill, and part science. bility, the work in the field of comparative Basically, the identification of a tool as micrography has not progressed to the point responsible for an evidence or a crime scene of feeding sufficient individual identity data mark is possible only because each tool has into a formula and producing, automatically, many randomly situated, microscopic im- a determination of positi^'e identification, perfections that are not duplicated, even by The merits of the decision must be weighed machine production, on any other tool or against the skill and experience of the tech- workmg surface. These minute imperfections nician in each and every identification, are then transferred to any materials softer An example of tool mark identification is than the working edge, leaving on the sur- shown in Figure 1, where a comparison is face what is equivalent to a fingerprint of made between a drill bit mark in lead, as the the tool that was used. The final comparison standard, and a silicone rubber reproduction 339 FORENSIC MICROSCOPY Fig. 1. A comparison photomicrograph of a silicone rubber reproduction of an evidence tool mark (left) and test tool mark (right). tion procedure is fairly standard throughout comparative micrography and the universal instrument employed for this work is the comparison microscope, made up either as a single unit or constructed by spanning two metallurgical microscopes with a comparison bridge. It is necessary that all optics em- ployed be carefully matched as to magnifica- tion and field and in some instances, it is desirable to employ matched vertical il- luminators such as represented by the Ultrapaks shown in Figure 2. For the most part, low magnification of the order of 10 to 20 diameters will be employed for this work. For special problems, it is useful to have paired optics capable of producing compara- tive examinations as high as 200 to 300 diameter magnification. Although the techniques employed in fire- arms identification are similar to, and prob- ably an offshoot of comparative micrography, nevertheless, this technique is usually given separate identity. Firearms identification — or what is popularly, though erroneously, called "ballistics" — entails more than the simple matching of bullets or cartridge cases. For the sake of simplicity, let us say that a Fig. 2. Comparison microscope using metal- lurgical microscopes, comparison eyepiece and paired Ultrapaks. of a similar drill bit mark made in soft sheet metal. While the nature of the mark is dic- tated by the tool used, the over-all identifica- 340 FOHKNSIC MICROSCOPY weapon will be examined, proved in working order, and will be fired into some collecting medium, such as water, cotton batting or waste, oiled sawdust, or anything soft enough to stop the projectile without muti- lating the surface. On the surface of the fired bullet there will be parallel or slightly diverging lines, inclined slightly from the vertical axis of the bullet. The more promi- nent visible grooves are the impressions of the rifling of the barrel. These are the result of a manufacturing operation which puts a very slow-pitched series of grooves in the interior of the barrel as shown in Figure 3; the pur- pose of these is to cause the bullet to rotate and produce gyroscopic stabihty in flight. The raised portions in the barrel are referred to as "lands" and the depressed areas as "grooves". These, of course, produce their negative counterpart on the fire projectile so that a depressed area or groove on the bullet corresponds to the land of the barrel. The number of lands and grooves and their inclination, as well as other dimensional characteristics depend upon the manufac- turer's notions as to what will produce the most accurate and highest velocity projectile when the gun is in use. Figure 4 shows fired projectiles from two popular American hand- guns. The identification procedure regarding bullets involves first determining that all class characteristics are in conformity. These consist of the caliber of the weapon, the number of lands and grooves and direction of twists, the relative speed of twist, and the width of the lands and grooves. Any major deviation observable in class characteristics immediately suggests nonidentity. The sec- ond step in the identification or comparison of the bullets is to place two tests from the same weapon mider the objectives of the comparison microscope and to study their surface characteristics in order to determine what sort of a "match picture" is presented by tests fired from this particular weapon. One of the test projectiles is then removed and in its place, the evidence or "fatal" pro- jectile is substituted. A study is then made of the surface of both bullets as viewed through the comparison microscope, and by suitable manipulations of their position the technician attempts to match or line up the individual characteristics until a pattern closely dupUcating the comparison of the two test bullets is achieved as shown in Fig- ure 5. The basis upon which an identification is made is not amenable to formula computa- tion. Although both bullets might have been fired in the same weapon, not all fine stria- tions on one projectile will be found repro- duced on the second projectile. The fine striations observed will consist of those made Fig. 3. Cross section of .38 cal. Smith and Wesson barrel showing land running at a slight angle to the barrel axis. Fig. 4. Fired projectiles from .38 Special Colt (left) and .38 Special Smith and Wesson (right) revolvers. Land impression appears in the center of each bullet. 341 FORENSIC MICROSCOPY Fig. 5. Comparison photomicrograph of two projectiles fired from the same weapon. by fixed structures in the interior of the barrel and those produced by transient ma- terial such as rust, unburned powder, ac- cumulated carbon and metallic deposits, and so forth. Consequently, it is necessary for the examiner to study thoroughly the pattern presented by two tests to differenti- ate between striations produced by signifi- cant, stable and permanent imperfections and those of a transient nature. Because of the complexity presented by this changing pattern, it is not possible to establish any percentage criteria of matching lines by which an identification can be assured. Only through the experience of having studied the comparison of numerous projectiles can a technician arrive at a properly qualified opinion as to the identity of the source of two projectiles. After a "match" has been determined, it is possible to take photographs through the comparison microscope. As a rule, this final procedure is not employed smce the presentation of photographs in court is not often helpful to the jury in mak- ing a decision in the case. Of equal importance in the field of firearms identification are the fired cartridge cases ejected at the scene from either hand-oper- ated or semi-automatic or automatic weap- ons. Wlien a cartridge is ignited in a weapon, the high pressures developed force the rear face or head or the cartridge case against a vertical surface, called the breech-face. At the time of this action, any imperfections, striations, pits, and so forth on the surface of the breech-face are recorded as negative im- pressions on the primer of the cartridge case. These breech block markings, in conjunction with the firing pin impression, represent val- uable areas for identity study. Their useful- ness and significance is on a par with rifling impression on the fired projectile. In addition to these two areas, an automatic weapon im- parts additional identification features in the form of extractor and ejector marks resulting from the cyclic operation of the weapon. All these features, separately or together, can af- ford sufficient individual characteristics to substantiate an identification of two car- tridge cases as having a common origin. In a broad sense, the same philosophy and pro- cedure is followed as would be followed in an identification of a fired projectile. Figures 6 and 7 show examples of identification through breech block markings and ejector markings. The results of firearms identification and comparative micrography examinations are expressed in court as opinions. As indicated, it is not possible to subject the observations made to the rigorous procedures of mathe- matics. Therefore, the opinion rendered is 342 FIBERS (TEXTILE) based upon the experience and qualifications of the individual expressing it. Only through Fig. 6. Comparison photomicrographs of breech block markings on two cartridge cases fired in the same weapon. Fig. 7. Comparison photomicrograph of ejec- tor marks on two cartridge cases fired in the same weapon. a proper background and intimate famili- arity with procedures and literature in the field can this information or opinion be ac- cepted as well-founded in court. REFERENCES Journal of Criminal Law, Criminology and Police Science," Northwestern School of Law (Chi- cago), Williams & Wilkins Company, Balti- more, Md. Journal of Forensic Sciences, American Academy of Forensic Sciences, Callaghan & Company, Chicago. Analytical Chemistry , American Chemical Society, Washington, D. C. The Microchemical Journal, Interscience Pub- lishers, Inc., New York. Mikrochimica Acta, Springer-Verlag, Vienna. Kriminalistik, Hamburg, West Germany Hatcher, J. S., Jury, F. J., and Weller, J. "Firearms Investigation, Identification and Evidence," Stackpole Company, Harrisburg, Pa. Davis, John E., "An Introduction to Tool Marks, Firearms and the Striagraph," C. C Thomas, Springfield, 111. O'Hara, C. E., and Osterburg, J. W., "An In- troduction to Criminalistics," Macmillan Company, New York. Kirk, Paul L. "Crime Investigation," Inter- science Publishers, Inc., New York Gonzales, T. A., Vance, M., Helpern, M., and Umberger, C. J., "Legal Medicine, Pa- thology and To.xicology," Appleton-Century- Crofts, Inc., New York. Joseph D. Nicol General microscopy FIBERS (TEXTILE) Microscopic methods are used in fiber technology for: (1) Fiber identification, (2) detection of damage, (3) structural investiga- tion, (4) fiber measurement. On the whole, normal light microscopic methods are in most general use; electron optical methods are important only for fine structural examination. With the bright-field microscope, fibers are examined longitud- inally or in cross section usually after stain- ing. The projection microscope is much used for fiber research. The special instruments used for this purpose are known as lanameters or 343 GENERAL >IICK()SCOPY Table 1. Birefringence Values of Principal Fibers ny na Hf-na A in m^ Wool 1.556 1.546 0.010 10 Wool, chlorinated 1.556 1.548 0.008 8 Silk 1.591 1.538 0.053 53 Casein wool 1.534 1.538 0.004 4 Ardil 1.545 1.545 0.000 0 Cotton 1.578 1.532 0.046 46 Cotton, bleached 1.578 1.532 0.046 46 Cotton, mercerised 1.554 1.524 0.030 30 Linen 1.595 1.522 0.073 70 Linen, bleached 1.594 1.530 0.064 60 Jute 1.577 1.536 0.041 40 Viscose rayon 1.547 1.521 0.026 26 Copper rayon 1.548 1.527 0.021 20 Acetate rayon 1.478 1.473 0.005 5 Nylon 1.580 1.519 0.061 58-60 Perlon 1.582 1.525 0.057 57 Orion 1.500| 1.515/ 1.500\ 1.518/ -0.003 4 Acrilan 1.520 1.524 -0.004 4 X51 1.516 1.520 -0.004 4 Dynel 1.530 1.525 0.005 5 Dacron 1 1.745 1.630 0.115 115 Terylenej Kuralon 1.535 1.526 0.009 9 Arnel 1.472 1.471 0.001 1 Creslan 1.517 1.521 -0.004 3 Verel 1.538 1.539 -0.001 1 Zefran 1.512 1.519 -0.007 7 fibroscopes. In projection they give a con- stant magnification of about 700 X. Measur- ing tables are supplied with this instrument, which permit rapid determination of fiber thickness. The phase contrast and interference micro- scopes are used for structural investigation. This method can be recommended mainly for longitudinal and cross sections. Especially with the interference methods it is possible to observe and measure even minute refrac- tive index differences in the fibers. Birefringence measurements of fibers are very important both for fiber identification and for structural investigation. These meas- urements can be made with a polarizing microscope with the aid of a capillary ro- tator. Very accurate measurements can also be made with the immersion method. The average birefringence values for a number of fibers are listed in Table 1. Fibers are rather seldom subjected to a fluorescent microscopic examination. The fluorescence inherent in most fibers differs only slightly and is moreover fairly weak. The method can be successfully employed, however, in examining cotton for mcrceriza- tion. These fluorescent colors may be im- portant in conjunction with other data. In some cases the use of fluorochromes has pro- vided important data. In the case of wool, for instance, the nature of damage can be demonstrated very well by staining with a mixture of Rhodanine 6 GD and Coriphos- phine. Data so far obtained do show that fluores- cence investigation is definitely useful for fiber microtechnology. Both electron and x-ray microscopy are widely used for fiber research. They are of great importance especially for more funda- mental structural investigation. Investiga- tions are made with replicas, thin sections and breakdown products. In surface exami- nation, good results have already been ob- tained with the electron scanning micro- scope. It appears that this method, together with x-ray microanalysis and ion etching is going to be very important in fiber research. Techniques of Fiber Microscopy The technical part of fiber microscopy can be divided into: preparation methods, stain- ing methods, mounting methods and micro- scopic techniques. Most of the preparation techniques can be used for examining all types of fibers, but this is not the case with staining methods. Subdivision according to the nature of the fibers is necessary. Preparation Techniques. Making Cross-sections. Three methods are employed for making cross-sections. (1) Plate Method. A bunch of fibers is pushed through a 0.5-mm hole in a 0.5-mm thick copper plate. The opening must be well filled so that the fibers are tightly packed. The fibers are severed on both sides with a 344 F1B^:KS (TEXTILE) sharp razor blade. The cut surface is now examined under a microscope without fur- ther preparation. (2) Hardy Microtome. Relatively thin sections can be made with this simple in- strument. It consists of two rectangular plates held together by lateral ridges. In the middle of one plate there is a very narrow slot. From the middle of the other plate a thin metal lip protrudes which fits exactly into this slot. If the two plates are fitted to- gether a very small part of the slot remains open. By means of a micrometer screw" a second metal lip can be pressed into this open part from underneath. In order to make a section, the slot is filled with a bunch of fibers. Next, the second plate is affixed. The fibers are severed with a razor blade above and below the plate. The mi- crometer screw is fitted in such a way that the lip is immediately under the slot. By means of this, the bunch of fibers can now be pressed several microns out of the slot. After the protruding fibers have been coated with cellulose lacquer, the cut is made with a sharp razor blade pressed as close to the plate as possible. Sections of lO^u can easily be made, and with sufficient experience sections of 3ju are not unusual. (3) Freezing Microtome. If a bunch of fibers is imbedded in celloidine or gelatin, it it possible to make a l-dfj. section with a sledge freeze-microtome. On the whole the sections are more regular than with the previous methods and hence more suitable for examination with phase contrast and in- terference microscopes. Making Longitudinal Sections. If a bunch of fibers laid exactly parallel is mounted in gelatin, thin longitudinal sections can be made therefrom with a sledge freeze-micro- tome. It is essential for the knife to be mov- able W'hile the specimen holder is stationary. It must, however, be possible to set the holder three ways relative to the knife. If a freezing microtome is not available. it is quite possible to make good longitudinal sections with Stove's method, as follows: A cellulose acetate film is applied to a slide. This film is moistened with xylene. On top of this a bunch of fibers is laid exactly parallel. It is advisable to apply some ten- sion by affixing a 2-gram weight at each end of the bunch. When the xylene has evapor- ated, a drop of ethyl lactate is added, while next a solution having the following compo- sition is applied in abundance: Cellulose acetate film 5gr Tricresyl phosphate 3 ml Ethyl lactate 10 ml Amyl acetate 75 ml The whole must then dr^^ for 36 hours at room temperature and can next be mounted in paraffin with a melting point of 50°C. It is better to use a mixture of paraffin and bees- w^ax, the hardness then being adapted to that of the fibers. Surface Examination. The surface struc- tures of fibers can be examined by the fol- lowing methods: (1) replica, (2) semi-embed- ding, (3) interference microscopy, and (4) electron scanning method. (1) In the repUca method, cellulose ace- tate (12 g) is usually used, dissolved m ethyl lactate (10 ml) and amyl acetate (78 ml). A film of this solution is made on a slide by the smear method customary for microbiol- ogy. The fibers are placed carefully on this film. After about 30 min. they can be pulled off the film. The replica can be examined forthwith under a microscope. Instead of cellulose acetate solution a normal cellulose lacquer can be successfully used. (2) In semi-embedding, the fiber is care- fully pressed into a swollen film with about the same refractive index as the fiber. In the microscopic image the hifiuence of the lower half of the fiber is thus eliminated. Protein glycerin can be effectively used for protein fibers. For wool and keratin fibers ROX 1 (Chroma) is particularly suitable. For cellulose fibers ROX 2 (Chroma) or cellulose lacquer may be used. 345 GENERAL MICROSCOPY For synthetic fibers, cellulose lacquer can be used. In these cases it is advisable, after the lacquer has dried, to apply a thin shadow coating to the upper surface of the fibers. This coating is especially necessary if the fibers are matted. (3) The normal interference methods can usually be employed for surface examination of fibers. (4) A very good method for fibers is the electron scanning method. On the whole the images are much better than those obtained with the reflection electron microscope (q.v.) The foreshortening of the image is not a serious drawback. This method can be par- ticularly important for examining keratin fibers and paper suspensions. Staining Methods and Chemical Reactions. Of the numerous staining methods and chemical reactions used in textile and paper microscopy, only the most important will be mentioned. A complete list of all the customary reactions can be found in the references. Cuoxam (copper oxide ammonia) reagent for cellulose. Ammonia is added to an aqueous solu- tion of copper sulfate. The precipitate is washed and dried and put in as little concentrated am- monia as possible. Herzberg's Solution (zinc chloroiodide) (A) 20 g dry zinc chloride in 10 g water; (B) 2.1 g KI and 0.1 g iodine in 5 cc water. Solutions A and B are mixed together. Allow precipitate to settle, de- cant and keep in the dark, a crystal of iodine be- ing added. I ado --potassium iodide (KI3). 0.6 g KI and 1 g iodine dissolved in 100 cc water. Phloroglucine-acid. An approx. 5% solution of phloroglucinol in absolute alcohol. Always used with concentrated HCl. Reagent for lignified cell walls. Millon's reagent. 1 part mercxu-y is dissolved, with application of heat, in 2 parts HNO3 (cone). Next 2 parts of water are added. Reagent for proteins. Ruthenium Red. 0.1 g ammoniacal oxychloride of ruthenium is dissolved in water (10 cc). After staining, the fibers are mounted in glycerol. Gen- eral reagent for best fibers. Victoria Blue B. Reagent for distinguishing raw and bleached cotton. 3% Victoria Blue in 10 cc water. The fibers are boiled in this solution. Wash in cold water until no more stain comes off. Colotex B. Stain made by Union Chemical Com- pany, New York; similar to Neocarmine W. The fibers are stained for 3-5 minutes in the solution, washed in water and then washed again in water containing several drops of ammonia, washed once again in distilled water and then dried. Shirlastain A (I.C.I.). The fibers are eluated for 1 minute at room temperature, stained in the solu- tion, washed in distilled water and examined un- der the microscope. lodo-sulfuric acid. 3 g KI is dissolved in 60 cc water, while 1 g iodine is added. 10 parts of water are added. The sulfuric acid solution is prepared by adding together 3 parts glycerin, 1 part water and 1 part concentrated sulfuric acid. After the fibers are stained in the iodine solution they are mounted in the sulfuric acid. Allworden's reaction. Reagent for damage to wool (see Fig. lb). The wool fibers are treated with saturated chlorine or bromine solution. Depend- ing upon the thickness of the fiber, the tempera- ture, and the acid or lye damage, a number of small blisters occur on the fiber. The time needed for bringing about the blisters indicates the state of damage. In the case of wool damaged by acid the blisters occur earlier than normal, while in the case of lye-damage the reaction takes longer than normal. The normal reaction times at 17° C are: for fine wool for wool for coarse wool 20-25 M diam. 25-30 M " >.30m " 30 sec 70-80 sec 120 sec The reaction time must always be compared with a control to eliminate influences of tempera- ture, reagent and fiber diameter. K.M.V. reagent. (Fig. 1.) 20 g KOH is dissolved in 50 cc liquid ammonia; after about 1 hour the liquid is decanted and stored in a stoppered flask. The reagent keeps for 6 to 8 weeks. Pieces of wool not longer than 3 mm are treated with the reagent. The moment the reagent begins to act a stopwatch is pressed. The reaction time is the time elapsing before the first blisters ap- pear on the fiber. This depends upon: (a) the thickness of the fibers (thinner fibers have a short reaction time). (b) temperature (higher temperature — shorter reaction time). (c) state of damage (acid-damage shortens re- action time; lye-damage lengthens it). (d) age of reagent (the reaction time is in- creased through loss of NH3). (e) chroming of wool (chromed wool reacts more slowly than unchromed). Factors (a), (b) and (d) can be eliminated by using a control of the same thickness. 346 ^^ v_ • i V FIBERS (TEXTILE) -1 !K— f^*^ . -^^'te^' ' ^ ■^^^'''■■"^TBP'^S"*"^'*™*^ V 1 ^"^..- ■; ^^ /-\ "--< A ■J ^ ^^ 0 u u Fig. la. Wool fibers (from top down): ASTM Fig. lb. Wool fibers (from top down): ASTM 40, cross section, 450X. ASTM 56, longitudinal, 56, KMV reaction, llOX. AUworden reaction, llOX. ASTM 56, cross section, llOX. ASTM 58, 450X. Damaged by heat, llOX. Re-used wool, longitudinal, llOX. 225X. 347 GENERAL MICROSCOPY The I'ullowiiig reaction times at 18°C can be taken as normal for several thicknesses of wool : Dutch wool (37m) 12 min New Zealand (24^) 8 min Cape wool (20^) 8 min The first indication of blistering is at the ends of the pieces. Cumulative damage may result in a normal re- action time. Methylene Blue Test. 0.002-0.003% methylene blue in distilled water. Pieces of wool about 5 mm long in methylene blue are microscopically examined after the dye has acted for about 10 minutes. This reaction is used mainly to dis- tinguish damage bj' heat from damage by alkali. Methyl Orange Test. The normal indicator solution is used. In the case of wool this test is used for: (a) detection of damage by light (b) retention of chemicals distinguishable by color and formation of crj^stals of the acid form of methyl orange, which is far less water-soluble than the sodium salt itself. Congo Red Test. The Congo Red test is based on: (1) The varying diffusion at room temperature of stains in the cuticula and in the secondary cellu- lose layer of cotton fiber after this has been made accessible to the stain to a greater or less extent by swelling in caustic soda solution, with the re- sult that the cellulose layer is stained a deeper shade. (2) The behavior of cotton fiber when swollen with caustic soda solution, especially as regards the constricting effect of the cuticula. (3) Spiral splitting of the cuticula. The speci- men of cotton for examination is : (a) soaked for 3 min in a sodium hydroxide solution of a certain strength; this may con- tain a wetting agent if desired. (b) washed with water. (c) placed for 10 min in a concentrated solution of Congo Red. (d) rinsed. (e) placed in an 18% caustic soda solution. This test is used for detecting various forms of damage to cotton. Although on the whole it al- lows of explicit conclusions, it should be stated that the result is entirely qualitative. Cotton Blne-Lactophenol. The fibers are heated for 1 min in lactophenol, then stained for 5-10 min with a warm aqueous solution of cotton blue. After rinsing with warm lactophenol they are microscopically examined. This test is used mainly for damage by molds. Mounting Methods. It is of major importance for the fibers to be mounted in a medium whose refractive index does not vary too much from that of the fibers. The following media are very satis- factory: (1) Glycerin-gelatin: 7 g gelatin is soaked for 2 hr in 42 cc distilled water; 50 g glycerin and 0.5 g phenol crystals are added; it is heated on a water bath for 10-15 min, while stirring, and is filtered through a glass filter, (n = 1.474.) (2) Canada balsam. (3) Broadfoot and Schwarz. Constituent parts: isobutyl methacrylate (Du Pont), Aroclor 1242 (Monsanto), Xylene. IBM (cc) 10 18 18 18 18 Aroclor (cc) Xylene (cc) 30 3 27 6 24 9 21 12 18 Refractive index 1.495-1.506 1.500-1.505 1.505-1.510 1.5ia-1.520 1.520-1.525 (4) Aquamount (Gurr) (5) Glycerin 85% Special Micro-techniques. The fol- lowing techniques are important in fiber re- search : Micro-elongation tests. The fibers are fixed tight at one end, and run over a pulley at the other, on which a weight may be sus- pended. To ensure that the same piece is examined throughout the test, the method may be varied by having a pulley and a weight at each end. Micro-torsion tests. Here again the fibers are fixed tight at one end. The other end is fixed in rotating metal jaws. Great care must be taken to ensure that the bearing and cen- tering of the rotating part are very accurate to avoid the fiber disappearing from view during the test. Capillary rotator. For exact birefringence measurement of fibers the fiber is placed in a glass capillary. The capillary is filled with a hquid whose refractive index is the same as that of glass, and it can be rotated in the same liquid. It is possible with a simple disc to rotate the capillary with the fiber through an angle of 90°. The double refraction of the 348 FIBERS (TEXTILE) fiber is now measured, while the rotator is turned 90± to measure the fiber thickness. Special measuring methods. Special meth- ods that may be ment ioned are the microme- ter plates for determining denier and fullness and the measurements for determining lumen percentage. Various methods have been suggested in the course of the years but few have become widely used (e.g., the "Universalokular nach A. Herzog" marketed by VEB Optik, Jena). In this ocular, a reticulated micrometer makes it possible to determine the titre of viscose fibers direct from the cross section. Calibration of the micrometer also makes it suitable for all other fiber sections. The fullness of fibers of irregular section can be calculated by determining the true surface with a planimeter and calculating the area of a circle described from the sec- tion. The fullness V, as a percentage, is then V = (400 F/ttB-') = 127.32 {F/B'-), in which F is the fiber-section and B the greatest di- mension of the section. If a projection micro- scope is not available the true area can easily be determined with a reticulated micrometer. The lumen percentage can be determined most easily using a fixed magnification pro- jection microscope. The fibers are compared wdth a number of round "ideal" sections, in which both lumen and sections are circular. Preparation ^lethods for Electron Microscopic Examination. Duglosz' semi- embedding method jor making replicas. As fibers are usually bigger than lOju in diame- ter, it is difficult to make replicas that have retained their original form after the neces- sary manipulation. With the semi-embedding method, fibers are dropped on to a photo- graphic plate, the gelatine layer of which has been swollen. After the gelatine layer has dried, a microscope sUde covered with a marco-resin film with catalyst and acceler- ator is placed on the gelatin layer. When the plastic has hardened, the photographic plate is pressed off. The sHde with the resin and the fibers fixed in it are now used to make a replica by the silver-carbon method. Pagers replica method for extensive surfaces. Fibers out of a suspension are put to dr}^ on a clean slide. A film of meth}^ met hacry late swollen at the surface in plasticizer is pressed on the slide. After hardening, the film is re- moved together with the fibers. With the aid of a film of polyvinyl alcohol, allowed to form from a 10% solution, the fibers are pulled from the methacrylate. Lastly, a posi- tive replica is formed, again with 10 % poly- vinyl alcohol. Other replica methods. (1) Impressions in polyisobutyl methacrylate which softens at the surface at 100°C. (2) Semi-embedding in polystyrene. The upper surface is then used for making a replica. (3) Bradley's two-stage replica technique is also widely used for fibers. Fiber embedding for idtra-microtomy. Cut- ting ultra-thin sections of fibers necessitates embedding them in a synthetic resin. This raises the following problems: (1) The embedding medium must in- filtrate properly into the fiber and must not change its structure. (2) During hardening the plastic must not loosen from the fiber. (3) It must be possible to orient the fiber properly in the sj^nthetic resin. Embedding media that are used are 50/50 mixtures of butyl and methyl methacr\'late or epoxy resins. The latter are more suitable because they do not loosen from the fibers so readily. The fact that the epoxy resins have a rather distinct structure under the electron microscope is a drawback however. Electron staining reactions, ^"arious stain- ing reactions are recommended for electron microscopic examination of fibers. Good re- sults are obtained with osmium tetroxide, ferrialum and ZnCU . For cellulose fibers, 20 % AgNOs for several hours also gives good results. Hess successfully used KI3 (Iodine 5%, KI 10%, pH 5). He got very good results 349 GENERAL MICROSCOPY with thallium ethylate in benzene. All these methods can be used successfully for cellu- lose fibers and polyvinyl alcohol. Fiber Microscopy and Morphology Fibers can be classified as: (1) protein fibers, (2) cellulose fibers, (3) synthetic fibers. Protein Fibers. The most important pro- tein fibers are the keratin fibers. As regards textile techniques wool is the chief one of these. In addition to wool, are hairs that are used for "effect" or for other special pur- poses, e.g., mohair, cashmere and goat hair, from the Angora goat, the Cashmere goat and the common goat, respectively, camel hair, llama, alpaca and Vicuna, allied to the camel. There are many kinds of fur-produc- ing animals, for instance : the rabbit, musk- rat, skunk, mink, wolf, ocelot, seal, tiger cat, lamb. Brush fibers come from the horse, cow, pig, marten, skunk and squirrel. It is important to be able to identify all these fibers microscopically. No good sys- tematic method for this exists as yet. An im- portant contribution has been made by Wild- man who systematically classified the scales, according to the shape of the margin, as fol- lows : 1. (a) Smooth: Margins which are on the whole smooth with little or no inden- tation. (b) Crenate: This means "notched" and is restricted here to margins which have fairly shallow indentation and in which the majority of the "teeth" taper to a relatively sharp point. (c) Rippled: Margins having indentation, the majority of which are deeper than those of the crenate type and in which most of the "peaks" are rounded in- stead of ending in sharp points. (d) Scalloped: Scalloped margins are rela- tively rare among the textile fibers. 2. The distance apart of the external margins of the scale comparison is at a stand- ard magnification, (a) Close: Successive scale margins along the fiber which are very close together l:h (length: height) = >10. (b) Distant: Scale margins which are far apart in the direction of the long axis of the fiber l:h = 3:10. (c) Near: Successive scale margins which are spaced at distances intermediate between the above extremes l:h < 3. 3. Type of Over-all Pattern. All the following patterns may be qualified by the terms defined above to indicate the form of the scale margins and their spacing intermediate patterns can be described by combining the names of the basic patterns. (a) Mosaic: The term is self-explanatory but is qualified as follows: Regular mosaic: The units of which are very approximately the same size. Irregular mosaic: The units of which are obviously not of the same size. (b) Waved: Some form of wave occurs commonly in scale pattern and the term is applied to any pattern which appears to be wavy; where the waves follow a continuous course around the fiber their uninterrupted character is inferred from the term waved without further qualification. Interrupted wave. A pattern which has a general wave form but which is clearly interrupted falls into this category. Simple regular wave: This type has waves which are continuous and are of approximately the same wave- length and amplitude. Streaked wave: The waves are inter- rupted at regular intervals by longi- tudinally running bands of steeply inclined scale margins. Shallow, deep and medium waves: These adjectives indicate the rela- tive depth of the majority of the waves in the pattern irrespective of the distance apart of the scale mar- gins. 350 FIBERS (TEXTILE) (c) Chevron: This pattern consists of waves, usually regular, but some times irregular, with either the crests or the bottoms of the -troughs or both, relatively narrow and V-shaped. There are two varieties of chevron pattern, namely. Single chevron: in which only one of the crests or troughs is narrow and V- shaped. Double chevron: in which both the crests and the bottoms of troughs of the waves are narrow and V-shaped. (d) Pectinate (comblike) : The term is seK explanatory, but two basic varieties are recognized. Coarse pectinate describes the pat- tern in which the "teeth" of the comb are relatively large and set wide apart. Lanceolate is the pattern in which the teeth of the comb are long, narrow and rather more pointed than in the coarse pectmate type. (e) Petal patterns: The general appear- ance of these resembles a series of overlappmg flower petals due to the imbricate or overlapping appearance of the scales. There are several varie- ties of petal patterns and they are really modified wave forms, some- times interrupted, regular waves and sometimes interrupted irregular waves. It must be clearly understood that a pattern does not qualify as a petal pattern unless its appearance strongly suggests a series of overlap- ping or imbricate elements. Irregular petal: This is a form of in- terrupted irregular wave. Diamond petal: The wave crests over- lap the troughs of the succeeding series of waves toward the tip end of the fiber. With the aid of this scale system it is cer- tainly possible to identify the types of hair, especially if intermediate forms are con- cerned and the differences between root end, shaft and tip of the hair are examined sepa- rately. The medulla of fibers can be used for identification in only a few cases. We can distinguish : 1. Continuous thin medulla — diameter <32 the fiber diameter. 2. Continuous full medulla — diameter > }-'2 the fiber diameter. 3. Continuous, regularly pectinate. Cell partitions divide the medulla into regu- lar chambers (a) chambers in one row (b) chambers in several rows. 4. Continuous spiral. 5. Interrupted at regular intervals, pre- dominant type. 6. Interrupted at irregular intervals, pre- dominant type. 7. Sporadic types. 8. Irregular shapes. Many of these types occur with one kind of fiber. Only 2, 3 and 4 have any systematic significance. Structure of Wool and Other Animal Hair (Figs, la and lb). In general animal hair consists of the fol- lowing parts: The medulla, the innermost part of the hair, recognizable by less elongated cells not highly pigmented. After cornification, fairly large air cavities occur in the medulla, which may help to bring about the color of the hair (interference colors). As a rule the medulla is fairly easy to see in wool preparations. The cortex, a zone consisting of spindle- shaped cells. Each of these cells is enveloped in membrane about 500 A thick. The con- tents consist of fibrils (tonofibrils). These tonofibrils are well oriented along the fiber axis. In between the fibrils there is an evap- orated protein substance, the cement. The tonofibrils consist of keratin. The cortex outer layer is formed by a less oriented fibril- lary mass: the cortex mantle. This cortex mantle forms a screen in the fiber against the penetration of dyes and chemicals. 351 GENERAL MICROSCOPY The scale layer, a layer of flat overlapping The scale cells can best be examined with elements. The scales also have a fibrillary the semi-embedding methods discussed under structure. They consist of macro-fibrils which Surface Examination (p. 345) or by making can be seen in isolated scales with a phase a rephca. If the hair is pigmented it must microscope and micro-fibrils observable only first be bleached in hydrogen peroxide, with an electron microscope. In between The cortex cells. The course of the cortex these fibrils is the scale cement. layer can most suitably be examined with This heterogeneous structure of the scales normal glycerin preparations. The individual has led some authors to classify them as en- cortex cells can be examined by disintegrat- docuticula and exocuticula. But this distinc- ing the hair in caustic potash or by the tion has no significance in practice. trypsin-sulfuric acid method. The scales play an important part in the The caustic potash treatment is effected occurrence of crimp. Outside a curve the in 5% KOH for 2-3 hours at 50°C. The scales are thicker than inside and there is drawback is that the cells swell greatly, also more matter there. Shape and size are rather uncertain. Chemicals and dyes can penetrate better The better way is first to dissolve the into the scales inside the curve than into cement substance by the trypsin treatment those on the outside. and then to make a preparation in sulfuric The epicuticula, a membrane about 100 A acid. thick completely enveloping the wool hair „ . , ^. « ^r ^ • ^ no ^ '^ r- o Trypsin solution: 0.75 g trypsin and 0.3 g so- on the outside. It is very resistant to attack ^ium bicarbonate are dissolved in 100 cc distilled by acids and bases. Its composition, however, water. (The sodium bicarbonate must be very does not differ fundamentally from normal pure). The solution is fully active for only one keratin hour. The hair is defatted properly in petroleum ri ,1 J. . . 1 X J.T- • ±- ^ • ether or benzene, dried, washed and cut in 0.5-1 Some authors state that the epicuticula is . n on en u • m .^ • i ^ mm pieces, rer 30-50 mg hair, 10 cc trypsin solu- the cortex cell membrane. Kassenbeck tion is used. The suspension is left standing for 3 showed in ultra-thin sections with an electron hours in a closed Erlenmeyer flask at 40°C ± 3°C. microscope (1) that the scale edges which are Sulfuric acid: 85 parts analytically pure sulfuric visible under the normal microscope are ^^.i^, s.g. 1.84 and 15 parts distilled water are ridges on the actual scales which envelop the fiber like a cuff ; (2) that the epicuticula was The cortex mantle can best be prepared by membrane enveloping the scale. treating the wool fibers with 1.6% peracetic acid followed by extraction with 0.1 iV am- Methods of Examining the Various Fiber j^onium hydroxide. The insoluble portion is Elements 7-10% of the fibers. A cross-section shows The epicuticula. The structure of the epi- that besides the cortex mantle the cell mem- cuticula can be examined with Von All- brane of the cortex cells can also be seen, worden's reaction. The defatted wool is while in phase contrast part of the scales is placed in saturated bromine solution and is also found to be present, examined through a microscope until the The medullary sheath. The size and course first blisters occur. Water is then sucked of the medullary canal can best be studied through, the blisters usually increasing in from a longitudinal preparation. To study size. The blisters may be stained with 0.1% the size and shape properly, the air in it methylene blue. The time elapsing before the must be expelled. first bhsters occur and the shape and size of The most effective way to do this by lay- these bhsters are important and may be used ing short pieces in KOH or turpentine. A as important characteristics to detect dam- 17 % KOH solution at room temperature is age. mostly used. If the 17% KOH solution is 352 FIBERS (TEXTILE) just brought to the boil, this if necessary be- (keratin) which is Hmitcd in quantity as ing repeated, scale layer and cortex disin- compared with celhilose, microorganisms tegrate. The cement substance between the capable of attacking this complex substance medullary cells also disintegrates. are not found as commonly as cellulose de- Bilateral structure of wool. The Japanese, composing organisms. Nevertheless, the Horio and Kondo, found that if frizzy hair keratin-attacking organisms are sufficiently was stained with Janus green or Ponceau R represented to necessitate adequate precau- the hairs showed a bilateral structure, one tions in handling and storing woolen prod- lengthwise half showing greater affinity for ucts. Ponceau R and the other a greater affinity Wool can be easily attacked by bacteria for Janus green. They concluded that it had and by molds. The most severe is caused by a bilateral structure. Mercer has called the several strains of Penicillia and/or Asper- two halves the para-cortex and the ortho- gillus, which cause colored spots. These cortex. spots, which are serious because they are The difference between the two parts of difficult to remove and may moreover cause the fiber is very easy to see if the wool hair trouble in dyeing, are a fairly superficial at- is first stained with methylene blue and if tack usually caused through microorganisms this is followed by Allworden's reaction. The growing on fiber such as wax, soap, oil, fat or inside curve of frizzy hair is then much nitrogenous waste products, more strongly affected than the outer. Molds of the genera Alternaria, Stem- Examination of cross-sections shows that phylium, Oospora and Penicillia cause the on the whole the ortho-cortex consists of spots on wool but to a certain extent attack larger cells. the wool as well. Other strains causing such The difference between para and ortho- effects are Aspergillus, Fusarine, Tricho- cortex is found to some extent with all kera- derma and Cephalothecium. tin fibers. Although mohair was originally The aerobic bacilli include several which described as 100 % ortho fiber, close examina- may attack wool. One of the best known is tion has shown that a part of this is also para Bacillus mesentericus. Furthermore, there is cortex, although the percentage is low. Bacillus subtiHs. A very strong attack is The dividing fine between the para and caused by Actinomycetes which quickly ortho parts is usually very irregular. In some brings about a loss of mechanical character- cases there may be two ortho parts with the ig^jcg Qf ^^e wool. There are strong indica- para cortex between, while fibers have also ^-^^^^ ^j^^^ ^j^ig damage is caused only if the been observed with a central ''ortho" strand ^^^^ ^^^ already been damaged mechanically enveloped in para cortex. ^^ chemically. Fluorescent microscope exammation shows proteolytic bacteria and molds rapidly that besides the differences between the cor- ^^^^^^ ^^^^^ ^^^^ ^^^^^ favorable conditions, tex parts there are also fairly strong differ- ^^^^^^^ ^^^ ^.^^^^^^ ^^^ intercortical ences between the scales 01 the outer and in- , , ,. • , ,• • . ^i • ji , , J., T xi, r J J cement, and disintegration into the spindle- ner curves of wool fiber. In the case ot dyed , ' „ "= ^, ^ ^ fibers, the scales on the ortho are found to shaped cortex cells occurs. These effects can allow dye to penetrate better than those on be seen directly with microscopic examina- the oara side ^^°^" '^^^^ ^^^^ should be distinguished from normal mechanical damage because hyphae Microbiological and Biological Attack a^^ci sometimes spore carriers are usually on Wool present. Microbiological Attack. As wool and Especially after staining with 0.1 % meth- other animal hair contains a specific protein ylene blue, the hyphae are easy to see, while 353 GENERAL MICROSCOPY mold attack can be rendered very clearly- visible with a phase contrast microscope. Even a fairly slight attack causes a loss in double refractivity and therefore a polariza- tion microscope can also be used. Bacterial attack can be observed very well and very quickly Avith a phase contrast mi- croscope. Staining methods also make the bacteria clearly demonstrable, e.g., with methylene blue or fuchsine (2-3 min.) or with carbol fuchsine or carbol gentian violet (20-30 sec). Finally, it should be stated that attack by molds and bacteria can occur only if there is sufficient moisture, whilst soluble nitroge- nous material is likewise essential. Biological Attack. Biological attack is largely accounted for by moths and the larvae of the carpet beetle {Anthrenus ver- hasci L.). Here again, it may be mentioned that neither moths nor carpet beetles can attack wool unless fats and salts are present. Attack by insects can be detected micro- scopically immediately, owing to the typical morphology of the damage. Moreover, threads of the moth larva's web are often present and beetles often leave hairs behind. Diagnosis therefore causes no difficulty. Morphology of Fibers Silk (Fig. 2a). Raw silk consists of two filaments of fibroin stuck together with seri- cin. The filaments are triangular. Longitud- inally, the fiber is very irregular. It shows constrictions, creasing, folding and thicken- ings. Both the shape and dimensions of the triangular fibroin filaments are important. If the cross section is slightly rounded the fibers take dye less readily than the larger purely triangular ones. Degummed silk is a smooth, almost struc- tureless fiber. A barely visible longitudinal striation can be observed. Zinc chloroiodide sometimes stains silk fight yellow. In Mil- Ion's reagent the fibers turn red. Provided it is not weighted, silk dissolves in Cuoxam. Tussah silk (Fig. 2a). The filaments are flat and wide, show pronounced longitudinal striation and cross-wise imprints of fiber in- tersections due to the fibers hardening. Af- ter maceration in cold chromic acid the "fibrils" can easily be isolated. The diame- ter of the fibrils is 0.3-1.5iu. Furthermore, there is coarser striation caused by air canals. The cross section is somewhat cuneiform. In zinc chloroiodide the fiber remains colorless. In Millon's reagent it turns red. In Cuoxam it dissolves. Cellulose Fibers. Cellulose fibers may be classified as bast, leaf, and seed. Bast fibers. This category comprises flax, jute, hemp and ramie. The fibers are found in the fibro-vascular region of the phloem. The fiber bundles are bound together with cellu- lar tissue and waxy substances. The fibers are usually known as "soft" fibers. Leaf fibers. These are obtained from the leaves of monocotyledonous plants. The fi- bers occur in bundles, i.e., accumulations of individual cells overlapping at the ends and thus forming continuous filaments. These cells are also held together by waxy sub- stances. Unlike the bast fibers they are called "hard." The complete bundles are used for textile purposes. Most plants producing leaf fibers are related. The main fibers belonging to this group are sisal and manilla. Plant hair. The seed hairs of various plants can be used for textile manufacture. The main kinds used industrially are cotton, coco- nut and kapok. Microscopic identification of the various cellulose fibers is not particularly difficult. Characteristics such as "fiber bundles," shape of lumen, whether the lumen is air- filled and the shape of the cross section are adequate. For microscopic examination generally only the ordinary light microscopic methods are used. The electron microscope is used only for fundamental investigation of the cell wall. Cotton (Fig. 2b). The microscopy of cot- ton is very important both as regards mor- 354 FIBERS (TEXTILE) \^='*2 Fig. 2a. Natural fiber structures (from top down): Tussah silk, phase contrast, cross section, 450X. Tussah silk, phase contrast, longitudinal, llOX. Silk, phase contrast, cross section, 225X. Flax bleached, phase contrast, longitudinal, 225X . H. ( Fig. 2b. Natural fiber structures (from top down) : Linen, phase contrast, cross section, 450X. Cotton, phase contrast, cross section, 225X. Cot- ton, phase contrast, longitudinal, 225X. Hemp, phase contrast, cross section, 450X. 355 GENKKAL MICKOSCOI'Y phology and examination with special meth- ods such as micro-chemical reactions and examination in polarized light. The normal bright field methods can be applied very suitably. Phase contrast microscopy with special mounting methods is very important for examining micro structures and for trac- ing mercerizing. Dark field microscopy can also be used for examining surface structures, and the same applies to incident light micros- copy. The fiber is a single cell and looks like an irregularly twisted, collapsed tube with a central canal or lumen. No lumen can be seen at the top of the unbroken fibers. In its undamaged state the basal part of the fiber shows a membrane (part of the fiber that was underneath the seedcoat). The twists often change from Z to S spirals. There are three cross-sectional shapes — round, elliptic and hnear. Dead cotton fibers are often U-shaped in cross section, no lumen being visible. The cross section varies very greatly from fiber to fiber. In this cross section the pri- mary wall, the secondary wall, and the lumen can be distinguished. The outside of the fiber is covered with a cuticula consisting of wax and pectin. The primary wall consists of fibrils oriented at random. The secondary wall consists of vari- ous layers deposited within each other, which are easiest to see in swollen material. The secondary wall also consists of fibrils oriented spirally. The usual swelling medium is copper oxide ammonia. This swells the fiber irregularly. At certain intervals there are zones which do not swell. Often a spiral structure is also left, probably originating from protoplasm resi- dues. In iodo-sulfuric acid the fibers swell and turn blue; protoplasm residues turn yel- low. Ruthenium red, together with copper oxide ammonia, has the result that the cu- ticula, the wall of the lumen and the proto- plasm residues turn red. Zinc chloroiodide stains the fiber reddish violet. Mercerized cotton and bleached cot- ton color with this reagent more strongly than raw cotton. With copper oxide ammonia mercerized cotton forms no globules. The fibrils lie next to one another in each layer of the secondary wall and are oriented spirally relatively to the long axis. The direc- tion of the spiral is often reversed and coin- cides with a reversal in the external twisting. The dimensions of the fibrils have been de- termined with an electron microscope. Damage to Cotton Fibers. Mechanical Damage. The forms of mechanical damage that may occur are irregular cross fracture, bruises and places where the fiber has been unevenly torn off. Chemical Damage. This results in gradual decomposition of the cuticula, shown by an emphasizing of the spirals. In the case of chemical attack by acids, the spirals can be seen all over the fiber. Photo-chemical de- composition is usually more highly localized. Biological Attack. Morphologically this closely resembles chemical damage. As a rule, however, the hyphae of the fungi or bacteria colonies are demonstrable. Flax (Fig. 2a). Flax consists of cylindrical cells which are usually smooth except at the location of the usually X-shaped internode. The cross section is round to polygonal; the cell wall is thick. There is a narrow, well-de- fined lumen. At the end of the cell the lumen is lacking. The lumen has an independent existence. In Cuoxam it forms a small tube that remains even after the fiber is com- pletely dissolved. Zinc chloroiodide colors unbleached flax pale violet, while bleached flax colors dark violet. In Neocarmine W flax turns blue. Hemp (Fig. 2b). Hemp is often difficult to distinguish from flax. The differences are: (1) hemp cells are blunt-ended, forking later- ally; (2) in iodine-sulfuric acid hemp shows cross striation and a l)lue-green color (flax: blue) ; (3) hemp fibers are not as transparent as flax fibers; (4) the cross section is different; (5) the parenchyma tissue often attached to the fiber contains calcium oxalate crystals. With iodo-sulfuric acid a blue-green color 356 IIBEKS (TEXTILE) occurs. With zinc chloroiodide a blue or violet color occurs. With ammoniacal fuchsine a pale red color occurs. In Cuoxam the fiber does not dissolve as quickly as flax, while the primary wall remains. Jute (Fig. 3a). Jute consists of fiber bun- dles overlapping at the ends, thus creating a continuous filament. The fibers are held to- gether with gum, wax and lignin. The surface of the fiber is smooth with few internodes. The lumen is bigger than in flax. The cross section is polygonal. In Cuoxam the fiber does not dissolve, but swells. In zinc chloroiodide a pronounced yel- low coloration occurs. With phloroglucino- hydrochloric acid the fiber turns red. Ramie (Fig. 3b). A flat, ribbon-like fiber. The elementary fiber is 8-10 cm long. The surface is characterized by small cross stria- tion and folds. The cell wall is thin, the lumen flat. In Cuoxam the fiber dissolves completely. With zinc chloroiodide it turns blue-violet. Sisal. Stiff fibers, roughly cylindrical in shape, with a characteristic widening in the middle. Ends blunt and thick. Cross section polygonal. Lumen usually fairly big. The fibers often contain air bubbles. lodo-sulfuric acid gives a yellow color. Defatted fibers afterwards bleached in sodium hypochlorite and rinsed in 90%-alcohol give a red color in NHs vapor. Manila (Fig. 3b). Cylindrical fibers with pointed ends. Cross section irregular to poly- gonal. Lumen round. Cell walls thin. In the fiber bundles there are often stigmata, i.e., thick, silicon-containing plates. They can be easily detected by macerating fibers in chro- mic acid. With iodo-sulfuric acid a golden yellow to green color occurs. Kapok (Fig. 3a). Air-filled, thin-walled, long, smooth cells. The lumen is large. The cross section is round to elliptical. With zinc chloroiodide a yellow color occurs; with phloroglucino-IICl pink. Regenerated Natural Fibers. Besides protein fibers this category contains the cel- lulose fibers. The protein fibers are made from casein, zein, groundnut or soyabean protein. All these fibers show very slight double refrac- tivity and have irregular structures. Un- dissolved or prcmatiu'ely coagulated protein particles can be distinguished in the fibers. Usually they have an irregular skin: air pockets, spherulitic structures and crystals are often found. More important tlian the protein fibers are the cellulose fibers. Most kinds of regenerated cellulose have a skin, which is formed directly in the coagu- lation bath, and a medulla. In cross section the fibers are irregularly sinuate. They show clear longitudinal striation. The structure, skin, medulla, are particu- larly easy to see with the polarizing micro- scope as the skin has a higher specific double refractivity than the pith. Irregular double refractivity is also to be seen in the cross section. The refractive index differences occurring in the cross sections are readily demonstrable with an interference microscope. The skin is then particularly easy to see. Phase contrast microscopic examination of cross and longi- tudinal sections is also useful for structural examination. With the electron micro.scope it was possible, in addition to the skin, to detect a very thin membrane, the cuticula (Kassenbeck). In the medulla, fibrillar struc- tures are found. The elementary fibrils are 70-90 A in diameter. These elementary fibrils form packs of various sizes. Electron microscopic examination shows the skin usually to have a spongy structure. With thallium and iodine reactions Hess showed that viscose fibers had crosswise stri- ation in the fibrils. In addition to a "small" period of approx. 150 A, this .striation showed a large period of 550-670 A. In many cases, both light and electron microscopic examination shows the skin to consist of several layers. Casein and Milk Wool. The fiber section is round with several scallops. There is slight longitudinal striation. 357 GENERAL IVIICHOSCOPY r\ \ ' ' __ J , ^^r; 5^^ 'y V ^ -^f« ! ^^U §^t»m£3r%,0m«i^ Fig. 3a. Natural Fiber Structure (from top Fig. 3b. Natural fiber structures (from top down): Jute, phase contrast, cross section, 450X. down): Manila, phase contrast, cross section, Jute, phase contrast, longitudinal, 225X. Kapok, 450X. Manila, phase contrast, longitudinal, 225X. phase contrast, cross section, 450X . Kapok, phase Ramie, phase contrast, cross section, 450X . Ramie, contrast, longitudinal, 225X. phase contrast, longitudinal, UOX. 358 FIBERS (TEXTILE) Reactions: In zinc chloroiodide the fiber Usually it is necessary to study the fibers turns yellow. In Millon's reagent a red color by means of longitudinal or cross sections, occurs. In Cuoxam it is insoluble, but turns It is advisable to assess the importance of all blue. divergencies and inhomogeneities by means Wscose jRai/on (Fig. 4b). The best medium of microtorsion, microtension and micro- for examining viscose rayon and copper wear tests. rayon is glycerin, mineral oil (n = 1.46) or In synthetic fibers the following diver- monobromonaphthol (n = 1.66). Striation gencies and inhomogeneities in structural is pronounced in longitudinal specimens. This details may occur: varies considerably. The cross section shows (1) Places with a lower specific bire- many irregular scallops. A "skin" is clearly fringence than the rest of the fiber. In both visible on the cross section. torsion and tensile tests these places prove to In Cuoxam viscose and copper rayon are be weak. Their effect on fiber characteristics soluble. In zinc chloroiodide the fiber colors is least if many such places occur in close blue to red-violet but eventually turns black, proximity. In all specimens necks form at With Neocarmine W viscose rayon turns red- these places, violet and copper rayon blue. (2) Divergencies in birefringence caused Acetate Rayon (Fig. 4a). The surface of the by the skin. A skin of irregular thickness is fiber is smooth. The cross section shows some found in many cases with this group of fibers, round lobes visible as longitudinal ridges in Especially if the skin has a higher double longitudinal specimens. In polarized light the refractivity than the medulla, an mterrup- fibers are weakly double refractive. In cross tion in it may greatly affect the strength, section some parts are dark and others light (3) Irregular hirefringence divergencies. between crossed Nicol prisms. This can be Small places which are either more bire- explained by the orientation of the elemen- fringent or less double refractive than the tary micelles. rest of the fiber may greatly affect its char- Acetate rayon is soluble in acetone. If acteristics. These divergencies often occur Vinyon, which is also soluble, is present it is together with great or small divergencies in better to use acetic acid, which leaves Vin- cross sectional shape. yon unaffected. With zinc chloroiodide a pro- (4) Stress concentrations. WTiere there are nounced yellow coloration occurs (nylon also many inclusions in the fiber, for instance in turns yellow!). In Cuoxam the fiber is in- the case of some large matting particles, soluble. Acetate rayon is also soluble in satu- stress figures occur. At such places fibrillar rated phenol solution (as is nylon). spHtting of the fiber may occur. Synthetic Fibers. Microscopic examina- (5) Spherulites. Spherulitic structures are tion of synthetic fibers employs practically found fairly commonly in these fibers. When all microscopic methods. Birefringence meas- small they nearly always cause fibrillar split - urements are important for detecting differ- ting. Large specimens or a number of small ences in specific birefringence. These differ- ones in proximity cause cross breakage. If ences have been partly correlated with there is a skin they cause less harm, physical-mechanical characteristics. (6) Isotropic particles. On the whole the In addition to this birefringence deter- same applies to these as to spherulites. Fi- mination it is very important to judge the brillation usually follows the occurrence of homogeneity of the fiber with the phase con- these inhomogeneities. Their effect is great- trast and interference microscopes. The finer est in skinless fibers. If no stress concentra- structures, however, can be rendered visible tion exists around these particles they are of only with an electron microscope. Uttle importance. 359 GENERAL MICROSCOPY Fig. 4a. SjTithetic fiber structures (from top down): Acetate rayon, phase contrast, cross sec- tion, 450X. Acetate rayon, phase contrast, longi- tudinal, 225X. Cuprammonium raj^on, phase con- trast, longitudinal, 450X. Cuprammonium rayon, phase contrast, longitudinal, 225X. 360 Fig. 4b. Synthetic fiber structures (from top down): Viscose rayon, phase contrast, cross sec- tion, 450X. Viscose rayon, phase contrast, longi- tudinal, 225X. Nylon, phase contrast, cross sec- tion, 450X. Nylon, phase contrast, longitudinal, 225 X. FIBKKS (TEXTILE) (7) Air pockets. Both elongated, spindle- shaped and unelongated air pockets occur. Usually they are characterized by a wide light refractivity zone and a pronounced stress figure. Fibrillar spUtting in torsion and tensile tests is usually the result. With orig- inally negatively birefringent fibers, such as Orion, a positive form of birefringence may occur through strong fibrillar splitting if there is a skin. (8) Fibrillar fiber structure. Most synthetic fibers have a fibrillar structure. In most cases this can be seen only with an electron micro- scope from rephcas or thin cross sections. In Orion and polyvinylalcohol the fibrillar structure is to be clearly seen with a Ught microscope. With Terylene and Dacron it is visible with a Ught microscope, but is clear only under an electron microscope. The fi- brillar structure of nylon can be seen only with an electron microscope. (9) Surface structure. This refers in the first place to variations in cross section. In addition, however, small structures such as the fairly common spiral grooves are im- portant. Frosted fibers may have an irregu- lar surface through the protrusion of frosting crystals. These can best be detected by shadowing the fiber slightly and examining it with a normal reflector microscope. (10) Lacunose effect. Elongation of syn- thetic fibers until necks are formed may cause more or less spindle-shaped opaque spots in the fibers. These spots consist of a large number of small cracks in close prox- imity. Except in the case of cold drawing, the effect is also found normally in some fibers. Nylon (Fig. 4b). Homogeneous, optically "empty" fiber. In polarized light, variations in double refractivity A = 56-62. There is sometimes a skin. The cross section is round. In zinc chloroiodide nylon turns yellow and later orange to brown. In saturated phenol solution it is soluble. In 50% formic acid nylon is insoluble at 80°C, Perlon is soluble. Orion (Polyacnjlic) (Fig. 5a). The cross section is dumb-bell shaped to irregular. This causes striation in longitudinal speci- mens. In polarized light a positively bire- fringent skin and a negatively double refract- ing pith can be disthiguished. Double refractivity varies from A = -f-3 toA = —10. In various types the "pith" contains dark particles which are spindle-shaped to round. These may be double refractive, but also iso- tropic. Air pockets are also found. In satu- rated phenol solution the fiber is hardly visi- ble. Upon heating, visibilit}'' increases. In dimethylformamide the fiber dissoh-es when heated. Acrilan {Foly acrylic). The cross section is mostly kidney-shaped. Usually there is a pronounced skin. The pith often contains isotropic particles and air pockets. The fiber is weakly negative birefringent A = approx. 4. X 51 (Polyacrylic) . The surface is fairly smooth. The continuous filament is optically clear, the staple fiber is less transparent. The cross section is round with a very pro- nounced, fairly thick skin. The birefringence is fairly low: A = approx. 4. Dynel (Polyacrylic) (Fig. 5a). Fairly smooth fiber with a ribbon-shaped cross sec- tion. Under a polarizing microscope bire- fringence varies from A = approx. +6 to A = approx. —6. Polyvinylalcohol 11 {"Kuralon") (Fig. 5a,b). Fiber with the most pronounced skin and pith. The pith contains manj'' spindle- shaped to round bodies which are partly double refractive, partly amorphous. In some cases there are air pockets. The cross section is dumb-bell shaped to irregular. The fiber is weakly birefringent and })oth positive and negative birefringence is found. Dacron, Terylene (Polyester) ^ Fiber with a fibrillar structure which can only be seen in longi- tudinal sections. Otherwise ho- mogeneous. Birefringence 115 Cross section : Round 361 GENERAL ^MICROSCOPY ,^ r:3« :--^-'^-i^3^: .•w*" ''« 'v'9'r^ i#Jii£4^2i3 s-< / W^»^ f/f'<, ,f wiiiidiiiiwuoM (Ill I mv''SmimmtlKmimmimtm Fig. 5a. Synthetic fiber structures (from top Fig. 5b. Synthetic fiber structures (from top down): Orion, phase contrast, cross section, 450X. down): Kuralon, phase contrast, longitudinal, Orion, phase contrast, longitudinal, 450X. Dynel, 225X. Perlon, phase contrast, cross section, 450X. phase contrast, cross section, 450X. Kuralon, Dacron, phase contrast, cross section, 225X. phase contrast, cross section, 450X. Dacron, phase contrast, longitudinal, 225X. 362 LNDUSTRIAL RESEARCH Creslan. Cross section: Round, bean or dogbone; negative birefringent, A approx. 4. Verel. Cross section: Dogbone; bire- fringence low negative, ap- prox. 1. Zefran. Cross section: Round; bire- fringence low negative, ap- prox. 1. REFERENCES Matthews, Momersberger, "Textile Fibers," Wiley, New York, 1954. LuNiAK, "Die Unterscheidung der Textilfasern," Zurich, 1954. Herzog, "Handbuch der Mikroskopischen Tech- nik fiir Fasertechnologen," Berlin, 1951. Herzog, "INIikrophotographischer Atlas der Tech- nisch-wichtigen Pflanzenfaser," Berlin, 1955. Stoves, J. L., "Fiber Microscopy," London, 1957. WiLDMAN, "The Microscopy of Animal Textile Fibers," WIRA, 1954; "Proceedings Inter- nat. Wool Te.xtile Res. Conf." (Austr. 1955, Vol. F). Wolff, Tobler, F.v.G., "Mikroskopische Un- tersuchung Pflanzlicher Faserstoffe," Leip- zig, 1951. Turner, "The structure of Textile Fibres," Man- chester, 1953. "Proceedings Electron Microscopy Conference," Stockholm, 1956. "Proceedings Electron Microscopy Conference," Berlin, 1958 (at printers). Annates Scient. Textile Beiges, Nos. 1-4, 1955. J. ISINGS INDUSTRIAL RESEARCH Both light and electron microscopy are being used extensively for research in many technical operations which are far removed from the long-established applications in the biological sciences. Almost every industry, in some phase of its operation, has need for these tools and associated techniques. Ac- tually, microscopy has been used in industry for a long time. The light microscope has contributed to the solution of problems in control, development, and research ever since the technological revolution. Recently the electron microscope also has become es- tablished as a supplement to the light micro- scope. Electron microscopy is covered else- where in this Encyclopedia. However, it must be emphasized that both techniques are employed in industrial research in a complementary fashion, and are strongly intcrdcpondcnt. Although application of microscopy is hkel}^ to vary with the particular industry, it is possible to provide basic information about the instruments and techniques in use, which should be of common interest to all. The microscopist has at his disposal a very powerful tool which can be adapted to meet his special needs through the use and development of appropriate teclmiques. It is essential, however, that he not only be familiar with the instrumentation, but also have an adequate knowledge of the disci- plines appropriate to his specific problems. The purpose of this article is to discuss briefly the more important aspects of the microscope as a tool, together with related techniques, and to provide references which deal comprehensively with these subjects. The most obvious function of microscopy is to reveal directly the fine-scale structure of a subject by means of combinations of lenses which resolve and magnify structure. Each of the various types of microscope shares this function. Some are capable of revealing ex- ceedingly fine structure, whereas others can obtain higher contrast in the image, make certain measurements, or accept special specimens. The foremost problem in microscopy is the obtaining of contrast to reveal significant structure at high magnification. Much progress has been made in this area through new techni(iucs of specimen preparation as well as b}' advances in optical design of the microscope. The maximum useful magnification of lenses has not been greatly extended in recent times, but modern microscope con- struction has made it easier to realize the full potential of the instruments. No longer is it 363 GENERAL MKKOSCOPY necessary for the microscopist to be pre- the numorifal aperture of the system by occupied with optics; nevertheless, there arc 1000. This rule is based upon the relationship certainrules which must be observed in order between the resolving limits of the eye and to realize the best instrumental performance, the microscope system. The resolving limit of the human eye is 0.2-0.3 mm at the Resolving Limit, Useful Magnification, normal viewing distance of 25 cm. Therefore and Contrast ^q render visible the structure resolved by a Detailed accounts of the factors governing lens of 1.4 numerical aperture, a minimum resolution, and, in turn, useful magnification magnification of 1000 X is required. If a have been given in many publications (1, 2). photomicrograph is going to be viewed at The resolving limit of a microscope may be distances greater than 25 cm, then a mag- defined as the ability to register two objects nification exceeding that indicated by this whose centers are separated by the least rule is warranted. distance, d, at which the objects may be Until now^ we have been dealing with detected as distinctly separate. The mini- lateral resolution, ignoring vertical resolu- mum object diameter is usually not less than tion, or as it is usually termed, "depth of twice the least detectable distance of separa- field." As lateral resolution increases, depth tion, the latter varying directly with the of field decreases. Therefore, to perform an wavelength of light, X, and inversely with observation of structure in depth, usually the refractive index, n, of the medium in some compromise is necessary. Obviously, which the object is immersed and with sin it is not always of greatest advantage in a a, where a is one-half the angular aperture study to exploit upper resolving limits of a of the objective lens. Thus, system. The rendition of contrast also is of prime d = - — : — . consideration. Although a subject may con- 2m sin a . , , , i . i ■ tarn structure on a resolvable scale, this The resolution of a microscope is controlled structure may not be visible microscopically. simultaneously by the numerical aperture This lack of visibility is frequently due to (N.A. = n sin a) of the objective and con- the inherent optical properties of the speci- denser lenses. In fact, the numerical aper- men, namely, that the structure possesses ture of a system may be expressed as the the same light-transmitting or reflecting average of these two numerical apertures: qualities as the surrounding or background ,, . , XT A medium. Contrast can be improved, both by TVT . / . \ N.A.obi + N.A.cond . , , , N,A. (system) = . optical methods and by specimen prepara- tion. For highest resolution, oil immersion lenses A microscope "sees" structure because of are available today with numerical apertures the optical characteristics inherent in a given up to 1.4, and a corresponding resolving limit specimen preparation which interact with of the order of 0.2 micron. Since no lens, the incident illumination and modify it. objective, or condenser can exceed a nu- Consequently, contrast may be improved by merical aperture of 1.0 in air, it is essential better optical monitoring of these altera- to use high index immersion oil on the con- tions or by enhancing the degree of light denser as well as on the objective to realize alteration by the sample through suitable the ultimate in resolution and magnification preparative methods. The most commonly from a light microscope. encountered optical factor controlling con- i The useful magnification of a microscope trast is the numerical aperture of a system, may be roughly calculated by multiplying As with depth of field, there is a loss of con- 364 IM)ISIHI\L HKSEAKCH trast accompanying an increase in the light intensities observed in the field, as in numerical aperture. Frequently there must phase microscopy. Interference microscopy be a compromise toward lower numerical is very new and has hccii used principally by aperture, this time to achieve desired con- the biologist, but should find application in trast. Optical aids to increase contrast have industry, especially in the study of trans- been de^'eloped and have pro\'ed of special parent films comprised of structures ha^'ing interest to the biologist whose living prepara- small differences in refractive index or thick- tions are best studied in that condition, ness. Nevertheless, the industrial microscopist also Polarized-light microscopes are useful in has a persistent need for these special tools, increasing contrast and in making optical which include phase contrast, interference, measurements on crystalline and paracrys- and polarizing microscopes, and filters. talline subjects. Differences in orientation of Phase, Interference, and Polarizing crystalline and fibrous materials are es- JMicroscopes. The phase microscope takes pecially well revealed. A polarizing micro- advantage of the optical-path differences scope is unicjue in that specimens are studied within the specimen which result in the while illuminated with polarized light. A existence of phase differences among the polarizer, transmitting light vibrating in one light waves transmitted by the various por- plane, is usually placed between the con- tions of the specimen. Further phase changes densing lens and the source of illumination, are introduced in the light passing through and an analyzer, also of polarizing material, the optical sj^stem and add to the phase is located in the tube between the eyepiece differences created by the specimen and and objective. Generally the polarizer and thereby render the object visible (3). Briefly, analyzer are used in the crossed position so the diffracted rays are altered by a phase- that their respective planes of vibration are retardation coating at the back focal plane perpendicular. In this position, the field of of the objective, the intensity of the un- view is completely dark, and the sample is deviated light is reduced by a filter in the visible only by virtue of any effect it might form of a ring at the back focal plane of the have on the original plane of vibration of the objective. The deviated beam interferes with light with which it is illuminated. Figure 1 the diffracted rays at the focal plane of the shows an image produced by anisotropic eyepiece to form an image, the contrast of crystals under polarized light. Optical which is derived from these exploited phase properties of substances have been measured differences. Special phase microscopes, both for many years by means of polarized light, of transmitted- and vertical-illumination Mineralogists, chemical microscopists, and types, or components for conversion of exist- crystallographers use ciualitative and quanti- ing microscopes, are available commercially tative optical data such as refractive index, from several manufacturers. The resolving birefringence, optic sign, extinction, etc., as limits of phase objectives are not as high as an adjunct to other information in the identi- comparable objectives for conventional fication of crystalline materials. Although bright-field work, but the increase in con- these methods are highly useful in industrial trast may overshadow the lack in resolving studies, no attempt will be made to present power. any detailed information in this area. Ex- The interference microscope, as the name cellent coverages of these techniques appear implies, produces contrast in the image by in the literature (4, 5). Polarized-light optical interference. Contrast is apparent microscopy is eciually applicable to either because of differences in color (when white transmitted or ^'ertical illumination. In light is used) rather than in differences in metallographic and ceramic research, the 365 GE.N EH A 1. >I I ( KOSCOP Y Fig. 1. A photomicrograph of starch grains under polarized light. The black cross exhibited by each starch grain indicates that it has anisotropic crystals radiating from the center. Such a struc- ture is called a spherulite. latter is employed to enhance contrast and to compare orientation among the grains of polished and etched specimens. It has been found especially useful in locating corrosion and other reaction products. Dark -Field Illumination. Dark-field illumination is effective in obtaining contrast and for observing the colors of structures. Incident dark-field has been used very suc- cessfully in the detection of small quantities of materials lying on the surface of a sub- strate, particularly when sparsely dis- tributed. The dark-field condenser is con- structed to prevent any light from entering the objective but that which is scattered by the specimen. The specimen is illuminated by a hollow cone of light of too large an angle to permit the direct beam to enter the ob- jective. The dark-field image is characterized by a dark background with more or less bright structures revealed in the subject. There is an intermediate case of hollow conical illumination in which a portion of the undeviated beam is allowed to enter the objective, in order to enhance contrast. This is discussed in further detail later in this article. Improvement of Contrast by Speci- men Preparation. As mentioned previ- ously, contrast can be improved by speci- men-preparation methods. Staining, etching, metal shadowing, selecting appropriate mounting media, and replicating are some of the sample manipulations used to increase the contrast of the preparation. Staining (6) has been used in biological preparation for years, and it is feasible to extend staining techniques to industrial samples, even metallurgical preparations. Some techniques are already developed, but frequently it is required that a technique be devised to satisfy a specific need. In view of the numerous stains now available and also the known reactions which yield colored reaction products, usually it is not difficult to select or develop a suitable staining pro- cedure. Stains are employed to dye selec- tively certain structures of interest. They may be merely selective in the sense that they reveal structure without being recog- nizably indicative of chemical composition, or on the other hand, they may be strictly specific for a chemically reactive ion or group. In the latter case it is obvious that much is to be gained in relating the dis- tribution of the chemical substance which reacts with the stain to the over-all structure of the specimen. Thin sections sliced with a microtome (6) are quite appropriate for staining. Even in the unstained condition, microtome-pro- duced thin sections are inherently more contrasty subjects with less confused struc- ture than is the larger sample from which they were cut and thus permit the observa- tion of the internal morphology of the sample. The biologist has pioneered in the field of microtomy, but again this is an effective technique for the industrial worker who need not be limited to employing the techniques of the biological micromotist. Because, in nonbiological problems, it is not 366 INDUSTRIAL RESEARCH generally required that the morphology of a the best possible instrumental performance living specimen be preserved, it is permissible cannot be realized unless certain precautions to be less inhibited as to embedding tech- are taken regarding adjustment, alignment niques, in particular, the solvents, tempera- of components, and the choice of optics and tures, and embedding media employed. illumination. Metal-shadow casting (7) is a techniciue by Fundamentally, a microscope consists of which a thin film of metal is vacuum evap- an objective, eyepiece, and condenser, and, orated onto the surface of a sample or a as such, is called a compound microscope, replica of that surface. This method creates a These components are held and manipulated subject on which differences in surface on a stand which is usually equipped with a morphology are evident due to an increase stage to hold the specimen. The condenser in reflectivity or absorption of the surface directs light to the specimen, the objective and/or a shadow effect. Metallization of receives the light changed by the specimen specimen surfaces by vacuum evaporation and produces the initial magnification of the has been used extensively in electron mi- image. The eyepiece, through which the croscopy; however, there are many applica- specimen is viewed, magnifies the image tions of the technique in light microscopy, created by the objective. The total magnifi- particularly in conjunction with surface cation of the viewed image is a product of replicas. Details of replication, metallizing, the magnifications of the objective and eye- and shadowing by vacuum evaporation are piece. There are many variations in types of given in the section on specimen preparation, objectives, condensers, and eyepieces. For This discussion of resolving limits, useful example, in vertical illumination a single magnification, and contrast is by no means lens may serve at the same tune the purposes complete; nevertheless, it is believed that the of condenser and objective. Again, since factors most significant to industrial mi- many types of illumination are required, croscopy were mentioned. Further detailed condensers are individually designed to information may be fomid in the references achieve specific results, cited or in other sections of this publication. Objectives are the most expensive com- ponents of a microscope and from this Alignment, and Related Methods of standpoint, as well as from that of their role Illuniination j^^ image formation, deserve the concentrated Microscopes and methods of illumination attention of the microscopist as to then- selec- (1, 2) used with them cannot very well be tion. Objectives are classified with respect to divorced in any discussion of the principles optical correction. Some are chi-omatically involved. Consequently, they will be treated corrected to bring two wavelengths of light together; however, emphasis will be placed into focus at the same plane, and also cor- first on microscope function and construe- rected spherically for one wavelength. Ob- tion. Later, attention will be shifted to jectives having this degree of correction are methods of illumination, and finally the called achromatic objectives and are only selection of microscopes for nonbiological about one-third as expensive as apochi-o- application will be considered. matic objectives which are chi-omatically Although there are many types of micro- corrected for three wavelengths and spher- scopes, they are closely related in function, ically for two. Unless the microscopist is well and possess comparable components. There- grounded in the use of the microscope, fore, an understanding of the basic construe- images inferior to those obtained from tion and operation of one microscope may be achromats are likely to be realized from apo- applied in some degree to others. However, chromats. Achromats have smaller numerical 367 GENEKAL MICKOSCOPY apertures than corresponding apochromats, but greater depth of field, less curvature of field, and a better lendition of contrast. These factors make achromats easier to use. There are other objectives which are of iiitennediate correction, i.e., they are not as well corrected as apochromats but better so than ahcromats. These are called fluorite objectives. Generally lower magnifications do not require the employment of apochi'o- mats; however, the latter must be used at the highest magnification in order to attain the ultimate resolution. Consequently, it is often advisable to acquire a set of objectives by which the lower magnifications are achieved with achromats, intermediate mag- nifications wtihout oil immersion, by fluo- rites, and the ultimate resolution at the highest magnification by an oil-immersion apochromat. Objectives are classed also as to magnifica- tion, working distance, focal length, and numerical aperture. All of these are inter- related. Generally, the longer the focal length, the smaller is the magnification and the numerical aperture, but the greater the working distance. The exception is a special type of long-w^orking distance objective of high N.A., which has a front lens of un- usually large diameter. Even in this case, the depth of field is low. Some objectives are especially designed for the observation of particular samples. Some are optically cor- rected for use on a sample covered by a coverglass of a specified thickness ; others are not corrected in this sense, and are to be used on uncovered specimens. Some are constructed so as to be strain-free and suit- able for polarized-light studies, and yet others are made to perform observations by phase contrast, dark field, etc. It is of ut- most importance that the potential worker in microscopy be cognizant of the demands to be placed upon the objectives, so that a judicious selection may be made. Excellent publications are available, which provide detailed information on microscope optics (8). Newly computed lens formulas have made it possible for several manufacturers to place high-quality flat-field objectives on the market. These are not mentioned in the above-cited publications, but one should look into their description by the manufac- turer. They yield images which are in focus over the entire field of view, making it pos- sible to obtain photomicrographs of large areas with objectives of high numerical aperture. Eyepieces (oculars) are designed by the makers to be used with their microscopes and objectives. An eyepiece of one manu- facturer should not be combined in a sy.stem with, the objective of another unless suitable corrections in tube lengths are possible. The following chart illustrates the magnitude of discrepancies which exist among microscope tube lengths and ocular focal plane positions supplied by various manufacturers. Make Mechanical Tube Length, mm Location of Lower Focal Plane of Ocu- lar Below Upper Rim of Tube Leitz 170 18 Zeiss 160 13 Bausch and Lomb 160 12 Reichert 160 15 Several types of oculars are available, but it does not fall within the scope of this article to describe them in detail. However, it should be pointed out that specific objec- tives may require matching oculars. For example, to compensate for the chromatic undercorrection of apochromatic objectives, certain eyepieces are overcorrected. Log- ically these are called compensating oculars; they should always be used with apochro- mats. Other oculars which are intermediately corrected are recommended by several man- ufacturers to be used with their objectives. Examples are Hj^perplane and Periplan eyepieces. Oculars may be provided with net reticules, scales, etc., to facilitate making 368 iNUisntiAi. |{i;si;ai{<:h measurements under llic rnicroHCf^fM!. Care should be exercised in e()in})iriiiif^ oculars and objectives to juoid exc(!(;diiif5 useful magnifications h)ased upon th(! previously discussed criterion of numerical aperture. This is especially true when measurements of images are made. The halo around an over- magnified image may cause the structure to appear much larger than it is in reality. Condensers direct light to the specimen and, along with the objective, determine the numerical aperture of the microscope sys- tems. Therefore, their numerical aperture must be adjustable to that of the objective. The condenser lens system should also be corrected chromatically, spherically, and be free of coma. Coma results when different regions of a lens have different magnifica- tions, creating spurious tailed images. A condenser which has been corrected for all three is referred to as an achromatic-aplan- atic condenser. Ordmarih" the effective numerical aperture of a condenser is subject to alteration by means of an adjustable iris- aperture diaphi-agm placed in an appropriate position to define the angle of the cone of light illuminatmg the specimen. This is im- portant to reduce glare as well as to achieve resolution. Alignment and Illumination. The use of the condenser is so closely allied to meth- ods of illumination and microscope ahgn- ment that at this pomt it becomes desirable to consider all three. In fact, condensers are designated according to methods of illumina- tion. There are both bright-field and dark- field condensers, phase contrast, and hght- polarizing condensers. Certam requirements for alignment and illumination have been expressed by several authors (1, 2, 9). One of the best step-by-step charts available is one published by Sliillaber (1). Briefly, the centration of light som-ce and optical com- ponents is accomplished, a field-limiting aperture is imaged by the condenser onto the sample plane and adjusted to coincide with the field of view, and the aperture (condenser) diaphragm is used to control glare. In general, these steps are suited to bright-field aiiginnent in both transmitted and \'('i-tical illumination. When a sample is illuminated, the light nujdified by the sample creates the image. If excessive amounts of undeviated light pass through a specimen, inherently low in con- trast, and enter the objective, the deviated beam is unable to compete favorably and, therefore, the image is comparati\'ely weak. If the ratio of the intensities of the undevi- ated to the deviated beams is reduced, the image can be made relatively stronger. The latter condition may be achieved by illumi- nating the sample with a hollow cone of light, whose inner angle is too great for undeviated light to enter the objective lens; only the fight widely scattered by the sample is seen in the image. This is dark-field illumination. Although highest possible numerical aper- tures are employed in dark field, the image is not as informative m some respects as others observed under conditions of lower resolution. This is because much of the devi- ated light from the specimen is also pre- vented from entering the objective. Dark-field condensers suited to both trans- mitted and incident illumination are avail- able. It is possible to employ varj-ing degrees of hollow conical illumination approaching dark field, and permitting some of the un- de\'iated beam to enter the objective. More often for expediency in obtaining contra.st, the undeviated beam is reduced by adju.st- ing the bright -field conden.ser diaphragm to illuminate the specimen with a narrower solid cone of light. Resolution suffers in the latter method, and not m the former. Several manufacturers make pha.se-contrast con- densers which illuminate the sample with a hollow cone suitable for semidark field, and which may be u.sed effectively in producing image contrast in conjunction with ordinary objectives. Besides dark-field and bright- field effects, there are special condensers for optical staining, that is, they create different 369 GENEKAL MICROSCOPY colors in the image to promote visibility and The universal microscope is designed foi rendition of contrast photographically. photomicrography as well as for viewing. It was stated that all components of the This is essential since the microscopist'a illuminating and microscope system should recorded data are in the form of written be centered to obtain good alignment; how- observations and photomicrographs. The ever, this is not always true. Sometimes it Leitz Ortholux is a typical microscope of this is advisable to illuminate the specimen type. It may be attached to the Aristophot obliquely. Several condensers are provided II photographic camera. This camera is with centerable aperture diaphragms, which equipped with a reflex housing which facili- when decentered, direct light at an angle tates the composition and focusing of the onto the specimen. The effect of optical image prior to taking the picture, and in- shadowing is thus achieved. eludes the conventional shutter, bellows, Equipping a Microscopy Laboratory, and camera back. The microscope body con- The requirements for equipping an indus- sists of an inclined binocular tube for viewing trial microscopy laboratory are different for and a single vertical ocular tube for photo- each situation. Nevertheless, the authors micrography. The stage, nose pieces, il- hope that the following comments, taken luminators, tubes and condenser rack are from their experience, will be helpful. The attached to the stand by dovetail devices versatile universal research microscope is which make it simple to change setups. The very well suited to industrial application, new Ortholux UAM stand has two illuminat- Interchangeability of parts, freedom of ing systems, one for incident illumination alignment, choice of optics and illuminations and the other for transmitted illumination, impart the flexibility required to examine the which may be used individually or in com- wide variety of samples found in most non- bination. Either illumination may be ar- biological laboratories. Of course, if one ranged for dark- or bright-field, phase- type of sample must be dealt with continu- contrast, fluorescence, polarized-light, etc., ously, it may be advantageous to acquire a studies. more specialized instrument. For example, a In addition to the one illustrated, Reichert, metallograph is definitely needed in a metal- Zeiss, and American Optical, among others, lurgical laboratory. This is a microscope of provide instruments which are somewhat inverted design which will accept relatively similar, large polished metal specimens. Metallo- graphic examination falls within the ca- Approach to the Problem and Applica- pabilities of the universal microscope, but the tions latter is not so convenient to use routinely for this purpose. Assuming a reasonable knowledge of avail- One objection to the universal-type instru- able techniques and the operation and ment is that, although various arrangements limitations of the equipment, increased pro- of optics are available, it can be used only ficiency in the field of microscopy can only for one operation at a time. This is not be gained through research experience. In serious when there is only one microscopist, fact, it is difficult to learn otherwise of the and when more become involved, the prob- diverse applications within a particular area lem can be solved by acquisition of another of investigation. Many phenomena, although stand of the same make. At this time, all manifest on a larger scale, actually take the optical components need not be dupli- place on a microscopic scale, and therefore, cated; only those necessary for special are most profitably studied microscopically, problems could be added as indicated. The background, interest, and ingenuity of 370 INDUSTRIAL RESEARCH the investigator influence greatly the course of the research. For example, a chemist using a microscope may be motivated to set up a wet analytical laboratory centered about a microscope under which he 'observes reac- tions between minute (juantities of sample and reagents. In contrast, a metallurgist may make microhardness tests on his par- ticular specimen with the aid of a micro- scope. Whatever the specific purpose, one of the most important abilities de\Tloped by experience is that of devising techniques which will yield more significant data from the samples to be studied. Frequently, the development of techniques involves only the adaptation of existing ones to specific needs, but sometimes leads to a drastic departure from conventional methods. Appropriate techniciues are often the products of crossing disciplines, instrumentation, or sample ma- nipulation. It maj^ be found that there is a need to resort to one or all of these expedients during the course of a research project. This article is based on experience which can be classed as applied industrial research, part developmental and part trouble shoot- ing. There is little difference in the general approach to either type problem. Sample selection is an important pre- requisite to the experimental work. Because of the limitations m the size of the micro- scopical sample, utmost attention must be devoted to the problem of taking a repre- sentative sample. Sampling should be ac- complished with critical awareness of the prior history of the sample, the seciuence of events occurring throughout the entire process from which the sample is extracted, and the nature of the chemical and physical environments in all stages of the sample history. In general, it is safer to take large samples first, and then later, as decided by more critical examination, to reduce the sample size. This necessitates that observa- tions of gross structure be made first at unit or low magnification and subsequent study carried out at higher magnification. Initially, Fig. 2. The Leitz Ortholux microscope shown in combination with the Aristophot photomicro- graphic camera. for example, the binocular stereomicroscope is indispensable. The decision in the choice of samples is usually followed by sample preparation and the selection of types of illumination. There- fore, the remainder of this section will be confined to comments upon combination of disciplines and sample preparation. The employment of a microscope to study struc- ture does not exclude the application simul- taneously, or at another time, of other disci- plines such as wet chemistry, photometry, absorption and emission .spectroscopy, to name but a few. There should be no restric- tions which prevent the microscopist from bringing any suitable tools to bear upon the problem. For example, microscopy is an excellent tool to select and isolate specimens suitable for identification by X-ray diffrac- tion and other microanalytical methods, and it should be possible to use these freely, where indicated. 371 GENERAL MICKOSCOPY Combination of Disciplines. Chemical Microscopy. Chemistry and microscopy have been used in combination for many years (10). Chamot and Mason in their "Hand- book of Chemical Microscopy" have pre- sented one of the most authoritative works in this field. It is felt, therefore, that no lengthy discussion of their techniques be- longs here. These authors have emphasized the identification of inorganic ions in small specimens by means of microscopical recog- nition of characteristic crystalline forms in reaction products. They recommend using a polarizing chemical microscope to facilitate identification. These methods are covered in some detail in pages 13 to 72 of this Ency- clopedia. Wet Chemistry and Microscopy. Another combination with microscopy is straight wet chemistry on a reduced scale to provide di- rect microscopic observation of the course of the reaction. In this area, the specimen con- tained in small vessels, pipettes, etc., is handled with a micromanipulator while un- der observation. Strikingly quantitative methods have been developed; even ti- trimetry is possible. Spot-Test Reactions and Microscopy. Al- though the use of spot-test reactions (12, 13) in conjunction with microscopy is not gen- erally recognized, it has been found to be quite informative in the authors' laboratory. Spot tests yield colored reaction products which indicate the presence of ions or groups, and are most applicable m inorganic micro- scopical analysis, but also have been found useful for organics. If these reactions are chosen judiciously, distributions as well as species of ions or organic groups frequently may be shown. Some biological stains fall into this category, but very few of these have proved applicable to industrial specimens. More often greater success has resulted from adapting spot tests which had not been used previously in connection with microscopy. The criteria for selecting a spot test have been that it must have the desired selectivity within the bounds of the system, it must produce a very intense color which is, in low concentration, visible microscopically, and it should proferal)ly form a product which does not migrate seriously from the site of reac- tion. Adsorption complexes and precipitates, for example, are the most generally suitable reaction products in this sense. Although many of these reactions have been success- fully employed here, only one typical ex- ample will be cited. The following detailed description of adaptation of a spot test to a metallographic specimen serves to illustrate the nature of the specific difficulties as well as advantages. A steel to which lead had been added to improve machinability was being examined to locate the lead. The samples were polished and etched with picral. The most predom- inant structures were elongated manganese- sulfide inclusions in both the leaded specimen and the control specimen containing no lead, but the leaded specimen exhibited small structures, not evident in the control speci- men, at the ends of the manganese-sulfide inclusions. Consequently, it was immediately suspected that the observed structures were lead-bearing. A search made for a suitable spot-test reaction resulted m the choice of the reagent diphenylthiocarbazone (dithi- zone). Reacted with lead, it yields a water- insoluble red inner complex (11). First the dithizone was applied in chloro- form solution to the sample after the lead had been reacted superficially with potas- sium chromate solution to produce a film of lead chromate ; however, the lead dithizonate was soluble in the chloroform. Further litera- ture search (12) provided a satisfactory means of preserving the reaction product at the site of the reaction. Dithizone is itself water insoluble ; however, it was learned that dithizone would transfer from a chloroform solution to a water solution of potassium cyanide. The chloroform and water are im- miscible, and the dithizone is only sparingly soluble in the aqueous potassium cyanide solution. The reaction product is insoluble 372 INDUSTRIAL RESEARCH in the aqueous solution. If the sohitions are placed together in a container, the chloro- form solution sinks to the bottom of the dish, and acts as a reservoir to replenish the aqueous KCN with reagent 'as it is depleted from solution in the latter. When the surface of the polished and etched specimen was ex- posed to the aqueous solution no migration of the reaction product occurred. The pres- ence of lead chromate was later shown to be unessential in providing sufficient quantity of lead-salt reactant; in fact, it was observed to have migrated disturbingly. Therefore, the potassium chi'omate reaction was omit- ted. Instead it was found that the picric acid and alcohol etchant had created a surface film of nonmigrating reaction product which was present in quantities quite adequate to yield microscopically visible lead dithi- zonate (13). Figure 3 show^s a photomicrograph of a leaded steel specimen treated in this manner. The structures at the ends of the man- ganese-sulfide inclusions were stained red indicating that they are indeed lead-bearing. No other structures reacted to yield this stain. Due to the random nature of the plane of metallographic sectioning, it is possible for some of these structures to ap- pear dissociated from the manganese sul- fide. Examples of other spot-test reactions used to show chemical distributions by micros- copy are rubeanic acid for copper, 5(7>-di- methylaminobenzylidene) rhodanine for sil- ver, and manganous sulfate and silver nitrate for hydroxyl ion. The last example is one in which the hydrox}^ ion triggers a redox reac- tion between manganous and silver ion, pro- ducing precipitates of black manganese dioxide, and black finely divided metallic silver. No insoluble reaction product forms between the reagent and the ion for which the test is performed. This ion merely influ- ences the components of the reagent to react with one another. This illustrates the point that indirect as well as direct reactions which result in insoluble products should be V ^ ' A Fig. 3. A polished and etched leaded steel specimen stained with dithizone to show the dis- tribution of lead. The arrows indicate the location of the lead bearing structures which stained red. investigated when attempting to develop a microscopical spot-test technicpe. New spot-test reagents are rapidly being discovered; however, most of them are used in the conventional manner. It is hoped that increased selectivity by means of more ap- propriate reagents for microscopy will make it feasible to devise relativelj^ simple system- atic methods of analysis. Ultimately, it may be possible to place these tests on a quantitative basis through microspectro- photometry. Physical Methods. Microspectrophotome- try is an example of the combination of physical methods (14). Some absorption spectra may be obtained by this means from microscopical specimens for purposes of identification and anal3^sis. Biologists have been able to gain insight into the chemistry of components of single cells with this tool. Spectra in the ultraviolet, ^'isible, and the near infrared have been studied. This science is in the beginning stages of development, and, therefore, suffers from rapidlj^ changing instrumentation, but holds great future promise. (See the "Encyclopedia of Spectros- copy".) 37a GENEKAL MICROSCOPY In general, microspectrophotometry is ac- complished in two ways. One is by illuminat- ing the specimen with white light, and subse- quently analyzing the light transmitted by the subject. The other is by illuminating the specimen with monochromatic light and monitoring the sample transmittance as the wavelength is varied. Jelley's microspectro- graph (15) is an example of the first method. He uses a transparent replica of a grating to break down into its component wavelengths the light transmitted by the specimen, and records the spectra photographically. The second method requires a monochromatic light source, special optics if the range is to be extended into the ultraviolet and infrared, and suitable sensing devices for the wave- lengths encountered. As usual, as the refine- ments are made, the complexity of instru- mentation becomes formidable. Some of the Perkin-Elmer spectrophotometers are suited to alteration for microspectrophotometry through a commercially available micro- scope accessory. Beckman also has provision for a similar modification of its equipment. It is anticipated that specialized equipment of this type soon will be developed and made available. Microspectrophotometry is especially use- ful for the identification of organic com- pounds or, more particularly, the chemical groups found in these compounds. Spectra in the infrared are more definitive ; neverthe- less, significant comparisons can be made by means of ultraviolet and visible spectra among certain specimens giving information as to the course of chemical reactions. Physical properties of small quantities of materials can be determined microscopically. Melting point, sublimation temperature, hardness, solubility, deformation behavior, and tensile strength are some of these. For example, the tensile strengths of single textile fibers have been measured with the aid of a microscope and a modified micro- tensiometer. The torsion wire tensiometer, conventionally used for surface tension measurements, was modified by the addition of a lever arm. The extension of an attached fiber could be observed directly in the mici'o- scope. This apparatus was found (juite suit- able for exerting force of the proper magni- tude and with the desired control to make tensile tests on individual wood pulp and cotton fibers. The fibers were attached to the tensiometer under a stereobinocular micro- scope (IG). Melting points, sublimation temperatures, and fusion data in general can be accumu- lated with a heating or cooling stage. W. C. McCrone (17) in his book "Fusion Methods in Chemical Microscopy" has presented many of the possibilities in this area. He covers observations of temperature-depend- ent phenomena of crystalline compounds individually and in mixtures. Mixed fusion of the unknown with a standard reference compound has been emphasized. These methods are suited to organic chemical analysis. Fluorescence microscopy (19) involves physicochemical phenomena from which chemical inferences are drawn. Aside from inherent fluorescence to be found in certain specimens, fluorescence can be created by suitable dyes which selectively attach to specific compositions within a structure. These dyes are called fluorochi^omes, and are used by biologists, for example, in im- munological chemistry, to locate structures in organs which are most active in antibody production. The fluorochromes are designed to form true chemical bonds (not adsorption complexes) with the antigenic protein sub- stances which are collected by certain organs. It is proposed that by this fluorochi-ome tech- nique, selected nonbiological substances could be traced thi-ough a chemical and physical process. One of the outstanding advantages of this method is the high sensi- tivity. Dyestuffs in concentration as low as jQ-is g pg^j^ \yQ detected. Microscopes for fluorescence observation are supplied with ultraviolet illumination to 374 INDUSTRIAL RESEARCH excite fluorescence in the specimen. Normal from artifact by the establishment of this optics may be used in most appHcations; relationship. Interpretation is thereby facili- however, sometimes it is desiral)lc to utilize tated when structure is ultimately seen at the shorter wavelengths which are absorbed the same and higher magnifications from the by glass. Then it is necessary to substitute perspective of the image produced by elec- quartz or reflecting optic components plus trons rather than by light, first -surf ace aluminum mirrors in the il- Further discussion of the crossing of dis- luminating system. Even the microscope ciplines could lead to an inappropriately slide should be made of quartz. Most often lengthy section. These examples have been an adequate illuminating system for fluores- cited here, not only because of their indi- cence studies can be set up by employing a vidual value, but also because they ser^'c to mercury arc lamp with an appropriate filter exemplify an effective approach in micro- as a source, together with a dark-field con- scopical research. denser. Only the fluorescent light from the Specimen Preparation. It must be specimen and small amounts of scattered emphasized that specimen preparation is one radiation are thereby permitted to enter the of the most important aspects of micro- objective. If bright-field illumination were scopical research, and therefore, deserves used, the undeviated beam entering the ob- considerable attention. Fine structure on jective would be highly objectionable from one scale or another is present in almost any the standpoint of harm to the eye as well as object. When the order of magnitude of of exciting fluorescence in the cement of the some detail or inhomogeneity of that struc- objective lens. ture falls within the resolving range of the Not always is it necessary to excite with microscope, such material is amenable to ultraviolet light. In some cases, visible light study by the techniques outlined above, which is shorter than that of the anticipated However, without some prior treatment, the fluorescence may provide the excitation. object does not usually exhibit the detail Combining Light and Electron Microscopy, required for exhaustive examination. The Light and electron microscopy can and treatment given a sample to create a speci- should be used in combination. Ordinarily men appropriate for microscopj^ is construed it is inadvisable to proceed directly to elec- here as specimen preparation, tron microscopy without first having done Ordinarily specimen preparative methods light microscopy. This statement is con- are performed to confine observation and to sistent with the recommendation that a reveal structure in the chemical and phj-sical microscopical study should be initiated at sense, both specific with regard to certain lower magnifications than that of which the constituents, and nonspecific. In order to electron microscope is capable. It finds delineate suflficient detail for interpretation further justification in the narrowly limited of the morphology, it is often possible to nature of electron-microscope specimens, use nonspecific preparations. These most Obviously, the sampling problem in electron frequently include treatments which modify microscopy is more acute, in view of the the optical properties of the sample. How- required smaller specimen size. Furthermore, ever, they may not entail changing the a suitable specimen preparation for electron sample, but rather defining certain features microscopy is likely to exhibit structure in of the sample for microscopical viewing. An an unrecognizable form unless the transition example of the latter is surface replication of in the rendition of structural detail is related an untreated sample, and of the former, the first by light microscopy. Very often much reduction of optically confusing surface of the true structure can be distinguished roughness by polishing. These general 375 GENERAL MICKOSCOPV methods are often followed by specific ones the decision concerning specimen prepara- unique to the problem under investigation. tion as has the transparency. Once a representative sample has been There is no hard-and-fast rule which must decided upon, it is advisable, even before be observed stating that all opa(iue samples preliminary preparation, to obtain from it are best studied by reflected light, or con- as much information as is possible. This versely that transparent samples must should be done to decide upon a preparative be studied by transmitted illumination. In procedure, as well as to ascertain the ap- many instances, it is advantageous to make pearance of the unaltered structure, and specimens of opacjue materials to be studied thereby assist in the recognition of spurious by transmitted illumination (e.g., opaque phenomena (artifacts) W'hich may be created particles to be studied in profile) and vice by the specimen preparation. versa; nevertheless, in general, the examina- The data collected from a specimen prep- tion will be carried out using the conven- aration are limited by many factors, some of tional combinations. It may be added that which can be anticipated and others which not all specimens are strictly opaque or are not easily predicted. One of the foremost transparent, and they may be viewed profit- problems is learning to distinguish true struc- ably in either or both illuminations. All of ture, especially that which is significant, these various situations will be discussed from the above-mentioned artifact intro- here. duced by the sample treatment. Obviously, Specimen Preparation for Reflected Light. an accurate prediction of the type of artifact Metallographic, some mineralogical, and likely to be encountered in a specimen would ceramic specimens fall into this category, be quite helpful. This is possible to a certain Several available publications describe in degree in almost any situation, but never detail the conventional techniques in this entirely so. Recognition of artifact is of prime area (19, 20) Briefly, the sample is mounted importance, both during specimen prepara- in a plastic material which aids in holding it tion and in the interpretation of data. It is during subsequent grinding, polishing, etch- hoped that some assistance will be provided ing, and viewing. The mounting medium b}^ mentioning some of the inherent artifact may be thermoplastic and, therefore, can be hazards incidental to the preparative meth- formed in molds under pressure and tem- ods discussed in this section. Nevertheless, perature, or thermosetting, as for example, major emphasis is devoted to the proper the epoxy resins, which can be polymerized choice of technique from the perspective of at low temperatures and under no pressure sample character. to produce a suitably hard mount of soft Logically, specimen preparations may materials. These mounting materials should be classed in two categories according to be examined critically to look for materials whether the material is transparent or which may be mistaken for constituents opaque. The importance of this classification of the sample. For example, some thermo- is brought out by the existence of the two setting formulations contain highly reflective major types of microscopes, which have been particulate structures. The effects of the designed for the express purpose of examin- mounting temperatures and pressures upon ing opaque or transparent specimens. There the sample also deserve consideration. They are other qualities of samples which deter- may induce changes which render the speci- mine specimen preparation, some of which men useless. are size, hardness, roughness, solubility. Sectioning of materials after mounting can elasticity, fluidity, and crystallinity. How- be done with a cutoff wheel. The grinding ever, none of these has the marked effect on and polishing can be carried out with abra- 376 INDUSTRIAL RESEARCH sive papers, and the metallographic polishing metallographic procedure (20, 21). Even performed on wheels covered with cloth textiles have been etched. However, when bearing different grades of abrasives. In reflectivity is low, image contrast has been cutting, grinding, and polishing, flow or cold improved by metallization, work is usually produced, but may be essen- Metallization technicjue (7) consists of the tially eliminated in most specimens by re- vacuum evaporation of metal onto the speci- peated etching and repolishing with the finer men. The metal is evaporated from a heated grades of abrasives. The pulling out of struc- filament, and during the evaporation travels tures is another common difficulty in grind- in straight paths from the filament to the ing and polishing. When this takes place, the surface of the specimen where it deposits holes left in the surface may be misconstrued only on the exposed portions. If the specimen as inherent porosity in the sample. These is inclined to the direction of evaporating are but a few cases, but they serve to exem- metal, a shadow effect is produced. The as- plify a typical artifact incurred by metal- perities receive a heavier metal coating and lographic preparation. the pits less or none. The difference in re- Even small quantities of finely particulate flectivity and resulting image contrast is minerals (21), fibers, textiles, and papers can expressed in terms of surface topography, be prepared by metallographic methods. It is Even when metallizing is done at normal often expedient to study a polished mount of incidence, and no shadow effect is produced, a material rather than to attempt to cut the a more contrasty image will be seen by sample into thin sections. The plane of sec- vertical illumination. Surface relief is present tion is not difficult to control, and if a trans- in some specimens in the polished condition parent mount is made, it is sometimes due to differences in rates of abrasion among convenient, by looking into it, to view the specimen constituents. Metallization has surface of the specimen and relate it to the been used extensively in electron microscopy internal structure visible at the plane of of surface replicas (23). section. The epoxy resins, in particular, Replication plays an important role in the Epon 828* catalyzed with triethylene tetra- microscopical examination of surfaces in a mine, have been found quite satisfactory in wide variety of industrial problems. Replicas mounting soft materials prior to sectioning, may be made from plastic, most often cellu- grinding, and polishing. Frequently these lose acetate in the authors' laboratory, and samples must be impregnated with the they are employed to reproduce surface medium before mountiiig and sometimes this structure. An impression or imprint is made can be done by vacuum impregnation. Im- by pressing the surface of a plastic sheet, pregnation is also possible by an adaptation softened by solvent, onto the surface of a of embedding techniques used in the prepara- sample. The plastic is allowed to dry while tion of biological specimens for electron mi- in contact with the sample. When they are eroscopy (22). By this method, the sample is separated, the plastic retains a negative soaked in a liquid which is a solvent for the imprint of the sample; that is, the topog- epoxy monomer, then in a solution of mono- raphy of this replica is a mirror image of the mer in this solvent, and later in catalyzed sample. Not always is it necessary to metal- monomer which is subsequently hardened by hze replicas to achieve adeciuate contrast, polj^merization. Enough light is sometimes reflected, by sur- Most metals and some minerals are sufh- faces bearing no metal, to give instructive ciently reflective to produce images of high images. Artifacts most commonly found in contrast when etched, and this is a common replicas are bubbles resulting from incom- * Available from Shell Chemical Company. plete contact while taking the imprint. It is 377 GENERAL MICKOSCOPY ;^'»- » «.' Fig. 4. Photomicrographs of replicas of an area on a cylinder liner showing progressive damage during a wear test. Top, Step 1; Center, Step 5; Bottom, Step 6. of interest to note that when replicas are metal shadowed, there is a lack of contrast in structure within the shadows. This arti- fact may be avoided, if necessary, by shadowing from opposite directions. The most widespread application for repli- cas is in the study of surfaces of opaque specimens; however, they are also useful in confining observation to the surface struc- ture of transparent materials. Surfaces of curved objects, large objects which cannot be removed for examination, or parts of an intact mechanical device can be studied by the replica method. Replicas from curved surfaces, especially from cylinders, can be flattened without seriously distorting the structure rendered. An example of the repli- cation of a curved surface in a mechanical device, which was not dismantled, is the study of scuffing and wxar on the cylinder liners in a diesel engine (24). The progress of the wear was followed by replication. An engine test run was interrupted at certain intervals and replicas were taken of the cylinder-liner walls. Because it could not be anticipated precisely w^here the most sig- nificant scuffing would appear, the entire w^all was replicated. These replicas were compared after the test w^as completed, and areas of interest were selected and studied more critically. Figure 4 illustrates that most of the wear in one of these selected areas took place catastrophically between Steps 5 and 6 and very little occurred be- tween Steps 1 and 5. This type of retrospec- tive study offers many advantages. Specimen Preparation for Transmitted Light. Objects viewed by transmitted il- lumination are usually transparent and are studied by means of a biological-type micro- scope. The specimen preparation is or- dinarily held on a microscope slide and covered with a cover-glass. In fact, as was previously mentioned, the objectives of the microscope employing transmitted light are corrected for use with a cover-glass, ^^ery 378 IiNDLSTKlAL RESEARCH often the specimen is surrounded by a mounting medium, such as balsam, placed between the slide and cover-glass. The re- fractive index of the mounting medium often determines what is visible in a specimen. For example, if the external features of a transparent subject are to be visible, the refractive index of the mounting medium must differ from that of the subject. If, on the other hand, the internal structure is to be studied, confusing external features may- be made invisible by placing the specimen in a medium having a matching index of refrac- tion. Figure 5 shows an example; e.g., a sodium chloride crystal which was mounted in a matching medium to reveal brine inclu- sions. Because different levels are visible in transparent or semitransparent specimens, it becomes necessary to make thin specimens so as to confine observation to a thickness within the range of the depth of focus of the objective. Otherwise, out-of-focus structure detracts from the image quality. Small par- ticles and liquids are easily made into thin preparations, but reducing the thickness of large rigid specimens poses somewhat of a problem. Thin sections of materials of the proper consistency to be cut can be pro- duced with a microtome, or sections of minerals and ceramics can be made by modi- fied metallographic techniques (25). The latter involves cutting slices of the sample as thin as possible with a cutoff wheel, attach- ing the slice to a microscope slide and thin- ning it further by grinding and polishing the exposed side. If the section is mounted in a medium with a matching refractive index, distracting surface irregularities are invisible. Only the internal structure is visible. When transmitted and vertical illumination can be employed simultaneously, a suitably highly polished and/or etched surface is not covered with a mounting medium or cover-glass. Thin sectioning with a microtome has been well established in microscopical science Fig. 5. A .salt crystal mounted in a matching refractive index medium to reveal brine inclu.sions. 130X. for many years; however, renewed interest in the development of refined techniques for the purpose of electron microscopy has pointed the way to extending those for light microscopy. Especially the embedding of materials in plastics has been found appli- cable to microtomy for light microscopy. The glass knife, also used routinely by man\^ in electron microscopy, has been found useful in cutting thicker sections for light micros- copy of materials embedded in these plastics. Sections of paper and textiles as thin as 3 microns have been cut routinely by these techniciues. Figure 6 shows a typical 3-mi- cron section of paper. Embedding procedures can be modified according to the nature of the material. In this respect, often it is possible to become less inhibited as to the emfjedding procedures than in the case of biological specimens. Many materials can withstand higher temperatures or solvent and chemical actions which are ordinarily avoided in embedding procedures. For ex- ample, the procedure used in the authors' laboratory for embedding paper commences by soaking it in alcohol. In fact, when the paper was first soaked in water, a swelling 379 GENERAL MICROSCOPY /•'/)/' / ■> ' f'i i ? 'Hi^i^iA . -^^M, £ud ' , ^ >^V,i . / / \ d.4 Fig. 6. A thin section of newsprint. 500 X of the fibers occurred and persisted through- out the embedding process. This swelhng was undesirable from the standpoint of the study being made; but prior perfusion with a sol- vent was necessary to promote impregnation of the paper by the plastic; therefore, alcohol was chosen as a suitable substitute. Deformed structure constitutes most of the artifact found in thin sections, and frequently it can be recognized by examining and measuring the sample both in the em- bedded and unembedded condition. In conclusion, it is apparent that the successful application of microscopy depends upon a great many factors which are con- trolled by the interaction of the microscopist with his instruments. It should again be emphasized that the realization of significant results depends upon the broad experience and ingenuity of the worker, and in his interest in applying various techniques, both orthodox and unorthodox, the latter es- pecially tailored to the problem. REFERENCES 1. Shillaber, Charles Patten, "Photomicrog- raphy in Theory and Practice," John Wiley & Sons, Ltd., London, 1944. 2. Allen, RayM., "Photomicrography," Second Edition, D. Van Nostrand Company, Inc., Princeton, New Jersey, Toronto, New York, London, 1958. 3. Bennett, Alva H., Osterberg, Harold, JuPNiK, Helen, and Richards, Oscar W., "Phase Microscopy," John Wiley and Sons, Inc., New York, 1951. 4. WiNCHELL, A. N., "Elements of Optical Min- eralogy, Part I. Principles and Methods," 5th Edition, John Wiley & Sons, Inc., New York; Chapman & Hall, Ltd., London, 1951. 5. Wahlstrom, E. E., "Optical Crystallog- raphy," 2nd Edition, John Wiley & Sons, Inc., New York, 1951. 6. Krajian, Aram, A., "Histological Technic," C. V. Mosby Company, St. Louis, Mo., 1940. 7. Williams, R. C, and Wyckoff, R. W. G., "Applications of Metallic Shadow-Casting to Microscopy," J. Applied Physics, 17, 23- 33 (January, 1946). 8. Rossi, Bruno, "Optics," Addison-Wesley Publishing Company, Inc., 1957. 9. Stoves, J. L., "Fibre Microscopy," National Trade Press, London, 1957. 10. Chamot, Emile, and Mason, Clyde, "Hand- book of Chemical Microscopy," Vol. 1, Third Edition, John Wiley and Sons, Inc., New York, 1958, and Vol. 2, 2nd Edition, 1953. 11. Feigl, Fritz, "Spot Tests, Volume I, Inor- ganic Application," Elsevier Publishing Company, New York, 1954. 12. Evans, B. S., Analyst, 69, 368 (1944). 13. Gerds, a. E., and Melton, C. W., "New Etch Spots Leaded Steels," Iron Age, 178, No. 9, 86 (August 30, 1956). 380 MICROSCOPISTS AND KESEVKCII MANAGEMENT 14. OsKR, Gerald and Pollister, Arthur W., "Physical Techniques in Biological Re- search," Vol. 1, Academic Press, Inc., 1955. 15. Jelley, E. E., J . Royal Microscope Soc, 56, 101 (1936). 16. Authors' unpublished work. 17. McCrone, Walter C. Jr., "Fusion Methods in Chemical jSIicroscopy," Interscience Pub- lishers, Inc., New York, 1957. 18. Haitinger, "Die Fluoreszenzaualjse in der Microchemil," E. Hain & Co., Leipzig, 1937. 19. Kehl, G. L., "Principles of Metallographic Laboratory Practice," McGraw-Hill Book Co., Inc., New York, 1949. 20. Greave, Richard H., and Wrighton, Har- old, "Practical Microscopic Metallog- raphy," Fourth Edition, Chapman and Hall, Ltd., London, 1957. 21. Dillinger, Lee, and Sclar, Charles B., "A Method of Mounting Minute Particu- late Samples of Opaque Ore Minerals for Quantitative Microscopic Analysis," Eco- nomic Geology, 55, 187-191 (1960). 22. Newman, S. B., Borysko, E., and Sw^erd- Low, yi., J. Res. Nat. Bur. Standards, 43, 183 (1949). 23. Wyckoff, Ralph W. G., "Electron Micros- copy, Technique and Applications," Inter- science Publishers, Inc., New York, 1949. 24. Young, A. P., and Schwartz, C. M., "A Replica Method for Examining Wear and Scuffing in Cylinder Liners," to be pub- lished. 25. Weatherhead, A. Petrographic microtech- nique (thin sectioning), A. Barron, London, 1947. C. W. IVIelton and C. M. Schwartz MICROSCOPISTS AND RESEARCH MANAGEMENT There is no question but that inckistry needs microscopy. It is a tool of science that enables us to get closer to things and ob- tain a better understanding of the how and why of nature. It must be remembered, how- ever, that it is only a tool and that a scientist is needed so that the human brain can inter- pret what is seen by the eye. One ciuestion to be discussed is whether the use of mi- croscopy is best accomplished by a trained specialist called a microscopist or by a scientist with microscopical training. An- other ciuestion relates to the position of the specialist in industrial research. The specialist learns a great deal al)out a limited area of science and in our present world certainly some specialization is neces- sary to solve the complicated problems of science. But on the other hand, a knowledge within several disciplines often is the key to the solution of problems. In some cases there is a point of no return in knowing more and more about less and less. It is not important whether the microscope is used by a scientist called a microscopist or by a physical, organic, bio, inorganic, etc., scientist who is well informed on its use po- tential. The important factor is that the scientist have the knowledge to solve the problem; knowledge in the fundamentals of optical science so that he can use the micro- scope and all its accessories to their best advantage; knowledge from experiences which can be correlated with experimental observations; knowledge to relate what he sees to the theory of what could be the mechanism of a reaction or the phj'sical structure of a material; knowledge which makes the scientist a specialist. In our industrial research organizations, advancement has been up an administrative ladder. Each step requires supervision of more people. In research organizations a general scientific knowledge of all the work being supervised is expected of the leader. This means that administrative personnel cannot specialize but must broaden their knowledge. The eventual desire for adminis- trative positions is probably one reason why it has been difficult to interest college stu- dents in a profession as specialized as mi- croscopy. On the face of it, it would appear that it does not have a great financial future. A number of research organizations, recog- nizing that the future of the specialist was being lost in such an adniini.strative hier- archy, ha\'e developed professional as well as administrative ladders. This makes it 381 GENERAL MICROSCOPY possible to develop or grow in a professional way without going into administrative work. C^^ananiid has such a program so that we are able to recognize the ability of the specialist and reward him accordingly. In designing such a program, the question arose: What kind of program can we design that will retain, develop and reward a crea- tive scientist in his field of competence and not force him into manager ranks as the only means of obtaining these rewards? If the administrative progression is the only way for advancement, then we may lose a good creative scientific specialist and obtain a poor manager. Many scientists can and should become managers, but those who have neither the desire nor the ability should not be forced to become managers in order to advance in their profession. One way to accomplish this is to change the conditions which force many scientists to become administrators in order to get more money and recognition. We know from attitude surveys that more money alone is not the answer to the problem. The profes- sional scientist is interested in the nature, scope and structure of his assignment and the relationship with his scientific colleagues, both within and outside his own company. He wants and has the professional right to an opportunity to contribute to scientific journals, participate in technical conferences and to expand his scientific horizons. He wants to know how his supervisors view his present performance and competence and what his future possibilities are in the or- ganization. In setting up a new program for profes- sional growth at Cyanamid, we began by establishing certain minimum requirements or standards for competence. Within these broad standards were various levels of per- formance, knowledge, experience and poten- tial through which the scientist would grow at various stages of his career. These levels of competence are in turn related to salary. It is possible for a man to move up from level to level without changing his job. He is not limited or hemmed in by a series of compartmentalized job descriptions. These new levels of professional accomplishment measured against the existing levels of managerial accomplishment form the basis for Cyanamid 's Professional Development Program. Here the man is evaluated, not the job. Growth possibilities depend upon the ability of the man, not the confines of a job. Figure 1 shows the progression of develop- ment in Cyanamid 's Central Research Divi- sion. The B.S. graduate enters at level 1 as a Scientist (i.e., Chemist, Physicist, Biologist, Chemical Engineer, etc.). After a minimum of one year's experience, it is possible for the outstanding man to reach level 2. This level has qualifications similar to that for the M.S. degree, and the M.S. from college enters at this level. After a minimum of five years' ex- perience from graduation the outstanding man can reach level 3, and at such time he is advanced to Research Scientist. The Ph.D. enters into the organization near the top of level 3. The outstanding Ph.D. moves into level 4 after one year of experience. A minimum of five years after re- ceiving his Ph.D. he may move into level 5 and advance to Senior Research Scientist. The outstanding B.S. or M.S. in science or engineering can progress through the same levels. For those who wish to develop further professionally there are the higher levels of Research Associate, Research Fellow and Senior Research Fellow. On the management side the Group Leader is the first position, and these are chosen from levels 4 and 5, depending upon the responsi- bilities of the position. The advancing posi- tions in this program are Manager of a Sec- tion and Director of a Department. It is possible to move from the professional side to the management side and vice versa. It is no longer necessary for research scientists to become managers to enjoy status or income. A word here about the concept of expe- 382 Level 8 Level 7 Level 5 & 6 Level 3 & 4 fl 1 & 2 Management or Administrative MICROSCOPISTS AND KESEAKCII MANAGEMENT Professi'inal or Scientific Direct')]- of Department ^- —> Research Fellow Senior Manager of Section Oroup Leader ^- > Research Fellow Senior Research Scientist B & A Research Associate -<- <- Research Scientist B & A Scientist B & A ^ Fig. 1. American Cyanamid Company's Research Division Development Program. rience in professional work. One year older does not necessarily mean that one year of experience or development has been gained. It should be clear that five years in a repeti- tive job might mean one year of similar ex- perience repeated five times rather than five years of new and broadening experience. The scientist who grows professionally may advance through the professional devel- opment program until he reaches the level equivalent to his maximum knowledge, ex- peiience and potential. He will have to con- tent himself with this growth level unless he is able to get more knowledge or experience. Sometimes the work done b}^ a man may not utiUze his full abilities. If he has the poten- tial, it may be advisable for him to study more or transfer to a more challenging as- signment. It is research management's responsibility to utilize all personnel at their highest po- tential and make sm^e that the work being done stimulates rather than impedes, a man's creativity and progress. Proper place- ment is our goal and, although it is not easy to achieve, we recognize this and work toward it at all times. The levels which have been discussed are related to compensation. Realistically speak- ing, compensation — whether at the profes- sional or hourly level — is based on certain general characteristics. No management has yet been able to come up with a method of administration which can function in a vac- 383 GENERAL >IICROSCOPY uum, iiidcpcMident of what is generally paid by other firms. Accordingly, as in most management prac- tice, the top level of each research salary range is determined by surveys in coopera- tion with similar research organizations. Once the broad salary ranges have been es- tablished, it is possible to fit in the various levels of competence required in research work. Here again despite numerous experi- ments, psychologists and testing experts have never been able to come up with a screening method which can evaluate the ability of professional personnel independent of the two major factors of education and experience. For this reason, the compensa- tion for new research personnel coming di- rectly from the university is based on their level of education. In attempting to fulfill its responsibility of utilizing personnel at their highest poten- tial, our research management has developed a combined evaluation and counselling pro- gram. This involves the use of a Counselling Guide and a Performance Review form, in conjunction with two employee interviews. The objective of this method of review of an employee's performance is twofold: 1. For Management: To provide the im- mediate supervisor and higher levels of man- agement with a means of evaluating the performance and potential of each employee and of summarizing this information in re- ports. 2. For the Employee: To enable him to review his own performance objectively, both alone and with his supervisors; to learn how he is doing and where he is going, and to insure that he has the fullest oppor- tunity for maximum improvement and de- velopment. The Counselling program involves five general steps to complete the review of an employee's performance through forms and interviews. They are: 1. Distribution and explanation of forms to employee. 2. Com{)lcli(jn of the forms, insofar as possible, by both the employee and his supervisor, independently, prior to the counselling interview. 3. The coun.selling and performance re- view interview between the employee and his supervisor. 4. Discussion of the completed forms by the supervisor and the next level of super- vision. 5. An interview between the employee and his second-level supervisor to discuss results of the evaluation and the interview with the employee's immediate supervisor. Why do we suggest distribution of these forms to the employees as a first step? Pri- marily, because we know that people tend to dislike and distrust what they do not understand. By giving them an opportunity to look over the forms at their leisure — and by suggesting that they make an attempt at a self-analysis, we believe some normal psychological resistance may be removed. The Counselling Interview is probably the most difficult step in the over-all Counselling Program because the supervisor and em- ployee meet face to face to discuss personal opinions. The supervisor should conduct the interview skillfully and with great tact so that criticism will remain factual and con- structive rather than subjective and emo- tional. The two persons should meet in an atmosphere of mutual respect and with the same objective of further developing pro- fessional or administrative ability. In conclusion, I want to make it clear that the specialist or professional we are talking about is a scientist who is advancing his knowledge and not just going along on his own experience. He does not know all the answers and will try new approaches and experiments to try to solve even old prob- lems. The specialist who knows all the an- swers is not growing in his profession, he is a technician getting more proficient doing things over and over again with exacting skill. George L. Royer 384 MICKOTOMY MICROTOMY Thanks to the contributions of the Ger- man physicist Abbe the efficiency of the hght microscope was rapidfy improved in the latter half of the 19th century. This development in turn implied that completely different demands had to be made on prepa- ration of specimens. In order to make use of the increased resolution power of the mi- croscope, it was necessary to employ thin sections in the study of biological material. In this way the development of microtomy was started. This development was both a conseqvience and a necessary precondition of modern light microscopy. The limit for the most important property of the microscope, the resolving power, is determined by the wave length of the light employed. As far as the light microscope is concerned, we can therefore only expect technical improve- ments in the future. Regarding the technique for making microscopical preparations, it seems however justified to hope for consider- able progress. Today the routine work in human pathol- ogy constitutes the overwhelming part of light -microscopical work on biological mate- rial. The cutting technique for this purpose is well developed, and generally known, and further information is therefore not needed. Contrary to this, severe problems are en- countered, in quantitative morphological and c>4ochemical work. Like the other steps in the preparation of sections, cutting dis- torts the native tissue. If this is not taken into consideration, errors may arise in quan- titative work. The aim of the present article is to illustrate these problems. The Microtome. When classical his- tology was developed, the demands with respect to precision which were made on a microtome corresponded to the best which could be produced. Today the situation is different. The progress of modern precision industry makes the production of good mi- crotomes an easy task. Only in the produc- tion of ultramicrotomes for cutting of mate- rial for electron microscopy are modern resources fully exploited. In principle a microtome produces two different move- ments between the knife and the prepara- tion: (1) the cutting movement occurring in the cutting plane, and (2) the feeding move- ment perpendicular to this plane. In a mi- crotome of good quality both these move- ents must be as precise as possible. The true feeding should not deviate more than 0.1 M from the microtome setting, when the latter is in the range 1-20 m- Measurements on various slide and rotation microtomes show that this demand is met by most microtomes from well known firms. It is thus not the microtome which causes the problems in the cutting for quantitative analysis. The Knife. Besides a good microtome it is also necessary to have a knife with optimal qualities. Ever since the cutting of biologi- cal material began, much work has been devoted to getting as good knives as pos- sible. A large selection of suitable kinds of steel is a^^ailable, varying from pure carbon steel to various types of multialloyed and specially hardened stainless steel. Thi-ee qualities of the steel are of special impor- tance (1) the hardness, (2) the toughness and (3) the corrosion resistance. Unfor- tunately these parameters do not vary in a parallel manner. A very hard steel is brittle and increased addition of stainless metals reduces the hardness. Knife steel therefore represents a compromise which, however, thanks to modern metallurgy meets exacting demands. The edge of the knife is the section line between the two facet surfaces. In order to understand which qualities the edge must possess it is necessary to illustrate schemati- cally how cutting is performed (Fig. 1). During cutting, a distortion of the tissue occurs, causing the thickness of the section ti to be larger than the feeding of the block /i . For the cutting three angles are of im- portance, (1) the rake angle (r), (2) the bevel angle (b), and (3) the clearance angle 385 GENERAL MICROSCOPY \ Fig. 1. Schematic illustration of sectioning at an angle of 90° between the cutting direction and the edge line, c = clearance angle, h = bevel angle, r = rake angle, h = the feed of the microtome, <2 = section thickness. (c). The sum of these three angles is 90°. The larger the rake angle employed, the slighter the distortion of the section, i.e., the more the ratio hi a approach 1. It is therefore de- sirable to have the clearance and the bevel angles as small as possible. However, the clearance angle cannot be smaller than 6°, if the risk of compression of the tissue block is to be avoided. The possibilities of decreas- ing the bevel angle are also limited, because of the difficulties involved in producing a sufficiently even edge line. If the facet sur- faces meet in a small angle, even minimal abrasions may cause rather serious irregular- ities in the edge line. The solution must therefore be a compromise. To produce a bevel angle of 20°, e.g., which meets precise demands on the evenness of the edge line, a disproportionately large amount of work is required, while it is easy to make a good edge if the bevel angle is allowed to exceed 50°. The sharpening of a microtome knife in- volves imparting to the two facet surfaces such a planarity and finish that their line of sectioning is exactly even. Fig. 2 shows an interference microscopical picture of facet surfaces produced by different methods. The shape of the interference fringes reveal the irregularities of the surface. The interfringe distance represents a profile depth of 0.27 ^l. The rate of removal of the various methods is, of course, inxcrsely proportional to the resulting surface finish. The treatment of the facet surface, therefore, has to be made stepwise; the last step generally consists of lapping against glass in an immersion me- dium. To prevent this step from becoming too time-consuming, it is necessary that the facet surfaces be pretreated with coarser methods. Unfortunately, stropping is still widely used as the last step in the sharpen- ing of microtome knives. As indicated in Fig. 2b, the facet surface exhibits a rather good finish after stropping and the edge is therefore sharp. However, the decreasing distance between the interference fringes as the edge line is approached shows that the facet surface is curved, so that the bevel angle is considerably larger than the original one, and is not well defined. The other angles also are unknown, and accurate cutting be- comes undependable. Stropped knives there- fore are to be avoided. The control of the knife edge is a fre- quently neglected detail in microtomy. It is still customary to judge the sharpness by hair cutting or by passing the thumb over the edge. These procedures are unworthy of the instrumental equipment which we pos- sess today. A microtome knife used for quan- titative biological work ought to be Avell defined with respect to the bevel angle and to the width of the facet surfaces, and the evenness of the edge line should be con- trolled in incident dark-field microscopy at 700 times magnification. Measurement of Section Thickness. In all investigations intending to supply quantitative morphological and cytochemi- cal information, in which sectioned material is used, it is necessary to know the thickness of individual sections. This determination is often difficult. The lack of hardness in the sectioned tissue means that not even a mini- mal measuring pressure may be applied to the surface of the section without causing plastic deformation. Secondly, to control what is being measured, the method em- 386 MICKOTOMY ■V^ ; ■ '--^^^V ii i ' .' • s, H (a) (b) (d) (c) Fig. 2. Interferograms of different facet surfaces, (a) honed; (b) stropped; (c) lapped against cast- iron; (d) lapped against glass. (Hg-light, 300X). The interference fringes reveal irregularities in the surface. The distance between two fringes represents a profile depth of 0.27 fi. ployed must possess the horizontal resolvmg power of the light microscope, as well as the vertical resolving power required for thickness determination. In Fig. 3 is shown an apparatus for thickness determination which meets the.se demands. It consists of a commercial microscope for incident light. To increase the focusing accuracy, a thin metal wire has been moimted in the plane of the field iris. By means of a special ar- rangement with off-central incidence of the light, this wire produces a shadow in the object plane which moves back and forth across the field of view during focusing and defocusing. The line of symmetry for this movement is indicated with a hair cross in the eyepiece, and it is therefore possible to reproduce the focusing with great accuracy. The vertical movement of the object i\-e dur- ing focusing is registered by means of a mechanical measuring device. In this way it is possible to determine the thickness of a section by focusing first on the upper surface of the slide, and then on the upper surface of the section. By means of a special con- struction of the microscope stage, the slide may be moved in the horizontal plane with very small vertical deviations. If plane parallel slides are used, it is possible to per- form topographical analysis of sections. Un- der ideal conditions the instrument permits measurements with an accuracy of 0.1 n. Thickness Varialions in Sections. For anybody planning to use sectioned material 387 GENERAL MKIJOSCOrV Fig. 3. Apparatus for thickness measurements of microtome sections. for quantitative work the following questions arise. What are the optimal conditions for sectioning? Is there any difference between the thickness of the section and the feeding of the microtome? How large is the thickness variation within and between sections? It would be an easy task to select from the literature several papers, the value of which may be questioned, because the authors have neglected these questions. During an investigation comprising more than 15,000 thickness determinations of sections, the problem of cutting distortion has been analyzed from various points of view. On the basis of this investigation it is possible to answer the questions stated above. The type of microtome employed is un- important as long as it is of good equality. It is likewise of minor consequence whether it is driven by hand or by a motor, whether or not it is run continuously, and whether the rate of cutting is varied within rather wide limits. The type of tissue is also of little importance, except of course in the case of mineralized tissue. Among the embedding media employed the distortion is less with plastics than with paraffin. Indirect meas- urements indicate that the greatest risk of distortion exists with freeze sectioning. The properties of the knife are of decisive importance for the qualities of the sections. If the bevel angle is above 50°, if facet sur- faces more than 10 n wide are used, or if the edge line in 700 X magnification is uneven, a rapid increase is observed in the ratio io/n as well as in the thickness variation within and between sections. If a stropped knife is employed, the distortion equals that ob- tained with a knife having a 70° bevel an- gle. Even when the cutting conditions are optimal with respect to the properties of the knife, to the type of microtome, and to the preparation of tissue^ — a rare event- — con- siderable distortion may be anticipated. First of all, an increase of the thickness occiu's, which seldom is less than 10 per cent. Secondly, thickness variations arise within and between sections, and these are so large that they cannot be neglected in quantita- tive work. The coefficients for the variation within and between sections in a series of 5 fx paraffin sections are not less than 6 and 30 per cent, respectively. For frozen sections the corresponding value for variation be- tween sections is 45 per cent. Duplication of Sections. A complete correction for thickness variations of sections would require a topographical survey of each individual section. This is impossible, be- cause the necessary equipment is lacking in most laboratories; moreover, such a proce- dure would be extremely time-consuming, and the cytochemical and morphological analyses of the sections seldom permit them to be movmted on plane parallel slides for thickness determination. Since the variation between sections entirely dominates the 388 MICHOroMY Fig. 4. Schematic representation of tissue and plastic bar molded into same block. Fig. 5. Schematic representation of tissue in plastic hollow cj^linder chucked for freeze-.section- ing. thickness variation of sections, it is highl}^ important to correct for this factor. It is possible to do this by means of plastic dupli- cates. A plastic bar (Fig. 4) made of but 3d and methylmethacrylate (15 and 85 per cent, respectively) is used for embedded material. This is formed and embedded in intimate contact with the tissue to be sectioned. For frozen material a hollow cylinder is used, made of butylmethacrylate, which is filled out with the tissue (Fig. 5). In this way one obtains simultaneously a tissue section and a plastic duplicate which may be separated and measured independently, e.g., with in- terference microscopy. The error committed by allowing the setting of the microtome to indicate the thickness of the section is in this way reduced to about one half for em- bedded material and to one third for freeze- sectioned material. REFERENCES Abbe, E., Sitz.-ber. Jen. Gesell. Med. Naturw., 71, 207 (1880). Abbe, E., J. Roy. Micr. Soc, Ser. II, 1, 680 (1881). Apathy, S. von, "Neuere Beitriige zur Schneide- technik", Z. wiss. Mikr., 29, 479 (1912). Ardenne, M. von, Z. wiss. Mikr., 56, 8 (1939). Bailey, A. J., "Precision Sectioning of Wood", Stain TechnoL, 12, 159 (1937). Berek, M., Sitz.-ber. Gesell. Beford. ges. Nalur- wiss., (Marburg) 62, 189 (1927). Bishop, F. W., "New Automatic Sharpener for Microtome Blades", Rev. Sci. Insir., 25, 1190 (1954). Brattgard, S.-O., "Microscopical Determina- tions of the Thickness of Histological Sec- tions", J. Roy. Micr. Soc, 74, 113 (1954). Dempster, W. T., "Distortions Due to the Slid- ing Microtome", Anat. Rec. 84, 269 (1942). Dempster, W. T., "Properties of the Paraffin Re- lation to Microtechnique", Michigan Acad. Sci., 29, 251 (1943). Dempster, W. T., "The Mechanics of Paraffin Sec- tioning by the Microtome". Anat. Rec, 84, 241 (1942)'. Dempster, W. T., "Paraffin Compression Due to the Rotary Microtome", Stain TechnoL, 18, 13 (1943). Ekholm, R., Hallen, O. and Zelander, T., "Sharpening of Knives for Ultramicrotomy", Experientia, 11, 361 (1955). Fanz, J. I., "An Automatic Microtome Knife Sharpener and Methods for Grinding and Honing the Knife Satisfactorily", /. Lab. Clin. Med., 14, 1194 (1929). Hallen, O., "A New Method for Sharpening Microtome Knives", Lab. Invest., 5 (1956). Hallen, O., "On the Cutting and Thickness De- termination of Microtome Sections", Acta Anatomica, Suppl. 25, 26 (1956). Hallen, O., "Quantitative Analysis of Sectioned Biological Material". (In press). Heard, O. O., "The Influence of Surface Forces in Microtomy", Anat. Rec, 117, 725 (1953). HiLLiER, J., "On the Sharpening of Microtome Knives for Ultrathin Sectioning", Rev. Sci. Instr., 22, 185 (1951). Johnston, C, "Some Aspects of Microtome Knife Sharpening", /. Med. Lab. TechnoL, 10, 131 (1952). Kisser, J., Z. tviss. Mikr., 43, 361 (1926). Kisser, J., Z. iviss. Mikr., 44, 452 (1927). Lange, P. W. and Engstrom, A., "Determina- 389 GENERAL ]\Iir.KOSCOPY tion of Thickness of Microscopic Objects", Lab. Invest., 3, 116 (1954). Lendvai, J., Z. wiss. Mikr., 26, 203 (1909). Low, W., Z. wiss. Mikr., 48, 417 (1931). Malone, E. F., "Sharpening Microtome Knives", Anat. Rec, 24, 97 (1922). MoHL, H. VON, Bot. Z., 15, 249 (1857). Nageotte, J., Bull. cVHist. Awl-, 3, 258 (1926). Richards, O. W., "The Effective Use and Proper Care of the Microtome", Spencer Lens Co., Buffalo, N. Y., 1942. ScHMERWiTZ, G., Die Ihnschau, 36, 827 (1932). SsoBOLEW, L. W., Z. wiss. Mikr., 26, 65 (1909). Uber, F. M., "Microtome Knife Sharpeners Op- erating on the Abrasive-Gronnd Glass Prin- ciple", Stain TechnoL, 11, 93 (1936). Olle Hallen PLASTICS In comparison with metals and fibers, mi- croscopic investigation of plastic materials is still in its infancy. It is true that more has been done in this field in recent years, but there is still no question of extensive re- search. Microscopic methods are applied to plastics for: (a) examining fillers with a view to homogeneous dispersion; (b) examination of the mixing of various plastics; (c) studying fracturing of plastic materials; (d) stress in- vestigation ; (e) penetration and distribution of plasticizers. Investigation can largely be carried out with intact materials, but study of fracture and stress should be effected with sections or films. All microscopic methods are suitable for examining these materials. The normal Hght microscope is used for observing dis- persion of fillers. Sections or films are used for this purpose. For very finely ground fillers and for investigating the occurrence of cracks under the influence of filler par- ticles, a phase contrast microscope is used. Differences in refractive index, occurring particularly in mixing various plastics, can be successfully investigated with a phase contrast microscope for the purpose of quali- tative examination. Quantitative examina- tion of these mixtui'es can best bt; clfected with interference methods. For plasticizer penetration the interfer- ence methods are excellent, and it is often advisable to make use of so-called cine pho- tographs. The polarizing microscope is mainly use- ful for investigating stresses in plastic prod- ucts and the occurrence of spherulites and oriented structures. This method is especially important for polyethylene and polysty- rene investigation. The thickness of most plastic objects makes it necessary to use compensators which also compensate large phase differences. Ehringhaus' rotary com- pensator, of the Calcspar type, can be rec- ommended. In examining a mixture of two plastics it maj^ be advantageous to mcorporate a flu- orescent dye, so that dispersion can be exam- ined with a fluorescence microscope. In plastics research, surface techniques have a fairly wide range of application. The normal reflection microscope, both bright and dark field, is used for appraising surface irregularities and for examining surfaces of breaks. In the latter case the purpose may be to examine the breakage phenomena as such, and in the case of highly filled plastics, dispersion of the filler can also be easily ap- praised. The interference methods, both double and multiple beam, may be important aids to surface and fracture examination. Electron microscope techniques are still comparatively uncommon. Their range of applications is in examining fine structures. Fields where this method has already been applied are the investigation of spherulite formation in polyethylene and nylon films and the examination of break surfaces by means of replicas. The grain structure of suspension poly- mers can successfully be investigated with the electron microscope using ultra thin sec- tions. 390 PLASTICS Sectioning of Plastic Materials Hard plastics, such as polystyrene and methacrylates can most suitably be cut with- out further preparation, by means of a stable sledge microtome. The specimen holder must be as vibration-free as possible; the spiH'imen may protrude only slightly outside the holder and the knife must be secured in a heavy knife block. Moreover, only that part of the knife which is close to the point where it is secured should be used. Sometimes vibration cannot be entirely avoided. Especially in phase contrast or interference microscopic examination, care should be taken not to appraise specimens too quickly. As a rule, vibration becomes troublesome only when a number of succes- sive specimens is to be used for spatial reconstruction. With such objects the blade of the knife soon gets damaged and grooves occur in the cut surface. Every slice must therefore be judged immediately with the aid of a dissecting microscope, so that a different part of the knife can be used at once if any damage is found to have been started by the knife. The knife must be of higher quahty than is required for cutting biological material. The blade must be very accurately gromid. All slices tend to curl up. As soon as the sec- tions are obtained they should be tem- porarily unrolled with a couple of small artists' brushes. They can be fully extended by placing them on a drop of xylene. This is usually sufficient to obtain a completely flat specimen. The specimen can be mounted in Aqua- mount or Canada balsam. Elastic plastic materials, such as poly- ethylene, are cut with a freezing microtome; here again the sledge type is used. The re- quirements for cutting hard materials apply also to elastic materials. Vibration resulting from cutting is noticeably less than with hard material. Ai'tifacts due to cutting with a damaged knife blade are found, however. Regular inspection of the slices immediately after cutting reduces the adverse effects of this. Here again the sections can be stretched on xylene after the slices have first dried. This is a successful method especially for polyethylene and polypropylene. The production of thin sections for inves- tigating filler dispersion is preferable to the melting method (see below). Cutting with a microtome does not alter the distribution of the fillers. Large aggregates do not disinte- grate, nor are new ones formed. It is also possible to examine a spatial arrangement by means of a series of sections. Moreover, the influence of spherulite structures on filler dispersion can be appraised only from sec- tions. Examination of Films Obtained by Melting Small pieces of plastic are heated until they have just melted. They are then pressed between two microscope slides until they are the desired thickness. These films, after mounting in Aquamount or Canada balsam, can be used for microscopic examination. This method is fairly widely used for exam- ining filler dispersion in plastics. Its draw- backs are: (1) Spatial distribution of the fillers can no longer be determined. There are fairly strong lines of flow in the specimens and no opinion can thus be formed regarding the uniformity of dispersion. (2) Fillers often form a very labile system with the plastic. The big risk in melting is that the dispersion found in the films may not be the same as it was originally. Large aggregates may disintegrate and smaller granules may cake together. (3) Orientation effects are no longer found. The filler granules are often oriented in the plastic. This orientation is lost. (4) The influence of spherulite structures on filler dispersion can no longer be ascer- tained. The spherulites are either destroyed or in any case deformed. These structures 391 GENERAL MICROSCOPY have a considerable ei'tect on the unilorniity of filler dispersion. Hence it is very impor- tant to keep them intact. (5) If a mixture of two different synthetic components is to be examined, correct ap- praisal of the dispersion is very important. Disturbance of the original distribution is undesirable. Appraisal of Plastic Films with the Polarizing Microscope Most plastic films have a molecular orien- tation, even though they are not deliber- ately elongated. Even rolling causes a certain amount of double refraction. The intersec- tion of molecular strands gives rise to a film texture. Elongation in the main direction of the strands soon causes a parallel structure, but elongation perpendicular to this direc- tion widens the angle of intersection of the strands. In many films this film structure is demonstrable with a phase contrast micro- scope. In other cases, in which the refractive index difference between strands and sur- rounding material is slight, the phenomenon will hardly be visible with a phase contrast microscope. In these cases it can often be rendered visible by staining with I2-KI solu- tion or iodine tincture. The film is then ex- amined in its elongated state under a polar- izing microscope, the axis of elongation of the specimen being rotated from the diago- nal position to the orthogonal position. When the axis of the specimen forms an acute an- gle with the orthogonal position, the molecu- lar strands of the one axis fade out while those of the other axis do not. In this way the microscope shows a "herringbone struc- ture." This structure, which is regular in most films, also occurs in nearly all synthetic fibers, although the round or irregular sec- tions of these fibers renders examination difficult. Exact examination can therefore only be made with longitudinal sections. With synthetic fibers I2 • KI staining is also to be recommended. The texture occurring with the fibers may be: (1) Regular single heri'ingbone structure. This corresponds completely to the film tex- ture and can be further divided into coarsely fibrillary and finely fibrillar. (2) Irregular single herringbone structure. No clear arrangement of molecular strands. Both the angle of intersection, strand thick- ness and strand distance vary from place to place. (3) Regular film structure. The angle of intersection, distance of molecular fibrils and strand thickness show a regular variation from outside to inside, but from place to place the herringbone structure may still be regular. (4) Irregular film texture. The general character of the fibrillar structure is differ- ent on the outside of the fiber and in the middle. But the structure is everywhere irregular. (5) Mixed film texture. The outer layer of the fiber shows a regular herringbone structure, while the pith has an irregular texture, or vice versa. (6) Branched film texture. When the fiber is examined with its axis roughly correspond- ing to the orthogonal position, a fine second- ary film texture may sometimes be visible in the light bands, having a slightly different orientation from the main direction. The secondary fibrils are noticeably thinner than the principal fibrils. The above system applies in the first place to synthetic fibers, but is also found with films — though to a less extent. Investigation of Spherulite Structures For examining spherulite structures it is usually advisa))le to use thin films. Many high polymers form spherulite structures when they crystallize from the melt. The number of spherulites formed de- pends upon the number of crystallization nuclei present and the speed of cooling. They may be very suitably examined with phase contrast and polarizing microscopes. The 392 PLASTICS finest structures, however, become visible Ohsinicted Spherulitcs. The structure only with an electron microscope. is still rej2;ular, t»iil (lie purel}^ spherical or We can distinguish these crystal structures ellipsoid form is no loiijicr attained. This is microscopically according to (1) their type usually because the individual spherulites and (2) their shape. The most common types have impeded one another's growth, are described below. Misfornied Spln'i-iililes. If there are Single Spherulites. The polarizing mi- very many nuclei in the melt, the spherulites croscope show^s these as round to ellipsoid may obstruct one another to such an extent particles with a distinct polarization cross, during growth that no regular structures the arms of which are not uniformly thick, can arise. In this case it is often very difficult If the spherulite is rotated this cross often to recognize the above-mentioned types, splits into two hyperbolas. If these crystal structures are elhpsoidal the bigger axis is Melting Point Tests with High Poh- in the direction of elongation. mers Dendrite Spherulites. Spherulites with a To determine the melting points of various pronounced radial main structm-e. The ra- plastics the ordinary melting point micro- dial fibrils have pronounced featherwise scope can be used. It should be borne in branches. This branching is also clearly vis- mind that pieces of polymer are usually very ible w^th an electron microscope. poor conductors of heat. Only very small "Peacock"-Spherulites. Spherulites are granules, 10-15 m, are therefore used as sometimes found that show an alternation specimens. If bigger pieces are used it is of concentric bands. These bands have alter- often found that the granules are already nate positive and negative birefringence, melted on the outside, while the inside does Electron microscopic examination suggests not melt until a higher temperature is that this is a mixed structure of the single reached, as shown by the thermometer in and dendrite types. The negative bands are the microscope heating plate, said to be formed by the branched fibrils Even with very small granules the tem- and the positive ones by the unbranched perature must be allowed to rise only very pieces with their purely radial orientation. slowly. It is advisable not to increase it by A pattern resembling the preceding one is more than 2°C a minute, fomid with polyethylene, in which spheru- ^ . „ . . lites occur with alternate negatively hire- Surface Examination fringent rings and isotropic bands. This Examination of the surface of fractures gives a "peacock" effect under the polarizing causes little difficulty on the whole. With microscope. transparent plastics it may happen that Irregular Spherulites. These are under- cracks continuing to several microns below- stood to be structures without any of the the surface cause interference colors, above-mentioned regularly geometrical fibril If it is desired to examine filler disp(Msion orientations. Sometimes zig-zag structures from fracture surfaces, it is often adxisable are found, while in other cases there is no to use polarizetl light. For many plastics a regularity at all (e.g., Terylene). specific angle of incidence can be determined A classification of spherulites according to at which polarized light depolarizes when their shape includes the following: reflected. As (lci)()larization does not occur Regular Spherulites. Hound to ellip- with the liller particles or occurs at a differ- soid in shape these have originated independ- ent angle of incidence, filler particle disper- ently and have not been obstructed in their sion can be effectively studied with an ana- o-rowth. lyzing filter. It should be remembered that 393 GE.XKK AL MICROSCOPY surface unevennesses cause pronounced shadows which make appraisal difficult. Hence this method is applied only if filler dispersion cannot be appraised from sections or if it is desired to know the direct cause of the fracture. Examining Plastic in Non-Plastic Ma- terials For investigating the presence of polymers in textile or leather materials, fluorescence microscopic methods are used. If the material containing the plastic fluoresces only slightly or not at all, micro- tome sections can be examined direct by fluorescence microscopy. Most plastics flu- oresce bright blue-white or yellow-green. This method is very exact on the whole. Fluorescence microscopic examination of plastics in natural textile fibers is not usually directly possible because most natural fibers have a fairly strong fluorescence of their own. As these fibers generally have fairly strong hydrophilic properties, however, it is very simple to use fluorochromes. Staining with a mixture of coriphosphine 0.1% and Rhoda- mine 6 GD 0.1% for 15 minutes, thorough washing in distilled water and drying, fol- lowed by fluorescence microscopic examina- tion, enables the plastic coating of the fibers to be rendered clearly visible in nearly all cases. J. ISINGS PULP AND PAPER Among the methods for investigating pulp and paper, microscopic examination occupies an important position. Microscopic techniciues are used for: (a) identifying paper fibers; (b) investigating coatings on paper; (c) investigating sheet formation by the fi- bers; (d) investigating fillers; (e) morphologi- cal examination of fibers. In the practice of paper technology, iden- tification of the various paper fibers is of primary importance. As a rule it is sufficient to use a normal light microscope. With a total magnification of (iO and 200 X stronger magnification is usually superfluous. Coat- ings can also be most suitably examined from their cross sections with a normal light microscope. These cross sections are usually made with a Hardy microtome or a freezing microtome. The pulp condition of the various fibers can most suitably be studied by examining the fibers in acjueous suspen- sions with a phase contrast microscope. This method has been very widely applied, espe- cially in recent years. Birefringence of fibers with a polarizing microscope is not deter- mined; this method is usually limited to in- vestigating fillers. The same can be said of the x-ray micro- scope. Although it is possible with this method to see the mutual coherence of the paper fibers it is especially important for examining filler distribution. Good results can be achieved, particularly with the x-ray projection microscope (q.v.) as there is no need to make cross sections for this. The fact that stereo photographs can easily be made with it makes this method still more attractive. Fluorescence microscopy is still infre- quently applied for paper fiber investigation. An indication of the usefulness of this method is to be found in Herzog, who noted a fairly pronounced difference in fluorescence color between soft wood sulfite and soft wood sulfate pulp. More important than the flu- orescence of the pulp itself however, is the method whereby the coloring of fibers with fluorochromes is used. Definitive data re- garding this are not available in the litera- ture, although it is clear that straw and esparto pulp can easily be distinguished by this method. Lastly, electron microscopy must be men- tioned as an important method of paper investigation. This method is particularly important for exact investigation of the fibrils of pulped fibers. These can be exam- ined with replica techniques under an ordi- 394 PULP AND PAPER nary electron microscope or with thin sec- tions, the contact zone between the pulped fibers being chiefly examined. The reflection electron microscope has also been used with great success. It seems that still more results can be achieved in this direction with elec- tron scanning techniques. Paper Fiber Research Preparation Technitjues. The tech- niques used for preparing specimens for pa- per investigation are usually very simple. First, fiber suspensions are investigated which are obtained by shaking pulp or paper in water, possibly with the application of heat, after which it readily disintegrates. For x-ray microscopic examination the paper can remain intact, but for contact micro- radiography sections about 50 microns thick are used. The cross sections for examining coatings are made with a Hardy microtome (see Fibers (Textile) -Techniques). The speci- mens are usually embedded in a mixture of equal parts of water and glycerin. For the electron microscope normal replica tech- niques are used, while evaporated suspen- sions are suitable, after shadowing, for examination with a reflection microscope. Full details of this can be found in Emerton. Staining Techniques. The main stains for paper fiber iin^estigation are: (1) Zinc chloroiodide solution in water (see Fibers (Textile) Techniques, p. 344). (2) Solution of iodine in potassium iodide (I2-KI). Both normal staining and I2-KI staining followed by treatment in glycerin- sulfuric acid give good results in paper in- vestigation. The formulas are given in the article on Fibers (Textile) -Techniques (p. 346). (3) Staining according to Lofton and Merrit wnth fuchsine and malachite green. Composition of the stain: 2 parts bj' volume of aqueous 1% fuchsine solu- tion. 1 part by volume of aqueous J^% malachite green solution. The sohition will keep for abovit eight days. A paper fiber mash is freed from water as much as possible and stirred with the solution on a slide. After removal of the solution the specimen is treated for 10 to 30 seconds with 3 to 4 drops of 0.1% HCl and rinsed. Result. Unbleached sulfite wood pulp: purple red; unbleached soda and sul- fate wood pulp: blue with a fairly pronounced red haze. Ground wood: stains like unbleached soda and sulfate pulp. (4) Staining according to Graff (TAPPI standard method). Two stains are suitable for paper examination. The most widely used is Stain C: Solution I: 40 g aluminum chloride in 100 cc wa- ter S.G. 1.15 at 28°C. Solution II: 100 g CaCU in 150 ml water S.G. 1.36 at 28°C. Solution III: 50 g dr\^zinc chloride in 25 ml water S.G. 1.80at28°C." Solution IV: 0.40 g dry KI 0.65 g dry 12 in 50 ml water S.G. 1.80 at 28°C. The mixture for use is 20 ml of solution I, 10 ml of solution II, 10 ml of solution III and 12.5 ml of solution IV. A precipitate is allowed to form, the clear top liquid is pipetted. Add a crystal of I2 and store the mi.xture in the dark. The results for the various pulps are listed in Table 1. Paper Morphology The fibers used in paper manufacture are of three kinds: (1) Rags originating from textile material. (2) Wood, straw, esparto, etc., pulp. (3) Highly lignate fibers obtained by grinding wood. In all three cases there are unpulped and pulped fibers. Rag Fiber. Cotton. Cotton is nearlj' always very easy to identify as a fiber in jxiper. Even with thorough pulping, it is hardly ever possible to split all the fibers into the highly d(^formed single dimensional fibrils. The twisting of the ribbon-shaped fiber is a clear means of identification. Flax and Hemp. Both flax and hemp may be used as paper fiber. As they are usually found in only small quantities in rag paper and are, moreover, greatly deformed by pulping, it is very difficult to estimate their 395 GENERAL MICROSCOPY Table 1. Stains for Pulps (Graff C) Unbleached (Colors in order of Bleached lower lignin content) Rags claret Ground wood yellow bright yellow orange Soft wood pulp sulfite yellow — green yellow pink gray bright purple gray — pale red purple refined sulfite pale brown — blue gray light reddish orange — dull red sulfate weak green yellow, strong yellow brown — mild yellow green — dark yellow gray dark blue gray — dull purple Hard wood sulfite dull yellow green weak purple blue — bright purple gray refined red orange — dull red soda and sulfate light blue green — dark blue-green — dark red gray dull blue — dull purple Manila 3'ellow gray — light blue — purple gray Jute bright orange yellow bright yellow green Straw, esparto bright green gray, dark blue gray — purple gray Kotzu pale pink green Gampi bright green yellow — light blue green Besides the C stain, stain A as modified by Sutermeister can also be used. This stain is obtained by mixing 45 ml of solution II from stain C with 5 ml solution IV from stain C and storing this in a dark colored bottle, a crystal of I2 being added. Table 2 shows the colors that then occur for the various pulps. In appraising these, however, Graff's color chart is important. percentage alongside the other rag fibers. Only large, hardly damaged pieces of flax are clearly recognizable from the displace- ments. It is usually more difficult to identify hemp fibers in paper from their morphologi- cal characteristics. The fiber tends very greatly to fibrillation and little can be seen of the large lumen. For the characteristics of the fibers when unpulped or only slightly pulped see the section on the Morphology of Textile Fibers, p. 350. Hard Wood Pulp. The hard woods used for pulp are chiefly poplar, birch, eucalyptus and beech. Chestnut, maple, willow and ash are used to a less extent. Hard wood pulp consists of vessels, vascular tracheides, fiber tracheides, libriform fibers and parenchyma constituents. In the eucalyptus there are also secretion canals. In pulp it is often difficult to distinguish the various parts. Character- istics distinguishing them from other pulps are mainly the large vessels. As far as the paper industry is concerned it is usually of importance only to know whether a pulp consists of hard wood or soft wood. Further- more, in the case of a mixed pulp the per- centage of hard wood usually has to be de- termined. The type of hard wood is less important. Beech (Fagus sijlvatica). The large vessels are usually fairly long and vary in width from 18 to 80 n. The ends are cut off ob- liquely and are usually digitaliform at one end. The pits are easy to see. The thinner vessels are usually scalariform. The number of scalar rings is 6 to 20. Sometimes the rings are broken, but even with fairly thorough pulping they can still be identified. The types of pits are: (1) Simple (small, scattered in medulla) (2) Bordered (close together and small — the "border" is usually difficult to distin- guish). (3) Large medulla (closely defined and in rows). 396 PULP AND PAPER Libriform fibers are the main constituent of pulp. There are many wide, short pits on the cell wall. The tracheides are verj^ broad and have wide simple pits. The border of these is difficult to see. There are invariably many parenchyma elements in the pulp. These are elongated, thin-walled and very porous. Birch (Betula alba). The vessels are char- acterized by bordered pits very close to- gether. The border is not visible. The wood tracheides show simple pits, likewise very close together. The cell wall is only moder- ately lignate. Parenchyma cells are very thin and occur only sporadically. Poplar (Populus alha). Bordered pits with hexagonal simple pits without a border and large pits in crosswise rows. The vessels have large, pronounced end-protrusions. The pits of the fibers are not arranged in long rows. Some fibers have no pits. The paren- chyma elements are often elongated. (3) Soft Wood Pulp. In the soft woods one must distinguish between the fiber tra- cheides, the medulla tracheides and the parenchyma elements. The fiber tracheides are 4 or 5 sided thick-walled elements. On all walls. Semi-bordered pits are also found locally, i.e., where the tracheides border upon the parenchyma cells. Their walls are thicker in places and in cross and tangential sections these often look like tangential bars. The medulla parenchyma has ordinary pits on the walls adjoining other parench3^ma cells. The walls adjoining the tracheides have semi-bordered pits. In sulfite pulp the parenchyma cells also have pronounced resin particles. These par- ticles, which are easily identifiable, are im- portant in distinguishing between sulfite and sulfate pulp. The surface of the primarj^ wall of the Spring tracheides is greatly wrinkled during pulping of the wood and subsequent drying, while cracks also occur in it. Moreover, small fibers readily split off. This effect is particularly easy to see in sulfite pulp. In Autumn tracheides this crazing effect is far less pronounced. Sulfate pulp is usually obtained from firs, the bordered pits appearing as large, square holes. Besides the above differences between sulfite and sulfate pulps the following may apply: Sulfite pulp Sulfate pulp Unbleached Bleached Unbleached Bleached Zinc chloroiodide blue violet blue violet brown violet brown violet reticulated no reticulation no reticulation I2.KI green yellow no color weak green j'ellow no color Lofton and Merrit purple red no color red violet no color Graff C solution green yellow purple gray j^ellow brown — yellow green blue gray two walls they have large round bordered pits in rows. Besides these large bordered pits, the Spring tracheides, which have a wide lumen, show also small, ordinary pits irregularly distributed over the wall. The width of these tracheides is about 75 ju- The Autumn tracheides have only bordered pits. Their width is about 40 n. The medulla tra- cheides have small, round bordered pits on Straw pulp. Straw pulp consists mainly of bast fibers somewhat resembling flax fibers. Unlike flax, however, they are always pres- ent in paper almost undamaged. Further- more, the X-shaped displacements are lack- ing. In addition there arc large, round parenchyma cells without any contents. To a less extent annular, spiral and reticular vessels are found. Of ]irimary importance to 391 GENERAL MICROSCOPY fiber identification are the strongly pebbled, crenelated skin cells. Esparto pulp. This pulp is made from tlie leaves of Stepa tenacissima and Ly genus spartus. This pulp consists mainly of short, thin bast fibers closely resembling flax. But, here again like straw, the X-shaped displace- ments are lacking, while the fibers are hardly damaged even when pulped. These bast fibers are considerably thinner than those of straw. Small, sac-like skin cells are found throughout the specimen. The parenchyma cells, which are also present to a considerable extent, are usually badly damaged by pulp- ing. The specimens also contain pieces of annular and spiral vessels and also short bits of hair, shaped like commas, which are very useful for a diagnosis. It is difficult to estimate the percentage of esparto pulp mixtures which also contain straw pulp. After staining with Rhodamine 6 GD/Coriphosphine, an estimate is possible with the aid of the secondary fluorescence. Straw pulp fibers give a yellow green color and esparto a brownish green color. Ampar (Bagasse). This pulp is made from the sugar-cane waste. Characteristics of this paper material are finely porous parenchyma cells and differentially formed sclereides. The bast fibers have the following forms: (a) greatly thickened, fairly wide fibers (approx. 20-30 fx). (b) short, uniformly wide fibers with thin walls, (c) narrow, tapering fibers largely resembling straw fibers. The skin cells differ greatly in size and are not frequently found; they are usually badly damaged. Ground wood. The ground wood mostly found is soft wood. In Italy ground poplar is often also incorporated in paper. Under the microscope the normal image of soft wood or hard wood anatomy is recognizable in the ground wood particles. Ground soft wood is characterized by the wide fiber tracheides with their almost roimd bordered pits. Crosswise to these tracheides are the medullary rays built up of narrow and fairly long parenchyma cells. The drop- lets of resin characterizing soft wood are usually easy to see. Characteristic of groimd hard wood are the large vessels and the ab- sence of pits. Pulped Fiber. The morphology of fibers is greatly changed by pulping, the extent depending upon the degree of pulping and on the type of pulping machine. On the whole the changes are fairly great in the case of rags, hard wood and soft wood. Straw and esparto pulps change to a less extent. In the case of highly pulped material it may often be difficult to distinguish the various kinds with a microscope. Staining methods are then often fairly difficult and are conse- quently of little value as evidence. Table 2. Stains for Pulps (A) Unbleached (Colors in order of lower lignin content) Bleached Soft woo( sulfate Hard woe sulfite sulfate soda Rags Straw, es Bamboo Manila Jute Ground \ i (sulfite) and soda )d parto, tood gray green — pink brown green — gray green — gray violet gray — blue violet — violet gray green — green gray blue — blue yellow yellow red violet — red brown violet — blue violet blue violet — violet — light pink violet — blue red brown — brick red — lilac gray blue — blue yellow green 398 PULP AND PAPER (a) (c) Fig. 1. Pulps. Lower left : Soft wood sulphite pulp, phase contrast, 300X. Upper left : Hard wood pulp beaten in hollander to 57° S. R., phase contrast, 300X. Lower right: Cotton rags beaten in hollander to 50° S. R., phase contrast, 200X. Upper right: Straw pulp beaten in hollander to 38° S. R., phase con- trast, 350 X. When pulping is less thorough it is usually possible to identify the constituent fibers. In the case of rag fibers pulping causes in the first place internal fibrillation often at- tended by cross fracture. Despite this frac- ture, however, the typical fiber characteris- tics, such as the X-shaped displacements in the case of flax, usually remain fairly \'isible. This internal fibrillation is followed by exter- nal fibrillation, whereby a network of fine fibrils ari.ses. Cell wall membranes are not found. In the case of hard wood fiber, too, pulp- ing causes considerable fibrillation. In the first instance, similarly to rags, this is inter- nal. Besides this there is greater damage which loosens membranes, both from the fibei's and the vessels, which lie as two-di- mensional fibrils in the fiber suspension. These membranes ai"e clearly visible only 399 INDUSTRIAL HYGIENE MICKOSCOPY ^\ilh a phase contrast microscope. With a normal microscope these fibrils are visible as structureless slime. The fibers are often so badly affected by pulping that they can be distinguished only by staining methods. The large vessels are also torn apart. Often, however, the fragments of these vessels are sufficient for identification. On the whole, soft w^ood pulp shows a similar picture. Here again it is mainly the occurrence of two-dimensional fibrils that distinguishes them from rag fiber. As a rule, soft wood fiber disintegrates more than hard wood fiber. For strongly pulped specimens, pieces of tracheides with the typical single row bordered pits are important for identifi- cation, in addition to staining methods. Pulping of straw and esparto fibers has much less effect. Straw fibers are only very slightly affected (Fig. lb). Comparatively slight internal fibrillation, in addition to ex- ternal fibrillation of the ends of the fibers to fibril bundles is the sole result, even after protracted pulping. The skin cells j^how a strong tendency to ci'imible, while the paren- cyma cells are badly torn and pulped into large pieces. Difficulties in identification are hardly ever caused by pulping. REFERENCES Herzog, "Handbuch der Mikroskopischen Tech- nik fiir Fat^ertechnologen," Berlin, 1951. Herzog, "Mikrophotographischer Atlas der Teeh- nisch-wichtigen Pflanzenfaser," Berlin, 1955. Stoves, J. L., "Fiber Microscopj'," London, 1957. Wolff, Tobler, F.v.G., "Mikroskopische Untersu- chung Pflanzlicher Faserstoffe," Leipzig, 1951. J. ISINGS HiSTORADIOGRAPHY. See X-RAY MICROSCOPY Industrial hygiene microscopy (including refrac- tive INDEX MEASUREMENT) The microscope is used as a qualitative instrument in the field of industrial hygiene to determine w'hether the type of dust preva- lent in the working environment of the in- dustrial worker is of a toxic nature and thus may be injurious to health. Determination of the amount of the toxic dust in the air is also usually necessary to evaluate the degree of hazard. Approximate percentages, such as in the case of quartz dust, can sometimes be determined with the microscope. How- ever, in most instances, use of wet chemical analysis or additional instrumentation such as the spectrograph or x-ray diffraction is also necessary for more accurate quantitative results. Following this preliminary cjualitative and quantitative investigation, the microscope is used for counting the number of dust particles in the air and determination of their particle size. Results obtained may in- dicate that a definite health hazard exists. For example. Threshold Limit Values adopted at the yearly meetings of Govern- mental Industrial Hj^gienists state that in the case of dust containing more than 50% free silica, the number of dust particles in the breathing area of the worker should not exceed five million particles per cubic foot of air. This is particularly important in the case of small particle size dust which presents a greater health problem. Micro- scopic examination indicates that in pneu- moconiosis, the particle size of dust in the 400 INDISTI{IM, IIY(;iK\E MFCHOSCOI'Y lung is mainly below 5 microns in size, the largest not exceeding 10 microns. Information obtained as a result of micro- scopic investigation is usually reported to the Medical and Safety Departments of the company involved. In the case of conditions exceeding the Threshold Limit Value, an exhaust system to remove the dust from the worker's breathing area may be recom- mended. If an exhaust is already present, results may indicate that it is inadequate in size or not operating efficiently. Qualitative and Quantitative Analysis Identification of the dust is not always necessary since in many industrial opera- tions, the nature of the product is known. However, many materials such as abrasives, polishing and buffing compounds are sold under trade names with little information as to their chemical constituents. Prerequisite to microscopic examination, spectrographic analysis is often of value for determination of the elements present. For example, if it is found that the sample is largely silicon, there is the possibility that it is primarily silicon dioxide. Confirmation as to whether the material is silicon dioxide and its molecular form can be made with the microscope. Determination of molecular form is important since investigation has served to indicate that crystalline forms of silicon dioxide such as cjuartz, cristobalite and tridymite are more toxic than crypto- crystalline and amorphous forms. Preliminary microscopic examination should be made with the petrographic mi- croscope (Fig. 1) or with a laboratorj^ mi- croscope equipped with polarizing accessories (cap analyzer and disc polarizer). Examina- tion of the sample should be made with the analyzer and sub-stage polarizer in a crossed position. Particles observed will appear bright or dark on a dark background indi- cating whether they are single refracting (isotropic) or double refracting (anisotropic). Fig. 2 shows two forms of silicon dioxide as Fig. 1. Petrographic microscope. Fig. 2. Mixture oi {[uartz (briglit particles) and opal (dark particles). observed with the petrographic microscope. Amorphous opaline silicon dioxide appears dark due to the fact that the ray of light 401 INDLSTIU A!. IIY(;iENE iVIICHOSCOPY emitting from single refracting particles is lower index. Complete disappearance of the rejected by the analyzer located above the particle being examined, indicates that it has specimen. Quartz in the same preparation an index of 1.544 for yellow light, the index appears bright. This is explained by the fact for the ordinary ray of crystalline quartz, that double refracting particles divide the The analyzer is now reinserted and the ray of light from the polarizer into two rays stage rotated to the second extinction posi- (ordinary and extraordinary ray). As in the tion and estimation made on the same par- case of the opal, the ordinary ray is rejected tide as to whether the refractive index for bj'' the analyzer but the extraordinary ray, the other component vibration is higher than vibrating at right angles to the ordinary 1.544. If examination by the Becke line ray, is transmitted through the analyzer. As method proves this to be true, the above a result, the quartz particles appear bright described procedure, as used for the 1.544 on a dark background. index, is now repeated using a liquid of This preliminary microscopic examination 1.553. Disappearance in this liquid indicates serves to classify the unknown as to whether that the particle in question has a second it belongs to the isotropic or anisotropic index of 1.553 for the extraordinary ray and group of chemicals and minerals. Additional can thus be considered crystalline cjuartz. optical properties such as refractive index Additional identifymg procedures such as or indices are necessary for identification. In examination of interference figure and de- view of the possibility that the bright aniso- termination of optic sign used by the petrog- tropic particles (Fig. 2) might be crystalline rapher for mineral specimens are difficult or quartz, a Handbook of Chemistry or Text- not possible on small particle size dusts, book of Mineralogy is consulted. Informa- A simplified procedure for the identifica- tion obtained indicates that quartz has two tion of dust particles is to examine the sam- refractive indices, 1.544 for the ordinary ray pie with a dispersion staining dark-field mi- and 1.553 for the extraordinary ray. croscope. This method (1) of illumination is Determination as to whether the unknown obtained by the addition of the accessories has these refractive indices can be made by shown in Fig. 3 to a laboratory type micro- use of a petrographic microscope illuminated scope. For small particle size dust, a combi- with a sodium light source. The sample is nation of a 20 X (10.25mm) 0.40 N.A. ob- mounted under a cover glass in a liquid of jective with a 20X hyperplane eyepiece is refractive index 1.544 for yellow light, and suggested. The condenser employed is a 1.40 examined with the analyzer and polarizer of N.A. achromatic condenser but with the top the microscope in a crossed position. The element removed to give an N.A. of 0.59. stage of the microscope is rotated until a The diameter of the dark-field stop used particle being observed appears dark. The below the condenser is very critical, a diam- analyzer is then removed and the index of eter of 17 mm usually giving good results, the particle with respect to the liquid is de- Use of a cap analyzer over the eyepiece is of termined by the usual Becke line procedure, value for the identification of anisotropic Using this method, a halo of bright light dusts such as quartz. may be observed moving into or away from As in the usual petrographic method de- the particle as the focus of the microscope scribed above, the sample is placed in a re- is slightly changed from exact focus. If the fractive index liquid which in the case of halo moves into the grain on the up focus, quartz would be 1.544 for yellow light. In the particle has a higher index than the place of a sodium lamp, a 6 volt — ^108 watt liquid; movement of the halo away from the ribbon filament lamp is used for illumination, particle into the index liquid signifies a Using this dark-field method, yellow light 402 INDUSTRIAL HYGIENE MICROSCOPY Fig. 3. Microscope accessories for dispersion staining. for which the quartz and Hquid are equal m refractive index will pass straight through the particle oblique to the optic axis of the microscope and thus not enter the objective. Blue and red light for which the quartz is not equal in refractive index are refracted at the particle-liquid interface to a degree as to enter the objective. Due to the greater bending of blue light as compared to red, the particle when equal to the liquid for yellow light appears largely blue with a trace of red. If the small amount of red is somewhat diffi- cult to observe, increased visibihty can be obtained by very slightly decentering the dark-field stop. Success with this method, especially in the case of small particle size dust, is based on the use of index liquids of much greater dis- persion than the sample examined. The greater the difference in dispersion between the sample and the liquid, the more brilliant the colors. For the identification of quartz, styrene stabilized with ter t-huty\ catechol has proved to be of value. It has an unusu- ally high dispersion for its refractive index of approximately 1.544. As examined in this liquid with the dispersion staining dark-field microscope, using a cap analyzer over the 20 X hyperplane ej^epiece, quartz particles oriented for the ordinary ray, 1.544 appear largely blue with a trace of red. Particles oriented for the extraordinary ray, 1.553 appear blue with a large amount of red or are homogenously red in the case of very small particle size. Rotation of the cap analyzer 90 degrees results in particles oriented for the ordinary ray (appearing blue with a small amount of red) to change in color to blue with a large amount of red, or all red. Particles oriented for the extraordinary ray, change in color to those significant for the ordinary ray. Additional dispersion colors can be ob- served which are of value for identification. In place of a 1.544 index Hquid, a 1.553 Hquid equal to the extraordinary ray of quartz can be used and notation made as to the change in color on rotation of the cap anah'zer. Dispersion colors can be observed indicating whether the dust sample is above or below the index liquid in refractive index. A pure blue with no trace of red is obtained when the particle is slightly lower in index than the liquid; light blue signifies a still lower 403 iNDi SUM VI. in<;ii:M: mickoscoi'Y index. A l;iifi;(' uiiioiiiit of red iiidicjilcs a iiifi; ai-ca may not always repres(?iit affurate slifrlilly IiI^Ikt index than the li(|nid and inloiniat ion as to 1 ho amount of toxic dust orange and \'cll(i\v a still higher index. If hrcat lied hy t lie op(!rator. For more accurate solid and liiiuid arc \'ci'y fai' apart in index, results, the sample can l)e colle(;ted in water dispersion coloi's will not l)e observed, tlic or alcohol in a ^ li-cenl)Ui'}i;-Smith or Midget sample appearin}:; white. To deteiinine as inipingei- at t he breathing area of the opera- to whether the while appearance indicates tor. an index difference far above <)V fai' below, After collection, the iinpin}j;ei- tube is the dark-held stop must be removed and the placed in a water bath to evaporate off the I^eeke line metliod of index detei'ininat ion water or alcohol leaving the dry dust in the (Muployed. impiiig(!r tube. A small amount of index This same procedure of dark-field illuini- Tuiuid such as styrene, used in the identifica- nation can be used for the determination of t ion of (juailz, is placed in the impinger tube oilier toxic dusts sucii as isoti'opic opal (I'ig. and thoioughly mixed with the dust sample. 2). This foi'm of silicon dioxide having only A rubber policeman is of value for l)o1h mix- one i'efracti\-e index shows no change in ing and transferring the sampk; to a micro- color as the cap analyzer is rotated since it scope slide. The preparation is covered with has the same optical properties in all direc- a cover glass and inspection made using the tions. Posit i\'e identification of opaline silica dark-field dispersion staining microscope to is more dilliciilt I lian ot hei- forms of silicon detei'mine the approximate percentage of dioxide since it has a variable refractive in- toxic dust such as (juartz present in the (lex. That ind(\\ can vaiy from l.ll to l.f(i preparation. More accurate (}uantitative re- depending on the amount of water present suits can be made by comparison with a ranging fi'om 1 to '20 per cent. Another prob- standard sample containing a known amount lem with the lower index dusts such as ()j)al, of toxic dust. Two microscopes can be used cristobaIit(> and tridymite is to find identify- in this determination. The stage of one mi- ing index li(|ui(ls that have a much greater eroscope contains the unknown sample and disp(M'sion than the sample such as to give the stage of the second microscope the known biilliaiit dispersion colors in the case of small sample. The eyepieces of both microscopes particle size. In general, the lowei- index are connected by means of a comparison eye- liquids have low dispersion, in sonu> cases piece (Fig. 4) which permits examination of iiaving a dispiM'siou only slightly exceeding both samples in a circular field of view di- that of the sam))le examined. This probUnn vided vertically, one half being the unknown can be largely solved by the use of ethyl sample and the other half the known, cinnamate. Like styrene, it has a high dis- Other methods of (luantitative micro- persion foi' its index of 1.557. When mixed scopic analysis of mixtures are described by with ethyl phosphate, a series of licjuids Chamot and Mason (2). Ross and Sehl (3) ranging in index fi'om l.flO to 1.557 can be describe a method for determination of the prepared having higlu^r dispersions than the percentage of cjuartz in a mixture using the index licjuids commonly used for the Becke usual Becke line methods of identification, line method of index determination. F.mploying this procedure, particles are As indicated in the fii'st part of this articl(% counted and assigned weights dependent on determination of the approximate amount of their size as compared to the size of scjuares the toxic dust can in some instances be made of a Whipple disk placed in the eyepiece of by use of the microscope. This analysis is the microscope. This procedure should be sometimes made on ledge or rafter samples more easily and accurately done by use of which altiiough collected close to the work- dispersion staining. By proper selection of 404 INDLSTKIAL J1V(;IE.\E MICROSCOPY Fig. 4. Comparison eyepiece. refractive index liquids, the toxic dast such as quartz could be obser\'ed in one color, material not c^uartz in other colors of the spectrum or white if far removed in refrac- tive index. A variation in the impinger method of collection ls to use the identifjang index liquid as the collecting liquid in place of water or alcohol. In this case, Uquids of less volatility than st\Tene should be used. In- dex Uquids cormnonh' emploj'ed are a mix- ture of two hquids which may differ .slight h' in vapor pressure. As a result, during the procedure of collection of the .sample, there may be greater evaporation of one of the constituents resulting in a slight change in refractive index. For accurate results, after collection of the dust .sample, a drop of the collecting hquid should be checked on a refractometer and adjustment of its index made if necessarv bv addition of either a small amomit of the lower or higher index components. Using the identifWng hquid as the collecting medium, evaporation on a water bath is not neces.sary. The impinger sample can be stirred with a rubber police- man and a drop transferred to a shde for microscopic examination. Quantitative results obtained should not be based on the e.xamination of one sample. Although di.spersion staining is probably the simplest method of microscopic examination, considerable experience is required in the case of small particle size du.st. It has been our experience that more reliable information can be obtained by preliminary examination of a known sample, such as quartz, previous to e.xamination of the miknown, in order to determine exact dispersion colors on rotation of the cap analyzer. The unknown is exam- ined in the same way to determine whether particles present show the same shade of col- 405 INDUSTRIAL HYGIENE MICROSCOPY Fig. 5. Phase microscope. ors as the known. Both samples should be examined at approximately the same tem- peratm'e because the refractive index of a liquid changes with a change in temperature. For many liquids, the rate of change may have a value of around 0.00045 for each degree centigrade of variation. This temperature factor is of value in some cases to obtain brighter dispersion colors which aid in increasing the visibility of small particle size dust. For example, if the preparation containing quartz is exam- ined in a warming stage in a liquid of 1.544 adjusted to that index at 25 degrees centi- grade, increasing the temperature will de- crease the index of the liquid. As a result, quartz particles appearing blue with a trace of red can be made to appear homogene- ously red, orange or yellow^ dependent on the temperature used. Consideration must be made, that although the contrast on a dark background is greater, identification by dispersion colors other than blue with a trace of red results in decreased accuracy. However, this possible error is reduced pro- vided a sample of known material is simul- taneously examined with the unknown by means of comparison dispersion staining microscope, comparing the change in color of both samples as the temperature is in- creased. Counting Dust Particles The number of dust particles in the air is usually determined by first collecting the dust in an impinger tube using water or al- cohol as the collecting liquid. After collec- ion, a 1-mm aliquot of the sample is placed in a dust counting chamber such as a Sed- wick-Rafter or Dunn cell and counted under the light-field microscope employing a lOX objective in combination w^ith a 7.5 X or 10 X eyepiece. Higher magnification oculars can be used but the resulting magnification should not exceed 1000 times the numerical aperture of the objective. An objection to the use of the light-field microscope is the fact that dust particles of small size, having little or no color and close to the index of the collecting liquid, will be difficult or impossible to see. For this reason, we have suggested (4) the use of the dark contrast phase microscope (Fig. 5) for count- ing dust particles which results in greater accuracy and ease in counting. Fig. 6 is a comparative photomicrograph of magnesium fluoride dust in isopropyl alcohol as observed by both light-field and phase microscopy. It is rare that dusts of such low index and proximity to the index of the collecting liquid will be encountered in the usual work- ing atmosphere. However, these photomicro- raphs serve to illustrate the effectiveness of the phase microscope for rendering visible dust particles having a refractive index differing only slightly from the refractive index of the examination liquid. According to Chamot and Mason (2), the use of the phase microscope gives values to an addi- tional decimal place beyond that obtained by the usual Becke line method. A number of papers have been published in the last few years suggesting the use of 406 INDUSTRIAL HYGIENE >lICROSCOPY molecular filters in preference to the im- pinger method for collection of dust. Molec- ular filters are cellulose ester gels resembling ordinary filter paper in appearance but have a pore size such that they can retain dust particles of small particle size. One of the main advantages of their use, as pointed out by First and Silverman (5), is the fact that there is negligible penetration of the filter. Dust particles are deposited on tin surface in the same state in which they existed when suspended in air. In using these filters, the usual procedure is to place the filter in a holder designed for that purpose and collect the dust in the air at 0.1 to 0.5 cubic foot per minute using an elect ricall}^ driven pump, Freon-powered equipment, or hand pump. After collection, the filter or a section of it is placed dust side down on a clean slide or on the ruling of a counting chamber such as a hemacytometer. Since the filter is opaque and transmitted light using the light-field microscope has been the usual method of counting, it is necessary to render the filter transparent by applica- tion of a drop or two of liquid equal to the refractive index of the filter. A major objection to the above procedure, as pointed out by Drinker and Hatch (6), and by Paulus, Talvitie, Fraser and Keenan (7), is the fact that toxic dust particles such as quartz and diatomite close to the index of the clearing liquid will not all be included in the count, especially if of small particle size. A solution to the problem, as in the case of the impinger method of collection, is the use of the phase microscope. The superi- ority of phase illumination as compared to the usual light-field method is illustrated in Fig. 7 of diatomaceous earth particles col- lected on a molecular filter (Millipore type AA). The filter was rendered transparent by appHcation of a drop of liquid of 1.507 at 25°C and covering with a cover-glass. Em- ploying the phase microscope, the liquid used should be equal or very close to the index of the filter in the third decimal place Fig. 6A. Magnesium fluoride. Light-field illu- mination. ULJi. i* awiMiii :'-^k. iLJLJLJul 11 ' -, i-. ■ » . *- *.* 3' . - ,X.\ t ' ^ " - '■]7'!.t: M ^\ .; ; ^ A. . » * -.^ ' \ nr:'' !«-3 ■ t, ■ws«."T| »'*-~^«1| l(P~™«« »w j; m Fig. 6B. Exactly the same area as figure 6A. Phase illumination. in order that the texture of the filter will not be revealed. Fig. 8 is a photomicrograph of diatomaceous earth collected on a Milli- pore Filter, tj^pe AA using a liquid of index 1.515 as the clearing liquid such as has been suggested for the usual light-field method. Although satisfactory for the light-field technique of counting, it cannot be used with phase illumination. As noted in the photomicrograph, it is difficult or impossible to differentiate small diatom particles from the texture of the filter. The liquid used for clearing must have no solvent effect on the filter. IMolecular filters are solubk^ in ('(Mtaiii ketones, esters and methanol. A mixtiu'e of light mineral oil with Ai'oclor 1242 adjusted to an index of 1.507 on a refractometer at 25°C has proved 407 INDUSTKIAL IIYGIKNE IVIICKOSCOPY tfS" Fig. 7A. Diatomaceous earth particles on a Millipore type AA filter. Cleared in a liquid of refractive index 1.507. Light-field illumination. k'i f of litjuids vary with temperature. Slight dif- ferences are not important but use of a 1.507 clearing liquid corrected for that index at 25°C and used at a much lower or higher temperature may result in some objection- able visibility of the filter. Deterniiiiatioii of Particle Size Determination of the particle size of the dust at the approximate breathing area of the worker can be made by examination of the impinger sample collected for the purpose of making the dust count. If a hemacytom- eter was used as the counting chamber, some idea of the particle size can be obtained by comparison with the ruled areas of known size. Increased accuracy is obtained by use of a micrometer disc in the eyepiece or a Filar Micrometer eyepiece (Fig. 9). Before use, the divisions of the micrometer must be calibrated by use of a stage micrometer hav- ing a scale of known area, the smallest divi- sion being .01 mm (10 microns). Calibration of the eyepiece disc is made by first focusing on the stage micrometer. The stage microm- eter is then moved to a point so that one line of it coincides with a line to left center of the eyepiece scale. Count is then made Fig. 7B. Exactlj^ the same area as figure 7A. Phase illumination. satisfactory. Another possibility if this . 1.507 liquid is not available, is isoamyl salicylate. It is a colorless liquid with pleas- ant odor, has no solvent effect on the filter and its refractive index is usually close enough to 1.507 at 25°C as to render the r»W5f«^4 filter quite transparent. In using these lici- uids, consideration must be made as previ- ously mentioned, that the refractive index mA Fig. 8. Diatomaceous earth particles on a Millipore type AA filter. Cleared in a liquid of refractive index 1.515. Phase illumination. 408 INDUSTRIAL HYGIENE MICROSCOPY across the eyepiece scale to the right from this point to another point where a line of the eyepiece scale coincides with a line on the stage micrometer scale. , If a large number of dust samples arc to be examined, use of a micro-projector (Fig. 10) has proved to be the preferred method. In this case, the dust sample is projected onto a screen. Particle size distribution is deter- mined by comparing the size of the projected images with a ruled area on the screen of known size. Another procedure is to measure the projected particles with a millimeter rule. Knowing the magnification, the value in microns for each division of the rule can be determined. For example, if the choice of optics and projected distance is such as to give a magnification of 1000 X, a particle having a diameter of 1 mm as measured with the rule has an actual size of 1 micron. In preparation of a particle size distribu- tion curve, the total number of grains meas- ured may vary from 200 to 2000 dependent on whether the dust sample contains a few sizes or many sizes. The number of each micron size counted is multiplied by the cube of the diameter to obtain the relative amount of that size present in the sample and the percentage of the total amount is calculated for each micron size. To obtain an accumulative curve, the percentage of the total sample larger than each size is plotted against the diameter in microns. Considerable variation in the sizes ob- tained can result dependent on the method of collection. For example, disintegration of particles by high-velocity impingement may give an excess of fine particles. The use of liquid collecting medium such as water may result in solution of many particles. As pre- viously indicated, due to proximity in index of the collecting liquid and the dust, small particle size material as examined by the usual light-field microscope will be difficult or impossible to see. As a result, these par- ticles will not be included in the preparation of a particle size distribution curve. Fig. 9. Filar micrometer ^ Fig. 10. Micro-projector. Drinker and Hatch (6) suggest the use of the Molecular Filter and Thermal Precipi- tator for collection of dust for particle size determination. Collecting efficiency is high over the entire range of interest and the particles are deposited without being subject to physical stress. In the case of molecular filter samples, as previously described, a liquid of approximately 1.507 refractive in- dex must 1)0 used for rendering the filter transparent. Approximate information as to size can be made by comparison with the rulings of the hemacytometer and more 409 INDUSTRIAL HYGIENE MICROSCOPY accurate information by use of a calibrated ej^epiece micrometer. As in counting, the use of phase microscopy should increase the accuracy in determination of particle size. According to Richards (8), more precise measurements can be made with phase mi- croscopy owing to the sharp edges of speci- mens free from indefinite diffraction pat- terns. Also, because of their definite edges, one can more readily determine whether a particle is single or aggregate. A limitation of the usual particle size dis- tribution curve obtained by use of the opti- cal microscope, is the fact that particle size below its resolution is not included. Also the fact that little or no information is usually obtained as to the size of various constituents of a mixture. For example, particle size infor- mation on dust consisting mainly of clays and quartz sand in a ceramic plant may indicate that a large percentage of the dust is below 3 microns in size. It is important to know whether the small particle size dust is mainly quartz or clay. If present as quartz, the degree of hazard is considerably greater than in the case of small particle size clay. Simultaneous measurement of particle size and identification of the particles measured can be made in many instances by use of the dispersion staining dark-field microscope. Unfortunately, samples collected on mo- lecular filters cannot be used for this deter- mination due to the fact that the index liquid employed must be selected for the purpose of rendering the filter transparent rather than for the purpose of identification of the particulate material collected. An exception is the rare case when the dust on the filter has an index equal to or very close to the index of the clearing liquid. Samples col- lected by other methods can be measured by use of the dispersion staining dark-field microscope. For example, in the case of dust from the cerami(^ plant containing primarily quartz sand and clay we are mainly inter- ested in the size of the quartz particles. If examined in styi'ene, the quartz particles measured will appear blue with a trace of red changing in color on rotation of the cap analyzer to blue with a large amount of red or homogeneously red dependent on their particle size. Clay particles such as kaolinite present in the preparation, having different indices, appear in other colors of the spec- trum or white and are thus readily distin- guished from quartz. REFERENCES 1. Crossmon, G. C, "New Developments in Dis- persion Staining Microscopy as Applied to Industrial Hygiene", Am. Ind. Hyg. Assoc. Quart., 18, 341 (1957). 2. Chamot, E. M. and Mason, C. W., "Handbook of Chemical Microscopy", 3rd Ed., John Wiley & Sons, Inc., New York, (1958). 3. Ross, H. L. AND Sehl, F. W., "Determination of Free Silica-Modified Petrographic Immer- sion Method", Ind. Eng. Chem., Anal. Ed., 1, 30 (1935). 4. Crossmon, G. C, "Counting of Dust Particles b}^ Phase Microscopy", A. M. A. Arch. Ind. Hyg. Occup. Med., 6, 416 (1952). 5. First, M. W. and Silverman, L., "Air Sam- pling with Membrane Filters", A. M. A. Arch. Ind. Hyg. Occup Med., 7, 1 (1953). 6. Drinker, P. and Hatch, T., "Industrial Dust", 2nd Ed., McGraw-Hill Book Co., New York, N. Y. (1954). 7. Paulus, H. J., Talvitie, N. A., Eraser, D. A., AND Keenan, R. G., "Use of Membrane Fil- ters in Air Sampling," Am. Ind. Hyg. Assoc. Quart., 18, 267 (1957). 8. Richards, O., "Photomicrography with the Phase Microscope," Photographic Soc. Amer- ica J. Sec. B., 16, 94 (1950). Germain Crossmon 410 Infrared microscopy The design requirements of infrared mi- forms a reduced image of the exit sht at the croscopes, or more accurately infrared micro- sample space. The objective collects the en- sampling attachments, have been rather ergy which has passed through the sample thoroughly explored in recent years. All have and forms a magnified image of the sample the common property of measuring, as a at an adjustable diaphragm. The energy then function of wavelength, the infrared absorp- passes to a second field mirror which forms tion of minute samples. The resulting ab- an image of the Littrow mirror near the cen- sorption spectra are similar to the infrared ter of curvature of the thermocouple con- spectra of macroscopic samples so that the denser, which forms a reduced image of the well known applications of infrared (see En- slit on the thermocouple detector, cyclopedia of Spectroscopy) — qualitative The microscope can be placed hi the spec- and quantitative analyses and molecular trometer either between the source and the structure investigations — are possible. Infra- entrance slit of the monochromator, or be- red microscopes have unique importance tween the exit slit of the monochromator and where (a) the amount of sample is small, the detector; the latter avoids difficulties in (b) the dimensions of the sample are small samples sensitive to heat or photochemical and (c) the sample is not homogeneous. As effects. Special precautions may be taken to little as 0.1 microgram of sample in the field eliminate absorption from atmospheric wa- of the microscope can yield useful spectra, ter vapor which may seriously interfere with Samples which have been studied using spectra for ver}' small specimens (5). such equipment include natural and syn- A Cassegrain-type condenser permits the thetic fibers, single crystals, biological tissue substitution of photoconductive cells, photo- sections and bacterial cultures. Instruments multiplier tubes or other types of detectors of this type allow the study of liquid extracts in place of the thermocouple, and solutions in cells of extremely small vol- In practice the required thickness of the ume. This means that it is now possible to sample is about the same as for macroscopic use samples of compounds separated by work, frequently about 25 microns. The chromatography. The technique for compress- minimum sample area is that required to ing finely ground samples mixed with KBr provide sufficient energy for satisfactory de- powder into optically clear pellets of optimum tection — that is, inversely proportional to dimensions is now a very useful sampling source brightness, the transmission effi- procedure for infrared microspectropho- ciency, the detector sensitivity and the tometry. Polarization effects permit investi- square of the effective numerical aperture gation of molecular orientation in dichroic of the rnicroscope objective. Thus the area samples. may be decreased by the use of such intense In the Perkin-Elmer infrared microscopic sources as carbon arcs, zirconium arcs or attachment (4) to their infrared spectrom- tungsten glowers. Other controllable instru- eter, energy from the exit slit of the mono- mental factors of course provide limits of chromator is incident on a field mirror which size. If the sample does not cover the entire directs it upward and forms a reduced image field of the microscope, some radiation will of the pupil of the monochromator (Littrow reach the detector without being subjected mirror) near the convex mirror of the con- to absorption by the sample. This "dilution" denser, so that radiation from the entire use- of the spectra is also increased by aberra- ful slit is directed to the condenser, which tions and diffraction introduced by the ob- 411 INTKHKKKKIVCE MICROSCOPY jectivc. It is important to avoid impurity radiation in quantitative work on long, narrow samples such as fibers. A nimiber of excellent infrared spectra made with the microscope attachment have been published, including tissue sections, single fibers (17- 20 fi), licjuids in short lengths of AgCl capil- lary tubing (3), and single crystals (6, 8, 9). The determination of optical constants in the infrared especially for metals is a special contribution of Beattie and Conn (1, 2). Military surplus infrared image electron con- verter tubes continue to find application for the infrared examination of opaque dyes and pigments, and silicon ingots. In the latest work according to news reports the infrared microscope is adaptable to the examination of substances which can be differentiated on the luisis of their specific infrared emissivity and reflectivity, as well as absorption (7). REFERENCES 1. Beattie, J. R., Phil. Mag., 46, 235 (1955). 2. Beattie, J. R. and Conn, F. K. T., Phil. Maq., 46, 222 (1955). 3. Bloxjt, E. R., Parrish, M., Bird, G. R., and Abbate, M. J., J. Opt. Soc. Am., 42, 966 (1952). 4. Coates, V. J., Offner, A., and Siegler, E. H., /. Opt. Soc. Am., 43, 984 (1953). 5. Eraser, F. D. B., J. Opt. Soc. Am., 43, 929 (1953). 6. LowENTHAL, S., Rev. Opt., 34, 29 (1955). 7. Maresh, C, Coven, G., and Cox, R., Anal. Chem., 30, 829 (1958). 8. NoMARSKi, G., Rev. Opt., 34, 29 (1955). 9. NoRRis, K. P., /. Sci. Inst., 31, 284 (1954). G. L. Clark Interference microscopy FIBERS (TEXTILE). See GENERAL MICROSCOPY, p. 343. INDUSTRIAL RESEARCH, APPLICATION TO. See GENERAL MICROSCOPY, p. 363. INSTRUMENT CLASSIFICATION AND APPLICATIONS Interference microscopy is a techiiicjue for the microscopic observation of specimens, whereby two superimposed fields of view are presented to the observer. One field con- tains an image of the specimen and is called the "image field"; the other (the "reference field") differs from the image field in some way, but the two fields are mutually co- herent at every point. Contrast is produced by the effect on the interference phenomena of the path-differences caused by the optical thickness (for transparent objects) or surface contour (for reflecting objects) of the speci- men. It will thus be seen that interference mi- croscopy makes use of the same properties of the specimen as does phase contrast mi- croscopy. The advantages of the interference method are that certain artefacts (the "phase-contrast halo") are avoided, that measurements of path difference can be made precisely, and that the nature of the contrast can usually be altered by varying the overall path difference between the two interfering fields to obtain the best condi- tions for the detail being studied. A further advantage is that in some types of inter- ference microscopes the aperture of illumi- nation is not restricted. Classification of Instruments It is convenient to classify interference microscopes by the way in which the image and reference fields differ from each other. Three classes can be recognised, as follows: I. The reference field is formed by light which has had no contact with the object. IT. The reference field contains an image 412 INSTRUxMExNT CLASSIFICATION AND APPLICATIONS of the object which is out of focus or which has suffered very heavy aberrations. III. The reference field contains an image of the object which is displaced laterally. Each of the above three classes can also be divided into instruments for transparent or for opaque objects. As all these types of instruments possess both advantages and disadvantages, the selection of a type suit- able for a given kind of work is a matter reciuiring some care. In instruments of Class I the information given by the interference pattern is quite explicit, i.e., the relationship between in- tensity and optical thickness of the object does not depend on the properties of the object at points other than the one being considered. This may be contrasted with the state of affairs m phase-contrast microscopy where the intensity is related to the mean optical thickness in the neighborhood of the object point. This is not necessarily the case in instru- ments of Class II, where an effect similar to that seen in phase-contrast microscopy may arise due to the influence of the image in the reference field. The effect of this image may be removed, however, by suitable design, as described later. In instruments of Class III the informa- tion may be explicit if the object is so small that the images in the two fields do not over- lap; otherwi.se the intensity is a fmiction of the difTerence of optical thicknesses at two points at a distance apart equal to the sep- aration of the two images. If this distance is made very small the two images will appear almost as one and the intensity will then very nearly be a function of the gradient of optical thickness in the direction of the dis- placement of the two images. The Formation of Contrast In order that the image and reference fields should be coherent it is neces.sary to fulfil two requirements. Firstly, both fields must be illuminated by light from the same cross-section of the light beam from the .source and which is distributed in exactly the same way in the two fields. Secondly, the two fields mast be aligned so that points in each field which correspond from the point of view of the source are superimposed with an accuracy which, it can be shown (1), is equal to the resolving limit of a perfect objective of aperture equal to that of the illuminating condenser. For good fringe con- trast it is also necessary that the intensities of the two fields be equal. Under these conditions, if there is a path difference p between the two fields, which are taken each to be of unit intensity, the intensity given by the combined fields is given by 7 = 2(1 -1- cos 27rp/X) a) where X is the wavelength used. If a detail in the topograph}^ of the specimen introduces an additional small path difference dp, the corresponding change, dl, in intensity is (4ir/\)dp sin 2irp/\. The contrast C is then given by C = dl 2irdp wp tan — X X {2) It is evident from {2) that the contrast at any point can be brought to the best value by suitable adjustment of p, the "back- ground" path difference. The contrast will be greatest when p/X is an odd multiple of X, but will not in practice become infinite becau.se of parasitic hght in the instrument. In general, however, such values of p will not give the best conditions, for then the intensity will depend on the square of dp, giving reduced weight to details with small values of dp and concealing the sign of the path-difference variation. Furthermore, as these values of p correspond to dark-field conditions, the intensity of the image is greatly reduced. It is generally better to depart somewhat from the.se conditions to obtain a more nearly linear variation of in- tensity^ with path difference and retain a reasonable image brightness. It can be seen 413 INTERFERENCE MICROSCOPY from (^) that either positive or negative vakies of C can be obtained. If white hght is used color contrast is also present, which in many cases is useful in resolving the ambiguity which occurs if the optical thickness of a specimen exceeds one wavelength. It also can give rise to images of great beauty. Interference Microscopes for Trans- parent Objects Instruments of Class I. Some types of instrument designed for opaque objects can be used for transparent objects and fall into this class. One of these is the Linnik (2) microscope which consists of a Michelson interferometer with a microscope objective built into each arm (Figure 1). The trans- parent object is placed on a plane mirror and is imaged by one objective; the other images a plane mirror only. This arrange- ment has some advantages; the optical ar- rangement of the Michelson interferometer allows good control of the background con- ditions, giving the possibility of introducing fringes of controllable width into the field and of varying the background path differ- ence, and the intensity in the two beams can be controlled for good contrast of the fringes. Against this, the two beams pass through physically separate optics, which must be selected for equality of aberrations and chromatic dispersion of path difference. Also, as the beams are widely separated, the instrument is very sensitive to vibration. In consequence it must be built into a very ^SI Dooble. f^ocus le.ns <> E BecD rn Spl if te.r jL yep leoe Tnonspanent spe-oirnen. Fig. 1. Linnik interference microscope. Objective Object C O ncJ e rt se. r- Fig. 2. Smith's interference microscope with double focus lenses. heavy and rigid stand, which is necessarily costly, and cannot be used for path-differ- ence measurements of great precision. A further difficulty which arises when using this instrument for transparent specimens is caused by the finite distance between the specimen and its image in the mirror. Other instruments in this class will be described under the section on "Instruments for Opaque Objects." Instruments of Class II. Instruments of this class have the great advantage that both beams traverse the same optics and are never very widely separated. This confers considerable immunity from the effects of vibration and also eliminates the necessity for matching optics for aberrations and path difference. It is very desirable that splitting and recombination take place in the space between condenser and objective. One instrument of this class is the polariz- ing interference microscope of Smith (3) (Figure 2). Double-focus lenses of bire- fringent material are placed in conjugate planes as shown, forming two axially sep- arated images of the light source; the object is placed in one of these, and the two beams are re-combined after passage through the objective. Two images of the object are seen, one of which is out of focus and causes an effect similar to the "phase-contrast halo," making a measurement of optical thickness at a 414 INSTRUMENT CLASSIFICATION AND APPLICATIONS point depend on the mean path difference over a surrounding area equal to the out-of- focus disc corresponding to the axial dis- tance between the two image.s. This effect would not be serious if this distance could be made large, but in practice it is limited by considerations of aberrations of the back focal plane of the objective, so path differ- ences measured by this instrument should be treated with some reserve. To avoid this effect another type of po- larizing microscope due to Dyson (4) (Figure 3) employs two thick plates of Iceland spar cut parallel with the axis, one above and one below the object, the working distance of the objective being increased by a suitable op- tical system. The out -of -focus disc is now strongly astigmatic. As both sphtting and recombination take place in the object space the limitations on the size of the disc are removed and it can be made larger than the field of view, which considerably reduces its effect on the micertainty of measurement. As both these instruments make use of beams which are polarized in mutually perpendicular directions, polarimetric means may be used to measure the path difference with considerable accuracy. The effect of the out-of-focus disc may be removed entirely by another instrument due to Dyson (1) (Figure 4). Two glass plates, one or two millimeters thick and partially silvered on both surfaces, replace the Iceland spar plates shown in Figure 3. The ray paths are similar to those in a Jamin interferom- eter and the out-of-focus disc can be made Ot>jec1-ive / Wolf- splver-ed t / Sur.'Paces,. " 1 - Shghf 1^ v^eoge "^^ / Shaped plates ^ \' ZJ^ XceloncJ span plat'G.s c:u+ parol lei Wii^t^,^- Opl'ic axis UJ Objective Qloss block w/itl-i upp*in suntcJCe Silvered Object >/2 Plate Fig. 3. Dyson's interference microscope with crystal plates. Fig. 4. Dyson's interference microscope with half -silvered glass plates. many times the diameter of the field. As the reflecting system introduces a small central stop, a dark center, which can be made larger than the field, occupies the center of the out-of-focus disc, which therefore has no influence on the measurement, which is quite explicit. The background path difference and the number of fringes in the field are both varied by making the two plates slightly wedge- shaped. The number of fringes is varied from zero (uniform field) to a maximum by rota- tion of one plate about the microscope axis, and the path difference by translation of the other plate by means of a micrometer screw. By means of a photometer eyepiece measure- ments of path difference may be made with an accuracy of X/300. The highest powers of objectives can be used and the aperture of illumination is unrestricted. As the separation of the images is so large, the leveling adjustment of the plates is quite critical, and the instrument is therefore more difficult to adjust than the others in this class. This is the price paid for unambiguous measurement and freedom from optical arti- facts. Instruments of Class III. Instruments of this class are usually known as "shearing" microscopes. An early instrument due to Lebedev (5) (Figure 5) made use of two Iceland spar plates cut at 45° to the optic axis, with an intervening half-wave plate to interchange the planes of polarization. Two images are thus seen, sheared laterally with respect to each other, one being somewhat 415 IMKKI KKENCE MICROSCOPY >■/. 2 Rote Ci^ystal plc3l"es cut at| T -45° with optic axes Fig. 5. Lebedev's microscope. V^oMQs1"on pnism Objective Con Wollas+On pnism Fig. 6. Smith's shearing interference micro- scope. astigmatic. A similar principle has been used by Smith (6) for a high-power instrument, replacing the lower plate by a plate of ciuartz of increased thickness, the half-wave plate being no longer required. This is of impor- tance in high-power instruments where very large convergence angles are employed. The amount of shear is restricted by the thickness of the plate which can be employed, so both images appear in the field. In consec}uence only a limited area of the edge of a large specimen can be examined without overlap and loss of explicitness. On the other hand, the limited shear makes for increased ease of adjustment. A critique of this and the Dyson instru- ment in Figure 5 has been published by Davies (7). Another type of shearing microscope has been described by Smith (3), (Figure 6), using two Wollaston prisms in conjugate planes as shown. This avoids the astigma- tism of one of the images, but this time the shear is limited by aberrations of the con- jugate planes. Tl is interesting to notice that the lower Wollaston may be dispensed with if the condenser apertvu'e is reduc^ed to a nai-row slit parallel with the apex edges of the prism, but this will of course give rise to some deterioration of the image (luality. The shear is produced in an instrument due to Frangon (8) by means of a Savart plate just below the eyepiece, using slit illumination. Shearing microscopes of low power have considerable advantages of simplicity and stability, making them perhaps the best instruments for the important field of the measurement of thin films, as described later. Interference Microscopes for Opaque Objects The distinctions between the three classes of instruments for opaque objects are if anything of greater significance than in the case for transparent objects, for path differ- ences produced by reflection are usually greater than those produced by transmission. Instruments of Class I. The Linnik instrument described above is typical of this class, and the discussion of this instrument as applied to transparent objects applies here also. It is possible to construct microscopes of this class in which both beams traverse the same objective. An instrument due to Dyson (9) (Figure 7) uses a small reference flat separated axially from the specimen, with a splitting plane midway between them. The Objective Small silvered spot •Half s.ilvered surface r777r,'rr7T7 Fig. 7. Dyson's interference microscope for opaque objects. 416 LNSTKUMENT CLASSIFICATION AND Al'l»LICATIO.\S image is transferred with a small magnifica- tion to the image plane of a conventional objective by means of a mirror system. This instrument can be used with the highest powers, but application to low powers is not easy without restricting the field of view. An instrument due to Krug and Lau (10) (Figure 8) uses an oblique splitting surface below the objective with a reference surface out to one side. This is suitable for low powers and, in fact, cannot be used with high powers because of the large space re- quired by the oblique splitting surface. Several low-power instruments have been described which form fringes between the specimen surface and a partially reflecting plate pressed into contact with it. By suit- able selection of the reflectivity multiple- beam fringes can be obtained, giving an enhancement of sensitivity, but in such instruments control of the fringe position is diflacult. The instruments so far described in this class are optically satisfactory, but the fact that the mechanical connection between specimen and reference surface is via the microscope limb and slides makes the direct measurement of path difference somewhat difficult because of thermal drift and vibra- tion. (The magnetic adjustment incorpo- rated in the instrument shown in Figure 7 helps a little in this respect.) For this reason, fringe measurements are better made from a photograph. In many cases met with in engineering the sensitivity of the interference method is too high; thus, an ordinarily machined surface gives a confused mass of fringes which is difficult to interpret. A method was devised by Zehender (11) to deal with this situation A replica of the surface is made in a trans- parent plastic, which is then examined by transmission in a low-power interference microscope. A suitable instrument is that described by Dyson (12), in which a modified Mach-Zehnder interferometer is built into a microscope stage and very low powers are Holf- silvened Sunfoce Fig. 8. Interference microscope of Krug and Lau for opaque objects. used. By immersing the replica in a liciuid of suitable refractive index, the sensitivity can be lowered as far as may be desired. Thus, using polymethyl methacrylate as the replica material and water as the liquid, each fringe corresponds to a height difference of 0.00013" which is a suitable value for engineering purposes. Instruments of Class II. An instrument of this class, using an out-of-focus image of the specimen as the reference field, would be very promising, as it would be free from the mechanical troubles mentioned above. How- ever, reflecting objects usually introduce large phase differences over their whole sur- face and this would have an effect similar to that of placing a piece of ground glass in the reference beam at a point distant from the focal plane, thereby very seriously re- ducing the fringe contrast. It is for this reason that instruments of Class II are conspicuous by their absence. Instruments of Class III. Because of the fact that in this class of instrument the path difference measured is that between two points of the specimen separated by the dis- tance of shear, the applications of such in- struments are limited to the examination of small areas of imperfection on otherwise regular surfaces. An important class of such observations is that of the determination of the thickness of thin films by measurement of the height of the step formed at the edge of sach a film laid on a reflecting surface. 417 INTERFERENCE MICROSCOPY For this application low powers are desir- object, such as a cell, by measurement of its able for two reasons. First, it is sometimes area and optical thickness. not easy to make the film of constant thick- Assume that the cell is a parallel-sided ness right up to its edge, so the distance of disc of thickness t and area A , its refractive shear is required to be larger than the width index being Mc , immersed in water of re- of the imperfection ; the required large shears fractive index /x„ . Its optical thickness is are more easily obtained in low-power in- then struments. Secondly, parasitic effects, such _ , _ . as aberrations of the rear focal planes of objectives, are usually less pronounced in The cjuantity Mc — Atir is equal to 100 ac, low-powered optics so these give mterference where c is the concentration of dry substance effects of higher contrast. in cell and a is the quantity known as spe- Polarizing instruments such as the shear- cific refraction. The total weight M of dry ing microscopes of Smith and Frangon de- material in the cell is equal to cAt, and from scribed above can be modified in an obvious equation (3) can be given as manner for opaque specimens and are par- m = 4 /inn ticularly suitable for this type of measure- ment. By suppressing the image altogether The value of a is remarkably constant for and virtually converting the microscope into the types of substance which are found in a polarimeter, Dyson (13) has shown that cells, and so from equation (4) a good ap- it is possible to make settings with an error proximation to the dry weight can be ob- not greater than one Angstrom unit of tained in spite of the fact that the cell con- thickness, tents constitute a heterogeneous mixture. It is evident from this review that the As the cell has not in fact the elementary situation is less satisfactory for opaque ob- shape suggested above it is necessary to re- jects than for transparent objects. No in- place Ap in equation (4) by an integral of strument is available for opaque objects path difference over the cell surface. This which allows the use of a full range of powers integral can be evaluated by a number of onany type of object and also makes possible methods. The path difference can be meas- the precise measurement of path difference, ured at a number of equally spaced points, using a graticule in the eyepiece, or a photo- Applications of Interference Microscopy graph can be taken. By setting the back- Biological Applications. The inter- ^^'^^"^^ P^^^ difference approximately mid- ference method is useful for qualitative ob- '^^^ between the values for maximum and servations in biology because of the freedom "i^^i^^^"^^ brightness, the mtensity can be from optical artefacts which it gives. Thus, a TJ^^ very nearly a hnear function of path oi^^r • +• £ 4.- 1 i.1,- 1 dmerence for values of the latter small slow variation of optical thickness across a . , , ,, , , , ,•«. 1, 1, ,1, .• 1 , against a wavelength; hence the path difier- cell would probably pass unnoticed when u r i i_ i -x ^^ xi , 111 , , ■. . ^n(?6 can be found by densitometry on the observed by phase contrast methods, but is . • •,, • , j.- r ,. , . , ' negative, with appropriate corrections for immediately evident by interference con- ^^^ ^^^^^^^^^ characteristics. Methods are also available, due to Davies A great impetus to the use of interference and Deeley (16) and Mitchison et al. (17), microscopy was given, however, by Davies whereby the dry mass can be evaluated im- and Wilkins (14) and Barer (15), who dis- mediately by photo-electric means at the covered independently the possibility of microscope without the necessity of taking measuring the dry weight of a biological photographs. 418 INSTRUMENT CLASSIFICATION AND APPLICATIONS All excellent and exhaustive account of this subject with an extensive bibliography has been published by Davies (8). It is also treated extensively by Hale (18). Both these accounts discuss the errors which may be encountered in this work. In biological applications it is usually necessary to make use of the highest possible powers, so only instruments which allow this can be used. In particular, instruments in which slit illumination is used are of little value. The nature of the object may also have a bearing on the type of microscope to be used. If the object is in the form of small isolated patches, such as a widely scattered distribu- tion of cells or groups of cells, each small as compared with the diameter of the field, a shearing type of instrument can be used, such as that described by Smith (6). If, on the other hand, the object is of the same size as the field or somewhat larger, the Dyson (1) microscope may be more useful. If the object is not too thick, interference contrast can still be obtained even if the object occupies the whole cross-section of the reference beam. Under these conditions it is still pos- sible to make meaningful measurements of the path difference. The reason for this is that the object in the reference beam is so far out of focus that the effective background path difference does not vary appreciably over the limited area of the cell being meas- ured. Applications to Opaque Specimens The literature on the use of interference microscopy to opaciue specimens is much less extensive than that covering biological applications. However, it is being used to an increasing extent for the measurement of surface finish in engineering (12) and to the control of the groove profile of diffraction gratings. For opaque specimens the interference method has considerable advantages over the phase contrast technique. It is funda- mental in phase contrast that the surface of the specimen acts as part of the optical train which images the illuminating annulus on to the phase ring, and it obviously will be quite unsuited for this purpose if the surface introduces large path differences, as for example in the case of a machined sui'face. Even if the surface is relatively smooth, a spherical or cylindrical object will introduce obvious difficulties. In addition, in view of the large phase differences commonly observed in reflecting specimens, the image seen by phase contrast may bear very little resemblance in detail to the actual surface topography. An ex- ample of this (a diffraction grating) is shown in Reference 9. Interference microscopes of Class I are immune to both these difficulties. An investigation has been made by Tol- mon and Wood (19) of an error which may arise in any high-power interference micro- scope, but Avhich is most often met with in connection with opaque specimens. This concerns the relationship between the fringe spacing and the angle of slope of the surface; they show that errors of the order of 10% may arise if the angle of obliquity of the reflected beam is not taken into account. A discussion of the micro-interferometry of opaque objects with a description of several instruments is given by Perry (20). REFERENCES 1. Dyson, J., Proc. Roy. Soc, (London), A204, 170-187, 1950. 2. LiNNiK, W., see Kinder, W., Zeiss Nachr., August, 1937. 3. Smith, F. H., British Patent No. 639,014, June '21, 1950. 4. Dtson, J., Nature, 171, 743 (1953). 5. Lebedev, A. A., Rev. Opt., 9, 385 (1930). 6. Smith, F. H., Research, 8, 385-395 (1955). 7. Davies, H. G., in "General Cytochemical Methods," Ed. J. F. Danielle, New York, pp. 55-161, Academic Press, 1958. 8. FRANgoN, M., "Le Microscope a Contraste de Phase et le Microscope Interf^rentiel." Paris: Editions du Centre National de la Recherche Scientifique, p. 113, 1954. 419 INTEKIERENCE >IICKOSCOPY 9. Dysox, J., Proc. Roy. Soc. (London), A216, 493-501 (1952). 10. Krvg, W. and Lau, E., Ann. Phys., 6 (8), 329 (1951). 11. Zehender, E., see Kohaut, A., Werkst. v. Betr., 86 (12), 725-732 (1953). 12. Dyson, J., Engineering, 179, 274-276 (1955). 13. Dyson, J., Phijsica, 24, 532-537 (1958). 14. Davies, H. G. and Wilkins, M. H. F., Nature, 169, 541 (1952). 15. Barer, R., Nature, 169, 366 (1952). 16. Davies, H. G., and Deeley, E. M., Exp. Cell Res., 11, 169 (1956). 17. Mitchison, J. M., Passano, L. M., and Smith, F. H., Quart. J. Micr. Sci., 97, 287 (1956). 18. Hale, A. J., "The Interference Microscope in Biological Research," E. and S. Livingstone Ltd., Edinburgh and London, 1958. 19. Tolmon, F. R. and Wood, J. G., J. Sci. Instr., 33, 236-238 (1956). 20. Perry, J. W., Research, 8, 255-261 (1955). J. Dyson PLASTICS. See GENERAL MICROSCOPY, p. 390. PULP AND PAPER. See GENERAL MICROS- COPY, p. 394. THEORY AND TECHNIQUES Although interference of Hght waves has long been used for high precision measure- ments in physics and technology, only recently has it been applied to any large de- gree in microscopy. The invention and wide- spread success of the phase microscope em- phasized the point that much could be gained by application of the principles of physical optics. One result has been the development of a number of interference microscopes which allow three dimensional study of ob- jects, both qualitatively and quantitatively. Opaque objects can be examined by inter- ference microscopes employing reflected light. Transparent objects, normally in- visible in an ordinary microscope, can be studied by transmitted light. As in the phase microscope, this permits living material to be examined without staining or other special preparation. Both the phase and interference micro- scopes function by causing light waves to interfere. In each the waves are first split apart so that they follow different physical paths; they are then treated differently; and finally they are brought together to inter- fere. With the phase system irregularities in the object itself cause some of the light to deviate from its original direction. These de- viated rays are then treated differently by the phase plate than are the undeviated rays. Thus the resulting image is character- istic of irregularities and discontinuities in the specimen. In the interference microscope the light ray separation is accomplished by means of a beam-splitter. The object modifies one beam with respect to the other by retarding or advancing it. The image which results when the beams are recombined is therefore characteristic of the light-retarding proper- ties of the object. Aii area of uniform optical path appears with a uniform brightness, or if white light is used the color is constant. The really distinctive feature of the inter- ference microscope is that it allows measure- ments of the optical path to be made. From these measurements information on the thickness of objects, the index of refraction of solids and liquids, the concentration of protein solutions, and the mass of cells can be deduced. Interference of Light The constructive and destructive inter- ference of waves, illustrated in Fig. 1, follows certain fundamental laws. First of all, the two beams must be coherent, which in practice means that they must have come from the same source. Second, they must have the same wavelength. Third, if the beams are plane polarized, they will not interfere when the planes of polarization are mutually perpendicular, but only when they are brought to the same plane. The concept of a wavefront is very useful. For the common plane wave, the wavefront 420 niKOKY AM) TECH.MQIES Constructive Interference (Phase difference: zero) Wave Destructive Interference (Phase difference half wavelength) Wave Wave 2 Resultant General Interference Wave I Wove 2 Resultant (zero) "\ T \ / N Resultant Fig. 1. Examples of interference of two waves of equal amplitude. In constructive interference, where the waves are in phase, the amplitudes add at each point. The resultant amplitude is twice the original, as shown on the third line. In destructive interference, the phase difference is one-half wave- length, and the algebraic sum of the two waves is zero at each point. In the general case, where the phase difference is neither zero nor a half -wavelength the resultant can also be obtained by adding al- gebraically the amplitudes of the two waves at each point. is a plane perpendicular to the direction of propagation, over which the phase is every- where the same. For a wave diverging from a point source or converging toward a point image the surface of constant phase is spheri- cal. The optical path between two points is the product of the physical distance and the index of refraction of the medium. If there are several different media along the light path the total optical path is the sum of the optical paths through each medium, meas- ured along the path followed by the ray. In interferometry, where two beams are separated and later recombined, the optical path difference, or OPD, is just the difference in total optical paths encountered by the two beams while they were separated. Several types of interference microscopes use polarized light but in a manner somewhat different than in the polarizing microscope. In the latter, one beam of plane polarized light is incident on the specimen. On enter- ing a birefringent specimen, plane polarized light is broken into two components, which are in phase with each other as they enter the specimen (Figs. 2a and 2b). If the optic axis of the specimen lies in the plane perpen- dicular to the original direction of the light one of the beams consists of vibration paral- lel to, the other perpendicular to, the optic axis of the specimen. These two beams travel through the birefringent specimen at differ- ent velocities, so that they emerge out of phase (2a). In general their resultant is no longer plane polarized light, but it is ellip- tically polarized (1). This light cannot be extinguished by an analyzer. Therefore in the polarizing microscope, such a specimen appears bright in the field. In interference microscopes which utilize polarized light, relatively thick plates of bi- refringent material are used to separate physically the two beams, as shown in Fig. 421 INTERFERENCE MICROSCOPY oa I ^H4H fH-H-l + Birefringent Plate Component parallel to~^' optic axis Original vibration Component perpendicular to optic axis H t < M o.a\ Biretnngent 'Plate Fig. 2. Action of birefringent plates on polarized light, (a) Optic axis parallel to face of plate. The incident polarized light has a component vibrating parallel to the plane of the paper, indicated by the dashes, and a perpendicular component indicated by the dots. Before entering the birefringent plate they are in phase. On entering the plate the parallel component moves at a higher velocity, so that on emerging the parallel component is ahead of the perpendicular component, (b) The resolution of the incident plane polarized vibration into two components which occurs at the entrance face of the plate, as seen by an observer viewing the oncoming light, (c) Optic axis inclined with respect to the surface. In this case not only is there a phase difference introduced between the two components, but one com- ponent is refracted differently than the other. Crystals of calcite (Iceland spar) found in nature exhibit this effect, which results in a double image of everything viewed through the crystals. 2c. The two beams emerging from such a plate are not capable of interfering even if they are brought into coincidence, because their planes of polarization are mutually per- pendicular. They can be brought to the same plane of polarization by means of an ana- lyzer with its transmission axis at 45° to the plane of vibration of each beam. It can also be done by means of various compensa- tors which permit the precise determination of the phase difference between the two com- ponents. Another birefringent device frequently used is a half -wave plate. As shown in Fig. 3 it retards one component of polarized light by one-half wavelength relative to the other component. The net effect is to rotate the plane of polarization from azimuth — « to The action of a quarter-wave plate is shown in Fig. 3b. If the incident plane- polarized light is oriented at 45° to the optic axis, the two components formed at the first surface are equal in magnitude. When they emerge, 90° out of phase, the resultant vibra- tion is circularly polarized. Principles of Commercially Available Interference IVIicroscopes IVIultiple-beani. The multiple-beam sys- tem can be set up with as little equipment as a standard microscope, a slide and cover glass with partially reflecting coatings, and a pinhole at the first focal plane of the con- denser (2, 3, 4). A typical system (5) is a 0.5 mm diameter pinhole at the focal point of a 25 mm objective. Monochromatic light, such as the green radiation of a mercury arc isolated by a suitable filter, is used to illumi- nate the pinhole. Light is partially trans- 422 THi:OKY AND TECHNIQUES mitted and partially reflected at each metal- lized surface, as shown in Fig. 4a. In areas where the optical path difference between successive transmitted rays is an integral number of wavelengths, construc- tive interference occurs, and a bright fringe is seen in the field. A single fringe therefore traces out a contour line of equal optical path, and the deviation of a fringe as it passes through a specimen can be used to measure the optical path of the specimen. The higher the reflectance of the metal- lized surfaces, the sharper is each fringe (6). For biological work with living specimens, an important requirement is that non-poison- ous metals such as mconel or titanium be used. If the index difference between speci- men and mounting medium is considerable, additional multiple reflections may occur which add complexity to the fringe pattern (7, 8). For the examination of surfaces and steps on flat surfaces the system can be used in reflection, through the use of a vertical illu- minator. If the surface to be tested does not have sufficient reflectance it can be over- coated with silver, which contours the sur- face (9). For highest precision, such as in measuring glass or metal polishing defects or the thickness of evaporated films, fringes of equal chromatic order are often u.sed (10, 11). As shown in Fig. 4b, a white light source is used and the image of the specimen is projected onto the slit of a spectrograph. The continuous spectrum is crossed by dark fringes, and the displacement of a particular fringe as it crosses the specimen is a direct measure of its physical height. Precisions of c D a. H Slow Fig. 3. Effects of wave-plates on plane-polarized light, (a) Half-wave plate, C, rotates the plane of polarization from azimuth - a (at A) to azimuth + a (at E). At B are shown the components of the incident light which are parallel (solid wave) and perpendicular (dashed wave) to the slow axis of the crystal. At D the parallel component has been retarded one half wavelength relative to the perpendicu- lar component. The sum of these two vibrations gives a vibration which lies along the line shown at E. (b) A quarter-wave plate, H, converts linearly polarized light at F to circularly polarized light, shown at J. The two components of incident light, shown at G, are equal and in phase. After passing through the waveplate the components are out of phase by one quarter wavelength (I). As these two waves cross a reference plane, their instantaneous sum can be found by vector addition as shown by the arrows at I. As the wave progresses across the reference plane the tip of the vector traces out a circle, as shown at J. 423 INTERFERENCE MICROSCOPY Slit of Specirograph Totally reflecting mirror To eyepiece Specimen Objective Metallized surfaces I j/V-Semi-reflecfing mirror Wtiite light source Pinhole Reflecting surfaces Specimen * Condenser Monochromatic source a. b. Fig. 4. Multiple-beam interference schemes, (a) The system used for transparent specimens. Al- though the condenser collimates the light from the pinhole, the rays are shown here at an angle so that the interreflections can be displayed, (b) One system used for opaque specimens. Rays incident on the interferometer are shown bj^ solid lines and those reflected toward the spectrograph are shown by dashed lines. Both the angle of the reflected rays and the height of the specimen are greatly exaggerated. image Semi-reflecting surface Monochromatic source Reference surface (Flat) M Surface to be examined Fig. 5. The essential components of the Linnik interference microscope. For best contrast the source should be conjugate to the specimen surface, and the objectives Li and Lj should be interferometrically matched. a few Angstroms can be obtained by this method. Two -beam Interference Microscopes for Examination of Surfaces. The system shown in Fig. 5 was described in 1933 by Linnik (12). It is essentially a Michelson interferometer with a microscope objective in each arm. A normal image of the sm-face, A, mider test is formed by objective, Li . The coherent reference beam returning from the flat reference mirror, M, combines with the object beam to produce interference contrast in the image. Irregularities in the surface under test are thereby revealed and can be measured. More recently a system due originally to Mirau (13) has been modified and developed (1-i, 15). As shown in Fig. 6 it employs only one objective and a relatively small inter- ferometer path, which has certain advan- tages in stability. The vertical illumination mirror R directs the monochromatic light down through the objective. At the top sur- face of plate T the beam is divided by a semi- reflecting coating. The transmitted com- ponent is reflected from the surface A which 424 THEORY AND TECIIMQl ES is under test. The light reflected from T strikes a fully silvered spot, S, which is lo- cated at the same optical distance from T as is the surface A . The two components are mage Fig. 6. The Mirau system for surface examina- tion. R: Partiality reflecting mirror, O: objective, T: glass plate with semi-reflecting surface, *S; fully silvered spot, A: surface under test, X and Y are, respectively, the points at which the beam divides and recombines. recombined at Y and travel through the ob- jective together. These two-beam s3^stems are somewhat easier to use than the multiple-beam meth- ods. However, the two-beam precision is lower, being the order of Ko fniige, or 0.025 micron. Two-beam Interference Micro.scopes for Kxaniinalion of Transparent Ob- jects. The Dyson Interferometer Microscope. The Dyson interferometer microscope (Hi), manufactured by Cooke, Troughton, and Simms, Ltd., utilizes semireflecting coatings for beam-splitters. As shown in Fig. 7 the upper surface of plate A transmits part of the incident beam, which then proceeds through the object. The reflected portion of the incident beam is again reflected by the lower surface of plate A. This reference beam then passes through an area of the slide away from the specimen. Plate B serves to reunite the object and reference beams. The fully silvered spot on the upper surface prevents rays having small values of y from reaching the objective. The reason is that for these rays the reference beam is so close to the object beam that both may pass through the specimen. Slide, specimen and cover glass Spherical surface Fully silvered Immersion liquid -^Ct^^-^ Micrometer screw Condenser Semi-reflec+ing surface Fully reflecting surface Fig. 7. The Dyson interferometer microscope. Plates A and B are very slight wedges, and the wedge directions can be aligned either in the same or opposite directions to give a fringe field or uniform field, respectively. Plate A is movable transversely for measurements. 425 INTERFERENCE MICROSCOPY The spherical mirror together with the With it visual settings arc made by matching upper surface of plate B is essentially a unit- the luminance in the specimen or surround power relay system which increases the work- with that of a reference spot superimposed ing distance of the microscope objective. An in the field. image of the object is formed by this relay The above methods measure only the system below the objective and is magnified fi'action of a wavelength of optical path. If in the usual fashion by the ol)jective and the OPD is greater than one wavelength, or eyepiece. if it is not certain whether the path differ- As can be seen by comparing the solid ence is positive or negative, the use of white and dashed paths, the distance between the light and the fringe field can usually settle reference and object beam depends on the the question. angle, 7, of the incident ray. Thus when a The advantages and disadvantages of this full cone of light is incident from the con- microscope have been discussed by Hale denser, the reference beams are passing (18). through various portions of the slide sur- The AO-Baker Interference Microscope. The rounding the specimen. AO-Baker interference microscope (19, 20) In order to eliminate poor contrast and utilizes a birefringent plate located above the errors due to nonuniformities in the slide condenser as a beam-splitter. Plane polarized and coverglass, immersion fluid is used in light incident on the birefringent plate is the three places shown in Fig. 7. divided into two mutually perpendicular For quantitative measurements, several polarized beams, as described earlier. The methods are available. If plates A and B have extraordinary beam, consisting of vibrations their wedges aligned in opposite directions in the plane of the optic axis, is the object the field is uniform, except for the specimen, beam. As plate A is moved laterally by means of In the shearing system, shown in Fig. 8a, the micrometer screw the field goes through the optic axis is in the plane of the paper but maxima and minima of luminance. The mi- inclined at an angle of about 45° to the sur- crometer dial is turned to obtain minimum faces of the plate. This causes the object luminance in the specimen and then in the beam to be refracted to the left, or sheared, background immediately adjacent to the as shown. The reference beam, the ordinary specimen. The difference in micrometer beam, passes through the plate as ordinary readings is proportional to the optical path light passes through a glass plate, difference between the specimen and the On passing through the half-wave plate surround. If the plates A and B have their both beams have their planes of polarization wedges aligned other than exactly opposite, rotated by 90°. This allows the beams to be the field is crossed by fringes. Their spacing reunited by a second birefringent plate depends on the relatively angle between the placed before the objective. It is identical wedge directions, and they move as plate A in thickness and orientation with the first is adjusted laterally. The micrometer dial plate. As shown in Fig. 8a one beam has is turned until a dark fringe is located first passed tlirough the object while the other in the specimen and then in the background has passed through an adjacent area of the immediately adjacent to the specimen. slide. After being reunited by the second bi- Another method of measuring is to photo- refringent plate these beams travel together graph the field with uniformly spaced fringes, through the objective and to the image. In Photographic photometry can then be used general there is a phase difference between (17). the beams. It represents the optical path A photometer eyepiece is also available, encountered by the object beams minus that 426 THEORY AND TECHNIQUES yepiece Rotafable analyzer Objective Opfic GxiSv^p 1— Quarfer-wave plate Calcite plates — * Specimen Half-wave plate— ^ Calcite plates— I \f -y- — "Condenser Polarizer piaie ~~7 / / ^ \ tates-H f V^ Optic axis b. Mirror Fig. 8. The AO-Baker interference microscope, (a) The shearing system. The transmission axis o the polarizer is set at 45° to the axis of the first calcite plate. The calcite plates are essentially cleavage sections, similar to those found in nature. Each ray is split into two components as shown in Figure 2c. The half-wave plate axis is parallel to the polarizer axis, as is the quarter-wave plate axis. Measurements are made by rotating the analyzer, (b) The double-focus system. The only difference from the shearing sj^stem is that here the axes of the calcite plates are parallel to the faces of the plates. encountered by the reference beam in the region in which they were separated. In the double-focus system shown in Fig. 8b, the optic axis of the birefringent plate A is parallel to the surfaces. For a set of object rays passing through the specimen, the cor- responding reference rays pass through an area above the specimen. As before, the phase difference between the two beams after they are reunited represents the difference in optical path encountered by the two beams. In both systems the method for detecting and measuring this phase difference is shown in the upper part of Fig. 8a. A quarter-wave plate is oriented with its effective optic axis at 45° to the plane of polarization of the ob- ject beam. The two beams are thereby con- verted into circularh^ polarized light, one rotating clockwise, the other counter-clock- wise. The resultant of these two vibrations is simply plane polarized light. If the two beams are in phase, the plane of polarization is parallel to the axis of the quarter-wave plate, which may be called zero azimuth. Otherwise, the azimuth angle of the plane of polarization is just half the phase differ- ence between the object and reference beams.* If the object beam is behind the reference beam, as in the case of a phase re- tarding specimen, the plane of polarization is rotated counter-clockwise from the zero azimuth. The opposite is true in case the surround has a greater optical path than the specimen. If an analyzer is oriented perpendicular to the zero azimuth the background, where the optical paths are equal, will appear dark or black. Objects having greater or less op- * Phase difference is often measured in degrees, 360° corresponding to one wavelength of phase difference. 427 INTEHFE HENCE MICKOSCOPY tical path than the surround will appear bright. If the analyzer is turned until it is perpendicular to the plane of polarization in the image of the specimen, the specimen will have a minimum of luminance. The ori- entation of the analyzer is then read from a scale to determine the optical path in the specimen. As in the other two-beam systems, if white light is used the specimen appears in color contrast. The colors change as the analyzer is turned. This is often useful in studying and photographing complex speci- mens. For measurement of optical path, mono- chromatic light is necessary. The half- and quarter-wave plates are made for the green radiation emitted by a mercury arc. This wavelength, 546 m/x, was chosen partly be- cause it is very near the peak of spectral sensitivity of the eye and partly because convenient, very bright mercury arcs are available and the green radiation can be iso- lated easily by means of gelatine, glass, or interference filters. The human eye has great sensitivity for comparing the luminance of adjacent por- tions of a field of view. In order to take ad- vantage of this sensitivity half-shade eye- pieces have been designed for use with this polarizing type of interference microscope (21, 22). For half -shade eyepieces employ- ing the optimum half-shade angle (23) and used with the 100 X shearing system, ana- lyzer settings reproducible to 0.5° standard deviation can be obtained on specimens with areas of uniform path difference. The object is set so that its image strad- dles the dividing line. As the analyzer is turned the two halves of the specimen image change in luminance relative to each other. At the match setting, the analyzer orienta- tion, ^1 , is noted. Then the analyzer is turned until a match is obtained in the back- ground adjacent to the specimen, and the reading, 02 , is noted. The optical path difference between the specimen and surround is computed accord- ing to the eciuation 9i — Oo , measured by any of the preceding methods, can be used to calculate a number of other quanti- ties. The OPD is given by = tini — 111) a) where t is the thickness of the object, W2 its index, and ni the index of the medium through which the reference beam passed, the mounting medium. Thickness can be determined directly from equation (1) if the index difference is known. If the object is spherical, cylindrical, or some other regular shape so that its thickness can be determined by lateral measurements, then the index difference can be measured from equation (1). If one is free to change and to measure in- dependently the index Ui of the mounting medium without changing that of the speci- men, a second equation can be obtained. ni — ^'rii 4>- 4>' (3) U) Since these equations utilize the differ- ences between measm-ed quantities, care must be taken that the desired precision is not lost. A sensitive method for determining the index of a specimen is to change the index of the mounting medium until the object has no interference contrast, i.e., until 0 = 0. Of major importance for biological studies is the general rule that the index of a solution of protein in water increases hnearly as the concentration and that the specific refractive To image " and eyepiece Beam- splitting prism Parallel plate compensators Object slide Parallel plate ■ ' compensators Objectives Comparison slide Condensers ^-=^^ Step compensator * Wedge compensator c Beam-splitting prism ntermediate lens Fig. 11. The Leitz interference microscope. In this sj'stem the object and reference beams are completely separate and pass through separate condensers and objectives. The horizontal distance between these two complete microscopes is 62 mm. Coarse adjustment of the optical path is provided by the step compensator. This is a rotatable disc on which ten glass plates are arranged so that they can be swung into path of the reference beam. The wedge compensator provides fine adjustment and measurement of the optical path. The motion of the center wedge is measured by a micrometer. increment, a, is essentiallj^ the same for all proteins (30). The latter is defined by the equation ns — riv C (6) where n^ is the index of the solution, n^, the index of water, n^, = 1.33, and C the concen- tration in grams of dry material per 100 ml of solution. For proteins the average value of a is 0.0018. If the specimen of thickness t is mounted in water, the measured OPD is related to the above quantities by (6) Therefore = (ria — nw)t /a = Ct. (7) 431 IINTKKFEKENCE MICROSCOPY The quantity Ct is the dry mass per unit area of the specimen. The total dry mass of cells or iiucloii (;aii therefore be determined by integrating the dry mass per unit area over the area of the specimen. If the thickness, t, can be measured by one of the methods described earlier, then the concentration, C, can be determined from equation (7). Or if the index ris can be measured as described, equation (5) can be used to determine concentration. Precautions and Precision The greatest potential source of random and systematic errors is the observer him- self. In the extinction method it is difficult and treacherous to set the object at mini- mimi when the background is changing lumi- nance simultaneously. The tendency often is to set on maximum contrast rather than on extinction. The half -shade or photometer eyepieces help to overcome this difficulty by changing the task to one of matching. In this method the observer should adjust back and forth through the match position sev- eral times in order to avoid errors due to after-image effects. For the most exacting work it is desirable to have an iris or other diaphragm which can be adjusted to limit the field of view to the specimen. In the discussion of optical path and thick- ness measurements it was assumed for sim- plicity that the rays all passed through the specimen parallel to the optical axis of the microscope. In interference microscopes of the Leitz, Dyson, Linnik, Mirau, and Baker types a cone of rays is incident on the specimen. Oblique rays generally receive a different retardation than normal rays. In transmission microscopes a ray inci- dent at an angle 71 , on a flat plate-like speci- men receives a retardation of 8 = t(n2 cos 72 — Wi cos 7i) relative to the corresponding reference ray (81). Here no and ni are the indices of the sample and of the immersion liquid; 72 and 7i are the respective angles of the rays meas- ured from the optical axis. Averaging this fiuantity over the whole cone of rays gives approximately 5„ = K«2 — «l) 1 + 4nin2 where da is the measured, average path dif- ference, (NA)c is the numerical aperture of the condenser system used, which may be different from that of the objective. Up to (A^^)c = 0.4 this expression is ac- curate to 0.1 % or less and may be used to correct for oblicjuity errors. The equation suggests two ways to reduce the error. A low condenser NA produces less error, but it also reduces resolution and image lumi- nance. Whenever the sample permits, one should use the highest possible index 7ii of immersion medium, thus reducing obliquity error without sacrificing resolution. It is interesting that for objects of index greater than the surround the transmission microscopes give a measured optical path which is slightly too large. Whereas in the reflection interference microscopes it has been found that the readings are slightly too small (32, 33). For precision work it is necessary to pay as much attention to the path taken by the reference beam as to that of the object beam, since measurements involve them both. For the Dyson system methods have been given (18) for assuring that the reference area which is mostly outside the field of view, is sufficiently free from inhomogeneous mate- rial. In the shearing systems (AO-Baker, Frangon, Johanssen and Afzelius) the ref- erence beam passes through an area which is within the field of view, and the slide should be rotated if necessary to make sure this area is clear. As seen in the eyepiece of an AO-Baker shearing microscope the sharp image is at B (Fig. 12). An out-of -focus image is seen at A . This means that the ref- erence beam for area A went through the specimen, which is a distance d to the right 432 TIIIOKY AND TECIIMQl ES of area .4. Hence the reference beam for area B goes through an area C, whicli is a distance d to the right of B. This area C should be kept clear when making measurements in area B. Because of the separation of the object and reference beams, any non-parallelism such as a wedge in the slide, coverglass, or moiniting medium becomes part of each reading. However, since the optical paths are determined by subtracting a reading in the surround from a reading in the object, the contribution of a uniform wedge drops out. In order to be safe the reading in the surround should be close in distance and in time to the reading on the object. To detect such a wedge the slide may be turned, pref- erably by means of a circular stage, and any change in luminance of the background noted. In the systems employing polarized light a birefringent specimen should be aligned with its optic axis parallel to the vibration direction of the polarized light (parallel to the direction of shear). For manj^ biological specimens the birefringence is too small to cause appreciable error. For highly bire- fringent specimens, methods have been de- scribed by Faust (34). It has been shown (22) that with the AO- Baker shearing system there is a condenser aperture which produces the best contrast. The user can determine this condition of best contrast either by qualitative observa- tions or by the use of a photomultiplier. With the above precautions precisions of from one tenth to one three-hundredth of a wavelength can be expected, depending on the system used, the size and uniformit}^ of the specimen, and the care of the observer. Special -Purpose Modifications As mentioned earlier, the total dry mass of cells or nuclei can be determined by inte- grating the measured dry mass per unit area over the area of the specimen. In cases where the measurements must be done c^uickly or Fig. 12. The fiekl of view of an AO-Baker shearing micro.scope. The object has a sharp image at B, a blurred image at A. Area C should be kept free of material which would cast a blurred image on area B. the specimen is very irregular, it may be possible to employ an automatic integration technique (35, 36, 37). In the .system of Mitchison, Passano, and Smith (36) the in- tegration is done opticallj'^, but care must be taken that the maximum optical path differ- ence is less than about X/8. The index of liquids available only in small quantities can be measured by a technicjue developed by Smith and Iverson (38). The liquid is allowed to fill a small Avedge, and the interference bands seen in the image are measured. The optical absorption of specimens can be determined by employing a modification of the AO-Baker system (39). If the object beam passes through an absorbing specimen but the reference beam does not, the emerg- ing beams will be unequal in amplitude as well as ill phase. Polarimetric technicjues, including a sensitive half-shade device and/ or photoelectric comparison, yield both the amplitude ratio and phase difference. Meas- urements at various wavelengths in the visible spectrum ar(^ possible. REFERENCES 1. DiTCHBURN, R. W., "Light," p. 360, Blackie and Son, Ltd., London, 1952. 2. Merton, T., Proc. Roy. Soc. (London), A189, 309 (1947). 3. Mellors, R. C, Kupfer, A., and Hollen- 433 LIGHT (OPTICAL) MICROSCOPY DER, A., Cancer, 6, 372 (1953); 7, 779, 873 22, (1954). 4. Richards, O. W., J. Biol. Photogr. Assoc, 19, 23. 7 (1951). 5. OsTERBERG, H., "Phase and Interference 24. Microscop3%" in G. Oster and A. W. Pol- lister, "Physical Techniques in Biological 25. Research," Vol. 1, p. 378, Academic Press, New York, 1955. 6. Jenkins, F. A., and White, H. E., "Funda- mentals of Optics," 3rd ed., p. 270, McGraw- 26. Hill, New York, 1957. 27. 7. Glauert, a. M., Nature, 168, 861 (1951). 8. Lindberg, p. J., Optica Acta, 4, 59 (1957). 9. ToLANSKY, S., "Multiple-beam Interferom- 28. etryof Surfaces and Films," p. 63, Clarendon Press, Oxford, England, 1948. 10. ibid., p. 96. 29. 11. KoEHLER, W. F., /. Opt. Sac. Avi., 43, 743 (1953); 45, 1011, 1015 (1955). 30. 12. LiNNiK, W., C. R. Acad. Sci. URSS, 1, 18 (1933). 13. Delaunay, G., Rev. optique, 32, 610 (1953). 31. 14. The Microscope, 11, 163 (1957). 15. /. Sci. Instr., 35, 189 (1958). 32. 16. Dyson, J., Proc. Roy. Soc. {London), A204, 170 (1950). 33. 17. Da vies, H. G., Wilkins, M. H. F., Chayen, J., AND LaCour, L. F., Quart. J. Micro. 34. Sci., 95, 271 (1954). 18. Hale, A. J., "The Interference Microscope in 35. Biological Research," E. & S. Livingstone, Ltd., Edinburgh and London, 1958. 36. 19. Lebedeff, a. a., Rev. optique, 9, 385 (1930). 20. Smith, F. H., Research, 8, 385 (1955). The ob- 37. jectives and condensers made for this inter- ^8. ference microscope by C. Baker, Ltd., Lon- don, are used on the AO-Baker microscope of the American Optical Co., Buffalo 15, N. Y. 21. Smith, F. H., Nature, 173, 362 (1954). Koester, C. J., J. Opt. Soc. A7n., 49, 560 (1959). Inoue, S., and Koester, C. J., J. Opt. Soc. Am., 49, 556 (1959). R. C. Faust and H. J. Marrinan, Brit. J. Appl. Phys. 6, 351 (1955). Francon, M., Rev. opt., 31, 65 (1952). Manu- factured by Barbier, Benard, et Turenne, Paris, and Optique et Precision de Levallois, Paris. FRANgoN, M., J. Opt. Soc. Am., 47, 528 (1957). Johansson, L. P., and Afzelius, B. M., Na- ture, 178, 137 (1956). Manufactured by Jungnerbolaget, Stockholm 14, Sweden. Ingelstam, E. & Johansson, L., Optica Acta, 2, 139 (1955). E. Ingelstam, Exp. Cell Res. Suppl. 4, 150 (1957). Grehn, J., Leitz Mitt. Wiss. u. Techn., 1, 35 (1959). Barer, R., Nature, 169, 366 (1952); Da vies, H. G., AND Wilkins, M. F. H., Nature, 169, 541 (1952). Ingelstam, E., and Johansson, L. P., J. Sci. Instr., 35, 15 (1958). TOLMON, F. R., AND WooD, J. G., J. Sci. Instr., 33, 2dG (1956). Ingelstam, E., Johansson, L. P., and Bruce, C. F., J. Sci. Instr., 36, 246 (1959). Faust, R. C, Quart. J. Micro. Sci., 97, 569 (1956). Davies, H. G., and Deely, E. M., Exp. Cell Res., 11, 169 (1956). MiTCHisoN, J. M., Passano, L. M., and Smith, F. H., Quart. J. Micro. Sci., 97, 287 (1956). SvENSON, G., Exp. Cell Res., 12, 406 (1957). Smith, F. H., and Iverson, S., Quart. J. Micro. Sci., 98, 151 (1957). Koester, C. J., Osterberg, H., and Will- man, H. E., Jr., /. Opt. Soc. Am., 50, 477 (1960). C. J. Koester Light (optical) microscopy COMPARISON MICROSCOPES. See ENGINEER- ING MICROSCOPES, p. 438. DESIGN AND CONSTRUCTION OF THE LIGHT MICROSCOPE Base and Illuminator. The traditional horseshoe base with mirror has generally been supplanted by a base in the form of a hollow casting enclosing some form of built-in illuminator. Mirrors for use with ex- ternal sources or natural lighting are avail- able optionally, but the trend is toward the convenience and ease of operation resulting 434 DESIGN AND CONSTRUCTION OF THE LIGHT MICROSCOPE from the permanently correct alignment of a built-in source of illumination (See Fig. 1). In some instances, instead of building the source into the base, an attachable source is made available which is interchangeable with the mirror. Such sources are generally of somewhat lower intensity than the built-in sources, sufficient for most visual (bright- field) tasks, but inadequate for photomicrog- raphy, or special requirements, such as dark- field or phase contrast. They are less expen- sive than the built-in source and easier to use than the mirror, since alignment is fixed. The built-in source, on the other hand, is generally a very brilliant source, suitable for photomicrography, darkfield, and phase. A low wattage compact coil filament lamp is employed using the traditional Koehler illumination system, in which the lamp condenser is imaged in the field, and the coiled filament in the aperture of the system. Filters and variable voltage taps on the transformer provide means of controlling the intensity of the illumination. The Arm. The traditional arm with in- clination joint for tilting the microscope has in large measure been supplanted by a rigid non-tilting arm, which indeed is frequently an integral part of the base casting. The arm is normally hollow, enclosing the focusing mechanisms. At its upper end, it is fre- quently made in the form of a ring support permitting mounting of interchangeable microscope bodies, and rotation of the bodies to any desired orientation. Bodies. Tilting the microscope on an inclination joint has generally been sup- planted by the use of inclined body tubes, giving the same comfort of viewing, but without the attendant undesirable tilting of the stage. The trend is toward inclined binocular bodies with their more natural fatigue-free type of viewing. Some manufacturers supply binocular bodies, which maintain a constant tube-length despite changes in the inter- pupillary setting. The advantage of this is that it keeps the objectives working accur- FiG. 1 . This microscope depicts a good many of the trends in the modern microscope, such as built-in illumination, inclined binocular body, built-in means for photography, and low position coarse and fine adjustment. (Photo courtesy Ameri- can Optical Co.) ately under nominal tubelength, and rigidly maintains the factory setting for objective parfocality. The so-called 'Triocular' body (Fig. 1) is also becoming more popular. This is an in- clined binocular with a third (vertical) eye- piece tube for photomicrography. These con- tain either a removable prism for directing the light to either visual or photographic systems, or a beam-divider, which directs a portion of the light to each system. Built-in Cameras. Small built-in cam- eras are finding increased favor over the larger external photomicrographic cameras. 435 LIGHT (OPTICAL) MICROSCOPY These are normally 35 mm cameias, with their own eyepiece system built-in. Like the trend toward built-in illumination, they rep- resent the outcome of a demand for quick convenient photography, resulting from the permanently correct alignment built-in at the factory. Focusing the picture is done either via the binocular viewing system, or by means of a small focusing viewer system. Coarse Adjustment. The ancient and honorable rack and pinion persists as the most popular means of coarse focusing. One major difference in the modern microscope is that many stands now focus the stage rather than the body tube. The position of the coarse adjustment has accordingly been lowered to a more convenient height, per- mitting operating it with the hand resting on the table surface. In some stands the coarse and fine focus controls are made con- centric for greater ease, and rapidity of op- eration. Fine Adjustment. Like the coarse ad- justment, the fine adjustment is almost uni- versally located in a low position permitting use with the hand resting on the table sur- face. On many stands it is now made con- centric with the coarse adjustment. Unlike the coarse adjustment, the fine adjustment has undergone many changes through the years, and no consistent design pattern exists among various manufacturers. Most microscope makers have, however, adopted ball bearings not only for the fo- cusing slide, but for the thrust bearings on the rotary control. Another trend, is to the use of mechanisms, which permit using a single focusing slide for both coarse and fine focus, rather than the older style of a slide carrying a slide. The basic mechanism which produces the fine focusing motion is generally one of the following: 1 . Screw and nut with reducing lever. 2. Screw and nut with reducing cam. 3. Screw and nut with worm wheel. 4. Cam and lever, friction clutch on pin- ion. 5. Planetary ball-bearing drive. 6. Face cam with limits which engage coarse adjustments. None of these has established a clear-cut superiority over the others, and accordingly, there seems to be no trend toward conform- ity among microscope manufacturers in this design feature. Stages. As has already been mentioned, the stage in the modern microscope has been made the focusable member, supplanting the focusable body tube, and making feasible the low position concentric coarse and fine adjustment. Mechanical stages for accurately moving and locating the specimen are supplied either as accessory attachments or as built-in de- vices. Li some cases both the 'east-west' and 'north-south' motion are accomplished by sliding the specimen slide across the plain stage surface. In others, this is true for only one of these motions, the other being ac- complished by moving the entire upper stage surface. Another means of moving the specimen is offered in the so-called 'GHde Stage', in W'hich the entire upper stage surface is mov- able by finger pressure in any direction. The plate slides on a film of grease, and can be moved with quite good precision as desired. Substages. The simplest substage equip- ment, used on student microscope, consists of a rotatable metal disc containing a num- ber of different apertures, used to control the N.A. of the illuminating beam. An iris diaphragm supplants this disc in a slightly better student model. Another simple form of substage contains a 2-lens abbe condenser with iris diaphragm, held either in a fixed focus mount, or a helical focusing mount, in which focusing is done by rotating the condenser. Once beyond the student level, however, most microscope substages employ rack and pinion focusing for the condenser. The con- denser is normally clamped in a sleeve mount, usually factory precentered. An iris 436 ENGINEERING MICROSCOPES ENGINEERING MICROSCOPES diaphragm is always available for control of good centration of the optics, although eye- the illuminated N.A. piece centration is still much less critical Nosepieces. The rotating nosepiece, used than objective centration. to interchange objectives, is a remarkably precise mechanism. Repeatability of cen- James K. ±5enford tering is commonly held to within 0.00005", a value attained chiefly by a precisely fitted bearing and a sturdy and well designed click- stop. The bearing may be either a central There are many applications of standard cone bearing, or a peripheral ball bearing, and specialized microscopes to the field of The click stop may be either a V-shaped engineering in production and testing of spring clicking onto a peripheral ball stop, or manufactured articles. A few of the impor- a spring-loaded rotatable ball clicking into a tant examples are cited to show^ the great notch on the rim of the nosepiece. In either importance of microscopy in engineering, case, the spring is a very flat leaf spring, Surface Examination. Perhaps the most having good rigidity in its wide dimension direct application is the examination of sur- to hold centering precisely, but yielding faces particularly of metallic specimens, for smoothly in a radial direction to give just the detection of flaws and cracks and for the correct force to click the stop mecha- the distinction of surface finish. When a nism home. crack is of the order of 0.0001 in. in width, Objectives. Objectives are almost uni- or when a sm-face structure is equivalent to versally threaded in conformance with the a series of rulings of the order of 10,000 to Royal Microscopical Society standard, so the inch or less, a microscope is required. A that objectives of one manufacturer may be conventional microscope discloses the extent interchanged with those of another in a or width of markings in a plane view, but nosepiece. The virtue of this standardization depth measurement requires a special type is somewhat lessened by the fact that the of microscope, the Schmaltz Profile Micro- objectives are not standardized for shoulder scope discussed in a later section. Require- to specimen distance, hence the convenience ments are adequate illumination of the sam- of parfocality is lost when intermixed objec- pie, preferably from a vertical illuminator; tives are used on a multiple nosepiece. and adequate resolving power (R.P.) of the The lens elements are commonly bur- optical system (see Optical Theory of Light nished into individual cells, which in turn fit Microscope). This R.P. or the minimum dis- into a common bore in the objective barrel, tance between tw^o points which appear as The fit of cells into the bore and the centra- two separate images is given by R.P. = tion of lens elements in their cells is the most 0.61X/NA, when X is the w^ave length of light difficult part of making a good objective, the used and NA is the numerical aperture, tolerances being unbehevably tight. w^hich is a measure of the maximum cone of Eyepieces are not greatly different in the light which the microscope objective can modern microscope than in the microscope of take in and refract to the observer's eye. For 50 years ago. The simple 2-lens Huygens the majority of cases of examination of ma- construction has stood the test of time and chined surfaces the microscope is suitably is still the most widely used eyepiece in mi- fitted with a % m. objective of NA 0.28 croscopy. Probably the most significant and an eyepiece of powder 20 X. The labora- change in eyepieces is the requirement for tory microscope is suitable when specimens better centering today because of the in- are not too large, but in other cases special creasing percentage of binocular models. To workshop microscopes of rugged design are get proper results binocularly requires quite available. 437 I.ICIIT (OPTICAL) MICROSCOPY Hardness Tests. A special case of surface power of the photographic emulsion instead examination is the determination of the di- of the acuity of the eye. This of course is a ameter of the depression formed in the test special case of photomicrography (q.v.) piece during a hardness test of the Brinell Schmaltz Profile Microscope. It has type. The depression is illuminated by a been shown that in surface examinations the vertical illuminator and the magnified image conventional microscope is not able to make formed in the focal plane of the eyepiece can possible depth measurements. If a normal be measured in two ways. The simplest is by section of the specimen could be cut the comparison with a glass scale mounted in depth could of course be measured directly the focal plane of the eyepiece upon which as a simple lateral measurement. In the Zeiss the image of the indentation is focused and Schmaltz Microscope this section cutting is the diameter measured, as though the scale effected by optical means. A line of light is were an ordinary rule, down to 0.01 mm. focused on to the surface by means of an The magnified image may be measured more illuminating unit inclined at an angle of 45° accurately with a micrometer eyepiece. Two to the surface inspected. This line of inter- wires at right angles are mounted on a slide section is then viewed with a microscope also in the focal plane of the eyepiece, and the inclined 45° to the surface and 90° to the slide is moved by a micrometer screw from illuminator. The appearance in the eyepiece one side to the other of the indentation consists of a band of light rumiing across a image. It is easily possible to measure within black background. The microscope is fo- 0.001 in. on the magnified image which with cnsed on one edge of this band, and the a magnification of 10 X means 0.0001 in. in image thus forms a magnified contour of the the actual depression. Excellent equipment surface. If the magnification is M, then the is available in which the microscope is built pitch or frequency of variations along the into the hardness testing machine. band is magnified M times, and the departure Comparison Microscope. In the estima- from straight lines, representing depths, 3=2 tion of surface finish by comparison with M times. Measurements are made with a mi- standards, the conventional microscope suf- crometer eyepiece. The instrument is suit- fers from the disadvantage that only one able for measurement of surface finish of specimen can be viewed at a time and ap- turned, bored and ground surfaces in which pearances have to be memorized. It is not the scratches all run in one direction, practicable to arrange specimen and stand- The Introscope. The examination of the ard side by side for simultaneous viewing, so interior surfaces of long tubes such as rifle that special comparison microscopes have bores and gun barrels is a problem for which been designed, the objects being well sepa- the conventional microscope is unsuited. The rated while the images are formed in adja- Introscope design for this purpose consists cent positions in the eyepiece. This is done of a low-power microscope with built-in il- with prisms which form semicircular images luminator, and a special copying system of on each half of the eyepiece. lenses which enables the objective at some Projection Microscope. It is possible to adjustable distance down the tube to be well make comparisons of specimen and standard separated from the eyepiece at the end of by projection in which a screen takes the tube. The objective forms an image of /i of form of a photographic plate. The two ob- part S of the surface under examination; jects are not viewed simultaneously but a light from the lamp scattered from S reaches photographic record is made of each sepa- the lens after reflection by a prism. Copying rately and thus the choice of the eyepiece units consisting of pairs of collimating lenses magnification is determined by the resolving successively transfer the image to I2 , h , h 438 ENGINEERING MICROSCOPES and finally to h , where it is viewed by the file grinding machines designed to finish- eyepiece, grind regular and irregular forms of all types. Stereoscopic IVIicrosccpe. Since the eyes Thus in one type of grinder a drawing of the are separated on the average by 64 mm, required form is made 50 times size, and the each sees an object from a different view- outline is followed with the stylus of a pan- point, and the images formed on the respec- tograph. A microscope is fitted to the short tive retinas are correspondingly different, arm of the pantograph w^hich reduces 50/1, By the stereoscopic process the brain com- and thus the microscope cross-lines trace out bines the two images into one composite pic- the required form in the plane of the work, turc with an accompanying sense of solidity An accuracy of at least 0.0004 in. or better or depth. So the stereoscopic microscope ob- is possible. Manufacturers can provide de- jectives form two images and each eye views tails on a wide range of screw and profile one with an eyepiece. A prism system per- projectors and travelling microscopes for the mits correct adjustments of the stereoscopic toolmaker, workshop microscopes for inspec- image. For high-power microscopes two ob- tion processes, and many closely related op- jectives are not practicable so that two view- tical instruments employing principles of the points are secured by using the light re- optical lever and interference effects (for ex- fracted by }'2 of the objective to form the ample, for the evaluation of surface flatness), image for one eye, and light from the other Heating Microscope. Of both engineer- half to form the other image, the division ing and chemical interest are microscopes being accomplished by a prism close behind which provide necessary information con- the objective. cerning the behavior of materials at elevated jNIeasuring Microscope. For various temperatures. In some types hot stages of measurements in machine shops for example, various designs are provided directly on the requiring the highest attainable accuracy, a microscope, especially for fusion analysis of considerable number of microscopes of vari- crystals at relative low temperatures and ous types are available which utilize glass similar observations. In other cases the heat- scales or micrometer eyepieces. The Zeiss ing microscope consists of 3 principal units Universal Measuring Microscope and Verti- mounted on an optical bench comprising (1) cal Comparator employ glass scales for the a light source; (2) electric furnace with measurement of length. A scale is viewed specimen carriage; and (3) observation and with the microscope and subdivision is ef- photo-microscope. fected bj' means of a spiral micrometer eye- Fuel ashes, slags, ceramics, glazes and the piece. The master scale is graduated in 0.05 like are more or less complex mixtures of the in. divisions; a fixed glass scale or graticule most diverse inorganic compounds. Being in the microscope divides the image of one non-homogeneous substances, they do not division into tenths or 0.005 in. Thus small have clearly defined melting points, but divisions are further divided into hun- fuse and soften as they are heated and melt dredths, or 0.00005 in. by means of a special over a more or less wade temperature range graticule graduated on a separate glass disk depending upon their chemical composition and rotated around a center to one side of before finally reaching the liquid state. In the microscope axis. The microscope on a order to establish the typical softening and measuring machine as an optical vernier is melting behaviors small samples of the ma- also an optical pointer, which may take the terial to be examined are pressed into speci- form of a bent microscope as in the case of mens of specific shapes and the characteristic that used as jig borers, tool-room millers and physical changes determined as the specimen other precision machines, as well as in pro- is heated. 439 LIGHT (OPTICAL) MICROSCOPY Fig. 1. Microscopic silhouettes of ash sample at increasing temperatures. With the aid of the heating microscope the individual phases of the process may be ob- served and recorded photomicrographically. The image of the specimen located in the electric furnace is magnified by the micro- scope and projected onto a ground glass screen, where it is observed. The characteris- tic phases are easily photographed. Thus the entire cycle can be observed, the series of photomicrographs giving an impression of the behavior pattern of the material being investigated. By evaluating the volume changes of the specimen a melting curve in relation to temperature is easily plotted. One of the principal advantages of this method of investigation is that the specimen at no time is subjected to pressure by a prob- ing rod. Thus, since no load is applied to the specimen, not only the softening process, but also any desired phase of the alteration in shape of the specimen can be investigated and recorded, such as the swelling of such material as ashes as well as the "melting point," the so-called hemisphere point. In Figure 1, a series of photomicrographs made with a Leitz instrument showing sil- houettes of the specimen, the individual characteristic alterations in shape necessary for judging the thermal behavior pattern of fuel ashes are clearly seen. The original shape of the specimen is shown in the first photo- micrograph of the series. The next ones show that sintering occurs between 960° and 1160° C; the fact that both diameter and height of the specimen are reduced shows that sintering and not softening has taken place. At 1180° C the outlines of the silhou- ette have changed noticeably, indicating that the ashes of this stage have passed into the softening phase. The softening temperature, i.e. the temperature at which softening com- mences is thus found to be between 1160° and 1180° C. After swelling slightly at 1210° C. the specimen eventually melts. The photomicrograph made at 1270° C. shows the "hemisphere point," the officially accepted melting point. At 1430° C the ashes begin to liquefy and at 1470° C they have flowed uniformly to all sides. (From a brochure published by Ernst Leitz, Wetzlar, Ger- many.) REFERENCES 1. Habell, K. J., AND Cox, Arthur, "Engineer- ing Optics," Pitman and Sons, Ltd., London, 1956, pp. 411. G. L. Clark FIBERS (TEXTILE). See GENERAL MICROSCOPY, p. 343. HARDNESS TESTS. See ENGINEERING MI- CROSCOPES, p. 438. HEATING MICROSCOPES. See ENGINEERING MICROSCOPES, p. 439. INDUSTRIAL RESEARCH, APPLICATION TO. See GENERAL MICROSCOPY, p. 363. INTROSCOPE. See ENGINEERING MICRO- SCOPES, p. 438. MAGNETOGRAPHY: THE MICROSCOPY OF MAGNETISM* Although the phenomena of magnetism have been known for many centuries and * Reprinted by permission from the Bell Lab- oratories Record, 35, No. 5, pp. 175-8, 1957. Photo- graphs courtesy of Bell Telephone Laboratories. 440 MAGNETOGRAPHY Fig. 1. Iron filings oriented along the magnetic lines of force around and between the poles of a small horseshoe magnet placed under surface shown. their forces applied to practical problems with tremendous advantage, they have de- fied study by ordinary visual methods. One of the early methods of observing magnetic effects was to sprinkle iron filings above a magnet and allow the filings to orient them- selves along the paths of magnetic force. As illustrated in Figure 1, the filings are aligned with the greatest concentration near the poles. It was with this demonstration and the simple compass test that every student of elementary physics encountered the law — -"like poles oppose and unlike attract." This method of study, although interesting from the gross viewpoint of the external forces, gave no light on the internal forces that must be present and in equilibrium. The first relatively successful attempts to- ward "seeing" magnetization, or visually ap- praising the forces that occur in the indi- vidual crystal of a ferromagnetic material, were not made until recent years. In 1931- 32, Professor Francis Bitter of the Massa- chusetts Institute of Technology described a method by which it was possible to see the separate so-called "magnetic domains" in a specimen. When a piece of ferromagnetic material is demagnetized, a nearby compass needle will be unaffected, and iron filings sprinkled on the sample will not indicate the presence of magnetic fields. When the sam- ple is examined microscopically, however, it is found to consist of small regions, each of which is magnetized to saturation in a cer- tain direction. These regions are called "mag- netic domains". In using Bitter's method, which was later used and greatly improved by W. C. Elmore of Swarthmore College, and L. W. McKee- han, formerly of Bell Telephone Labora- tories, a suspension of colloidal magnetite is employed. This is an extremely fine iron ox- ide in a soap solution. When a drop of the colloidal suspension is placed on a freshly polished magnetic surface and covered with a thin glass disc, the magnetic particles are attracted by the strong flux at the domain boundaries. A metallurgical microscope which magnifies about 100-200 diameters is used to see the outlined domains, and the specimen is studied with either "brightfield" or "darkfield" illumination. In the brightfield system, the light rays from the source are directed through the microscope objective lens to illuminate the specimen. On return, the reflected light from the specimen passes through the objective lens, thence to the eyepiece to form the final image (Figure 2). In the darkfield system, I REFLECTED PAYS I TO EYEPIECE ^ Min LIGHT RAYS FROM SOURCE TRANSPARENT MIRROR IN VERTICAL ILLUMINATOR OBJECTIVE COVER GLASS . :'0| LIGHT Alb SOURCE CONDENSER COLLOIDAL MAGNETITE ON SPECIMEN Fig. 2. The "brightfield" system of illumina- tion. Reflected light returns through the objective lens, transparent mirror and eyepiece to the ob- server. 441 LIGHT (OPTICAL) MICROSCOPY used most frequently by the Laboratories, the rays from the Hght source are directed around the objective, strike an annular mir- ror or condenser, and illuminate the speci- men. This is illustrated in Figure 3. Since a polished magnetic specimen is highly reflect- ing, all incident rays striking the surface are reflected away from the objective, and the field in the microscope appears black. How- ever, the small colloidal magnetite particles, which have already aligned themselves along the domain boundaries, are illuminated by the light impinging on their irregular sur- PLANE MIRROR SURFACE DARK CENTER-STOP -f LIGHT SOURCE OBJECTIVE SURFACE OF ANNULAR MIRROR CONDENSER COLLOIDAL MAGNETITE ON SURFACE OF SPECIMEN Fig. 3. "Darkfield" system of illumination. Light is reflected from the specimen surface through the microscope lens system to the eye. faces. Since no light emanates from the back- ground, the reflecting particles appear white. This provides maximum visual or photo- graphic contrast as illustrated in the domain micrograph (Figure 4). Over the years, the colloidal magnetite method has been used extensively. In this work, through the efforts of W. O. Baker and F. Winslow of the Laboratories, further re- finement in the preparation of colloidal mag- netite made it possible to see the more deli- cate delineations of the domains, and to observe the changing domain patterns over a much longer period of time. Of particular interest was the change in domain patterns caused by various physical influences such as inclusions or small surface defects. It was further observed that irregu- larities in specimen preparation, such as sur- face straining, would cause a type of domain pattern that was not representative of the underlying structure. In addition to present- ing a static picture of domains under the in- fluence of an applied magnetic field, these domains would move with a change in field Fig. 4. Magnetic domain patterns obtained on silicon iron by H. J. Williams using the colloidal magnetite method with darkfield illumination. 442 MAGNETOGRAPHY intensity or direction. As a result, the domain patterns could be photographed and re- corded in still or motion pictvu'es. (The lat- ter were taken by F. Tylee in collaboration with H. J. Williams and J. G. Walker.) Although much was gained in the study of magnetic effects with colloidal methods, which are still used to great advantage, there are certain inherent restrictions. Among these limitations are the time recjuired for the particles to collect along the domain walls and the failure to delineate the fine structure that the microscope can resolve. During the latter part of the nineteenth century, a useful effect was observed by the Scottish scientist, John Kerr. He found that, when a beam of polarized light is reflected perpendicularly from a surface that is mag- netized normal to itself, the plane of polari- zation of the light is rotated. The direction of rotation depends upon the polarity of the reflecting surface. Thus, if the north pole of a material rotates the plane in a clockwise di- rection, the south pole will rotate the plane in a counter-clockwise direction. It was thought that a study of the domains in a magnetic material might be possible using this effect. The phenomenon, known as the Kerr magneto-optic effect, was in 1950 ap- plied to the microscopical study of magnetic domain structure in cobalt by H. J. Williams, E. A. Wood and the author. In 1938, L. V. Foster of the Bausch and Lomb Optical Company described an illumi- nating system for a metallograph, using a cut and cemented calcite prism as a polariz- ing illuminator. Later, a polarized light com- pensator was developed to be used in con- junction with this metallograph for the study of opaciue minerals. The combination of these two illuminating devices is shown in the dia- gram of Figure 5. With this device, either the north or south poles in the surface of the specimen can be made to appear bright while the other set re- mains dark. This is done by adjusting the elliptical and rotation compensators, shown in Figure 5, to extinguish the light from one set of domains. When there is no compensa- tion, both sets of poles or domains have the same intensity and they cannot be distin- guished. By means of this polarizing attachment for the metallograph, it was believed that the Kerr magneto-optic effect could be used to study the domain structures in magnetic materials. In the initial studies, using this effect, a single crystal of cobalt was selected. Subsequently, a cobalt crystal was made in the form of a half disc. With this shape crys- tal it was possible to observe domains be- tween positions parallel to the c-axis, or easy direction of magnetization, to positions per- pendicular to this axis. To minimize the pos- sibility of a disturbed sm'face of cold worked material, the crystal was carefully electro- polished. Cobalt has a hexagonal structure. In this element, the domains tend to lie with the direction of magnetization in either a posi- tive or negative sense along the c-axis. The photomicrographs that are shown in Figure 6, starting with a displacement of 4 degrees SPECIMEN BLACKENED OBSERVATION _./ elliptical > '^'-IcompensatorJ / rotation ^ ^compensator; EXTRAORDINARY RAY CALCITE PRISM LIGHT BEAM Fig. 5. Adaptation of elliptical vibration com- pensator used in conjunction with metallograph for studies of domains in magnetic materials. 443 LIGHT (OPTICAL) MICROSCOPY from the c-axis and udvuiiciiifi; through posi- tions to 20 degrees, 49 degrees and 70 de- grees, illustrate the change of domain pat- terns from rosettes to elongated areas. On the surface perpendicular to the c-axis (Fig- ure 6) the magnetization forms small regions having poles of opposed polarity. This gives rather complex domain patterns on the sur- face, as can be seen in the comparison pho- tomicrograph. Figure 7. Photomicrograph Figure 7, left, shows a collection of magnetite Fig. 6. Domain structure in a cobalt crystal cut in the form of a half disc. Arrows show direc- tions of magnetization in the domains. No mag- netic field applied. particles (darkfield illumination) and Figure 7, right, is a direct view (polarized light) of the polished surface. The latter shows many fine details that cannot be seen in the colloi- dal magnetite picture. In the microscopy of the magnetic do- mains of cobalt it has been found that the contrast is very low and, at times, has caused considerable difficulty in visual study. Pho- tomicrography Avith high contrast emulsion films and special development has further aided this study. Recently, B. M. Roberts and C. P. Bean of the General Electric Research Labora- tories have studied the ferromagnetic inter- metallic compound manganese-bismuth. This also has a hexagonal crystal structure with the direction of easy magnetization along the c-axis. In this compound, a much greater contrast is achieved with the do- main patterns appearing virtually black and white. A study of this compound is currently in progress at Bell Telephone Laboratories ; a domain pattern showing rosettes that indi- cate a surface normal to the c-axis is shown in the reference on page 440. Such patterns are an aid in determining the orientations of the crystallites in a polycrystalline specimen. The importance of microscopy as an aid in the fundamental studies of magnetic do- •*^-^^Vf ■hm^ «■ (Old) Pattern obtained using colloidal magnetite Fig. 7. Comparison of techniques. (New) Polarized light pattern showing greater detail on same area of the cobalt crystal. 444 OPTICAL THEORY OF LIGHT MICROSCOPE mains cannot be overlooked. The colloidal microscope, and the second (or eyepiece) magnetite method has long been established stage results in the very large "virtual inl- and has become an important adjunct as a age" shown near the bottom. Fig. 1 also is microscopical method applied to magnetism, helpful in showing how the mirror directs This method gives strong delineation of the the entering light upward toward the con- domain patterns with a minimum of speci- denser, which in turn concentrates the light men preparation and outlay for specialized into a brightly illuminated spot on the ob- equipment. On the other hand, the polarized ject. light method has the advantage that it yields instantaneously to field intensity or Resolving Power directional changes, and employs no surface The statement that the magnification of additives. Also, it may be possible to use the light microscope ranges from about 10 X this method to study specimens at elevated to 2500 X raises the question as to why it temperatures. It is believed that with fur- cannot be extended beyond 2500 X. Actu- ther development of the polarized light ally it is quite possible to do this, but very method a more intensive study of the geo- little is gained by doing so, due to the limit metrical configuration of the magnetic do- in resolving power imposed by the finite size main will be possible. of light waves. Thus resolving power is in ^ ^ ^ the final analysis the fundamental quality F. Gordon Foster , • , • .• i ^■ ■^. ^.u which imposes a practical upper limit on the .^.^.■-....^ ...^r^^^^^^-o r, .-...^■...— r, magnifying power of the microscope. MEASURING MICROSCOPES. See ENGINEER- ^ ^. .^ ^ , \^ ,.,.^ iM^ xAirDncrriDcc a to Resolving power is a measure of the ability ING MICROSCOPES, p. 439. » ^, . ^■ .■ ■ -u c ^ ^ -i of the microscope to distinguish hue detail. OPTICAL THEORY OF THE LIGHT It is, in quantitative terms, the distance be- MICROSCOPE tween two points in the object which can ,, ._ . just be detected as being separated and not Maenmcation *. , single, ihe resolving power oi a microscope The light microscope is an optical instru- depends generally on the design of the ob- ment which produces highly enlarged images jective. An objective capable of utilizing a of very small objects. The microscope large angular cone of light coming from the achieves its magnifying power in two stages, specimen will have better resolving power The objective performs the first stage of the than an objective limited to a smaller cone magnification, forming a magnified image of of light. This is demonstrated in Fig. 2, the object near the top of the microscope where the same specimen area has been pho- tube. This image is further enlarged by the tographed with two different objectives of eyepiece, which acts as a magnifier. Objec- quite different designs as indicated below tives range in power from about 2X to each photomicrograph. 100 X. Eyepieces range in power from about The quantity A^ sin U in Fig. 2 is called 5X to 25 X. Thus the total magnification, the Numerical Aperture (N.A.) of an ob- being the product of objective and eyepiece jective. By definition: magnification, ranges from about 10 X to 2500 X. Fig. 1 is a ray diagram of a typical micro- where N is the lowest refractive index be- scope, and illustrates the two stages neces- tween the specimen and the objective, and U sary to achieve the final magnification. The is the half-angle of the cone of light as shown first (or objective) stage results in the "pri- in Fig. 2. The objective shown at the right mary image", shown near the top of the has a higher N.A. and has the superior re- 445 NA = AT sin U LIGHT (OPTICAL) MICROSCOPY Retinol Imoge MecharNcol Tube Length (160 mm) Projection Distance (250 mm) Fig. 1. The path of light in a microscope. solving power. The smallest detail h which can be resolved by an objective is given by the formula h = \/2 N sin U = \/2NA. As the above formula indicates, there are three ways in which to decrease the least re- solvable separation h. The first is to decrease X, the wavelength of light; the second is to increase the angle U in the object space; and the third is to increase the refractive index N in the object space. The wavelength can be decreased to a rather modest extent by using the blue-violet region, the short wavelength end of the white light spectrum. Still further shorten- ing of the wavelength is possible by using the ultraviolet region. This, however, re- quires a completely different optical system from the light microscope. Increasing the angle U in Fig. 2, increases resolving power in proportion to the sine of the angle. Highly complex objective lens systems result in values of sin U as high as about 0.95, and this value represents about the maximum attainable limit for the N.A. of a top-ciuality dry objective. Normally however 0.65 N.A. represents a practical limit for the great majority of microscopes (see Fig. 3). To go still further in resolving power "im- mersion objectives" are used, giving a value of N greater than unity in the formula h = \/2N sin U Immersion objectives, using oil as an im- mersion fluid, result in N.A.'s as high as 1.40, although 1.25 is more common. Monobromo- 446 OPTICAL THEORY OF LIGHT MICROSCOPE U-6" 54' UM4-29 Fig. 2. The dependence of resolving power upon numerical aperture. The same object is here photo- graphed with two different objectives. The picture at left was obtained with a 0.12 N.A. objective. The much sharper and better resolved picture at right was taken with a 0.25 N.A. objective. naphthalene immersion objectives with N.A.'s of 1.60 have been made, but they are very uncommon, and are difficult to use. Depth of Focus At any given focus setting of a microscope only a limited thickness of the specimen ap- pears in sharp focus. This thickness is called the "depth of focus" and, like resolving power is largely dependent on the objective design. The depth of focus is (approximately) inversely proportional to the scjuare of the N.A. of the optical system, and is extremely small for high N.A. systems, e.g., a 1.40 N.A. objective has a depth of focus of about 3^ wavelength of light (0.00025 mm). Aberrations A first-class microscope is remarkably free from aberrations, or image defects, at least in the central portions of the field. However, some understanding of the seven different basic lens aberrations is essential to the Fig. 3. Objective types. The objective at the right is a 4 mm, 0.65 N.A. Achromat hav- ing 5 glass elements, and no fluorite. The central objective is a 4 mm, 0.85 N.A. semi-apochromat, having 5 glass lenses and 1 fluorite lens. The ob- jective at the left is a 4 mm, 0.95 N.A. apochromat, having 5 glass elements and 2 fluorite lenses. proper selection and intelligent use of a microscope and its accessories; hence a brief discussion of the aberrations as they apply to a microscope is given here. (1) Spherical aberration exists where light from a single object point on the axis is more strongly refracted by either the inner or the outer portion of the lens aperture, resulting in failure of the light to come to a common 447 LIGHT (OPTICAL) MICROSCOPY focus. The visible effect is a loss of contrast, jectionable in photomicrography where eye- Well made objectives are normally quite well accommodation and re-focusing in scanning corrected for spherical aberration. High the field are not possible. Special fiat-field N.A. dry objectives, however, are quite sen- objectives are sometimes used here, and also sitive to the effect of cover-glass thickness on special negative lens systems can sometimes spherical aberration. Most objectives are de- by used to replace the eyepiece, to improve signed to work with 0.18 mm cover-glass the performance of regular objectives, thickness, and even small departures from (5) Distortion is the lens aberration in this nominal value will result in loss of con- which straight lines are imaged as curved trast in the image in high N.A. dry objec- lines. This aberration is generally under good tives. The experienced microscopist uses control in a microscope, cover-glasses very close to 0.18 mm for opti- (6) Chromatic aberration is the defect mum image quality. Immersion objectives which causes light of different wavelengths are much less sensitive to cover-glass thick- to be brought to different foci. Chromatic ness, since immersion oil is quite close to the aberration is normally under good control glass cover slip in refractive index and forms in a microscope, particularly in the highly essentially a homogeneous optical medium complex "apochromatic" objectives (see between specimen and objective. "Objective Types")- (2) Astigmatism is the defect whereby the (7) Lateral color results in light of one image of a point is drawn out into two sepa- color being imaged at a greater magnification rate line images at 90° to each other. In a than light of another color, causing the image well made microscope, astigmatism will of an off-axis point to be spread out into a sometimes be present to a minor extent near tiny spectrum. This defect is greatest in the the margin of the field, but not at the center, higher-power objectives. It can be compen- (3) Coma is the lens defect in which dif- sated by proper choice of eyepiece, and ferent circular concentric zones of the lens again the manufacturer's recommendations system have different magnifications. It re- should be followed. suits in a comet-shaped image of a point object. Like astigmatism, in a well made Objective lypes microscope it will sometimes be present to a Three types of objectives are normally minor extent near the margin of the field, but made available by microscope manufactur- not at the center. ers. These are called "achromats", "semi- (4) Curvature of field is the defect in apochromats", and "apochromats", named which a flat object is imaged as a curved in order of increasing excellence and com- surface. This defect is the most difficult one plexity (see Fig. 3). The achromats are the to deal with in a microscope design. The simplest and least expensive. For most pur- high-power objectives in particular tend to poses achromats do an adequate job, and have strongly curved fields, so that when consequently they are popularly used on the central image is in sharp focus the mar- most medical and laboratory microscopes, ginal image is out of focus, and vice versa. The achromat is corrected for chromatic A certain amount of compensation for the aberration at two wavelengths, one in the objective curvature of field is possible by red and one in the blue, and is fully corrected optimum choice of eyepiece, and in this re- for spherical aberration at one wavelength gard the microscopist is well-advised to fol- in the yellow-green. At other wavelengths low the manufacturer's recommendations on in the visible spectrum the correction is good, objective and eyepiece combinations. but not complete. Normally, curvature of field is more ob- By combining fluorite lenses with glass 448 OPTICAL THEORY OF LIGHT MICROSCOPE lenses it is possible to get better color correc- Illuminating Systems tion in an objective. The semi-apochromat uses fluorite to a limited extent, and repre- sents a compromise between the achromat Brightfield Illumination. Most micros- scopy is done with "brightfield illumination" in which the illuminating beam is a solid and apochromat in performance. The apo- ,. ,. , , , , , ^ a ■. ^ ■ , • cone oi light concentrated on the specimen chromat uses several nuorite lenses in combi- , , , i i- , ,. -.u 1 1 XI- by a condenser lens system, mounted directly nation with glass lenses, to achieve a very , , , • m, • i ,.,, c ,. ^ ,. » beneath the specimen, ihis lens system is high degree oi correction. Correction for „ , ,, , , ,, , • , .■■,.■■ u- 1 . xu called a substage condenser and is nor- chromatic aberration is achieved at three „ . , . , . . ,. , ,,,.,, J , , , mally equipped with an ins diaphragm to wavelengths m the red, green, and blue „ , , , , , ,. , J 1 • 1 u X- • allow the observer to control the angular respectivelv, and spherical aberration is ,.,,•,,•• ^ i, n^, • , , , , , "^ J 1 , 1 xu 1 J cone 01 the illuminating bundle. 1 his control also held under closer control throughout , ,• , • • , ,, . ., , , ,, . ., 1 . ^1 on the light permits one to attain the opti- the visible spectrum than is possible m the . , , , , . , , 1 „ mum compromise between contrast and re- achromats. Apochromats are also normally , . „, . . ■,, . r ^ , . , . T.y . ,, ,. solving power, ihis setting will be found to higher in JN.A. than corresponding powers ,-rr i i- i r- i . ,11, diner depending on the nature of the speci- in the achromats. r- o t- men. T^ . rr. Sometimes the light source will be external Eyepiece Ivpes , . , , , to the microscope, but the tendency is to- Far and away the most popular eyepiece ward built-in sources, for obvious reasons of for general microscopy is the Huygens eye- convenience and fool-proof operation. How- piece. This comprises two separated simple ever, whether the source be built-in or ex- plano-convex lenses, with field diaphragm ternal, the principle of Koehler illumination located between the lenses. Despite its sim- is normally employed in this system, the pUcity and inexpensive construction it is light source is imaged by a "lamp condenser" well corrected for lateral color and works into the substage condenser. The latter im- quite well with low and intermediate power ages the lamp condenser onto the specimen, objectives. The virtue of this system is that it results in For use with higher-power objectives, eye- a nice even field of illumination, even pieces of more complex form are superior to though the source is of uneven brightness, the Huygens eyepiece. These more complex as for example a coiled filament lamp. A eyepieces compensate the lateral color aber- small amount of diffusion is sometimes added rations of the higher power objectives, and to completely even out the illumination. Fig. are particularly suitable for use with apo- 4 compares Koehler illumination with the chromats. These eyepieces are sometimes formerly much-used critical illumination, called "compensating" eyepieces, and in The most commonly used condenser is the other instances are given specific trade two lens "Abbe condenser", having an N.A. names. The various manufacturers recom- of 1.25 or 1.30 (see Fig. 5). The upper ele- mend optimum combinations of eyepieces and ment is a plano-convex hemisphere, and the objectives to achieve best correction, partic- lower element is bi-convex in shape. While ularly in regard to lateral color. the aberrations of this system are sizeable, For photomicrography special negative it nonetheless does a very creditable job of (dispersive) lens systems are available, which illumination for any normal routine tasks, markedly improve the correction of the im- It is sometimes made with the lower ele- age for flatness of field. These negative sys- ment independently focusable as shown in tems cannot be used as visual eyepieces, but Fig. 5 to provide better illumination of large are strictly for photomicrography. fields in low-power work. 449 LIGHT (OPTICAL) MICROSCOPY SOURCE IMAGED IN SPECIMEN PLANE CRITICAL ILLUMINATION LAMP CONDENSER IMAGED IN SPECIMEN PLANE FIELD DIAPHRAGM SOURCE IMAGED IN APERTURE DIAPHRAGM KOEHLER ILLUMINATION Fig. 4. Ray-paths in critical and Koehler illumination. The Koehler form of illumination is most widely used, since it gives even illumination despite lack of uniformity in the source. <::T:ri FOCUSABLE ELEMENT ABBE CONDENSER VARIABLE FOCUS CONDENSER ACHROMATIC CONDENSER Fig. 5. Condenser types. The simple Abbe type at the left and the modified variable focus Abbe in the center are the most commonlj' used forms. The more complex achromatic form however gives more color-free illumination and higher N.A. 450 OPTICAL THEORY OF LIGHT MICROSCOPE For more exacting work in microscopy or for photomicrography, an achromatic 1.40 N.A. condenser is often used. This is ciuite well corrected for spherical and chromatic aberration and permits more accurate con- trol of the illuminated field and aperture, as well as giving more color-free illumination than is possible with the Abbe condenser. Its construction is shown at the right in Fig. 5. Darkfield Illumination. This is another common system of illuminating microprepa- rations. In this system the illuminating cone is again concentrated on the specimen, but in this case is hollow, having a dark central core, as shown in Fig. 6. The objective lies in this dark central core, and "sees" only objects which scatter light onto it. The clear background appears black, and the specimen shines bright against this dark background. This very striking form of illumination is useful on very small transparent living ob- jects such as spirochetes, bacilli, etc. The most common form of darkfield illu- minator is the Paraboloid Condenser, shown at the left in Fig. 6. A somewhat more com- plex, and slightly superior design is the Car- dioid Condenser, shown at the right in Fig. 6. The Paraboloid delivers a hollow cone lying between the approximate N.A. values 1.15 and 1.40. The cardioid delivers a somewhat better concentrated hollow cone lying be- tween the approximate N.A. values 1.20 and 1.40. Since in both cases the N.A.'s exceed unity, it is obvious that the condensers must be oil-contacted to the glass object slide. Since the objective N.A. must not be great enough to accept any of the direct cone of light, it is necessary to either use objectives lower in N.A. than about 1.0, or else reduce the N.A. to about 1.0 by a small baffle, called a "funnel stop", which fits into the back of an objective. The effectiveness of a darkfield sj^stem is dependent on the use of an intense, nondif- fused, light beam from the lamp condenser. Carbon arcs, ribbon filaments, or compact coil filament lamps are normally employed as the light sources. Extreme care must also be taken to get the object slide and cover glass clean, since any foreign objects will scatter light and destroy the contrast. To carry the darkfield principle to its ex- treme, a very intense beam of light may be directed onto the object at approximately right angles to the optical axis. Such an ar- rangement is called an ultramicroscope, the name being derived from the fact that it is possible with this instrument to see particles which are well below the resolving power of I OBJECTIVE CARDIOID CONDENSER PARABOLOID CONDENSER Fig. 6. Two forms of darkfield condensers. Note that, in each case, the direct light from the con- denser is not picked up by the objective. The scattered light from the object, indicated by the dashed rays, produces a bright image seen against a dark background. 451 LIGHT (OPTICAL) MICROSCOPY the microscope. The shape or nature of such particles cannot, of course, be determined, but they can be detected, and some estimate made of their size by inference from their brightness, and the speed of their Brownian movement, if the particles are suspended in a hquid. Since the limitation on the smallest par- ticle which can be made visible is imposed by the level of energy of the illuminating beam, the ultramicroscope uses an extremely bright light source, such as a carbon arc, and concentrates the light energy by means of a medium-power microscope objective into a horizontal beam at right angles to the IMAGE PLANE REAR FOCAL PLANE OF OBJECTIVE CONDENSER - UNDIFFRACTED ORDER LIGHT SOURCE Fig. 7. The phase contrast microscope. The illumination is restricted to a hollow cone by the Annular Diaphragm, which is imaged into the Phase Shifting Element by the lenses. Interfer- ence, producing a visible image, occurs between the diffracted light passing around the Phase Shifting Element, and the direct light passing through this element. microscope optical axis. The object slide is of special construction to admit this con- centrated beam from the side. Phase Contrast. Phase contrast is an- other form of illumination useful on trans- parent objects, and in particular on live prep- arations. The normal microscopic object is seen because it has regions of different opac- ity. In brightfield illumination a completely transparent specimen is very difficult to see in any detail, as all parts are equally dense. Darkfield illumination shows up border ef- fects in such specimens, due to edge scatter- ing, refraction, and diffraction. Phase con- trast, is of more value with transparent media, due to its ability to reveal internal detail. This method has found extensive use in the study of transparent living prepara- tions. Phase contrast is really more than just a change in the form of illumination. It is bas- ically a method of separating the diffracted and undiffracted parts of the light, treating them differently, and then re-combining them under conditions such that they pro- duce controlled visible interference effects. The scheme produces a visible image of an otherwise almost invisible object. The arrangement necessary for phase con- trast is shown in Fig. 7. A clear annulus in the focal plane of the condenser is imaged at infinity by the condenser, and then re- imaged by the objective in its upper focal plane. The undiffracted light all passes through this image, and, by means of an annular phase pattern located in this focal plane, is both reduced in intensity and given a quarterwave phase shift with reference to the diffracted light. The end effect of these two changes in the undiffracted light, when combined with the diffracted light passing around the phase annulus, is to simulate the phase and intensity distribution which would be present in the objective focal plane if the specimen had density variations rather than refractive index variations. As a conse- quence, the image formed by the interfer- 452 OPTICAL THKORY OF LIGHT MICROSCOPE ence of the diffracted and iindiffracted por- tions simulates that of a specimen having density variations. For a more thorough understanding of phase contrast the reader is referred to "Phase Microscopy", by Bennett, Jupnik, Osterberg and Richards. Polarized Light is another method of bringing out visible structure in transparent materials. Light energy is transmitted by transverse waves, generally vibrating in any direction at right angles to the direction of the light ray, but it is possible by means of devices known as "polarizers" to restrict the vibrations to a single plane. If two such polarizers are inserted in the light beam in such a manner that the second one trans- mits in a plane at right angles to the first, the light will be extinguished. Such an ar- rangement is called "crossed polarizers". The lower element is called a "polarizer"; the upper one the "analyzer". In practice, the polarizer is located beneath the microscope condenser j and the analyzer just above the objective. If now, in between these crossed polarizers we insert an object which is crys- talline in nature, it will appear bright against the dark background caused by the crossed polarizers. Furthermore, if the crystal is ro- tated about the optical axis of the micro- scope, it will alternately brighten and darken every 90°, usually with attendant strikingly beautiful changes of color. The reason for this is that crystalline materials have differ- ent properties in different directions, and as a conseciuence they alter the state of polari- zation of the light, and thereby effectively "uncross" the polarizers. Polarized light has long been a mainstay in the study of crystals, minerals, chemicals, and fibers. Special microscopes called "pet- rographic" or "chemical microscopes" are used for any serious work in these fields. These are equipped with accessories for mak- ing quantitative measurements of the polar- izing characteristics of the material under study. Fig. 8 shows a typical Petrographic Microscope. It is equipped with a circular rotating stage, divided in degrees, with a vernier permitting reading of crystal orien- tation of 0.1°. The objectives are mounted in centerable mounts, since it is necessary COARSE FOCUS 8ERTRAND SWING "OUT AND IRIS ANALYZER SWING "OUT AND ROTATION MECHANICAL STAGE FINE FOCUS- CONDENSER SWING -OUT CONDENSER FOCUS BERTRANO FOCUSING ACCESSORY SLOT CENTERABLE OBJECTIVE GRADUATED ROTATING STAGE CIRCULAR POLARIZER POLARIZER ROTATION k Fig. 8. A Typical Petrographic Microscope. This specialized microscope for use in the study of crystalline materials differs in many respects from the conventional microscope, as indicated by the many special control knobs in the figure. 453 LIGHT (OPTICAL) MICROSCOPY to have the specimen stay accurately on center as it is rotated. The polarizer, mounted beneath the substage condenser, and the analyzer mounted above the objec- tive are rotatable, and provided with circular scales to read their orientation. Beneath the analyzer is an accessory slot, in which one may insert measuring acces- sories, such as a quarterwave plate, a sensi- tive tint plate, a quartz wedge, or a tilting calcite plate compensator. The first two of these devices are used to categorize crystals into one of several classifications. The quartz wedge refines this classification into semiquantitative sub-classifications, while the tilting calcite plate gives actual quan- titative measurements for more refined clas- sification. Above the analyzer is an important ele- ment of the Petrographic Microscope, known as the 'Bertrand Lens'. This lens forms an image of the back aperture of the objective in the eyepiece focal plane, so that the ob- server sees the aperture rather than the field of the microscope system. A crystal, located between high N.A. condenser and objective, is traversed by polarized light in many di- rections. In each direction, its properties dif- fer, and it accordingly transforms the state of polarization differently, so that between crossed polarizer and analyzer the crystal ex- hibits a characteristic light pattern which varies with direction. Looking at the aper- ture of the objective, one sees this pattern of light and shade, known as the ''interference figure", and gains a clue as to the nature of the crystal. The interference figure of a crystal may be further investigated in a quantitative manner by means of an accessory called the 'Universal stage'. This device, located on the stage of the microscope, permits turning and rotating the crystal about several different axes. Crystals have either one or two direc- tions called 'optic axes' in which they behave as non-crystals. By means of the Universal Stage, the positions of these optic axes may be determined, giving a very good clue as to the composition of the crystal. For a more complete understanding of polarized light microscopy, the reader is referred to the following texts: "Crystallog- raphy and Practical Crystal Measurements", A. E. H. Tutton. "Microscopic Character of Artificial Minerals", A. N. Winchell. "Hand- book of Chemical Microscopy", E. M. Cha- mot and C. W. Mason. "Optical Mineral- ogy", Rogers and Kerr. James R. Benford ORIGIN AND HISTORY Development of Optics and Microscopy from About 1600 to 1723 During the first years after the invention the words "perspicillum" or "conspicilium" were used both for the telescope and the mi- croscope. What was meant in each case had to be understood from the context. For the microscope also "smicroscope" was used or "engyscope" and other words. Eventually the names familiar to us were accepted about 1615. They originated with the Academica dei Lyncei which had been founded in 1003. Allegedly the names were invented either by Giovanni Faber or Giovanni Desmici- anus, who were both members of the Acad- emy and whose names appear in writings from that time. Johannes Kepler (1571-1630) is better known for his achievements in astronomy but he also studied intensively optical phe- nomena. In his Paralipomena, printed in 1604, he treated the nature of light and color, the position of images, refraction, function of the eye, and other subjects. Color is ex- plained as light partially buried in the ma- terial of the medium which can vary in dens- ity, transparency, opacity and the various degrees of "the light inherent in the ma- terial." Kepler labored hard to determine the law of refraction. He experimented with many transparent media, as air, water, glass, turpentine, vinegar, wine, several oils, and 454 ORIGIN AND HISTORY others. He established the fact that refrac- famous Discours de la Methode in Holland in tion does not depend on the intensity of light. 1637. It is in its main part a philosophical Quantitatively, he thought that for the ratio treatise which established the fame of its of the angles of incidence and refraction a author all over Europe but it contains also relation involving the secant of the angle of essays on geometry, meteors, and optics in incidence would hold. which applications the usefulness of his The anatomical structure of the eye was method is demonstrated. In the optical part, unknown at Kepler's time. He made the dis- called Dioptrice or Diopfrique, Descartes CO very that the image of an object observed strives at a deeper understanding of optical is focused on the retina by the refractive experiments from mechanical considerations, power of the crystalline lens. He is also Already Hero of Alexandria had felt the among the first to arrive at distinct prescrip- necessity of explaining the law of reflection tions for defective vision, namely convex by assuming that it was a property of light lenses for farsightedness and concave lenses always to take the shortest path between for shortsightedness. two points. Kepler tried to understand the In 1611 the Z)fop^nce was published, which path of the reflected ray from the specific is a "demonstration of those principles which resistance which the light finds in the dense relate to vision and visible objects on account media. of those glasses onl}^ lately invented." In Descartes, however, tried to establish a the preface the author states that the book complete system of optics on mechanical was inspired by the wonderful astronomical principles based on the assumption of an discoveries which Galilei had made with his ethereal fluid penetrating all bodies and the telescope and reported in the preceding year space around them extending throughout the 1610. The book brings some new results in whole universe. The particles of this fluid all fields. The critical angle of total internal were thought to be very small and to act reflection is defined in the chapter dealing like balls bouncing off other bodies. In this with refraction. way Descartes explains the equality of the Kepler used the camera obscura for study- angles in reflection as a plausible mechanical ing experimentally the properties of some phenomenon. By suitable assumptions on lenses in order to confirm the results of his impact and resistance of those balls imping- mathematical considerations. Kepler arrived ing on the surface of optical media he arrived at the conclusion that the internal surface at the law of reflection which he properly of the crystalline eye lens must be a hyper- formulated as the constancy of the ratio of boloid for the best focusing conditions on the the sines of incident and reflected angles, retina, thus anticipating Descartes. Kepler Descartes does not mention the names of discussed combinations of convex and con- any other investigators in this connection, cave lenses both for each kind in itself and Usually the law of refraction is ascribed to with the other kind, which includes the Willibrord Snell, 1591-1626, a professor of Galilean telescope; what we call today the mathematics at Ley den University, who had astronomical or Keplerian telescope is de- taught about his law in his courses on optics, scribed. However, no reference to the use of a He had prepared a publication but it was combination of lenses as a microscope is never printed. Huygens is reported to have made, although the mathematical treatment seen the manuscript in which the law of re- of lens properties is quite thorough. It in- fraction was expressed in mathematical eludes for example also an investigation of terms, stating that for two different optical spherical aberration. media the ratio of the cosecants of the inci- Rene Descartes (1596-1650) published his dent and the refracted angles stays constant 455 IJGIIT (OPTICAL) MICROSCOPY if the angle of incidence varies. Later, also the back through a compound microscope by Descartes, this law was expressed by the the tube of which is put through a hole in respective sine functions. the vertex of the mirror. An interesting fea- The publication of Descartes' Dioptriqiie ture of this scheme is Descartes' proposition was followed by a violent controversy which to use a Galilean type arrangement of one continued even after Descartes' death. The convex and one concave lens, both of which theoretical derivation of the law was assailed, have to be hyperboloid. Descartes had stated that the velocity of Still another proposition of Descartes has light in the denser medium was increased, the purpose of diminishing spherical aberra- whereas his opponents argued that it would tion in a unique way. He suggested putting be retarded. Hero had argued for the law of a water-filled tube with a membrane at one reflection that a ray would follow the short- end right at the eyeball. The membrane est path. For the law of refraction, Fermat would then adapt itself and its shape to the assumed that a ray following an intrinsic contour of the eyeball. The other end of the law of nature would travel that path in a tube would be closed with a curved glass disc denser medium which requires the shortest shaped similar to the eye curve. The effect time. For this argument it must be assumed would be to increase the depth of the aqueous then that the velocity of light would be re- humour of the eye because the refractive duced in a denser medium. In the eighth indices of this Hquid and of water are about chapter of the Dioptrique Descartes showed the same. From such an arrangement Des- how the aberration of a spherical lens could cartes expected larger images of the objects be avoided by non-spherical surfaces as for observed, but he realized the inconvenience example ellipsoids or hyperboloids. He gave in practical use. He suggested replacing the detailed instructions how to grind and polish water-filled tube by a long glass cylinder this kind of lens and found one or two glass with a suitably curved surface at both ends, makers who tried to make them, but without or a hollow tube with lenses at both ends, much success. Their surfaces had to be shaped in a special Descartes suggested also an application of way as required by his idea of increasing the nonspherical lenses for single lens micro- effective depth of the human eye humour by scopes. One lens was to be mounted in the optical means. vertex of a spherical concave mirror so Athanasius Kircher, lGOl-1680, wrote the that it would protrude a little through a book Ars Magna Lucis et Umbrae the first hole in the mirror, the object being held at edition of which was published in Rome its focus by means of a prong. The whole 1646 and the second edition in Amsterdam apparatus had to be held close to the eye and 1671. It contains one section "De Mira aimed at the sun. The light rays would illu- Rerum Naturalium Constitutione per Smi- minate the object and it would appear croscopium Investiganda." The word Smi- greatly magnified if observed through the croscopium used here comes from the Ionian lens. This instrument was intended for hand dialect in which Herodotus wrote. Kircher is use only. the only author using this word. It fell out Descartes described another more refined of use in later years. In his book he gives an one to be mounted on a high stand in the account of all the microscopes of his day: same way as telescopes are mounted. A para- (1) Two convex lenses are used, bolic mirror is provided, in the focus of (2) Glass spheres filled with water are which the object has to be held. The illumi- used as single lenses. nation is enhanced by means of a convex lens (3) Tiny glass spheres at the end of a tube in front of the object. It is observed from are used with the object to be put right on it. 456 ORIGIN A\D HISTORY (4) One lens is used which has one spher- ical and one hyperboloid surface. (5) Two hyperboloid glasses are used, one convex and one concave, as described also by Descartes. Caspar Schott (1608-1666) wrote Magia Universalis Naturae et Artis with a chapter: "Microscopes and Their Wonderful Power in Revealing the Constitution of Natural Objects" which appeared in 1657. In this book the author gives a classification of microscopes of his day: (1) A short tube with a fiat glass at one end and a glass sphere at the other end. The object had to be put inside of the tube on the flat glass. (2) A glass vase into which small objects are put at the bottom and the neck of which is closed with a crystalline lens of a spheri- cal shape. Other examples for microscopes are given which repeat what Kircher had listed pre- viously. Peter Borelius made in 1655 a contribu- tion to microscopy of historical importance in his book De Vero Telescopii Inventore. He collected important evidence concerning the inventors of the microscope which is used in the foregoing text. He also gave a classification of the microscopes of his time: (1) Two glasses are put at the ends of a small tube. One is convex with spherical surfaces and the other is plane and carries the object. When this was a fly "enlarged to the size of an elephant", the instrument was called "conspicilium muscarium", or fly glass. But it would be "conspicilium pulicarium" or flea glass when the object was a flea enlarged "to the size of a camel", much to the enjoyment of the observer. (2) One single spherule was mounted at the end of a tube. At the other end would be a small glass box containing sundry tiny objects. (3) Two convex lenses were used with much superior properties. (4) Several coaxial tube parts were mounted in such a way that the whole could be lengthened and shortened. It would contain three or four different lenses. Borelius mentioned also the names of some of the more important opticians of his time and described the manufacture of their in- struments, including the grinding and polish- ing of lenses. In another part of his book Borelius gave a record of microscopical observations in- tended to be a guide to those who "wished to study the Majesty of Cod in minute cre- ation and an offering of respect to the worthy citizens of Middelburg among whom the microscope had been invented." Listed are milk, blood, vinegar, two kinds of lice, fleas, six kinds of worms, spider eyes and eggs, pimples, ants, moss, ferns, caterpillars, but- terflies and other objects up to one hundred exactly. The attitude of Borelius demonstrated in his work and the details both of his instru- ments and observations classify him more as an amateur of optics; he was a physician at the court of Louis XIV. Several other writers in the 17th century were delighting in listing the innumerable objects of the new world invisible to the naked eye. Among them was Robert Hooke, who gave a list of some 60 microscopical observations in his J\Iicrographia, written in Enghsh and pubhshed in 1665. This book is remarkable because it also contains a de- tailed description of his compound micro- scope which had two convex lenses and, what would be called later, a field lens which he had learned to use with advantage for some observations. He also built a special illuminator in order to get brighter images of his objects. He complained about the un- avoidable smallness of the objective lens which will not "admit a sufficient number of Rayes to magnifie the Object beyond a determinate bigness." He mounted on a stand an oil lamp, then in the proper dis- tance, "a pretty large Clobe of Class fiU'd with exceeding clear Brine," and a convex 457 LIGHT (OPTICAL) MICROSCOPY lens. With the flame of the lamp, the glass ball and the lens coaxially lined up he at first got enough light, "but after a certain degree of magnifying they leave us again to the lurch." In one of his models Hooke used two lenses and filled the space between them with water. Such an arrangement was an improve- ment of Descartes' water tube described about 30 3^ears before. Hooke praised the brightness of the image obtained in this way, but he did not use it much "for other in- conveniences." The pictures in Hooke's book are considered by some historians the first precise and detailed technical observations depicted which we possess. Another important microscopist of this time was Johannes Hevelius, (lGIl-1687) a German astronomer who devoted much time to the study of other scientific fields. His Machina Coelestis was printed in two vol- umes, the first of which, published in 1673, contains descriptions of his microscopes. He expressed his admiration for the English microscopes, probably of Hooke's design, but improved it by his invention of a micro- metric fine focusing device. It consisted in a vertical screw-threaded rod, rotating in a spherical nut which pushed a sliding sleeve along another rod upon turning of the screw. The body of the microscope was fastened to the sleeve Avhich, upon traveling along the rod, would gently lift or lower the micro- scope tube. With this "fine adjustment" microscope design made another important advance. In the following century further progress in design and applications of the microscope was made with substantial participation of amateur microscopists, among whom Antonj van Leeuwenhoek must be mentioned as the most famous and successful. He lived all his life from 1632-1723 in Delft, Holland, where he was active in business on a small scale as a haberdasher, draper. Chamberlain to the Sheriffs, and surveyor. He was an amateur without any scientific education and special- ized in optical lens grinding and polishing, in which art he acquired a high professional skill. All his observations and discoveries were made with homemade single lens mi- croscopes, but no description of the manner of manufacturing or of using them was left by Leeuwenhoek. He never divulged his secret methods with which he could out- perform all other microscopists for at least a century. However, a few of his original microscopes are still in existence and have been studied. They have all very small double-convex lenses mounted in a socket between two metal plates, about 1.5 X 3.5 centimeters, riveted together. The object was held by a needle near the socket by means of which the object could be brought into the proper position. The lenses had a focal length of from fractions of a millimeter to about five millimeters, and a magnifying power from about 50 times to 270 times. Coarse and fine focusing were accomplished by means of two screws, one longer and vertical and another shorter and horizontal, the latter being fastened crosswise on the former by a nut piece which held the object needle. Upon proper operation of these two screws, the object could be accurately focused and rotated. Twenty-six of Leeuwenhoek's microscopes were bequeathed to the Royal Society, where they all vanished about a century ago. The rest Leeuwenhoek left with his daughter when he died. After her death in 1745, they were auctioned off and scattered; most of them were acquired by Dutchmen. From the sales-catalog Leeuwenhoek seems to have left 247 complete microscopes with lenses and objects still in place and in addi- tion to these 172 lenses mounted in their plates. Three lenses were made from quartz. The plates were mostly made from silver but three were of gold and sold by weight. His observations and discoveries Leeu- wenhoek described in letters (165 of them) mostly directed to the Royal Society in 458 ORIGIN AND HISTORY London, which made him a Fellow in 1618. His letters have been translated from the old-fashioned Dutch which was Leeuwen- hoek's mother tongue into English and Latin; many of them were published in the Philosophical Transactions from 1673 to 1723. A complete edition of Leeuwenhoek's letters was never published, but a collection of them appeared from 1695-1719 under the title ^^ Arcana Naturae ope micro scopiorum detecta" or "Mysteries of Nature Discovered by Means of Microscopes." Some historians blame Leeuwenhoek for his naive ignorance; others praise him as the founder of bacteriology and protozoology. By the end of the 17th century Leeuwenhoek was the only earnest and scientific micro- scopist in the world. He seems to have been the first to really see and describe "ani- malculae" (the "little animals"), including spermatozoa and bacteria. There were many contemporaries and followers of Leeuwenhoek who tried to imi- tate him but they were seriously hampered by the low quality of the optical lenses avail- able at that time. Neither had they the excellent mechanical skill for making their own lenses as Leeuwenhoek had done, nor did they have the exceptional keenness of his eye. There is no wonder then that this whole area was not ciuite respectable among scientists and remained so for cjuite a long time. Numerous indications of this attitude can be found, for example in the London Encyclopedia in 1829 in an article entitled "Optics:" "Microscopes, though but toys compared with telescopes, nevertheless deserve to be rendered as perfect as possible; for they yield not to them in the quantity and variety of rational amusement which they are ca- pable of introducing to us, though not of the sublime description of the wonders of the heavens. Compound microscopes, though not so much to be depended upon for the purposes of discovery and philosophical in- vestigation as single lenses, are still best adapted for recreation." The preference of single lenses expressed in this article over compound microscopes has its reason in increasing awareness of aberrations, of which especially chromatic aberration is very harmful to the quality of the images attainable with a compound microscope. The so-called spherical aberi-a- tion, brought about by the use of spherical lenses, could be remedied to a large extent, as Newton had shown in his ''Optics," pub- lished 1704. But he also thought that the chromatic abberation was unavoidable be- cause the production of colors by dispersion seemed to be inseparable from the refraction necessary for the formation of images by lenses. This had been found by Newton experimentally. He also derived a mathe- matical formula for dispersion which was criticized by Euler in 1750 and a new formula was proposed. However, about 80 years later, Cauchy showed that Euler's formula was untenable and suggested another. This field of mathematical optics has been worked on by many other scientists and mathe- maticians up into modern times. Many experimental and theoretical in- vestigations were made in the 100 years following the publication of Newton's "Optics" and it was found that the disper- sion and refraction change in a different way in going from one medium to another and the problem was recognized to consist in finding the proper combination of two or more optical media. Development of Microscopy from 1723 to IVIodern Times The problem of manufacturing achromatic lenses, at first for telescopes only, was solved by several amateurs and scientists, either entirely or partially independently. Among them were Ch. M. Hall in 1729, a barrister and amateur in optics; L. Euler, a mathe- matician, in 1747; S. Klingenstierna, a math- ematician, in 1755; J. Dollond, an optician, 459 LIGHT (OPTICAL) MICROSCOPY in 1755. The latter was producing lenses The method of finding the best combina- with reduced chromatic aberrations on a tions of lenses for the desired effect was commercial scale from that time on. more or less trial and error, leading to a The decisive experiments had been made tremendous waste, and it was Carl Zeiss in by Hall, who found that flint glass differs Jena, Germany, who was no longer willing only slightly in refractive power from or- to accept this situation. In order to bring dinary crown glass, though its dispersion is to an end "das ewige bei uns Optikern ge- more than twice as large. This led to a co- brauchliche Probieren" meaning "the eternal axial combination of a strong positive crown trying out of lenses so customary with us glass lens with a weaker negative lens of flint opticians," Zeiss formed a partnership with glass, which compensates the chromatic Ernst Abbe, a young physicist at the Uni- aberration of the other lens but reduces its versity of Jena, who cleared the theoretical magnification to only about one-half. This background and calculated better and better is the type of lens which Dollond brought lenses in the future. From 1883 on apochro- on the market for telescopes from about 1760 matic objectives were made which were on. The difficulty in making perfect chro- corrected both for achromasy and for spher- matically corrected lenses for the microscope ical aberration at three different wave seemed unsurmountable because of their lengths. Because glass with new optical necessary exceedingly small size. properties was needed, Schott's Optical Onlv very few experimenters were success- Glass Works were founded at Jena. It pro- ful in their efforts at that time. An exam- duced all the optical glass needed by Zeiss. pie is an achromatic microscope objective This cooperation turned out to be so ad- in the possession of the Utrecht Uni- vantageous that the Dutch and English versity Museum which had been made monopoly in optics was no longer unchal- by an amateur in 1791. He was a lenged after the second half of the 19th eavahy colonel, named F. Beeldsnijder. century. Even professional opticians, such as Fraun- An important contribution to the optical hofer in Munich, Amici in Modena, Charles art was made by Abbe in 1878 by the design in Paris, and others, were failing in the first of immersion objectives, the principles of quarter of the 19th century. However, as which had been well known for about 200 early as 1807, Harmanus van Deijl, optician years. In the 17th century Hooke had made in Amsterdam, Holland, had produced satis- the observation that he got much clearer factory achromatic objectives for sale and pictures of the "animalcules" living in water had published his efforts and results. Ob- when the front lens of his microscope touched viously, it had been impossible for a long the water so that there was no air between time to learn from his experiences. the lens and the object. The explanation is In 1824 a new principle for making higher- that the resolving power of an objective is powered achromatic objectives for the mi- proportional to its aperture or to the vertex croscope was introduced by the French angle of the cone of light which the front physicist Selligue, who suggested screwing lens admits. This aperture is much reduced several low-powered achromatic lenses to- by total reflection of the light coming from gether. In this way, a higher magnification the object and going in one case, either was obtained without grinding and polishing through water and air or, in the other case, lenses with very short focal length. This through water, a cover glass, and air into principle was used and improved by the the front lens. commercial opticians of that time. Chevalier In Hooke's experiments this loss Avas in France, Amici in Italy, and Lister in avoided because the interspace between the England. object and the front lens of his microscope 460 ORIGIN AND HISTORY was entirely filled with water. Hooke nat- special requirements of diversified applica- urally did not know this explanation but tions leading to detailed specializations of Amici in 1850 tried to make use of this microscope design and methods of use which principle. Obviously the effect would be the deviate more or less from tradition, more useful the more similar the refractive Examples of this development are such power of the medium of the interspace be- fields as interference microscopy, polarizing tween the object and the front lens came to microscopy, X-ray microscopy, and the so- the refractive power of glass. Amici tried called flying spot microscopy. Also belonging other liciuids, e.g., glycerin and oil, with in this group are the reflecting microscope, some success. the fluorescence and the phase contrast mi- These problems were brought to the atten- croscope, which are briefly described in the tion of Abbe by English friends with a re- following. quest for investigations. He complied only The Reflecting Microscope. The prin- hesitantly because the great amount of work ciples of a reflecting microscope were first necessary did not seem to him to be justified described by Isaac Newton (1672) in a letter by only limited apphcability, as for instance to the Royal Society in London. He might in metallurgy. But he measured and tried a have realized that the objective of his re- great number of liquids, about 300, and fleeting telescope, described twenty years found the best to be cedar oil. The next step before, could be used as a microscope objec- was to modify the objective lens so that it tive of long focal length if the light travel would be better adjusted to the optical data through the instrument would be reversed, of the oil. In this way the old idea of Des- In the following 100 years other inventors cartes of optically homogenizing the path tried this principle also on the other reflect- of light through several media, which he had ing telescope types known at that time, tried with his crude water tube, found a re- However, about 1824 this line of develop- vival in a refined manner in modern times. ment practically came to an end because of With apochromatic and homogeneous the introduction and further improvements immersion systems the optical microscope of achromatic lenses. was approaching the limit of its power of A new phase began in 1931, induced bj^ the resolution, which is about 250 millimicrons important advantages which reflective sys- for visible light, permitting the observation tems have over refractive systems for special and identification of most bacteria. This applications. These are the absolute achro- limit was shown by Abbe to be determined matism, the long working distance, and its by the wave length of the light used for unlimited applicability for a very wide range observation. of wavelengths. Depending on the special About the results of his experimental and conditions several distinct groups have been theoretical investigations Abbe reported in developed during the last 30 years, a great number of publications beginning in For single mirror objectives materials like 1869. His complete theory of the microscope quartz, lithium fluoride and other synthetic appeared in periodicals from 1888-1895. inorganic crystals are used for the ultraviolet and the infrared region. The practical use of Recent Lines of Development these systems was found to be awkward in The microscope had been found to be an some cases because the customary micro- instrument with a very wide general appli- scope stand and its accessories, which had cability in many fields of research and devel- their own long line of development, usually opmental technology. It could be expected could not be used. Active in this field were that with the further growth of these fields men Hke R. Smith, 1738; K. Schwarzschild, needs would arise which would demand 1905; H. Chretien, 1922; B. K. Johnson, 461 LIC;ilT (OPTICAL) MICROSCOPY 1934; A. H. Linfoot, 1938; D. D. Maksutov, founder of this field of microscopy, F. Zer- 1944; J. D. Dyson, 1949; A. Elliot, 1950; nike, used a principle known from the study and others. . of diffraction gratings. This principle was at In two-mirror objectives a concave mirror first used for the testing of telescope objec- is used for compensation of the aberrations tives. In 1942, the discoverer recognized its of a second, convex, mirror. This principle importance for microscopy and the number was refined by the application of immersion of its applications grew with every year, and non-spherical systems, monochroma- In the usual microscopic observation the tors, interferometers, and spectroscopes, objects are visible due to their stronger or Workers in this field were Hershgorin and weaker absorption, in other words, the re- associates in 1941; D. R. Burch, 1943; E. duction of the ampHtude of the light waves. M. Brumberg, 1943; A. H. Bennett and In the optical system of a phase contrast associates, 1948; W. E. Seeds, 1949; D. L. microscope another property of the object Wood, 1950; E. R. Blout and associates material is used which consists in shifting 1950; D. W. Dewhirst, 1951; S. Miyata, the phase of the light waves more or less 1952; and many others. depending on the material constants. Upon In solid objectives several reflecting sur- leaving the object the Hght does not appear faces are used but the space between them to be changed in any way when observed is filled with glass, or, in some cases, quartz, with the usual optical means. However, Some authors feel that further development when it is combined again with another part of this type of optical systems might well of the original light bundle which has suf- improve the standards of the old microscope fered a fixed change of its phase by a so- construction as they are followed at the called phase plate, interference occurs which present time. Contributors to the field of renders visible details of the image which solid objectives were D. D. Maksutov, 1932; could not be seen under a traditional micro- D. G. Wynne, 1952; A. Bouwers, 1952, and scope. others. An especially important field of this tech- Fluorescence IMicroscopy. The advan- nique is the observation of details in living tage of this special field of microscopy which cells which must not be harmed or changed was first mentioned in publications in 1929 by the usual methods of selective staining, is the possibility of making visible or study- Those interested in the development of the ing finer details of the structure of living mathematical theory and the contributions protozoa or cells of tissues which might to optics of men like Descartes, Newton, differ from their surroundings in fluorescing Huygens, Gauss, Hamilton, Maxwell, Seidel, power under illumination by ultraviolet Helmholtz, Abbe, Gullstrand, and T. S. light. This necessitates the use of special Smith are referred to articles in the proper optical systems, as for example lenses made periodicals as for example published by the from quartz and aluminized mirrors. Also Royal Microscopical Society in London. A special optical filters were used for a better comprehensive article, 22 pages, about the definition and a suitable choice of a particu- theory of optical systems was published by lar section of the spectrum. Contributions M. Herzberger in Zeitschrift filr Instrumen- to this field were made by M. Haitinger, P. tenkunde Vol. 52, 1932. Ellinger; 1929, J. Smiles, 1933; J. Dyson, 1949; J. R. Benford, 1947; B. K. Johnson, Historical Review 1949; F. Brautigam and associates, 1949; In the course of the development of mi- C. A. Thorold, 1954, and others. croscopy from its origin to modern times, Phase Contrast Microscopy. The distinct periods can be recognized. They are 462 ORIGIN AND HISTORY usually initiated by a technical break- discovered by Pollender in 1849. Establish- through, followed by a whole series of dis- ment of histology as a special field by Kolli- coveries in new fields not accessible by the ker in 1852. The first microtome was de- optical means in use before. signed by Welcker in 1856. Lactic acid This mode of progress is well" known from bacteria were discovered by Pasteur in 1857. other fields of technical and scientific en- Introduction of antisepsis based on micro- deavor but it is especially intriguing to out- scopial evidence by Lister in 1864. Methods line the respective phases in the development of selective staining developed for tissues by of a field the study of which was denied so Gerlach in 1858 and for bacteria by Weigert long to unaided human senses. In the field of in 1875. microscopy the following periods can be Fourth Period: From 1878 to 1931. Intro- recognized, in each of which the technical duction and further development of the progress and the achievements are named: homogeneous immersion, ^c/iteyemen^s; Gon- First Period: From about 1600 to 1723. orrhoea cocci found by Neisser in 1879. Invention of the microscope. Discovery of Leprosy germs seen by Hansen and typhus the law of refraction by Snell and Descartes, germs by Ebert in 1880. Malaria parasites 1637. Establishment of geometrical optics seen within red blood cells by Laveran in by Fermat, Newton, Huygens and Euler. 1880. In the two decades from 1880 to 1900 Achievements: Discovery of the capillaries numerous other pathogenic protozoa and by Malphigi in 1660. Discovery of plant bacteria were discovered, as tuberculosis and cells in fossil wood by Hooke in 1664 and cholera by Koch in 1882, plague by Kitasato the first magnified drawings of small ani- in 1894, tetanus by Nicolaier in 1884, Sj^h- mals, the flea and the louse, by the same in ills by Schaudinn in 1905, of foot and mouth 1665. Discovery of the mammalian ovarian disease using ultraviolet light by Frosch in follicle by De Graaf in 1672. Descriptions by 1924. Leeuwenhoek of protozoa in 1674, striation Fifth Period: From 1931 to the present of muscle fibers in 1682, bacteria in 1676, time. Invention and development of the spermatozoa in 1672, yeast cells in 1688, electron microscope. Further development of leukoc3rtes in 1689, axon and sheath of the light microscope into numerous new nerve fibers in 1717. models for special applications. Achieve- Second Period: From 1723 to 1830. Al- ments: The first micrographs of virus cor- most no progress from Leeuwenhoek's death puscles, not visible with the light microscope, at the beginning of this period until the end. were made by Kausch and Ruska of the This was the high time of the hobby micro- tobacco mosaic virus in 1939. scopist, best exemplified by the title of a It is obvious that these periods overlap book which appeared in 1763. It is Leder- each other widely and many a discovery miiller's "Mikroskopische Gemiits- und might just as well be assigned to the preced- Augenergotzung," which might be trans- ing or the following period. However, a gen- ial ed by "Microscopical Pasttime for Eyes eral pattern is certainly recognizable, and Soul." Third Period: From 1830 to 1878. Inven- references tion and further improvements of achro- Magie, W. F., "A Source Book in Physics," Mc- matic lenses. Achievements: Fermentation is Graw Hill Book Co., New York, 1935. explained from the action of living cells by Sakton, G. "A History of Science," Harvard ^ . , 1 , rr^ • -.oo^ rT^^ n j_ 1 University Prcss, 1 . vol . 1952, 2. vol . , 1959. Cagniard de la Tour m 183/. The first clear g^^^^^^ ^^ "Ancient and Medieval Science," description of the division of the cell nucleus University of Pennsylvania Press, Philadel- is given by Leyden in 1848. Anthrax bacteria phia, 1953. 463 LICTTT (OPTICAL) MICROSCOPY "Moments of Discovery" ed. by G. Schwartz AND Ph. W. Bishop, Basic Books Inc., New York, 1958. King, H. C, "The History of the Telescope" Sky Publishing Corp., Cambridge, Mass., 1955. "Modern Methods of Microscopy" ed. by A. E. ViCKERS, Butterworth Scientific Publica- tions, London, 1956. RoosEBOOM, M., "Microscopium," Leiden, 1956. DoBELL, C, "Antony van Leeuwenhoek and his Little Animals," Russell and Russell, Inc., New York, 1958. "Origin and Development of the Microscope" as illustrated by the Catalogues ... of the Royal Microscopical Societj'. ed. by A. N. Disney. The Royal Microscopical Society, London 1928. HOPPE, E., "Geschichte der Physik," F. Vieweg AG, Braunschweig, 1926. RosENBERGER, F., "Geschichte der Physik," (3 vols.), F. Vieweg AG, Braunschweig, 1890. Dannemann, F., "Die Naturwissenschaften in ihrer Entwicklung und in ihrem Zusammen- hang," 2. ed. W. Engelmann, Leipzig, 4 vol- umes, 1920-23. Articles in Isis, Scientific American, American Journal of Physics, Science, Forschungen zur Geschichte der Optik, Zeitschrift fur Instru- mentenkunde, and others. E. K. Weise PARTICLE SIZE AND SHAPE MEASUREMENTS AND STATISTIC (See also pp. 406, 408) Particle Size Methods The microscope for making particle size measurements should be equipped with a graduated mechanical stage, achromatic or apochromatic objectives including an oil immersion objective, a substage condenser, and a graduated fine focusing adjustment. A revolving nosepiece is desirable which can be used to attach three or four objectives, preferably parfocal, to the microscope. A monocular microscope is preferable to the binocular for highly critical work. A phase contrast microscope is often useful for meas- uring particles of low contrast, and a polar- izing microscope is useful when differentia- tion among various types of particles or identification of particles is desirable. The light source should be equipped with a con- densing lens and a field diaphragm, and should be focusable. A number of accessories are either neces- sary or highly convenient. Particle diameters are generally measured with eyepiece (ocu- lar) micrometers which consist of scales to be placed in the focal plane of the ocular. A number of types are available and the type used depends somewhat on the definition of diameter which is employed. If the diameter is defined in terms of actual dimensions of the particle, a linear scale is desirable. "Globe and circle" micrometers are con- venient for rapid sizing, the diameter being defined as the diameter of a circle whose area is equal to the projected area of the particle. Filar micrometers have either a movable scale or a movable cross hair and fixed scale. Ocular micrometers must be cali- brated for every combination of objective, ocular, and tube length used, and the cali- bration is usually accomplished with a stage micrometer with linear rulings. Another useful accessory is the dark field condenser, which often provides increased contrast. A condenser with central stop is useful for fairly large particles, but the "Cardioid" condenser is preferred for par- ticles having diameters close to the limit of resolution of the microscope. Special devices which produce vmiform il- lumination of opaque objects by reflected light are often very helpful. The illumina- tion may be vertical or annular. Particle size distributions are sometimes made from photomicrographs. Advantages are less eyestrain and a permanent record of the appearance of the particles; disadvan- tages are the additional time-consuming steps and less flexibility in the examination of a field. Microprojection equipment is often useful, especially if the images can be pro- jected on a disposable screen such as a sheet of white paper. The images can be checked off as measured, thus avoiding duplication. 464 PARTICLE SIZE AND SHAPE MEASUREMENTS AND STATISTICS The proper preparation of slides to pro- duce representative, well-dispersed samples is especially important. It is, of course, necessary to start with a small representa- tive sample of the powder, a few milligrams of which are usually dispersed in a drop of liquid on a clean slide. A wooden toothpick is a convenient dispersing tool which has the advantage over glass or steel that it is not apt to crush or shatter the particles. The suspension is spread over the glass slide as a thin film and covered with a clean cover glass to complete the preparation. The fol- lowing are a few of the many liquids which have been used for dispersing powders. Water Water and ammonia Water and potassium citrate (2%) Water and Calgon (0.1%) Glycol Glycerol Ethanol Acetone Cyclehexanol Butanol Dammar Turpentine Duco cement Glucose syrup Xylene Polymethylmethacrylate Polystyrene Solutions of surface-ac- tive agents Considerable experimentation with differ- ent liquids is sometimes required before one is found which is satisfactory for a given powder. A sufficiently large number of particles must be measured so that the size distribu- tion obtained is representative of the original powder. Usually at least 200 particles are measured. If the size distribution is very wide there may be hundreds or thousands of small particles for each of the large ones. In this case it is often helpful to measure all the larger particles in a relatively large area with a low-power objective and all the small particles in a much smaller area with a high- power objective; the results are combined arithmetically for any arbitrary area. The number of fields which should be ex- amined depends largely on the number of particles per field. The choice of fields can be random or according to a pattern, but should be made prior to observing the fields through the microscope in order to avoid unconscious preferential selection. When measuring the sizes of particles whose diameters are close to the theoretical limit of resolution of the microscope, proper alignment and illumination are particularly important. Thus, when illuminating with transmitted light ("brightfield" illumina- tion) critical or Kohler illumination should be used to obtain the full benefit of the capa- bilities of the microscope. Images of particles observed with the mi- croscope are larger than the particles by an amount approximately equal to the limit of resolution of the microscope. The amount of this enlargement varies somewhat with the illumination, the type of material, and the ability of the observer to distinguish various degrees of contrast. Some microscopists sub- tract the limit of resolution from the meas- ured diameters when determining particle sizes which are close to the limit of resolu- tion. The thickness of particles can be measured with the fine-focusing adjustment. The grad- uations on this adjustment must be cali- brated and this is readily accomplished with small beads or cylindrical fibers. The diame- ter of the bead or fiber is measured with an ocular micrometer and the microscope is focused on the glass slide and then on the top of the fiber or bead, noting the readings on the fine adjustment. The difference in readings corresponds to the diameter. Ob- viously the thickness of an irregular object can then be obtained by focusing on the slide and on the top of the object and noting the difference in readings. Many special techniques are available for determining particle sizes and size distribu- tions. Electronic methods have been devel- oped for scanning the image formed by a microscope. For example, a television rastor can be placed at the ocular of the microscope so that a minute spot of light scans the parti- cle field on the microscope shde, while the 465 LIGHT (OPTICAL) MICROSCOPY transmitted light is picked up by a photo- sidered to be the loose aggregates, the "ulti- multiplicr tube. A similar result can be mate" particles of which the aggregates are achieved by illuminating the sample with, a composed, or even individual mineral grains very narrow fixed beam of light and scan- in the ultimate particles, ning the sample by automatically moving The term particle size is also usually am- the stage. The beam, after passing through biguous unless defined for each type of ap- the sample, passes through the microscope plication. Particle size is usually described and into a photomultiplier tube. In each in terms of rather artificially defined "diame- case the signal from the photomultiplier ters" which generally fall in one of two tube is analyzed electronically and presented classes. Diameters of one class, called sta- as a size distribution. tistical diameters, are defined in terms of the Equipment has been developed for ob- geometry of the individual particles and are taining photomicrographs of aerosol parti- determined for a large number of particles, cles without having to remove them from Definitions of some statistical diameters are: the air or other gas. Actual images of the (1) Martin's diameter: the distance be- particles are obtained in one technique, while tween opposite sides of the particle, meas- in others diffraction patterns are obtained ured crosswise of the particle, and on a line whose intensities are functions of the particle bisecting the projected area. sizes. Such methods are especially useful for (2) The diameter of the circle whose area obtaining the size distributions of particles is the same as the projected area of the such as fog droplets which might be changed particle. during a collection process. (3) The shorter of the two dimensions ex- Particles which are smaller than the limit hibited. of resolution of a microscope can be observed (4) The average of the two dimensions ex- as unresolved dots of light using the ultra- hibited. microscope. A mean particle size can be de- (5) The average of the three dimensions termined by suspending a known weight of of the particle. powder in a known volume of some liquid (6) The distance between two tangents to and determining the number of particles per the particle, measured crosswise of the field unit volume. A hemacytometer- type cell is (for example, of a microscope), and per- often convenient for holding the suspension pendicular to the tangents, during the counting process. The bottom of Diameters determined by sieving are also the cell is ruled off into squares and "snap statistical diameters. counts" are made of whether there are none. Diameters of the other class are defined in 1, 2, 3, etc., particles per square at any mo- terms of the physical properties of the parti- ment. cles. Examples are diameters determined by Theory and Statistics of Analysis.* sedimentation or elutriation. The word particle, in its most general sense. Numerous definitions of the mean particle refers to any object having definite physical size of a powder may be used, the choice de- boundaries, without any limit with respect pending largely on the use to which the data to size. Particles varying from about 0.01 to are to be applied. It is convenient to define 1000 microns are considered in this discus- such means in terms of data which have sion. The term is often ambiguous unless been classified into groups (classes) which carefully defined for a specific case. For ex- are defined by means of particle size limits ample, the particles in a soil may be con- called class boundaries. The midpoint of *From "Encyclopedia of Chemistry Supple- each interval is called the class mark (rf,), ment," Reinhold, New York. 1958, p. 210. the number of particles in each interval is 466 PARTICLE SIZE AND SHAPE MEASURE>IENTS AND STATISTICS called the frequency (/»•), and the total num- ber of particles measured is denoted by n. The following table is a summary of the defi- nitions of various means. Name Symbol Definition Arithmetic mean d - Z dj, n Geometric mean d. dJn)Un n (fi\] -1 Harmonic mean dna d. n \dij Mean surface di- i/^r ameter r "• Mean weight di- ameter dw /Zfidiy^ Linear mean di- ameter dx Zfidi' Zfidi Surface mean di- ameter d,. Zfidi' Zfidi' Weight mean di- ameter d'wm Zfidi^ Zfidi' The median and the mode are also often used as a measure of central tendency. The median is the middle value (or interpolated middle value) of a set of measurements ar- ranged in order of magnitude. The mode is the measurement with the maximum fre- quency. The mode is sometimes reported with the arithmetic mean or the median to give an indication of the skewness of the distribution. Some quantitative indication of particle shape is often desirable. Harold Heywood has suggested the following method. The particle is assumed to be resting on a plane in the position of greatest stability. The breadth (J5) is the distance between two parallel lines tangent to the projection of the particle on the plane and placed so that the distance between them is as small as possible. The length (L) is the distance between parallel lines tangent to the projection and perpendicular to the lines defining the breadth. The thickness (T) is the distance between two planes parallel to the plane of greatest stability and tangent to the surface of the particle. Flakiness is defined as B/T and elongation as L/B. The sizes of particles in a powder or other particulate system may be represented by some mean value, as indicated above. Usu- ally some indication of size distribution or spread is desirable and can be provided by the standard deviation, / -^{di- dYSi, n or by means of quartiles if the median is used as the measure of central tendency. Quartiles are the values completing the first 25% and the first 75% of the values when they are arranged according to increasing order of magnitude. However, it is often necessary to provide a more complete indica- tion of the particle size distribution. Classi- fied data can readily be represented by a bar type of graph called a histogram. The posi- tion of the bar locates the class interval and the length of the bar corresponds to the fre- quency. Suppose that in preparing the histogram an infinite number of particles are measured and the class intervals made to approach zero. The ends of the bars would become a smooth curve, called the size frequency curve. The function describing the curve is the distribution function. Probably the most familiar distribution function is the normal distribution, defined by the equation where s is the standard deviation. The stand- ard deviation for particles which are dis- tributed normally is such that the interval 467 METALLOGRAPHY (d — s) to (d + s) includes about 68 % of the particles. Other functions which fit distributions often encountered are (1) the log normal distribution : /(d) = log Sg-\/2ir exp ■ _ / \ogd - log dg\ ^ \ log Sg ) (2) the Rosin-Rammler function: dio -— = lOOcM^-'exp {-hd'), d{d) (3) fid) = a exp i-bd'), and (4) fid) = ad-' where Sg is the standard deviation of the logarithms of the diameters, dw/d(d) is the weight falling within a narrow size range d{d), and a, h, and c are constants. Cumulative curves are often used to rep- resent particle size distributions graphically. These are curves which result from plotting the percentages of particles greater than (or less than) a given particle size against the particle size. The ordinate can represent the percentage of total surface or of total weight instead of the percentage of the number of particles. When the diameters are normally or log normally distributed the cumulative curves are straight lines when plotted on ''probability" or "log probability" graph paper, respectively. Such paper is commer- cially available. A very large number of methods have been proposed for determing the sizes of particles, but most of them are based on one or more of a small number of principles. Optical microscopy is often used for particle size determination. It has the advantage that the particles are observed directly and that the basic equipment is relatively inexpensive. However, it is a very tedious method if ac- curate results are to be obtained and is sub- ject to large sampling errors unless adequate precautions are taken. Richard D. Cadle PLASTICS. See GENERAL MICROSCOPY, p. 390. PROJECTION MICROSCOPES See ENGINEER- ING MICROSCOPES, p. 438. PULP AND PAPER. See GENERAL MICROS- COPY, p. 394. SCHMALTZ PROFILE MICROSCOPE. See EN- GINEERING MICROSCOPES, p. 438. STEREOSCOPIC MICROSCOPE. See ENGI- NEERING MICROSCOPES, p. 439. Metallography INDUSTRIAL RESEARCH, APPLICATIONS TO. See GENERAL MICROSCOPY, p. 363. TRANSMISSION ELECTRON MICROSCOPY OF METALS— DISLOCATIONS TION. See p. 291. WEAR AND LUBRICATION. MICROSCOPY, p. 308. AND PRECIPITA- See ELECTRON MiCROMETRON AUTOMATIC MICROSCOPE The micrometron is a new microscope images. It is often necessary to scan sys- which provides a fully automatic stage tra- tematically over large areas of tissue section verse mechanism combined with accurate in order to reveal a simple central truth rela- photoelectric observation of the microscopic tive to normal or pathological state, or to 468 ml croscope 1 mage dl ssectf opt 1 ca 1 head □- r - rrul t ! pi ler phototubes pulse amp I I f ler channel s Digitizing Single Pulse Selector i pul se counter Fig. 1 MK.KORADIOGRAPHY pulse converter pulse width a n al yz e r multi- channel pulse width recorder count a number of blood cells, for example, in a much larger field than is provided by observation of a single area through the microscope. As several authorities have pointed out, large numbers of single, quantitative ob- servation events which can be gathered at great speed by an automatic microscope have a total content of significant infor- mation that a human observer simply can- not obtain in any reasonable length of time. The basic principle of operation is a quantitative digitized pulse enclosing the whole desired element of measurement within a single instrument-selected or in- strument-processed pulse, which becomes the information signal. The advantage of the digital-computer automatic microscope lies in the fact that it can measure, compute and interpret at the rate of thousands of single observations per second as compared with 50 or 100 observation events that the microscopist can study in a reasonable period. The operation of these electronic circuits is a consequence of the radiation- observing properties of Geiger, proportional and scintillation counters. The block dia- gram of the first commercially produced instrument, the Micrometron (Optronic Re- search, Inc., Cambridge, Mass.) is shown in Figure 1. Routinely it can count 10,000 blood cells in 10 to 20 seconds. When used as a counter it can take the place of 6 to 10 hematology technicians with an accuracy equivalent to the averaged result of the count of 20 hu- man observers working simultaneously on the blood specimen from a single patient. As a research instrument it can establish unequivocal diagnostic indices or basic standards of toxicity levels, or study fluores- cence of nuclei of cells or their spectral ab- sorbance under widely varying conditions. Unquestionably also a wide variety of in- organic and industrial materials may be routinely observed, measured and controlled by such a technique. In the words of Rovner, one of the developers of the micrometron, "It is exciting to realize that the individual blood cells have the quahties of a digitized scanning modality, probing into contact with every living cell in the human organism, and that we can now study them in large enough numbers to learn a great deal more of what they are trying to say." REFERENCE 1. Rovner, Leopold, "Microscopy: Automatic," Medical Physics, Vol. Ill, p. 369, Year Book Publ. Inc., 1960. G. L. Clark Microradiography. See x-ray microsocopy, p. 56i 469 Optical mineralogy (petrographic microscopy) Optical mineralogy utilizes the theory and mineral specimens of identical gross chemical methods of optical crystallography for the composition, but with different thermal his- description, classification, and identification tories, commonly show noteworthy differ- of opaque and nonopaque minerals, espe- ences in their optical properties, cially by observations under the polarizing Very few minerals are "pure" in the sense microscope. Prior to about 1900 the polariz- that the chemical composition can be ex- ing microscope was used principally for the pressed by a formula in which the ratios of determination of the texture and mineralogy the numbers of the atoms can be expressed of rocks, and emphasis was placed on the in simple, whole numbers. Isomorphism is identification of minerals by properties that common in most minerals and comes about could be measured in thin sections mounted from mutual substitution (diadochy) of on glass slides. Since 1900 instrumentation atoms or groups of atoms in crystal struc- and techniques have improved rapidly, and tures. For example, monoclinic amphiboles examination of minerals under the polarizing have a generalized formula which can be microscope by a variety of methods has be- expressed as WsCX, Y)5Z8022(OH, F)2 , in come routine procedure in many fields of which W = Ca and Na; X = Mg, Fe", and science and technology. Mn; Y = Al and Fe'"; and Z = Si and Ti. An ultimate purpose of optical mineralogy Each element contributes more or less to the is the accurate correlation of optical proper- optical and other properties depending on ties of all minerals with crystal structure, its relative amount in a crystal, the proper- chemical composition, and other physical ties of the individual atoms, and their posi- properties, and to arrive at an understanding tions and manner of coordination in the of the influence of various physicochemical crystal structure. factors on the optical properties. With such Correlation of optical properties with information the optical technique is a valu- chemical composition and physical state able tool for identification of mineral sub- differs in complexity from one mineral series stances, estimation of chemical composition, to another. Quartz, for example, has a and determination of the natural history of crystal structure that, within ordinary tem- crystallization. Optical studies, together with perature ranges, is incapable of accepting x-ray examinations and crystallochemical in- foreign atoms except in trace amounts. Ac- vestigations within a large range of pres- cordingly, the optical properties of quartz sures and temperatures provide a complete are constant from specimen to specimen, and and powerful approach to the study of the are diagnostic. Minerals in which isomor- nature and genesis of minerals. phism (diadochic substitution) is character- The physical properties of a mineral in istic show wide variations in optical proper- equilibrium with its environment are con- ties. Some appreciation of the contributions stant and correlate directly with the chemi- of individual elements or radicals to optical cal composition. The physical properties of a properties is gained from a study of the spe- mineral of the same chemical composition cific refractive energies (Jaffe, 1956) or which is not in equilibrium differ from those specific refractive capacities (Allen, 1956), of the equilibrium mineral to an extent de- particularly as the amounts of the elements pending on the departure of the atomic con- or radicals in crystals influence the refractive figuration of the crystal structure from that indices, for the equilibrium state. For example, two A convenient method for showing how op- 470 OPTICAL MINERALOGY tical properties vary with composition is afforded by plotting data of various kinds on charts. In simpler mineral series, such as the olivine series, [(Alg, Fe)2SL04], the com- position of a particular member of the series can be expressed in weight or molecular pro- portions or percentages of Mg2Si04 and Fe2Si04 . Using Mg2Si04 and Fe2Si04 as "end members" simple diagrams can be pre- pared which show how the optical properties vary with composition. Mineral series for which compositions can be expressed in terms of three "end members" are plotted on triangular diagrams, and so on. However, mineral series with more than four end mem- bers refiuire polyhedral plots, and the diffi- culty of plotting data and interpreting diagrams becomes very great, if not almost insurmountable. In complex mineral series an effort generally is made to group similar end members in such a way that data can be plotted on simple diagrams, and it is un- derstood that compositions determined from optical data plotted on the diagrams are only approximations. Laboratory investigations in optical min- eralogy are of three general kinds: (1) study of crystals or fragments in liquid immersions, (2) examination of single crystals or aggre- gates in petrographic thin sections ground to a thickness of 0.03 mm and mounted on a glass slide, and (3) examination of opaque substances in polished sections by reflected light, together with systematic etching and microchemical analysis. All techniques em- ploy the polarizing microscope as more or less modified for a specific purpose. The theory and practice of the immersion method (Wahlstrom, 1960), the thin-section method (Moorehouse, 1959), and the polished-sec- tion method (Short, 1940, and Freund, 1954) are reviewed in several books and in numer- ous articles in chemical, mineralogical, and geological journals. Summaries of properties of nonopaque minerals (Winchell and Win- chell, 1951; Larsen and Berman, 1934) and of opaque minerals (Short, 1940) are found in standard referances and handboods. The fundamental optical properties of nonopacjue minerals are the principal re- fractive indices measured for one or more reference wave lengths of light, the crystal- lographic orientation of the vibration direc- tions of light corresponding to each of the principal refractive indices, and the amount and manner of absorption of light for various directions of vibration and transmission of light by a crystal. All other optical prop- erties can be calculated or predicted if the fundamental properties are known. The immersion technique permits accur- ate measurement of refractive indices by comparison of the mineral with liquids of known refractive indices, observation of the manner of absorption of transmitted light, and, under favorable circumstances, de- termination of optic orientation. Other properties can be calculated or can be meas- ured with special optical accessories. In special applications of the immersion tech- nique crystals or fragments are observed on special multiaxis auxiliary stages that per- mit rotations of preparations into any de- sired position. Valuable additional equip- ment is a monochromator for varying the wave length of the source of illumination, and temperature control apparatus. In the thin-section method refractive in- dices are not measured directly, but optic orientation and differential absorption of light can be determined by direct observa- tion. Other diagnostic properties are noted or measured with the aid of a variety of op- tical measuring devices. As in the immersion method a multiaxis stage assists in many measurements. Polished sections of opaque minerals are examined in reflected light from a vertical illuminator attachment. Various optical ef- fects are noted, and properties such as re- flectivity, color, and rotation of polarized light may be estimated visually or measured with spectrophotometric equipment. In prac- 471 OPTICAL MINEKALOGY tice mineral identification is expedited by merits such as Chayes point counter or Rosi- systematic etch tests with corrosive reagents wal stage. and by microchemical and x-ray analysis of (7) Index liquids for immersion studies. powder scratched from the surface of the A complete set includes liquids in the index polished section. range 1.40 to 2.0, approximately. To stand- For simple determination of minerals by ardize liquids an Abbe-type refractometer optical techniques apparatus requirements with temperature control suffices for the in- are not great, but for investigations of new dex range 1.40-1.80. Higher index liquids are minerals or for the purpose of establishing standardized in a hollow prism mounted on a accurate correlation of optical properties spectrometer. with composition, crystal structure, and (8) Equipment for sample preparation and genetic history, extensive facilities are re- physical analysis. Included should be all ap- quired. A well-equipped laboratory for the paratus necessary for cutting, grinding, and examination of minerals by the methods of polishing thin sections and polished sections mineralogy, including optical mineralogy, and placing in appropriate mounts for micro- should include the following: scopic examination by transmitted or re- (1) Polarizing microscope with rotatable fleeted light. Crushing equipment and screens stage, attachable 4-axis and 5-axis universal or elutriation equipment for particle-size sep- stages, and attachable vertical illuminator aration or analysis are useful. for examining pohshed surfaces of opaque (9) Equipment and reagents formicrochem- substances. Appropriate optical accessories ical analysis, particularly for use with pol- and lenses. ished sections of opaque substances. (2) Temperature control equipment. This (10) Miscellaneous laboratory equipment, should include apparatus for varying tem- including reflecting goniometers for crys- perature of stage and hemispheres of uni- tal-angle measurement. Stereographic and versal stage and a heating stage that can be equal-area nets for plotting crystallographic attached to the microscope stage. A high- data, including optical data, temperature furnace may be useful for (11) For correlation of optical data with studies of certain thermal effects. chemical composition and crystal structure (3) Light sources. Included should be a facilities should be available for quantitative white light source, a monochromator, and a chemical analysis, spectrographic analysis, series of color filters. Also useful are a so- and x-ray study of single crystals and pow- dium-vapor lamp, mercury-arc lamp, and a ^ers. Occasionally useful are differential hvdroo-en-discharo-e tube thermal equipment, an electron microscope, (4) Spectrophotometric eciuipment for ^^^^ ^^ electromagnetic or electrostatic sep- analyzing colored light transmitted by '^'^*^', ^^' 'Reparation of mineral fractions. mi • ■ J. 1 111 1 (12) For correlation of optical properties microscope. Ihis equipment should be de- .,, , . , . • -.i , , , ., . r ^^■ With complex miueral SBries, comparisoii With signed also to permit measurement oi ellip- ^i • i • i • ^10^.1- ,. ., „ . , _ , ,. , ^ synthesized minerals is useful. Synthesis ticity of transmitted or reflected light, meas- , • , ui <• '^ ' commonly requires apparatus capable 01 gen- urement or rotation of plane of vibration of ^^^^-^^^ ^igh pressures and temperatures and, reflected light, and measurment of reflectiv- ^t times, withstanding the effects of corrosive ity of opaque specimens. fluids (5) Photomicrographic apparatus suitable for photographing specimens by transmitted , ^ , ^ , ,. , ■ Allen, R. D., "A New Equation Relating Index or reflected light. ^^ Refraction to Specific Gravity," Am. Min- (6) Counting or measuring stage attach- eralogist, 41, 245-257 (1956). 472 PETUOGRAPIIIC THIN SECTIONS Freund, H., "Handbuch der Mikroskopie in der Technik," Vol. II, Umschau Verlag, Frank- furt, a. M., 1954. Jaffe, H. W., "Application of the Rule of Glad- stone and Dale to Minerals'-, Am. Miner- alogist,^!, 757-777 (1956). Larsen, E. S. and Berman, H., "The Microscopic Determination of the Nonopaque Minerals," U.S. Geological Survey, Bull. 848, 1934. MooREHousE, W. W., "The Studj- of Rocks in Thin Section," Harper and Bros., N.Y., 1959. Short, M. N., "Microscopic Determination of the Ore Minerals," U.S. Geological Survey, Bull. 914, 1940. Wahlstrom, E. E., "Optical Crystallography," 3rd Ed., John Wiley and Sons, N. Y., 1960. WiNCHELL, A. N. AND WiNCHELL, H., "Elements of Optical Mineralogy, Pt. II," "Descriptions of Minerals," John Wiley and Sons, N.Y., 1951. Ernest E. Wahlstrom PETROGRAPHIC THIN SECTIONS The industrial chemist often has to deal with hard or brittle solids which are essen- tially transparent. In the microscopy of such solids, the ground thin sections developed by the petrographer for the stud}^ of rock specimens (1) may be used to good advan- tage, either alone or in conjunction with other sampling methods (2). Such thin sections are particularly useful in developing the follow- ing information: number and population of the various species which are present; the structure of the specimen and the spatial distribution of the entities of which it is composed; evidences of physical and chem- ical changes which have occurred in the specimen; microscopic characteristics of the specimen in specific orientation. Typical industrial materials which are well adapted to thin section microscopy are: abrasive wheels, coal, charcoal, ceramics, re- fractories, clinkers and slags, glass, resins, scales and deposits, pelleted and molded materials, catalysts, caked and cemented solids, and solid raw materials. Literature references to several of these applications are given by Chamot and Mason (2). The petrographer's standard procedure for thin section preparation involves the fol- lowing basic steps: preparation of a slice or chip of suitable area and i^ inch or less in thickness; grinding of one side of the slice to provide a smooth, though not polished, surface; cementing of the smooth surface to a glass slide; grinding of the other side of the slice with progressively finer abrasive until the desired thickness (usually .02-.03 mm) is reached and the free surface is smooth; cementing of a cover glass on the section unless it is desired to polish the surface and combine polished section and thin section microscopy in the same slide (3). Poorly con- solidated materials are embedded in a suit- able resin before grinding and either water or a hydrocarbon, depending on the solu- bility characteristics or the specimen, is used as a grinding lubricant. Typical basic equip- ment includes: a diamond or carborundum saw, a variable speed mechanical grinder with interchangeable lapping plates for coarse and fine abrasives, a flat glass plate for fine lapping, and an adjustable hot plate or oven. Rogers and Kerr (1) describe these techniques in detail and Chamot and Mason (2) list a wide variety of modern cementing and embedding media. The petrographer's standard procedures may usually be applied to anj^ thermally stable industrial materials. Modification, as dictated by the specific case, is required for extension to organic and thermally unstable inorganic samples. This involves specific choice of the combination of cementing and embedding media, abrasive, lubricant and general conditions based on the chemist's knowledge or best guess concerning the inter- action of these factors with the sample. For instance, excessive temperatin-es must be avoided during grinding and handling of the specimen and a lubricant must be chosen which will not dissolve it. Thin sections of samples which are soft enough to retain abrasive during the customary w^et grinding but which, for any reason, may not be sec- 473 OPTICAL MINERALOGY tioned by microtomy may sometimes be and Wolff (9) describe the use of thin sec- ground suitably by substitution of a gradu- tions in the study of the degeneration of ated series of garnet and emery papers for alumina-silica firebrick during use in the the usual lubricated, powdered abrasives, thermal cracking of natural gas to hydrogen The use of a low curing temperature and a and carbon. Free and combined silica in the somewhat viscous cementing agent will bricks were being reduced to volatile silicon minimize solution of the specimen. For in- monoxide, leaving corundum as a granular stance, a liquid epoxy resin, such as Shell residue. Changes in the pattern of the attack "Epon" 815 with a room-temperature curing from level to level in the checkerwork as re- agent such as diethylenetriamine , may be vealed by thin sections indicated carbon, used for the embedding and cementing at rather than hydrogen or carbon monoxide, room temperature of materials such as urea to be the principal reducing agent, and Bisphenol A which would readily dis- Other typical industrial applications of solve in and react with the resin at higher thin section microscopy, taken from actual temperatures. Some compromise with desired experience, are described briefly below, refractive index and penetration of the spec- Catalyst Degradation Studies. A clay- imen may be involved. bonded, diatomaceous catalyst carrier was The petrographer's use of the standard subjected to a corrosive environment in thickness of .02-.03 mm for his thin sections which blackening of the pellets, due to ear- not only insures adeciuate transparency for bonization of organic reactants, and eventual most nonopaque rock specimens but is also disintegration of the pellets took place. Ex- indispensable to his direct identification of amination of the used pellets in thin section mineral species in the section (1). In many revealed distinguishable crystalline species, industrial studies, particularly those in resulting from attack of the clay binder, which the species are identified or charac- which w^ere not present in the virgin catalyst, terized through microscopy of separate, pow- The species were either identified or charac- dered portions of the specimen or those in terized by microscopic examination of the which only the structure of the sample is of powdered, used catalyst and further details interest, the thickness of the section is not concerning the degradation were developed critical and may be dictated only by the through systematic study of thin sections of transparency of the material and the diame- pellets taken from selected locations in the ter of the smallest feature which is of inter- catalyst bed. Clusters of carbon were defi- est. In some cases the section may be 0.25 nitely associated with an iron-containing mm or more in thickness. On the other hand, crystalline species and the disintegration of for semi-opaque materials such as cements the pellets was tentatively tied to the ob- (4) and filled plastics (5, 6) the section must served disparity between the large diame- be as thin as .01-.015 mm; in such cases the ters of the clusters of new crystals and the microscopist may elect to use, w^here applica- smaller diameters of the skeletal pores, ble, the metallographer's polished sections Deposition of Silica Gel on Catalyst or, for soft materials, the petrographer's Carrier. Attempts to deposit a uniform coat- "peels" (7). ing of silica gel on the component particles The ceramics and refractories industries of a catalyst carrier consisting of bonded have probably drawn more heavily than quartz fragments yielded a product which other industiies upon petrographic thin sec- did not have the desired surface properties, tions in their microscopic studies. Tj^pical Examination of thin sections of these grains applications are illustrated by Insley and revealed that the deposited silica gel was Frechette (4) and by Norton (8). Wright merely plugging the large external pores of 474 PETKOGRAPHIC THIX SECTIONS the grains and was not distributed uniformly over the full carrier surface (Figure 1). Structure of Urea Prills. In one case, the reason for the abnormal, weakness of urea prills was sought. Thin sections revealed that the crystal bundles of which the indi- vidual prills were composed did not, in the case of the weak prills, meet along uniformly narrow contact zones in the normal manner but were irregularly separated by relatively wide void spaces (Figure 2). This observa- tion could be correlated with operating con- ditions. In another case, the prill structures of a number of competitive brands of urea were examined by means of thin sections. In three brands, round cross-sections and com- pact structures characteristic of prills cooled directly from the melt were observed. In the fourth brand, the cross-section was irregular and the rather loosely bonded urea crystals showed a strong peripheral alignment ; it was deduced that these prills were probably formed by mechanical means. Examination of Heat Exchanger De- posit. A thick scale removed from a heat exchanger tube was found through thin sec- tion microscopy to consist of a single, readily identifiable species in definite, even bands. The interpretation of the observed structure was that the scale was deposited during dis- crete intervals between which neither erosion nor deposition of scale occurred. These ob- servations could be correlated with plant operating conditions. Internal Structure of Crystals. In an investigation of the effects of environmental conditions on the crystallization of am- monium sulfate, internal twinning and other imperfections were studied by means of thin sections. Strain Patterns in Resin Test Pieces. Thin sections were helpful in the study of strain patterns in deformed resin test pieces, since (1) strain patterns which were too in- tense for resolution in the full thickness of the specimen were easily resolvable at re- duced thickness and (2) patterns could be Fig. 1. Silica gel plugging. Coarse catalyst car- rier pores. (XlOO). ^t^v^c;^ Fig. 2. Mechanically weak urea prills. (X40). Fig. 3. Strained Izod specimen — end of notch. Ordinar}^ transmitted light. (X14). studied in any desired orientation of the specimen. Figure 3 shows, in ordinary trans- mitted light, a portion of a section approxi- mately .25 mm in thickness of the notch area of a notched Izod impact test specimen 475 PHASE MICROSCOPY Fig. 4. Same field as Figure 3. Polaroids crossed. 3 of transparent resin which had been placed in the Izod test vise and hit with the test hammer only hard enough to deform the specimen slightly without breaking it. Ob- serve (1) the incipient tears at the notch corners and (2) the two sets of fracture cracks on the left. Figure 4 shows the same field between crossed Polaroids. The intense strain pattern, due to plastoelastic deforma- tion, is divided into two sections, each of which converges on one of the incipient tears at the notch corners. The fracture cracks extend into the strain patterns but distort them only very slightly. In the full 3^^^-inch thickness of the specimen, these features were not readily resolvable but the adjoining strain patterns of lower order due to elastic deformutiun, which were obliterated in grind- ing to the .25-mm thickness, were fairly leadily resolvable. At half-thickness of the specimen, the latter patterns were quite clear and the former was just beginning to emerge. REFERENCES 1. Rogers, A. F. and Kerr, P. F., "Optical Min- eralogy," McGraw-Hill, New York, 1942. 2. Chamot, E. M., and Mason, C. W., "Handbook of Chemical Microscopy," Chapter 5, John Wiley & Sons, New York, 1958. Kennedy, G. C, "The Preparation of Polished Thin Sections ," £;con. (?eoL , 40, 353-59 (1945) . 4. Insley, H. and Frechette, V. D., "Micros- copy of Ceramics and Cements." Academic Press, New York, 1955. 5. RocHOw, T. G., "Microscopy in the Resin In- dustry," Ind. Eng. Chem. Anal. Ed., 11, 629- 34 (1939). 6. RocHOW, T. G. and Gilbert, R. L., Chapter on "Resinography" in Volume V of Matiello's "Protective and Decorative Coatings," John Wiley & Sons, New York, 1946. 7. Weatherhead, a. V., "A New Method for the Preparation of Thin Sections," Mineralog. Mag., 25, 529-33 (1940). 8. Norton, F. H., "Refractories," McGraw-Hill, New York, 1946. 9. Wright, R. E. and Wolff, H. I., "Refractory Problems in the Production of Hydrogen by Pyrolysis of Natural Gas," /. Am. Ceram. Soc. ,31, 31-38 (1948). R. E. Wright Phase microscopy (See also pp. 365, 452) ANOPTRAL MICROSCOPE* In 1944 the author began searching for possible means of detecting more clearly in the microscope the structure of living cells and tissues. Most excellent microscopes have been in use for nearly a hundred years, * Exerpt from paper in Mikroskopie, 9, 1954, Nos. 1-4, pp. 80. and they reproduce very well the structure of objects which are fixed and stained and usually embedded in Canada balsam. How- ever, there are certainly many unknown fundamental facts that can only be revealed by studying the living preparation directly. Life is sometimes characterized much better by motion, growth, propagation and ex- change of the cellular constituents than it is 476 ANOPTRAL MICROSCOPE •s 'c ^ f^^ > . A< < Fig. 1. In the left-hand picture living yeast cells were photographed by means of an ordinarj^ micro- scope objective (Reichert achromat 60X); in the picture to the right the same specimen is shown with the objective treated as described in the text. In the latter picture the cells with a good nutritional status appear dark while others remain more or less pale (old, degenerating, poorly nourished cells). There are no halo effects around the cells as there would be with an ordinary phase contrast. by the multiplicity of structures. There are, however, difficulties in the way of studying unstained preparations. All those who have worked with the micro- scope must be aware of the fact that a nor- mal stained biological preparation provides an excellent picture when viewed through a microscope with large objective and con- denser apertures, whereas in the absence of staining very little if anything can be seen since the details are only distinguishable then by their refractive power. Not until the illuminating aperture has been reduced, e.g., by stopping down the iris diaphragm, do the details begin to appear one after an- other, but even then they are surrounded by concentric diffraction fringes which have a detrimental effect in that they reduce the resolving power. Nor does dark-field microscopy help very much. Abrupt changes in the refracting properties of the details of the object cer- tainly show up clearly, but small differences or gradual changes in the refractive index remain invisible. These drawbacks led the author to under- take extensive experiments with the "schlie- ren" method discovered by Topler nearly a hundred years ago, and to apply them to microscopy. It was soon discovered that an annularly defined aperture in both the condenser and the objective was the most suitable one. The realization of annular de- fined apertures in objectives, how^ever, was very difficult. After experimenting with thin coatings of silver, without much success, the author tried coatings of soot applied by re- peatedly exposing the surfaces of an objec- tive lens to a candle flame. A microscope modified on the above lines provided much better images of living speci- mens viewed through it. The obvious course, then, was to proceed on purely empirical lines. ^ It was soon evident that excellent re- 1 It is interesting to note that Leeuwenhoek was not familiar with the theory of optics. He was 477 PHASE MICROSCOPY Fig. 2. The first contrast pictures (Fig. 1.) were made by using a circular aperture in the con- denser and a likewise circular, clear area on a lightly-sooted objective lens surface. The left hand picture shows the path of light in the system; above and below are the objective and condenser apertures in a perpendicular view. In the picture to the right is shown the anoptral method devel- oped later. The non-diffracted rays penetrate a heavily light-absorbing, non-refiecting area in the objective. suits could be obtained by using an annular diaphragm opening in both the condenser not college-trained, understood only Dutch, and did not even write his own language like an edu- cated man. The microscope consisting of objective and eyepiece had been discovered and used before Leeuwenhoek was born, but he never possessed one. He saw blood corpuscles, sperm-cells and even bacteria with his instrument, which con- tained only a single tiny lens. Yet he was able to make such a number of microscopical observations which he communicated to the Royal Society of and the objective, the transmission of the objective diaphragm being 50 per cent. Liv- ing specimens such as 3'east cells appeared very clear and sharply defined with this sys- tem, and had an agreeable brownish tint on a bright background. The cells showed darker the more the refractive power of their content exceeded that of their surroundings, i.e., the better their state of nourishment (Fig. 1). Later the author found that the phase- contrast microscope, developed on a theoret- ical basis by Zernike, had been constructed by the Zeiss Optical Works. When the author looked into a phase-contrast microscope for the first time he was surprised to note the similarity of the images obtained with it and with his own ecjuipment. Both of them cer- tainly had annular condenser apertures, but the phase-contrast objectives had light - absorbing annuli whereas the author's had an absorbent coating recessed in the form of a ring. But the effect produced by the images was not identical. Round each light-refract- ing detail the phase-contrast microscope ex- hibited haloes which had no counterpart in reality. The author's microscope showed none of these haloes, and the gradation of the different intensities of contrast, depend- ing on their light-refracting properties, was agreeably soft. During his first experiments the author made an observation which was afterwards to prove significant. If the coating of soot in the objective was not recessed in the form of an annulus, but complementarily applied so that the light coming from the condenser diaphragm passed through this annulus, an image was obtained which exhibited the op- posite properties (Fig. 2). The author re- sumed his experiments in this direction in London that the Dutch have recentlj^ published them in four large volumes. Nor were the inven- tors of the first achromatic microscope objectives, Aepinus and Beelsnyder, professional opticians. The former was a Privy Councillor and the latter a cavalry officer (see also p. 458). 478 ANOPTRAL MICROSCOPE 1952, partly because the available phase- contrast microscopes had not been developed further. The halo effect in them, mentioned above, continued to prove disturbing, and often led to false interpretations. When attempts were being made to im- prove this second method it became clear that increasing the light absorption of the soot ring to over 90 per cent gave a particu- larly well contrasted image with a golden- brown background against which the details of the object usualh^ appeared bright (be- cause of their higher optical density), but were surrounded by narrow, darker zones. The image was, so to speak, the reverse of an ordinary phase-contrast image, with dark borders instead of the bright haloes. The golden-brown tint of the image was a happy chance, for it is agreeable and restful to the eye — a fact long known in art photography. The question is now: Are the dark shaded zones of the picture less disturbing than the bright haloes of ordinary phase-contrast microscopy? The answer is a definite affirmative. We recognize the shape of most objects better if we see them lighted as they usually appear in our everyday surroundings. The objects are not as a rule presented in transmitted light, or as silhouettes, but appear in our visual field illuminated mostly from above, from the side, or from several directions at the same time. The diffuse light from the sky is probably the most original and natural form for the human eye. If we look at a stone on sandy ground under this mode of illumination we can see a lightly shaded zone around it. If we saw the same thing as a negative, in which the half-shadow borders were replaced by haloes, it would be difficult to apprehend the shape of the object. The normal image of the stone on the sandy ground bears the same relation to the negative as the microscopic image produced by the author's second method has to the image obtained by '^positive" phase-con- trast. The disturbing haloes — simulating phosphorescent edges — of the latter, are con- verted into a virtue by their reversal, since this half -shadow bordering assists the apper- ceptive faculty and produces an almost plastic effect. (Fig. 3). / ^^*. svib^_ .'»: Fig. 3. Parts of two (2) epithelial cells from the oral mucosa seen by ordinary piiase contrast (a) and by the anoptral method (b). From the latter we obtain an illusion of third dimension (depth) which is due to the shadow-effects around the cell borders. Reichert achromats lOOX. Magnification 1350X . 479 POLARIZING MICROSCOPE The writer has also been able to familiarize himself with the "negative" phase-contrast objectives of different makes. The image ef- fects obtainable with them are of course somewhat similar to those obtained by the author's second method, but the pictures give rather a cold impression because of their grayish-white color accompanied by a metal- lic sheen and a strange fogginess. As the phase annuli are made by vacuum evap- orization and especially as metals are used, the light-absorbing layers tend to intro- duce reflecting surfaces into the system. Since the degree of absorption of the phase annuli must be considerable in order to ob- tain a very high degree of contrast, it is no wonder the stray light from these reflections becomes nearly equal in brightness to the image itself. The soot surface, on the con- trary, reflects very little. For the commercial production of the new objectives, the coating of soot, as it is not very resistant, must be replaced by other suitable substances which produce the same effect. Messrs. Optische Werke C. Reichert A. G., Vienna, Austria, have solved this problem by using means available only to a large firm of microscope makers. They are now manufacturing these objectives and their accessories on a com- mercial basis under the name of "Anoptral Contrast Equipment" (anoptral = nonre- flecting). A. WiLSKA FIBERS. See GENERAL MICROSCOPY, p. 343. INDUSTRIAL RESEARCH, APPLICATIONS TO. See GENERAL MICROSCOPY, p. 363. PLASTICS. See GENERAL MICROSCOPY, p. 390. PULP AND PAPER. See GENERAL MICROS- COPY, p. 394. THEORY AND MICROSCOPE CONSTRUCTION. See OPTICAL THEORY OF LIGHT MICRO- SCOPE, p. 445. Polarizing microscope (See aUo chemical micros- copy, pp. 21, 31, 52; INDUSTRIAL HYGIENE MICROSCOPY, p. 400; OPTICAL MINERALOGY, p. 470) BASIC DESIGN AND OPERATION. See OPTI- CAL THEORY OF LIGHT MICROSCOPE, p. 453. DESIGN FOR MAXIMUM SENSITIVITY General Description The polarizing microscope is a compound light microscope used for studying the aniso- tropic properties of objects and for render- ing objects visible according to their optical anisotropy. To this end a polarizing micro- scope is generally equipped with a polarizer and an analyzer (or "polars" following the terminology of Swann and Mitchison, 30), strain-free condenser and objective lenses, compensators, a Bertrand lens or a telescopic eye piece, and a graduated revolving stage. Both transmitted light and vertical illumina- tion are used and the image may be studied orthoscopically as in ordinary microscopy or conoscopically by viewing the interference pattern at the back aperture of the objec- tive. Depending on its application a polariz- ing microscope may also be called a pet- rographic microscope, a metallographic microscope, a chemical microscope, etc. Many types of interference microscopes are also essentially modified polarizing micro- scopes. The general principles and applica- tion of regular polarizing microscopes can be found in several texts (1, 2, 4 to 7, 17, 22 to 480 DESIGN FOR :MAXIMI >I SENSITIVITY 28, 32, 33; see especially 5 for general refer- ence and 33 for a thorough theoretical treat- ment of the polarizing microscope). This article will give special attention to methods and devices for obtaining maximum sensitivity with the polarizing microscope. As described later, the combined sensitivity, resolution, and image quality of the polariz- ing microscope has been vastly improved in the past few years, and retardation (coefR- o cient of birefringence X thickness) of 0.1 A unit can be detected on objects 0.2m wide. These advances now permit application of polarization microscopy to new realms such as analyses of fine structure in living cells, or experimental studies of rigorous diffrac- tion theory of the kind which could not be performed with equipment available in the past. Specification of the System for Ob- taining ^Maximum Sensitivity For the sensitive detection of birefringence (BR) and dichroism, precise measurement of retardation, optical activity (rotation of plane-polarized light), extinction angle, etc., detectable contrast must be introduced with only minimal differences in a parameter. Contrast is maximized for a given object bj^ reducing stray light in the system (11, 30), and by using appropriate compensators or half -shade plates (3, 7, 11, 13, 15, 18, 27, 28, 30, 31). The degree of success in reduc- tion of stray light, or increase in extinction factor (EF), can be expressed numerically as EF = intensity of light with parallel polars, divided by intensity of light with crossed polars (30) . The effect of various components on the EF is discussed in the next section. The ability of the eye to perceive contrast is drastically lowered at low levels of image brightness, a condition which prevails at high EF's. A maximally bright light source is therefore required. Contrast discrimina- tion can also be improved bj^ darkening the room and by masking off unwanted sources of bright Ught in the image (e.g., retardation can be measured to X/1000 with a quartz wedge when the regions outside of the dark fringes are masked). Specification of the Components Polars. ^Modern sheet polaroids (general properties outhned in 21) can, but do not always, have EF as high as 3 X 10^ for green light at normal incidence. For most work EF of this magnitude is adequate, but two addi- tional factors should be borne in mind for critical application. The parallel transmis- sion of a pair of high extinction polaroids is of the order of 10 ^ 15 %, hence a considerable light loss takes place. Most sheet polaroids have microcrystalline textures or inclusions which can cause "hot spots" affecting the diffraction image. Calcite polars coated with a low reflection film may show virtually no light loss (paral- lel transmission '=50^c) and an extremely high EF (^3 X 10«). Of the various t>T)es of calcite prisms available (17, Lambrecht 20 has supplied excellent quality prisms), the square-ended Glan-Thompson prisms ap- pear to be the most effective. The optical qualitj' and EF of calcite prisms depend to a great extent on the skill of manufacture and subsequent care of handling since the soft calcite surfaces are prone to scratching and pitting. The beam passing through the analyzing calcite prism should be made par- allel by the stigmatizing lenses to minimize the astigmatism resulting from the double refraction of calcite. Condenser and Objective Lenses. All optical components lying between the polar- izer and the analyzer must be scrupulous^ clean. Not only may any dust particle, greasy surface, etc. scatter enough light to lower the EF and seriously reduce object contrast, but by acting as loci having high transmittance, such entities would give the effect of oblique illumination and distort the diffraction image. The deleterious effect of large amounts of strain BR in the condenser and objective 481 POLARIZING MICROSCOPE (a) (b) (C) Fig. 1. Appearance of the back aperture of a 97 X 1.25 NA strain-free coated objective with and without rectifiers. The condenser, which is identical with the objective, is used at full aperture. Colli- mated mercury green light (546 ni/j) used for illumination, (a) Crossed polarizers, no rectifier, (b) Po- larizer turned 2°, no rectifier, (c) Crossed polarizers, with rectifiers in both condenser and objective. Photographs a, b, and c were given identical exposures. From Inoue, S. and W. L. Hyde (12). lenses has long been recognized and polariz- ing microscopes are generally furnished with "strain-free" lenses. The magnitude of strain BR in these lenses is frequently still considerable and for realizing the utmost sensitivity, lenses with exceptionally low strain BR must be selected. The procedure for selecting these follows. Preparation for the Test. Provide a polariz- ing microscope with a very high intensity light source, such as a high pressure mercury arc lamp (General Electric A-H6 or Osram HBO 200) or other lamp of equivalent brightness. The components of the micro- scope should be adjusted or repaired so that the EF of the system without the lenses is greater than 10^. Close down the field stop and use Kohler illumination, employing a nominally identical objective as the con- denser. It is convenient to screw the test lens into an objective holder which in turn is placed inverted on the rotating stage and centered to the optical axis. This acts as the condenser (or objective in an inverted sys- tem as in Fig. 2) and the selected mate re- mains fixed. In addition prepare a removable rotating mica compensator of the order of X/20 to be placed in the slot above the ob- jective or between the condenser and the polarizer. Test for Freedom from Strain BR. When the pair of lenses is aligned, the polars crossed and the image of the field stop properly focused a dark cross (Fig. la) should appear at the back aperture of the objective. In the absence of strain BR the cross is observable with objectives of NA as low as 0.1 wath no object lying between the condenser and ob- jective lenses. The cross must be symmetrical and very dark between crossed polars and should open symmetrically into two hyper- bola-hke fringes the "V's" (Fig. lb) when the polarizer or analyzer is rotated. The arms of the Vs should remain dark until they dis- appear beyond the edge of the aperture. If the lenses are free from local and lateral strain these dark fringes remain undistorted when the stage is rotated and the test lens is examined at various orientations. Some lenses seemingly passing the above test may still suffer from radially symmetric (strain) BR. This is checked by inserting the mica compensator. When the lenses are free from radial as well as lateral BR the cross loses contrast and fades away with little change in its shape or position when the compensator is rotated. If lenses possess ra- dial BR the cross will open into two V's as though with a strain-free lens the analyzer had been turned. Rectifiers. Lenses selected for freedom from strain by the methods described can give extremely high EF at low NA's (e.g., 3 X 10^ at condenser NA = 0.1). However, 482 DESIGN FOR MAXIMUM SENSITIVITY they still exhibit a dark cross which be- comes progressively more prominent as the NA is increased (Fig. la). Light from the areas between the arms of the cross is in- troduced by rotation of the plane of polariza- tion at glass air interfaces in the lenses and in microscope slides. The rotation is a result of differential reflection losses of the parallel and perpendicular components of polarized light and may be as large as 7° ^^ 8° for high NA objectives even after low reflection coating and oil immersion (2, 10, 33). This lowers the EF (^lO^ for NAobjective = NAcondenser = 1.25) and furthermore dis- torts the Airy diffraction pattern (14, 19). Much of the rotation and therefore light giving rise to the cross can be eliminated by the use of "polarization rectifiers" built into the objective and condenser lenses (Fig. 2R). The rectifier consists of a X/2 BR plate which reverses the above mentioned rotation and a zero power meniscus which introduces suffi- cient additional rotation to cancel out the reversed rotation (12). Suitably rectified objectives and con- densers used with a proper microscope (last section) give verj^ high EF (2 X 10^ for NAcondenser = NAobjective = 1-25) aild Uiakc detectable very small BR (<0.1 A°) of ob- jects which in themselves are barely resolva- ble wdth the light microscope (<0.2yu). The objective back aperture being uniformly dark (Fig. Ic), spurious diffraction is reduced to a minimum and the image is therefore more reliable (12, 14). Determination of the polar- ization angle and other parameters can be made with great accuracy by using rectified optics whereas such measurements with or- dinary polarizing microscopes contain in- herent and significant errors (33). Compensators. Many types of compen- sators and half -shade plates have been used for increasing the sensitivity and precision in determining the ellipticity of polarized light, polarization angle, etc. Their construction, sensitivity and operation may be found in references 3, 4, 7, 9, 11, 13, 15 to 18, 23, 24, 25, 27, 28, 30, 31, 33. As mentioned above, the property of the eye or other detectors influence their performance. Photoelectric or photographic photometry can also aid in in- creasing the sensitivity. When algebraic computation of compensa- tor actions are difficult, e.g., when more than two anisotropic components lie between the polars, advantage can frequently be taken of the geometrical construction available with the Poincare sphere or its two dimen- sional analog (16, 29). Optical Layout of a Polarizing Micro- scope with Exceptional Sensitivity. Fig- ure 2 shows the essential optical layout of a transilluminating polarizing microscope de- signed to give maximum sensitivity and image quality. The system illustrated is in- verted with the light source S on top and detectors (EM and E) at the bottom. Light from the bright source is filtered and is focused by Li and L2 onto a pinhole aperture A 2 . This restricts the size of the source image so that after projection by L3 it just covers the condenser aperture dia- phragm Ae . The polarizing Glan-Thompson prism POL is placed behind the stop A4 away from the condenser COND, to prevent light scattered by the polarizer from entering the condenser. Half-shade plates are placed at the level of A5 , and compensators COMP above the condenser. Both the condenser and the objective OBJ lenses are rectified by Ri and R2 and are low reflection coated. The image of the field diaphragm (A3 or Ag) is focussed onto the object plane OB by the condenser, whose NA can be made equal to that of the objective. Stigmatizing lenses Sti and St 2 are low re- flection coated on their exteriors while their backs are cemented directly onto the ana- lyzing Glan-Thompson prism ANAL to pro- tect the surfaces of the prism. Aperture stops Ai-As are placed at critical points to minimize scattered light from entering the image forming system. The final image is cast by OCi on a photographic or photoelec- 483 POLARIZING MICROSCOPE 1 r " ::i:r 1 FILTERS \ c ^ Li A / / POL A4 Ae Z3 i::==r:3 tzzzzii COMP A, COND OB OS«t Rt ST, ANAL -i^. e ■^ ilrirn M OC. OC. EM Fig. 2. Optical layout of an inverted transil- luminating polarizing microscope designed to give maximum sensitivity and image quality. See last section of text for details. trie sensor EM or to the eye E via a mirror M and ocular OC2 . This optical system was used for develop- ing and testing the rectified objectives men- tioned earlier and may be considered to be the basic system necessary and ample for obtaining maximum sensitivity and resolu- tion with the transilluminating polarizing microscope. REFERENCES 1. Ambronn-Frey, "Das Polarisationsmikros- kop, seine Anwedung in der Kolloid-for- schung in der Farberei," Akademische Ver- lag, Leipzig, 1926. 2. Barer, R. in Mellors, R. C, "Analytical Cytology" Chapt. 3, McGraw-Hill, New York, 1955. 3. Bear, R. S. and Schmitt, F. O., "The measure- ment of small retardation with the polariz- ing microscope," /. Opl. Soc. Am., 26, 363- 364 (1936). 4. Bennett, H. S., "The microscopical investi- gation of Biological materials with polarized light," in McClung: Handbook of Micro- scopical Technique, p. 591, P. B. Hoeber, Inc., New York, 1950. 5. Chamot, E. M. and Mason, C. W., "Hand- book of Chemical Microscopy," John Wiley and Sons, New York, 1949, second edition, Vol. 1. 6. GiBBs, T. R. p., "Optical Methods of Chemi- cal Analysis," McGraw-Hill, New York, 1942. 7. Hallimond, a. F., "Manual of the Polarizing Microscope," Cooke, Troughton and Simms, York, England, 1953. 8. Hstj, H. Y., RicHARTZ, M., and Liang, Y. K., "A generalized intensity formula for a sys- tem of retardation plates," /. Opt. Soc. Am., 37, 99-106 (1947). 9. Inoue, S., "A method for measuring small re- tardation of structures in living cells," Exptl. Cell Research, 2, 513-517 (1951). 10. Inoue, S., "Studies on depolarization of light at microscope lens surfaces. I. The origin of stray light by rotation at the lens surfaces." Exptl. Cell Research, 3, 199-208. 11. Inoue, S. and Dan, K., "Birefringence of the dividing cell," /. Morph., 89, 423-456 (1951). 12. Inoue, S. and Hyde, W. L., "Studies on de- polarization of light at microscope lens sur- faces. II. The simultaneous realization of high resolution and high sensitivity with the polarizing microscope." /. Biophys. Bio- chem. CytoL, 3, 831-838 (1957). 13. Inoue!, S. and Koester, C. J., "Optimum half -shade angle in polarizing instruments," J. Opt. Soc. Am. 49, 556-559 (1959). 14. Inoue, S. and Kubota, H., "Diffraction anomaly in polarizing microscopes," Na- ture, 182, 1725-1726 (1958). 15. Jerrard, H. G., "Optical compensators for measurement of eliptical polarization," /. Opt. Soc. A7n., 38, 35-59 (1948). 16. Jerrard, H. G., "Transmission of light through birefringent and optically active media: the Poincar^ sphere," /. Opt. Soc. Am., 44, 634-640 (1954). 484 ANGLE REFRACTOMETRY 17. JoHANSEN, A., "Manual of Petrographic Methods," McGraw-Hill, New York, 1918, second edition. 18. KoHLER, A., "Ein Glimmerplattchen Grau I. Ordnung zur Untersuchung^ sehr schwach doppelbrechender Praparate," Z. wiss. Mikroskop., 38, 29-42 (1921). 19. KuBOTA, H. AND Inoue, S. "Diffraction images in the polarizing microscopes," J. Opt. Soc. Am., 49, 191-198 (1959). 20. Lambrecht, K., See Catalogue "Polarizing Optics." (4318 N. Lincoln Ave., Chicago 18, 111.) 21. Land, E. H., "Some aspects of development of sheet polarizers," /. Opt. Soc. Am., 41, 957-963 (1951). 22. Oster, G., "Birefringence and Dichroism," in Oster and Pollister: "Physical Techniques in Biological Research," Vol. I, Chapter 8, Academic Press, New York, 1955. 23. Pfeiffer, Hans H., "Das Polarisationsmik- roskop als Messinstrument in Biologic und Medizin," Friedr. Vieweg and Sohn, Braun- schweig, 1949. 24. Rinne, F. W. B. and Berek, M., "Anleitung zu optischen Untersuchvmgen mit dem Polarizationsmikroscop," Leipzig, 1953. 25. RucH, Fritz, "Birefringence and Dichroism of cells and tissues," in Oster and Pollister: Physical Techniques in Biological Research, Vol. Ill, 149-176 Academic Press, New York, 1956. 26. Schmidt, W. J., "Die Bausteine des Tierkor- pers," F. Cohen, Bonn, 1924. 27. Schmidt, W. J., "Die Doppelbrechung von Karyoplasma, Zytoplasma und Meta- plasma," Bd. II, Protoplasma-Monogra- phien, Berlin, 1937. 28. Schmidt, W. J., "Die Doppelbrechung des Protoplasmas und ihrer Bedeutung fur die Erforschung seines submikroskopischen Baues," Ergeb. Physiol., 44, 27 (1944). 29. Skinner, C. A., "A universal polarimeter," /. Opt. Soc. Am., 10, 491-520 (1925). 30. Swann, M. M. and Mitchison, J. M., "Re- finements in polarized light microscopy," J. Exp. Biol., 27, 226 (1950). 31. TucKERMAN, L. B., "Doubly refracting plates and elliptic analyzers," Univ. Nebraska Studies, pp. 173-219 (1909). 32. Wahlstrom, E. E., "Optical Crystalog- raphy," John Wiley and Sons, New York, 3rd. Ed., 1960. 33. Wright, F. E., "The Methods of Petrographic Microscopic Research," (Carnegie Institu- tion of Washington, Washington, D.C., 1911). Shinya Inoue FIBERS (TEXTILES). See GENERAL MICROSCOPY, p. 343. INDUSTRIAL RESEARCH, APPLICATION TO. See GENERAL MICROSCOPY, p. 363. PLASTICS, See GENERAL MICROSCOPY, p. 390. PULP AND PAPER. See GENERAL MICROS- COPY, p. 394. Refraction of light, refractometry and interferometry* ANGLE REFRACTOMETRY (See also p. 400) The direct application of Snell-Descartes' law led to the design of many prismatic in- struments utilizing angular measurements — the so-called transmission refractometers. * Objections may be raised that this topic is not properly included within the framework of Mi- croscopy. However, the fundamental phenomenon of optical refraction of light in microscope lenses, the microscopic measurement of refractive index, The subsequent discovery by Augustin Fres- nel of the quantitative relationship between refraction on a surface, reflectivity by this surface and absorption by the medium, fi- and the intimate relationships between micro- scopes, refractometers and interferometers as optical instruments, completely justify inclusion of the four articles extracted from an exhaustive monograph and prepared expressly for this En- cyclopedia by Dr. Raymond Jonnard. — Ed. 485 KEFKACTIO.N OF LIGHT, KEFRACTOMETRY AND INTERFEROMETRY nall}^ led to the design of another type of re- nients embodied in the instruments hsted are fractomcter adapted to the study of opaque based also a number of methods which can be substances, namely, reflective refractome- carried out with such conventional instru- ters. The following account of this field of ments as the microscope, the goniometer and measurements is intended to serve as a guide, the photometer. These methods include: the In each case it is imperative to refer to the Becke lines displacement, the Christiansen original references, giving the exact detailed effect, the Merwin (46) and the Emmons procedure to be followed. (For references, re- methods for crystals, and the Wood and fer to article "Historical Introduction.") Ayliff method for solids in general. Transmission refractometers include: the Prismatic refractometers utilize either Abbe simple and double types (26)* from one of the two basic properties of trans- which are derived the various designs of parent media: (1) the existence of a mini- "dipping refractometers": the Pulfrich re- mum deviation angle of emergent light, or fractometer (27); the Lob instrument (28); (2) the existence of a critical angle, the Pfund total reflexion device (29) and its Instruments based upon the determina- modification by Countryman (30) ; the Ferry tion of the minimum deviation angle have hollow prism refractometer; the Jelley re- only an historical interest. Their passing into fractometer (31); the Jelley-Fischer (35) in- oblivion is perhaps unjustified, for it is now strument adapted for observations on molten possible to measure — and to record or con- substances ; the Nichols double reflection in- tinuously monitor — ^with them small angular strument (32) and its modification for micro- deflections with an approximation better by analysis (33), both of which must be used several decimal places than was possible half conjointly with a microscope; the Hilger ul- a century ago. Such instruments, being traviolet refractometer designed after the basically extremely simple, might be worth original instrument of Ch. Henri (34); the reconsidering for certain industrial applica- Fredriani instrument (36), really a modifica- tions. For these reasons, additional com- tion of Jelley 's; the recording modification of ments on the goniometric method and of that Ferry's design used by Cruikshank and Fair- of Biot and Arago are given, weather (37) in their chemical work; the Methods of Newton, and of Biot and photoelectric multiple reflection device of Arago. Only in rare instances is it possible Karrer and Orr (38) ; the Emmons spherical to measure the angle of refraction directly, refractometer for solids (39) ; the recording Such measurements, of limited accuracy, are refractometer of Barstow (40) ; the "chemical convenient when working with properly refractometer" of Barnes (41); the Jacob- shaped transparent solids. They are easily sohn instrument for opalescent fluids (42) ; effected with a goniometer, sometimes called the Stamm et al. differential recording in- a "spectrometer." The conventional goni- strument (43), using either three Wernicke ometer is described in practically every text- hollow prisms, or a double Zenger hollow book. More elaborate instruments of the one; the Clamann pneumatic refractometer "transit" type, when equipped with reflec- for gases (44), in which a beam of light is de- tion or internally illuminated auto-collima- flected at the interface of two gas phases, tion devices, are usable for this purpose, imparted with a parallel rapid laminar flow; In all cases the measurement method is the Broumberg refraction monochromator, based vipon the property of a prism of ex- which can be adapted to refraction measure- hibiting a minimum deviation angle. The ments (45) ; and a few others mentioned else- existence of this angle is a direct consequence where. of the critical refraction angle. The method On the principles of angular measure- originally designed by Newton, and by Biot 486 1 ANGLE REFRACTOMETRY and Arago is still valid. The original prism of Arago and Biot was large, yielding directly the absolute n value of gases or liquids rela- tively to vacuum. The apex angle is conveniently made 143°. (Instruments of this type are no longer com- mercially a^'ailable.) The path of light is first determined with the prism either evacu- ated or filled with dry air. In a second opera- tion, one measures with the specimen filling the prism. Refractive Index Measurement with the "Autocollimator."A small instrument constructed especially for the measurement of small angles or small defects of parallelism in mechanical parts and machines, known as the "Tuckerman Autocollimator"*, can also be used for the measurement of the refrac- tive index of transparent substances. The autocollimator consists essentially of a highly corrected objective lens, in the focal plane of which is a high-precision reticle. Part of this reticle is illuminated to form a fiduciary object. The parallel light emerging from the lens can be reflected by anj^ external plane surface back into the instrument, and the reflected image can be made to fall on the main scale of the reticle. The position of the reflected image of the reticle is a function of the orientation of the mirror, and this can be measured by means of another scale in the fiducial image, which acts as a vernier to the main scale of the reticle. An eyepiece is provided to observe the images. The distance from the mirror to the instrument does not affect the readings, so that the instrument can be used to meas- ure the degree of parallelism between two plates separated by a transparent medium. If two such plates form a small angle a (un- known), the interposed medium having a refractive index n, the measured angle be- ing d, one has: Conversely, if a and d are known, one calculates n: n = sin d/s\n (2a) (S) The Goniometer. For the measurement of small angles involved in a number of methods mentioned herein, it is often con- venient to use the goniometer. The goniome- ter method is based upon a property of the refracting prism already described, i.e., the existence of a minimum deviation angle. The determination of the refractive index 7i re- quires measuring the minimum deviation angle. This involves moving both the prism and the telescope, as in the Biot and Arago method. However, a greater accuracy is possible with the goniometer by rotating the prism on the turntable so as to give it suc- cessively two exactly symmetrical positions. A modification of the method permits measurements on liquids. Instead of a solid prism, the goniometer is equipped with the double rectangular liquid compartment of Hallwachs. The table is oriented so that light falls at grazing incidence on the plate of glass separating the two fluids and one observes the angular deviation relatively to the direction of the rays when the two com- partments are empty: sin- d = Hi 2 _ n2 (9) a = arc sin (sin d/n) (7)t * Mfd. by Aminco, Silver Spring, Md. t (For equations 1-6 see next article). (?ii and n2 are indices of the fluids filling the two compartments, respectively). By using a double switching method and taking read- ings in two opposed directions at 180°, dif- ferences of refraction of the order of 1.5 X 10~^ can be measured. Other Methods of Angular Measure- ments. A number of refractometric meth- ods involving angular measurements to be carried out with non-specialized instruments have been developed primarily for the ana- lytical chemist. The required implements are usually a microscope, a goniometer, or a simple type of photometer. Microscopic methods utilizing the Becke lines displacement are described in textbooks 487 REFKACTION OF LIGHT, REFR ACTOMETRY AND INTERFEROMETRY of microscopy. They were critically reviewed prism which also supports the specimen un- by Say lor (47). The utilization of the Chris- der investigation. A suitable telescope is tiansen effect falls into the same category, placed on the small side of the prism, near In general, such methods have been brought its right angle. This telescope is mobile to a high degree of perfection mostly for the around an axis perpendicular to the plan of requirements of petrographic studies (48 refraction of the rays of light, and centered 49, 50). They culminated in the very perfect around the right angle of the prism. It is microscope stage for crystallography de- attached to a graduated circle, so that the signed by Emmons and Winchell (39). The angle of emergence of the refracted rays can measurements can be extended to any sub- be measured. When the direction of emer- stance exhibiting metallic reflection (51), as gence is found, the field of the telescope ap- well as to plastics and to resins (52). The pears divided into two zones: one bright, basic procedures remain those of earlier in- and one completely dark if the incident vestigators (53, 54, 55, 56, 57) and others, light is monochromatic. Several of these procedures can be combined The original Abbe refractometer was im- with the "immersion method" familiar to proved to permit measurements on both microscopists (46, 58, 59). The Merwin transparent and opaque liquids as well as on method for crystallography has long been solid specimens. To the hypotenuse face of standard. Those of Wood and Ayliffe (60), the original Abbe prism a second prism ex- and of Wood, Ayliffe and Williams (61) are actly identical is attached. The hypotenuse rather general for all solid specimens. face of this second prism, in contact with that The procedure of Kienle and Stearns (62) of the former, is finely ground and is provided for computation of refractive indices of with a small cylindrical cavity where the opaque specimens from iso-reflectance dia- specimen under investigation is placed. To grams and the Fresnel's foraiulas are illus- this combination is added a trapezoidal piece trative of the use of a spectrophotometer of glass of such a shape that the incident (American Cyanamid Co. automatic re- ray of light falls upon the surface of the corder) for such a purpose. specimen. Refractometric measurements by means The methods available for measuring the of the microscope are considered by Addey characteristic constants of the critical angle (63). refractometer are often successfully used in The Critical Angle Refractometer: setting up other types of optical instruments. Abbe Instrument. Measuring instruments such as spectrometers. The original Abbe of this type are based upon Snell's formula refractometer has been modified and im- (1). The measurement consists in observing proved in numerous ways by ingenious the critical angle of emergence Vc with light manufacturers. Precision type instruments falling at grazing incidence {i practically are currently manufactured in this country equal to 90°) upon the boundary of two by Bausch & Lomb Optical Co., American transparent media, one of which has a known Optical Co., Precision Scientific Co., In- refractive index, rig . dustro — Scientific Co., and by the Gaertner The simplest instrument of this category Co. The latter, originally developed by Ja- is the Abbe refractometer. It is essentially cobson is suitable for turbid fluids. The formed of a flint prism whose angles are re- Ostwald refractometer (Zeiss Mfg.) is a dif- spectively 90, 60, and 30 degrees. A slightly ferential instrument based upon the same convergent beam of light is furnished by a principle. suitable condenser and falls at grazing in- The Fery Refractometer. A different cidence upon the hypotenuse face of the version of the application of the principle of 488 ANGLE REFRACTOMETRY the critical angle to the measurement of re- ing (and approximately evaluating) the dis- fractive indices is embodied in the Fery re- persion. fractometer. In this instrument, the refract- Various attachments, which need not be ing prism is a hollow body to be filled with mentioned here, permit one to adapt the the liquid under investigation. The faces instrument for measurements on either bulk comprising the refracting angle are made liquids, single drops, properly cut trans- of two half -lenses. This body is mobile trans- parent solids, or opaque liquids, versely to the path of light, so that a narrow The accuracy of the measurement is of the beam of collimated light can be made to pass order of 0.00003 to 0.00004, according to the through it at various distances from its re- position of the dividing line in the field of the fracting angle a. A telescope provided with a telescope. cross-air reticule permits one to observe the The Pulfrich Refractometer. This de- apparent position of another reticule dis- vice utilizes the total reflection principle. A posed inside the collimator tube. This ap- glass prism of high refractive index has two parent position depends upon the refractivity polished faces perpendicular to one another, of the liquid contained in the prism. The one of them being horizontal. This horizon- image of the collimator reticule is brought tal face has a small glass cell cemented to it back in coincidence with that of the reticule for retaining the liquids to be examined, of the telescope by moving the prism by The beam of incident light is directed so as means of a micrometric screw. This displace- to strike the prism at grazing incidence after ment is a simple function of the refractive in- traversing the liquid. The emergent beam dex of the liquid. passes through the vertical face of the prism. This instrrmient really makes use of a It is sharply outlined by those rays which "variable refraction angle prism." It has actually graze the prism surface. The sharp known a very prolonged success in chemistry, boundary is observed with a telescope at- probably because of its low cost of construe- tached to a divided circle and vernier. Read- tion and its ease of operation, but is less ings are possible to 6 seconds of arc. For satisfactory than the Abbe refractometer. accurate temperature control, the optical The Dipping Refractometer. The "dip- parts of the instrument are surrounded by ping refractometer" has received very wide a water jacket. acceptance in industry. Since it can be ap- The instrument is intended for measure- plied to measurements even in flowing fluids, ments on liquids, the refractive indices of it is the nearest thing, in its class, to a truly which are lower than that of the glass of the industrial apparatus. prism. To extend the range of measurements, The dipping refractometer is designed on interchangeable prisms of different indices the principle of the Abbe refractometer. The are available. The accuracy on the refractive refracting prism is similar to the upper index of liquids is about 0.00001. It reaches prism of the Abbe apparatus, but there are 0.00002 on the measurement of the dis- provided usually six interchangeable prisms persion with monochromatic lights, of various indices covering the range of Solid samples must be cut so as to present measurements from 1.32539 to 1.54409. two perpendicular faces meeting at an un- The telescope is of conventional design broken line, while one of the faces must be butof longer focus. The dividing hne between fairly well polished. Optical contact is ob- the two halves of the field is formed in the tained in the usual way. plane of a divided scale and its position is For convenience, the manufacturers sup- read directly; no moving parts are involved, ply tables giving the value of n vs. the meas- A single Amici prism is provided for correct- ured angle i, for various spectral lines. 489 REFKACTION OF LIGHT, HEFRACTOME I RY AM) INTERFEROMETRY Lob and Pfund Refractometers. The ferences. A Nichols refractometer placed on principle of the Lob refractometer is identi- the stage of such a modified microscope then cal with that of the Pulfrich refractometer, becomes a small interferometer, the beam but all the readings are made with the same spacing of which depends upon the refractive telescope used for indicating the direction of index of the substance under study, critical emergence. The eyepiece is fixed, For work in the ultraviolet, M. M. Hilger and it includes an accessory prism for com- manufactures a photographic angle refrac- pensating the spectral dispersion. Instead of tometer which appears to be but a commer- the Pulfrich prism, a double Abbe prism is cial version of an instrument originally de- used, signed by Charles Henri (34). The Pfund refractometer utilizes the The chemical micro-refractometer of Jel- principle of the total reflection of light at ley (31) is intended for the rapid determina- normal incidence at the boundary of two tion of the refractive index of immersion different media. A convenient instrument of liquids used for the identification of crystals this type for the study of the solidification by one of the immersion methods employing process of varnishes and lacquers has been the Becke lines or the Christiansen effect, described by Countryman and Kunerth (30). The essential part of the instrument is a Other Prismatic Refractometers. An small glass wedge of refractive index rig , sup- early modification of the Abb6 refractometer ported by a plate. The wedge supports a was the Fery instrument, already mentioned, drop of the specimen, forming a liquid prism In this instrument the curved faces introduce of index n^ and of refracting angle a. This "caustics" which produce greater uncertain- prism is about 1 mm thick. A narrow slit is ties on the readings. Despite this defect, this adjusted so that a thin beam of light falls instrument has been used for an extraor- upon the prism about the middle of the dinarily large volume of data relative to or- wedge. An observer sees a sharp virtual ganic chemistry. image of the slit formed upon the scale which Cruikshank and Fairweather (31) have is graduated directly in terms of refractive modified the Fery refractometer to adapt it indices, for continuous recording. The instrument can also be used con- The Nichols (32) double reflection micro- jointly with a heating device for measuring scope attachment for refraction measure- the refractive indices of melted substances, ments in reahty must be classified with the up to about 300°C. prismatic refractometers. It has been The Jelley refractometer is particularly adapted to microanalytical work by Alber well adapted to chemical studies when a and Bryant (33). This instrument, requiring modest sensitivity is required. Its commer- only 5 to 10 mm of fluid, may be operated cial version, the Jelley-Fisher refractometer, at high temperature (melting point refrac- described by Edwards and Otto (35), is cur- tivity) and is most convenient for chemical rently used to measure indices at the melt- studies. Its sensitivity can be considerably ing point. For measurements at still higher increased by using an interference method for temperatures (up to 300°C), one can use measuring the displacement of the two beams Fredriani's modification (36), which permits of light. It is possible to place a double simultaneous observation of the melting Young's slit over the conventional micro- points. scope condenser. If the aperture is limited to Broumberg's focal refraction monochroma- the production of the so-called marginal tor (45), although not using a prism, is rays, then the emergent light is coherent, based upon Snell-Descartes' formula, and and suitable for the production of inter- therefore it may be considered in the same 490 ANGLE REFRACTOMETRY light as the prismatic refractometers. In fact, tained by auto-collimation. The two-dimen- it can be used for refractometric measure- sional plot may be directly projected on a ments. screen, or it may be magnified. Extensive data relative to the refractivity The Grauer angle refractometer is an in- of organic substances and the application of teresting adaptation of a two-prism system such measurement methods to the problems employed at the National Bureau of Stand- of organic analysis have been collected (64- ards for the study of solid specimens (76). 68). Applications of refractometry to chemi- It could be easily adapted to the study of cal production control became really prac- fluids. tical with the advent of rugged recording "The instrument consists of a light source types of instruments. One of the early de- with a horizontal filament, a collimator, a signs is the multiple reflection instrument of diaphragm, a pair of 45-degree-90-degree Karrer and Orr (38), now obsolete. It was cemented prisms forming a 90-degree hol- based upon the principle of Fery's hollow low at the upper half and a parallel plate prism. The Barnes recording refractometer at the lower half, a filar micrometer, and a (41) is now superseded by the commonly telescope. All the components are centrally used industrial instrument of Barstow (40). aligned along a common axis." At present there is a tendency to extend Compensation of the Spectral Dis- a renewed interest in differential types of in- persion. In the preceding discussion it was struments. A differential refractometer prism assumed that the incident light was mono- is described by Kegeles (69). It consists of chromatic. With composite light, white or two identical, hollow rectangular prisms of variously colored, the critical angle of refrac- refracting angle A adjacent on their hypot- tion varies with the wavelength. If such enuse faces. One compartment is used for a light is used for the measurements, the reference fluid of index no , the other re- boundary line seen in the telescope appears ceives the test solution of index n. Parallel as a hazy colored band, the colors being ar- light enters normal to one of the small faces, ranged symmetrically in the two halves of passing through the test fluid first, and is the field. The dispersion of the light in the deflected from the normal by an angle given instrmnent is due to the combined action by Snell-Descartes' law. Similar prismatic of two factors: (a) the constant effect of the refractometers have been described by Tho- prisms, the physical dimensions of which are vert (70), Zuber (71), Longworth (72), Debye fixed, and (b) the variable effect of the speci- (73), and others. men under investigation. The first factor The advantage of such a design lies in its could be corrected once for all; the second re- small size, so that sufficiently rugged meas- quires adjustment for every specimen in- uring cells may be built for direct measure- vestigated. ment of sedimentation equilibrium in the In order to cancel out this phenomenon, ultracentrifuge (Kegeles (74)), by a modifi- the modern refractometers are provided with cation of Lamm's schlieren scale method two composed Amici prisms disposed inside (75), with the introduction of suitable cor- the telescope. These prisms are cut at a rections indicated by Kegeles. With suffi- suitable angle so that the light from the so- ciently high prismatic channels and suitable dium Dsa lines passes undeviated. optics, it is possible to observe directly a A rough estimation of the spectral refrac- two-dimensional refractive index distribu- five dispersion of substances under investi- tion in a heterogenous — say, for instance, re- gation can be derived from the conventional fraction increment dn, vs. a concentration expression: gradient. Greater deflections may be ob- d = no - l/Oip — nc) {lO) 491 REFRACTION OF LIGHT, REFR ACTOMETRY AND INTERFEROMETRY where the n's are the refractive indices suc- cessively measured with the hght from the Z)Na hne of sodium, the Ca line of hydrogen, and the F^ line of hydrogen, respectively. For this measurement, the two Amici prisms of the refractometer are mounted in such a way that they can be rotated in op- posite directions about the optical axis of the telescope, their respective positions being read on an arbitrary scale. With all types of critical angle refractom- eters, certain factors limiting the accuracy of the measurements are not entirely con- trollable by the operator. These factors are: (1) The extreme temperature sensitivity of the instrument, producing a hazy, shift- ing, dividing line, even when a thermostatic jacket is provided. (2) The difficulty in setting the exact in- cidence and proper convergence for the light. This factor affects the contrast at the divid- ing line in the field of the telescope and limits the reproducibility of the settings. (3) The limited power of the dispersion compensation prisms. These prisms are effi- cient only for the D lines of sodium light, and they exactly correct images that are formed only near the center of the field. With liquids of high refractive indices (oils, etc.) the boundary is always somewhat broad and iridescent. (4) The accuracy on the reading of the divided circle (critical angle). This last point is discussed in physics textbooks. Some of these difficulties are reduced by designing refractometers covering certain limited ranges of indices, for special indus- trial applications (oils, sugars, resins, wines, etc.), and by using readily interchangeable monochromatic sources of radiant energy, now commercially available at reasonable price and directly operated from the power line (filtered carbon arc, mercury arc, helium and hydrogen tubes, Na, Sr, K, Cd, Kr, Ne, lamps, etc.). A convenient table for the se- lection of suitable radiation sources will be found in an article bv Tilton and Taylor (79). Microscopic Methods Based on Change of Focus. The method of deter- mination of the true thickness of an im- mersed object by means of the microscope was originally invented by the Due de Chaul- nes, and was completely discussed by Jo- hannsen (53), the Winchells (54), Addey (63), Blunck (65), and summarized by Cha- mot and Mason (58). This method is very widely used in chemistry. A glass plate is provided with a hollow cell 0.5 to 2.0 mm deep and about 3 mm in diameter. The bottom of the cell is polished with parallel faces and scratched. The cell is filled with a liquid to be studied and cov- ered with a cover glass, itself scratched. The apparatus is placed under the microscope, and the apparent vertical distance between the virtual images of the scratches is meas- ured with the micrometric adjustment of the microscope. The true distance between the two faces of the cell is also measured in the absence of liquid in the cell (n = 1). The refractive index of the liquid is given by the relation : n = true thickness/apparent thickness (H) The micrometric adjustment of the micro- scope can be calibrated directly in terms of refractive indices by the use of liquids of known refraction. When the solid specimen available is too small (powders), the above mentioned crys- tallographic methods cannot be applied. It is, however, possible to evaluate the refrac- tive index of these solid specimens by immer- sion in liciuids of convenient refractive index. When the index of the immersion liquid equals that of the immersed specimen usually some sharp modification of the microscopic image of the latter can be observed. The re- fractive index of the liquid is then measured by one of the methods described. The equal- ity of the indices can be ascertained by the microscopical observation of either one of 492 ANGLE REFRACTOMETRY the following phenomena: (a) the displace- ment of the Becke lines; (b) the Christian- sen effect. The Becke Lines. The displacement of the Becke lines when a crystal is immersed in a liquid is explained as follows. When a crystal is seen through the microscope, the image of the edge will move away from the crystal when the objective is moved upward, and its index is smaller than that of the im- mersion fluid. It will not move at all when the refractive indices are equal. The method applies also to crystals other than those of the cubic system. In this case, several Becke lines are visible, but, by add- ing a polarization attachment to the micro- scope, all but one of these lines at a time can be immobilized and the corresponding refractive index thus determined. With posi- tive uniaxial crj^stals one index, n,^, , the lowest, is the same in all directions of the crystal. The other index varies between the value above and a maximum value Ue . The inverse relationship holds for negative uni- axial crystals. With biaxial crystals, both indices n^, and Ue , respectively, vary with the direction, but a maximimi and a minimum value also can be recorded when a mixture of crystals of various sizes is observed. However, in this case the final interpretation requires another set of observations with conoscopical light (convergent polarized light). A very complete discussion, including all operational details and bibliography, will be found in Chamot and Mason (58). The Christiansen Effect. The Christian- sen effect takes place when two contacting media (a crystal immersed in a liquid) have the same refractive index for a given wave- length, but a different spectral dispersion; the Becke lines then have the aspect of small bright spectra. The image of the object is surrounded by colored fringes moving in one direction or the other with the movements of the microscopic objective. Observation of the Christiansen effect leads to very valuable information concern- ing the spectral dispersion of the material under investigation, if that of the immersion liquid is known. Extensive lists of immersion liquids which have proved their worth exist in the litera- ture (46, 56 58). Let us mention only the tetrasodium dioxypentathiostannate of Jel- ley, which, being soluble in water, is valu- able for the study of water-insoluble organic materials. The methods available for adjusting the refractive index of the immersion liquid to equal that of the immersed specimen are : (a) the dilution method, utilizing known mix- tures of liquids (tedious, but very accurate); (b) the Merwin (46) method. Use is made of the different spectral dispersions exhibited by the liquid and by the sepcimen. A mono- chromator is used to change progressively the wavelength until the Becke lines seen in the microscope disappear; (c) the Emmons method, using the different relationship exist- ing for the immersion liquid and the speci- men, between the refractive index and the temperature: the microscope must be pro- vided with a hot-stage attachment; the method can also be combined with method (b); (d) the Leitz-Emmons hemispherical refractometer, which fits over any standard microscope. The specimen and immersion liquid are enclosed in a small hemispherical cell together with a concentric hemispherical hollow rotatable from outside, and having a different diameter. When the indices of the liquid and specimen, respectively, have been matched by adjusting the temperature of the hot stage, the refractive index of the liquid is immediately determined by rotating the hemispherical hollow until the position of total reflection of light is found ; the appara- tus serves as its own refractometer. The crystallographic immersion methods described here are applicable to metallo- graphic and metallurgical specimens. Most of them can be carried out on rough speci- 493 REFRACTION OF LIGHT, REFR ACTOIMFTRY AND INTERFEROMETRY Table 1 dn Wavelength Nichols refractometer .001 Na Jelley .001 Na Edwards-Otto .002 Na Microscope .003 Na Beck lines .003 Na Christiansen effect .003 Na Pfund .0001 Na Henry .0004 Na Barstow (recording) .0004 Na Cruikshank-Fairweather .0004 Na Fery .0005 Na Tuckerman (auto-collima- .0005 Na tor) Ahh6 .00001 Na Abbe .00002 (White) Pulfrich .00002 Na Grauer .00004 Na Dipping (B. & L.) .00004 Na Lob .00005 Na Interferometers .0000401 Na Phase-contrast interferom- .000000005 Na-594 eters mens in a few minutes, and for these reasons are widely used industrially. Selection of a Refracto metric Tech- nique. The last significant figure of relative refractive index obtained with the available instruments serves as a guide in selecting the proper refractometric technique. Routine fig- ures currently accepted for some commer- cially available instruments are given in the Table 1. REFERENCES See next article. R. JONNARD HISTORY OF LIGHT REFRACTION (See also p. 454) The term "refraction" refers to the change of direction, or "breaking" of the rectilinear path of light at the surface of separation of two transparent media. This paradoxical phenomenon has at- tracted the attention of philosophers since early Greco-Latin anticiuity, and it was cor- rectly opposed to the rectilinear propagation observed when light travels through unob- structed space. It is worthy of note that the word Optikos ("I see") existed only in the Greek language. Thus, the study of optics appears to be an intellectual phenomenon properly Greco-Latin in origin, arising sometime toward the end of the Hellenic period of history (1). The historical development of knowledge relative to the refraction of light encompasses the whole of geometric optics as well as the nature of light and the mechanism of vision. Reflection phenomena, linked with recti- linear propagation were well described in Plato's "Times" (427-347 B.C.), in Euclid's "Optica", Vol. VII (2), (circa 325 B.C.), in Appollodorus's "Philosophy" (3), and in Claudius Ptolemy's "Physics" and "Catop- tries" (2nd century A.D.). All these authors knew the elementary laws of optical reflec- tion. Simultaneously with the development of knowledge about the reflection of light, the apparentl}^ contradictory notion of refraction was neatly distinguished by the peripateti- cian of Chalcis, Aristotle (322-284 B.C.) in his "Rhetoric" and "Meteorology." His views were extended by the Stoic Posi- donius, of Apamee (135-50 B.C.). Lucius Annaeus Seneca's "Natural Questions" (circa 25 A.D.) contains abundant allusions to the consequences of the refraction of light through flasks, bottles, glass, menisci and the hke. In fact, Mediterranean antiquity knew how to utilize these effects practically. The full significance of these discoveries appeared after Aristotle, referring to the fifth century (B.C.) philosopher Empedocles, postulated that light is a vibrational phe- nomenon traveling with a finite velocity through fluids of definite densities. Such views were subsequently expoimded by Aris- toxene of Tarente (circa 4th century B.C.), to be finally developed by the Arab Averrhoes in the 12th century A.D. As early as the 10th century A.D. one 494 sin i/sin r = n = Cte {!) HISTORY OF LIGHT REFRACTION finds a most explicit mathematical theory of was able, in 1644, to interpret correctly Willi- refraction in the Ibn-Al-Haitham's "Treatise brod Snell's experiment (Leyden, 1621) and of Optical Geometry" after this author, rec- to formulate the basic law of refraction which ognizing the limitations of Ptolemy's ele- summarizes so well most of geometric op- mentary law, discovered the principle of the tics: inverse return of spherical waves, which is the basis of modern prismatic refractome- ters. During the Middle Ages the laws of where n is the refractive index characteristic light refraction were widely applied. For in- of the medium traversed. This law made stance, Roger Bacon in his "Opius Mains" modern optics possible. (1270) describes glass lenses without stating Thus, the early history of the refraction of their basic properties. Indeed, one finds in light can be traced from early Greco-Latin Al-Hazem's Treatise all necessary funda- antiquity, along two main lines of thought, mentals which might have been desirable at One is concerned with the studies derived the time. Thus, contrary to a common asser- from the reflection of light, running from tion and to the opinion of Joseph Bertrand, Empedocles (5th century B.C.) through Aris- the discovery of refraction considerably pre- totle, Plato, Euclid, Apollodorus, Ptolemy, ceded Tycho-Brahe, Vitellio (circa 1270), Bacon, Magnus, Kircher, Snell and Descar- and Kepler. tes. The other is concerned with the phe- Johannes Kepler (6) must be credited with nomena related to the non-linear propagation what may be considered the first practical of light, from Aristotle again, to Aristoxene, refractometer. Its principle is well described Posidonios, Cleomede, Lucretius, Seneca, in his "Dioptrice," published in 1611. The Ibn-Al-Haitham, Averrhoes, Bacon, Tycho- instrument permits one to determine the re- Brahe, Vitellio, Kepler, Hooke, Snell and, lationship between the ratio of length of the again, Descartes. shadows and the ratio of incidence and of For the remaining conformal picture frag- refraction angles, the two rays traveling ments, the historical development, the reader through air and an unknown medium, re- is referred to the classical works of Leonardo spectively. Although Kepler labored under da Vinci (10), Huygens (11), Fresnel (12), the wrong assumption that light has an in- Arago (13), Mach (14), and others. A very finite velocity, it is with such an instrument clear exposition of the phenomena of refrac- that he discovered the important phenome- tion from the standpoint of geometric optics non of the internal total reflection in glass is given by J. W. Forest in 0. Glasser's book for an angle of about 42°. Thus, Mees (7) (15), while the references in Reymond's book erred when he stated that Kepler did not (16) somewhat fill up some lacunae in Wile's know the law of refraction. basic work. Other important reference works There is some evidence that Robert Hooke include: Delambre (17), Doublet (18), (8), in England, also invented a refractome- Marion (19) and Tannery (20). ter of some kind, but no trace of Hooke's Almost simultaneously with the publica- instrument has been discovered. tion of Descartes' momentous synthesis, The 17th century continued to labor under Pierre de Fermat showed that the "sine law" the assumption of an infinite velocity for could be deduced from the philosophical light. To Descartes (9), light was simply a principle of "optical extremum path." The "pressure transmitted to the eye," with an principle proved much broader than the ex- infinite velocity which could "still be perimental data then available, and its in- greater" when traveling in denser media fluence in science may be felt in the work of than in lighter ones. Despite this error, he Leibnitz, Kant, Hamilton, Gauss, and even 495 REFRACTION OF LIGHT, REFRACTOMETRY AND INTERFEROMETRY of ]Mach. This lends support to a remark where abouts of the rays ultimately emerging made by Destouches (21) that "the essential, from the optical system considered. This is in order to build a theory, is to possess simple given by Malus' theorem, according to which and schematic ideas, of origin rather intui- all the rays issued from a point source *Sr, tive than purely experimental, and a long after undergoing multiple reflections and/or work of the mind is required." refraction, remain normal to a family of de Fermat's principle is often expressed parallel surfaces E, each point of which is mathematically as a "stationary time in- equidistant from S (conjugate) by the same tegral": optical length L. The surface defines the "wave-front surface" or surface of equal / n ds = 0 {2) phase. It is that which is reached by light after a unique time T (equations 3 and 4) \vhere n is the cartesian refraction index, ds from all the possible trajectories so that: T = L/C = S(Z/F) (-^) / is the geometric path length, and the quan- tity nds defines the optical path length. It is interesting to note that a very similar where I is the geometric path length. This is principle of optimum path length was al- in strict conformity with de Fermat's prin- ready proposed in the third century by Hero ciple. A bundle of light rays satisfying such of Alexandria who implicitly admitted a conditions is said to form a congruence of finite light velocity (22). The principle is normals when their direction of propagation sometimes more useful in the equivalent is defined by a series of parameters (X, Y, Z) form : which depend upon at least two independent variables (such as V, U, etc.). In the limited ds/d'K = 0 {3) case of geometric optics (n = Cte) with only one variable, one has a totality of curves In all its forms, the principle states that orthogonal to a family of wave surfaces in- the sum of all optical path lengths followed stead of a congruence. Huygens construction by the light rays through any succession of (1690) is merely a convenient geometric isotropic, homogeneous transparent media demonstration of this theorem, separated by stigmatic surfaces, is station- It is perhaps clear to those conversant with ary, that is, two such adjacent rays differ in Huygens' work that the ideal stigmatic sur- optical length only by an infinitesimal quan- face of a given optical system may differ tity at least of second order. The principle from the actual physical surface of its diop- never states whether this length is a maxi- tic elements. Thus, three situations may mum, a minimimi, or an inflexional quantity, arise, according to which the actual surface but only that it is constant. The writer, is either tangent, internal, or secant to the among others, has cautioned about the in- stigmatic surface. Accordingly, the optimum discriminate interpretation of this principle path length of de Fermat is either a maxi- (23). The argument can be summarized as mum, a minimum, or an inflectional quantity, follows. In the relatively simple case of a respectively, and in the order given above, beam of parallel rays of light, the algebraic The implications of de Fermat's principle sum of all of the partial path lengths of in- pervade much of contemporary physical- cident plus reflected rays remains constant chemistry. Almost simultaneously with its regardless of the angle of incidence on a publication, Pierre Maupertuis showed, in- given surface, be it an ellipse, an hyper- dependently, that when a "material point in boloid or a parabola. action" is analyzed, the sum of all the ener- It is obviously important to know the gies involved is stationary, the representa- 496 HiSTOKY OF LIGHT REFRACTlOiN tive equation being: P J V(U + E) ds = 0 (5) where U is the potential energy, E is the kinetic energy ds represents the distances involved. It remained for the genius of Sir William Hamilton to discover the analogy between equations (5) and (-5) and to bring it forth by postulating: y/{U + E) = n (6) The fundamental importance of Mauper- tuis' discovery is perhaps clearly stated in the following citation: "Thus, the refractive index n expresses the force function or field existing in matter, which results in slowing down the velocity of light and determines the trajectory of rays or streams of information, to use modern terminology. In this sense, n is a measure of interaction between matter and light, and its determination contributes to a knowledge of the structure of matter. The abstract concept of a field of forces re- mains only a convenient but arbitrary form of language to explain the properties of space and to predict the future behavior of the ma- terial particles, the local sources of the physi- cists, or points of particular chemical inter- est contained in space. Thus, field of forces, space properties, and the chemical proper- ties of the elements are theoretical views of the mind that become real, like shadow of a tree, only when some interaction takes place. One sees that the concept of generalized in- teraction is contained in that of field, as was perceived by Paul Langevin when he wrote in 1903 : Tt is always matter which contains the charges whose field divergence becomes different from Zero.' In the absence of mat- ter, indeed, V{U + E) = 1, and it is the nature of the sources that deter- mines the properties of the field: if there is no matter, there are no charges, no sources, no force, and no field in space." (1) Following this discovery, the great amount of experimental work stimulated b}^ the theories of T. Young, A. Fresnel, and C. Maxwell, all based upon the fundamental concept of "field of force," revealed that, in fact, the refractive dispersion of transparent substances is fully governed by their ability to absorb radiant energy. It is possible to develop newer, more satisfactorj^ theories of electromagnetism without utilizing the ab- stract concepts of field of force and of action at distance (24) but the antiquated field re- mains a convenient means of expression. Today, it is admitted that the proposition represented by equation (5) is valid if n varies little and continuously. The "rays" considered are no more than trajectories of electromagnetic energy, and de Broglie re- marks (25) that, in this case the treatment is compatible with hath Newton's strangely modern corpuscular theory and Fresnel's undulatory theory. But, if n varies suddenly, diffraction takes place and the concept of "ray" vanishes. However, that concept of refraction in the sense of Hamilton's equa- tion remains valid. It is apparent from the foregoing, that the determination of refrac- tive indices may be accomplished essentially in two ways: (a) by application of Kepler- Descartes formula, involving angular meas- urements, (b) by application of de Fermat's principle, involving velocities, or phase, measurements. The first method involves all the prismatic refractometers and derived pro- cedures. The second method concerns all the interferometric, differential methods. From this point on it is a matter of textbook record that the instrumental development based upon one method or the other fluctuated Avith the fortunes and temporary relative impor- tance of the underlying theories of light. REFERENCES 1. JoNNARD, R., "Optics, a Greco-Latin Mira- cle," Sigma-Delta-Epsilon Lect., N. Y. Academy of Sciences, Jan. 20, 1958. 2. Teubnee, S. H., "Euclidis Optica," Heidel- berg, 1895. 497 REFRACTION OF TJGIIT, KEFRACTOMETRY AND INTERFEROMETRY 3. Wilde, H., "Geschichte der Optik, Berlin, 1838. 4. DE ToNQUEDEc, J., "Ciitique de la Connais- sance," G. Beaiichesne, Paris, 1929. 5. JoNNARD, R., "Principles of Rationalization in Biology and Medicine," Proc. 7th Nat. Conf. I.S'.A., Cleveland, 0., Sept. 1952. 6. Kepler, Johannes, "Ad Vitellianem Para- lipomena quibiis astronomiae pars optica traditur," I, p. 20, Frankfort, 1604; Also: Dioptrica, Frankfort, 1654. 7. Mees, C. D., and Baker, J. R., "The Path of Science," J. Wilej^ & Sons, New York, 1946. 8. Hooke, Robert, "Micrographia, or Some Physiological Descriptions of Minute Bodies Made bj' Magnifying Glasses, with Observations and Inquiries Thereupon," London, 1664. 9. Descartes, Rene, "Dioptrique," 1637; "Discours de la Method: Principa Philos- phae," 1644. 10. da Vinci, Leonardo, "Ueber die Malerei," German ed. by Ludwig. 11. HuYGENS, C. V. Z., "Traite de la Lumiere," Leide, 1690. 12. Fresnel, a., "Oeuvres Completes," Paris, 1858. 13. Arago, F., "Oeuvres Completes," Paris, 1865. 14. Mach, E., "Principles of Physical Optics," transl. Anderson, J. S. and Young, A. F. A. Dover Publications, N. Y., 1926. 15. Forrest, J. W., "Refractometry," cf. in Glaser, O., "Medical Phj'sics," Year Book Publ., Chicago, 1944. 16. Reymond, a., "Histoire des Sciences Exactes et Naturelles dans I'Antiquite Greco-Ro- maine," A. Blanchard, Paris, 1924. 17. Delambre, J. B., "Histoire de I'Astronomie Ancienne," Delagrave, Paris, 1817. 18. Doublet, E., "Histoire de I'Astronomie," Doin, Paris, 1922. 19. Marion, F., "L'Optique," L. Hachette, Paris, 1869. 20. Tannery, P., "Memoires Scientifiques: Sciences exactes et naturelles dans I'An- tiquite," Gauthier-Villars, Paris, 1912. 21. Destouches, J. L., "M^thodologie de la Physique Theorique Moderne. Notions de Logistique," Tournier et Constans, Paris, 1946. 22. Hero, "Dioptra," ed. by Schone, H. Teub- ner, 1903. 23. JoNNARD, R., "Random Selection System for Automatic Dynamic Biochemical Analy- sis," Ann. N. Y. Academy of Sciences, 87, ()69-728, (19G0). 24. Moon, P., and Spencer, D. E., "The Cou- lomb Force and the Ampere Force," J. Franklin Inst., 257(4), 305-315 (1954); Am. J. Phi/sics, 3, 120-134 (1954); ./. Franklin Inst., 257(4), 369-382 (1954). 25. DE Broglie, L., "Matter and Light, the New Physics." Dover, New York, 1946. 26. Abbe, E., "Neue Apparate Zur Bestimmung des Brechungsund Zerstrenungsver Fester und Pluessiger Koerper," Naturwiss., 8, 96- 174 (1874). 27. PuLFRiCH, C, "Ein Neues Refraktometer," Zeitschr.f. Instr., 7, 16-27 (1887). 28. Lob, P., "Eine Neue Universal Refraktome- ter mit Innenablesung," Zeitschr. f. In- str., 53, 27-30 (1933). 29. Pfund, C, J. Opt. Soc. Am., 24, 121-125 (1934). 30. Countryman, A., and Kunerth, W., "The Construction and Use of a Pfund Parallel Plate Refractometer," /. Opt. Soc. Am., 24, 25-28 (1934). 31. Jelley, E. E., "A Microrefractometer and its Use in Chemical Microscopy," Proc. Roy. Microscopic. Soc. London, 54(3), 234- 245 (^535, 317-8-XVl) (1934); "The Prep- aration and Constitution of Thiostannates. II: tetra- and octa-dioxypentathiostan- nates," J.A.C.S., 78, 1076 (1934). 32. Nichols, L., "Double Reflexion Refrac- tometer," Nat. Paint Bull. 1, 12-13 (1937); 1, 14-16 (1937). 33. Alber, H., and Bryant, J. T., "S3'stematic Qualitative Organic Micro-analysis. Deter- mination of the Refractive Index of Liq- uids," Ind. Eng. Chem. Anal. Ed., 12, 305- 307 (1940). 34. Henri, Ch., "Etudes de Photochimie," Gauthier-Villars, Paris, 1919. 35. Edwards, A. E., and Otto, C. E., "A Micro- refractometer of Simple Design. Labora- tory^ Construction," Ind. Eng. Chem. Anal. Ed., 10, 225-226 (1938). 36. Fredriani, H. A., "Refractive Index Meas- urements at and Above the Melting Point of Solids," Ind. Eng. Chem. Anal. Ed., 14, 439 (1942). 37. Cruikshank, C, and Fairweather, W., Brit. Pat. 555,928, Sept. 13, 1943. 38. Karrer, E., and Orr, R. L., "Recording Photoelectric Refractometer," /. Opt. Soc. Am., 36, 42-46 (1946). 39. Emmons, R. C, and Winchell, A. N., "Some Methods for Determining Refractive In- dices," Am. Mineral., 11, 115-118 (1928). 498 HISTORY OF LIGHT REFRACTION 40. Barstow, O. E., "A New Recording Refrac- tometer," Instr. 23(4), 396-398 (1950). 41. Barnes, R. B., "Refractometer for Chemi- cal Reactions," U.S. Pat. 2,413,208, Dec. 24, 1946. 42. Jacobson, S., U. S. Pat. 237',1625, May 20, 1945. 43. Stamm, R. F., Mariner, Th., Barnes, R. B., AND Stryker, Ch. R., U. S. Patent 258,- 3973, Jan. 29, 1952. 44. Clammann, E., "Pneumatic Refractometer," Instr., 26(5), 740-742 (1953). 45. Broumberg, E. M., "Une Nouvells Methode Pour Rendre la Lumiere Monochroma- tique," C.R.U.R.S.S. Acad. Sci., 2, 464- 469 (1935;) Rev. d'Opt. Theor. et Experim., 15 (2), 69 (1936). 46. Merwin, H. E., and Larsen, C. S., "Mix- tures of Amorphous Sulfur and Selenium as Immersion Media for the Determination of High Refractive Indices with the Micro- scope," Am. J. Sci., 34(4), 42-47 (1956). 47. Saylor, Ch. P., "Accuracy of Microscopical Methods for Determining Refractive In- dices by Immersion," Nat. Bur. Std. J. Res., 15, 277-294 (1935). 48. TsuBOi, S., "A Dispersion Method of Deter- mining Plagioclase in Cleavage Flakes," Mineral. Mag., 20, 108-122 (1923). 49. TsuBoi, S., "A Dispersion Method of Finding the Principal Refractive Index of Crystals in Powders," /. Geol. Soc. Tokio, 32, 1-6 (1925). 50. TsuBOi, S., "A Dispersion Method of Dis- criminating Rock Constituents and its Use in Petrographic Investigations," J. Fac. Sci. U. Tokio, 1(2), 139-140 (1926). 51. Valasek, J., "Optical Constants of Sub- stances Which Exhibit Metallic Reflexion," "Int. Crit. Tables," 5, 248. 56. West, C. D., "Measurement of Refractive Indices of Resins and Plastics," Ind. Eng. Chem. Anal. Ed., 10, 627-628 (1938). 53. JoHANNSEN, A., "Manual of Petrographic Methods," p. 238. McGraw-Hill, N. Y., 1914. 54. WiNCHELL, N. H., AND WiNCHELL, A. N. W., "Elements of Optical Mineralogy," p. 69, J. Wiley & Sons, N. Y., 1922. 55. Ollivier, H., "Cours de Physique Gen6- rale," p. 252, Hermann, Paris, 1923. 56. Wright, F. E., "The Methods of Petro- graphic Microscopic Research," Carnegie Inst. Wash. Publ., 158, 95 (1915). 57. Wright, F. E., "Measurement of the Re- fractive Index of a Drop of Liquid," J. Wash. Acad. Sci., 4, 269-279 (1914). 58. Chamot, E. M., AND Mason, C. W., "Hand- book of Chemical Microscopy," Vol. 1, 2nd Ed. p. 358. J. Wiley & Sons,"N. Y., 1938. 59. Martin, L. C.,, "Optical Measuring Instru- ments," p. 164, Blackie & Sons, Phila., 1924. 60. Wood, R. G., and Ayliffe, A., "Method of Analyzing Small Crystals for Optical Prop- erties," J. Sci. Instr., 12, 299 (1925); Phil. Mag., 21, 321 (1936). 61. Wood, R. G., Ayliffe, A., and Williams, G., "Method of Analyzing Small Crystals for Optical Properties," Proc. Roy. Soc. London, A-177, 140 (1940-41). 62. KiENLE, R. H., AND Stearns, E. I., "Adapta- tion of Automatic Spectrophotometer for Special Measurements," Instr., 20(11), 1057-1063 (1947). 63. Addey, F., "A Note on the Measurement of the Vertical Dimensions by Use of the Graduated Fine Adjustment of the Micro- scope," /. Qekett Micros. Club (2), 14, 279 (1922). 64. Benedetti-Pichler, A. A., and Spikes, W. F., "Introduction to the Microtechniques of Inorganic Qualitative Analysis," Doug- laston, N. Y., 1935. 65. Blunk, G., "Quantitative Bestimmung Physikalische Chemischer Eigenschaften Mikroscopish Keiner Mengen," Zeitschr. wiss. Mikros., 37, 138-40 (1920). 66. Eversmann, Th., "Bemerkung zur Mittel- baren Langennessung mit Distanz faden nach Reichenbach," Zeitschr. f. Instr., 52, 525-528 (1932). 67. Kirk, P. L., and Gibson, C. S., "Refractive Index Measurements in Qualitative Or- ganic Microanalysis," Ind. Eng. Chem. Anal.Ed.,ll,\^'i (1939). 68. Reimers, F., "Investigations of Microchemi- cal Methods. IV Identification by Means of Microdetermination of Refraction," Z)ans^-. Tids. Farm., 15, 81-99 (1941). 69. Kegeles, G., Costing, L. J., Hansen, E. M., AND Morris, M., "Equipment and Experi- mental Methods for Interference Diffusion Studies," Rev. Scient. Instr., 20(3), 209- 215 (1949). 70. Thovert, — ., Ann. Chim. Phijs., 2(9), 369 (1915). 71. Zuber, R., Z.Physik., 79, 280-290 (1932). 72. Longsworth, L. G., "Optical Methods in Electrophoresis," Ind. Eng. Chem. Anal. Ed., 18, 249-269 (1946). 73. Debye, p., "A Photoelectric Instrument for Light Scattering Measurements and a Dif- 499 REFRACTION OF LIGHT, REFRACTOMETRY AND INTERFEROMETRY ferential llefractometer," J. Appl. Phys., 17(5), 392-398 (1946). 74. Kegeles, G., and Gutter, F., "Determina- tion of the Sedimentation Constants from Fresnel Diffraction Pattern.s," J.A.C.S., 69, 1302-1305 (1947); 73, 3770-3777 (1951). 75. L.-VMM, O., "Measurements of Concentration Gradients in Sedimentation and Diffusion by Refraction Methods. Solubility Proper- ties of Potato Starch," Novo Acta Regiae Soc. Sci. Upsaliensis, IV,10(6), 1-115 (1937). 76. Grauer, B., "Prismatic Refractometer," Nat'l. Bur. Std. Tech. News Bull., 37(9), 535 (5953). 77. TiLTON, L. W., AND Taylor, J. K., "Refrac- tive Index Measurements," in: Berl, W. C, "Physical Methods in Chemical Anal- ysis," Vol. 1, Academic Press, N. Y., 1950. 78. Lauer, J. L., and King, R. W., "Refractive Index of Liquids at Elevated Tempera- tures," Anal. Chem., 28(11), 1697-1701 (1956). 79. RoTHEN, A., "Optical Properties of Surface Films," Ann. N.Y. Acad. Sci., 153, 1054- 1064 (1951). 80. Dayhoff, E. S., "Correction in High Ac- curacy Fresnel Region Microwave Inter- ferometry," N. B. S. Tech. Rep. 4514 (Feb., 1956). 81. Raveau, C, "Etude des Franges des Lames Crystallines au Moyen de la Surface des Indices," Bull. Soc. Fr. de Mineral., 34, 6-12 (1911); "Sur la Visibilite et les Singu- laritds des Franges d'interference," Resum. Comm. Soc. Fr. de Phys., p. 40, (April, 1901). 82. LowRY, T. M., AND Allsopp, C. B., "A Pho- tographic Method of Measuring Refractive Indices," Proc. Roy. Soc. London, A-126, 165 (1929). 83. Lord Rayleigh, "On Some Physical Prop- erties of Argon and Helium," Proc. Roy. Soc. London, 59, 200-217 (1896). 84. Williams, W. E., "Studies in Interferome- try," Proc. Phys. Soc, 45, 699-727 (1933). "Applications of Interferometry" Methuen, London (1928). 85. HiRSH, P., "The Refractometer and the In- terferometer," Z. Nahr. Genuscm., 43, 65- 78 (1922); "Die Abderhalden-Reaktion Mittels der Quantitative!! Interferome- trischen Methods," Deutsch. Klin. Wchn- schr., 4, (1365-1366 (1925); Deutsch. Klin. Wchnschr., 4, 1560-1562 (1925). 86. Cotton, A., "L'lnterf^rometre de Jamin a Faisceaux Polarises," Rev. d'Opt. Theor. et Instrum., 5, 153-166 (1934). 87. Brillouin, M., "Recherches sur I'ellipti- citd du G^oide," C. R. Acad. Sci., 137, 786- 788 (1903); "M^moires des Savants Stran- gers," 23(3) (1908). 88. Barchewitz, P., "Etude sur I'interf^rome- tre a rayous Polarises de Jamin," Rev. d'- Opt. Theor. et Instrum., 5, 167-178 (1934). 89. Smith, F. H., U. S. Pat. 2,601,175, June 17, 1952. 90. Sagnac, G., Le Radium (July 8, 1911). 91. Ramsay, B. P., "A grating interferometer" /. Opt. Soc. Am., 24, 253-268 (1934). 92. Barus, C, "Elliptic Interferences," Carnegie Inst. Wash. Publ. No. 149: (1911-1914). 93. Marton, L., "Electron Interferometer," Phys. Rev., 85(4), 1057-1058 (1952). 94. Marton, L., Simpson, J. A., and Suddeth, J. A., "An Electron Interferometer," Rev. Scientific Instr., 25 (10), 1099-1104 (1954). 95. Simpson, J. A., "Electron Interference Ex- periments," Rev. Modern Phys., 28(3), 254- 260 (1956). 96. Marton, L., "Electron Physics," Nat. Bur. Std. Circ. 527. 97. Jonnard, R., "R^fractometrie Interf^ren- tielle et Structure du S^rum Sanguin," Maloine, Paris, 1937. 98. Jonnard, R., "R^fractometre Interferen- tial Pour Usages Biologiques," Rev. d'Opt. Theor. ei Instrum., 15, 425-430 (1936). 99. DuFFiEux, N., "Fente Simple Pour Spectro- graphies," Rev. d'Opt. Theor. et Instrum., 15, 298 (1936). 100. Lendel, E., Zimmer, A., and Fehlow, W., Munch. Med. Wchnschr., 37, 1-16 (1927). 101. Bruins, H. R., Rec. des Trav. Chim., 50, 121- 128 (1931); Koll. Zeitschr., 54, 265-271 (1931); 54, 272-275 (1931); 57, 152 (1932); 59, 263 (1932). 102. Jonnard, R., "Automatic Electronic Record- ing Interferometer. Il-Performance," O. N. R. Technical Report, March, 1957. 103. Ollivier, H., "Cours de Physique Gen6- rale," 3, 252, Hermann, Paris, 1923. 104. Jonnard, R., "The Orthoptic Microscope Interferometer. Theory, Standardization, Applications to Precision Interferential Refractometry." Proc. 7th Nat. Conf. I.S.A., Sept. 8-12, Cleveland, O., 1952. 105. Jonnard, R., U. S. Patent 2,858,728, Nov. 4, (1958). 106. Stockbarger, D. C, and Bitrns, L., "Line Shape as a Function of the Mode of Spectro- graph Slit Irradiation," J. Opt. Soc. A7n., 23, 379 (1933). 107. Jonnard, R., "Random Selection System for 500 HISTORY OF LIGHT REFRACTION Automatic Dynamic Biochemical Analy- sis," Ann. N. Y. Acad. Sci. Conf. on Auto- matic Analysis, Nov. 12, 1959. 108. Ferranti, a., Brit. Pat. 760321, 1958. 109. Peck, E. R., and Obetz, S. W,, "Wavelength or Length Measurements by Reversible Fringe Counting," J. Opt. Soc. Ain., 43(6), 505-509 (1953). 110. Ingelstam, E., "A Phase Contrast Refrac- tometer With High Accuracy for Gases and Liquids," Ark. f. Fysik, ^6(29), 287-316 (1953). HI. JoNNARD, R., "Recent Advances in Inter- ferometry." Trans. N. Y. Acad. Sci. II, 15, 269-280 (1953) ; "Biological Applications of Interferometry," Seminar, Brooklyn Polytech. Inst., Dec. 1953; "The Nature of Biological Information," Chem. Eng. News, 32, 595 (1954). 112. HuNTOON, R. D., Weiss, A., and Smith, J. R., "Electronic Fringe Interpolation for an Optical Interferometer," Nat. Bur. Std. Report 1228, May 15, 1952; /. Opt. Soc. Am., 44(4), 264-269 (1954). 113. Decker, M. M., and Mueller, H., "Trans- mitting Data by Light Modulation," Con- trol Eng. (July, 1957). 114. JoNNARD, R., "Automatic Electronic Re- cording of Refractive Dispersion by Inter- ferometry," Proc. First Internat. lustrum. Cong, and I.S.A. Conf. Sept. 13-23, Phila., Pa., 1954. 115. JoNNARD, R., "Automatic Electronic Record- ing Interferometer. I-Design and Construc- tion O. N. R. Technical Report," August, 1956. 116. JoNNARD, R., "Adventures in Laboratorj^ Au- tomation," Cybernetica, 2(3), 152-188(1959). 117. JoNNARD,R.,L'analysedes Composes Lourds par R^fractometrie Interf^rentielle. Bull. Soc. Chim. Biol, de Paris, 21, 1185-1193 (1939). 118. LoRENTZ, H. A., "Ueber die Beziehrung Zwi- schen der Fortplanzungsgeschwindigkeit des Lichtes und der Korper Dichte." /. Wiedinann Ann. des Phys. und Chem. 9, 641-665 (1880). 119. Gibson, R. E., and Kincaid, J. F., "The In- fluence of Temperature and Pressure on the Volume and Refractive Index of Benzene." J.A.C.S., 60, 511-518 (1938). 120. Kurtz, S. S., and Ward, A. L., "Refractivity Intercept and the Specific Refraction Equa- tion of Newton. I. Development of the Refractivity Intercept and Comparison with Specific Refraction Equations," J. Franklin Inst., 222, 563-592 (1936). II. The Electronic Interpretation of the Refrac- tivity Intercept and of the Specific Refrac- tion Equations of Newton, Eykman and Lorentz-Lorenz. J. Franklin Inst., 224, 697-728 (1937). 121. SwiETOSLAWSKi, W., "Atomic Refraction." J.A.C.S., 42, 1945-1951(1920). 122. Lagemann, R. T., "A Relation Between Vis- cosity and Refractive Index." J.A.C.S. 67, 498-499 (1954). 123. Debye, P., "Polar Molecules." Chem. Cat., N. Y., N. Y., (1929). 124. Nutting, P. G., "Dispersion Formulas Ap- plicable to Glass," /. Opt. Soc. Am., 2, 61 (1919). 125. Wright, F. E., "The Petrographic Micro- scope in Analysis." J.A.C.S. 38, 1647-1658, (1916). /. Opt. Soc. Am. 5, 389-395 (1921). 126. Chaudron, p., "Gaseous Equilibria Measure- ments." Thesis, Univ. of Paris, fac. Sc. Mason, Paris (1921). 127. Thompson, W. R., Hussey, R., Tennant, R., AND Campbell, N. O., "Effects of Ra- diations on Biological Systems. I. Studies in Respiration." J. Gen. Physiol. 16, 207- 220 (1932). 128. Bell, F. K., and Krantz, J. C, Jr., "An Interferometer Method for the Assay of Nitrous Oxide." /. A. Pharm. Assoc, 29, 126-130 (1940). 129. ScHONROCK, L., "Refractivity of Sucrose Solutions." Int. Critical Tables, 2, 336. 130. Raw^ R., "The Control of Fractional Distil- lation by Means of Refractive Index Meas- urements." J. Soc. CAem./nd., 66, 451 (1947). 131. Kirk, P., "Density and Refractive Index. Their Application in Criminal Identifica- tion." A. Thomas, 1951. 132. JoNNARD, R., "Determination of Total Nitro- gen in Proteins and Their Hydrolyzates." Ind. & Eng. Chem. 17, 246-248 (1945). 133. Paic, M., AND Deutsch, v., "Dosage R^- fractometrique des Prot^ines Seriques." C. R, Ac. des Sc, 199, 1306-1308, 1934. 134. Cleason, S., Ark. Kem. Mineral. Geol. 23- A(l), (1946). 135. TisELius, A., and Cleason, S., "Adsorption Analysis of Amino Acids and Peptides," Arkiv. Kemi. Mineral. Geol., 15-B(6), 18-23 (1941). 136. HoLMAN, R. T., AND Hagdahl, L., "Tiselius- Cleason Interferometric Absorption Ap- paratus," Anal. Chem., 23, 794-797 (1941). 137. SvENSON, H., "Some Improved Forms of the 501 KEFRACTION OF LIGHT, REFRACTOMETRY AND INTERFEROMETRY Differential-Prismatic Cell." /. Opt. Soc. ylm.,44C2), 140-146 (1954). 138. Kegeles, G., "A New Optical Method for Observing Equilibrium." J.A.C.S., 69, 1302-1305 (1947). 139. GiBERT, R., /. Chim. Phys.,U, 37-39 (1947). 140. Calvet, E., and Chevalerias, R., "A New Interference Method for Studying Diffu- sion in Liquids," J. Chim. Phys., 43, 37-53 (1946). 141. CouLsoN, C. A., Cox, J. T., Ogston, G. T., AND Philpot, J. St. L., Proc. Roy. Soc. London, A-192, 382 (1948). 142. GuiLLAUMiN, Ch. O., "Remarques Tech- niques sur I'interferometre de Zeiss et la Methode de Hirsch." Bull. Soc. de Chim. Biol., 15, 1393-1414, 1933. La Reaction S^- rointerferometrique de Hirsch. Conf. Soc. Pharm. Paris, June 6, 1934. 143. JoNNARD, R., "Titration des Antigenes et des Anticorps par Refracom^trie Interf^ren- tielle." C. R. Soc. Biol., 127, 418-421 (1938). 144. JoNNARD, R., "La Titration du Serum Antifievre Scarlatine par R^fractom^trie Interferentielle." C. R. Soc. Biol, 128, 263- 265, 1938. La Titration des Antigenes Mi- crobiens et des Scrums Immunes par R^- fractometrie Interferentielle. Biol. Med., 28, 469-512 (1938). 145. JoNNARD, R., AND LouviER, R., "L'affinit^ du S6rum Sanguin Pour les Hormones Sexuel- les Determinee par R^fractom^trie Inter- ferentielle." Rev. Pathol. Comp. et Hijg. Gen., 512, 1-4 (1939). 146. JoNNARD, J. R., "L'analyse des Composes Biologiques Lourds par R^fractometrie Interferentielle." Bull. Soc. Chim. Biol, de Paris, 21, 1185-1193 (1939). 147. Ingelstam, E., Djurle, E., and Johans- son, L., "Precision Concentration Analysis of D2O/H2O by Means of Phase Contrast Refractometry." /. Opt. Soc. Am., 44(6), 472-77 (1954). 148. Jonnard, R., "Etude Interf^rom^trique de la Refraction du Serum Sanguin en Fonction de la Concentration." C. R. Ac. Sc, 203, 124 (1936). 149. Jonnard, R., "Etude de la Stability du S^rum Sanguin." C. R. Soc. de Biol., 121, 841-843, (1936). 150. Jonnard, R., "Le S^ro-Diagnostic Inter- fdromdtric Devant la Physique," Rev. de Pathol. Comp. et Hyg. Ge«.,483, 1-7, 1936. 151. Jonnard, J., "Etude Interferomdtrique de la Refraction du S^rum Sanguin en Presence d'eiectrolytes." C. R. Soc. de Biol., 122, 1305-1306 (1936). 152. Jonnard, R., and Zuckerkandl, F., "Etude de la Refraction du Serum Sanguin en Presence d'eiectrolytes." C. R. Soc. de Biol, 122, 1315-1316 (1936). 153. Jonnard, R., Faillie, R., and Zucker- kandl, F., "Etude du Pouvoir Reactionel du serum Sanguin par la Methode Refrac- tometrique." Biol. Med., 28, 1-23 (1938). 154. Jonnard, R., "Refractometrie Interferen- tielle du serum de Cancereux." Bull. Soc. de Chim. Biol de Paris, 19, 893-897 (1937). 155. Jonnard, R., "Individualidad Fisico-qui- mica de los Constituyentes del Suero San- guines." Bol. Inst. Med. Exp. Buenos Aires, 43, 529-436 (1937). 156. Jonnard, R., "InvestigacionesFisico-quimica Sobre el Suero de Cancerosos," Bol. Inst. Med. Exp. Buenos Aires, 44, 141-158 (1937). 157. Jonnard, R., and Ruskin, S. L., "Studies in Calcium Metabolism." Am. J. Digest. Dis. 5, 676-680(1938). 158. Jonnard, R., and Ruskin, S. L., "Etude Physio-chimique Comparee du Glu- conate de Calcium, du Cevitamate, et de la Vitamine C." C. R. Soc. de Biol, 128, 286- 288 (1938). 159. SvENSoN, H., and Forsberg, R., "A New Optical System for Simultaneous Record- ing of Refractive Index and its Gradient in Stratified Solutions." J. Opt. Soc. Am. 44(5), 414-416 (1954). 160. Conn, G. K. T., and Eaton, G. K, "On a Systematic Error in the Measurement of M Optical Constants." /. Opt. Soc. Am. ' 44(6), 477-480 (1954). See additional bibliography, p. 522. R. Jonnard INTERFEROMETRIC METHODS'' Besides the angle refractometer (q.v.) a second group of refractometrie methods in- volves direct determination of the velocity variation which radiant energy undergoes * The writer's experimental work with the recording interferometer was entirely supported by the Physics and Physiology Branches, Office of Naval Research, U. S. Navy, under a series of contracts granted to the Paterson General Hospital, Paterson, N. J. and Columbia University College of Physi- cians and Surgeons, New York, N. Y. from 1954 to date. 502 INTERFEKOIVIETRIC METHODS when traveling through transparent media. While absolute measurements of the velocity of light are difhcult, evaluation of relative variations is relatively easy.- The instru- ments used for this purpose are generally called interferometers. Today, radiant energy is considered as an electromagnetic perturbation the two fields of which are normal to one another and also normal to the direction of propagation of the perturbation: the vibrational field is trans- versal. A monochromatic radiation is char- acterized by its 'period, T, or time in seconds required by one energy wave to effect a com- plete oscillation, or in Maxwell's theory for the corresponding line of force h to effect a complete rotation around the point 0. During this time T, the perturbation travels in a direction x with the velocity C (light velocity in a vacuum). The distance traveled is: C-T = X m this being called the wavelength, measured in Angstrom units.* The period T is not directly Only in a vacuum, C being the radiations, is X directly propor The frequency, v, of a radiation ber of oscillations contained in CT over a period of one second measurable, same for all tional to T. is the num- the distance V = C/\ US) Frequencies are expressed in Fresnel units (10~^2). The wave numbers, v, are more con- veniently handled than either X or v. They are the number of vibrations contained in 1 cm: 1 _ l.lQs cm X ~ X( in A) (cm-i) (M) The quantities C and X vary with the me- dium traversed, but v (and T) do not. The latter quantities characterize the conditions under which the radiation is generated ( ex- cited orbital, atomic number, temperature, * See equations 1-11 in two preceding articles. etc.). The quantity C (or X), on the contrary, characterizes the conditions under which the radiation travels. The value of C can be calculated in an- other way, by means of the classical dimen- sional equations. Both Coulombic electrostatic and mag- netic attraction forces between two unit charges e can be measured. For the electric attraction one has: F = ± 1 ^ or or in dimensional form: [el = Ko^'Hl^i^Ui^T-^ (15) (16) Since a moving charge e generates a mag- netic field, the latter may also be a measure of e: [e] = [Mi/2Li/Vo-^'2] (17) (since force = mass X acceleration = MLT-^, and d^ = U). The two above quanti- ties are equal: [M^i2.L^i2.^-ii2] = [Ko^im^i^L^i^T-^] (18) or [LT-^[m-"^-K,- 1/2 1 = Va'o- = V (19) MO Hence, the quantity V found has the di- mension of a velocity whose value is equal to the ratio of the e s u and emu units. It was experimentally found equal to the ve- locity of hght ill vacuo (Weber, 1856) : C = 2.9986 X IQio cm/sec. The value of C calculated above repre- sents a maximum. It is not the only charac- teristic velocity of electromagnetic phe- nomena which needs to be considered. The group-velocities of Rayleigh are useful in interpreting the peculiarities of refractive dis- persion within an absorption hand, even when "monochromatic light" is used, for even the finest spectral lines have always a finite 503 REFRACTION OF LIGHT, REFRACTOMETRY AND INTERFEROMETRY width within which the spectrum is con- tinuous. A discussion of group-velocities falls out- side the scope of this article, except where the refractive index is concerned. The group-velocity U is related to the re- fractive dispersion by the formula: U = C/[n - {\-dn/d\)] {20) Since dn is large compared to d\, U tends to become much smaller than C, which repre- sents a maximum: the refractive index n = C/U for the bundle becomes greater than that for any one of the radiations it contains. The group-velocity concept has recently acquired a fundamental importance, as the cornerstone of a generalized theory of Re- fraction. This new theory is now capable of encompassing the situations arising when the radiation emitters are in fast motion relatively to the medium of index n. These phenomena include the emission of the Cherenkov-Vavrilov polarized radiation, the Doepler effect at super-light velocitj^, and others. They have an immense practical im- portance for the interpretation of all high- energy high-velocity phenomena involved in nuclear energy investigations. The com- plete theory of interferences at superlight velocities how^ever, remains to be fully de- veloped. The optical pathway Li of light through a transparent medium is related to the time t required to traverse it by If one of the media is a vacuum (no = 1), it results that the optical path length is the product of the geometrical length of a me- dium l\y its refractive index relative to vacuum (ni). This important relation is the basis for the determination of relative re- fractive indices by the interferometric methods, involving a direct comparison of the products nL in two media, one of which is known. In Huygens vibrational theory it is postulated that "something along the path of a light ray is vibrating according to a sinusoidal function of the general type : L = A sin 2t ^^ - ^ + a) (24) t = Li/Fi , or Li = t-Vi (21) A similar relation holds for any other me- dium in which the velocity will be V2 . If during the same time t, light traverses a layer Li of medium 1 and a layer L2 of medium 2, one can write: The correctness of this theory became ap- parent when Thomas Young succeeded (1813) in composing the vibrations issued from two synchronous light sources (pin- holes lighted from the same primary source) whose interference thus extinguished the light. By application of Huygens theorem to the space between the primary source of radiant energy and the plan containing Young's twin apertures, it is possible to demonstrate that the rays issuing from these apertures are formed of synchronous vibra- tions. Such rays are said to be "coherent". In this memorable experiment two syn- chronous point sources of light A and B, sep- arated by a distance a produce two narrow beams of light falling upon the same area c on a screen placed at the distance d. The amplitude L of the synchronous vi- bration X issued from A and B is: L = An sin 27r iH (25) UIU = ^Fl/F2 {22) Since F1/F2 = n2 or relative index of me- dium 2 compared to medium 1, one has h\ = TfltLl {2S) The difference p between the paths of the light rays Ac' and Be' falling on c' (c — c' = z) is: p = Be' - Ac' = X2- xi (26) and the difference of phase <^, of the two 504 IMEKFEKOMETRIC METHODS superimposed \ibrations in c' is equal to the difference of amplitude of the two vibrations falling in c' at the same time :

/^ ^ v'..aa •Jii^* Jjriy.. Fig. 4. Polyacrylonitrile as thin, self-support- ing film cast from dilute solution in ethylene carbonate. Shadowed at 10° with U and electron micrographed. Shows Type I particles some as monolayers partially covering holes (broken bub- bles) in the film. 72,000X ilarly in natural or synthetic latices, emul- sions and suspensions, the colloidal particles are involved in failures such as by freezing, in consolidation such as by smoking natural rubber latex or in performance such as in painting or printing. Particles of Type II may be perceptible light microscopically and generally are perceptible electron micro- scopically. Type III. These may be aggregates of Type II particles. (Figure 7) More and more resins are being formulated to meet useful specifications by "alloying" two or more phases of the same or different component (monomer) in a system. For example, a latex may be added sometime during the poly- merization of styrene to make a more im- pact-resistant material. (Figure 8). During the consolidation with polystyrene the rub- bery colloidal particles may be greatly en- larged or a rubbery phase may be milled into an adliesive to give resilience to the finished joint. A crystalline phase may be introduced to a system to give rigidity. That is particles of Type III may be amor- phous or crystalline and of the same or differ- ent stereoisomer. The amorphous phase (s) may be in the rubbery and/or glassy states (Figure 9). Type III particles and parts are generally visible in the light microscope. It should provide illumination as bright-field or dark- field, by transmitted or reflected light, polarized or unpolarized, as needed. Ability to change from one kind of illumination to another should be quick and easy (1, 2). Figs. 5(A) Emulsion polymerized polyacrylo- nitrile, as dried and electron micrographed. Type II particles, useful as a latex or in consolidation. 34,000X . (B) Same polj^mer after pressure-molding and fracturing. Positive electron micrograph of replica shows that resin came apart between same type of particles as were molded together. 34,000X . 528 RESINOGKAPIIY Type III units of each phase not only can be related to their origin and functions by their light microscopic morphology, but also by their optical properties such as relative and specific refractive indices * and kind and degree of double refraction. Some other physical properties can generally be deter- mined such as softening temperature, scratch and indentation hardness. Such tests usually can be done on the discrete phase in situ ("intangible isolation (6)")- Some other de- structive tests can be devised if the discrete phase is tangibly separable from a suffi- ciently large sample. Type IV. This may be a system of phases of the whole material. The phase(s) may be amorphous, continuous or discontinuous, rubbery or glassy, crystallizable or non- crystallizable. If a grafted part shows a boundary with the original polymer, it is best described as a separate unit of Type R". So are other coatings and layers of one phase each. For example, there may be a precoating of phenolic resin on each of two anodized aluminum sheets with an inter- mediate layer of rubber in between (1). In every case, each unit has its own physical- chemical origin and purpose, and it is im- portant that each be observed separately during process-control and development of use. Fibers and foils are special examples of Type IV; they are so important that there are special technologies, societies and schools to study them. Yet the study of fibers and foils must be considered as resinography. They are highly polymeric in origin, have important uses in resins as reinforcing parts, have characteristic directional morphology and many optical properties. These proper- ties are generally those of particles of Type I or II which have been oriented to some extent, usually by the application of exter- nal stress or by artificial conditions (e.g., spinning and drawing a fiber; casting a film) (Figure 10). Fig. 6. Phenol-fonnaklehyde polymer, experi- mental fracture surface. An electron micrograph of a positive replica. The larger type of particle is classified as Type II and the smaller is probably Type I. Fig. 7. Polyacrylonitrile in a very earl}^ stage of bulk polymerization. An electron micrograph showing particles of Type III composed of par- ticles of Type II (5). Crystals on the other hand are composed of macromolecules (Type I) which grow by their own attractive forces into geometric or skeletal shapes of characteristic habit and phase. The crystalline habits of any one phase are characteristic of variations in crystallizing conditions. Transformation of one crystalline phase into another is far less common than it is among metallic alloys. 529 RESINOGKAPHY •v' » jA •**■ v^^ r 4f , ra 0 s»«r •ii •^ 7 * s- - t. "" .4' ^ ^. :/. •-^- /ao/^ -*!• "H Fig. 8. Two "alloys," high in polystyrene containing small amounts of synthetic rubber as particles of Type III, illustrating also two methods of preparation and two methods of lighting. Left: Polished and etched (benzene in methyl alcohol ca. 1:4); bright-field reflected illumination. Right: Microtomed thin section; bright-field, slightly oblique, transmitted illumination. Fig. 9. Polystyrene varjang across the field in proportion of isotactic (crystallizable) and atactic (non-crystallizable). Light micrograph of unstirred sample between partially crossed polars, by trans- mitted illumination. Melt quenched in air, warmed slowly until glassy phase, such as (G) was almost entirely transformed to continuous rubbery (R) phase, held at slightly lower temperature to form some colonies (X) of crystals, quenched again to freeze the 3 co-existent phases, not in equilibrium. Some cracks are isotropic (shown black); some are anisotropic (shown white). The crystalline parts of Type IV are gen- (e.g., undulate extinction between crossed erally aggregates of unicrystals (Types II & polars), and practical properties (e.g., inter- Ill) (Figure 11). Such aggregates have their granular fracture). There is some evidence characteristic habit (e.g., spherulitic), origin that at least some kinds of aggregates are a (e.g., self -nucleating) , optical properties combination of crystalline (internal) forces 530 RESLNOGRAPHY and applied (external) stress. Starch grains are natural aggregate units quite character- istic of the botanical species. Paper sheets, woven or nonwoven fabrics, roving threads, tire cords, and reclaimed rubber are examples of systems of phases or materials that may be only part of the fab- FiG. 10. An experimental fiber of polyacrylo- nitrile. A positive electron micrograph of a silica replica. The fiber represents a unit of Type IV. The ridges of the "bark" running parallel to the fiber-axis apparently contain both Types II and I particles. So does the core, exposed in the upper left part of the micrograph but here the particles are less ordered. ricated whole. It may be very important to recognize their areas and boundaries and to determine how they may be reacted with or anchored to the contiguous materials. Con- versely, fibers, foils, sheets, tubes, fabrics and so forth may have another resin phase attached by grafting, impregnating or coat- ing. A very important kind of Type IV is the tnolded grain. This is generally supposed to represent the material of the molded piece and to contain the resin, fibrous and/or particulate fillers, pigments, lubricants, etc. Yet after molding, the original pellets of the molding grains can be seen before or after preferential relief polishing, etching, stain- ing or breaking at grain boundaries. The resinographic problem can be com- plicated. For instance, starch grains may be incorporated into a urea-formaldehyde ad- hesive binding cellulosic and metallic sheets. To differentiate the starch grains from the synthetic resin would one examine thin sec- tions (with reflected fight) or thick sections (with transmitted light) ? Combination with Other Materials. If the resin is combined with metal the method //^ •■ -^ ♦- / A^ Fig. 11. Commercial foil (Type IV material) manifesting three types of particles. Electron micro- graphs of positive replica of rougher side, no shadowing metal. Left: Spherulitic grains (T3^pe III) com- posed of crystallites (Type II). Right: Higher magnification. Apparently there are small particles (prob- abl}'- Type I) arranged roughly in line within the variously oriented crystallites. 531 RESINOGRAPHY of examination of the combination would to examine the surfaces of Type IV particles, probably follow the techniques of metallog- inserts, laminates, possible cracks and voids, raphy and microscopy by reflected light. If etc. of fabricated, cut or fractured surfaces the resin is combined with rock, mineral during manufacture, test or use. Only mor- or ceramic, the techniques would be related phology is revealed and such information is to petrograph}'^ and transmitted light, or min- limited ; optical properties or patterns can- eralograph}^ and reflected light. not be studied without provision for polariz- If common amounts of pigments or fillers ing the light, are added to the resin, the product is gen- In the resinographical laboratory three erallj'^ so opaque that reflected light is the general kinds (2) of compound microscopes only light-microscopical illumination feasi- will be essential if all 4 types of units are to ble. It is very satisfactory. be considered. These microscopes, in the recommended order of their acquisition are: Equipment (^^^ Stereoscopic, hiobjective (Greenoiigh) If the resinographer is in a narrow field light microscope for preliminary work and of making, fabricating, testing or using res- some units of Types IV and III (q.v.,p. 538). ins, his microscopical equipment, accessories (2) Monoobjective, 'polarizing light micro- and techniques may be definitely specified, scope ecjuipped for both transmitted and re- If, however, the resinographer is in research, fleeted illumination for more precise work development, consulting or a very broad with units of Types IV and III and some field of manufacture or use, he will need a units of Type 11. wide variety of equipment, accessories, tech- (3) Electron microscope, especially for niques, experience and personal help. Fur- seeing morphology of units of Type II and, thermore, the manufacturer of resins which at highest powers and best performance, are to be sold for many purposes may have some units of Type I. a much broader interest in the texture. Stereoscopic Microscope with Polars. structure and combination of a resin than The kind of microscope recommended for the fabricator or consumer who has a specific preliminary examination, preparation, mi- interest. Indeed, the resinographer of a cromanipulation, and experimentation on primary manufacturer has broad interests resinous materials is the Greenough type, and responsibilities of health, safety, com- biobjective, stereoscopic microscope fitted petition, "security" and fundamental re- with polars (2). This type of microscope is search ; his microscopical equipment must be generally excellent for the examination of as versatile as he is. particles (IV) of moldable or molded grains, Away from his laboratory the experienced crystalline aggregates, some fillers, most resinographer finds that he can recognize inserts, many foreign particles, most 1am- and study resins a great deal better with a inates; some single crystals, multi-phase simple magnifier (3) than with the unaided systems, most cracks, many voids. The ob- eye. For example, a lOX pocket magnifier, jectives have such long depth of focus {ca. correct for aberrations, can be carried at all 25 m) that the object might be studied as it times. It gives an erect, virtual image that is is or after fracturing, cutting, sawing, and/or easy to interpret. Its chief disadvantage is abrading into thick or thin sections to reveal that it must be held close to both the object all of the pertinent views. and the eye. This simple magnifier or a Reflected illumination, whether direc- pocket compound microscope (3) is recom- tional, or symmetrical, is typically dark- mended for field work, within limits and field in the Greenough microscope. That is, within the experience of the resinographer, the illuminator is outside the objectives, and 532 RESINOGRAPHY light gets into the microscope and eye as a walls, floors and table-tops, so as to examine result of scattering by the object. Conse- them without destructive sampling. Special quently, paper, for example, appears white stands may be improvised, and a glass slide appears dark. Dark-field The biobjective (Greenough) microscope images by reflected light are the easiest to has inherent limitations of resolving power, interpret because they are of the same type The two objectives must be so close and at as received by the unaided eye. such an angle that their numerical aperture Transmitted illumination is typically is not much over 0.1. In fact, objectives of bright field in the Greenough microscope, all available magnifications have about this Paper, for example appears dark and a glass same numerical aperture, which is not much, slide appears bright. if any, greater than a hand lens. Therefore Reflected and/or transmitted light may the stereoscopic microscope cannot reveal be used with adequate control and inexpen- structure separated by much less than 5 ix sive lighting. With the simultaneously ro- (0.005 mm). The kinds and quantities of tatable polars and transmitted light, double properties are limited not only by the ob- refraction can be detected and studied during jective and focusing device but also by the heating or cooling under stretch or compres- illuminating systems. sion, or with other physical-chemical treat- Therefore, the resinographer's next micro- ments. scopical requirement would probably be a With its two objectives and an eyepiece high powered, polarizing light microscope. for each, the Greenough microscope is two Polarizing Light Microscope of High compound microscopes, one for each eye at Power. In resinography a light microscope of about the angle for maximum stereoscopic moderate power and quality of image is re- eft'ect. There should be little eyestrain or quired for the smaller sizes of Type IV units fatigue, provided that the prisms are in (Figure 12, right). The highest powers and adjustment. The Greenough microscope has qualities may be needed for Type III; they reinverting prisms so as to give erect images, certainly will be needed for Type II when- This feature is especially convenient in ever such small units can be made visible micromanipulation and experimentation: with light. In order to cover the whole Something pushed "north" and "east" does range of samples from transparent to not seem to go "south" and "west." Erect opaque, high powers must be available by images are interpreted as they would be either transmitted or reflected illumination with the eye unaided: If a texture is bright (2). That is, there must be both a substage on the side toward a lamp the texture is an condenser and a vertical illuminator and elevation, not a depression. each should be of high aperture, of good The two stereoscopic objectives have long image quality and with a polarizer. The working distances so that physical or chem- analyzer can be common to both illuminat- ical operations may be performed on the ors. A binocular eyepiece (so much used by object; for examples: scratch or impression biologists) is not used with polars and so is hardness testing ; tensile or compression test- not recommended for resinography.. The ing; electric, magnetic or electrophoretic combination (2) of "petrographic" and "ore" experiments; chemical reactions; etching or microscopes is recommended. The "petro- staining. graphic" part with its condenser and polar- The commercially available stands are izer for transmitted illumination is used versatile and interchangeable. The micro- for emulsions, suspensions, fibers, films, scope on the simple mount can be placed mounted grains and prepared thin sections directly on flat objects such as sheets, boards, (Figure 8, right). The "ore" part with 533 RESINOGRAPHY Fig. 12. Experimental fibers of polyacrylonitrile showing variation in cross-sectional texture. Left: Electron micrograph of ultrathin section cut after coating with a resilient polyacrylate latex (R) and then mounting in methacrylate (4). Right: Light micrograph of a thin section showing that these ex- perimental fibers varied in the number of rings. vertical illuminator and polarizer is used for surfaces, opaque grains, polished thick sec- tions, opaque inserts, and optical sections of transparent objects (Figure 8, left). Along with other optical methods for en- hancing contrast, darkfield illumination by either transmitted or reflected illumination is handy, if not required. Dark field by re- flected light permits the distinction of pig- ment or streak color from specular color. In general, images by dark-field reflected illumination are easier to interpret because they are more natural than bright-field images. Among high polymers and resins obser- vation and measurement of optical proper- ties are important adjuncts to descriptive and interpretive morphology or texture. In order to observe most optical properties the light microscope must have variable aper- tures from low to very high, in both objec- tive and condenser, the polars must be crossable and uncrossable, and their direc- tions of vibrations must be known. There should be a rotatable object-stage. With transmitted illumination specific or critical refractive indices may be measured. Double refraction can be detected and meas- ured. Distinction can be made among the origins of the double refraction, such as single crystals, crystal aggregates, strain (Figure 9), deformation, orientation of aniso- tropic particles of resin, or orientation of tiny rods or plates of isotropic particles in an isotropic medium of different refractive index (2). The direction(s) of vibration of slower component (s), the sign of elongation or the optical sign can be determined di- rectly. The two or three specific refractive indices may be determined with a com- pensator and crossed polars. With only one polar and knowing its vibration direction, pleochoism and sign of color absorption can be determined quali- tatively and quantitatively in colorful, dyed or stained materials such as fibers. Rotated between crossed polars the be- havior of a polymeric specmien may be very different from that of a single, unstrained crystal. Examples are : crystalline aggregates, such as spherulites; natural aggregates, such as starch grains and natural fibers ; synthetic aggregates such as special polarizing films, wrapping foils and textile fibers. Crystals and other orientations of highly polymeric particles have characteristic dif- fraction patterns (interference figures) which are separate sources of determinative and 534 RESINOGRAPHY interpretive information from transmitted fabricating interface may be unknown — light. for example, with commercial fibers, foils, By reflected light, interference patterns laminates and moldings, (polarization figures) are being studied on The internal texture may be entirely dif- ore minerals, but such a study on resins is ferent from the external one. The problem to date onl}^ suggested. of representing the interior of the specimen Electron Microscope. High polymers, may be very difficult. Since the area needed alone or simply formulated with plasticizers, is small, a small part of a fracture surface curing agents and/or stabilizers are com- may be sufficiently flat (Figures 5B, 6, 14). posed generally of very small particles For such purposes fracturing is generally (Types II and I). Such particles are smaller simple when the sample is cooled to brittle- than the wavelengths of light and, for this ness. Even a film can be fractured in a planar reason, are not observable with a light micro- section if the substance can be obtained as scope. A stream of electrons at a potential of a liquid and used as an adhesive between 55 kv has a wavelength of about 0.05 A halves of two L-shaped pieces of sheet iron and even though the numerical aperture of or steel. After the "adhesive" is hardened an electron lens is extremely low (about and cold-embrittled it is usually fractured 0.02) a resolution as relatively poor as 100 A by holding one "leg" of the composite L in is more than ample to separate colloidal a vise and striking the other leg sufficiently particles (Type II) . Electron microscopes, of with a hammer. The fracture-surface is then or over 50 kv in potential, somewhat cor- replicated as usual. rected in astigmatism, clean and in good op- In a corollary manner a replica can be eration, can resolve distances of 20 to 10 A made of a single microtomed surface of fibers, or, perhaps, smaller. This kind of resolution foils, films or other soft materials. Grooves should resolve macromolecular particles will probably appear in the replica but these (Type I) of above approximately 20,000 in are readily related to microscopic nicks in molecular weight. The remaining problem is even the best cutting edge. A glass cutting to obtain adequate contrast in the bound- edge and a cantilever rocking microtome aries or among loci of influence. may be preferred to rotary microtome and If the sm-face of a high polymer or its a steel knife, resin is to be examined electron microscop- Mounting specimens for microtomy is ically, a negative, positive, or pseudo replica even more important in electron microscopy will probably serve the purpose (Figures 5B, than in light microscopy. Special precautions 6, 10, 11, 14). In any kind of replication of may have to be taken to avoid distortion, a high polymeric material, the primary repli- chattering and other effects. A very success- eating medium must not modify or stick ful method for fibers is to treat them w^ith a to the material. Therefore, replicating media synthetic rubber latex before embedding in are generally limited to aqueous media, such methacrylate (4) (Figure 12, left). as a "solution" of gelatin, methyl cellulose The most direct method of examining a or polyvinyl alcohol. Control samples of the high poljoner is to cast it into a film which replicating and photographic media should is sufficiently thin for electron microscopy be compared with the electron micrograph (Figure 4). This presupposes that the speci- of the specimen. men is obtainable in a fluid state, which will A "surface" of a polymer is generally an flow to a thin film and cure or dry to a film interface of the polymer with air, liquid or which is self-supporting and able to with- solid, and the electron microscopical image stand the incidence of the electron beam, must be considered as such even when the Linseed oil cast on water can be prepared 535 RESINOGRAPHY '—Tread ■m^0^. ^ ^ *#.iG P^eaker Stock Breaker Cord Cushion Stock — Carcass Cord + Stock Fig. 13. A cross-section of an automobile tire of unknown manufacturer. The section, cut with a wet razor blade, was photographed by oblique, reflected light. Shows the kind, numbers and distributions of the layers and indicates places to sample for identification of the kinds of rubbers and fibers. *p» f '"■'^ -"T. '^m^ ,. '* .n- . ^' k» ^^ tt **»«»». H -/^ 4 Fig. 14. Natural rubber tire tread stock filled with carbon black (shown as black round areas and also as commensurate depressions and eleva- tions). An electron micrograph of positive replica; no shadowing metal. Matrix also manifests tiny rubber particles of Type 1(7). this wa}^ (Figure 3). Cellulose esters (e.g., collodion) and some vinyl polymers (e.g., "Formvar" polyvinyl acetal) meet the require- ments so well that they are used as replicat- ing media. Gelatin, polyvinyl alcohol and cellulose ethers are hydrophilic and are used as replicating media when organic solvents are detrimental. These organic replicating media and substrates are high polymers and as such they could manifest particulate texture as large as that of the specimen. Instead, carbon, silica and other inorganic replicating media should be considered. With any replicating medium or substrate, "blanks" should be run. For similar reasons "blanks" should be run on all photographic materials. The general techniques of electron mi- croscopy and replication are described in detail elsewhere in this encyclopedia, and limitations as well as advantages of the electron microscopy have been published (2, 3). The specific advantage of the electron microscope is that it makes visible prac- tically all particles of Type II and significant ones of Type I. The limitations are related to the problems of representing the texture, gaining sufficient contrast and avoiding polymerization, scission or any other change due to the electron beam or vacuum. There are also the general limitations of resinography with respect to its infancy. 536 RESINOGRAPHY The technology is almost certain to develop as the microscopy, interpretations and con- clusions develop. Nature of Information Obtained Micro- scopically The nature of resinographical information has been discussed in terms of four kinds of architectural units and four kinds of micro- scopes. In conclusion, reference is made to the finished product, in terms of fabrication, end-use and gross-texture (1). These can only be suggested because of restricted space. In rubber tires, for example, (Figure 13) one may be interested in the thickness of tread or retread, occurrence of reclaimed rubber, distribution of carbon or white re- inforcing agent, layers of gum-stock, num- ber of plies of cord, kinds of cordage fibers, construction of bead, etc. In these and other rubber products one may be interested in variations in texture of the rubber due to poor distribution of accelerator, retardant or antioxidant. The distributions of sizes and shapes of pigments, fillers or abrasives may be important in rubber products, (Fig- ure 14) writing boards, floor and table cover- ings, pavements, and paints. Related to paints is the resinographic examination of inks, lacquers, pohshes, ad- hesives, mastics and casting resins. Mold- able resins are generally solids, as powders, grains or preforms. Some polymeric cements and adhesives are also marketed dry. In painted, inked, varnished or lacc^uered surfaces there may be the question of num- ber, thickness, uniformity and adherence of layers. There are similar problems with laminates of wood, paper, cardboard, fabrics, poljTiieric foils and adhesive tapes. Coatings may also be important on leather, fabrics, and other materials of decoration, safety, clothing, upholstery, housing, packaging and industry. Resins and/or high polymers are an inherent part of non-woven fabrics and papers of high wet and dry strength. Whatever the state of fabrication and use of the resin or high polymer the resinog- rapher must be prepared to show the texture and structure of its units in situ. REFERENCES 1. RocHow, T. G., AND Gilbert, R. L., "Resin- ography," Chap. 5, Vol. V, "Protective and Decorative Coatings," edited by J. J. Mat- tiello, Wiley, New York, New York (1946). 2. Clark, G. L., Editor-in-Chief, "The Encyclo- pedia of Chemistrj'," pp. 220-3, Reinhold Publishing Corporation, New York, (1957). 3. Chamot, E. M., and Mason, C. W., "Handbook of Chemical Microscopy," Vol. I, 3rd edition, Wiley, New York, (1958). 4. BoTTY, M. C, Felton, C. D., and Anderson, R. E., "Microscopy of Experimental Fibers," /. Textile Research Institute (1960). 5. Thomas, W. M., Thomas, A. ]M., and Deich- ERT, W. G., "Microscopical Study of Heter- ogeneous Polymerization," International Symposium on Macromolecules (Interna- tional Union of Pure and Applied Chemis- try), Wiesbaden, Germany, October, 1959. 6. Saylor, C. p., "The intangible isolation of con- taminants within purified substances," Gordon Research Conference on Separation and Purification, 1954, Science 119, 6 (Apr. 16, 1954). 7. RocHOW, T. G., RocHOW, E. G., "The Size of Silicone Molecules," Science, 111, 271-275 (1950). 8. Moore, L. D., Peck, V. G., "Study of Particles Present in Some High Pressure Polyethyl- enes," J. Polymer Sci., 36,141-153 (1959). 9. RocHow, T. G., Thomas, A. M., Botty, M. C, review on "Electron Microscopy'," Anal. Chem., 32, 99R (1960). T. G. RocHow 537 Stereoscopic microscopy (See also p. 532) BASIC DESIGN, OPERATION AND USE The stereomicroscope differs from the conventional compound laboratory micro- scope in several respects: (1) it is actually two compound micro- scopes aimed at a common object; (2) the image is erect and non-reversed; (3) the image is truly stereoscopic, or three-dimensional, as contrasted with the binocular body of a conventional microscope which presents identical images to both eyes, and hence is not stereoscopic; (4) the useful magnification limit is much lower, generally 100 to 150 X, as contrasted with 1000 to 1500 X in a normal compound microscope. From the foregoing, it is apparent that the stereomicroscope gives an image which PAIRED EYEPIECES PORRO ERECTING PRISM SYSTEMS SPECIMEN Fig. 1. The basic optical elements of a stereo- microscope in its simplest form. The Porro prisms serve not only to erect the image, but provide a simple means for changing the interocular dis- tance to suit the observer. is much more easily interpreted than those obtained with the laboratory microscope. The low magnification is a factor in this easy interpretation, since one sees a good deal of the actual object, and can easily gain the impression of simply being brought closer to a familiar object. True stereoscopy is also a potent factor in creating realism in the image. The observer sees the depths and heights standing out clearly and quite naturally, and the transi- tion from normal viewing to microscope viewing is easily made. The stereomicroscope has quite a large depth of focus, so that the sense of depth perception is heightened accordingly. The erect and non-reversed image also creates a natural and easily understood knage. It also permits one to easily manipu- late the specimen on the stage, without the attendant frustration involved in attempts to do the same thing on the laboratory microscope. This permits use of the stereo- microscope in many appUcations where very precise manipulations must be made, such as micro-dissection of biological material, or assembly of tiny electronic components in the industrial field. Structure of the Stereomicroscope Figure 1 shows the basic structure of a simple stereomicroscope. The left and the right microscopes aim toward a common object field. Each microscope comprises an objective, a prism-erecting system, and an eyepiece. Each prism system is mounted on a bearing concentric with the objective axis, around which it may be rotated to adjust it to the observer's interocular distance. Normally the angle between the two mi- croscope axes is in the neighborhood of 10°, which is about the normal convergence angle for viewing an object at 15 inches distance. 538 BASIC DESIGN, OPERATION AND USE The eyepieces are normally three-element construction, with external focal plane. They have about 25% larger fields than conven- tional microscope eyepieces, and are often termed "widefield" eyepieces. The objectives are each achromatic dou- blets. To change power, another objective pair of different focal length must be sub- stituted. This is normally done by mounting several objective pairs on a sliding or re- volving nosepiece. Instead of changing objectives, it is pos- sible also to change lenses within the body of the microscope to achieve a power change. One such system employs parallel light beams beyond the objectives, into which small Galilean telescopes are inserted in either normal or reversed direction. A 2X Galilean system, so used, multiplies the power by either 2 or 3^^, depending on its orientation. A second pair of Galilean tele- scopes may similarly give 3 X and V3 X . A more elaborate power changing system is to make the objectives continuously vari- able in power. Such systems are known as "zoom" lenses. With these, the field of view is never lost during power change, so that there are no blind spots in converting gradu- ally from low to high power. This is an im- portant point, particularly in teaching, where the student may more easily grasp the rela- tionship of the part to the whole in a bio- logical specimen. Figure 2 shows a side view of the optical system of a zooming stereomicroscope. The complete system, of course being actually two such systems, arranged to aim at a common object point, as in Figure 1. The zooming lens system comprises three pairs of doublet lenses, the upper two pairs being accurately cam-driven to hold the image in focus as the power is changed. Figure 3 shows a front view of the cam system used to drive the movable lenses. The power change knob has a gear on its mounting shaft which drives two identical gears on two identical cam shafts. Thus the /vEPicce PIECE FOCAL Fig. 2. The optical system of a continuously variable power stereomicroscope. The movable lenses are shown in the two limiting positions of their cam driven motion. INTEROCULAR ADJUSTMENT AXLE INCLINED PORRO MIRROR SYSTEM (4 MIRRORS) CAM DRIVEN UPPER LENS MOUNT CAM DRIVEN LOWER LENS MOUNT MOUNT FOR STATIONARY LENS Fig. 3. The mechanical structure of the micro- scope shown in Figure 2. A single knob drives both cams, thereby synchronizing the power change in both left and right sides of the microscope. cam-driven lenses move in accurately syn- chronized motions so that left and right 539 STEREOSCOPIC MICROSCOPY Fig. 4. The assembled zooming stereomicro- scope, with built-in fluorescent lamp illuminator. images are always matched for power, and maintain a constant focus as the image is zoomed. An interesting psychological phe- nomenon occurs when one zooms such a microscope, in that the image definitely appears to approach at high power and recede at low power, despite the fact that on the grounds of pure physics or geometrical optics, the image lies in the same plane at any power setting. A small insect placed on the stage and viewed at low power assumes a somewhat frightening aspect as one zooms to high power, making the insect appear to be a large creature approaching at a startling rate of speed. Stereomicroscopes are supplied with a wide variety of stands to hold them and focus them on a great variety of objects, large and small. Also various forms of illu- mination are made available, such as fluo- rescent lamps for illumination of transparent objects or spot-light forms of illumination for top or oblique lighting of paque speci- mens. Figure 4 shows a typical arrangement for a stereomicroscope intended mainly for transmitted light work. The substage illu- minator contains two small fluorescent lamps, with a frosted screen to even out the lighting on the specimen. Historical Background Early attempts to make stereoscopic microscopes concerned themselves more with the binocular properties than the stereo- scopic properties. Indeed many early de- signs, such as those of Cherubin d'Orleans (ca. 1660) were actually pseudoscopic sys- tems, i.e., the left image went to the right eye and vice versa. Wheatstone in 1852 wrote a treatise on binocular vision which stimulated increased interest in the stereo- microscope. Wenham, 1860, is credited with the construction of the first truly stereo- scopic microscope, but the forerunner of the modern systems was the twin objective sys- tem (Fig. 1) devised by H. S. Greenough in 1897. Improvements in the 1900's consisted of opening up the fields of view and making power change more convenient, culminating in the continuously variable power system introduced in 1959. J. R. Benford ENGINEERING MICROSCOPE. See LIGHT (OP- TICAL) MICROSCOPY, p. 439. "SOLID-IMAGE" MICROSCOPE For many purposes, such as tracing net- works of neural structure in the brain, a normal microscope has the serious limitation that only a thin layer of the specimen can be studied at one time, owing to the limited depth of focus of microscope objectives. When a thin section is examined, only those structures which happen to lie in the plane of the section can be observed. This plane may be confused by such objects as fibers coming away from or toward the plane of the section. A microscope which gave a large 540 "SOLID- IMAGE" MICROSCOPE Fig. 1. Gregory "solid image" microscope: left, microcsope in which slide is mounted on steel tuning fork (not shown); right, screen mounted on matched tuning fork upon which synchronous vibrating image is projected. The 2 processes thus provide scanning of specimen in depth giving a "solid image." depth of field, and whicli presented the image as a solid in a luminous block in which the structures could be observed in depth would have important advantages over existing optical microscopes. We have designed and constructed a primitive prototj^De of an instrument giving such a "solid image". It is possible to observe this from any position and to see the structures with appropriate parallax. The instrument involves two processes. (1) The focal plane of the objective is effec- tively made to "scan" up and down through the depth of the section. In the prototype instrument, this occurs at a rate of 50 double scans/sec. The slide is mounted on a steel fork tuned to 50 c./s. which is driven by a polarized solenoid energized b}^ a 50 c./s. sine wave. Thus the slide is carried up and down through the focal plane of the objective 50 times/sec. (2) The image is projected on to a screen mounted on a second tuned fork which vibrates at the same rate and in the same phase as the slide carrying the speci- men. Scanning serves to extract the information in depth from the specimen ; the second proc- ess reconstitutes it in depth, giving a "solid image" in the space swept by the screen. Its frequency of vibration is greater than the fusion frequency for the eye, so that little or no flicker is observed. Figure 1 is a photograph of the apparatus, in which the tuning fork does not appear. In improved designs an aperiodic drive is superior to the power-economizing tuned fork, for this will not drift in phase with temperature changes as does the tuned fork. The vibrating screen is replaced b}^ a helix driven by a synchronous motor, the image to be projected on a sector near the edge of the helix. A solid glass helix or "circular wedge" is being experimentally tried for changing the focal length of the objective to give scanning without physical movement of the objective or the specimen. With a sinu- soidal scan, as used now, it is essential (a) to modulate the intensity of the light by using a high-pressure mercury lamp off mains, and to phase this with velocit}^ ; and (b) to chop the hght at each half-scan, to cut out the image produced either by the upward or downward scan to prevent the possibility of mis-registration of the 2 sets of images given by a phase error or by asym- metrical wave form of the scan. The sub- stage condenser is extremely important; it must be of high N.A. (numerical aperture) and correctly adjusted. The instrument was exhibited for the first time in London on May 23, 1960. REFERENCE Naiure, 182, 1434 (1958). Richard L. Gregory 541 Television microscope A very simple principle is the basis of the image be visible from all parts of the audi- television microscope. In photomicrography torium. However, since the brightness of the a real microscopic image is formed on the image decreases as the square of the mag- photographic emulsion; here it is formed on nification, a corresponding increase in the the receptive surface of the tube of a tele- intensity of illumination of the object is vision camera. In other words, in compari- required to obtain sufficientl}' bright pro- son with the customary manner of using a jected images. Even in lecture halls of aver- television camera, the microscope takes the age size the limits of a permissible illumina- place of a photographic objective as imaging tion load for the object are already reached system. The advantages of such an arrange- with medium microscopic magnifications, ment are due to the characteristics of the Nevertheless, one often encounters the er- television installation and the principles roneous opinion that the determining limit upon which it is based. Thus the application for microprojection is given by the efficiency of electronic image transmission permits a of available light sources, spatial separation of the object and the mi- One often encounters the false conception croscope on one hand and the reproduction that it suffices, for protection of the object, of the image on the other hand. This at the to filter out the invisible light, especially the same time makes it possible to show the heat radiation. However, it is to be kept in microscopic image to any number of ob- mind, at least in microprojection of biological servers, also of such objects as are in rooms objects with transmitted light, that the which for various reasons may not be entered illumination load on the object is decisive, by the observers. This for example can be the A structurally determined absorption by the case when microscopic images of infectious object is the presupposition for production material are to be shown, or conversely, to of image contrast in the most commonly avoid the danger of observers carrying in- employed bright-field illumination with fection to tissue cultures. The television transmitted light. Independent of its wave- microscope can also be advantageously used length, the total radiation absorbed by the in the microscopic examination of radio- object is finally converted into heat. If now active objects or objects subjected to radio- one compares the size of the receptive surface activity. of a television camera tube having an effec- Application of the television microscope tive image diagonal of about 15 mm with in the morphological branches of natural the diameter of the customary microprojec- science and medicine have brought about a tion image of about 1-1.5 mm, and if one substantial advance. It is generally recog- further considers that the illumination in- nized that microprojection is a most valu- tensity required on the receptive surface for able and almost indispensible aid in these operation of the camera tube is very low, subjects. This is specially true today w^here then it is understandable that one arrives at the number of instructors stands in a pro- requirements for illumination of the object gressively less favorable ratio to the increas- which are not substantially greater than ing number of students. The size of the those for direct subjective observation, projection screen must be increased with the Herein lies the basis for the decided ad- size of the lecture hall and the average dis- vantages of the television microscope. As a tance of the observers from the screen in consequence of the slight illumination load order that details of the microprojection on the object and the possibility of a large 542 TELEVISION MICROSCOPE Fig. 1. Diagram of the path of rays in the Zeiss-Siemens television microscope. Left with trans- mitted, right with incident light. number of viewing screens, practically every- thing visible in the microscope can be shown simultaneously to an audience of any size. Naturally that holds not only for micro- scopic imaging in bright field with trans- mitted light, but also for the various miag- ing procedures with unfavorable light yield, as for example dark-field, phase contrast, or epimicroscopy. This optical system projects the micro- scopic image at suitable magnification on the receptive surface of the television camera tube and provides for a continuous mag- nification regulation in the ratio of 1:3.2 without change of optics as shown in Fig. 1. For incident illumination an auxiliary lens is inserted in the front opening of the lamp carrier in place of the quadruple filter holder. Unexceptionable maintenance of the Koeh- ler illumination principle is assured in each case. The imaging beam is deflected at a right angle from the projective to the tele- vision camera by a prism located in an adapter which is screwed into the threaded objective ring of the television camera. A light-excluding tube provides for lightproof connection between adapter and projective. Operation of the Ziess-Siemens microscope takes place in customary manner. A small remote control desk is located on the table for regulating the television installation. Cables pass from it as well as from the cam- era to the central power supply of the tele- vision installation. The control monitor with its 17-cm picture tube, set up on the table, is likewise connected with the central power supply. Additional viewing apparatus of optional screen size, number, and distance, and also a television projection receiver are connected by means of a coaxial cable op- tionally to the central power supply or to a control receiver. Up to three television cam- eras can be directly connected to this central power supply. They can be applied m a most diversified manner. Since the Siemens tele- vision installation is equipped with a so- called automatic gain control, its operation is restricted to switching on and off, and the occasional refocusing of the camera tube and regulating the image brightness of the re- ceivers according to the brightness of the room. This means that during actual use its operation is essentially restricted to manip- ulation of the microscope and consequently is simpler than that of a microprojection apparatus. Basic considerations and the experience gained thus far indicate unequiv- ocally the advantages of the television principle of microscopy. Adapted from an article by Helmut Haselmann in Zeiss Werkzeitschrift, No. 29, p. 42 (Sept., 1958). G. L, Clark 54a Ultramicroscopy DESIGN AND OPERATION. See OPTICAL THEORY OF THE LIGHT MICROSCOPE, p. 451. Ultrasonic absorption microscope The desire to explore the world of small Consideration of the propagation charac- scale structure (e.g., biological systems) has teristics of acoustical energy in tissue or sus- furnished at least a portion of the incentive pensions of biological material suggested the for the development of such devices as the development of an instrument which can be light microscope and the electron micro- expected to yield information on the struc- scope. In each case much new information ture of biological systems not obtainable on the microstructure of the system or mate- from either light or electron microscopy rial studied has resulted. studies (2, 3). This follows from the fact The light microscope (absorption type) that the interaction of the sound waves with exhibits certain structural features of cellu- the tissue structure is of a different nature lar and subcellular biological organization from that of light or electrons. Experimental by means of the contrasting pattern of light results (4, 5, 6) indicate that the protein and shade resulting from the absorption of constituents of tissue are largely responsible electromagnetic energy to different extents for the magnitude of the absorption of acous- by various parts of the structure under tic energy in the ultrasonic frequency range. examination. The use of various selective This work also shows that some different staining materials such as dyes, which dis- types of protein molecules, at equal concen- play different affinity for acidic and basic trations, absorb sound at different rates (7). parts of the tissue structure, permits other Therefore, a suitably designed acoustic in- aspects of structure to be detected and also strument could yield information on both raises the possibility of correlating mor- spatial distributions and identification of phology, not directly observable in unstained types of protein in tissue. Since this device material, with chemical properties. With the jg ^^ly in the early stages of development, advent of phase contrast microscopy it be- ^his article is concerned with outlining the came possible, without an increase in resolv- principles of operation, briefly describing iiig power, to Identify structural features not ^^^^^^^ ^^ measurements on filament models observable with the light microscope em- ,, i,. j-j-j.- j.u u , . , , . * . . , ^ at low resolution, and indicating the results ploying the absorption principle. » ,, ^. , , i ^- r ^^ • ui TTT-,r .1 • 1 r 1 . • of theoretical calculations of attainable With the arrival of electron microscopy, new details of cellular structure could be ' . . , . detected because of the higher available re- ^he results of an approximation analysis solving power and suitable impregnation ^^ ^^^ resolving power of this device are methods for producing contrast. These selec- g^^en below. The details are included m tive impregnation methods result in the reference (3) in which formulas are derived retention of salts of heavy metals at certain permitting a calculation of the maximum sites in the structure and thus produce con- resolving power attainable with present trast in electron transmission because the technology. It should be noted here that scattering power of the elements increase as considerable additional knowledge of tissue the atomic number becomes larger (1). structure can be expected from an examina- 544 ULTRASONIC ABSORPTIOX MICROSCOPE tion by an "ultrasonic microscope" of less resolving power than that of the light mi- croscope. This follows from the fact that many structural components of biological materials or systems, with different ultra- sonic absorption coefficients, may not be de- tectable at all by the light microscope. There is no a 'priori reason why materials with greatly different ultrasonic absorption co- efficients should have either detectably dif- ferent absorption coefficients or indices of refraction for light. The principle of operation of the ultra- sonic microscope is illustrated in Fig. 1. High frequency sound waves are generated in a "coupling" medium by a piezoelectric crystal vibrating in a thickness mode. The coupling liquid, which fills the irradiation chamber, serves to conduct the sound to and from the movable specimen which is inter- posed between the crystal and a small ther- moelectric probe. (An array of probes can be incorporated for more rapid acquisition of data). The piezoelectric crystal is excited electrically at a mechanical resonant fre- quency by voltage pulses with a rectangular temporal envelope (carrier frequency equal to the mechanical resonant frequency) of short duration. The small probe detects the acoustic energy level of the sound which passes through the portion of the specimen in its immediate neighborhood. The varia- tion in this transmitted energy level, as a function of the position of the specimen rela- tive to the probe constitutes an acoustic image of the ultrasonically detected struc- ture. Two mechanisms are involved in the de- tection of the ultrasound by the thermo- electric probe (8, 9, 10). First, an increase in the temperature of the wire results from the conversion of acoustic energy into heat by the viscous forces acting between the wire and the imbedding fluid medium. Sec- ond, acoustic energy is converted into heat by the absorption of sound in the body of the coupling medium which surrounds the SPECIMEN THERMOCOUPLE PROBE X-CUT QUARTZ' PLATE COUPLING LIQUID Fig. 1. Schematic representation of the ultra- sonic microscope showing the transducer plate which is excited to produce pulses of ultrasound in the coupling liquid, the specimen under exam- ination which is movable in the coupling liquid, and the thermocouple probe whose junction is placed immediately adjacent to the specimen. probe and specimen. The thermoelectric emf of the probe acts as the input to a DC chopper amphfier (noise level less than 0.01 microvolt), the output of the latter driving the vertical deflection plates of a cathode-ray oscilloscope (see Fig. 2). Thus, when the sound source is driven by a suitable radio frequency pulse, the cathode-ray beam is transiently deflected from its equilibrium position and the magnitude of this deflec- tion is a measure of the relative amount of acoustic energy detected by the probe. As the specimen is moved parallel to the radi- ating face of the crystal through the pulsed acoustic field, the changing deflection of the cathode-ray beam is observed and recorded. The data are then plotted and a contour "picture" of the disturbance to the sound field distribution, caused by the presence of the specimen, is obtained. The contour pic- ture constitutes an acoustic image of struc- ture in the specimen. The electronic instrmnentation used in the first niodel of the ultrasonic absorption mi- croscope is illustrated in the block diagram of Fig. 2. A commercial signal generator is used to obtain the radio frequency energy of predetermined frequency. The signal gen- erator is controlled by a keying unit which activates the generator to produce a pulsed output. The keying unit is, in turn, con- trolled by a digital tuning device which 545 ULTRASONIC ABSORPTION MICROSCOPE CRYSTAL SPECIMEN THERMOCOUPLE JUNCTION POWER lAMPLlFlER JCOUPLING 1 UNIT SIGNAL GENE RATOR TEMPERATURE SENSING ELEMENT HEAT EXCHANGE UNIT SOUND TANK KEYING! I UNIT I DIGI AL TIMER |dc. amplifier |oscilloscope| Fig. 2. Block diagram of the electronic instrumentation of the ultrasonic absorption microscope. permits accurate duplication of the duration of the acoustic pulse. A power amplifier stage, driven by the signal generator, is used to provide the power required to drive the piezoelectric transducer. A coupling unit is placed between the power emplifier and the transducer element in order to obtain elec- trical impedance matching. The electronic driver system must be stable in frequency and amplitude so that artifacts are not in- troduced into the "picture." An example of the type of data obtained at relatively low resolution using filament models as test specimens is shown in Fig. 3. Here, a nylon monofilament 0.003 inch in diameter is moved in a plane between the sound source and the probe (which has a maximum dimension of 0.0005 inch in the vicinity of the junction) starting from a position in the field where its presence does not appreciably influence the level of acous- tic energy detected by the probe. The direc- tion of movement is parallel to the crystal face and perpendicular to the filament axis. During this motion, the quartz plate is ex- cited to radiate pulses of 12 mc/sec ultra- sound with a duration of 0.1 sec. The ab- -20 +40 RELATIVE DISPLACEMENT (in units of 0.001") Fig. 3. The detection of the presence, in the coupling liquid, of a 0.003 in. nylon filament bj^ the ultrasonic microscope operating at a fre- quency of 12 mc/sec. 546 ULTRASONIC ABSORPTION MICROSCOPE scissa of the figure indicates the position of the filament specimen in the plane of move- ment parallel to the crystal face. The or- dinate indicates the relative .acoustic inten- sity detected by the probe, the scale being in units of deflection on the oscilloscope screen. The minimum in the curve corresponds to the position of closest approach of the nylon filament to the probe as the specimen tra- verses the acoustic field. The presence of the filament at the position of closest approach causes a reduction of 30% in the acoustic intensity below the undisturbed level. At half of this reduction, the curve is 0.007 inch wide and the width is equal to the fila- ment diameter (0.003 in.) at 0.8 of the total reduction. Not all of the observed reduction is produced, in this case, by absorption of sound within the nylon. Since there is a mis- match in acoustic impedance between the coupling liquid and the filament of at least 10%, some of the incident acoustic energy is scattered. In addition, some of the sound energy is converted to heat at the interface between the nylon filament and the coupling liquid as a result of the viscous forces acting there. Measurements similar to those shown in Fig. 2, but obtained with copper filaments of 0.001 inch diameter in the field, demon- strate that at an operating frequency of 12 mc/sec, model structures of this type with a diameter of 25 microns can be resolved. An approximation analysis (3) (based on formulas derived to describe the behaviour of a thermoelectric probe in a pulsed acoustic field) (8, 9, 10) of the operation of the ultra- sonic microscope indicates that a structure with a "radius" of 0.4 micron and having an acoustic intensity absorption coefficient per unit path length differing from the average value of the specimen by 5% should be detectable if the following acoustic and other parameters are employed: frequency — 1,000 mc/sec, ultrasonic intensity — 1,000 watts/ cm2, acoustic pulse duration — 0.1 micro- second, and thermocouple lead diameter-0.1 micron. A convenient pulse repetition rate for rapid assimilation of data would be 1000 pps. The "radius" of the structure is defined to be that distance from its center at which the deviation of the absorption coefficient from the average value of the surroundings is down to 0.7 of the maximum deviation. If the absorption coefficient of the structure differs from the average by a greater per- centage, then a smaller structure can be detected. Note added in proof: Since the preparation of this article, tech- niques have been developed for producing sound fields of appreciable amplitude in high absorption liquids to 500 mc-sec and for fabricating thermoelectric detectors having maximum dimensions in the vicinity of the junction of 5 microns (11). REFERENCES 1. See appropriate sections of this encylopedia for comprehensive descriptions of light, phase-contrast and electron microscopy. 2. Dunn, F., and Fry, W. J., J. Acoust. Soc. Am., 31, 632-633 (1959). 3. Fry, W. J., and Dunn, F., "Physical Tech- niques in Biological Research," ed. W. L. Nastuck, Academic Press, New York (to be published, 1961). 4. Carstensen, E. L., and Schwan, H. P., "Ultrasound in Biology and Medicine," ed. E. Kelly, Am. Inst. Biol. Sci., Washing- ton, D. C. (1957). 5. Schwan, H. P., and Carstensen, E. L., WADC Tech. Rept. 56-389, Wright Air De- velopment Center (1956). 6. Carstensen, E. L., and Schwan, H. P., J. Acoust. Soc. Am., 31, 185-189 (1959). 7. Carstensen, E. L., and Schwan, H. P., J. Acoust. Soc. Am., 31, 305-311 (1959). 8. Fry, W. J., and Fry, R. B., /. Acoust. Soc. Am., 26, 294-310 (1954). 9. Fry, W. J., and Fry, R. B., /. Acoust. Soc. Am., 26, 311-317 (1954). 10. Dunn, F., and Fry, W. J., I.R.E. Trans, on Ultrasonic Engineering, PGUE-5, 59-65 (1957). 11. Dunn, F., /. Acoust. Soc. Am., 32 (1960). Floyd Dunn and Wm. J. Fry 547 Ultraviolet microscopy BASIC PRINCIPLES AND DESIGN The wavelength region of the visible spec- trum is 400 to 700 m/i. The adjacent shorter wavelength region (100 to 400 m^) is the (invisible) ultraviolet region. Below 100 niM, air absorption restricts the transmission of ultraviolet, and this region is called the "vacuum ultraviolet" region. Inasmuch as resolving power is pro- portional to wavelength, it is possible to resolve finer structure in ultraviolet than in visible radiation. Also, since many biological materials have absorption bands in the ultra- violet, their microstructure may, by the use of the ultraviolet microscope, be rendered visible without resort to staining. The ultraviolet microscope differs from the conventional light microscope in that it includes means for converting the invisible ultraviolet unage into a visible image. Furthermore, since optical glasses do not transmit ultraviolet, other materials, notably fluorite and fused quartz, are used for lens elements, and mirror systems often are used in place of refracting optics. These same differences also apply to the illuminating system, and of course, the source itself is different, since it must emit in the ultra- violet rather than the visible, and cannot be enclosed in a glass envelope, since glass ab- sorbs the ultraviolet. Conversion of the ultraviolet to visible energy is accomplished by (1) photography, (2) fluorescence, (3) television type receptor tubes, (4) flying spot television techniques, or (5) photoemissive image-converter tubes. Early History. Koehler (1) developed the first optical systems for ultraviolet microscopy in about 1904, his work being aimed toward the goal of increased resolving power due to the use of shorter wavelengths. Microscopes made in accordance with Koehler's design were made by the Zeiss firm The objectives were monochromats, i.e., corrected for only one wavelength in the ultraviolet. Photography was used to render the image visible. Work with this instrument was difficult, and apparently very little real application was made of it. Spectral Absorption. In the 1930's, in- terest in ultraviolet was revived by the work of T. Caspersson (2, 3) in Sweden, and J. Loofbourrow (4, 5) in the United States, in studying the spectral absorption properties of various biochemicals. With ultraviolet, these men found that it was possible to photograph live unstained tissue sections and tissue cultures, and to differentiate vari- ous normal and abnormal cells by their absorption characteristics. Objective Designs. Studies of this nature led to a re-activation in the field of ultra- violet microscope design improvement. L. V. Foster (6), described in 1946, an achromatic objective composed of fused ciuartz and fluorite for use in a sufficiently broad range in the ultraviolet, so that a fluorescent screen could be used for focusing. Interest in spec- tral absorption studies, however, led to the demand for still greater wavelength range, and mirror type objectives were designed by Brumberg (7), Burch (8), and Grey (9) to fulfill this requirement. The most widely used of these are the Grey designs, which employ a combination of reflecting and refracting optics to achieve good correction over a large wavelength range with very little central obscuration by the mirror system. The refracting elements are of fused quartz and fluorite Fig. 1, shows the construction of the 0.72 N. A. Grey design as made by Bausch & Lomb. This objective is corrected for the complete spectral range from 254 mju to 700 m/x, hence is useful in applications where focusing is accomphshed in the visible for photography in the ultraviolet. 548 BASIC PRINCIPLES AND DESIGN Sources of Ultravaiolet Energy. The noisy and odorous cadmium spark, originally used as the source for ultraviolet, has been replaced by the quartz- jacketed mercury arc. Combination glass and liquid chemical filters have been replaced by the grating monochromator, which provides simple and convenient means for changing and select- ing desired ultraviolet wavelength regions. Viewing Systems. Perhaps the most diffi- cult technical problem in the design of the ultraviolet microscope has been that of achieving a receptor to permit seeing the image. All of the early work used photog- raphy with trial and error search for best focus. Later the fluorescent screen was em- ployed, but the image was generally too grainy and dim for much useful work. The advent of high-speed processing of photographic film led to the possibility of only a very short time delay between ex- posure of the film and seeing the image. The "color translating microscope" developed by the Polaroid Corporation utilized this technique. Three adjacent frames of ultra- violet sensitive film were exposed rapidly in sequence to three pre-selected wave- lengths in the ultraviolet. The film was rapidly developed, fixed, and projected onto a viewing screen. Each frame was projected through a primary color filter, and the images, optically superimposed on the screen, formed a composite three-color image. This instrument was, of course, ex- tremely expensive, and its use, accordingly, severely restricted. Television image scanning techniques have also provided means for seeing ultra- violet images. Two basically different ap- proaches have been used in applying tele- vision apparatus and methods. In one, called the "flying spot microscope" (q.v.) an ultra- violet-emitting cathode ray tube face is located in what is normally the image plane of a photomicrographic set-up. The micro- scope, acting in reverse, forms a greatly reduced image of the cathode ray tube face Fig. 1. The 53X, 0.72 N.A. Catadioptric objec- tive for ultraviolet, designed by D. S. Grey and manufactured by Bausch & Lomb. Refracting ele- ments are of fluorite and fused quartz. on the specimen. An ultraviolet sensitive photomultiplier tube picks up the signal transmitted by the specimen, as it is scanned, and the energy signal from the photo- multiplier is converted to a visual image on a conventional cathode ray tube, which is tuned to the same sweep frequency as the illuminating cathode ray tube. The particu- lar virtue of this "flying spot" system is the low dosage of ultraviolet concentrated on the specimen, since prolonged exposure to ultraviolet is lethal to most living specimens. Its disadvantage is the technical difficulty of attaining a satisfactory means of varying the wavelength of the ultraviolet emitted from the cathode ray tube. Another television type solution to the problem is to use a conventional ultraviolet source and monochromator to illuminate the microscope, and to pick up the image on an ultraviolet sensitive image-orthicon, and by means of a closed circuit television send- ing and receiving set-up, view the image on a conventional television screen. Both of these television systems share the basic problems of the color translating microscope, that is, they are extremely costly and elaborate to build, and require con- siderable training and skill to keep in good operating condition. Because of the com- plexity of these systems, their use has been 549 ULTRAVIOLET MICKOSCOPY I Fig. 2. The RCA "Ultrascope" image-converter tube used as an ultraviolet viewer and pre-focusing device for photography in the Bausch & Lomb Ultraviolet Photomicroscope. limited to a very small number of research organizations. In 1959, RCA introduced an ultraviolet image converter tube, which greatly simpli- fied the technique and equipment needed to see an ultraviolet image. The tube called an "Ultrascope", employs a photoemissive cathode, a single-stage electron imaging system, and a fluorescent screen to convert the ultraviolet image into a light image. The double conversion is thus ultraviolet to electrons to visible. Figure 2 shows the manner in which this converter tube is mounted on an ultraviolet microscope. A small removable mirror de- flects the ultraviolet image to the Ultrascope tube, or when withdrawn permits photog- raphy of the image, as seen and focused by the Ultrascope tube. This equipment repre- sents a tremendous simplification over previous ultraviolet viewing systems, and gives promise of creating new and expanding interest in this field of research. REFERENCES 1. Zeit. f. Wiss. Mickros., 21, 129, 273 (1904). 2. Skand. Arch. Physiol., 73 (1936). 3. Nature, 143, 602 (1939). 4. Bull. Basic Sci. Res., 5, 46 (1933). 5. J. Opt. Sac. Am., 29, 53.5 (1939). 6. J. Opt. Soc. Am., 38, 689 (1948). 7. Bull. Acad. Sci., USSR, 6, 32 (1942). 8. Proc. Phys. Soc. {London), 59, 41 (1947). 9. /. Opt. Soc. Am., 39, 719, 723 (1949). J. R. Benford COLOR TRANSLATING TV ULTRAVIOLET MICROSCOPE Aim of the New Design. The electronic color translating TV ultraviolet microscope, designed by scientists of Neutronics Re- search Company, is the most recent means 550 COLOR TRANSLATING TV ULTRAVIOLET MICROSCOPE for studying the topography, chemical simi- larities and dissimilarities, and the absorp- tion spectrum of cellular components, syn- thetic and natural fibers, crystals, and amorphous substances. Its usefulness enters fields of investigation in biology, medicine, physics, and metallurgy as well as organic, inorganic, and physical chemistry. Operation. The Model ME- 10 la TV ul- traviolet microscope essentially consists of a triple monochromator system, which se- quentially illuminates a specimen at three easily preselected ultraviolet wavelengths down to 2400A. The microscope proper pro- vides coarse and fine adjustments for the condenser and objective, which consist of 53 X apochromatized catadioptric lenses furnished with the microscope. The objec- tive fine focus is in steps of one micron. The specimen sUde rests on a microscope stage, which allows circular and x-y move- ments. The three sequential UV beams are then picked up by the projection eye- piece (3.5 X and 10 X) which projects the absorption images of the specimen on to the three UV-Vidicon tubes of the color television cameras. The JTV wavelength separation is achieved by means of a rotating mirror system. Each vidicon is associated with a primary color in a closed loop color television system. The presence of a specific color in the image on the 15" television monitor screen indicates transmission of the corresponding ultraviolet wavelength. A camera is provided to obtain color photo- graphs of the television monitor screen display. Magnifications from specimen to screen of about 4,000 X to 25,000 X are possible. Resolution, under favorable condi- tions, is of the order of 0.3 micron. Varia- tions in hue correspond to differences in transmission spectra and thus indicate differ- ences in chemical composition. Variations in optical density (with no color difference) indicate various amounts of material of the same absorption spectrum in the light path. The line analyzing microspectrophotom- eter simultaneously displays the point-to- point absorption variation for any selected ultraviolet wavelength by extracting any one of the horizontal scan hues in the color television monitor for display on the line analyzer 5" screen. A dry processing camera is provided to obtain permanent records of the transmission curves. Since specimen illumination for any wave- length is not continuous but occurs in the form of about 14 millisecond long pulses, the problem of irradiation damage to cells is minimized. The use of three TV cameras overcomes the serious problem of color carry-over. Applications. The microscope is intended for quantitative and qualitative investiga- tions from 2400A to 6000A. To a major ex- tent, investigations to date have been in the medical field in the examination of cells (living, stained, or unstained) for determin- ing chemical similarities and dissimilarities of objects within the specimen as well as the structure and topography of the cell. Microchemical analysis can be performed by this technique. Unstained and living specimens, which are entirely colorless when viewed in visible light, give results comparable to a selectively stained sample. For instance, at 2800A, absorption coincides closely with areas rich in proteins containing amino acids, which would take an acid stain such as eosin; on the other hand, areas which absorb heavily at 2600A often cor- respond to those that take basic stains indicating the presence of nucleic acid. Focusing and area selection are performed as in ordinary microscopy except that eye strain is reduced to a minimum since the operator need only watch the color TV screen during these operations. G. L. Clark 551 ULTKAMOLET AIlCHOSCOrY IMAGE FORMATION BY A FRESNEL ZONE PLATE* This article describes experiments in the use of a new type of Fresnel zone plate for image formation with visible light and ultra- violet radiation. Tests with visible light and ultra\'iolet down to 2537 A are described below followed by a description of our plans to extend these tests down to about 100 A. There is good reason to believe that the same zone plate will operate over a large range of wavelengths from the soft x-ray region through the infrared because its transparent zones are completely open. Therefore, it is opportune to speculate on the possibility of using the zone plate to focus even other radiations and particles. First, however, let us consider the motives that prompted us to construct this zone plate for soft x-ray and extreme ultraviolet radiation. Early research in the means of focusing light was motivated in part by interest in the construction of microscopes and tele- scopes. These instruments were, after all, the earliest "space probes" into both the intermolecular and extragalactic worlds. Later, the properties of the electromagnetic spectrum outside of the visible region prompted natural extensions of image-form- ing devices into the near infrared and the ultraviolet. % ' 80 CRYSTALS <2 MM. THICK) METALS (0 1 ,. THICK) LITHIUM FLUORIDE CRYSTALLINE QUARTZ [, FUSED QUARTZ - 60 - / CALCIUM / FLUORID / V /. CORUNDUM - 40 - ALUMINUM \ 1 1 ! j / f STANDARD BAND PASS FILTER - 20 '■■ INDIUM , BISMUTH ♦,-.. i\ \ j A , / NiSO, COMPLEX Tranmissivity (I/Io) of Various Transparent Substances Fig. 1. The transmissivity of several materials in the ultraviolet. The crystalline materials that might serve for lenses cease to show any apprecia- ble transmission below 1000 A. The thin metal films could serve as filters. * See also "X-Ray Telescope for 1-100 A Re- gion," in "Encyclopedia of Spectroscopy," p. 778. Today similar motives exist for the con- struction of microscopes and telescopes to work in the soft x-ray and extreme ultra- violet (hereafter referred to as euv) region. While no definite bounds exist for the euv region we shall limit our study to wave- lengths from 10 A to about 1000 A. Al- though the developments in x-ray micros- copy and microradiography have already been responsible for two international congresses (1, 2), the new interest in tele- scopes that can operate in the soft x-ray and euv regions stems chiefly from discoveries in astrophysics. Rocket experiments have already demon- strated that solar, stellar and interstellar sources of euv exist (3), but the earth's atmosphere prevents most of the radiations shorter than 3000 A from reaching its surface. The advent of rocket and satellite-borne experiments justifies serious consideration of telescopes and spectrographs sensitive to wavelengths between 1 A and 3000 A. We distinguish between instruments, such as lenses and mirrors, whose main function is to form optical images and those such as prisms and gratings, whose purpose is to produce dispersion. Our interest for the present is primarily in image formation. The classical means of focusing visible light for image formation involve reflection (mirrors) and refraction (lenses). Proper choice of materials allows the use of reflectors and refractors in the ultraviolet from 3000 A to about 1000 A. Between 1000 A and 10 A refraction is out of the question because most materials are opaque to radiation in this region. Special types of glass (4) have been developed which can transmit about 70 per cent at 2200 A through a thickness of 1 mm. But this is exceptional since the transmission of most t3T3es of glass drops essentially to zero below 3000 A. Although transmission of wavelengths below 10 A does occur, the index of refraction is so close to unity that refraction is impractical. The transmission of several materials is shown in Fig. 1 (5). Except for thin fihns 552 IMAGE FORMATION BY A FRESNEL ZONE PLATE 60 r I r 1 I 1 II T -~r r 1 1 r T- T ALUMINUM ^ T 1 ••'' 7^""' V ■■■■■ — / \ - \ - / / V / \ ALUMINUM OVERCOATEO WITH / / MAGNESIUM FLUORIDE ^^ 1 1 BERVLLIUM^ \ ^ """--J \ / I / / "~~-— -V-.-.. - — / / / \ / / / ^ PIATlfdIIU * _ '"'^ * /^ \ , - <:2m-/ r SILVER--^ ■, \ / - \ \ INTEBFEBENCE 1 — -7^^^ / COATING ^' < ^. i^:ll 1 L 1 1 1 I 1 L 1 1 1 1 1 L ' -■"'y 400 800 1200 1600 2000 2400 2800 3200 3600 \ A, Reflectivities op Evaporated Metal Film Fig. 2. Normal incidence reflectivities of various materials in the ultraviolet. Aluminum overcoated ^•ith magnesium fluoride gives useful reflectivity even below 1000 A. Below 400 A there is a dearth of experimental information on reflectivity. of aluminum, tin, indium, and bismuth, as shown, (22) the materials of which lenses might be made cannot transmit below 1000 A. Below this wavelength, then, lenses are apparently not worth considering. How about mirrors? Anastigmatic image formation with a single mirror requires good reflectivity at normal or near-normal inci- dence. Recently Hass (6) and his co-workers have developed coatings of evaporated metal films which, when deposited on glass or other materials, produce high reflectivities down to fairly short wavelengths. Fig. 2 (5) summarizes the characteristics of some of the best coatings. The data show trends rather than authoritative values. For these the reader is referred to the modern litera- ture.^- -^ For wavelengths shorter than 585 A the Fresnel zone plate may find applica- tion as a recurring device. Nevertheless, one must not exclude the possibility of using very large mirrors with very low reflec- tivities. For angles of incidence close to 90° it becomes convenient to speak of the "grazing incidence angle," that is, the complement of the usual angle of incidence, which is the angle subtended by the incident ray and the normal to the reflecting surface. For x-rays, the phenomenon of total reflection at graz- ing incidence is well known (7, 8). The reflectivity is 100 per cent because the index of refraction for most materials is less than unity. Point-to-point grazing-incidence sys- tems for forming images have been built by Kirkpatrick and his students (9, 10, 11), but reflection from a single curved mirror at grazing incidence is highly astigmatic and can be corrected only by the difficult tech- nique of crossed mirrors. Systems of this type have not been developed for wave- lengths as long as 100 A but they may have to be considered. If w^e restrict ourselves to the less formid- able process of image formation by normal incidence reflection at a single surface, we can probably conclude that a region exists, somewhere between 10 A and 1000 A, where image formation by reflection or refraction is either difficult or impossible. Zone Plate for Focusing Extreme Ultra- violet and Soft X-Radiation Diffraction offers still another way of bending light to produce focusing. Several such methods have been suggested (12, 13) but even the simplest, the Fresnel zone plate (14, 15) has not received serious use in any region of the spectrimi, let alone that between 10 and 1000 A. Several writers have noted that Fresnel zone plates might be used for image formation in this difficult region, but no successful attempts to build a zone plate for x-rays or euv have come to this author's attention. This paper describes the construction and 553 ULTRAVIOLET MICKOSCOPY use of a Fresnel zone plate whose opaque and their differential absorption is therefore bands are made of thin gold and whose open greater in this region (1(5). On the other bands are completely transparent to radia- hand, the wavelengths from 10 to 100 A, tion between 10 A and 1000 A, because they which are very long x-rays, are considered are empty. very short compared with the wavelengths A Fresnel zone plate consists of alternately of ultraviolet radiation, and are of potential opaque and transparent bands l)ounded by interest in astrophysics. Also, the lines circles of radii around 150 A and 300 A emitted by the sun's /- , „ o /v^ corona would be of interest as sources for r„ = ri\/n; n = 1, 2, 3, • • • (i) . . . telescopic miage formation (19). Its principal focal length, /, obeys the simple Now consider the focal length, /. In relation microscopes we want / to be small, but in a j^ ^ ^ 2 (2) telescope a long focal length is an advantage ; in fact, a focal length of 100 cm would be where X is the wavelength, and n is the practical and one of 1000 cm, though some- radius of the central circle. what large, is not completely beyond possi- In considering the focusing of x-rays for bility if we could build the right kind of microscopy we see that the product /\ can platform. The reader must bear in mind that be disturbingly small. For example, in an the ultimate platform for telescopes operat- x-ray microscope one might consider / = 1 i^g in the soft x-ray and extreme ultraviolet cm and X = 1 A. This requires n^ = f\ = regions must be a satellite orbiting above 10-8 em2 or n = lO"'' cm or one micron. By the earth's atmosphere. Eq. i w^e find that the width of the nth zone is Xq choose figures of the right order of sn = (V^m - V^)n, (3) "magnitude, consider / = 100 cm and X = 100 A. Using n = 25 we get s„ = ri/2\/n = which, for large n is approximated by iq-z cm or ten microns. This value is about ^^ ten times the resolving power of the finest- *" ^ 2\/n ' grained photographic films and makes the construction of a zone plate for soft x-rays If, for example, w = 25 and ri = 10"^ cm, seem feasible. then Sn = 10-^/2\/25 cm = 10"^ cm or 0.1 However, a zone plate made by exposing micron. Since the best photographic emul- a photographic film would be useless in the sions (e.g., Eastman 649 spectroscopic region between 10 and 100 A because the plates) have a resolution of about a micron, base on which the emulsion rests, and even a zone plate made to these specifications the emulsion layer itself, are practically looks impossible indeed. The product /X opaque to these wavelengths, needs to be considerably larger before zone Recently we have succeeded in producing plates look feasible. Fortunately, the cur- a zone plate whose opaque elements are thin rent interest in the use of x-rays and euv in concentric bands of gold made self-support- microscopy and astrophysics can lead to ing by the use of thin radial struts (see Fig. larger values of both f and X. 3). The streaks in the photograph are im- Consider X first. Wavelengths much longer perfections in the form of very thin filament- than 1 A are important for different reasons, like pieces of gold. The outer diameter of In x-ray microscopy, for example, there has the zone plate is 0.2596 ± 0.0002 cm. The been a trend towards wavelengths approach- central circle has a diameter of 0.0426 ± ing 100 A because microscopic objects of 0.0002 cm. The bands are bounded by circles biological interest have low atomic numbers whose radii obey the relation r„ = 554 IMAGE FORMATION BY A FRESNEL ZONE PLATE 0.02lo\/7i, 71 = 1, 2 • • • 38. For a zone plate with a large number of rings the more pre- cise formula for r„ would be needed, r„ = VfnX Vl + nX/if. (5) The narrowest band was designed to meas- ure 0.0017 cm in width. Actually it varied between 0.0010 and 0.0020 cm. The zone plate was made for us by the Buckbee Mears Company of St. Paul, IMinnesota. As far as we know this is the first zone plate with trans- parent regions that are completely open (except for the supporting radial struts). The images formed are very sharp, compar- ing favorably with those made by lenses of similar focal length and aperture (see Fig. 9). We may suggest one explanation for the good quality obtained with these zone plates as compared with that of the photo- graphically based zone plates. The phase condition that requires an increment of exactly one wavelength between successive transparent bands must be difficult to meet with a film base many microns thick that is not held to optical flatness. This condition is probably less difficult to meet with the new zone plate that has no film base at all. Since the transparent regions of the zone plate are completeh" clear they should transmit radiation of all wavelengths. The gold bands are practically opaque to all radiations between 10 and 1000 A; there- fore, the zone plate should be able to focus soft x-rays and euv in this region. Of course, it can also operate in the visible and infrared regions. It should also work with particles having wave-like characteristics of the proper wavelength, but our present interest is in the soft x-ray and euv region. Our first zone plate, with which all the tests reported in this article were made, was designed to have a focal length of about -400 cm at 100 A. At 4000 A its focal length is 10 cm; it thus lent itself very conveniently to tests with visible light. All the tests de- scribed below with the exception of those at 2537 A were made with visible light. We Fig. 3. An enlarged photograph of the self -sup- ported gold zone plate used in these experiments. The diameter of the outer circle is 0.2596 ± 0.0002 cm. The central circle has a diameter of 0.0426 ± 0.0002 cm. The thickness of the gold is estimated as 10 microns. The white bands representing the transparent regions are completely open and hence will transmit electromagnetic radiation of all wavelengths. plan to continue tests at 100 A and 1000 A with newly acquired sources. Resolution Theoretically, the smallest angular separa- tion, 0min , between two monochromatic point sources at infinity that can just be resolved when imaged by a zone plate, can be shown (14) to obey the Rayleigh criterion for a lens: sin 0nnn = 1-22 (X/D) . (6) Here X is the wavelength and D the diameter of the outermost circle of the zone plate. To test this relation we used fine mesh screens as objects, illuminating them by transmission as shown in Fig. 4. In Eq. 6 we could vary X, the wavelength being used, but not D, as all our zone plates had the same diameter, approximately 0.26 cm. A zone plate is a highly chromatic device, having a focal length that is inversely proportional to the wavelength. We used filters to monochromatize the radiation or 555 ULTRAVIOLET MICKOSCOPY LIGHT SOURCE '.:■'. 's?^ ,' \. , — '7'? Fig. 4. The layout of the apparatus on the optical bench. The light source consisted of a tungsten filament lamp and a condensing lens system. For the experiments at 2537 A the source was a 4-watt General Electric germicidal lamp, No. GT4/1. The filters are described in the text. The zone plate was mounted at the open end of a bellows extension. The camera was an Exakta VX Ila. Eastman Kodak Microfile Contrast Copy film was used throughout, processed in D 19. Object and image distances are labeled p and q, respectively. • • • • e/f.9.9.9.f.f.9>.9 • • • ,,,,. ::::»■•• %*♦►*»«•».»..,•-■•• •! s fll ■ ^m m mf***»***»f-*»* * ••••' ^ ^^ ^ {• • • f»***«^**««<*««««t ft***! M| gft «| . ^ — -*»««v»«»*w«o«««*-^ — ^ ^9 flp mqf ■■ — _. -.^- ™ • • •■•«**#«*##^»»»»*»«,»» • •' ^ -w ■iw • ••!• E • •••■::;;:":::;:::::♦•••» ;»••• • ••• Fig. 5. Pictures of a mesh with 4 lines per mm taken with the zone plate at three different wave- lengths. Working out from the center we see pic- tures taken at 6700 A, 4358 A, and 2537 A, in sizes proportional to the original image sizes obtained with a fixed object-to-zone plate distance of 47 cm. The shorter wavelength results not only in a longer focal length and hence a larger image size but also in improved resolution. See also Fig. 6. relied on the strength of a spectral line, as in the case of the 2537 A source, to produce approximately monochromatic radiation. The object distance, p, and the image dis- tance, q, (Fig. 4) were chosen to exhibit the improvement in resolution that accompanies the use of shorter wavelengths. We operated in a region close to the limit of resolution. The parameters were purposely chosen to exhibit the improvement in resolution from red light to ultraviolet. For example, in Fig. 5, three pictures of the same object mesh made with the zone plate are shown superimposed on one another to exhibit the ratio of the sizes in which the original pic- tures appeared when taken with red (6700 A), blue (4358 A) and ultraviolet (2537 A) light. The increase in image size results from the fact that the focal length of the zone plate increases as the wavelength decreases, while the object distance, p, remains fixed at 47 cm. This increase in image size alone, can bring about an improvement in signal- to-background ratio, but notice that an improvement in resolution has also taken place. The ultraviolet picture is not only the largest; it is the best resolved because it was made at the shortest wavelength. To exhibit this improvement in resolution apart from the change in magnification, the pictures of Fig. 5 were enlarged photo- graphically so that the distance between squares remained constant. The improved resolution becomes clearer. In Fig. 6 (left), p equaled 47 cm and q was adjusted to 7.2 cm, a value that produced the best focus experimentally for a wavelength of 6700 A. A gelatin filter was used with a pass band of about 400 A. The object mesh had 4 lines per mm. The angle subtended at the zone plate by two successive centers of open mesh 556 IMAGE FORMATION BY A FRESNEL ZONE PLATE was 0.00053 radian, which is 1.68 ^„,in for that wavelength. The center picture of Fig. 6 was made at p = 47 cm, q = 12.5 cm, at a wavelength of 4358 A. A Baird interference filter with a total pass band of about 100 A was used. Here the angular separation between adja- cent centers is 2.59 ^min . An improvement in resolution is apparent. Fig. 6 (right) shows an image of the same mesh taken at 2537 A. The value of p re- mained fixed at 47 cm, q was 27 cm and the angular separation between object centers was now 4.44 ^min • The source was a General Electric germicidal lamp number GT4/1. The manufacturer states that this lamp yields 60 % of its energy at 2537 A. A filter was used to remove the visible radiation, but the monochromatic effect at 2537 A is due only to the relative intensity of this line. This demonstrates a point about zone plate focusing that may be useful in ultra- violet and x-ray astronomy: a relativel}^ in- tense and isolated spectral line may preclude the use of filters. These pictures lead us to believe that the resolution will continue to improve in a predictible way as we go down first to 1000 A and later to 100 with our zone plate. Comparison of Zone Plate with Pinhole and Lens Next we compared the image-forming qualities of our zone plate with those of a pinhole. The pinhole size was calculated to give the optimum resolution for the chosen values of X, p, and q. The optimum diameter for such a pinhole is given by the expression (17) # t • ^ « # ' • «**-••. d = 2 0.9pg (7) This may also be written d = 2\/o.9/X which, interestingly enough, is very close to the diameter of the innermost circle of a zone plate that gives good focus at the same distances and for the same wavelength. In Fig. 6. The pictures of Fig. 5 are here enlarged to produce equal image sizes. From left to right, they were made at 6700 A, 4358 A, and 2537 A. The angles subtended at the zone plate by two adja- cent midpoints of the open area of the object mesh were 1.68 (9min , 2.59 ^min , and 4.44 ^min , where ^minis the theoretical minimum angle of resolution computed at the respective wavelengths. The im- provement in resolution with shorter wavelength is apparent. other words, the pinhole behaves like a zone plate with a single circular opening. Fig. 7 is a picture of a mesh with 4 lines per mm framed by the lines of a coarser screen with 0.7 lines per mm, made with the pinhole at a wavelength of 4358 A. With the same screens as an object and all other factors the same, a picture was taken with the zone plate instead of the pinhole, with the results shown in Fig. 8. The angular separation, at p = 26.7 cm, between adjacent centers of the smaller screen pattern was 5.3 X 10~* radian or 4.5 ^min , as computed for the zone plate. Theoretically, the im- provement in resolution of the zone plate over that of the pinhole is a factor \/n or 6.23 for our zone plate of 38 rings. The actual improvement as demonstrated in Figs. 7 and 8 is quite apparent. The zone plate has another advantage over a pinhole in that the light flux reaching the image is greater with the zone plate, so that much less exposure time is required if all other factors have been left unchanged. In our notation, n stands for the number of circles in the zone plate pattern. Hence n/2 is the number of open zones or bands. The expected improvement in exposure should 557 ULTRAVIOLET MICROSCOPY Fig. 7. An object consisting of a coarse grid, one rectangle of which measures 1.43 mm by 1.72 mm, within which there is a fine mesh with 4 lines per mm, photographed through a pinhole whose aperture was chosen to produce the optimum reso- lution for the wavelength and distances involved. Only the coarse grid is resolved. Compare this with Fig. 8. be w/2 since all the zones have equal areas. For our zone plate, n/2 = 19. The actual exposure time used for the pinhole was 40 times greater than that required for the zone plate. Since the densities of the plates were compared only by eye the experimental values do not seem out of line. Next we compared the resolution of our zone plate with that of a lens of similar focal length, of aperture equivalent to that of the zone plate. For both the lens and the zone plate / = 10.2 cm and D = 0.26 cm. Fig. 9 (right) shows the image of a screen with 4 lines per mm photographed with the lens, and Fig. 9 (left) shows the same screen photographed with the zone plate. Adja- cent centers of the object screen subtended an angle equal to 2.6 d^.i^ ■ Other constants were p = 47 cm, g = 13 cm, and X = 4358 A. The lens had a simple plano-convex form. If a highly corrected camera lens had been chosen the comparison might have favored the lens, but since it was stopped down to //1 8 and the field covered was not very large, it was not subjected to undue demands ••#« fii29c^io) showing sigma phase in austenite matrix after heating at 900°C. for 500 hours. Fig. 7 indicates segregation of molybdenum between the two phases. ' Or, •-, . . ..k * \ • X • J^ ^ S- A. »- « .♦ Fig. 8. Microradiograph X160 Fig. 9. Photomicrograph X160 Alloy steel (15%Cr, 18%Ni, 3%Mn, 2}i%Mo, 7%Co, 2}i%lSih). Contains carbides and intermetellic compounds. (Segregation of alloj^ elements was further elucidated by use of other radiations). Note: Effect of specimen thickness. 590 MEDICO-BIOLOGIC RESEARCH with the same radiation at different angles of incidence. Examples are given in the literature (5). If microradiography by pro- jection is employed diffraction can produce images which correspond to a reduced inten- sity of transmitted radiation and simulta- neous images from intensity diffracted in forward directions with Bragg angles less than 45° (10). This effect is similar to the phenomenon of diffraction from coarse grains in metal specimens giving mottling effects on macro radiographs (11). REFERENCES 1. Sharpe, R. S., "X-Ray Microscopy and Mi- croradiography," pp. 591-602, Academic Press, New York, 1957. (Edited by V. E. Cosslett, A. Engstrom and H. H. Pattee, Jr.) 2. Johnson, W. and Andrews, K. W., Iron and Steel, 31, 437-444, 1958. 3. Betteridge, W. and Sharpe, R. S., /. Iron and Steel Inst., 158, 185-191 (1948). 4. This Encj'clopedia Section, p. 561, 574. 5. Andrews, K. W. and Johnson, W., pp. 581- 589 of Reference (1). 6. MoRLEY, J. I., Iron and Steel hist. Spec. Report No. 43, pp. 195-205. (Esp. Figs. 48-51 and p. 200). 7. KiRKY, H. W. AND MoRLEY, J. I., J. Iron and Steel Inst., 158, 289-294, 1948 (Esp. Figs. 34-40). 8. Clark, G. L. and Gross, S. T., Ind. Eng. Chem. Anal. Ed. 14, 676-683 (1942). 9. GoLDSCHMiDT, H. J., "Tlie Mechanism of Phase Transformations in Metals," pp. 105-119, 1956, The Institute of Metals, London. (Monograph and Report series, Xo. 18.) 10. VoTAVA, E., Berghezan, a., and Gillette, R. H., pp. 603-616 of Reference (1). 11. Glaisher, W. H., Betteridge, W., and Eborall, R., J. Inst. Met., 70, 81 (1944). K. W. Andrews W. JOHNSOX MEDICO-BIOLOGIC RESEARCH BY MICRO- RADIOGRAPHY Theory and Technique Microradiography in medico-biological re- search is that method by which a magnified radiographic image (microradiograph) of an object is obtained. It is desirable to bring the minimum magnification into this defi- nition because there are many workers using powers as low as 2X or 3X, claiming such slightly enlarged radiographs to be micro- radiographs. On the other hand, the power may be called "micro" if it brings into view such structures as cells, thin fibers, capil- laries, etc., which are not seen with the un- aided eye or with the low power mentioned above. It is proposed to consider 35 X to be this minimum power of a microradiograph. Magnification of an x-ray image may be effected with different methods discussed in several articles of this book. Most of these are of recent proposition and are now in the experimental stage; only contact microra- diography has been used in medico-biologi- cal research for more than 20 years (in rela- tion to this research this method is sometimes called "historadiography"). In 1951 Cosslett and Xixon proposed their "x-ray shadow microscope", later termed "x-ray projection microscope" for use in medico-biological research. Theory and tech- nique of the x-ray projection microscopy are described elsewhere in this book and will therefore not be discussed here. The method attracts with its possibility of obtaining a magnification of the radiographic image without interference from emulsion graini- ness. Principles of both methods are illus- trated in Fig. 1. Before it can be indefinitely applied in medico-biological research, however, the x- ray projection microscopy has to work out several problems of which the most impor- tant is that of the resolution of microstruc- o tures with x-rays 1 A or longer. These rays have a very low effective intensity especially at the distance necessary for a magnified image. This intensity is increased in the contemporary x-ray projection microscopy by making use of higher kilovoltage and, consequently, harder x-rays. Literature data on the x-ray projection microscopy show, however, that even these 591 X-RAY MICROSCOPY FILM — _ SPECIMEN X-RAY TUBE CMR PMR Fig. 1. Scheme of contact microradiography (CMR) and projection microradiography (PMR). Dependence is shown in P.M.R. of magnification upon the distance: specimen — film. hard projection micrographs (at 20-30 kv) are made at comparatively short distance be- cause their maximum true magnification (not that achieved by the following photo- graphic enlargement) is shown to be about 30 X. This is much lower than the working magnification of contact microradiography (around 120-150X). However, this difficulty of PMR may be overcome through making use of ultra-sensitive emulsions and by carry- ing out microradiography in vacuum. To our knowledge these experiments are now in progress. There are some reasons to beheve that they will be successful and the x-ray projection microscopy may soon be widely used in medico-biological research together with contact microradiography. Microradiography can take its place in medico-biological research equally with other methods if it complies with the following requirements : A. If it presents an image of details with satisfactory contrasts and sharpness which reduce the error to an admissible minimum. B. If a possibility exists of increasing x-ray absorption artificially via selective x-ray coloring. C. If there is a facility of comparing microradiographical patterns of details with those obtained by other methods of micro- scopical anatomy. It will be shown that all these require- ments are fulfilled in microradiography. Contrast and Sharpness in Microra- diography (Terminology relevant to mi- croradiography used in this article, mostly after Bronkhorst (37). Contrast: The amount of precipitated silver in the emulsion layer of a radiographed and processed film or plate which can be measured in units of blackness. Color of black velvet comprises approximately 2.0 of these units, color of a sheet of white paper — zero units, with all transitions from black to white having a definite numerical expression in these units. Difference in contrasts: Difference between two contrasts of any present in a micro- radiograph. Range of contrasts: The number of contrasts present in a microradiograph. Microra- diograph is poor in contrasts if their range is within the limits 20-30; 100 and more is the range of a microradiograph rich in contrasts. Density: Transparency of a given microra- diograph or part of it to the passing light. Sharpness (definition or detail): The exact- ness in picturing of details. One point of the image corresponds to one point of the object in an ideally sharp microradio- graph. Marginal sharpness: Sharpness seen on mar- gins of the image of a microstructure. Depth sharpness: Sharpness of image of mi- crostructures which are far from the emul- sion layer during microradiography. Unsharpness: Inexactness in picturing of de- tails. It is revealed as blurring either of margins or of structural details of the im- age. 592 MEDICO-BIOLOGIC RESEARCH O Brightness: ^Magnitude of light passing only x-rays from 1 to 10 A are used. through the microradiograph. These wave lengths of white radiation may Resolution of details: The visualization of be obtained at 12 kv and less in the x-ray structures invisible with the unaided eye. tube with tungsten target according to the Resolving power: Ability of an instrument Duane-Hunt law: 12,400 (e.g., microscope), photoemulsion, x-rays, etc. to bring into view structures not seen ^o (A) = — r^ (approximateh) {2) with the unaided eye. o It is obvious that the resolving power of X-rays 1 A long are absorbed by all mi- the microradiograph is closely connected crostructures containing a sufficient amount with both contrast and sharpness (see fur- of an element occupying a high place in the ther about the unsharpness due to "discon- atomic chart, e.g., by any tissvie containing trasting efTect" of the magnification). calcium or by blood vessels containing x-ray Microstructures may be only then vis- opaque media, artificially introduced. Eight ualized in a microradiograph if they absorb kv which generate white x-rays with the a certain amount of x-rays and the effective effective wave length 1.55 A appear to be intensity of rays which passed through these the lowest tension which can resolve details structures is sufficient to produce either a mentioned above with the necessary con- visible contrast within the emulsion layer or trast and sharpness. Harder rays (at 20-30 a conspicuous fluorescence of the screen. All kv) produce a blacker image than necessary these dependencies are well expressed in the and do not effectively show certain differ- equation: ences in contrasts, e.g., those between cal- /pr = /abs I reff cium-rich and calcium-poor elements. On the other hand, the effective intensity of x-rays The primary intensity comprises x-ray obtained at 3 kv and less may be too weak to factor, the absorbed intensity depends produce any differentiation. If we take for mostly upon the object factor and the effec- example again the calcium-rich tissue, it tive intensity produces the x-ray image. All should appear homogeneously white in an these factors will now be discussed in detail, ultra-soft microradiograph. X-ray Factor. X-rays may resolve in Lamarque (81-83) has already pointed out two ways: qualitative and geometrical. The that there is no essential difference between first is the direct x-ray resolution; it is con- a microphotograph and a microradiograph nected only with the quality of x-rays used, of tissues and cells low in absorbing x-rays. It will be discussed in this paragraph. The and this assumption is supported to some geometrical resolution is closely connected extent by the following experiment of Boha- with the divergency of x-rays. Some data on tirchuk (unpublished) : two plates of the this resolution will be found in this para- same thickness (about 1 mm) one of lead graph (focal spot size). Other data will be (A) and the other of aluminum (B) were in the next paragraphs on object and image, radiographed with x-rays of different pene- The direct resolving power of x-rays de- trative power (Fig. 2). It is obvious that in pends either upon their wave length, if rays the radiograph made with 6 kv x-rays (1) are monochromatic, or upon the length of there is no difference in contrasts between an effective wave, if rays are "white" or images of both plates; in other words these "continuous" (mixed). The diapason of wave- soft x-rays are alike to fight in their effect, lengths of soft and ultra soft x-rays (those The differentiation becomes visible at 9 kv, longer than 4.2 A) is quite large— from 0.5 but only at 20 kv (2) is it quite clear. Cer- to 100 A. However, in microradiography tainly there is no proved indication in the 593 X-RAY AllCROSCOPY Fig. 2. Magnified radiograph of lead (A) and aluminum (B), both about 1 mm thick. Radiograph o (1) is made with soft x-rays about 1.8 A, radio- o graph (2) with 0.5 A. The specific x-ray absorption is not conspicuous in the radiograph made with soft x-ravs. literature that some new or smaller details may be resolved with x-rays obtained at tensions lower than 3 kv. With 2.0 to 4.0 A the following microstructures can be resolved as described by different authors: cell, i.e., cytoplasm, nucleus (sometimes even nucleo- lus), cell membrane, sometimes cytoplasm inclusions, fibers, some blood elements, etc. In other words, all structures can be resolved which may be seen within the limits of mag- nification of a micro radiograph (see below). The resolution of such structures as chromo- somes, Golgi apparatus, etc. depends more on the resolving power of the emulsion than on the qualitative and geometrical resolu- tions. Therefore the tendency of some authors to use ultra-soft x-rays for every kind of microradiography does not appear sound from the point of view of micromorphology. The beryllium window of the Machlett tube AEG-50A is about 1 mm thick. Ac- cording to Lurie (86) (see Table 1) this win- dow reduces by half the intensity of a beam of x-rays with the effective wavelength o around 2.0 A. It is therefore possible to get o 2.0 A x-rays out of this tube. However, mindful of the great absorption of these rays by air (approximately 20 times more o than of 1 A rays) (110) one has to use a \ery short distance between the beryllium and the object-emulsion. At a longer distance the air between window and object has to be evacuated. X-rays with the wavelength o longer than 2.5 A are absorbed by 1 mm o beryllium. According to Lurie, x-rays 4.2 A long are absorbed by half, even by a beryl- lium plate as thin as 0.14 mm, which thin- ness can be achieved only under laboratory conditions. The industrial limit is now 1 mm. As was mentioned above, the air absorption o of x-rays 4.2 A and longer is very consider- able. For example, according to Victoreen o (110), this absorption for rays 4.2 A long is o almost 70 times more than those of 1 A. Evi- dently, all microradiographs with rays longer o than 2.5 A must be done in vacuum and the tube window, if any, has to be very thin and made of metal which occupies a low place in the atomic chart. Aluminum of 50 n thick- ness is used for this purpose by most workers. Lamarque pointed out that he used lithium but this metal, although a poor absorber of x-rays, is very difficult to handle. Linde- mann glass (containing boron, lithium and beryllium instead of silicon, sodiiun and calcium of ordinary glass) is used in Europe for windows. This glass is hard and resistant to atmospheric pressure even in thin sections. However, the x-ray absorption by Linde- Table 1 Wavelength Half- Value Layei -, in Millimeters (A) Air Beryllium .25 27400 23.8 .63 6440 15.1 1.0 1940 7.0 1.24 1040 4.05 1.54 550 2.41 1.93 287 1.26 2.5 136 .63 3.6 43.7 .21 4.2 30.6 .14 7.0 6.2 .026 9.9 2.3 .010 594 MEDICO-BIOLOGIC RESEARCH mann glass is a little greater than by beryl- lium. A scheme of the AF.G-50A Machlett tube with the vacuum camera is shown in Fig. 3. The similar camera was made for our tube by R. H. Archer (Ottawa, Civic Hospital). Until now we have discussed the continu- ous radiation produced by the tungsten target. Yet there is another way to obtain soft and ultra soft x-rays by using higher kilovoltages. These x-rays, so-called "char- acteristic", are emitted at a definite tension by a target made of a specific material. The above mentioned Duane-Hunt law (S) is not applicable to characteristic rays. Their wavelength depends upon other laws discussed in works on the subject where one can find both the target material and the necessary tension for the emission of this radiation. It is possible to make these characteristic rays monochromatic using special filters. Targets of chromium (X 2.287 A at 6 kv), iron (X 1.935 A at 7 kv) and o copper (X 1.539 A at 9 kv) are producers of characteristic radiation most frecjuently used (38). This characteristic monochromatic radiation is especially important for precise quantitative and qualitative analyses. These, however, meet with many practical difficul- ties in their application to biological speci- mens. The detailed description of methods may be found elsewhere (Clark) (38). The size of the focal spot is very important for obtaining sharp microradiographs. This size is 1.5 mm square in projection in the Machlett AEG-50A tube, and is 40 m in the Hilger & Watt microfocus tube. It may be reduced to still smaller size through the ap- plication of the condenser and objective lenses as is done in the Cosslett-Nixon tube for x-ray projection microscopy or in the electron microscope. The smaller the focal spot the more is the geometrical resolution of x-rays, i.e. the less is the penumbra formation. These depend- encies are discussed in detail in special pub- lications on the subject (11)(23)(38)(90)(99). VILCUUM CAMERA ADAPTER - VKUUM CAMCU ^ PHOTDOW>MIC PlJffC - GASKCTED JOMT SKOMCN I S TO 10 HOIONS THICK I Fig. 3. Scheme of Machlett AEG-50A tube with vacuum camera, specimen and plate in place as given by Lurie (86). The scattering in air of both kinds of secondary rays, modified and unmodified by the Compton effect, is quite considerable for soft rays used in microradiography (see equation 3). However, the influence of scat- tered rays on image formation may be re- duced to a minimum by using several dia- phragms on the path between tube target and object. Lead pipes covering the entire distance between the tube target and the cassette cover may be used with the same effect. Such a pipe is shown in Fig. 23. Object Factor. With a rough approx- imation the x-ray absorption (m) in biologi- cal tissues may be expressed in such an ec^ua- tion: 7n = k\^- Z* (approximately) -|- 0.2 (3) where X is the wave length, Z (AN) number of element in atomic chart, k the constant and 0.2 the coefficient of scattering. Hydrogen (Z 1), carbon (Z 6), nitrogen (Z 7), oxygen (Z 8) and calcium (Z 20) are the main elements present in animal (hu- man) body in considerable quantities; sulfur (Z 16), iron (Z 26) and phosphorus (Z 15) comprise only a small percentage of body components. Zuppinger (116) gives for ma- croradiology the following data on linear coefficients of attenuation (Table 2). From 595 X-KAY MICROSCOPY Table 2. Linear Absorption Coefficients (after table from Zuppinger) Substance Spec. Gravity Chemical Composition Linear Absorption Coefficient Water 1.0 H2O 2.506 X» Protein C 52%, H 7%, N 16%, 0 24%, S 1% 1.78X3 Fat (man) 0.9 C 75.6%, H 12.6%, 0 11.8% 1.135X2 Muscle 1.06 W 80%, Pr 18%, F 1%, Sa 0.9% 2.62X3 Blood 1.06 W 80%, Pr 19.1%, Sa 0.9% 2.61 X3 Transudate 1.008 W 96.8%, Pr 2.4%, Ash 0.8% 2.72X3 Exudate 1.02 W 93.3%, Pr 5.9%, Ash 0.8% 2.69X3 Pus 1.06 W 90.6%, Pr 7.8%, F 0.8%, Sa 0.8% 2.67X3 Tendon (conn, tis- 1.1 W 62.9%, Pr 34.7%, F and simihir subst. 1.9%, in- 2.37X3 sue) organic Sa 0.5% Fat tissue 0.92 W 11.7%, Pr 2.1%, F 86%, Ash 0.2% 1.37 X3 Calcium stone 2.6 CaCOa 22.85X3 Compact bone 1.9 W 26%, Pr 24.5%, F 2.3%, Ash 47.2% 13.24X3 Liver 1.06 W 76%, Pr 20%, F 3%, Ash 1% 2.61 X3 Hairs 1.3 C 50.5%, H 6.4%, N 17.1%, 0 20.7%, S 5%, Ash 0.3% 2.89 X3 Spleen 1.05 W 78%, Pr 16%, F 5%, Ash 1% 2.61 X3 Kidney 1.06 W 83.5%, Pr 15.7%, Ash 0.8% 2.62 X3 Lung W 80.1%, Pr 16%, F 2.7%, Ash 1.2% 2.69 X3 Thja-oid 1.06 W 82%, Pr 16%, Ash 1%, with I 0.016% 2.70X3 Iron 7.86 Fe 107.7 X3 Nerves 1.03 W 76%, Pr 7%, Ch and Lipids 13.5%, Ash 1.5%, SO.5%, P 1.5% W — water, F — fat, Pr — protein, Sa — salts, Ch — Cholesterol 3.12X3 this table one may see that x-ray ab- sorption comprises in muscle 2.62 X^, blood 2.61 X^ liver and spleen 2.61 X^ kidney 2.62 X^ ; in other words they have ec^ual or almost equal absorption coefficients. Only hairs (2.89 X3), thyroid and cartilage (2.70 X^), placenta (2.75 X^) and especially nerves (3.12 X^) possess more or less high coefficients distinguishing them from other tissues. One may find out from this table that some or- gans cannot be differentiated one from an- other in radiographs (for instance liver and spleen, or muscle and kidney, etc.) ; the differ- ence of absorption between some tissues is so minute that they also have little chance to be differentiated (e.g., muscle from liver and spleen). The situation with absorption improves considerably in microradiography due to another factor of equation 8 — X^ If, for instance, the difference in attenuation coefficients between carbon and oxygen is 0.802 for 1 A, this difference with 2.0 A wavelength comprises 5.973, in other words it increases about 7 times (110). However, this possibility of improving the morpholog- ical differentiation has its limits in medico- biological research due to the very slight penetrative power of ultra soft x-rays (see above). From the study of all these data one may conclude that the application of microra- diography to the study of genuine (non- contrasted) body tissues encounters many difficulties which are partly unsurmountable at this stage of development in microradio- graphical technique. Only calcium-contain- ing tissues possess an absorption coefficient so high that they can be differentiated in all their parts without the technique of coloring. One has not to forget, however, that con- 596 MEDICO-BIOLOGIC RESEARCH temporary histology was developed due to the introduction of coloring (or staining). The same possibility exists in microradiog- raphy. Until now the object of microradiography made with white radiation was approached mainly from a morphological point of view. An analytical physico-chemical approach to such a microradiograph is also possible to some extent. This idea was suggested by Lamar que (81) and worked out by Eng- strom (53), who proposed to radiograph simultaneously the specimen and, as the author calls it, the reference system. This consists of foils of nitrated cellulose having different thicknesses. The composition of cellulose is as follows: C— 43%, 0—41%, N — 7 % and H — 5 %, that is, it is quite close to the composition of protein. Fig. 4 shows how microradiography proceeds in this case. It is obvious that each thickness of foil produces a ciuite definite contrast which may be compared with those of tissue components. Consequently, according to this author, the mass (weight) of biological tis- sue may be determined. This interesting idea is connected, however, with some practical difficulties, limitations and possible errors, which are discussed in detail in the original works on this subject (see especially Bratt- gardt and Hyden) (36). The coefficient of scattering in biological tissue seems to be quite considerable (see equation 3). It increases proportionally to the wavelength of x-rays and to the percent- age of water in radiographed tissue. How- ever, since specimens are mostly dehydrated, cut in thin sections and are in close contact with the emulsion during radiography, the influence of scattering on the formation of penumbras may be neglected. The thickness of the object and its in- fluence on image formation will next be dis- cussed. X-ray Image Factor. A microradio- graphical image is the result of complicated interrelations among the qualitative resolv- PHOTOGRAPHIC EMULSION DARKENING OF EMULSION BY X-RAY STANDARDS -VARYING KNOWN MASS ALUMINUM FILTER Fig. 4. Scheme of Engstrom's reference system radiographed with specimen. Standards produce image of various contrasts depending upon known mass of foil. Identical image contrasts of reference system and specimen have evidently the identical mass (from Fitzgerald) (63). ing power of x-rays, absorption of x-rays by varied object components, geometrical reso- lution by x-ray beam, and the graininess of the emulsion. The last two items have a direct relation to the image formation and therefore will be discussed in this paragraph. Geometrical Characteristics of the X-ray Beam. Let us imagine that the immobile specimen is radiographed with divergent x-rays emitted from the point-like focal spot, on a practically grainless emulsion of negli- gible thickness and with no interference of secondary radiation. Then the unsharpness will be directly proportional: (a) to the size of the focal spot discussed above, (b) to the thickness of the specimen and, connected with it, the superimposition of images of microstructures, and (c) the distance be- tween specimen and emulsion (Fig. 5). It is obvious that the influence of the sec- ond factor may be diminished by making the object as thin as possible; the best remedy for the third is the closest contact between specimen and emulsion. There are many superimpositions of im- ages of microstructures lying at different levels of a three-dimensional specimen as seen in a two-dimensional microradiograph. The thicker the specimen the greater is the number of these superimpositions and the less is the chance to obtain an image of microstructure free of them. It is observed that the unsharpness caused by superimposi- 597 X-KAY MICROSCOPY ,f, A Y \ '^ // A I 1 b s / 1 ' ' / / V ik \ \ \\ ^ l\ \ \ " ' — >lu „_ Fig. 5. Scheme of geometrical unsharpness (from Engstrom) (54). S — specimen, F — film, A — focal spot, b — distance specimen — focal spot, a — • distance specimen — emulsion, U — penumbra. It is evident that penumbra (U) is the greater the far- ther the specimen is from emulsion. The closer it is, the less its image is blurred hy divergent x-rays. tions, even if it seems insignificant at low power, becomes more and more pronounced with increasing magnification. This loss of sharpness together with the "discontrasting" effect of magnification diminishes evidently the resolving power of the microradiograph. No literature data are available on the ratio between the thickness of specimen and the maximum possible magnification of mi- croradiograph. Cosslett (40) showed that the resolution of microstructures in an elec- tron micrograph of a section cannot be more than one tenth of the thickness of the sec- tion. In other words, one cannot expect to see details smaller than one tenth of a micron in an electron micrograph of a sec- tion one micron thick. Although there are no direct indications in the literature, ap- parently a similar ratio exists in optical microscopy. However, the optical resolution of the microscope may be improved to some extent by making use of the micrometer. The use of this is very limited in the study of contact microradiographs. The microm- eter may be applied only for the better view- ing of parts of microradiographical image lying in different planes of the emulsion coat. Evidently, the above-mentioned ratio has to be in microradiography other than 1 : 10 pointed out by Cosslett, probably 1 : 2 or even 1:1. According to our experience one may expect to obtain a sharp x-ray image of a microstructure, that is to say, an image which remains sharp at maximum magnifi- cation, if the radiographed section contains merely this structure or is at least no more than twice as thick as this structure. In other words, it is possible from the geo- metrical point of view to get a sharp image of a structure 5 m in diameter only in micro- radiographs of sections 5 ju or at the most 10 M thick. The contact between the emulsion and the object is achieved by pressure on all parts of the object. This may be done with a film of some polyethylene resin (e.g. "Teflon," "Alathon" of DuPont) 0.00001 in. thick (or even thinner if available). These films only slightly absorb x-rays, are resistant to alka- lies and acids, and are structureless as seen in microradiographs. In the cassette used by us (Fig. 6) the plastic film is stretched on the interior (in- ferior) side of the upper cassette cover. When the cover is in place the plastic film presses the object which is put on a fine- grain plate in the middle of the cassette base. After the plastic has worn out it may be changed easily. In the middle of the cas- sette cover there is a round opening 1 in. in diameter over which a light-proof film is stretched which does not absorb x-rays. Aluminum is not good for this purpose be- cause it is mostly transparent to light rays in thin sections and has a structure con- spicuous in microradiographs. Light-proof films may be made in one's own laboratory by bathing celloidine film (about 0.00001 in thick) in an alcoholic solution (0.25 %) of Sudan black. The entire upper part of the cassette is covered with lead 2-3 mm thick in order to protect the x-ray plate from any secondary radiation emitted by air. The 598 MEDICO-BIOLOGIC RESEARCH Round opening in lead covered with light proof filmsJ Top cover of cassette, with lead lifted (reduced size) Plastic Lead(2mni) Square opening in top ■Top cover Equal distance Base part of cassette, seen from side Depression (0,75mra) Base part, seen from above Scheme of Microradiographlcal Cassette Fig. 6. Scheme of MRD cassette used by Bohatirchuk (explanation in text). base and cover of the cassette are made of hard wood. The cassette proposed by Sherwood (104) is much more complicated. In this "vacuum exposure holder for microradiography" the contact between specimen and emulsion is achieved through the evacuation of air be- tween object and emulsion. This is perfonned with an air pump connected with the interior of the cassette (Fig. 7). Unfortunately, this cassette cannot be used in microradiography of medico-biological specimens because, ow- ing to the vacuum inside the cassette, its cover has to be made of a hard and thick plastic material resistant to atmospheric pressure. Such a cover absorbs soft and ultra soft x-rays used in medico-biological micro- radiography. O B C D E / / / / ^l Fig. 7. Schematic drawing of Sherwood's vacuum exposure holder. A — vacuum pump connection B — opaque cover, C — specimen, D — photographic emulsion on glass plate, E, F — grooved trap (from Sherwood) (104). 599 X-RAY MICKOSCOPY The most perfect contact is easily achieved in microradiography of sections of undecalci- fied bone (see below); the most difhcult is to get it in wet specimens. In the first case the bone section has a smooth, dry surface; in the second it may stick to the emulsion and the subsequent tearing off damages the emulsion. In other cases the emulsion is damaged by alkalis or acids present in wet specimens. Different methods are at hand to avoid this complication. Bohatirchuk (23) (26) (125) proposed to mount specimens on thin sheets of plastic (e.g., "Alathon"), and then place the mounted specimen plastic-down on the emulsion. Specimens mounted on plastic are prevented from shrinking ; they may undergo coloring proce- dures and may be preserved for a long time. Engstrom (54) proposed to cover the emul- sion with collodion, usually a few microns thick, prior to mounting the specimen. After radiography the collodion film is taken off easily and the plate developed afterward. In some cases wet specimens may be dried by putting them between two pieces of fine blotting (or filter) paper which in turn are pressed between two glass slides. This sim- ple method is very good, especially in han- dling thick sections. It is quite obvious that the system x-ray tube: specimen: emulsion has to be held motionless during microradiography. The shrinking and the resulting motion of the wet specimen during long exposures, especially inider vacuum, may be avoided only through the shortening of the exposure time. Vibra- tion of the building due to traffic or trembling of the tube due to boiling water in the cooling hoses are among the causes of unsharpness. They should be foreseen and avoided. It is possible in some objects, e.g., bone and teeth, to cut sections serially in planes perpendicular one to another. This allows the reconstruction of micro-anatomical in- terrelations from their microradiographs. It is necessary to mention here that the thickness (discussed in this paragraph) is considered to be ecjual in all parts of the ob- ject. This condition presents difficulties in microradiography of natural (not sectioned) objects. Evidently, a sharp microradio- graphical image may be obtained only of those parts of natural objects which are in close contact with the emulsion and are thin enough to permit the magnification of their miage. Fig. 8. Replica electron micrograph of unex- posed and processed grains of a NTB emulsion (re- produced from Boyd) (35). A. before, B. after de- velopment (explanation in text). Graininess of the Emulsion. The mosaic character of x-ray image is due to the silver grains of the emulsion. The graininess is revealed even at comparatively low magnifi- cation in the usual emulsions of high sensi- tivity. The emulsion for microradiography must have the finest grain possible. Many firms and companies in Europe and America nowadays produce such emulsions. Only the properties of the most frequently used Ko- dak and Gevaert emulsions will be discussed here. Eastman Kodak (USA) manufactures fine-grain spectroscopic plates 649-0 or GH and high resolution plates as well as other samples. Gevaert (Belgium) issues Lipp- mann films and Scientia 5e 56 plates. Emul- sions of these samples of both companies have very much in common. Regular size of grains of a fine-grain emulsion is within the limits of a few milli- microns (5-15 m/i). Grains acquire irregular shape after processing and their size in- creases (Fig. 8, reproduced from Boyd) (35). Even distribution of grains and dense con- centration are prerequisites of a sharp image 600 MEDICO-BIOLOGIC RESEARCH at high magnification. Unfortunately, these properties cannot be controlled easily and they vary therefore from one sample of emulsion to another, and even within the sample itself. There was an attempt to in- crease the density of grains in some experi- mental fine-grain emulsions (Blackett) (19). Although the microradiographs of this au- thor made on this emulsion appear to be a step forward, there is no indication in the literature as yet that the plates coated with this emulsion are produced commercially. Thickness of the emulsion coat varies in different emulsions, being in finest samples around 5-10 /x. Sensitivity of fine-grain emulsions is very low; if the sensitivity of an ordinary x-ray emulsion is taken for one, the sensitivity of Gevaert Lippmann emul- sion will be 1/10,000 and that of No. 649-0 or GH spectroscopic plates of Eastman Kodak about 1/15,000. Exposure time is therefore quite long, from 2 to 60 min- utes, depending in inverse ratio on the kilo- voltage used and on the thickness of the specimen. On the average, one tenth of a sec- ond exposure on x-ray film will require about 16 minutes on Lippmann film and about 25 minutes on Kodak 649 (see also below). Gevaert Lippmann films are more sensitive at lower kilo voltages than Kodak spectro- scopic plates. "Scientia" 5e 56 plates issued by Gevaert not so long ago have very fine grain and approximately the same sensitivity as Kodak 649-GH samples. Resolving power varies in different sam- ples. Producers of fine-grain emulsions claim that the finest samples resolve about 1,000 lines per mm. This means that this finest emulsion may resolve structures of 1 // size. However, what is right for an almost two- dimensional ruler-line is not always so in relation to a biological three-dimensional microstructure. A line is seen relatively sharp even in the presence of grains, but not the biological structure. Besides that, the resolu- tion of objects 1 M in size comes across the limitation of geometrical resolving power mentioned above. Processing of exposed fine-grain films and plates does not differ much from that of or- dinary material but it requires meticulous cleanliness. All the solutions have to be fil- tered every day, the glass ware thoroughly cleaned. Washing has to be carried out with distilled water, the drying to proceed in ex- siccators. Ready microradiographs have to be covered with cover glass for better pres- ervation. In general, darkroom accessories must be as in a chemical laboratory where ([ualitative and quantitative analyses are performed. D-19 or usual x-ray developer produce good results (the latter 50% di- luted). Time of development with these developers is from 3 to 5 minutes. No ad- vantage was found with other developers recommended for fine-grain emulsions, in particular microdot. The time of develop- ment with microdol is much longer (10 to 12 minutes) and the achieved range of contrasts is rather poor. A recipe for the preparation of a fine-grain emulsion in any laboratory is given in the paper by Bohatirchuk (125). Each magnification of any photo-image has a somewhat "discontrasting" effect. This is due to the fact that silver grains, which seem to be solidly packed at a lower magnification, appear separated at a higher one. It is obvious that the smaller the grains of emulsion are and the closer they are packed in the emulsion layer, the less is the discontrasting effect. It is accepted that the minimum difference in contrasts which may be observed in macroradiographs is about 30%. In other words, one cannot distinguish blackness 1.5 from 1.2; only 1.5 from 2.0 or from 1.0. This percentage is about 15-20 % in microradiog- raphy in spite of the discontrasting effect of magnification. Cytoplasm and cell mem- brane, cytoplasm and nucleus may be differ- entiated though their difference in contrasts is even less than 15-20%. This is shown by 601 X-RAY MICROSCOPY the work of Lamarque, Turchini (83) and others. It is necessary to mention here attempts to use for microradiography another ma- terial sensitive to x-rays. W. A. Ladd and M. W. Ladd (80) and Pattee et al. (95) found that a completely grainless x-ray image can be obtained on some plastic (e.g., vinylidene chloride) after a certain exposure time. Although its range of contrasts may be increased through bathing the exposed plastic in NH4OH (10%) the image obtained remains very indistinct. Authors express hope, however, that they will succeed in improving these preliminary results. Limit of magnification of microradiograph. It is obvious from the foregoing that the re- solving power of a microradiograph depends upon three components: qualitative and geometrical resolutions by the x-rays, and resolution by the emulsion. Only then can one expect the maximum resolution in a microradiograph when each of these com- ponents produces, in harmony with the others, the best effect. Both qualitative and geometric resolutions are to some extent under the control of the researcher, but the resolution of the emulsion is not. Summarizing the data discussed above one may conclude that the maximum mag- nification can be obtained if the microradio- , o graph IS made: (1) with x-rays 1 A or longer, (2) of a section 10 n and thinner and (3) on fine-grain emulsion like 649-0 or GH sample. This maximum is around 300-336 X, i.e., 20 X objective and 15 X eyepiece or 42 X objective and 8X eyepiece. As was men- tioned above the average (working) magnifi- cation is around 120 X to 150 X. Since the average thickness of sections used in micro- radiography is 10 ju, 5 M size is the smallest microstructure of which a sharp and contrast image may be expected. Only in the excep- tional case of a finest-grain emulsion may one observe a sharp and contrast image of detail 1 ju size, but this is possible only in microradiographs of sections 1 to 2 /x thick. Apparently one cannot expect notable improvement of the above mentioned limit of magnification in microradiographs for the near future unless a new sensitive material to x-rays will become available. One may consider that the little magnifi- cation possible in microradiography cannot reveal new data unknown to general micro- morphology. This is wrong, as will be shown later, because microstructures are seen from a new and peculiar point of view which is presented due to selective x-ray absorption. After all, knowledge of micromorphology is not so perfect that this new possibility may be rejected. One may conclude that both contrast and sharpness of contact microradiographs can be obtained sufficient for visualization of micromorphological structures within the limits of the method. These data can be used with certain reservations for qualitative and quantitative analyses of biological tissues. Selective X-ray Coloring of Vessels, Canals, Cavities, and other Elements of Biological Tissues After Baker, who proposed to use the term "coloring" instead of "staining" in histology, it was thought that the term "x-ray color- ing" could well designate any procedure con- nected with the artificial increase of x-ray absorbing capacities. Numerous x-ray coloring media (called here also x-ray opaque or contrast media, or simply OM or CM) are used in microradiog- raphy. Choice depends on the purpose of the research. Negative contrasting (through the injection of air) is not used in microradiog- raphy. Opaque media for microradiography may be introduced either in vivo ("vital injection") or post mortem. The introduc- tion post mortem may be carried out in both humans and animals; vital injection is used only in experiments. There are two kinds of vital injection used in microradiography: (1) direct and (2) indirect. The direct method is called microvasography or microradio- 602 MEDICO-BIOLOGIC RESEARCH vasography (Bohatirchuk) (22) or micro- angiography (Bellman) (12). In this method OM is injected into the blood stream and is retained in the blood vessels of the whole body or of a single organ for the duration of the experiment. In the indirect method OM is also introduced into the blood stream but with this it is only transferred to some organs or tissues which absorb this OM. Post mor- tem coloring may also be of two kinds: (3) through direct injection of 0^1, and (4) through selective absorption of OM. Direct Vital Injection of OM. Direct vital injection may be either total or local. The purpose of the total vital injection is to use the heart as a pressure pump for bring- ing OM to the minute vessels of any desired organ of the experimental animal. This method seems to be the most natural way of filling capillaries with OM. Accordingly, OM has to mix with the blood and must not cause any obstacle in the blood circulation, nor shock to the animal or immediate fatal effects. OM particles, if any, have to be much smaller than the diameter of the narrowest capillary (around 5-7 /x). The greater the particles are the greater is the danger of obstructing vitally important capillaries with sudden death of the animal resulting. One has to remember that par- ticles of all OM have a tendency to stick together producing an embolus of larger size than one particle. Thorotrast is the best OM for direct vital injection. It may be introduced into the animals blood in quantity up to 10% of its body weight. It does not irritate the walls of the blood vessels and does not pro- voke spasms as iodine freciuently does. Thorotrast, due to the high number of thorium in the atomic chart (290), produces contrast images even in so weak concentra- tions as 1 or 2 % (more details about thoro- trast see further). The experiment with direct total vital micro vasography proceeds as follows : (1) Blood is taken from the heart or aorta in the maximum permissible quantity (Wiggers) (115). (2) OM is administered. It is better to take out blood and inject OM by portions al- though it prolongs the experiment for 1 or 2 minutes. (3) Pathological agent (cold, heat, drug, etc.) is applied. (4) The animal is sacrificed no later than 5 minutes after the beginning of the experi- ment. The method of killing must not cause the displacement of the contrasted blood in capillaries. Therefore, death has to be achieved as fast as possible. Electric shock is probably the best method in this case. (5) Desired organs are taken out and pre- pared for microradiography according to general rules. Local vital injection is much easier to carry out. OM is introduced directly into the artery of a desired organ (e.g. kidney, ovary, etc.) and this is taken out after OM appears in vasa efferentia. Much greater concentra- tions of OM in the capillaries may be achieved with this method than in total vital injection. In addition to thorotrast other OM may be used (pantopaque, urocon, etc.). Indirect Vital Injection of OM. This examination is used to visualize the elements of the reticulo-endothelial system (RES). It is known that any particles, especially colloidal ones, being introduced into blood, if not excreted (secreted) are at first ad- sorbed and then absorbed by RES cells. Of all opaciue media, "Thorotrast" (Heyden, Germany, and an American prep- aration under the same name) is mostly used for this study. Thorotrast is the trade name for the colloidal solution of thorium dioxide. It contains from 22 to 26 % thorium by volume. Negatively charged particles of thorium are a few millimicrons in size. Thorotrast is well tolerated by ani- mals, mixes with blood perfect^, and its pH is about 7.2. Another preparation of tho- rium is umbrathor. It has an acid reaction (pH about 2.3). It is not so well tolerated as 603 X-RAY MICROSCOPY Fig. 9. MRD (microradiograph) of a normal rabbit liver in a case of indirect vital injection of thorotrast. Section 10 n, 2 hours after injection, approx. X80. Typical distribution of thorotrast in RES ele- ments of liver lobules. All the microphotographs (called MRD) of this article are reproductions of original microradio- graphs, i.e. their ma.ximum white color corre- sponds to the maximum absorption of X-rays in an object, the maximvmi black to the minimum ab- sorption. thorotrast and contains a little less thorium (about 22% by volume). Disadvantages of thorotrast are its slight radioactivity (half life about 25 years) and the perpetual re- tention in cells of RES. These drawbacks, however, are mostly of no importance in research with vital injections which last a short time. Figures 9 and 10 present microradio- graphs of the liver and spleen of rabbit and dog (both normal) after vital injection of thorotrast: 10 cc to rabbit and 15 cc to dog, two hours before sacrificing the animals. One may see rows of Kupffer cells in the liver and histiocytes in the spleen packed with thorotrast (Bohatirchuk) (21) (24). These pictures suggest that the thorotrast impreg- nation of RES cells may be used not only for the study of the morphology of RES ele- ments but also for the examination of their function. Thorotrast disappears completely from the blood in 4 to 6 hours. If, for some reason, thorotrast is contraindicated, one may use the colloidal preparation of iodine: iodocol sol (Mulford Colloid Laboratories, Philadelphia, USA). A similar preparation "iodsolen" was made in Germany (Schering Co.) shortly before the war. Both these preparations are not radioactive and are excreted from the RES cells within 24 hours. However, the x-ray opacity produced by both colloids is very slight and may be de- tected only in very soft microradiographs. In addition to liver and spleen small quantities of thorotrast are found after vital injection in the RES elements of lymphatic glands, kidneys and lungs. However, the data on these findings are very scanty. There are a few indications that iron (Barclay) (6), bismuth (Lamarque) (82) were found in the limgs of experimental Fig. 10. MRD of dog spleen, 24 hours after in- direct vital injection of thorotrast, approx. X180, 10 M. Thorotrast within histiocytes. Observe that OM is only in cytoplasm, nuclei remain black. 604 MEDICO-BIOLOGIC RESEARCH animals. These data also remain imelab- it is also necessary to prevent the outflow of orated. OM from the capillaries, which frequently Montichelli ef al. reported preliminarily happens during cutting. Mindful of this that in some cases of hyperthyroidism in- possibility it is better to use as dispersing creased absorption of iodine was found in the media solutions of starch, dextrose or gum thyroid gland via microradiography. Unfor- arable, because these media are mostly con- tunately, neither a further report by these verted into gels after formalin fixation of the authors about their results nor information specimen. It is also possible to prevent the about the techniciue used is found so far in outflow of OM by keeping the temperature the literature. of the section below freezing point for the OM is injected subcutaneously in the whole time of exposure. For this purpose the method called microlymphography. From interior of the microradiographical cassette this depot OM spreads into minute lym- has to be connected with the freezing gas. phatic vessels. After several hours the animal Although to our knowledge some researchers is sacrificed and the organ of which the use such eciuipment it is not described in the lymphatic capillaries should be studied is literature relevant to microvasography. taken out and prepared for microradiography Sometimes the outflow of OM may be indi- in the normal way. Thorotrast is con- rectly prevented by making use of thicker sidered the best OM for this method (14) (39). sections (up to 500 /x). Though microvaso- Some authors recommend stereo-micro- graphs of these sections permit only low angiography and stereo-microlymphography, magnification they present a good picture Since literature data do not mention any sufficient for general orientation in the achievement in medical research accom- morphology of capillaries. Some instructions plished by these methods their detailed de- on the subject of technique may be found in scription is omitted. Any one interested may works of Bellman (12) ; Barclay (7) ; Okawa find this not only in the works mentioned and Trombka (93) and others, above but also in publications by Engstrom Post Mortem Selective Coloring by (55) ; Oden et al. (92) ; Pattee et al. (95). OM. The first indication in the literature on Direct Injection of OM Post Mortem, the possibility of post mortem coloring may There is a wide choice of OM with particles be found in the work of Clark (38) and small enough to pass through the narrowest Mitchell (89). It was pointed out that some capillaries. Many iodine compounds (dio- qM are absorbed selectively by different drast, pantopaque, uroconsodium, tele- biological tissues. As was mentioned above paque, etc.); colloidal solutions (thorotrast, ^.^ ^^^^^^ ^^ ^^^ ^j^^ ^^^^ "adsorption" umbrathor, colloidal bismuth, etc.); emul- .^^^^^^^ ^^ "absorption" for the initial stages sions (micropaque, bismuth subcarbonate ^^ ^^ consumption because OM is at first etc.) ; various preparations of mercury and ^^^^^^.^^^ ^^^^^.^^ ^^^ ^^^^ membrane and only other OM are used for this purpose. ,^ . -^ u ^ j vu- +u ^, , . ., ,. I- r^^T • \ J. • after some time may it be found withm the The administration of OM into arteries or veins is carried out along general lines, ce as we • covered in textbooks on anatomical and ^^ ^^^ "^^^hod of coloring a ^ect.on of pathological techniques. One should be care- biological tissue is immersed into OM solu- ful to avoid any forcing during the filling of tion for several hours, then washed, dried the vessels with OM. Extravasation caused and radiographed. It was found that some by torn capillaries or aterioles can produce a embedding media (e.g., paraffin, ceUoidin) non-existing contrast in a large area of the impede the normal adsorption of OM. There- microradiograph. As was mentioned above, fore, they have to be removed before immer- 605 X-RAY MICROSCOPY Fig. 11. MRD of a normal rabbit heart after post mortem coloring with thorotrast, 25 m, ap- prox. X35. Selective absorption of thorotrast by contract- ing muscle substance is conspicuous. sion. Besides that, paraffin absorbs x-rays and may produce erroneous images. Bohatirchuk (unpublished) has immersed 15 M (or thicker) frozen sections of cat spinal cord for 24 hours into the following solu- tions: thorotrast, neo-silvol (colloidal silver iodide, 24% of silver), iodocol sol (iodine 8%), ferrocol sol (colloidal iron 3.5%), sequestrol (colloidal lead, 6%) and Lange's colloidal gold solution (0.01%). The most conspicuous selective adsorption found in microradiographs was that of thorotrast. All other solutions were either adsorbed very little (neo-silvol) or could not be detected in microradiographs at all. Note, however, that microradiography in this experiment was done with 1.2 A (8 kv) x-rays; it is possible that softer radiation would bring to view also tissues only slightly colored with OM (the description of morphological findings will be given later). Apparently, silver is selectively adsorbed by the same tissue ele- ments as thorotrast. However, final conclu- sions on the adsorption (or absorption) of silver and other OM (with the exception of thorotrast) depend on further study. Selective adsorption (or absorption) of OM depends probably upon the signs of electrical charges in both adsorbent (OM) and adsorber (element of biological tissue). Adsorber and adsorbent must have opposite charges in order to be attracted. This as- sumption may be proved in the case of thorotrast adsorption by positively charged or amphoteric tissues. Fig. 11 shows a macroradiograph of heart muscle of rabbit after a section of it was immersed into thorotrast for several hours. It is obvious that the amphoteric contracting muscle substance selectively adsorbs OM. Another example is the adsorption of thorotrast by positively charged collagen. It is known also that thorotrast is never absorbed by nega- tively charged tissues such as cartilage, chromatin, etc. One should mention that biological tissues do not show any affinity for neutral OM (e.g., for barium and bismuth). All these data support the old Ehrlich theory of coloring, according to which the sign of electrical charge is primarily re- sponsible for the adsorption (or absorption) of dye (Baker) (5). Summarizing data on x-ray coloring in general, we may say that this method can be used in microradiography and should bring, even with the limitations of magnifi- cation mentioned above, new data impor- tant for medico-biological research. Comparison of jVIicroradiographic Data with those of Other Methods used in Anatomy and Histology As was mentioned above, a microradio- graph presents usually a real image of a biological microstructure and not its ''shadow" as many authors think (some of them even use such a strange term as "shadowgraph") (47). Only in an ordinary photograph does one see sometimes an ob- ject and its structureless shadow; in the microradiograph each image is full of struc- ture. In addition to that, it is possible to color some microstructures, low in absorbing x-rays, with OM and to study their real image in microradiographs. The process of selective absorption of OM and the follow- 606 MEDICO-BIOLOGIC RESEARCH ing visualization of microstructures is very similar to that in histology. No one, how- ever, calls histological specimen "shadow" of the dyes used for their visualization. In order to avoid this misunderstanding we do not use the term "x-ray shadow" in our work, but call any radiograph of a structure an "x-ray image" of this. Resink (98) calls a real x-ray image of a real structure "supra- liminae" contrary to "infraliminae" or unreal image caused by superimposition of images of other structures or by pemmibras of other origin. If the microradiograph is made in accordance with the above rules its images will be mostly "supraliminae." However, real microradiographic images are not always easy to interpret; they have too many peculiarities characteristic only for x-ray images. The student of morphology as it is seen in microradiographs is able to compare radiographical images with known anatom- ical and histological patterns and to make the three-dimensional reconstruction of microstructures from their images in serial microradiographs (81) (82). It is necessary to emphasize the important works of Vincent (111) (112) (113) (145), Lacroix and Ponlot (79), who were the first to use microradiog- raphy, autoradiography and histological coloring of the same section of bone (bone was decalcified only for histology). Bohatirchuk (30) proposed "stain histo- radiography" for simultaneous study of un- decalcified bone together with surrounding soft tissues and bone marrow (unfortu- nately, it was impossible to call this method "color microradiography" because of pos- sible misunderstanding). In this method a section of undecalcified bone, usually 10 n o thick, is radiographed with x-rays 1.0 A long and subjected afterward to coloring procedures without previous decalcification or disembedding (Fig. 12). A Fig. 12. A — MRD, B — Stained specimen of the same human bone, 72, lOyu, approx. X256. The typical stain-historadiograph. The presence of decalcified tissue with cells is shown in those parts of specimen which do not reveal any calcium. Especially conspicuous this decalcified tissue is in those parts which are in close contact with the calcium-containing bone (dark black in microphotograph). 607 X-RAY MICROSCOPY Practical Achievcmciils Micro ra(lio«iraphy of liiological Tis- sues llijihiy Absorbing X-rays. It is natural to suppose that microradiography of calcium-containing tissues (bones, teeth and various calcifications) would be the main subject in which the selective absorp- tion of x-rays could prove its validity for medico-biological research, and yet there have appeared until recently comparatively few publications, probably because of the difficulty in obtaining thin slides of undecal- cified calcium-containing tissues for which the grinding was commonly used. One can get thin pieces of calcium-containing tissues with this method. However, it is sometimes impossible either to get ground sections thinner than 50 m (especially those of can- cellous bone and pathological calcifications) or to obtain serial sections of calcareous material. These drawbacks, together with the complicated techniques of grinding, appear to be the main obstacles to a wider use of this method. The introduction of plastic as an embedding medium and the manufacture of bone microtomes with heavy knives made of extra hard steel, allow the cutting of undecalcified bone in serial sec- tions 10 /i and less with minimal damage to calcareous material (about techniciues of cutting see below^ in Supplement). This cut- ting together with microradiography of ob- tained sections allow much more complete qualitative and quantitative evaluation of calcium distribution than was possible be- fore. The first approach to the study of bone problems was made by Amprino and Eng- strom (121) who used microradiographs of compact human bone as well as of bones from experimental material. Ground sections 20-50 fj. thick were radiographed with x-rays 2.5 to 4.0 A long. From their microradio- graphs these authors made at least two important conclusions: (1) they queried the validity of the Ebner-Gebhardt theory (114) which states that the cement substance is the only bearer of calcium salts in bone, because they think that fibers are also cal- cified; (2) they indicated that calcium im- poverishment of bone during atrophic proc- ess goes on without, at least to some extent, the simultaneous destruction of the organic matrix. Contemporary findings, arrived at with the improved techniciue of micro- radiography, confirm data of this early work. The next year Engstrom and Engfeldt (57) demonstrated in microradiographs the dif- ferent calcium content in neighboring la- mellae of the osteon. However, their tech- niques did not permit a thorough analysis of this finding. The different mineralization of lamellae was also found by Davies and Engstrom (43) in 1954. Again microradiographs were made of sections too thick to permit fur- ther conclusions on the structure of lamellae. Vincent, as mentioned above, combined mi- croradiography with autoradiography (af- ter administration of S^- to dogs) and his- tology (after previous decalcification). He confirmed the results of Amprino and Eng- strom, Davies and Engstrom in part re- ferring to different calcification of bone ele- ments. Stain historadiography (Bohatirchuk, 1957) (30) opened new possibilities in mor- phological bone studies. As was mentioned above, two properties differentiate this method from earlier ones : serial sectioning of undecalcified bone along with serial micro- radiography of the obtained sections, and the parallel coloring of the same sections with histological dyes. Accordingh^ the morphology of calcium-containing and cal- cium-free bone elements may be compared in microradiographs and histological speci- mens (the latter free of the artifacts caused by decalcification and disembedding). Bohatirchuk (33) in 1959 published re- sults of using this method in his studies on bone morphology. He showed that the main depot of calcium in fine-fibered bone is pres- ent in fibers or fibrils, which circle around 608 MEDICO-BIOLOGIC RESEARCH Haversian canals in compact bone or go parallel to the long axis in trabeciilae of cancellous bone. Bundles of these clacified fibers comprise "x-ray opatiue stratum." It alternates with another stratum called "x-ray rarefied." Both strata run parallel one to another. X-ray rarefied stratum con- sists of (1) thin calcified fibers crossing the bone in direction perpendicular or oblique to Haversian canal or to the long axis of trabecula and, probably, (2) of amorphous calcified ground substance in which intra- osseous lacunae are mostly located. These microradiographic data supplement and clarify those of electron microscopy and explain those differences in calcification of osteons which were seen by the above men- tioned authors. The work presents new data on the localization of intraosseous lacunae which, according to previous authors (e.g., Weidenreich) (114), should be localized along non-calcified fibers (Fig. 13, 14). Microradiographic studies of auditory ossicles were performed by Karlsson et al. {11). They showed that in the ossicles similar calcium distribution occurs as in fine-fibered bone of the human skeleton. Another group of works using microradiog- raphy as a principal method comprise those studying bone patholog3^ They were pub- lished mostly by Swedish authors. Engfeldt and Zetterstrom (52) investigated osteodys- FiG. 13. MRD of undecalcified human bone, 27 years old, 5;u, approx. X80. Stratification of fine- fibered bone, its elasticity are conspicuous. Fig. 14. MRD of undecalcified human bone, 72, 10 m, approx. X120. The pronounced stratification as well as disunited strata on the superior bone sur- face are clearly shown. In comparison with a young bone the aging bone appears to be blacker in I\IRD thus revealing the les.ser content of calcium. metamorphosis in one case of a ten months old child by microradiography and other methods (histology, autoradiography, mi- croscopy with polarized light, etc.). A de- mineralization of the skeleton together with the presence of calcium in kidneys is well shown in their microradiographs. Their find- ing was also very important in the fact that collagen fibrils are not destroyed when cal- cium is lost from them. In other words, the results of these authors contradict the theory of bone resorption accepted by most in the literature (see below). Two cases of Paget 's disease were investi- gated by Engfeldt et al. (58) making use of x-ray diffraction and microradiography. They found a considerable difference in mineral content between normal and Paget 's bone newly organized. This finding is in disagree- ment with- the existing theory according to which no pathological bone is produced in Paget's disease, only normal processes of growth and resorption are exaggerated (Dible, 1950) (45). Of course, on two cases no one can make final conclusions but these results must encourage fresh investigations especially with the new improved technique. Engfeldt ct al. (51) investigated bones of 4 cases of osteogenesis imperfecta. Their findings 609 X-KAY IMICHOSCOI'Y contradict the view generally accepted that the latter are calcified and the former not. the growth of long tubular bones in length is Serial microradiographs demonstrate also not affected in this disease. Their micro- fibers partiall}^ or completely decalcified, radiographs demonstrate clearly pathological From these obser\ations we suggest that the changes in epiphysial plates and metaphyses. decalcification of l)one fibers and not their We would like to emphasize especially the destruction is the characteristic sign of aging important work of Italian scientists in the bone atrophy. In other words, the old field of bone microradiography. Orlandini halisteresis theory is revived by stain histo- et al. (94), using microradiography as the radiographical findings (Fig. 12). main method, investigated formation and It is proposed in the same work to differ- evolution of callus of bone fractures, the entiate between osteoporosis and bone changes in ear ossicles in the course of atrophy using the first term only in case of chronic inflammation of the middle ear, the pathological destruction of bone, bone changes in osteomyelitis, leukemia, etc. As mentioned above, the physicochemical Starcich (108) proposed a new classification approach is also possible in the study of of myelomatous osteopathies based on his "white" microradiographs of calcium-con- microradiographical observations. Prevedi taining tissues and is effected by comparing and Margato (96) investigated the growth of x-ray absorption (1) of a standard wedge of bone. The main purpose of all the above which the thickness is increased step-wise, work was to show the importance of micro- and (2) of the desired part of the specimen, radiography in bone research, and this pur- Aluminum or collodion are used as material pose was excellently achieved. for wedges in this research. The Laboratory Microradiographic studies of resorptive for Morphological Ultra Structural Research processes in bone were initiated by Lacroix in Geneva uses this method (Baud) (10). The and Ponlot (79) with their work on post- reference system of Engstrom and Lind- traumatic osteoporosis, in w^hich this method strom may also be used for this analysis was used parallel with autoradiography and (Amprino and Engstrom) (117). We are of histology. This work demonstrated the im- the opinion, however, that change of bone portance of comparing microradiographs and density alone without some accompanying histological preparations of the same speci- morphological changes cannot be accepted men: large areas of the calcium -free fibrous as a decisive diagnostic sign, especially in tissue w^ere found in histological specimens cases of bone atrophy, because individual within the atrophic osteon. The same ob- variations are very great, servation was made by Bohatirchuk in his This quantitative approach to the deter- works on aging bone atrophy (27) (28) (29) mination of the calcium content via micro- (12G) (127) (128). Decalcified bone tissue was radiography is especially well worked out in found by him everywhere in aging atrophic dental research. Even in 1936 Hollander (76) bone. According to the Albright-Pommer described an original x-ray microdensitome- theory (1), this decalcified tissue is a new ter for qualitative analysis of teeth calcium, matrix produced by osteoblasts, and this The effective intensities were measured by a matrix onl}^ failed to calcify. The same double ionization camera after x-rays passed theory considers the complete destruction through the desired part of the tooth and of the old matrix to be the prerequisite of through the standard wedge of aluminum, any bone atrophy. However, stain histo- The author considered the possible error to radiographs reveal that fibers of atrophic be only 0.3%. He worked Avith tooth speci- tissue are the exact continuation of bone mens 300 m thick and with x-rays obtained collagen fibers; the only difference is that at 18 kv. 610 MEDICO-BIOLOGIC RESEARCH In a series of works on mineralized dental tissue Engfeldt et al. (49) used micro- radiography in combination with autoradi- ography. This work illustrates the impor- tance of the combined method not only in quantitative analysis but also in the finding of morphological changes (with vitamin D deficienc}^, at teeth germs, in the study of dental tubules, etc.). All this work con- tributes to the better knowledge of calcifica- tion both normal and pathological in teeth and to the understanding of the resistance of teeth to decay. There is no doubt that after overcoming the technical difficulty with the serial sectioning of undecalcified teeth, microradiography will make further progress in dental research. We should also mention important morphological contributions to this research by Roeckert (101) and by Applebaum (2) (3) (4). The first studied the structure of dentine in microradiographs, the second the morphology of so-called "mottled enamel" which, according to his observations" happens to exist quite fre- quently in people using fluoridized water." The research has only started in the microradiographical study of such calcareous tissues as tendon and ligament bones, larynx calcifications, etc. One tendon bone and several calcifications of larynx cartilages have been studied by Bohatirchuk. Both kinds of "calcifications" possessed the typ- ical structure of fine-fibered bone: alterna- tion of x-ray opaque and rarefied strata, typical distribution of intraosseous lacunae (colored specimens revealed even the pres- ence of osteocytes within lacunae), calcifica- tion of fibers, typical bone canalization, etc. (Fig. 15). Here microradiography at once answers the question whether the hard tissue is organized bone or a formless calcification. In one case, the hard tissue inside a tumor was found to be true bone. The complete results of these observations will be pub- lished elsewhere. Reasons are not clear, how- ever, why in one case true bone develops and in another a structureless calcium depot Fig. 15. Macroradiograph (A) and microradio- graph (B) of ossifying human rib cartilage, 80, 25 n ground section (courtesy of Dr. B^langer), ap- prox. X80. MRD presents the bone structure in the calcify- ing cartilage: intraosseous lacunae, stratification. In the right lower part of MRD several cartilage cells are seen with the partialh' calcified c}'topla.sm and cell membrane. remains. We believe that microradiography will help to explain these problems. Among other calcareous tissues urinary calculi were under study by microradiog- raphy. William H. Boyce et al. (34) found among other interesting data a striking resemblance between the stratified calcium impregnated fibers in fine-fibered bone and the similarly stratified calcium distribution in calculus stroma. This stratification is es- pecially obvious in calcium oxalate stones. They point out that probably an affinity exists between fiber or fibril of connective tissue or collagen fiber and by hydroxyapatite crystals. Sufficient has been said above to show that microradiography must be integral to re- search on calcareous tissues. Microi-adiography of Tissues per se Low in Absorbing X-rays in Medico - Biological Research. As mentioned above, this microradiography encounters many diffi- culties, particularly in its application to the study of morphology. We know only a few of these morphological contributions. First the w^orks of Lamarque and Turchini (83) should be mentioned, in which morphological 611 X-RAY MICROSCOPY description of microradiographical findings was given for normal and pathological hu- man tissues. Special attention must be given to the work of J. Lamarque and Guilbert (81) where parallel histological and histo- radiographical study was carried out on three tumors: skin epithelioma, mixed tumor of breast in dog, and sweat gland epithelioma of the scalp. The authors gave several histo- radiographical signs typical for each of these neoplasms. They found, for instance, that all negatively charged parts of stroma or cells do not absorb x-rays (e.g., chromatin of the normal nucleus) , but degenerated chromatin of cancer cells does absorb. They described also the microradiographical image of kerato- derma. But their data are very scanty as evidence on the importance of microradiog- raphy in this kind of research. Further stud- ies are necessary. Godlewski in a series of works (71) (72) (73) described microradio- graphical image of some cancer cells, giant cells, etc. These works are very important but, again, observations are very scarce. Other studies include those who apply the "reference system" proposed by Engstrom (53). The most important work on the subject was done in Sweden. The deter- mination of the dry weight of cytological structures (Lindstrom, 1954) (84), the localization and amount of water in biolog- ical samples (Engstrom and Glick, 1956) (59), the distribution of mass and lipides in a single nerve fiber (Engstrom and Liithy, 1949 and 1950) (61) (62), the changes in dry weight of squamous epithelium during car- cinogenesis (Lindstrom and Moberger, 1954) (85) the composition of the nerve cell (Brattgard and Hyden, 1954) (36) are the contributions most widely known in which the reference system of Engstrom was ap- plied. In the United States Fitzgerald (63) made use of the same method and has deter- mined in several works the mass of cancer cells and found that their mass is much smaller than that of normal cells of the same organ. Summarizing, one can say that despite some interesting contributions to our knowl- edge, microradiography of tissues yer se has been too little used and then only with scanty material. Further research is neces- sary in this most important field. MicroralEI)ICO-BIOLOGIC RESEARCH relations between hepatic artery l^ranches cadaver stomach. They found in nncro- and Hver cells, and also between sinusoids radiographs a considerable scarcity of capil- and branches of the portal vein. They sur- laries in the area of the stomach Avail close mise that branches of the hepatic artery are to the ulcer and they explained this local capable of regulating the amount of blood ischaemia as due to the spastic contraction brought to the sinusoids. Ayres de Sousa, of the capillaries in the ulcer area. They Cruz and Morais (106) investigated also mi- think that the blood circulation in this case crovasographical patterns of lungs. A. and proceeds through atriovenous shunt in the J. F. dc Sousa (107) studied biliary capil- submucous plexus and that the spasm of the laries after post mortem thorotrast injection mucosal capillai'ies in the ulcer area is caused into the common bile duct and gave patterns by the blocking influence of sympathetic to these structures. nerves which develops after the abdomen is In Canada, Saunders (102) investigated opened. In cases where this influence was ex- microvasographical patterns of muscles. His eluded no difference was found in the filling important data offer some facts on the blood of capillaries with opaque mediimi in any capacity of these organs. Saunders has in- part of the stomach wall. These authors vestigated also vasographical patterns of consider the important advantage of micro- tooth pulp in different ages (103). It is nee- vasography to be the possibiUty of examining essary to mention here that Saunders is comparatively thick sections of gastric wall alone on this continent in studying the ap- with its numerous capillaries, plication of the x-ray projection microscopy Another explanation of the phenomenon to medico-biological research (Cf. articles by mentioned above was given by Doran (46), Saunders). w^ho considered not the spasm but the The first publication on microvasography scarcity of blood vessels on the lesser curva- in pathological conditions appeared in 1943 ture to be primarily responsible for ulcer (Bohatirchuk) (24). Several experiments development (it is known that ulcers develop were carried out in this work with direct usually on the lesser curvature) . The inverse vital total macro and microvasography. picture was observed by him in the duode- After the injection of thorotrast, such num. However, he did not fill the vessels strongly acting agents as heat or cold were through a special device in which the pres- applied to some part of the body on one side, sure of the opacjue medium was controlled or introduced into the blood stream, e.g., by a manometer as Barclay and Bentley did, adrenalin, etc. The reaction of the big vessels but exercized this control by hand. Bentley to the agent was first studied in serial macro- and Barlow (1951) (18) indicated that this radiographs. After that the animal was sac- "poor" technique is a cause of, as they said, rificed and the capillaries of lungs, liver and Doran's "erroneous" conclusions, other organs w^ere studied. The action of the The scarcity of mucosal vessels filled with agents mentioned above on vessel tone was OM was demonstrated also by Benjamin clearly demonstrated in microvasographs. It (1951) (15) (16) and by Benjamin, Wagner was shown that heat is a vasodilator and and Zeit (1953) (17), who injected at first 50 cold a vasoconstrictor and also at the con- ml of 20 % neo-silvol followed by 50-100 ml siderable distance from the point of applica- of 10% bismuth oxychloride. Xeo-silvol tion. passed into the venous system while particles Barclay and Bentley (1949 and 1951) (8) of bismuth being too large to pass through (9) injected silver iodide into arteries of the arterial capillaries, remained on the ar- stomach resected because of peptic ulcer terial side. These authors presented very and, for comparison, into those of normal persuasive microradiographs in support of 613 X-RAY AllCKOSCOI'Y their opinion. Their technique was used by unknowns cannot be solved by a few investi- Barlow, Bentley and Walder (li)51) (9) in gations. their further microradiographical studies of Great progress was ac-hieved through the atrio-venous shunt in stomach wall; 400 microvasography in the knowledge of the n thick sections were used in this study. They vascular supply of the optic pathways in the re-asserted that shunts do exist and that ophthalmological clinic of Liege University their approximate size is 140 ix. by works of Frangois, Neetens and Collette Serious objections to the conclusions of all (()4) (65) (66) (67) (()8) (69). Making use of the above authors w^re brought by Bellman thorotrast, for injection these authors stud- (12) and especially by Herzog (74). Bellman ied the blood circulation of the chiasma, (1953) injected vessels of a resected stoma(;h geniculate body, optic nerve and other parts 30 minutes after surgery. His microradio- of optic pathways at capillary level. All their graphical findings are in disagreement with conclusions are supported by precise micro- the ones mentioned above. He found quite radiographs. They used comparatively thick adequate filling of the mucosal capillaries in sections (150-400 /x) in order to study large all his specimens. Herzog (1957) used 278 areas of the capillary bed. specimens of resected stomachs in his In one of their works authors studied study. Blood vessels were injected with 15% via microradiography the structure of silver iodide; 400 jj. thick sections were Schlemm's canal (66). They found numerous studied of which 3,000 microradiographs pores in the inner wall of this canal through were prepared. Herzog found the shunt men- which communication occurs between the tioned above only in one case. He could not vitreous of the anterior chamber and the find definite changes either in niunber of canal. These pores were seen in microradio- capillaries or in their morphological proper- graphs after the injection of thorotrast or ties. Herzog is rather pessimistic; he con- angiopac either into the anterior chamber or siders any conclusions about the vascular into Schlemm's canal. The authors surmise function via microradiography to be impos- that the obstruction of the pores prevents sible. He thinks many of the findings are un- free communication between the vitreous of certain. Making use of some drugs with a the anterior chamber and Schlemm's canal strong spastic action Herzog could not find with the resulting increase of intraocular any break in the blood circulation in areas of pressure, and this leads to the development the stomach wall close to the ulcer (drugs of glaucoma. A further contribution to glau- were injected either with silver iodide or coma research was made by the same au- bef ore it) . thors through their study of the influence of The contradictory results of the workers hyaluronidase on the pores (68). According mentioned above prove only the necessity of to Barany large molecules of polymerized further studies in this important branch of hyaluronic acid, w^hich is known to be pres- gastric pathology. First of all the technique ent within the wall, may block the narrow has to be identical if one would make con- pores and consequently provoke glaucoma, elusions of equal validity. In this technique This hypothesis was refuted by Frangois everything counts: opaque medium used, the et al, who found that on the contrary many method of injection and type of OM, period new openings make their appearance after of time between resection and injection and the injection of hyaluronidase into the between injection and microvasography, anterior chamber or after treatment of the type of narcosis used during surgery, his- wall with this enzyme. tological and microradiographical tech- Morphological findings of Frangois et al. niques, etc. Such an equation with so many were confirmed by Pattee et al. (1957) (95) 614 MEDICO-BIOLOGIC RESEARCH in stereo-microvasographical studies. These authors also injected OM into both the anterior chamber and Schlemm's canal and found free communication between these structures. Tumor vessels were under study by Bohatirchuk (19-13) (24) and Delarue et al. (1956) (-44). Bohatirchuk investigated ex- perimental tumors, mostly their metastases in Brown-Pearce rabbit cancero-sarcoma. Vital injection of thorotrast was used, in some cases complemented by post mortem filling. This author found that scarcity and irregularity characterize neoplastic blood vessels. Many sinuses, cavities and irregu- lar canals are present in the tumor stroma through which evidently blood circulation proceeds. This work is in apparent contra- diction to the work of Delarue et al. (44), who investigated the blood supply in tu- mors of the human large intestine. They found that the blood supply is abundant in tumors. However, the irregularity and scar- city of tumor vessels have only an indirect relation to the tumor blood supply. In one sinus or cavity there may be more blood than in many regular capillaries. In addition to that, the vital injection produces results more reliable than post mortem filling. The profuse hemorrhages w'hich happen fre- quently in tiunors of the large intestine are better explained by the presence of sinuses and cavities within the tumor stroma than by the presence of a plexus of regular blood vessels. Fig. 16 presents kidney metastases of Brown-Pearce tmnor in which large cavi- ties in tumor tissue are conspicuous. Of course, only two works on this subject are insufficient to give a complete answer to this interesting question. A very promising and original method of studying blood vessels in situ was used by Bellman and Engfeldt (13) in their experi- ments on the action of gonadotropin on the ovarian blood vessels in the rabbit. The authors put a film wrapped in aluminum foil under the ovaries of the Uving animal, Fig. 16. MRD of rabbit kidney with metastases of Brown-Pearce tumor (two of these are seen on the kidnej' surface, approx. in center of the pic- ture), direct vital thorotrast injection, 15 n, ap- prox. X80. Large irregular cavities and vessels are seen within both metastases filled with thorotrast. Note the great amount of thorotrast in cavities which indicates that despite the poor development of regular vessels the blood supply of tumor stroma is ampler than in normal organ tissue. after which OM was administered and one macroradiograph made, followed with gon- adotropin injection and another macroradio- graph taken on a new film. The considerable enlargement of vessels was found after gona- dotropin. Microradiography confirmed their finding. Bellman and Engfeldt (123) studied kid- ney lesions during hypervitaminosis D. Though changes in blood vessels are con- spicuous in the microradiographs presented, the material is not enough to make any final conclusions. Our review of works in microvasography is incomplete, yet it is adequate to show that with this method many problems of vascular morphology and of blood circulation may be 615 X-RAY MICKOSCOPY Fig. 17. MRD of cats spinal cord, PM thoro- trast coloring, 10/i, approx. X20. The selective ab- sorption of thorotrast by the white matter and fibers is shown. esses bccomo visible while us a rule nuclei (like nuclei of other kinds of cells) arc not conspicuous. Sometimes, however, some x-ray absorbinjj; inclusions arc seen within nucleus, e.g., nucleolus. Fibers with opatiue myelin sheaths are colored almost always and may sometimes be followed in micro- radiographs up to the periphery of the sec- tion. Fibrous network in neuroglia adsorbs thorotrast as do some glia cell elements. White matter generally adsorbs more thoro- trast than gray matter (Fig. 17, 18). Not much is known about selective coloring of other normal tissues besides the few facts pointed out above. Only one work was published on the prac- tical application of this method in patholog- studied in the normal as well as the patholog- ical. Indirect Vital Coloring with OM of Tissues Low in Absorbing X-rays. Although Gobi (70), Dauvillier (42) and Lamarque (82) suggested the possibility of indirect x-ray contrasting of organs and tissues via the blood stream, the first practical application of microradiography for this purpose was made by Bohatirchuk (1940) (22) who pro- posed the injection of thorotrast as a test to determine the function of RES. The idea of the test was that in case of stimulation or blockade of RES the adsorption of thorotrast has to be respectively accelerated or delayed and macroradiographs as well as microradio- graphs had to confirm this assumption. Al- though both acceleration and delay of x-ray opacity of organs and cells of RES was found in some experiments, they were discontinued due to the outbreak of war and were not renewed again. Post Mortem Coloring with OM of Tissues Low in Absorbing X-rays. As mentioned above, sections of the cat's spinal cord after coloring with thorotrast show in microradio- graphs selective x-ray absorption. The am- photeric cytoplasm of typical motor cells of the ventral horn, cell membrane and proc- FiG. 18. Microphotograph of a frontal part of the anterior horn of MRD No. 17 at the higher magnification (approx. X 180) . The selective thoro- trast absorption nervous in cells, fibers and glia elements is shown. In the right upper part of the microphotograph a large neuron with several dendrites is conspicuous. Note some inclusions within the dark nucleus which selectively absorb thorotrast (similar inclusions are seen also in other nuclei). The origin of these inclusions is not clear as yet. 616 MEDICO-BIOLOGIC RESEARCH ical cases. Bohatirchiik (1955) (28) immersed sections 35-70 n thick, both primary and metastatic from experimental and human tumors, into thorotrast for some time and radiographed them afterward. The selective adsorption of thorotrast by tumor tissue in case of hver metastasis (Fig. 19) is quite obvious (Brown-Pearce rabbit cancer). This increase of adsorption is conspicuous in spite of the fact that the normal liver tissue (RES elements) adsorbs OM selectively (see upper part of Fig. 19). It was also found that in some cases of liver metastases the adsorption of thorotrast was increased by liver tissue in which no cancer cells were present (Fig. 20). It is probable that this phenomenon is due to pre-cancerous changes in the liver parenchyma. This observation deserves at- tention because patho-histology has not Fig. 20. Alicrophotograph of :\IRD Xo. 19 at the higher magnification (approx. X200). Although the demarcation line between the tumor and nor- mal liver tissue is conspicuous, many areas in "normal" tissue are seen (e.g. in the upper right corner) which absorb thorotrast selectively. It is believed that these areas represent the precancer- ous stage of the tissue. Fig. 19. AIRD of rabbit liver with metastases of Brown-Pearce tumor. PM thorotrast coloring, 10 /x, approx. X80. Large metastasis is conspicuous in the right lower part of the specimen. Note the selective absorption of thorotrast by tumor cells and sharp demarcation line between the normal and cancerous tissue. Fig. 21. AIRD-s of cat's brain in ca.se of injury. (A) in place of injury (frontal lobe), (B) in some neighborhood of middle brain. PM thorotrast coloring, 10 n, approx. XSO. Selective adsorption of thorotrast by some elements of brain tissue is conspicuous not only in the place of injur}' but also at a considerable distance from it. The origin of elements selectively absorbing thorotrast is not clear as yet. It is believed however, that the de- generated myeline is this absorber. quite a reliable method for the morphological determination of pre-cancerous changes. In another work (unpublished) Auer and Bohatirchuk investigated the pos.sibility of x-ray coloring technique in case of degener- ated nerve fibers. Some signs were found of selective adsorption of thorotrast by these fibers (Figs. 21 A and B). However, results 617 X-RAY MICROSCOPY of these preliminary experiments awuit further study. Future of Microradiography in Medico - Biological Research All the advantages and disadvantages of microradiography relevant to mediro-l^iolog- ical research are pointed out in this chapter. One may see from this description that only in a few fields has microradiography brought evidence of its validity. Many fields are not explored at all. Research in calcareous tissues is one of the most promising fields for microradiography and this is shown by several works men- tioned above. But even in this field micro- radiography only begins to explore its possi- bilities. The vast field of heterotopic bones and calcifications is only slightly investi- gated via microradiography although more exact knowledge of this deviation from nor- mal is very necessary. Here stain historadi- ography in combination with serial cutting is applicable because it helps to understand the degree of participation of soft tissues and their cells in the process of ossification. Another field for the application of micro- radiography is bone growth. Much has still to be learned of the many interrelations between organic matrix and mineral de- posits, e.g., between cartilage cytoplasm nucleus and calcium. Parallel processes of ossification and resorption in course of bone growth are also far from being clear from a morphological point of view, especially the participation of osteoclasts and osteoblasts in bone reconstruction (rebuilding). Canaliza- tion of bone is another field in which micro- radiography was tried only in a few works, while making use of obsolete technicjues. No attempt is known from the literature to differentiate either between arterial and venous capillaries of bone or between cana- liculi of intraosseous lacunae and bone blood vessels. Bone pathology only starts to be investi- gated by microradiography. Here also micro- radiography in combination with coloring technique will be of great help. Differences of calcium impregnation have not been ex- plored at all by microradiography in malign and benign neoplasms nor the process of bone resorption during malignant growth. In our short review many bone diseases have been indicated which also are a "must" for micro- radiographical research. Especially impor- tant is the study of calcium behavior during various dystrophic and fibrotic processes in bone. Osteoporosis and bone atrophy were in- vestigated only in a few papers which brought forward a new concept of the mech- anism of calcimii loss. This new approach to the understanding of bone atrophic process is very important for the study of aging bone. Obviously, microradiographical data cannot be generally accepted until they have been confirmed by other methods. As was shown above microradiograph of bone presents a satisfactory image of calcium distribution even with comparatively low limit of magnification. Microradiography of x-ray colored tissues is elaborated only in a few fields. Micro- vasography, for instance, has neither a uni- form technique nor a generally accepted opaque medium. Every author tries his own method and this results in controversial conclusions as in the above cited works of Barclay and Bentley, Benjamin, Herzog, and others. The attempt to work out a satisfac- tory technique of injection is shown in works of Okaw^a and Trombka and Bellman. How- ever, as was mentioned above, these authors left many questions unanswered. The most detailed study of microvasographical tech- nique is also one of the "musts" for future research. One has to be always mindful that to discredit a method is much easier than to bring evidence of its validity. Microradiography of biological tissues low in absorbing x-raj^s colored with OM is also an acute problem for future research. Here even less is done than in microvasography. 618 MEDICO-BIOLOGIC RESEARCH A B Fig. 22. Magnified macroradiographs of A — bone section fixed (neutral formalin), and B — bone sec- tion immersed into osmium fixative (see text) for 12 hours. The lack of calcium in osmium fixed bone is quite obvious. The range of OM, methods of impregnation and the selective absorption of OM by dif- ferent elements of biological tissues, healthy and diseased, these are only a few of the many problems which will confront the re- searcher. The change of behavior of cellular elements during neoplastic diseases appears to be one of the most promising fields as well as the study of degenerated nerve fibers. The further elaboration of methods is necessary which permit the comparative study of microradiograph and its colored replica. Stain historadiography is such a method. It has proven its validity in research on cal- careous tissues. However, it is necessary to find a similar method also for other tissues low in absorbing x-rays. Ward's bio-plastic cannot be used as embedding mediimi in this case because it is too hard and cannot be cut with the usual microtome. It is therefore necessary to find some other embedding agent which (1) does not produce any image if radiographed, (2) is not too hard and (3) at the same time does not prevent any coloring with histological dyes. Coloring can be done also after the disembedding of radiographed sections but in this case there will be some structural dissimilarity between embedded and disembedded sections. All these and other problems of technique have to be solved. When starting to make use of microra- diography one has first of all to be clear that the working magnification will be 120X-150X and only in rare cases 300X. Consequently there is no sense in trying to "squeeze out" of a microradiograph those details which can only be seen with a higher power. Comparative study of the colored and the microradiographed specimen will be of help in case the higher power is necessary because the colored specimen has the same limits of magnification as a histological slide. Further research is necessarj'- in micro- radiography of tissues per se which are low in absorbing x-rays. Here, first of all, the simplification of equipment and technique is necessary. The expensive and complicated equipment cannot be recommended to anatomy and histology laboratories espe- cially if it requires a trained technical staff for its exploitation. A new, possibly grainless, emulsion is also one of the de- mands of microradiography of tissues low in 619 \-K\Y MICROSCOPY absorbing x-rays. Intracellular components require a greater magnification for their vis- ualization than that possible in microradi- ography. The x-ray industry must understand that now together with macroradiology (diag- nostic and therapy) there exists also micro- radiology (contact microradiography and x-ray projection microscopy) applied tomedi- cobiological research. Practical Suggestions for Starting Mi- croradiography in a Laboratory Choice of Equipment. The equipment depends on the kind of microradiographical research which one has in mind. The most complicated and expensive eciuipment is necessary for microradiography of non- calcareous tissues per se. This has to consist of a special tube with the inbuilt vacuum camera and a transformer which permits a tension so low as 3 kv (see above in "X-ray factor"). So far as we know" this ecjuipment is made only to special order and is not listed by any manufacturer (see equipment of Lamarque, Fitzgerald, Engstrom and others). A much simpler ecjuipment is possible for z •^ Fig. 23. Tube stand (A), tube (B) and cassette (C) in position. X-rays pass through a cone (D). E— Diffraction unit. microradiographical research of calcareous tissues, microvasography, etc. It has to consist of a tube and power supply. The above mentioned AEG-50A Machlett tube with 1 mm thick beryllium window may be used. The tube is supplied with hoses per- mitting cooling with flowing water from any water tap. A tube stand for this tube may be made in any workshop (Fig. 23). The low voltage necessary for the AEG- 50A tube may be obtained either from a diffraction unit or from a special micro- radiographical equipment made by some companies (e.g. Philips). In the latter the midget tube (with beryllium window and air cooling) is built into the transformer housing. The defect of this equipment is its low effectiveness due to the absence of water cooling and other reasons. Diffraction unit produces stabilized, recti- fied secondary current. Kv range is from 5 to 50. It is costlier than the former but is more powerful. It w^orks also with the Machlett tube. One may obtain low voltage also from the usual diagnostic transformer. It is only necessary to regulate input to the primary coil of this transformer. It is known that the coefficient of transformation is in trans- formers of this continent approximately 500. That is to say, 20 v input produces approxi- mately 10 kv. Since autotransformers of usual diagnostic apparatus cannot regulate so low an input, it is necessary to use an additional autotransformer for these low voltages. The switching of this autotrans- former and the calibration of the resulting kv output can be done easily by any en- gineer or even a skilled electrician. Since any obsolete transformer may be used for the equipment this self-made microradio- graphical unit will be within the financial limits of any laboratory. Cutting and Coloring of Calcareous Tis- sues. Specimens of calcareous tissues with surrounding soft tissue of the smallest pos- sible size are removed after death and fixed in 10% neutral formalin for at least 1 week, 620 MEDICO-BIOLOGIC RESEARCH followed by washing in running water for about half a day. Dehydration and embedding proceeds as follows : 4 changes of acetone of at4east 10 min- utes each 10% bio-plastic in acetone for 24 hours 20% bio-plastic in acetone for 24 hours 30 % bio-plastic in acetone for 24 hours 3 changes of 100% bio-plastic for 24 hours each. Mold releasing compound (Ward's) is ap- plied to the inner surface of a dish which aids in removing the block from the mold. When dry, fresh bio-plastic is poured into the dish and catalyst added (tertiary-butyl-hydro- peroxide). The correct ratio between the amounts of plastic and catalyst is important; 25 cc of plastic and 7 to 8 drops of catalyst is used in this laboratory. After the specimen has been immersed in the mixture it is allowed to polymerize at room temperature for 48 hours. When ready, the block is cut to approximately 2 cm^ size with a zig-zag saw and afterwards sectioned with the Jung microtome (Fig. 24). The block surface is slightly moistened with water and after cutting the section is unrolled on a piece of plastic. The upper sur- face of the section is blotted, turned around with the aid of a probe and blotted again. Afterwards the sections are stored between two glass slides. Sections are usually strong enough to stand this handling if care is ob- served. After radiography the sections are colored. Presently good results are achieved with the following procedure : Stain 5 to 10 minutes in 10 cc of sat. sol. Thionin (approx. 1 %) in 50 % alcohol and 90 cc of 1 % phenol in water (Nicolle 1871) wash in distilled water remove excess stain bathing for 5 seconds in 25 cc of dist. water to which 1 drop of HCl has been added wash well for few minutes counterstain for 10 minutes or longer in }/'2 % eosin in water Fig. 24. .Jung bone microtome type K. (Heidel- berg, W. GermanyJ. Technical Data: Knife 3H" xiH". Length of rails 40 cm 16 inches. Total vertical excursion. .30 mm l}i inches. Automatic adjustment of the thickness of the sections up to 50 microns. in steps of 2 microns. Maximum size of the ob- jects to be clamped . . 80 by 55 mm Ground space of the mi- crotome 80 bj' 40 cm Height of the instru- ment 40 cm Weight 90 kg approx. 210 pds. wash in distilled water dip shortly in 70 % alcohol (in and out) differentiation is completed in absolute alcohol till clouds of stain cease coming out of the section clear in xylene and mount in premount (Fisher Scientific). With this coloring the bone appears rosy or red, soft tissues in different shades of blue. This coloring is now used together with the previously reported Boehm-Oppel and Schmorl methods. SUPPLE IVIENT New Data For the current and previous years several contributions to micro-biological research 621 X-KAY mk:iu)scopy were made in which microradiography was used as a principal or partial method of examination, as well as many other publica- tions appeared on the subject of microradio- graphical technique. A new method of microradiography (al- pharadiography) was described by Belanger (122) in which the author used Polonium 210 as a source of alpha particles. A histological slide is put on a fine-grain emulsion plate or film and both are placed in a camera obscura on the distance from 20 to 30 mm. After exposure from 4 hours to 4 days, the obtained alpharadiograph may be studied under the microscope as a usual micro radiograph. The author showed several demonstrative alpha- radiographs he made with this method. There is no doubt that this method will contribute to the medical research in the near future. Bergandall and Engfeldt (123) described a review of various methods of preparing material for microradiography. They have especially elaborated methods of bone grind- ing, embedding, etc. Several critical remarks are made on x-ray projection microscopy. Some valuable suggestions for the prepara- tion of imdecalcified bone for microradiog- raphy was made by Pugh and Savchuk (137). Vincent (113), Engfeldt (48) demonstrated the stratification of fine-fibered bone in mi- croradiographs but did not give morpho- logical explanation to this phenomenon. Bohatirchuk reported results of his micro- radiographic studies on the reversibility of atrophic process in aging bones to the IVth and Vth Gerontological Congress (127) (128). He showed in the first report that certain signs indicate that the calcium impoverishment of aging bones is to some extent a reversible process. Experimental results reported to the ^^th Congress showed that the bone resorption and bone restitution proceed in aging humans and in rabbits along the same morphological lines, and that the experimental bone atrophy is also a reversible process. Bohatirchuk (unpublished) made experi- ments with various bone fixatives used by some authors. Microradiographs show almost the complete decalcification of bone by osmium buffered with veronal up to 7.3 pH during 12 hours of fixation (Scott and Pease) (141). Despite some impregnation of bone and soft tissue by osmium, there is almost no absorption of X-rays typical when cal- cium is present in bone (Fig. 22). The same decalcification may be found after fixation with 30% glacial acetic acid after 4 hours fixation. Both these fixatives were used by some electron microscopists who claimed that they have studied the "undecalcified" bone. The advice given to electron microsco- pists and microradiographists is: before using bone for study, check its calcium content via microradiography. Vascularisation of bone was the subject of microradiographic studies of Brookes (129) (130). Apparently this author has been the first who clearly demonstrated in his precise radiographs the large capillary bed in the rat bone proper under the normal and pathological conditions. Carreto (131), Trueta et al (143) and Vincent (145) studied partly via microradiography the calcification of bone organic matrix in endochondral type of growth. Trueta et al yet once more em- phasized the role of giant cartilage cells and vessels in the process of calcification. Two new contributions were made on the mass determination of cells by Grampp et al (13()) and Rosengren (140) obtained by reference system of Engstrom mentioned above. Findings are in agreement with works on this subject previously cited. Lagergren et al presented in two \\orks (136) (137) microangiographical patterns of the inflammatory hypervascularity and those of fibromatous and fibrosarcomatous tumors. In the latter work authors point out the connection between malignanc}^ of the tumor and its vascularity: according to their findings, the more vascular the more malignant. Amprino (117), (118), Strandh 622 MEDICO-BIOLOGIC RESEARCH (142) used microradiography in their studies of the process of bone remodelling. Amprino and Camanni (119), (120) studied micro- radiographical patterns of dental tlisues. Cosslett (132), One Sing -Poen (138), Grampp and Hallen (133), Istok et al (135) discussed several questions pertaining to the technique of the contact and projection microradiography. Acknowledgment. The work on microradiog- raphy in general and on microradiography of aging bone in particular has been started with the financial and moral support of the University of Ottawa. Afterwards it was supported by grants of the Atkinson Charitable Foundation (Toronto, Canada), Canadian Arthritis and Rheumatism Society, National Research Council of Canada, James Picker Foundation, and is supported now by the USA Public Health Service (Research grant No. A-2298). I express my sincerest gratitude to all these organizations. REFERENCES 1. Albright, F., "Osteoporosis," Annal. In- tern. Med., 27, 861-882 (1947). 2. Applebaum, E., "Mottled Enamel," Dental Cosmos, 78, 969-980 (1936). 3. Applebaum, E., "Grenz Ray Studies of the Calcification of Enamel," /. Dental Re- search, 12, \Sl-\90 (1938). 4. Applebaum, E., Hollander, Ch. F., and BoEDECKER, F., "Normal Pathological Variations in Calcification of Teeth as Shown bv the Use of Soft X-Rays," Dental Cosmos, 75, 1097-1105 (1953). 5. Baker, J. R., "Principles of Biological Mi- crotechnique," Methuen & Co., London, 1958. 6. Barclay, A. E., "Demonstration of Iron in Cells," Brit. J. Radiol., 22, 268-269 (1949). 7. Barclay, A. E., "Microarteriography," Ch. C Thomas, Springfield, 111., 1954. 8. Barclay, A. E., and Bentley, F. H., "The Vacularization of the Human Stomach," Gastroenterologij, 12, 177-183 (1949). 9. Barlow, T. E., Bentley, F. H., and Wal- DER, D. N., "Arteries, Veins, and Arterio- venous Anastomoses in the Human Stom- ach," Surgery Gyn. Obst., 93, 657-671 (1951). 10. B.\UD, Ch. a., "Radiographics et Micro- radiographies Osseuses Quantitatives," Praxis, 15, 329-331 (1957). 11. Beetlestone, a., and Thurner, G., "Some Considerations of Focal Spot Sizes," Brit. J. Radiol., 31, 492-497 (1958). 12. Bellman, S., "Microangiography," Acta Radiol., Supplement 102 (1953). 13. Bellman, S., and Engfeldt, B., "A Micro- radiographic Study of the Effect of Gon- adotropin upon the Blood Vessel System of the Ovaries of Rabbits," Acta Radiol., 43, 459-468 (1955). 14. Bellman, S., and Oden, B., "Experimental Micro-Lymphangiography," Acta Radiol., 47,289-307 (1957). 15. Benjamin, H. B., "The Neurovascular Mech- anism of the Stomach and Duodenum," Surgery Gyn. Obst., 92, 314-320 (1951). 16. Benjamin, H. B., "The Neurovascular Mech- anism of the Mucous Membrane of the Stomach," Surgery Gyn. Obst., 93, 672-675 (1951). 17. Benjamin, H. B., Wagner, N., and Zeit, W., "Intragastric Temperature: Its Varia- tions in Gastric Ulcers," Surgery Gyn. Obst., 97, 19-24 (1953). 18. Bentley, H. F., and Barlow^ T. E., "Aeti- ology of Chronic Gastric Ulcer," Lancet, 260, 470-471 (1951). 19. Blackett, N. M., "The Resolution Obtain- able with a Commercially Available Micro- radiographic Unit," Brit. J. Radiol., 31 368-371 (1958). 20. Blickman, J. R., Recourt, A., and Klop- per, p. J., "Microradiographical Investi- gation of the Vascular Pattern in the Liver of Rats," J. Belg. Radiol., 41, 122 (1958). 21. BoHATiRCHUK, F., "Thorotrast-Test for RES." In: Works on Conference on Con- nective Tissue, Acad, of Sciences of Ukr. S. S. R., Kiev., 1940. 22. BoHATiRCHUK, F., "Experimental Macro- andMicrovasographic Investigations," Dis- sert., Kiev, 1940. 23. BOHATIRCHUK, F., "Die Fragen der Mikro- roontgenographie," Fortschr. Geb. Ront- genstrahlen, 65, 253-263 (1942). 24. BOHATIRCHUK, F., "Totale experimentelle Makro- und Alikrovasographie, etc.," Fortschr. Geb. Rontgenstrahlen, 68, 159-180 (1943). 25. BOHATIRCHUK, F., "Makro- und Mikro- rontgenographie der Blutgefasse bei ex- perimentellen Geschwiilsten," Virchow's Ach. Path. Anat., 313, 216-228 (1944). 26. BOHATIRCHUK, F., "Historoentgenography (Microradiography) of Impregnated Can- 623 X-KAY MICROSCOPY cer Tissue," Am. J. Roentg., 70, 119-125 (1953). 27. BoHATiRCHUK, F., "Somc Microradiofr;raph- ical Data on Bone Ageing," Brit. J. Radiol., 27, 177-182 (1954). 28. BoHATiRCHUK, F., "The Ageing Vertebral Column (Barclay Prize Essay), Brit. J. Radiol., 28, 389-404 (1955). 29. BOHATIRCHUK, F., "Ageing and Osteoar- thritis," J. Canad. Med. Assoc. ,7Q, 106-114 (1957). 30. BOHATIRCHUK, F., "Stain Historadiog- raphy." Stain Technol., 32, 67-74 (1957). 31. BOHATIRCHUK, F., "Applied Microradiog- raphy in Medical Research," In: "X-Ray Microscopy and Microradiography," Aca- demic Press, New York, 1957. 32. BOHATIRCHUK, F., "Erfahruugeu der letzten zwanzig Jahre in der Anwendung der Mikrorontgenographie in der medizi- nischen Forschung," Fortschr. Geb. Ront- genstrahlen, 87, 44-56 (1957). 33. BOHATIRCHUK, F., "Microradiographical Data on Fine-Fibered Bone," Anat. Record, 133, 203-217 (1959). 34. BoYCE, W. H., Pool, Ch. S., Meshan, J., AND King, Jr., J. S., "Organic Matrix of Urinary Calculi," Acta Radiol., 50, 544-560 (1958)." 35. Boyd, G. A., "Autoradiography," Academic Press, New York, 1955. 36. Brattgard, S., andHyden, H., "The Com- position of the Nerve Cell Studied with New Methods," Intern. Rev. Cytology, 3, 455-476 (1954). 37. Bronkhorst, W., "Kontrast vmd Scharfe im Rontgenbild," G. Thieme, Leipzig, 1927. 38. Clark, G. L., "Applied X-Rays," McGraw- Hill Book Co., New^ York, 1955. 39. CoLLETTE, J. M., "LaMicrolymphographie," J. Belg. Radiol., 36, 293-312 (1953). 40. CossLETT, V. E., "How Thin a Section should be," J. Biophys. Biochem. Cytol., 3, 815- 816 (1957). 41. CossLETT, V. E., and Nixon, W. C, "Projec- tion Microradiography," 28, 532-536 (1955). 42. Dauvillier, A., "Realisation de la Micro- radiographie Integrale. Comptes rendus hebd. des Stances de TAcad^mie des Sciences," 190, 1287-1289 (1930). 43. Davies, H. G., and Engstrom, A., "Inter- ferometric and X-ray Absorption Studies of Bone Tissue," Exper. Cell Research, 7, 243-255 (1954). 44. DeLARUE, J., MiGNOT, J., AND BULLIARD, A., "Etude siu- la Vascularisation des Cancers du Gros Intestin," Presse Med., 94, 2164- 21()7 (1956). 45. DiBLE, H., "Pathology," J. & A. Churchill, London, 1950. 46. DoRAN, F., "Aetiology of Chronic Gastric Ulcer," Lancet, 260, 199-202 (1957). 47. Ehrenberg, W., and White, M., "Resolu- tion in Contact Microradiography," Brit. J. Radiol., Z2,55S (1959). 48. Engfeldt, B., "Recent Observations on Bone Structure," /. Bone & Joint Surg., 40 A, 698-706 (1958). 49. Engfeldt, B., Bergman, G., and Hammars- lund-Essler, E., "Studies of Mineralized Dental Tissues," Exper. Cell Research, 7, 381-392 (1954). i 50. Engfeldt, B., Engstrom, A., and Zetter- STROEM, R., "Osteopetrosis, Acta Paediatr., 43, 152-162 (1954). 51. Engfeldt, B., Engstrom, A., and Zetter- STROEM, R., "Biophysical Studies of the Bone Tissue of Osteogenesis Imperfecta," /. Bone & Joint Siirg., 36B, 654-661 (1954). 52. Engfeldt, B., and Zetterstrom, R., "Osteodysmetamorphosis, etc.," J. Paedi- atr., i5 ,V2^U0 (1954). 53. Engstrom, A., "A New Differential X-Ray Absorption Method for Elementary Chem- ical Analysis," Rev. Scient. histr., 18, 681- 683 (1947). 54. Engstrom, A., "Microradiography," Acta Radiol., 31, 503-521 (1949). 55. Engstrom, A., "Stereomicroradiography," Acta Radiol.. 36, 305-310 (1951). 56. Engstrom, A., Bellman, S., and Eng- feldt, B., "Microradiography," Brit. J. Radiol. ,38, 517-532 (1955). 57. Engstrom, A., and Engfeldt, B., "Lamel- lar Structure of Osteons Demonstrated by Microradiography," Experientia, 9, 19 (1953). 58. Engstrom, A., Engfeldt, B., Kelander, C. G., Wilton, A., and Zetterstrom, R., "Paget 's Disease," Acta Pathol. Microbiol. Scandin., 31, 256-261 (1952). 59. Engstrom, A., and Glick, D., "Localisa- tion and Quantitation of Water in Biolog- ical Samples by Historadiography," Sci- ence, 27-28 (1956). 60. Engstrom, A., and Lindstrom, B., "The Properties of Fine-Grained Photographic Emulsions used for Microradiography," Acta Radiol., 35, 33-44 (1951). 61. Engstrom, A., and Luthy, H., "Die Massenverteilung in der Markhaltigen 624 MEDICO-BIOLOGIC RESEARCH 79. 80. 81. Nervenfaser, bestimmt durch Roentgenab- sorptionsmessung," Experientia, 6, 244-245 (1949). 76. 62. Engstrom, a., and Luthy, H., "The Distribution of Mass and Lipids in the Single Nerve Fiber," Exper.'Cell Research, 1, 81-91 (1950). 77. 63. Fitzgerald, P., "Application of Soft X-Rays Methods to a Study of Normal and Neo- plastic Cells," Ann. New York Acad. Sciences, 63, 1141-1176 (1956). 64. FRANgois, J., Neetens, A., and Collette, 73 J. M., "Etude Microradiographit[ue de la Parol Interne du Canal de Schlemm," J. Belg. Radiol., 38, 1-15 (1955). 65. FRANgois, J., Neetens, A., and Collette, J. M., "Microangiographie Oculaire," In- ternat. J. Ophthalmol., 129, 145-159 (1955). 66. FRANgois, J., Neetens, A., and Collette, J. M., "Microradiographic Study of the Inner Wall of Schlemm's Canal," Am. J . Ophthalmol., 40, 491-500 (1955). 67. FRANgois, J., Neetens, A., and Collette, J. M., "Vascularization of the Optic Path- way," Brit. J. Ophthalmol., 39, 220-232, 1955; 40, 341-353; 40, 730-741 (1956). 68. FRANgois, J., Neetens, A., and Collette, J. M., "Microradiographic Study of the Influence of Hyaluronidase on the Perme- ability of the Inner Wall of Schlemm's Canal," Am. J. Ophthalmol., 41, 651-657 (1956). 69. FRANgois, J., Neetens, A., and Collette, J. M., "Vascularisation des Voies Optiques Primaires. Problemes Actuels d'Ophthal- mologie, S. Karger, Basel, 1957. 70. Goby, P., "Applications Nouvelles de la Microradiographie," Bull. Soc. Frangaise de Photogr., 5, 196-197 (1914). 71. CloDLEWSKi, M., Pelissier, M., and Leen- HARDT, P., "Radiomicrographie des Can- cers du Foie et des Voies Biliaires," J. de- Radiol. d'ElectroL, 38, 297-300 (1957). 72. GoDLEWSKi, M., Pelissier, M., and Leen- ^^■ H.\RDT, P., "Radiomicrographie de quel- ques Cellules Geants," J. de Radiol. d'ElectroL, 38, 296-297 (1957). 87. 73. GoDLEWSKi, M., Pelissier, M., Leenhardt, P., and Rovira, R., "Radiomicrographie et Frottis Cytologiques," J. de Radiol., d'ElectroL, 37,9(39-971 (1956). 88. 74. Herzog, W., "Theorie der functionellen Durchblutungsstdrungen beim Magen- geschwuer," Aerztl. Forschung, 3, 134-141 (1957). 89. 75. Herzog, W., "Zur Mikroangiographie," 83. 84. 85. Fortschr. Geb. Rontgenstrahlen, 86, 124- 132 (1957). Hollander, F., "The Degree of Calcification of Teeth: What it means and how we Measure it," Dental Cosmos, 78, 1143-1151 (1936). Karlsson, K., Engstro.m, A., and Eng- strom, H., "Microradiographic Studies of the Auditory Ossicles (Malleus and Incus) and of the Osseous Labyrinth," Acta Radiol., 42, 381-391 (1954). Kemp, F. H., and Nichols, A. F., "Focal Spot Sizes," Brit. J. Radiol., 31, 486-489 (1958). Lacroix, p., and Ponlot, R., "Remarques sur I'Histopathologie de TOsteoporose Posttraumatique," Acta Medica Belgica, J. Hilgers, Bruxelles, 1956. Ladd, W. a., and Ladd, M. W., "High-Reso- lution X-Raj^ Microscopy." In: "X-Ray Microscopy and Microradiography," Aca- demic Press, New York, 1957. LaMARQUE, J., AND GUILBERT, H. L., "NoU- velles Recherches Historadiographiques Appliqu^es a I'Etude des Tumeurs," Bull, du Cancer, 31, 13-28 (1939). Lamarque, p., "Historadiography," Radiol- ogy, 27, 563-568 (1936). Lamarque, P., and Turchini, J., "Une Nouvelle Methode d'Utilization des Rayons X: I'Historadiographie," Mont- pellier Medical, 2, 1-34 (19.37). Ltnstrom, B., "Absorption Spectro-Photom- etry with Extremely Soft X-Rays for the Quantitative Determination of the Dry Weight of Cytological Structures," Exper. Cell Research, 6, 537-539 (1954). LiNDSTROM, B., AND MoBERGER, G., "Stud- ies on the Qualitative Distribution of the Dry Weight in Squamous Epithelium dur- ing Carcinogenesis," Exper. Cell Research, 6, 540-542 (1954). LuRiE, W. B., "A New Advance in Histo- radiography," Cathode Press, 6, 26-27, 1949. Marinelli, L., Ferlazzo, G., Fitzgerald, P., AND Foote, G., "Historadiography of Thyroid Ciland" (Abstract), Am. J. Pa- thol.,25, 811 (1949). Meschan, I., Pool, Ch., Netteleship, A., Winer, M., and Zeman, W., "The Radio- micrography of the Autopsy Brain," Ra- diol., 65, 770-777 (1955). Mitchell, G. A. G., "Microradiographic Demonstration of Tissues Treated bv Me- 625 X-RAY MICROSCOPY tallic Impregnation," Nature, 165, 429 (1950). 90. Nixon, W. C, "X-Ray Microscopy," Re- search, 8, 473-483 (1955). 91. NUERNBERGER, J., EnGSTROEM, A., AND LiNDSTROEM, B., "A Study of the Ventral Horncells of the Adult Cat by Two Inde- pendent Cytochemical Microabsorption Techniques," /. Cell. & Comp. Physiol., 39,215-254 (1952). 92. Od]6n, B., Bellman, S., and Fries, B., "Stereo-micro-lymphography," Brit. J. Radiol., 31, 70-81 (1958). 93. Okawa, C, and Trombka, J. I., "The Tech- nic of Making Microangiograms of Rabbit Bone Marrow," Am. J. Clin. Pathol., 26, 758-764 (1956). 94. Orlandini, J., Bassini, G., Casacci, A., "La Formazione e I'Evoluzione del Callo Osseo di Riparazione delle Fratture," Anali di Radiologia e Diagnostica, 29, 254- 301 (1956). 95. Pattee, H. H. Jr., Garron, L. K., McEwen, W. K., and Feeney, M. L., "Stereo-Mi- crography of the Human Eye." In: "X- Ray Microscopy and Microradiography," Academic Press, New York, 1957. 96. Prevedi, G., and Margato, M., "Studio Microradiographico suU'Accrescimento del Osso Frontale e del Parietale nel Feto Humano," I'Ateneo Parmense, 6, 446 (1945). 97. Priwes, M., "Blood Supply of Long Tubular Bones," Vestnik Roentgenol., 20, 175-198 (1938). 98. Resink, J., "Roentgenogram of Fine Struc- ture," Acta Radiol., 32, 391-403, 1949. 99. Robertson, C. W., and Watson, G., "Pre- cise Measurements of Focal Areas in Diag- nostic X-Ray Tubes and Their Application in Tube Development." Brit. J. Radiol., 31, 489-492 (1958). 100. Robertson, J. K., "Radiology Physics," D. Van Nostrand Co., New York, 1944. 101. RoECKERT, H., "Microradiographic Studies of Teeth," Experientia, 11, 143 (1955). 102. Saunders, R. L. deC. H., "Microradio- graphic Studies of the Vascular Patterns in Muscle and Skin." In: "X-Ray Micros- copy and Microscopy and Microradiog- raphy," Academic Press, New York, 1957. 103. Saunders, R. L. deC. H., "Microradio- graphic Studies of Human Adult and Fetal Dental Pulp Vessels." In: "X-Ray Micros- copy and Microradiography," Academic Press, New York, 1957. 104. Sherwood, H. F., "Vacuum Exposure Holder for Microradiography," Rev. Sci. Instr., 18, 80-83 (1947). 105. SOUSA, A. DE, AND MiRABEAU CrUZ, J., "Mi- croangiographia e Microcolangiographia. Relatorio Apresentado a IV Congresso dos Radiologistas de Cultura Latina," Ber- trand (Irmaos), Lissabon, 1957. 106. SoUSA, A. DE, MiRABEAU CrUZ, J., AND Bello de Morais, J., "Microangiopneu- mographie," Am. Radiol., 1, 9-22 (1958). 107. SousA, A. de, AND SousA, J. F. DE, "An Ex- perimental Study in Microcholangiog- raphy," Brit. J. Radiol., 29, 377-381 (1956). 108. Starcich, R., "Histopathological and Micro- radiographical Observations on Myeloma- tous Osteopathy," Acta Haematol., 18, 113- 125 (1957). 109. Tirman, W. S., and Banker, H. W., "Fur- ther Studies in Microradiography," Am. J. Roentgenol., 75, 366-373 (1956). 110. Victoreen, J. A., "Absorption and Scatter- ing." In: "Medical Physics," O. Glasser, edit.. The Year Book Publishers, Chicago, 1944. 111. Vincent, J., "Les Remaniements de I'Os Compact Marque a I'Aide de Plomb," Rev. Belg. Pathol. Med. Exper., 26, 161-168 (1957). 112. Vincent, J., "Le Remaniement de I'Os Com- pact chez le Cercopitheque," Arch. Biol., 68, 561-579 (1957). 113. Vincent, J., "Microradiographie des La- melles de I'Os Haversien," Arch. Biol., 69, 561-575 (1958). 114. Weidenreich, F., "Das Knochengewebe." In: "Die Gewebe." In: "MoUendorf's Handbuch der speziellen pathologischen Anatomic und Histologic," J. Springer, Berlin, 1930. 115. Wiggers, C. J., "Physiology in Health and Disease," H. Kimpton, London, 1949. 116. ZuppiNGER, A., "Die theoretischen Grund- lagen und Moglichkeiten der Rontgen- diagnostischen Weichteiluntersuchung," Georg Thieme, Leipzig, 1935. 117. Amprino, R., "Rapporti fra Processi di Re- construczione e Distribuzione dei Minerali nelle Ossa," Zeitschr. f. Zellforsch., 37, 144- 183, 1952. 118. Amprino, R., "Struttura Microscopica e Ri- movamento delle Ossa," Atti della Societa Italiana Pathologia, IV, 9-68, 1956. 119. Amprino, R., Camanni, F., "Applicazione del Metodo Istoradiographico alio Studio dei Tissuti Mineralizzati del Dente," Minerva Stomatology, 6, 1-24, 1956. 120. Amprino, R., Camanni, F., "Historadio- 626 MICKOWGIOGRAPHY WITH PROJECTION MICROSCOPE graphic and Autoradiographic Researches on Hard Dental Tissues," Ada Anat., 28, 217-258, 1956. 121. Amprino, R., Engstrom, A., "Studies on X-ray Absorption and Diffraction of bone tissue," Acta Anat., 17, 1-22,-1952. 122. Belanger, L., "Microradiography with a Polonium Alpha Source," J. Bioph. & Bio- chem. Cijtology, 6, 197-202, 1959. 123. Bellman, S., Exgfeldt, B., "Kidney Le- isons in Experimental Hypervitaminosis D," A>7ier. J. Bad., 74, 288-294, 1954. 124. Bergandall, G., Exgfeldt, B., Preparing Material for ^Microradiography, "Acta Pathol. & Microbiol. Scandinavica," 49, 30-38, 1960. 125. Bohatirchuk, F., "Uber Ergebnisse der Mik- rorontgenographie," Acta Rad., 27, 351- 365, 1953. 126. Bohatirchuk, F., Microradiographical Data on Aging Bone Atrophy as Seen in Micro- radiographs and Colored Specimens," J. of Gerontology, 15, 142-149, 1960. 127. Bohatirchuk, F., "Reversibility of Bone Changes in Physiological Aging," Report to the IVth International Gerontological Congress, Merano, Venice, 1957. Tito Mat- tioli, Fidenza. 128. Bohatirchuk, F., "Experimental Bone Atrophy in Rabbits in Comparison with Bone Atrophy in Aging Humans as Seen in Historadiographs and Colored Specimens," Report to the Vth Int. Gerontological Con- gress, San Francisco, 1960 (in press). 129. Brookes, M., "The Vascular Architecture of Tubular Bone in the Rat," Anat. Rec, 132, 25-47, 1958. 130. Brookes, M., "The Vascular Reaction of Tubular Bone to Ischaemia in Peripheral Occlusive Vascular Disease," /. Bone & Joint Surg. (Brit.), 42b, 110-125, 1960. 131. Carreto, L., "Contributo alio Studio della Matrice Ossea Neodeposta," Archive "Putti," X, 211-230, 1958. 132. CossLETT, V. E., "Microscopy with X-rays," Nature, 183 1423-1427, 1959. 133. Grampp, W., Hallen, O., "Microphotometry in X-raj' Contact Microscopj%" Exp. Cell Research, 19, 83-92, 1960. 134. Grampp, W., Hallen, O., Rosenbren, B., "Mass Determination by Interference and X-raj' Microscopy," Exp. Cell Research, 19, 437-442, 1960. 135. Istock, J. T., Miller, C. W., Chambers, F. W., Lyon, H. W., "Historadiography of Hard and Soft Tissues, U.S.A. Armed Forces Med. J., 11, 497-506, 1960. 136. Lagergren, C, Lindbom, A., Sodergerg, G., "Hypervascularization in Chronic Inflammation Demonstrated by Angiog- raphy," Acta Radiol., 49, 441-452, 1958. 137. Lagergren, C, Lindbom, A., Sodergerg, G., "Vascularization of Fibromatous and Fibrosarcomatous Tumours," Acta Radiol., 53, 1-16, 1960. 138. Ong Sing Poen, "Microprojection with X- rays," Martinus Nijhoff, The Hague, The Netherlands, 1959. 139. PuGH, M. N., Savchuk, W. B., "Suggestions on the Preparation of Undecalcified Bone for Microradiography," Stain. Techn., 33, 287-295, 1958. 140. Rosengren, B. H. O., "Determination of Cell Mass by Direct X-ray Absorption," Acta Radiol. Siipplementu , 178, 1959. 141. Scott, B. L., Pease, D. C, "Electron Micros- copy of Epiphysial Apparatus," Anat. Rec, 126, 465-480, 1960. 142. Strandh, J., "Microchemical Studies on Sin- gle Haversian System," Exp. Cell Research, 19, 515-530, 1960. 143. Trueta, J., Morgan, J. D., "The Vascular Contribution to Osteogenesis," /. Bone & Joint Surg., Brit., 42b, 97-109, 1960. 144. Vincent, J., "Correlation Entre la Microra- diographie et ITmage en Lumiere Polaris^e de I'Os Secondaire," Exp. Cell Research, 12, 422-424, 1957. 145. Vincent, J., "Etude Microradiographique de rOssification Enchondrale," Acta Anat., 40, 121-129, 1960. F. Bohatirchuk MICROANGIOGRAPHY WITH THE PROJECTION MICROSCOPE The use of x-rays for microscopy derives from the fact that x-rays have a shorter wavelength than Hght and therefore a greater penetration and higher resolving power. Contact microradiography (1, 2) has long been available to biologists, but it is only recently that the principles of reflection (3) and projection (4) have been applied to x-ray microscopy. The first x-ray projection microscope was reported in 1952 by Cosslett and Nixon. The projection method of x-ray micros- copy rests on the fact that a point source of x-rays casts an enlarged image of a nearby 627 X-RAY MICROSCOPY object on to a distant fluorescent screen or striiment column is 80 cm (approx. 32 in.) photographic plate with a resolution ap- and so the operator can remain seated while proximately equal to the size of the point inspecting the fluorescent screen, source. The Cosslett-Nixon x-ray microscope Preliminary alignment of the microscope operates on this principle and routinely pro- column is carried out by lateral and axial vides a resolution of the order of a half to adjustments of the two lenses, with the ob- one micron. The best resolution obtained to ject of centering the electron beam upon a date with this type of instrument has been small circular viewing screen which is sub- about 0.1 micron. Since the magnification stituted for the target assembly within the depends on the ratio of the target-object to polepiece of the objective lens. A greatly target-plate distance, both of which are vari- demagnified image of the electron source is able, a high primary magnification of up to first formed and focused to a point upon X200 is obtainable depending on the nature the viewing screen. This screen is then re- of the object under study. placed by the target which consequently Projection x-ray microscopy has many ap- emits x-rays from what is virtually a point plications in biology, anatomy, and medi- source, so avoiding the difficult problem of cine. The first biological studies were made focusing x-rays. Further focusing is carried on drosophila (5) and insect flight muscula- out upon a fluorescent screen placed above ture (6) with interesting results. In the vas- the target, with the aid of a small XlO cular field it has been used to study the ocular or magnifier. In actual use the opera- microscopic blood vessels of a variety of or- tor judges the sharpness of the enlarged x-ray gans and tissues, including skin, muscle, image of a test grid placed upon the target bone, teeth, kidney, stomach, intestine, while varying the lens current. A fine silver brain, spinal cord and nerves (7, 8, 9). In mesh grid (1500 mesh/inch) composed of 3 the field of histology it has recently been micron bars and 17 micron spaces is used used to examine unstained sections of animal for this purpose, as well as for determining and human tissue (10) and is also being ap- the resolution and calibrating the magnifica- plied to the determination of the dry weight tion. When the grid image is as sharp as pos- (mass) and elementary analysis of cellular sible, the test grid is removed, and the in- structures. strument is ready for experimental purposes. Operation of the Instrument. The No further lens adjustments are usually x-ray microscope resembles an inverted elec- necessary during the course of the day, so tron microscope in that it consists of two that attention can be devoted to the experi- variable electro-magnetic lenses, termed the ments in hand. A specimen holder and condenser and objective, which are mounted camera can now be placed over the x-ray vertically above the electron gun. A v-shaped beam, an exposure made, and the plate de- tungsten filament within the gun produces veloped. an electron beam which is accelerated at a The instrument has a variable kilovoltage selected kilovoltage. The objective lens is (5-30 kv); hence it is possible to select that fitted with a special polepiece which produces voltage which gives the optimum penetra- a strong magnetic field that focuses the elec- tion and contrast. The choice of the target trons onto a thin target of metal foil sup- and the operating kilovoltage, are dictated ported by a target assembly within the by the study proposed. The target is inter- polepiece. The target gives rise to a beam of changeable, and is simply punched out of a x-rays over which a fluorescent screen or thin metal foil, 3 to 12 microns thick. Tvuig- specimen holder and camera can be placed sten, gold, silver, and various other elements as required. The total height of the in- may be used depending on the characteritsic 628 MICROANGIOGRAPHY WITH PKOJKCTION MICROSCOPE X-ray line deemed most suitable to the study, aperture, is particularly helpful when ex- A target of thin copper foil, either 5 or 10 amining large objects such as a gross section microns in thickness, used with an accelerat- of brain, tissue flap, or the ear of a living ing kilovoltage of 10 to 30 kv provides rea- rabbit. Both types of stage are used with a sonably good resolution and adequate pene- simple metal box camera (2.5 cm high) fitted tration for most vascular experiments upon with plate guideways which allow the selec- animals. Remarkably contrasty and detailed tion of different target to plate distances. The projection micrographs can be obtained with open bottom of the camera rests upon the such a target using standard lantern slide specimen stage, while a circular aperture at plates (e.g. Ilford Contrasty) while working the top accomodates a fluorescent screen for under atmospheric conditions. visual checking of the object or vascular Histological sections however require field prior to the introduction and exposure x-rays of long wavelengths, such as produced of a plate. by an aluminum target. Since better resolu- Specimen Preparation. Microangiog- tion is obtained with a thin target its maxi- raphy, or the x-ray study of microscopic mum thickness should be 4 to 5 microns. An blood vessels necessitates the use of a con- operating voltage of the order of (3 or 7 kilo- trast medium of high radiopacity and small volts or less is essential to the production of particle size, that is readily miscible with reasonable contrast in tissue sections, as is blood, and capable of traversing the capil- the use of a vacuum camera. Focusing is both lary bed. critical and difl&cult at these lower kilovolt- Excellent microangiograms of fresh surgi- ages owing to the low intensity and reduced cal or cadaveric material, both animal and image contrast, but it can be facilitated by human, can be obtained with Micropaque. reducing the target-screen distance before This is a colloidal suspension of barium with viewing the test grid. A new focussing aid a particle size of half a micron and less, which has recently been developed which reduces is capable of entering the smallest capillaries, the difficulties to a large extent (11). Its white color lends itself to anatomical A mechanical stage is placed over the tar- dissection and makes it easy to determine, get when viewing small objects under at- prior to x-ray microscopy, whether good fill- mosphevic conditions. The central aperture ing of the blood vessel net has been obtained, of the stage is large enough to allow a speci- A 10 to 25 % solution of Micropaque warmed men cup to be moved mechanically in the to body temperature and injected intra-ar- two horizontal axes across the x-ray beam, terially until whitening of the venous return and to be approximated to the target in order is detected, usually gives a clear demonstra- to obtain higher magnification. The bottom tion of the vascular pattern of the tissue flap of the specimen cup is either covered with a or organ under study. Attention should of thin supporting film of plastic (6 micron course be paid to injection pressures espe- Mylar) or fitted with rings which give sup- cially in the case of delicate foetal vessels, port to the specimen. All parts of the speci- This solution may even be injected into the men are in focus at once, and owing to the living animal although eventual clotting of great depth of field, stereographic views can the blood is to be expected. Specimens in- be produced either by tilting the specimen jected with Micropaque can be fixed in at low magnification between two exposures formalin. or traversing the specimen across the x-ray In the living animal the concentration, cone at high magnification. volume and toxicity of the contrast medium A simple stage of sheet plastic or metal, are important if vascular irritation and dis- fitted with pressure clamps about a central turbance of circulatory dynamics are to be 629 X-RAY MICROSCOPY avoided. Thr injection of contnist nuMliuni Conventionally fixed material, both ani- into small local arteries produces heautifully mal and linnian, can be readily examined by detailed microangiograms, but cannot be re- the projection method, but it- must be re- garded as other than a blood displacement membered that histological fixatives dena- techni(|ue unrelated to the normal circula- ture protein and appreciably alter x-ray tory picture. X-ray microscopy used in con- transmission. Also no firm conclusions should junction with circulating contrast medium be drawn regarding the functional status of provides a new method of observing the ves- the blood vessels pi'ior to death since fixatives sels of the microcirculation. themselves produce vascular changes. The The water soluble iodinated organic class use of such microangiograms should therefore of radiopaque compounds may be used for be restricted to the study of vascular anat- microangiography, but the best results in the omy. living animal have been obtained with The best projection micrographs are ob- Thorotrast, which is a colloidal solution of tained from freshly injected but unfixed ma- 25 % thorium dioxide. It is nonirritating to terial, it being possible to record large areas vascular tissue and owing to its small par- of moderately thick (1-5 mm) specimens tide size (0.1 micron) readily enters all parts without a blurring of the superimposed vas- of the vascular network. Its high radiopacity, cular images such as limits conventional combined with the fact that it persists in the microradiography. For example, thick sec- blood stream for several hours or more, tions (taken from the kidney, stomach and makes it very suitable for imaging small intestine of the freshly killed albino rat, after blood vessels in the living animal by x-ray retrograde injection of the aorta with Mi- microscopy. The dose should be based on cropaque) showed that the glomerular tufts the estimated plasma volume of the animal and intertubular capillaries of the kidney to reduce circulatory disturbances, and given could be imaged without diflficulty. The intravenously or intra-arterially at some arteriolar, capillary and venular components point distant from the vascular field under of the mural plexuses of the intestine, In- st udy. A dye should be added to the Thoro- eluding the capillary loops within the villi, trast to increase vessel visibility during could also be clearly visualized, operative procedures, since its accidental Projection studies of the rabbit ear in the escape quickly opacifies a tissue area. Speci- freshly killed animal (following bilateral mens injected with Thorotrast are not fixed carotid injection with Micropacjue) reveal by formalin, but can be fixed with alcohol or the extraordinary complexity of the periph- acids (12). eral vascular network. The continuity of its Vascular Results. Projection x-ray mi- small arteries and arterioles with the capil- croscopy is highly suited to the study of the lary bed and the draining venules are demon- small vessels which go to make up the micro- strated with almost diagrammatic clarity, circulation. All the blood vessels remain in The central artery of the ear and its sub- focus owing to the great depth of field of the sidiary branches are seen to give rise to x-ray microscope. This permits the visualiza- smaller arteries that intercommunicate by a tion of the volume pattern of the blood ves- series of arterial arcades, so forming a coarse sels and also its rendition stereoscopically by net or macromesh. Within this coarse periph- simply moving the specimen laterally be- eral net lies a more complex and finer net- tween two successive exposures. Such vas- work or micromesh formed by the arterioles, cular patterns are unobtainable by light capillary bed and ultimate venous radicals, microscopy even with the aid of clearing Micrographs of the ear margin show that techniques. the capillary loops and the small venules 630 MICROANGIOGRAPHY VI ITH PROJECTION MICROSCOPE draining them form a dense palisade, whose venous tributaries drain in turn into the adjacent marginal vein and a coarse sub- cutaneous venous network (Fig. 1). Central areas of the ear (Fig. 2) show the coarse and fine vascular nets, and also short circuits or arterio- venous anastomoses connecting ar- terioles and venules. These structures appear as S-shaped side branches from the ter- minal arterioles which after a short tortuous course run directly into an adjacent venule. Stereomicrographs of such tissue areas pro- vide a three dimensional view of the blood vessels and help to verify the course and connections of the arterio-venous anasto- moses. Also, since the size (diameter) of the vascular field and its magnification can be altered by varying the target specimen and target-plate ratio, it is possible to select and record arterio-venous anastomoses and other minute vascular features for detailed study and measurement (Fig. 3). ^i^M-t Mm%M 'r^*»^r-^&.i.^^L^ Fig. 1. X-ray micrograph of the margin of a rabbit ear showing the coarse arterial network or macromesh, fine capillary net or micromesh, and small marginal veins. X5.3 Af#^%^#^'' Fig. 2. X-ray micrograph of a more central area of a rabbit ear showing the vessels of the mi- crocirculation. The fine vascular net or micromesh is formed by the arterioles, capillary bed and ulti- mate collecting venules. X6.8 An unexpected finding, possibly of physi- ological interest, was that both large and small peripheral nerves could be located by their vascular patterns. In the case of the rabbit ear the longitudinally disposed capil- lary vessels of the intrinsic vascular plexus of the great auricular nerve can be easily seen. The veins draining the nerve trunk are more conspicuous than the arteries of supply, and can be seen to pass in a regular segmental manner tow^ard adjacent subcutaneous veins. Minor cutaneous branches of the nerve are likewise revealed by the vascular nets which lie about them (Fig. 4). Projection x-ray microscopy can also be carried out on the living animal either by taking a microangiogram immediately after the regional injection of contrast medium (blood displacement technique) , or by taking successive microangiograms following a sin- gle intravenous injection of contrast medium 631 X-B\\ Ml tit^. rit^tt aM lit^- y*rjn U> iitf, l^ft, X20 U> rt;('/>r(\ lit*'. f/at.t';rf) of lUt-, j^>*;riph*;ral va*- ^rtikr U;'] durinjf it* t,ran.«jt. ''cirrrijb.tjnsr *liiir W'}iu'u\ii*',). Ah nfffrtitht'jttiortf'A t\if', Ad^f-, or xlij^ rnay U; f:ii\cnWU'A m a« to rt-:\fr(i*f^i\, Wrtu(', krif/wn fraction «^/f th^r t<^/t,ail pla^Tna vr^imri^, with a vj^?w to Jifnitinji dr«'r«jktory m t^tt', f^r of th/T li'/inj( rabbit hav^r U^^i nUuiU'A (f^fj tftf: x-ffty fu'u:r\)*: for \)t',r\- <'k1* lip to two hour*, aft^rr a *inj(J<; d«'/<'/; of "niffffff.rfi.kt ry>rr^s>'fX;fidjr»J( to on*;-t'ffith ''/f thft fi^X'tmhU'A \>\h^:uin, voUiffX-, ('.'>% ffA/iqcj, hh/\ \>t^^i \fi'i*'/^/'A into ty: Uinr^Miti] v«ri of th*j c/fHirAnU'TfiS t-Hr. \)nnuit, thi.*, ftf^rUA the, cur itmy \>t', nnh'ji*'/rf^A to variou.*, ^rsrf^trini^^ital prrK^'/lurfr* .*;ijch hm trauma, fr^/:xinf?, or th^; inj'-/:t.ion of itUh.rfnh/:oUfU)f:h\ ajf^^it*. Ar^ras r/f inflammatory nrft-f^Utn h,fi<\ i,\i('. capil)ari*#, th^jr'rin ha'/<; U->;n imaj(^'/i, a?*, w<;ll a^t sit^rP, r/f art/?rial 'f\/^j'.ui tt/ijh/'j'jd to th^; \i\]nr*A axi^TA. (UmP-trifr^'umf, nuWc/^i'w*', (4 \(><',{x\mA (i,ri.*-;rifi,\ }>.fKi,>.{(t hav<; P^'-^^n rtvi'/trd'A froiu thh tim<; r/f onjvrt, t/> thdr <\'\*it\f\)(-;tir,xu<'A', h\uK&.i an hoiir Te^s,, oO min,; later. Control micrr^grapk-: ?\ifM\A be t^ken of both ears prior to reeord- irjg a series of ^afKnilar events. Vascular re- spofLse to vasomotor drugs may also V>e ob- servr^ by x-ray mirrr^iseopy. LvTnpl-jatic vf:#is^;ls cran als^i U- visualize^] in the living animal by x-ray micrff^-opy. fj()iA r^-sults hiave U;en obtained by the con- tact method, but t>j/; t^frrhnirral advan<^^^ an- tif^XtixuA (iZ) for the proje«^rtion methfxl hiave recently U^jn r^raliz*;'!. Suc^^f^ve x-ray mi- crograph.'' taken shr/rtly after 0'-4/» rnin.; th^; inje^rtir/n r/f contrast mtAhmt dire^rtly into the sul^Krutiin/roiia tissu/; of the living rabWt ear, show the peripheral \yui\)hn.Uf: f^iipiWhry pkrxijj^rs aU>ut the defx>t ar^ra of cor»tra«t mfAmm '^e.g., ,01 -.05 ml TTioro- tra«t; a« well as the collc^rting lymph '^:hjan- nels draining the arr^a ''Fig. 5;. Siirrh micr<^>- lyrnpFiangjr^ram.s al.«^> show the U^d-like silhou^^.te <^/f the valv^^; within the lymphatic f'/o, 4. X-ray rriicr';»f, and l*rft,, ft, i* druinf'A by a/lja- 652 MU KOVNC.KHiKvnU \MIH PKOJH HON MlCKOrk OrE within tiw» ^N^nt^-tiiVS iNTllJxh V^S^a^^l!»- \r: \>>ss«t^K and the vvri\'HS\nila.r ivtwork iM lymph vx>;^^ls A:i thev ^vtss ' alons: tho hrauohos of tho aurioxilar art or v. Rvth blvHxi Y\\«v>sn^l55 and l\niphativ"s haw Ixvn inia^xl s^inniluantvn:fil\' by i\>mhinh\fi: n\\orv>« aiva:iv'»4:rraph\o and m\OT\>l\mphanjjii>4iraphio ttvhniqnesi, Snoh Ttvhniqn«>s ma\' W ox^vvt^xi to add to vMir knowUxijjt^ of tho factors ix^u- lating Ixinph t\o\v in the hvitxjs: animal MioT\v\t\jiii'»4jrHph,v of hmnan n\atenai by tho pt\->joi'tion mothvxi has inohui^xi stnditN? of tho n\vxio of v^5>vnUariffativ>n v\f norvx^, as woll as of tho brnin and spinal ixmxI, Knowb txljiv of tho dtx^pl\' plaixxi at>d mion^six^Mo vossi^ls within tho human brain and sv^inal iX>rvi is n)\dorst.Htxdabl\- limitixl. sinvx> noithor dissix-tion nor traditional histoh^iv-^l moth- ixis ^x^nnit tht> traoit\jj of tht> ix^n^bral art<^rial tnx^ in itsi ontiroty. I'sitxg vXM\trast mtxiia of ix>lloidal v^imonsiv^ns, atul tho ow as a ji"anji>' of o^^pillary tillinjj;, it has IxxMt vxvssiblo to dotormii>o tho ixnirs*\ vxxnntx-t ioi\s, And vwl- vuno iv^ttorn of tho it\trfux^r<^bral wss^^ls of tho hnnuan brain, and rxxx^ui thom storo^v sivpioallx' i^V\ij. (>\ lntort>stii\jj: morphoU\jiiv\Hl foatnrvs v^f tho Fi^ ^, Mxcrvvuxvpi^^VAxn vM' thtf hun\*xx o\^ iv o\vT\l«xl by the* iNwts^ot w^thvxls I'V ^>^!K!!^^\^ vNf the iris CAtx K^ U5>«\i as au iuviex vv( sat-" • tivMx of the bvAVU The pv^^^^ is svn \^s5>e\s of the iris. ciUATY Kxt>' »«vi ehowivi> tv^^ X-RAY MICROSCOPY Fig. 7. Microangiogram showing the pial ar- teries, veins and capillaries on the surface of the human brain. The small club-like projections rep- resent the commencement of cortical and trans- cerebral arteries which descend vertically into the brain substance. Foetus CR.14.5. X5.5 cortical, transcerebral and central vessels of the cerebral microcirculation have been so recorded (14), for example the coarse dis- tributor network and fine capillary bed formed by the pial vessels on the brain sur- face, as well as the myriads of small trans- cerebral arteries which arise therefrom to traverse the entire thickness of the brain (Fig. 7). These last terminate in a subepen- dymal capillary plexus about the central cavity or lateral ventricle of the brain (Fig. 8). The vessels contributing to the intra- spinal circulation such as the radicular, pe- ripheral and central arteries of the spinal cord have been similarly recorded (Fig. 9). The blood supply of human peripheral nerves, such as the sciatic nerve and its branches, has also been studied by projec- tion microscopy. The vessels (arteriae nervo- rum) contributing to the extraneural and intraneural vascular pattern can be easily traced by taking successive micrographs along the length of the nerve trunk. The in- traneural arterioles and capillaries, by re- peated division and anastomosis, form a con- tinuous network along the length of the nerve. In view of the size and multiplicity of these vessels it is not surprising that pe- ripheral nerves bleed when severed. The vascular pattern of developing and adult dental pulp vessels of the human tooth can be demonstrated by both contact (15) and projection methods (16), and viewed stereoscopically, without destruction of the vessels such as occurs in histological section- ing. The vessels within the foetal jaws and dental germs are prepared by retrograde aortic injection with Micropaque at low pressure, prior to excision of both foetal jaws and removal of the contained dental germs for x-ray microscopy. The dental pulp vessels of the extracted adult human tooth can be filled with Thorotrast by suction in- FiG. 8. Microangiogram of human brain show- ing the long transcerebral arteries passing through the brain substance and terminating in a sub- ependj'mal capillary plexus about the lateral ven- trical. X17 634 MICROFLUOROSCOPY jection. The tooth crown is drilled via the occlusal surface and the dentine penetrated until several of the blood filled capillaries of the subdentinal plexus can be seen and sev- ered. Contrast medium is then aspirated through the pulpal vessels via the root canal vessels and the now ruptured capillary plexus. After fixation and decalcification the softened tissue of the tooth can be shaved off so that the final preparation for x-ray microscopy consists of the entire pulp with supporting side w^alls of dentine. X-ray mi- croscopy has provided the first complete pictures of the vascular networks lying within the dental papilla and about the den- tal sac of the developing human tooth; these have accordingly been named the intradental (intrapapillary) and peridental (perifollicu- lar) plexuses respectively. Projection x-ray Fig. 9. Micioangiograiu of human spinal cord showing the peripheral system of blood vessels sur- rounding the cord. Within the cord can be seen the brush-like central arteries, branching to right and left of the midline. Note the radicular vessels in the nerve roots and the capillary bed within each spinal nerve root ganglion. Foetus CR. 19 cms. X5 microscopy provides both overall and detailed pictures of the anatomy of the adult dental pulp vessels with facility. Moreover an intact specimen may be studied from different pro- jections and viewed stereoscopically before commencing sectional studies. The pulpal vascular patterns of all the teeth of the adult dentition have been so demonstrated includ- ing the finest capillaries of the peripheral or subdentinal plexus. REFERENCES 1. Goby, P., C. R. Acad. Set., 46, 686 (1913). 2. Engstrom, a., Bellman, S., and Engfeldt, B., Brit. J. Radiol., 28, 517 (1955). 3. KiRKPATRiCK, P., Nature, 166, 251 (1950). 4. Cosslett, V. E., AND Nixon, W. C, Proc. Roy. ^lOGR APHY x-ray source. When beams of increasing wavelengths pass successively through the same chemical element, each giving a radiograph in constant exposure time, perhaps the first one will produce only blackness of the film matter being trans- parent to short wavelengths; the follow- ing will have a slight impression (but by virtue of differential opacity of matter a corresponding image appears on the film); the last beam will show only a white image, which means the object is c^uite opaque towards the corresponding wavelength used. What seems to be a progressive phenomenon of absorption is due to the ciuality of the chemical element. Another one of different atomic number would have an analogous be- havior towards other wavelengths with a different threshold of opacity. 7. The image is formed by adjacent areas of several different simple elements differentially absorbing radiation as explained above. The diversity of phases each with its own opacity produces the contrast in the whole image. From the concept of opacity, therefore of absorp- tion, results this one of contrast. This kind of image exists for alloys but prac- tically never for living matter. 8. A simple element plays an important role in the constitution of plant tissue; living matter produces images composed by heterogeneous areas. Each of these is made of complex substances mixed with much water. When a simple ele- ment prevails in such a cell area, its presence is indicated by a notable opac- ity or notable transparency; this ele- ment can be identified and estimated in amount, but it represents only a more or less important fraction of the total substance from a given area. The mech- anism of absorption, phenomenon of opacity and radiocontrast occur indi- vidually for each simple element of a complex tissue. The entire image is a summation of their individual effects and the smallest area always appears as a very small entire image. In fact final interpretation of biological structures is relative, but a biologist works always with the complexity of manifestations of life. 9. Highest enlargement obtained today in contact microradiography is X1750 with a striated muscle of mouse, more than X2000 with Allium root tip cells (Engstrom) , which are an excellent per- formance in microradiography. 10. Limit of resolution of a projection mi- croscope is actually the best one with 0.1 m; that of the contact method de- pends on resolution power of light mi- croscope with only 0.2 m- These limits concern usual apparatus and no special experimental arrangements. 11. Total error of ciuantitative microanal- ysis from microradiographs is 2.5%, ten times less than some years ago. The reference system error is more important. 12. As pointed out above the two optical methods currently used with x-rays to observe microscopic structures are: (1) contact microradiography (CMR) ; specimen and film adhere to each other; x-ray image is further en- larged by photographic process; (2) projection microradiography (PMR); specimen and film are distant to each other; x-ray beam produces direct image enlarged. Both are employed by authors in plant field; respective advantages and disad- vantages decide choice of apparatus. Moreover a third method evolves from x-ray optical laws: (3) reflection microradiography (RMR); x-ray beam from speci- men incident at grazing angles on mirrors is reflected to form an en- larged image (cf. p. 672). This 645 X-RAY MICROSCOPY method is still in process of in sections as thick as 80 n, give statistical development at the present time, information of which other techniques are Besides the plant tissue artefacts men- incapable. MR is able to distinguish two tioned above, artefacts inherent in these (or several) groups of substances both hire- optical methods lower the quality of fringent, in cases where they show different radiograms if they are not corrected. x-ray absorption (by example lignin from Results drawn from medical or animal calcium oxalate crystals). Other results in investigations can incite plant cell research plant field show great utility of MR in path- to intensive application. It is known, among ologic and trophic problems: this method other examples, how mineral elements are discloses high absorption of tumoral meatus distributed in calcified bone tissue, how though these are not birefringent in polarized lipids, pentose-nucleoproteins and proteins light, mineral matter being in an unusual in nerve cells are quantitatively determined amorphous state in them. As MR provides by x-rays. Perhaps most botanists ignore information on motion of heavy ions during still these possibilities and also the limita- nuclear division, it serves in the same way tions of microradiographic techniques. for a cell such as pollen and for tissue either The role of microradiography is primarily normal or tumoral such as stem. Leaves fed to detect or localize opaque substances, with different nutrients have been distin- However it does not always take the place guished only by MR. of other techniques. If MR is able to be a If we compare now MR and electron worthy auxiliary method in botanical re- microscopy, both evidently apply to two search, it would be inadvisable to neglect different scales of magnitude. The latter information obtained by other methods such concerns purely descriptive cell morphology as crystallography, histology, cytology or with a resolving power of the order of a few microchemistry. Surely it is expedient to A, but it is closely dependent on the kind know the very interesting attempts to ob- of fixation. MR is on a microscopic scale of tain precise knowledge of living matter by complementary interest to the use of the x-ray techniques, as did Zeitz and Baez (23) ordinary optical microscope because the on Mn and Rb content of tea leaves; their resolving powers of both are similar. Thus physical work is necessary to future progress they contribute together in the solution in microanalysis. But a botanist knows that of the same problems in the plant field. spots in leaves generally are calcium salts Some typical examples are illustrated in and they will have to inquire with a crystal- the figures, each of which is described in a lographic method as to precisely what kind specific legend, of salt composes the crystals of spots. Mi- , . , ,, -J • r i- REFERENCES crochemistry usually provides miormation about ions present in situ. In this way both I- ^o^^^.^ '^"^ (ieneral Information about Micro- chemical composition and localization will be correlated Cosslett, V. E., Microscopy with X-raj^s, Nature, On the other hand there are advantages ^ \r t^ t? ■■ x' d ° Cosslett, V. E., Engstrom, A. and Pattee, of microradiography relative to other tech- h. H., X-ray Microscopy and Microradiog- niques in some results recently acquired. It raphy, Acad. Press, N. Y., 1957. discloses in normal plant tissue interesting Engstrom, A., Biological Ultrastructure, 326 p., architectures as crystallized components, Acad. Press, N. \., 1957. , ,, a Henke, B. L., High Resolution Microradiography, mipregnations, mtra- or extra-cell flow. Microradiographs, in which there is a sum- * ggg also other articles on X-Ray Microscopy mation of numerous elements of a formation in this Encyclopedia. 646 POINT PROJECTION X-RAY MICROSCOPY Technical Report No. 2, Ultrasoft X-Ray Physics, Air Force Office of Scientific Re- search, 1958. Nixon, W. C, in Modern Methods of Microscopy, Butterworths, London, 1956. Ong Sing Poen, Microprojection with X-rays, 132 p., Dnikkerijen Hoogland en Waltman, Delft, 1959. Trillat, J. J., Metallurgical aspects of Micro- radiography, Metallurgical Reviews, vol. 1, Part 1, 27 p., 1956. II. Technical Articles Referred to in Text. 1. Goby, P., C.R. Ac. Sc, 156, 686, 1913. 2. Dauvillier, A., C.R. Ac. Sc, 190, 1287, 1930. 3. Barclay, A. E., Brit. J. Radiol., 20, 394, 1947; 22, 268, 1949. Microarteriography, Oxford, Blackwell Scientific Publications. Barclay and Leatherdale, D., Brit. J. Radiol, 21, 5U, IMS. Leatherdale, Sci. News, Harmondsworth, 19, 61, 1951. 4. Lamarque, p., Radiology, 27, 563, 1936. 5. Clemmons, J. J. AND Aprison, M. H., The Review of Scientific Instruments, 24, No. 6, 444, 1953. 6. Mitchell, G. A. G., The J. of Photographic Science, 2, 113, 1954. 7. Pelgroms, J. D., Paper Trade J., 4, 25, 1952. 8. Legrand, C. and Salmon, J., /. des Recher- ches, CNRS, 26, 298, 1954; Bull, de Microsc. Appl., 4, No 1-2, 9, 1954. 9. Salmon, J., X-ray Microscopy and Microra- diography, 465, Acad. Press, N. Y., 1957. 10. Manigault, p. and Salmon, J., /. des Re- cherches, CNRS, 37, 319, 1956; Salmon, X- ray Microscopy and Microradiography, 484, Acad. Press, N. Y., 1957; Manigault and Salmon, Bull, de Microsc. Appl., 8, No 1, 14, 1958; Salmon, Bull. Soc. hot. Fr., 101, No 7-9,429, 1954; Salmon, C.R. Ac. Sc, 247, 510, 1958; Salmon, C.R. Ac. Sc, 248, 734, 1959; Manigault, Salmon and Rousseau, M., Bull, de Microsc. Appl., 9, 10, 1959. 11. Clark, G. L., Applied X-Rays, 4th ed. Mc- Graw-Hill, N. Y., 1955. 12. Nixon, W. C, X-ray Microscopy and Micro- radiography, 34, Acad. Press, N. Y., 1957. 13. Jackson, C. K., X-ray Microscopy and Micro- radiography, 487, Acad. Press, N. Y., 1957. 14. Ely, R. V., X-ray Microscopy and Micro- radiography, 59, Acad. Press, N. Y., 1957. 15. Ong, S. P. and Le Poole, J. B., Appl. Sci. Res., B, 7, 233, 1958. 16. Engstrom, a., Historadiography, in Analy- tical Cytology, McGraw-Hill Book Co., Inc., N. Y., 1955; Engstrom, A., in Oster G. and Pollister, a. W., Physical Techniques in Biological Research, 3, 489, Acad. Press, N. Y., 1956; Engstrom and Greulich, R. C.,J. Appl. Phys., 27, 758, 1956; Engstrom, Lundberg, B. and Bergendahl, G., Ultra- str. Rev., 1, No. 2, 147, 1957. 17. Greulich, R. C, and Engstrom, A.,Experim. Cell. Res., 10, No. 1, 251, 1956. 18. Engstrom A., and Lundberg, B., Experim. Cell. Res., 12, No. 1. 198, 1957. 19. Brattgard, S. O. and Hydien, H., Acta Ra- diol., suppl. 94, 1952. 20. Engstrom, A., in Oster and Pollister (see 16). 21. Dahl, a. O., Rowley, J. R., Stein, O. L. and Wegstedt, L., Experim. Cell Res., 13, 31, 1957. 22. Dietrich, J., Congress on X-ray Microscopy and Microradiography, Stockholm, 1959. 23. Zeitz, L. and Baez, A. V., X-ray Microscopy and Microradiography , p. 417, Acad. Press, N. Y., 1957. J. Salmon POINT PROJECTION X-RAY MICROSCOPY* (See also p. 661) The shadow projection method of x-ray microscopy is a recently devised technique for producing enlarged images ^vith x-rays. This type has developed rapidly by follow- ing the methods of electron microscope de- sign, leading to a resolution of 1 micron at an early stage, and 0.1 micron after further refinement. In this survey the fundamental principles are presented together with a few applications. Detailed recent results may be found in the four volmnes on x-ray micros- copy given as references. Physical Basis of Projection X-Ray Microscopy A comparison between the contact method of microradiography and projection micro- radiography is shown in Fig. 1. With contact microradiography (CMR) a normal type of x-ray tube can be used at low kilovoltage and * Revised version of a chapter in a book en- titled "X-ray Microscopy and Microradiography" published by Academic Press, New York, 1957. 647 x-KAY mk:koscopy FILM CMR PMR Fig. 1. Comparison of contact microradiog- raphy, CMR, and projection microradiography, PMR. (From V. E. Cosslett and W. C. Nixon,T/ie Times Science Review, London, 18, 10, 1955.) FILM OR FLUORESCENT SCREEN X-RAY BEAUR^ I // .<;aMERA STAGE TARGE OBJECTIVE LENS PUMPS ELECTRON GUN|]q£^ CONDENSER LENS .ELECTRON BEAM IN VACUUM Fig. 2. Essential parts of a Projection X-ray Microscope. the specimen is placed almost in contact with the photographic recording film. All of the enlargement is obtained optically with a light microscope from the original x-ray negative. This method is simple, quick and inexpensive and has been used by many workers, both in biology and metallurgy, since the original discovery of x-rays. With projection microradiography (PMR) the difficulty of focusing x-rays is circumvented by using a special x-ray tube to form a point source of radiation, and initial x-ray enlarge- ment is produced by simple geometrical pro- jection of the image of a specimen close to the source. The production of a magnified image using x-rays has encouraged the use of the term "x-ray microscope," quaUfied by the word "projection," although no x-ray lenses or mirrors are needed. This form of x-ray microscope is seen in detail in Fig. 2. An electron gun (hot tung- sten filament, biasing cap and anode at earth potential) forms a narrow beam of electrons accelerated to 5 to 20 kV. The space trav- ersed by the electrons is evacuated as in an electron microscope and the electron beam is focused by two magnetic lenses placed outside of the non-magnetic vacuum cham- ber. The objective lens with pole pieces re- duces the size of the electron source and the condenser lens, although weaker, conven- iently determines the amount of reduction over-all. The minute electron beam, from a few microns down to 0.1 micron or less, strikes the thin metal foil target that also forms the vacuum wall of the upper end of the x-ray tube. The x-rays generated in this way come from a spot similar to the electron beam size and are used to form an x-ray projected image of the specimen on the film or fluorescent screen. The specimen is held in a stage with three degrees of freedom for scanning the field of view in two directions and changing the magnification with the third. The camera for recording the x-ray image is merely a light tight box holding a cassette so that the plate can be exposed to the x-ray beam. The x-ray target and electron beam are shown schematically in Fig. 3 to demonstrate the method of image formation in the projec- tion x-ray microscope. The electron beam, coming from the left and focused by the ob- jective lens, is enlarged by the lens aberra- tions to a size 5, even if the total reduction in beam size would be much smaller than this. The semi-angular aperture of the lens, a, determines the value of 5 and also the total beam current striking the target. A strong lens with the lowest aberration is used for the final beam reduction although this means a short working distance and consequent placing of the target in the pole piece gap. The exact opposite arrangement is 648 POINT PROJECTION X-RAY MICROSCOPY ELECTRON BEAM \ FRINGES i>M(J^X^ Fig. 3. Image formation by projecting x-ray microscopj'. Unsharpness in the image plane = Ms. First Fresnel fringe half -width in the image plane = M(bX)i. (From W. C. Nixon, Proc. Roy. Soc. A232: 475, 1955.) found in the double condenser lens systems of modern electron microscopes where a long working distance is needed from the lens to the specimen stage. In this case the second lens is weak and aberrations occur but the electron intensity (unlike the x-ray intensity) is sufficient for high magnification on the final screen. The electron beam when striking the tar- get will be scattered and diffused within the metal foil to give a source of x-rays approxi- mately equal to s where s = 2p and p is the penetration distance of the electrons at the kilovoltage applied to the electron gun. A specimen represented by a straight edge a distance b away will cast an enlarged x-ray image onto the screen or photographic plate at distance a with the magnification M = a/h. The blurring due to the source size is also enlarged by M and so the "resolution" in these terms in the object is similar to the x-ray source size. This simple projective magnification means that a thick specimen is magnified by different amounts from front to back but the total specimen image is all in focus. This is easily demonstrated by using a point source of light, say a flashlight bulb and two bat- teries, and a coarse mesh grid. In a darkened room the projected image of the grid will be seen to vary in magnification as the grid is tilted but the image will remain sharp within the limits imposed by the size of the point source of light. There is no overlapping of out-of-focus layers of the specimen and stereographic techniques for qualitative viewing and quantitative measurement can be used even at the highest magnification hoped for in the future. These x-rays have a wavelength of 1 to 10 A, very much shorter than the 4000 A of blue light in an optical microscope. This means that the diffraction limit on resolution is much less serious in x-ray microscopy at the present time but as the resolution improves this effect will join the geometrical limits dis- cussed above. The magnitude of the Fresnel diffraction fringes is shown in Fig. 3 and the width of the first fringe at the plate, i.e., in the image, is D = AI(hXy'^ or in terms of the specimen, d = (bxy^. In this case b is the distance shown in the figure and X is the x-ray wave- length. The recognition of such fringes as a resolution limit is determined by the total magnification of the x-ray micrograph. This fringe width, d, in microns is plotted against the plate distance, a, in cm in Fig. 4. Vari- ous constants must be chosen for this equa- tion, such as the optical magnification of the 649 X-RAY MICROSCOPY lO 5 I FRINGE OS WIDTH _ IN ;j O.I Q05 O.OI -\\\ Mil 1 1 1 Mil 1 1 1 ^ — / ^ / — — /^ — E / ~ / i 1 1 M, \\\\ O.I 0.5 I PLATE 5 lO DISTANCE 50 lOO IN CM Fig. 4. Minimum plate distance from x-ray source to observe a fringe of width d, allowing lOX photographic enlargement. (From W. C. Nixon, Proc. Roy. Soc, A232, 475, 1955.) Fig. 5. Section of dog skull showing the diploe veins. (From C. G. Hewes, W. C. Nixon, A. V. Baez and O. F. Kampmeir, Science, 124, 129, 1956.) X-ray negative, and the actual values will determine the exact result. However, the values shown give the order of magnitude of the effect. It is seen that for a fringe of one micron width, the plate must be 10 cm from the x-ray source for that fringe to be visible in the final image. This distance is longer than usually used and the exposure time would be longer as well. In fact, with one micron reso- lution no one micron fringes have been seen due to Fresnel diffraction for this reason. As the resolution improves to 0.1 micron the plate distance falls to 1 cm and this is the distance most used in practice and indeed fringes of this size have been seen. These then are the main limitations on the performance of the projection x-ray micro- scope. All attempts at making an x-ray mi- croscope of this type quickly lead to a resolu- tion "less than one micron" and equal to the penetration limit for the kilovoltage and tar- get metal used. The different types of lenses that focus and reduce the electron beam merely affect the number of electrons in the final spot and thus the exposure time, with the poorer lens giving a longer exposure but at the same resolution. This implies enough intensity on the fluorescent screen to focus the microscope in the first place. Perform- ance can only be judged if all comparison figures are given such as spot size, beam cur- rent at the target, kV, exposure time, plate distance and specimen. Results An x-ray micrograph of a 5-mm section of dog skull is shown in Fig. 5, taken with Dr. Baez' x-ray microscope at the University of Redlands, California. The blood \'essels within the bone have been injected with a blue vinyl plastic and detail of the diploe veins is shown. This single 5 minute exposure presents the same information as would be obtained after some 25 hours of normal op- tical sectioning and examination. A detailed examination would take several months by normal methods, but a few 5-minute expo- sures with the x-ray microscope. Another example is the frozen-dried foot of a newborn mouse seen in Fig. 6, and taken by Mosley and Wyckoff at the National In- stitutes of Health, Bethesda, Maryland. In this case a study of developing bone and teeth is aided by the penetration possible with the x-ray microscope and freeze-drying is used both for fixation and removal of the water that would obscure the detail of the bone and tissue. Similar photographs of the 650 POINT PROJECTION X-RAY MICROSCOPY mouse head, both new born and embryonic, show the developing tooth buds in situ, and the spatial arrangement is seen with a stere- ographic view. Other metallurgical, mineralogical, chemi- cal and industrial applications may be found in the publications listed at the end of the article. Resolution Electron penetration in the target varies roughly as the square of the kilovoltage of the electron beam and reducing the voltage from 10 kV to 2 kV would imply an improve- ment of 25 X in resolution. In addition the x-rays generated at this lower voltage would be more easily absorbed by the thinner speci- mens and sufficient contrast would be ex- pected even for much finer detail. However, the x-ray intensity falls ofT very rapidly as the kilovoltage is lowered due to less efficient electron gun operation, less efficient x-ray generation, more absorption in the target and in the x-ray path to the photographic plate. The ideal solution is some form of soft x-ray image intensifier but until that time it is necessary to use the higher kilovoltage and overcome the electron limit in some other way. A thin layer of evaporated metal on a thicker beryllium backing will not give an improvement because x-rays are generated in the beryllium by the electrons that pass through the thin metal layer and this back- ground fogging obscures the improved resolu- tion. The alternative is to use a thin foil with no backing at all and this has been done for Figs. 7 and 8. In Fig. 7 the same type of test object as used previously by all three meth- ods of x-ray microscopy is shown at a higher magnification and with a resolution that is estimated to be close to 0.1 micron. A faint white Fresnel diffraction fringe of this width can be seen around the grid bars. An actual specimen as opposed to a test object is shown in Fig. 8. This is a bean section, also taken Fig. 6. Frozen dried new born mouse foot. (From V. M. Mosely and R. W. G. Wyckoff, un- published.) with a thin metal target and the same fringe can be seen around this specimen as well. These photographs represent about the best that can be done with the static x-ray projection microscope until some means of increasing x-ray intensity is found by any or all of the possible approaches including image intensifiers, field emission electron sources and better electron lenses corrected for spher- ical aberration. Other Methods of Point Source X-Ray Microscopy The difficulty of finding x-ray fringes was shown in Fig. 4 and the small size even at high resolution is seen in Figs. 7 and 8. This makes the practice of Gabor diffraction mi- croscopy with x-rays very difficult if not im- possible since it is necessary to have many fringes in the hologram if there is to be any information in the reconstruction. Baez and El-Sum show other difficulties as well, such as monochromatizing the x-ray beam, using opaque specimens and producing a mathe- matical point source. 651 X-RAY MICROSCOPY Fig. 7. Silver grid, 1500 mesh/inch, 3 micron bars, white Fresnel Fringe (From W. C. Nixon, Nature, 175, 1078, 1955.) Magnification given by 2 micron bar. Fig. 8. Bean section, high magnification using thin target. Magnification given by 10 micron bar. The X2 method of Le Poole and Ong is another way of getting a sHght gain in reso- lution if the film grain size and the x-ray source size are similar. The exact improve- ment is still arguable but the results they show are very good and the gold-shadowed and unshadowed bull spermatozoa taken in this way are superior to those obtained with the thin metal targets used for Figs. 7 and 8. Other projection methods are outlined by Rovinsky where in one case a pin-hole pro- duces the small x-ray source and in the other a fine tungsten point acts as the x-ray target. The resolution and exposure times are similar to the more standard methods and these variations on the basic theme may lead to improvements in the future, although not offering better resolution at present. Microanalysis X-ray microscopy in common with other methods of micro.scopy is intended to supply information on the microscale. With the electron microscope the scattering of elec- trons to produce the contrast in the image does not bear a simple relationship to the elements of the specimen. Selected area micro electron diffraction must be employed to identify the crystalline components of the object, from an area of one micron square or slightly less. A similar method can be used with the projection x-ray microscope where selected area x-ray diffraction can be used for areas down to a few microns in size. Lo- calization of the area is done from the x-ray image seen on the fluorescent screen; an aperture is brought into the beam so that a narrow pencil of rays strikes the specimen and the broader diffracted beam is recorded on fast x-ray film in place of the slower emul- sions used for microscopy. This method allows the identification of complicated crys- talline compounds containing many sepa- rate elements. In addition to this method, analogous to electron diffraction, it is also possible to per- form x-ray absorption and emission micro- analysis in a manner that has no parallel in electron microscopy. For the absorption method the techniques developed by Prof. Engstrom and his colleagues can be applied to projection x-ray microscopy with the ad- vantage of using an x-ray enlarged image for analysis. If this is recorded photographically the final accuracy with the present x-ray microscope resolution is similar to that of the contact method but it is more difficult to produce monochromatic radiation with the projection tube. The enlarged image does allow the use of radiation counter recording and this may speed up the whole process and make it more automatic. A more power- 652 PRODUCTION OF CONTINUOUS AiND CHARACTERISTIC X- RADIATION fill approach is to use the x-ray fluorescent emission from the specimen when the pri- mary x-ray beam has sufficient energy to ex- cite the characteristic hnes sought. This is the same method that is widely practiced on the macro scale for x-ray analysis of large amounts of material. The projection x-ray tube has a very high specific loading of the x-ray focal spot and this high brightness makes it specially suitable for micro-diffrac- tion, as mentioned above, and for micro- fluorescent analysis. In this case the pri- mary x-ray beam is collimated to a narrow beam and this strikes a very small area of the specimen. The emitted radiation is de- tected by a counter placed to one side of the main beam and for analysis a spectrometer is placed between the specimen and the counter. A similar but perhaps more powerful vari- ant of x-ray microanalysis is the use of the x-ray emission from the target itself. In this case the specimen replaces the normal x-ray target, is struck by the electron beam, and the emitted x-rays from an area equal to the size of the beam are analyzed by a spectrome- ter and counter. Point to point detection was achieved by mechanical scanning of the specimen in early models of this method with the position of the electron probe determined by means of an additional optical micro- scope focused on the same area. A recent more elegant approach is where the specimen is still moved for general viewing of the sur- face but the electron beam is scanned as well, as in a television raster, and the x-ray output modulates a cathode-ray tube that is scanned with the same speed. The resultant picture on the face of the tube represents the exact element analysis for each particular setting of the spectrometer detector. This method of scanning x-ray microanalysis seems to be the true strength of the scanning x-ray micro- scope and efforts are being made to extend the analysis to lower atomic numbers, lower concentrations and to determine the limit of detection. (See "Encyclopedia of Spectros- copy," pp. 745, 768.) These analytical tools use the same beam forming sj^stems as the static x-ray projec- tion microscope and the combination of x-ray microscopy and x-ray analysis although not at the resolution of the electron microscope, should become a well established technicjue of microscopical investigaion. REFERENCES Nixon, W. C, Research, 8, 473 (1955) (reprinted in "Modern Methods of Microscopj'", p. 92, Butterworths, London, 1956 and Interscience, N. Y.) Proceedings of the Cambridge Symposium "X-ray Microscopy and Microradiography", August 1956, Cambridge, England; editors, V. E. Cosslett, A. Engstrom and H. H. Pattee, Aca- demic Press, N. Y., 1957. Proceedings of the Stockholm Symposium "X-ray Microscopy and X-ray Microanalysis", June 1959, Stockholm, Sweden; editors, A. Eng- strom, H. H. Pattee and V. E. Cosslett, El- sevier Press, Amsterdam, 1960. Ong Sing Poen. "Microprojection with X-rays", Hoogland and Waltman, Delft, Netherlands, 1959. Cosslett, V. E. and Nixon, W. C, "X-ray ^Micros- copy", Cambridge Universitj- Press, 1960. W. C. Nixon PRODUCTION OF CONTINUOUS AND CHARACTERISTIC X-RADIATION FOR CONTACT AND PROJECTION MICRORADIOGRAPHY When discussing the production of charac- teristic or continuous x-radiation, by bom- barding a target with electrons, it is con- venient to separate the elementary processes of bremsstrahlung and inner-shell ionization from the modifying effects of electron scat- tering and energy loss. The so-called 'ideally thin' target is one in which only a small frac- tion of the electrons undergo any encounter (radiative or otherwise) within the target so that the electrons are undeviated and are of full energy. This ensures that any encounter resulting in the production of x-radiation takes place when the bombarding electron has a known energy and direction. Targets 653 X-KAY IVnCROSCOPY satisfying this condition would be imprac- ticably thin, but Von Borries (1949) and Lenz (1954) have calculated the distance in which an electron is expected to suffer 1 elastic collision*, and this is a convenient criterion for our purposes. At 10 kV it is ap- proximately 2 X 10~^ gm/cm^ for all ele- ments, rising to about 8 X 10~^ gm/cm^ at 40 kV. As the most probable angle of single scattering is only 0.01-0.1 radian, this cri- terion ensures that the electrons are usually deviated by only quite small amounts from their initial directions. The inelastic en- counters cause even smaller deviations (for example 1.4 X 10~^ radian in carbon at 20 kV) and so can be neglected in this connec- tion; and it can be readily shown (e.g., from the data of Lane and Zaffarano, 1954) that the amount of energy lost in this critical thickness if very small, being for example 0.04 kV for almninum, for an incident elec- tron energy of 10 keV. At 10 keV the full range is 0.26 mg/cm^ for aluminum, or 130 times the Aiifhellungsdicke. The range in- creases slowly with increasing atomic nmn- ber, and from the Bethe-Bloch expression (e.g. Paul and Steinwedel, 1955) can be shown to be greater for gold than for alu- minum by a factor of about 2. It is clear that work with thin targets is more fundamental and can be compared more readily with theory, whereas thick tar- gets are of more practical importance in that they provide maximum yields of x-rays. Thick targets can be subdivided into foils which are thin enough to transmit the x-radi- ation produced, as used in the projection x-ray microscope (e.g., Nixon (1955)), and the massive anticathode familiar in conven- tional x-ray tubes. We shall consider the production of the continuous and characteristic spectra in thin and thick targets, in the energy region of interest in x-ray microscopy. * Termed by them the 'Aufhellungsdicke', in connection with image formation in electron mi- croscopy. Continuous Spectrum from Thin Tar- gets Spectral Distribution. The spectral distribution from thin targets was first stud- ied by Nicholas (1929) using an electron energy of 45 keV and targets of aluminum and gold. Measurements were made at an- gles of 40, 90 and 140 degrees to the forward direction. In general the data showed that the intensity per unit frequency interval I^ was approximately constant up to the high frequency limit vo , and zero thereafter. A theoretical treatment by Kramers (1923) gave a similar relationship, of the form I,dv = —--- dv. This expression is not in general true, al- though it is a good approximation when con- sidering radiation emitted at right angles to the incident electron beam, using targets of high atomic number and electrons of low energy. The more rigorous quantum mechanical theory of the continuous spectrum is due to Sommerfeld and has been put into numerical form by Kirkpatrick and Wiedmann (1945) who gave theoretical spectral distributions for a wide range of conditions. According to their data, the shape of the spectral distribu- tion depends on the parameter Z^/V, and not upon either of these two variables sepa- rately, except at the extreme low frequency end of the spectrum, where the screening ef- fect of outer electrons, which reduces the x-ray intensity somewhat, depends on Z separately. The spectral shape depends markedly on the angle of observation. Fig. 1 (calculated from Kirkpatrick and Weid- mann's data) illustrates the spectral dis- tributions at 0° and 90°, and also averaged over all directions, for an electron energy of 10 keV, using targets for which Z = 13 (aluminum) and 58 (the highest available from their calculations, at this low energy). The simple expression of Kramers can be 654 PRODUCTION OF CONTINUOUS AND CHARACTERISTIC X- RADIATION 4 lO -SO 1 1 90°- \\ ^~~~~--— ____ay. \. - Z = I6 1 1 _ 1 -SO 40I0 J_____^ 1 1 90" 30 - ^^ -^ ay. ■ 20 - ^ \^ . O^ lO 2=58 ' 1 0-25 05 075 «/«. lO 0-2S 06 0-7S to Fig. 1. Calculated energy distribution of X-radiation from thin targets for an accelerating voltage of 10 kV. The units are ergs/steradian/unit frequency interval/electron/atom-per-cm^. (After Kirk- patrick and Wiedmann, 1945.) seen to be a useful approximation of 90°, al- though is not applicable in the forward di- rection. The intensity averaged over all angles, (but at the high energy limit only), is illus- trated in Fig. 2, as a function of ZV^', again after Kirkpatrick and Wiedmann. The sim- ple Kramer's theory predicts exact propor- tionahty; and this is indeed very nearly so, over a wide range of the parameters. Amrehn and Kuhlenkampff (1955) deter- mined the spectral distribution at 90° from thin targets of almninmn, nickel and tin. The effective thicknesses of their targets ranged from 4 X lO^^gm/cm^ (tin) to 1.2 X 10~^ gm/cm^ (aluminmn). By comparing this »-• •9 lOO / c 3 / 1 1 ,r / i / 1 / o o 3 3 lO 30 KDO XX>Z^ OO v/z'oo with the Aufhelhmgsdicke of 5-8 X 10-« gm/cm^ (for the electron energies of 25 to 40 keV used by these authors) it can be seen these represent a real approach to the condition of ideal thinness. It is perhaps more important to note that the energy loss in these targets would not exceed about 0.15 keV, a small fraction of the incident energy. The intensity from such targets is low, but by using a proportional counter and pulse height analyzer it was possible to obtain data of good statistical accuracy in reasona- ble times (30 seconds for each observation). Fig. 3 shows the spectral distribution ob- tained at right angles to the direction of the Fig. 2. Total intensity at the high energy limit as a function of ZVV (After Kirkpatrick and Wiedmann, 1945). The units are as in Fig. 1. electron beam, using an accelerating voltage of 34 kV, and a target of almninum. The corresponding theoretical curve is given (normalized to give the best fit with the ex- perunental data) and it is seen that the agree- ment in shape is very good indeed. When comparing aluminum and nickel, the total intensity was found to be closely proportional to Z^, (for equal atoms/cm^), although with tin the agreement was somewhat less good. Angular Distribution. The angular dis- tribution of the continuous spectrmn from thin targets has been investigated by several workers. The radiation from targets of alu- minum 6000 A in thickness was examined by Kulenkampff (1928) using electron energies from 16 to 38 keV. It was observed that the 655 X-RAY MICROSCOPY i-i-i-'-i-w-L.jji w 20 SO\q\I Fig. 3. Spectral distribution form a thin alu- minum target (accelerating voltage 34 kV) at 90° to the electron beam (Amrehn and Kuhlenkampff, 1955). Fig. 4. Angular distribution from a thin alu- minum target (accelerating voltage 34 kV) (Kerscher and Kuhlenkampff, 1955). intensity (i.e., energy flux per unit solid an- gle) in the forward direction is minimal and that the intensity passes through a maxi- mum at an angle dependent upon the elec- tron energy and the quantum energy. Bohm (1937-8) and Honerjager (1940) carried out more detailed measurements (in the same region of electron energy) on targets of alu- minum and magnesium and confirmed the essential aspects of the earlier investigations. The maximum occurs at smaller angles as the electron energy is increased; but for a fixed electron energy, it is the softer components of the spectrmn which have the smallest an- gle of maximum intensity. Bohm and Honer- jager observed additionally that, with de- creasing target thickness, the forward mini- mum became more marked, because of the decreasing effect of electron scattering and diffusion. It was noted that even at the high energy limit there was some emission in the forward direction but this was attributed to the ef- fect of (electron diffusion, even in the thin- nest targets used. The theory of Scherzer (1932) predicted that there would be no for- ward emission, l)ut this involved the assump- tion of small atomic numbers and high electron energies, and the more general cal- culations of Scheer and Zeitler (1955) show that there is expected to be some forward emission. The finite intensity observed by Honer- jager in the forward direction using his thin- nest (100 A) target affords evidence of this, because this thickness is only one half of the "Aufhellungsdicke" at this energy. Electron scattering would not be expected to affect the original distribution to any observable extent. Kerscher and Kuhlenkampff (1955) have used targets 250 A in thickness, and have established accurately the shape of the angu- lar distributions for several different photon energies within the continuous spectrum ex- cited in aluminum at 34 kV (Fig. 4). These workers used a proportional counter and sin- gle channel pulse anal^^zer to select particu- lar bands of photon energies. This method suffers from the disadvantage that the boundaries of the selected channels are ill- defined (because of the finite 'spread' in the heights of pulses produced by photons of a given energy) and Doffin and Kuhlenkampff (1957) have recently avoided this difficulty by using balanced filters, to select a well de- fined energy band. Massey and Burhop (1952) have given a comparison between the theoretical angle of maximum emission (at the high energy limit) and the experimental data of Kuhlenkampff and of Bohm. The agreement is very good, and the more recent data also fit well. X-ray Production in Electron -opaque Targets Spectral Distribution. When the x-ray target is thick enough to bring the electrons to rest, the observed spectral distribution is 656 PRODUCTION OF CONTINUOUS AND CHARACTERISTIC X-RADIATION a superimposition of thin target spectra for all electron energies up to the incident energy. The fomi of the thick target distribu- tion can be calculated only if the range- energy relationship is known. The scattering of the electrons causes a further modification of the spectral distribution in a given direc- tion, because the electrons in the beam are no longer travelhng in a defined direction with respect to the point of observation. The problem is not amenable to exact calculation but it is clear that the spectral distributions will contain a greater proportion of soft radiation than is observed from thin targets. Using the spectral distribution ly = const., together with the then accepted Thomson- Whiddington law of energy loss, Kramers obtained a thick target distribution of the form /, = KZ (vo — »') wliich is in fact very closely obeyed in practice, as will be shown below. The intensity per unit frequency in- terval is related to the intensit,y per unit energy interval, and to the somewhat more familiar intensity per unit wavelength in- terval, by the relations L = hL c The continuous spectrum from massive targets of several elements was investigated by Kuhlenkampff (1922), at an angle of ap- proximately 90° to the electron beam, using accelerating voltages between 7 and 12 kV. It was established that the Kramers rela- tionship was closely applicable down to about 0.4 j^o . A small additional term, independent of kilovoltage and photon energy but pro- portional to Z^, was found to be necessary. This second term is important only in the immediate vicinity of the high energy limit. These measurements w^ere extended to higher accelerating voltages (20-50 kV) by Kuhlen- kampff and Schmidt (1943), and to lower voltages (1-2 kV) by Neff (1951). Some of the latter data are reproduced in Fig. 5. The spectral distribution in the forward direction from electron-opaque targets con- Fig. 5. Spectral distribution from a massive target of platinum (Xeff, 1951). 2 4 6 8 10 Quantum Energy (kev) Fig. 6. Spectral distribution in the forward di- rection from an electron-opaque gold target (Dy- son, 1959a). sisting of foils in the region of 1 mg/cm^ in thickness has been investigated recently (Cosslett and Dyson (1957); Dyson (1959a)) 657 X-RAY MICROSCOPY Fig. 7. Angular distributions in the forward hemisphere from an electron-opaque aluminum target, for electron energies of 10.05 and 12.05 keV respectively. The curves in the upper quadrant represent observed data. Those in the lower quadrant have been corrected for self -absorption within the target. (Cosslett and Dyson, 1957). in the region of 6-12 kV. The data resemble very closely that obtained by Kuhlenkampff at 90°. Fig. 6 shows the spectral distribution in the forward direction from a gold target for four different accelerating voltages. Angular Distribution. The angular dis- tribution of the total radiation from an elec- tron-opaque target of gold was measured at 10 and 25 kV by Oosterkamp and Proper (reported by Botden et al. (1952)). At the higher electron energy there was a slight rise in intensity at increasing angles to a maxi- mimi at 15° from the forward direction, showing that electron scattering does not quite obliterate the type of distribution ob- served with electron-transparent films. At 10 kV there was no rise (i.e., no minimmn in the forward direction) , and it was considered that this was due to absorption in the target, the effect of which w^ould increase with in- creasing angle. Their measurements were made with an ionization chamber, for dosi- metric purposes; this would in fact accentu- ate the lower photon energies which show little anisotropy even when an electron-trans- parent target is used. In the work just referred to no attempt was made to discriminate between the dif- ferent energies within the continuous spec- triun, but a series of measurements using targets of aluminum, copper and gold has been described by Cosslett and the writer (1957, 1959a), in which a proportional counter was used for energy discrimination. Near the high energy limit the radiation is pronouncedly anistropic, although much less so than in the case of ideally thin targets. For example, for aluminmn at 10 kV, at a quantum energy of 9 keV, the anisotropy (defined as the intensity per unit solid angle in the direction of maximum emission di- vided by that in the forward direction) is calculated to be 4.1, for an ideally thin tar- get, whereas the "thick target" value ob- served in these measurements was 1.6 (see Fig. 7). At lower quantimi energies the anisotropy is even less and below about 0.5 of the high energy limit the radiation is vir- tually isotropic. Comparison with theory is difficult, but in principle it is possible to deduce the ex- tent of electron scattering by comparing the 658 PRODUCTION OF CONTINUOUS AND CHARACTERISTIC X-RADIATION anisotropies for thick and thin targets. It ap- mann's data. The agreement is in this case pears in fact that multiple scattering and dif- within the experimental error, fusion prevail even when the electrons have More recently the theory has been sub- lost only a small fraction (one tenth in this jected to a relatively comprehensive test by example) of their initial energy.- This implies the work of Amrehn and Kuhlenkampff that virtually all the x-ray output from a (1955) and Amrehn (1956). The latter au- thick almninum target for an incident elec- thor has made a detailed absolute com- tron energy of 10 keV is produced from elec- parison between his experimental energy trons Avhich are fully diffused. "Full diffu- distributions for five elements and the calcu- sion" and its relation to multiple scattering lationsof Kirkpatrickand Wiedmann. Agree- is discussed by Bothe (1932) and Paul and ment is often better than 10% and nearly Steinwedel (1955). always better than 20%. The general im- pression is that the theory is now well veri- Efficiency of Production of Continuous ^^^ ^^^ bombardment energies in the region X -radiation ^f 20-40 KeV. The main source of theoretical information Kirkpatrick and Wiedmann extended their regarding the efficiency of production in calculations to give an expression for the thin targets is the data of Kirkpatrick and thin target efficiency of production of x-radi- Wiedmann (1945), and for comparison some ation, by integrating their energy distribu- experimental data are available. tion curves over all energies and directions, Massey and Burhop (1952) have con- and by comparing this with the energy loss sidered the data of Smick and Kirkpatrick suffered by the bombarding electrons. The (1941) and Clark and Kelly (1941), con- thin target efficiency is given approximately verting it into absolute units where necessary, by and drawing upon Kirkpatrick and Wied- = 2 8 X IQ-^ZV (V in volts) mann's data in preference to the earlier cal- culations used by the experimenters. Smick although the calculations show that it is not and Kirkpatrick measured the radiation from strictly a function of ZV, but depends upon a thin nickel target bombarded by 15 keV these two variables separately to some small electrons, using balanced filters to select a extent. band of radiation of quantmn energy about Turning now to thick targets, the experi- 8 keV or 0.575 of the high energy limit. At "^^ntal data were reviewed by Compton and 88 degrees to the forward direction the in- ^^"^^^^^ ^1935), and the expression tensity was 2.2 X 10"^" ergs/steradian/unit ,, = 1.1 X lO'^ZV (Fin volts) frequency interval/electron/atom per cm^ ^^^^^ eonsidered to be accurate to about 20 %. The data of Kirkpatrick and Wiedmann yield ^^.^ eonfirmed the calculation of Ivramers a value of 6 X 10-^« in the above units, al- ^^^ obtained the result though Smick and Kirkpatrick give a calcu- lated value (after Sauter (1934)) of 2.9 X ' v = 0.92 x KT^ZV 10-^°. Clark and Kelly made a similar meas- ^^.^^ j^g theoretical expression for the energy urement using alimiinum at a bombarding distribution from thick targets. Kirkpatrick energy of 31.7 kV and a quantum energy of and Wiedmann calculated the efficiency by 0.82 of the high energy limit; at 60 degrees integrating their thin target data over all they found the intensity to be 6.2 X 10~^^ angles and energies and found the efficiency ergs, etc., (±33%), as compared with to be proportional to voltage and atomic 4.23 X 10- *i from Kirkpatrick and Wied- number to quite a high degree of accuracy, 659 X-RAY MICROSCOPY with a constant of proportionality of 1.3 X 10-^ The absohite intensitj^ of production in thick targets in the forward direction has been measured by the writer (1959a) for tar- gets of aliuninuni, copper and gold. The measurements extend from the high energy limit eo down to about 5.4 eo • From these data the absolute efficiency can be estimated by extrapolation to zero energy and integration over the whole energy range. Such an ex- trapolation is subject to some uncertainty but within this limitation the efficiencies in the forward direction were found to be ac- curately proportional to the accelerating voltage, and approximately proportional to the atomic number. The constant of propor- tionaUty was found to be 1.3 X 10^ which is slightly in excess of Compton and Allison's estimate but in good agreement with the calculations of Kirkpatrick and Wiedmann. Characteristic Radiation When discussing the production of charac- teristic X - radiation by electron bombard- ment it is convenient to define a cross-section Qk,l • • ' ior K ,L- ■ ■ ionization and to ex- press it as a function of the electron energy and of Vk,l--- the K,L-- excitation energy. The problem of finding a suitable expression has been discussed by Worthing- ton and Tomfin (1956) who express Qk in the form 0.77re2 - log. [4t7/(1.65 -f 2.65 exp (1 - U))], where U = V /Vk and V is the electron energy. The essential aspects of this expression are the inverse relation between Qk and V k^ (implying a strong inverse relation between Qk and Z), and a rapid rise in Q^c as C/ is increased from 1, to a maximum value at U = approximate^ 2.5, followed by a slow fall in Qk as U is further increased. Q kV k^ is the same function of U for all elements, and is illustrated in their paper. The actual out- put of x-radiation is obtained by multiplying Q K by w K, the fluorescence yield, and by ad hoc additional terms depending upon geome- try, self-absorption within the target, etc. Little absolute data are available for thin targets but the existing information for silver and nickel targets has been summarized and discussed by Worthington and Tomlin (1956), following Massey and Burhop (1952). The experimental and theoretical results do not fit well, although the curves agree in general shape. The output from thick targets has been M calculated by Worthington and Tomlin, but direct experimental observation by these authors and by the writer (1959b) on copper, and by Dolby (1960) on aluminum suggest that their calculations yield values which are high by factors of approximately 2 and 4, respectively. The experimental data and the theoretical expressions show that the efficiency and the intensity vary with kilovoltage in a some- what complex manner. For the intensity, empirical relations of the form / = kiVo Vk)' have frequently been observed to be of use over a restricted range of the variables, where q equals, for example, 1.6 (Worthing- ton and Tomlin 1956) or 1.65 (Compton and AlUson 1935). But formulas of wide appli- cability and methods of calculating the ab- solute yields on a systematic basis have yet to be found. REFERENCES 1. Amrehn, H., Z. Phys., 144, 529 (1956). 2. Amrehn, H., and Kuhlenkampff, H., Z. Phys., 140, 452 (1955). 3. BoHM, K., Physik. Z., 38, 334 (1937). 4. BoHM, K., Ann. Physik, 33, 315 (1938). 5. BoRRiES, B. v., "Die Ubermikroscopie", Saen- ger: Berlin (1949). 6. BOTDEN, p. J. M., COMBEE, B., AND HOUT- MAN, J., Philips Tech. Rev., 14, 165 (1952). 660 PROJECTION MICROSCOPY 7. BoTHE, W., Handb. Phys. XXII, 2, Springer: Berlin 2nd Edit. 1932. 8. Clark, J. C, and Kelly, H. R., Phys. Rev., 59, 220 (1941). 9. CoMPTON, A. H., AND Allison, S. K., "X-rays in theory and experiment", Van Nostrand, New York, 1935. 10. Cosslett, V. E., (1959) Unpublished. 11. Cosslett, V. E., and Dyson, N. A., contribu- tion to "X-Ray Microscopy and Microradi- ography" Academic Press, New York, 1957. 12. DoFFiN, H., and Kuhlenkampff, H., Z. Phys.,U8,49Q (1957). 13. Dolby, R. M., Brit. J. Appl. Phys. (1960). (in press). 14. Dyson, N. k.,Proc. Phys. Soc.,73, 924 (1959a). 15. Dyson, N. A., Brit. J. App. Phys., 10, 505 (1959b). 16. Honerjager, R., Ann. der Phys., 38, 33 (1940). 17. Kerscher, R. and Kuhlenkampff. H., Z. F/iys., 140, 632 (1955). 18. KiRKPATRICK, P., AND WiEDMANN, L., PhyS. Rev., 67, 321 (1945). 19. Kramers, H. A., Phil. Mag., 46, 836 (1923). 20. Kuhlenkampff, H., Ann. Physik, 69, 548 (1922). 21. Kuhlenkampff, H., Ann. Physik, 87, 597 (1928). 22. Kulenkampff, H., and Schmidt, L., Ann. Physik, ^Z, 494 (1943). 23. Lane, R. O. and Zaffarano, D. J., Phys. Rev.,M, 960 (1954). 24. Lenz, F., Zeit. Naturforsch., 9a, 185 (1954). 25. Massey, H. S. W., and Burhop, E. H. S., "Electronic and Ionic Impact Phenomena" Oxford, 1952. 26. Nepf, H., Zeit. Phys., 131, 1 (1951). 27. Nicholas, W. W., Bur. of Stand. J. of Res., 2, 837 (1929). 28. Nixon, W. C, Proc. Roy. Sac, A232, 475 (1955). 29. Paul, W., and Steinwedel, H., (Article in "/3- and 7-ray Spectroscopy", ed. K. Siegbahn, North Holland Publishing Co. Amsterdam), 1955. 30. Sauter, F., Ann. d. Physik., 20, 404 (1934). 31. ScHEER, M., AND Zeitler, E., Z. Phys., 140, 642 (1955). 32. ScHERZER, O., Ann. Physik, 13, 137 (1932). 33. Smick, a. E., and Kirkpatrick, P., Phys. Rev., 60, 162 (1941). 34. Worthington, C. R., and Tomlin, S. J., Proc. Phys. Soc, 69A, 401 (1956). N. A. Dyson PROJECTION MICROSCOPY (See a/so p. 647) The idea of using a small x-ray source for projecting enlarged images was mentioned by Sievert (1) in 1936. In order to limit the efifective emitting area, the use of a pinhole aperture was proposed. This method usually called the camera obscura or pinhole camera, was re-introduced by Avdeyenko, Lutsau and Rovinsky (2) at the Cambridge Sym- posium on x-ray microscopy and micro- analyses. It is equally suitable for both trans- mission and emission microscopy. In 1939, von Ardenne (3) proposed use of an electron optical system to make an x-ray source of very small dimension and high specific load. This type of microscope (which was successfully realized by Cosslett and Nixon (4) in 1951) is usually called the pro- jection microscope. Since 1951, the problems of obtaining better resolution and more reliable instru- mentation have been studied by Cosslett, Nixon and Pearson (5) (England), Le Poole and Ong (6) (Holland), Newberry and Summers (7) (USA) and Bessen (8) (USA), among others. The Projection Microscope Principle. The projection microscope is a microfocus x-ray tube, consisting essentially of an electron source, an electron optical system, and a transmission type target (Fig. 1). The electrons w^hich have an energy of some 5-15 keV are focused onto a 1-0.1 n spot. In most cases the target acts as a vac- uimi seal, allowing the specimen to remain in air. The electron optical system consists of a very, strong lens, usually called the ob- jective, and a weak one, the condenser. The function of the latter is to regulate the rate of demagnification of the lens system. As the current density of the electron spot largely depends on its size, the use of the condenser is indispensable for optimum working conditions. Due to the short life- time of both target and cathode filament, 661 X-RAY MICROSCOPY electrons electron source Fig. 1. Principle of the projection microscope. Fig. 2. 1500 mesh per inch silver grid, demon- strating the large depth of field. (Experimental Delft microscope.) these elements are replaceable. Thus, the projection microscope is a demountable sys- tem continuously evacuated to maintain the required vacumn. Properties. The advantage of using x- rays for microscopy is discussed elsewhere in this volume. This section will deal primarily with the specific aspects of the projection microscope as compared with the widely used contact method. (See B. L. Henke, p. 675.) These are: (1) The resolution is not limited by the resolution of either the recording material or the optical microscope. As the initial mag- nification can be adapted to the recording material, the choice of the film is more a matter of working convenience. Usually such a fihn is chosen to allow some 10 X optical magnification. For printing at this magnifi- cation a normal enlarger may be used. Le Poole and Ong (9) show that the use of ultra-fine film (of the Lippmann type) for projection microscopy has some advantages. (2) The resolution is determined by the source size and the diffraction phenomena exclusively. In the region where diffraction does not affect the resolution, the unsharp- ness is the product of source size and mag- nification. As a result, the depth of field is very large. (For diffraction error, see B. L. Henke, p. 677.) The magnification varies with the source to specimen distance result- ing in a perfect perspective of the specimen. See figs. 2 and 3. This fact is particularly im- portant for stereo microscopy. (3) The specimen and film are spatially separated. This fact allows us to: (a) study specimens at high temperature; (b) study Fig. 3. Plastic sponge, ca. 50X. (Experimental Delft microscope.) 662 I'KOJKCTION MICROSCOPY specimens which fluoresce under x-ray radia- tion ; (c) ehminate the effect of electron emis- sion and x-ray fluorescent radiation on the recording material when using hard radia- tion; (d) study specimens which may influ- ence the properties of the photographic emulsion. The specimen lies very close to the target of the projection microscope and great care must be taken to avoid reaction between specimen and target. The author notices, for example, that a Mercurichrome-stained specimen may give trouble, if it is used in connection with an aluminum target. (4) To reduce spherical and chromatic errors, the lens must have a short focal length (10, 11). As both the specimen and target lie almost in the focal plane of the lens, the specimen is in the active region of the lens field. If the specimen consists of ferromagnetic material it may introduce lens errors and change the focal condition. If the specimen is symmetrically aligned with re- spect to the optical axis and focusing is per- formed with the specimen in position, these errors can be minimized. Experiments, per- formed with the commercial T.P.D. micro- scope show that thin magnetic tape does not appreciably change the focal condition. Limitations: The Intensity Problem. As a microscope is aimed to reveal small details, it must have a good resolution. The resolu- tion, however, is among other things propor- tional to the image contrast (12), so the use of long wavelength radiation is imperative in connection with low absorbing specimen. Both the source size and the anode voltage are limited in their lowest values. This is caused by the relatively low specific emission of the cathode, resulting in a very low cur- rent density of the electron spot at the tar- get. Although it may be possible in principle to obtain an electron focus of less than 100 A with a 1 kV anode voltage, this is not practicable because of the low intensity. The current density of the electron spot is directly proportional to the brightness of the electron source and the square of the aperture of the objective lens. The latter is determined by lens errors, especially the spherical aberration. Calculations show that for white radiation, the x-ray energy flux at screen level is: p, a JVH»i^C,-^'%-^ (1) in which ./ is the specific emission of the cathode; V, the anode voltage; d, the diam- eter of the electron spot; Cs , the spherical aberration constant, and h, the target to screen distance. Focusing of the electron spot. The energy flux px determines the brightness of the fluorescent image and the exposure time of the film. At the expense of this figure, either the anode voltage V or the source size d can be decreased. The limit will be determined by the impossibility of visual focusing on the fluorescent image rather than by excessively long exposure. An image intensifier will not give considerable improvement as the con- trast and thus the visibility of detail will be limited by the number of x-ray cjuanta used for building an image element. This niunber is proportional among other things to the x-ray energy flux at screen level and the storage time of the eye or screen. By de- creasing the target-to-screen distance b, the brightness of the fluorescent image can be increased considerably at the expense of the field during focusing. The limited field of view is not serious in this case; as in fact, for focusing, one image element is sufficient. The condition that the magnification on the screen must be sufficiently high to insure proper focusing sets a limit on the smallest screen-to-target distance. A better approach to the "focusing on one image element" method is realized by Ong and Le Poole (14). Instead of x-raySj they used the elec- trons which are reflected from the target. The reflection coefficient of the order of 2 to 5% is greater than the x-ray efficiency. The electrons which have the same energy as the incident ones pass the lenses in oppo- 663 X-RAY MICROSCOPY ElECTBOSIAliC 00^- MAGNEIIC LENS ~ TRANSVERSE MAGNETIC Field PRIMARY IMAGE PRIMARY ELECT ELASTICALLV REF ELECTRONS SECONDARY IMA ELECTRON GUN MAGNETIC LENS Fig. 4. Principle of the focusing method, using reflected electrons. of the x-ray fluorescent screen under normal focusing condition. So, when using this focusing method, the Hmit will be set by insufficient stability during long exposure. An important advantage of this focusing method is that it works with the real speci- men in position and that the focus can be checked during exposure. The resolution, obtained by using this focusing method is demonstrated in figs. 5 and 6. Stability. While the problem of focusing is solved, the intensity problem still remains. Thus long exposure times are required and during this time, the microscope must be stable. The stability concerns: (a) Electrical stability. Changes in both anode voltage and lens current cause a broadening of the electron spot. If the elec- tron optical system is not well aligned, there will be an associated image shift. Although electrical stability can be achieved to a very high degree for a short time, it becomes more diflficult to obtain the same stability over long periods. The use of electrostatic lenses will not solve this problem because they give more spherical aberration, resulting in a Fig. 5. Gold shadowed bull sperms, ca 2000X. (Experimental Delft microscope.) site direction (see Fig. 4) and give an en- larged electron image of the spot. For sym- metry reasons it can easily be recognized that this image lies in the electron source and has the same size and shape when the target is in focus. By using a transverse magnetic field or by slightly tilting the lens, if it is of the magnetic type, this secondary image can be separated from the primary electron path and caught on a fluorescent screen. Calculations (14) show that the brightness of this image is some 10* higher than that perimental Delft microscope.) Fig. 6. 1500 mesh per inch silver grid, showing the resolution of the projection microscope. (Ex- 664 PROJECTION MICROSCOPY lower x-ray intensity and thus necessitating heat dissipation in the target is not a prob- a longer exposure time. And the combined lem (15). As the source size becomes smaller, electron optical and thermal effect is more the heat flow becomes more favorable. The serious. admissible specific load may be increased (b) Thermal stability. The electron spot on proportionally with the reciprocal value of the target is not the image of the cathode, the spot diameter. So, in a spot size of 1 n, but that of the crossover which has a more a load as high as 10^ W/cm^ can be tolerated, uniform intensity distribution and a smaller However, with the conventional hot cathode size. The apparent position of this crossover this is hardly realizable when low voltages depends greatly on the geometry of the are used. cathode, Wehnelt cylinder and anode as these elements act as an immersion lens. Description of a Commercial Projection Temperature changes effect the optical X-ray Microscope properties and thus the position of the cross- In comparison with the contact tube, the over. This effect is especially serious when it projection x-ray unit is a rather compHcated is accompanied with a disalignment of the electron optical and mechanical device. First immersion lens. Bearing in mind that the it must be continuously pumped to maintain cathode is at a high temperature and the the reciuired vacuum. (For ultrasoft x-ray surroundings have a relatively large mass, work, the contact tube cannot be made as a the temperature may still change consider- sealed off tube either.) Furthermore, it must ably during an exposure of more than 20 contain a highly stabilized anode voltage minutes. For high-resolution work, refocus- and lens current supply and an accurately ing before each exposure is necessary, even machined specimen and target stage. As the after several hours of continuous use. Pre- requirements are similar to those of an elec- heating the surrounding parts may reduce tron microscope some manufacturers provide this effect. A more effective solution will be adaptors which allow the use of the electron to use an illuminated pinhole aperture as microscope as an x-ray microscope. It is the electron source. This will, however, need encouraging to notice that complete com- one extra lens. mercial projection units are also available (c) Mechanical stability. Mechanical sta- now. bility may be achieved by appropriately Recently the Electron Microscope Divi- constructing and mounting the different sion of the Technical Physics Department, parts. The most serious error results from T.N.O. and T.H. at Delft, Holland, intro- relative movement between x-ray source duced its commercial microscope, which has and specimen during exposure, therefore, some special features. This instrument, a these parts need special attention. picture of which is shown in Fig. 7 and a Other limitations. The resolution limita- simplified cross section in Fig. 8, will briefly tion due to diffraction is discussed by Henke be described. The principal parts, i.e., the in this volume (p. 677). Another limitation electron source, electron lenses, target and of the resolution may be due to the depth of specimen stage and the camera may easily penetration and diffusion of the electron in be recognized. The electron reflection focus- the target. This effect can be reduced either sing method, mentioned previously, is used by using a lower voltage or a very thin tar- here. A simple but effective mirror optic in get (13). The effect of carbon contamination combination with a concentric corrective proves to be negUgible when the specific load lens and a 5 X ocular are used for observing is high enough (14). the focusing screen. The total magnification In contrast to conventional x-ray tubes, of this viewer system is 25 X. 665 X-RAY MICROSCOPY Fig. 7. Commercial X-ray projection micro- scope, manufactured by the electron microscope division of the T(echmcal) P(hysics) D(epart- ment), T.N.O. and T.H., Delft, Holland. The magnetic objective has a focal length of about 1.8 mm. It consumes 90 W and is water-cooled. As the specimen and target are inserted sidewards, the gap between the pole pieces is made relatively large (ca. 7 mm). The specimen carrier can be moved independently in three directions, the mag- nification on the film is variable between 150 X and 10 X. The admissible optical mag- nification of the film is approximately 10 X . To realize the two extreme values of the mag- nification the specimen must be put very close to the target in the first case and almost at the upper pole piece in the second case. As a result the magnification cannot be varied continuously over the whole range. Without breaking the vacuum, the specimen holder can be turned over 180° after a small outward movement. For a magnification greater than 150X either the specimen must be fixed on the target carrier or the target fixed on the specimen carrier. To accomplish this the target holder can be completely moved outward. Furthermore, it can be moved in a plane perpendicular to the optical axis, thus allowing the use of a clean part of the target for each exposure (See Fig. 9). The target holder can contain up to four different materials in the form of thin, evaporated or rolled metal films. The x-ray fluorescent screen is fixed on an aplanat and vie\ved with a binocular (two separate objectives) viewer (IG). As focus- ing is carried out with the reflected elec- trons and the fluorescent image is inadequate for visual observing, this part of the micro- scope should be considered as a view-finder. Hence, the image brightness is more impor- tant than the magnification. Thus the former is increased at the expense of the latter. The total optical magnification of this finder system is some 20 X. The screen and aplanat can be moved away so that the specimen can be viewed with the binocular for prehmi- nary positioning. In contrast with other projection units, the specimen as well as the camera is in vacuum for the following reasons. By using either very high or very low voltages, the target thickness must be equal to or less than the resolution. This is necessary to re- duce the effect of electron diffusion in the target in the first case and to avoid too much absorption of the x-rays in the second case. Such a target is inadequate as a vacuum seal unless its area is very much restricted. Fur- thermore, the air between source and film would absorb too much soft radiation. A direct advantage of a non-sealing target is that it can be easily exchanged during operation and that a clean part can be used for each exposure. The restriction that only dry specimens can be used is not considered to be a serious one. The new instrument, however, is designed in such a way so as to allow the target and specimen stage to be exchanged for another one with a vacuum sealing target. The author has no informa- tion yet about the design and execution of this particular part. The camera is designed for 20 exposures on 35 mm film. The film is transported auto- matically by moving the camera back and forth. A shutter is provided, which opens in exposure position. For short exposure, the intermediate target, which lies between ob- jective and condenser may be used to inter- 666 PROJECTION MICROSCOPY Unocutar «-ray scfean spedmenheUer concave mirror correction lens plane mirror focusing screen electron gun lODrnm Fig. 8. Simplified cross section of the microscope shown in Fig. 7. rupt the electron beam. The image field is about 25 mm on the fihn. (Angular field approx. 36°.) The anode voltage is variable in four steps from 5 to 20 kV. In each position the lens current is automatically adjusted so that the lenses maintain the same strength. The desk model housing contains the me- chanical and diffusion pumps, the high volt- age and lens current supply, and the elec- tronic stabilizer. Practical X-ray Microscopy In general, the brightness of the fluores- cence images is so low that they are not suit- able for visual observation. Another conse- quence of this low intensity is the poor contrast. As the visibility of a detail will be limited by quantum noise, a considerable gain cannot be expected from an image inten- sifier unless it has a large storage time, pref- erably with a contrast correction. Admit- tedly, such a device would bring the screen brightness to a convenient level, eye adapta- tion not being necessary, but an image intensifier is a much too complicated instru- ment to use it for this purpose only. The use of fluorescent screens with better resolution in combination with a high-aperture micro- scope objective may do the same. 667 X-RAY MICROSCOPY Fig. 9. Close up of the target-, aperture- and specimen holder. The required storage time can easily be achieved by using a photographic fihii. In the wavelength region normally used for projection microscopy, the quantum yield, i.e., the number of developed silver grains per absorbed x-ray quantum, is very close to 1. This results in a linear relation between film density and exposure. Contrast is pro- portional to the density, so for full visual information transfer a contrast correction may be necessary (17). Although the depth of field of the projec- tion microscope is large, a thick specimen may obscure important detail and further- more it needs a longer exposure time. So the thickness will be determined by the fact whether we want to see the absorption dis- tribution in a plane or in a volume. The last one is made possible by stereomicroscopy. As the anode voltage and target material determine the spectral distribution of the x-ray source, they must be adapted to the specimen to give an adequate contrast. When using monochromatic radiations of different wavelengths a qualitative chemical analysis can be carried out (18). The fact that the specimen is spatially separated from both film and target allows us to make a series of exposures with different wave- lengths while still projecting the specimen at exactly the same angle. This fixed source- to-specimen position is necessary to get an unambiguous relation between the specimen and its projected image. A direct consequence of the large depth of field is the fact that the magnification of the image is not well defined. It may vary widely over the different parts of the speci- men, especially when the thickness of the specimen is comparable with its distance to the source (see Fig. 2). For thin specimens, the magnification can be determined in different ways similar to those used in light microscopy. A method to determine the magnification of the various parts of a thick specimen will be discussed in the next section about stereomicroscopy. As in projection x-ray microscopy the use of long wavelengths is very limited by the intensity, the contrast for thin organic sam- ples is very poor. If the special features of this type of microscope are desirable for such specimen, contrast can be improved by an appropriate treatment. In many cases this kind of preparation and staining technique can be a modification of existing techniques 668 I'KOJECTION MICROSCOPY used in light and electron microscopy and x-ray medical diagnostic. For example, I2 or Li-KI solutions will stain cellulose, and OSO4 will stain unsaturated fatty acids (see Fig. 10). BaS04-sugar mixture can he fed to small insects (19). Shadowcasting with a heavy metal will reveal surface structure (see Fig. 5 and Fig. 11). Replicating allows study of the coarse surface of thick heavily absorbing specimens. For relatively thick specimens such as fabric, (Fig. 12), w^ood particles (Fig. 13), small insects (Figs. 14, 15) etc., an anode voltage of 10-12 kV can successfully be used. The technique of selective staining, like fat particles by OSO4 , is still in its initial stage. Using different staining agents and radiating with x-rays of different wave- lengths may show' the distribution of the concentration of some chemical compounds in the specimen. The different pictures ob- tained in this way may be superimposed after suitable coloring, and thus a color pic- ture can be obtained. Fig. U. Tissue paper, gold shadowed, ca 63X. (Experimental Delft microscope.) Fig. 10. A section of mouse skin, stained with OSO4 . (Experimental Delft microscope.) Fig. 12. Nylon stocking. (T.P.D. microscope.) Stereomicroscopy As a result of the large depth of field, it is often very difficult to get an idea of the 669 X-RAY iMICKOSCOPY iW^ fffl- ,r^ Fig. 13. Wood splinters. (T.P.D. microscope.) Fig. 14. Ant. (T.P.D. microscope.) spatial distribution of the specimen out of a print. This is clearly illustrated in Fig. 3. The large depth of field, however, is espe- cially important for stereomicroscopy in which the specimen is seen in its right pro- portion, but on a larger scale. The geometri- cal conditions for this can easily be deducted from Fig. 16. Assume that the M times mag- nified specimen is placed at a distance / from the observer (Fig. 16a). Let the maximum angle of convergence be (p. Thus tan (p = e/2l, in which e is the interocular distance. Fig. 15. Pu.sterior end of a mosquito. b •cts tftpitcts Mm c point sources ■ \ / Fig. 16. Condition for making stereographs. 670 PROJECTION MICKOSCOPY In Fig. 16b, the specimen can be thought of as being replaced by two fihii images, ob- tained by projecting the various points with each of the eye pupils as a projecting center. Fig. 16c shows that the same images can also be the result of two equal objects which are M times smaller (and thus have the dimen- sion of the real specimen under study). As the two objects are exactly identical, the two film images are the projections of one object, with the projection center shifted over a distance As (Fig. 16d). Fig. 16e shows that instead of moving the projection center, the specimen itself may be shifted over a distance Aa = As. The geometrical condition for making stereographs ^\i\\ be: As = Sa = ae/l {2) in which a is the source to specimen distance. In general, the value of a cannot easily be measured, especially for high magnification. The source to film distance 6, however, is in most cases a constant of the instrument. So, introducing the magnification on the film, or primary magnification equation (2) can be transformed to il/pAs = MjAa = he/l (4) in which the second member is a constant. The antecedent is the image displacement on the film or screen. So to satisfy this con- dition, neither the magnification nor the specimen displacement need be known. If, however, Aa is known, by means of a cali- brated specimen shift, Mp can be calculated from the relative displacement on a stereo- scopic pair of negatives. As this method of measuring the magnification is the only gen- erally reUable one, all projection microscopes should be provided with a calibrated speci- men shift. The source-to-specimen distance is related to the magnification by the following equa- tion: a = l/M (5) Mp = b/a This relation gives the order of magnitude of a. As the viewing distance I is not very critical, it is not necessary to know a accu- rately. Note that in equation (5) M is the final magnification. Inserting I = 300 mm (3) and M = 1000 gives a = 0.3 mm. So this Fig. 17. Depth distortion in contact stereomicroscopy. 671 X-RAY MICROSCOPY condition cannot be fulfilled by the contact mefhod where a is larger by at least two orders of magnitude. If equation {5) is not fulfilled, a distortion in the depth dimension occurs. This can easily be shown in Fig. 13. Let ^ 0 , -Bo , Co , and Do be four points equally spaced along a line. Let, furthermore, S> be the projection center (x-ray source), F be the film and Ap , Bp , Cp , and Dp the pro- jected points, li AqDo is a reasonable fraction of ^o*S, (Fig. 17a), then ApBp 9^ BpCp ^ CpDp . If, however, AS » AqDo (See Fig. 13b), the rays A^Ao \\ BoB, \\ CoCc \\ DoDc thus AcBe = B,.Ce = CcD, . When after processing and magnifying, the film is viewed, the resulting points seem to come from Av , By , Cv , and Dy ; now AvBy 9^ ByCv 9^ CvDv . For depth measurements, however, the last method is preferable be- cause the ratio of the distances on the film and depth dimension of the specimen re- main the same. This results in better ac- curacy. The depth of field must be a reasonable fraction of the source-to-specimen distance. An object with a depth dimension of some 15 cm and placed with the shortest distance to the observer of 30 cm can still be sur- veyed. In terms of our considerations, this means that the depth of focus of the micro- scope must be a reasonable fraction of the distance a. For microscopes using an optical system, the focal length / should be used in- stead of a. So this condition cannot be ful- filled. The only microscopes where all these conditions can be met are the camera obscura and the projection microscope. The use of x-rays for stereoscopic examina- tion has many advantages over using ordi- nary light. Besides the possibility of observ- ing opaque parts, the negligible refraction, and reflection make any part of the specimen observable. (Compare the visibility of a specimen behind a ground glass screen.) (The pictures shown in this article were kindly supplied by the Electron Microscope Division of the Technical Physics Depart- ment, T.N.O. and T.H., Delft, Holland.) REFERENCES SiEVERT, R., Ada Rad., 17, 299 (1936). ROVINSKY, B. M., LUTSAU, V. G., AND AVDE- YENKO, A. I. "X-ray Microscopy and Micro- radiography," Academic Press, p. 269, New York, 1957. Ardenne, M. von, Naturwiss., 27, 485 (1939). COSSLETT, V. E., AND NiXON, W. C, Natl. Bur. Stand. Symposium (1951), Circular 527, p. 257. Cosslett, V. E., Nixon, W. C, and Pearson, H. E., "X-ray Microscopy and Microradi- ography," Academic Press, p. 96, New York, 1957. Le Poole, J. B. and Ong Sing Poen, Ibid., p. 91. Newberry, S. P. and Summers, S. E., Ibid., p. 116. Bessen, I. I., Philips Electronic Inc., Engi- neering report #66, (1958). Ong Sing Poen and Le Poole, J. B., Proc. 4th Int. Conf. Elec. Mic, Berlin, (1958). Liebmann, G. and Grad, E. M., Proc. Phys. Soc, B64, 56 (1951). Dorsten, a. C. van and Le Poole, J. B., Phil. Tech. Rev., 17, 47 (1955). Ong Sing Poen, "Microprojection with X- rays," p. 71, Martinus Nijhoff, The Hague, 1959. Nixon, W. C, Proc. Roy. Soc, A232, 475 (1955). Ong Sing Poen and Le Poole, J. B., Appl. Sci.Res. B, 7, 233 (1958). Cosslett, V. E., Proc. Phys. Soc. B, BXV 782 (1952). Ong Sing Poen, "Microprojection with X- rays," p. 31 Martinus Nijhoff, The Hague, 1959. Ong Sing Poen, Ibid., p. 74. Mosley, V. M. AND Wyckoff, W. G., J. Ultracelstructure Res., 1, 337 (1958). BoTDEN, p. J. M., COMBEE, B., AND HOUT- MAN, J., Phil. Tech. Rev., 14, 114 (1952). Ong Sing Poen REFLECTION MICROSCOPY (KIRKPATRICK) Since Rontgen's first unsuccessful experi- ments in attempting to concentrate x-rays by lenses and mirrors, many similar efforts 6 7 8 9 10 11 12, 13. 14. 15. 16. 17. 18. 19. 672 have been made and the faihires recorded in the hterature. It has been evident always that a successful x-ray microscopy would open up fields of investigation closed to opti- cal microscopy because of limited resolution and to electron microscopy because of the very limited penetration of electrons, thus necessitating extremely thin specimens. All such attempts until the one announced in September, 1948, were unsuccessful, so that focusing and image formation in a micro- scope were generally considered to be im- possible. Accepting the known fact of total reflection of x-rays from mirrors at extremely Fig. 1. Arrangement of reflecting mirrors. REFLECTION MICROSCOPY small grazing angles of incidence. Prof. Paul Kirkpatrick of Stanford University has taken a great step forward toward a success- ful solution of a seemingly impossible problem. A concave spherical mirror receiving x- rays at grazing incidence images a point into a line in accordance with a focal length / = Ri/2, w^here R is the radius of curvature and i the grazing angle. The image is subject to an aberration such that a ray reflected at the periphery of the mirror misses the focal point of central rays by a distance given by S = l.bMr^/R, where M is the magnification of the image and r is the radius of the mirror face. The possible resolving power is such as to resolve points separated by 70 A, inde- pendent of wavelength. Point images of points and, therefore, extended images of extended objects may be produced by caus- ing the x-rays to reflect from two concave mirrors in series, particularly when crossed at right angles to each other. Figure 1 gives a schematic idea of the arrangement of the mirrors of the apparatus, Fig. 2 reproduces photographs of the microscope, Fig. 3 is a reproduction of the photographed enlarged images from a 350-mesh screen. This figure shows, besides the full image from both mir- rors (upper left), partial images from each mirror separately (H, upper right, V, lower A B Fig. 2. Reflection microscope: (a) end showing x-ray tube (right) specimen, specimen holder and mirror mount; (b) film end showing window for observing fluorescent screen. 673 X-RAY MICROSCOPY ■ »*% ■ •!• 1 • •■ 2. J. Opt. Soc. Amer., 38, 7G6 (September, 1948); also a manual prepared by A. V. Baez. G. L. Clark Fig. 3. Enlarged images of 350-mesh screen in x-ray microscope. (Kirkpatrick.) Upper left, full image from two mirrors; upper right, partial image from horizontal mirror; lower left, partial image from vertical mirror; lower right, direct radiation. left, and the large spot, lower right, caused by direct radiation). While this development is still in its early stages both in theory and in practice, it is evident that there is every prospect of a successful x-ray microscope. Elliptical sur- faces already have been found to be superior to spherical or cylindrical ones in tests with mirrors made by coating a spherical mirror with a continuously variable amount of gold. Magnifications of 70 diameters already ob- tained are bound to be exceeded. The optical system is also suitable for focusing of very soft x-rays Avith wavelengths up to 45 A used for diffraction analysis of structures of materials with very large d spacings and microradiography of single cells (cf. article on Ultrasoft X-ray Microscopy, by B. Henke) ; and for focusing of neutrons. As the very newest development is the use of a sys- tem of cross-reflecting mirrors to focus solar x-rays in an X-Raij Telescope, as described by A. V. Baez in the Encyclopedia of Spec- troscopy (q.v.). REFERENCES 1. Clark, G. L., "Applied X-rays," 4th ed., Mc- Graw-Hill Book Company, N. Y., pp 105- 106, 1955. TWO-WAVE (BUERGER) MICROSCOPE In the final analysis of ultimate crystal structures in terms of the motif, or complete configuration of a molecule serving as a point which is translated according to a definite repeating plan, a Fourier series is summed up. This is the same process as the super- position of many sets of interference fringes in a microscopic image. Hence optics pro- vides several methods of Fourier synthesis in place of mathematical calculations. The ultimate extension of the optical- analogue method is the two-wavelength mi- croscope. Since x-rays cannot be focused as conveniently as visible light (cf. article on Reflection Microscopy (X-Rays)) it is pos- sible only to collect a diffraction pattern. But W. L. Bragg also conceived the idea that, if visible light could be made to con- tinue in the paths of the x-ray diffracted beams, it could be focused to give an image of the crystal. Such a two- wavelength microscope (Fig. 1) has been successfully constructed by Buerger at IMIT, starting with a diffraction pattern photographed in his precession camera (an undistorted recip- rocal-lattice photograph) which is equivalent to the interference pattern that would be formed by visible light and photographed by a lens. A replica of the x-ray pattern is made by boring holes in a brass plate and is placed in an optical system that produces an interference pattern which is the image of the original crystal. The most complex part of the apparatus involves the use of mica plates behind the holes in the brass plate capable of being tilted to produce phase shifts by varying the length of optical path. Individual rows of holes produce opti- cal patterns as lines at right angles to the rows similar to the Bragg-Huggins fringes. The total effect is optical simmiation of all 674 ULTRASOFT X-RAY MICROSCOPY Pinhole ~^~^~-~S~~ Fig. 1. The Buerger Two-Wave Microscope showing atomic arrangement of FeS2 , marcasite, on the screen. patterns to give the structure pattern. The result for marcasite, FeS2 , is shown in Fig. 1. Thus the good features of x-rays, with wavelengths short enough to be compatible with atomic dimensions, are combined with the scaling up by substitution of visible light of the whole optical image-forming process, so that the image is magnified by a factor of the ciuotient of the wavelength of visible light to that of x-rays, or 10^ diameters. REFERENCES 1. Buerger, M. J., J. A^-pl. Phijs., 21, 909 (1950). 2. Clark, G. L., "Applied X-rays," 4th ed., Mc- Graw-Hill Book Company, N. Y., pp. 460- 462, 1955. G. L. Clark ULTRASOFT X-RAY MICROSCOPY X-ray microscopy has essentially the same resolution limit as does light microscopy and it is usually the slower and less convenient of the two methods. Nevertheless there are important possible advantages of x-ray mi- croscopy which arise mainly through the basic differences in the optics of image for- mation and in the image contrast mecha- nism. The simple method utilized in the projec- tion of x-ray images is illustrated in Fig. 1. Here it is seen that all planes of the three- dimensional object are imaged, with equal sharpness, into an essentially two dimen- sional plane — usually a thin photographic emulsion which becomes the "microradio- gram." Such complete depth of field can only be approximately realized with light microscopy through the use of very small objective apertures and with a consequent sacrifice of resolution due to diffraction. Sharp projection images, relatively free of Fresnel diffraction blurring, are possible only with the shorter wavelength x-radiations. This two dimensional microradiogram "model" of the object is often more ame- nable to high resolution measurement of structural detail and to quantitative stereo- graphic analysis of thickness or depth di- mensions. Unlike light images, x-ray images have contrast which is due mainly to photo-elec- tric absorption of the x-radiation by the sample; the complicating effects of refrac- tion, diffraction and reflection are either small or completely negligible. This leads to a relatively unambiguous interpretation of contrast and often permits quantitative microabsorption analysis. X-ray absorption analysis complements that by light absorp- 675 X-RAY MICROSCOPY PROJECTION GEOMETRY J- — t Source of Ultrcsoft \ \ X-Radiation PHOTOMICROGRAPH OF ERYTHROCYTE (RANA PIPIENS) I d Sample Film 20 Microns PHOTOMICROGRAPH OF CONTACT MICRORADIOGRAM 0.5x1^ SLIT-MICROPHOTOMETER TRACING OF CONTACT MICRORADIOGRAM Fig. 1. Illustrating the point of projection geometry utilized in x-ray microscopy. Interpretation of the structure of the biological cell from the light microscope micrograph is made difficult because of such effects as refraction and a very short depth of field. Contrast in the x-ray microscope picture is simply and directly related to the chemistry and mass-thickness of the cell. tion since it measures the mass and ele- mentary chemistry of the sample only, and it is not sensitive to the effects of molecular combination of the elements as is light ab- sorption analysis. It is also of importance to note that a wide range of x-ray wavelengths are available for microscopy, e.g., 1 to 100 A, with an associated range of absorption co- efficients which vary by a factor of 100,000. A considerable amount of research has been reported since the time of the discovery of x-rays on contact microradiography with radiation in the 1 to 10 A region. In this kind of x-ray microscopy, conventional x-ray sources are used and the samples are placed in contact with the photographic emulsion. In 1951 Cosslett and Nixon introduced pro- jection microradiography in which the sam- ples are placed very near a point source of x-rays of micron dimensions and generated by precisely focusing a demagnified electron image of an electron source onto the target material using magnetic lenses (see paper by Ong Sing Poen, this volume). For nearly all of this work with the soft x-rays (1 to 10 A) the relatively thick samples required for contrast limits the useful magnification to about 300 X. However, for dense materials of the heavier elements, resolutions have been obtained with these radiations which have permitted magnifications of the order of 1000 X. Quantitative x-ray microscopic analysis for both the mass and the elemen- tary chemistry of biological materials was first introduced by Engstrom in 1946. He was also one of the first to recognize the feasibility of using the ultrasoft x-radiations for high resolution microradiographic analy- 676 ULTRASOFT X-RAY MICROSCOPY sis. References to most of the work accom- plished to date in x-ray microscopy may be found in the several that are listed at the end of this paper (1). The wavelengths in the 10 to 50 A region are of particular value for the quantitative analysis of micron-size systems of organic and light element inorganic composition — i.e. for elements up through Ge 32. This follows from the fact (to be established in a later section) that at least 60% absorption is required for precisions in the one to five percent region in the measurement of the parameter nm, where m is the sample mass- per-unit-area "thickness" and m is the mass absorption coefficient as defined by the ex- pression for the ratio of transmitted to inci- dent monochromatic x-radiation 10 SAMPLE THICKNESS FOR 37% TRANSMISSION t = I/Io = e- (1) It is shown in Fig. 2 that in order to have 60 % absorption for either light element in- organic or organic materials with thicknesses of the order of one micron, wavelengths in the 10 to 50 A region should be employed. The discontinuities in these curves are due to the presence of critical K or L absorption edges. Ultrasoft radiations are of very great value for the microanalysis of this highly im- portant problem area of the lighter element samples not only because optimum absorp- tion signal may be gained, but also because either the K or the L absorption edges for the elements of atomic number 6 to 32 lie in this 10 to 50 A x-ray region. These absorp- tion edges form the bases for sensitive, differential absorption analysis. In the sections that follow, analysis and experimental results are presented in order to illustrate optimum methods and instru- mentation and application of high resolu- tion, ultrasoft x-ray microscopy for quantita- tive analysis. Resolution Limit for X-Ray Microscopy As stated above, the first requirement for resolving microscopic detail of fractional 5 10 50 A -ANGSTROMS Fig. 2. Indicating the necessity for using the ultrasoft wavelength in order to gain the sufficient absorption in micron-size systems for quantitative measurement. micron dimensions is sufficient contrast for detection and measurement. This must be met, for samples comprised of elements from the lower half of the periodic table, by the use of the ultrasoft wavelengths (10 to 100 A). It is important to note, however, that as the wavelength is increased in order to gain contrast, the diffraction blurring of the image also increases. The resolution limit is reached when this diffraction error becomes com- parable wdth the object dimensions which are required for minmiimi absorption con- trast. The characteristics of the diffraction error for an image formed by monochromatic radiation from a point source is illustrated in Fig. 3. The effect of having a finite source size and perhaps polychromatic radiation is simply the superposition and addition of many such intensity patterns so as to blur the edge of the image. The minimum error may be measured by the distance, /, from 677 X-RAY MICROSCOPY (A) (B) DIFFRACTION ERROR / = (XL)>'2x'/V(l + x) microns (2) PROJECTED SOURCE DIAMETER MONOCHROMATIC POINT SOURCE Fig. 3. Definingthe two major causes for image unsharpness in ultrasoft x-ray microscopy due to (a) finite source size and (b) Fresnel diffraction. the ideal image edge position, formed with- out diffraction, to the position of the first intensity maximum. This may be shown from diffraction theory to be given by the relation with L measured in cm and X in angstroms. L is the source to image distance and x is the ratio of sample-to-image distance to sample-to-source distance, d/c. It is evident from this relation, as plotted in Fig. 4, that the diffraction error is maximum for x equal to unity and primary magnification eciual to 2X (71/ = 1 + x), assmning the camera length, L, to be constant. As has been (2) pointed out, the diffraction limit for 2X magnification is actually 70% of that for high magnification projection microscopy for dimension c held constant, i.e., for con- stant sample field. In order to reduce this error one must use either unit primary mag- nification (contact method) or high primary magnification. For contact microradiography Eq. (2) may be written as / = (Xd)i/2 (3) and for projection microradiography (x » 1) CAMERA EFFICIENCY -S. AND DIFFRACTION ERROR - f 0.50 0.25 0.00 CMR* Fig. 4. Illustrating the very high efficiency and relative freedom from P'resnel diffraction error for contact microradiography. 678 ULTKASOFT X-KAY MICROSCOPY In order to compare the two methods, let us consider the error associated with an ob- ject of one micron dimensions and of compo- sition (organic or of the hghter elements) requiring typically about 25 'A for measur- able contrast. P'or contact microradiography the effective emulsion thickness is a fraction of a micron so that d is of the order of one micron and the error becomes from (3) about 0.05 micron. This is appreciably less than the diffraction error resulting from the light microscope used to measure the micro- radiogram which, for high resolution meas- urement, is about 0.2 n and which is conse- quently the limit of resolution for contact microradiography. To minimize diffraction error in projection microscopy the sample must be placed as close to the x-ray point source as is practical. For this distance equal to 0.1 mm the error is ec^ual to 0.5 ju and for the sample to source distance eciual to 25 microns or 0.001", the error becomes 0.25 M- If should be noted that even if the wavelength could be reduced to one fourth (6A) as permitted by heavy element sam- ples, this error is reduced by only one hah. It is therefore concluded that projection mi- croscopy has a resolution limit comparable to that of light microscopy and contact micro- radiography. Relative Efficiency of X-Ray Microscope Methods At present, the most serious practical limi- tation on x-ra}' microscopy is the very low efficiency of production of the required low- voltage x-radiations. With three or four kilo- volts and maximum beam current, present day projection microscopes yield barely enough intensit}^ for the required fluorescent screen focusing of the point source as gen- erated by magnetic lens demagnification of a point source of electrons. Exposure times are often longer than can be tolerated because of the instability of the focal spot position. To date no projection microscopes have been developed for the ultrasoft wavelengths. There is a possibility of successfully meeting the intensity problem by operating the pres- ent projection microscopes at normally high voltages with transmission targets of such thickness and composition that a sufficient amount of the desired ultrasoft radiation is generated along with the shorter wave- lengths. The harder radiations would permit precise adjustment for high resolution work and might then be rejected in the microra- diographic measurement by a recording ma- terial sensitive only to the ultrasoft compo- nent, e.g., the usual, very thin concentrated Lippmann emulsion mounted upon a thin plastic film, the combination being trans- parent to the harder radiation. The relative efficiency of the two meth- ods, defined as the reciprocal of the exposure time required per unit sample area, may be estabhshed as follows: The camera speed, 8, defined as the recip- rocal of the time rec^uired to record a given sample, is equal to the product of io , the radiant flux per unit solid angle and per unit projected source area; of 12, the solid angle of the x-radiation which is utilized; of the projected source area, which is proportional to the source diameter, w, squared; of the sensitivity of the recording material which is assmned proportional to the square of its resolution error, A^ ; and divided by the area of the irradiated recording material, M-A, where M is the primary magnification and A is the sample area. Thus, S = {5) For an optimum condition of operation, the penmubra error. A, is equated to the record- ing error, A^ . Also, from Fig. 3 it is noted that A = wx = Mb = (1 -t- x)d (6) where x = d/c and 8 is the error. A, as meas- ured at the sample. The maxim lun camera speed may then be written as S = Kio5'{il/A){l + 1/x)^. (7) 679 X-RAY MICROSCOPY Two important facts are immediately cvi- resolution limit of both contact and projec- dent from this result: (1) In order to gain tion microscopy are essentially the same and maximimi speed and because of the depend- equal to that of light microscopy, viz. about ence upon 8*, it is very essential that the 0.2 ju- The present practical limit of resolu- recording material and the source size be tion of both methods is dependent somewhat "matched" to the particular resolution prob- upon the resolving power of present day con- lem as required by Eq. (6). (2) The micro- centrated Lippmann emulsions and other re- scope efficiency, which is equal to SA, is cording materials for contact microradiog- very much greater for the contact method raphy and upon the intensity problem and than for the projection method as illustrated consequent instability problem for high-res- in Fig. 4. olution projection microscopy. For dense, For high-resolution projection microscopy heavy element samples, and with shorter (x ^ 1) the camera speed, S, is equal to wavelength x-rays, resolutions of about 0.1 ju Kiod^/d^ and the efficiency, SA, is a constant might be achieved with the projection equal to Kio8*Q. For contact microradiog- method as demonstrated by Nixon using a raphy, the camera speed is essentially con- silver grid (3). (2) The contact method can stant for a given sample thickness, d, and yield shorter exposure times (higher speed) equal to KioS^/d' and the efficiency is equal and shorter exposure-times-per-unit-sample- Kio8*Q/x^. (In the expressions for S, Q has area (higher efficiency) than are possible been equated to A/d^.) It should be noted with the projection method — this fact is of that the maximum speed, S, for contact utmost importance in the ultrasoft x-ray microradiography is higher than that for analysis. (3) Inasmuch as intense, sharply projection microscopy if the same source or focused sources are required for projection source brightness, io , is used since the sam- microscopy, and are not needed for contact ple-to-film distance, d, for contact micro- microscopy, the instrumentation for the con- radiography is usually less than the sample tact method is much simpler and is easier to to source distance, c, in projection micros- operate for high resolution work, copy. It should be emphasized here, neverthe- The specific intensity, io , is limited by the less, that the projection microscope becomes rate of heat dissipation in the conventional of considerable advantage for certain special large focal spot tubes. For the microfocus problems (see paper of Ong Sing Poen) and source, however, the efficiency of heat trans- also that the excellent electron-optical sys- fer is much higher for geometrical reasons terns which have been developed for projec- and io is not set by the target loading lunit tion microscopy are of very great and unique but rather it is set by the maximum beam value as tools in other kinds of microanalysis, current as allowed for a given resolution re- ultrasoft X-Ray Absorption Measure- quired for the electron optical system. In practice to may be as much as fifty times higher for the microfocus tubes. This fact is often utilized in order to reduce exposure time in contact microradiograph}^, at the sacrifice of sample field, by placing the sam- ple and film very close to the microfocus source. ment Most of the ultrasoft x-ray interaction is by photoelectric absorption, and, in general, the relatively small amount of the incident energy that is scattered is coherent and in the forward direction (low-angle scattering and diffraction) (4). The energy which is photoelectrically absorbed is re-emitted first Contact Method or Projection Method? as a photoelectron and subsequently, as the From the foregoing discussion, certain con- atom returns to its normal state, as fluores- clusions may now be summarized: (1) The cent radiation and Auger electrons. Because 680 ULTKASOFT X-RAY MICROSCOPY of the extremely low fluorescent yield for ultrasoft x-ray interactions the fraction of this energy which reaches the recording ma- terial in either projection or contact micros- copy is negligible. Because of the extremely short range of the emitted electrons within the sample and its supporting membrane, this component of the transmitted energy is also negligible. It is thus evident that projection micros- copy measures the amount of energy which is photoelectrically absorbed and low-angle scattered out of the direct beam. It is im- portant to note that contact microscopy, with ultrasoft radiations, measures only the amount of energy which is photoelectrically absorbed since low-angle scattered energy re- mains effectively in the direct beam within the short distance to the recording material. Because the amount of energy that is not scattered at the very low angles is so small, the contrast, or absorption signal, in contact microradiograms may be considered as re- sulting purely from photoelectric absorption. The measured sample transmission by Eq. (1) depends only upon the mass photoelectric absorption coefficient, ju, and the mass-per- unit-area-thickness, m, of the sample. The photoelectric absorption cross-section is a function only of the x-ray wavelength and the atomic nmnber of the absorbing element ; it is not sensitive to the molecular combina- tion of the elements, and, unlike the scatter- ing cross section, it is not a function of the physical structure of the sample. This simple and direct connection between the mass-and- elementary -chemistry of the sample and the x-ray transmission as easily measured from the microradiogram is probably the greatest and most distinctive advantage of this form of mi- croscopy. It is for this reason the writer has felt that it is important that work be carried out on the measurement and tabulation of photo- electric absorption coefficients for the ultra- soft x-ray region and on the development of efficient ultrasoft x-ray sources of monochro- matic radiation. A tentative tabulation (see Table 1) of light element mass absorption coefficients has been established (5) based upon a semiempirical method of interpola- tion from the best available absorption data and listed here for the several monochro- matic ultrasoft wavelengths which have been found to be useful in this laboratory for quantitative microradiographic analysis. Example of an Ultrasoft X-Ray Source and Camera for Contact Microra- diography There are three basic types of x-ray sources for contact microradiography: The first is a microf ocus source used for maximum speed contact microradiography, as men- tioned above, in which the sample and film are placed very near the source. The high speed is at the expense of sample field which is of the order of size equal to the source-to- sample distance. The second type is a source of conventional focal spot size, a few mm or less, and it permits sample fields of the order of one to two cm in diameter at the minimum working distance for high resolution. There are many such sources described in the litera- ture (1) and these have been applied pri- marily for polychromatic radiation. The third type, which has been under develop- ment in this laboratory, utihzes a large focal spot area and consequently requires rela- tively large anode power; but it is designed for quantitative, microradiographic analysis, presenting monochromatic radiation and a relatively large, uniform field. It is described below. Monochromatic ultrasoft radiation is gained by utilizing the K or L characteristic radiation from an appropriate target mate- rial. This radiation is isolated from the rela- tively low intensity background radiation by a proper choice of filter and excitation volt- age. In order to obtain such radiation it is extremely important that the target be kept free from tungsten and carbonaceous con- tamination otherwise the low-voltage elec- 681 X-RAY MICROSCOPY Table 1. Mass Absorption Coefficients for Some Useful Ultrasoft Wavelengths Derived Semiempirically' with a Probable Error of About ±2% Absorber Atomic Number Al Kai2 8.34 A CuLai! 13.3 A Fe Lai2 17.6 A Cr Lan 21.7 A 0 Kai2 23.7 A Ti La 12 27.4 A CKai2 44. A H 1 7.5 30 70 130 170 260 1100 He 2 30 120 275 500 660 1000 4300 Li 3 78 280 640 1200 1450 2300 9400 Be 4 151.7 581 1288 2292 2965 4532 17430 B 5 323.7 1233 2711 4784 6130 9200 32540 C 6 605 2290 4912 8440 10730 15760 N 7 1047 3795 7910 13120 16270 22590 3647 0 8 1560 5430 10740 16610 983 1473 5470 F 9 1913 6340 11600 1015 1301 1949 7280 Ne 10 2763 8240 1079 1863 2379 3575 13180 Na 11 3129 661 1402 2429 3100 4651 16650 Mg 12 3797 981 2085 3601 4592 6830 22850 Al 13 322.6 1146 2441 4189 5310 7840 24910 Si 14 510 1813 3812 6420 8040 11510 33840 P 15 640 2259 4661 7670 9470 13280 38610 s 16 814 2839 5710 9160 11190 15520 45230 CI 17 990 3364 6530 10210 12450 17330 50100 A 18 1163 3795 7110 11070 13540 18820 K 19 1429 4504 8310 12960 15820 22030 Ca 20 1706 5150 9450 14800 18030 24910 Zapon 998 3571 7270 11690 6500 9470 Parlodion 1177 4167 8390 13320 5450 7870 Animal proteins 854 3095 6391 10480 8719 12584 H2O 1388 4830 9554 14779 893 1338 4984 Zapon (H 5.3%, C 46.7%, N 6.6%, O 41.4%), Parlodion (H 3.6%, C 28.1%, N 11.5%, O 56.8%), Animal Proteins (H 7%, C 52.5%, N 16.5%, O 22.5%o, S 1.5%) tron beam will be prevented from ever reaching the target and exciting the desired characteristic radiation. This primary re- quirement has been met in two ways, as is illustrated in Fig. 5. Tungsten contamination has been elimi- nated by placing the tungsten coil emitter below the target and by electrostatically fo- cusing the electron beam along a circular path to the target surface. In order to avoid the space charge limiting of the beam current which usually accompanies the "hiding" of the anode from the emitter, a special grid at anode potential has been employed. The carbonaceous depositing under the in- teraction of the electron beam and hydrocar- bon vapors at the target has been minimized by maintaining a vacumii of about 10~® mm pressure, using water-cooled baffles and a charcoal trap between the oil diffusion pump and the x-ray source. A sample may be intro- duced without contaminating the x-ray source by using a separately evacuated cam- era and a vacuum isolation gate which is used to introduce a thin filter-window to the 682 ULTRA SOFT X-RAY MICROSCOPY ® ^^2 u ]b(B » ^ m I ® i THAP ANB 0 FFU8I0W PUMP K' 0 + (a) high vacuum isolation gate (i) water-cooled demountable anode (c) accelerating grid (d) helical TUNGSTEN EMITTER e) charcoal trap Fig. 5. The schematic design of an ultrasoft x-ray source of monochromatic radiation in which tungsten contamination of the target has been eliminated by placing the emitter below the anode. SAMPLE AREA ABSORPTION FOIL WEDSE n ROTATINS SECTOR WEDGE Rototlon RECORDING MATERIAL THIN WINDOW (FILTER) Fig. 6. Illustrating the method used for generating an exposure wedge for the precise calibration of a large-field microradiographic plate for quantitative analysis. source when the camera has been piunped microradiogram plate for an accurate ex- down, posure wedge. Such a wedge, as shown in In order to carry out precise quantitative Fig. 6, generated simultaneously with the microradiographic analj^sis, it is necessary to sample exposure, permits an accurate cali- have a sufficient area of uniform field on the bration of the emulsion, without the extreme 683 X-RAY !MICROSCOPY Fig. 7. A print taken directly from a micro- radiogram of 24 embryonic rat head sections illus- trating the relatively large uniform field available for an exposure wedge, multiple samples and con- trols. 25 20 .15 .10 .05 .00 TRANSMISSION FOR "CHRO VIIUM 1- UIL t-IL A ILK / \ / \ / \ Li Li Ln /3l ai2 / ev 679 588574 581571 K^ A 182 21.3 21,6 21.3217 / 1 J 5 10 15 20 25 30 35 40 X (A) Fig. 8. The transmission characteristic for a typical filter as used for isolating the Cr-L(21.7 A) and 0-K(23.7 A) radiations. One kilovolt excita- tion of these characteristic wavelengths places the relatively sharp peak of the continuous radiation within the absorption band, 12 to 18 A. care and control in photographic develop- ment as required by separate emulsion cali- bration exposures. (This same requirement for a large uniform field obtains also for the conventional step-wedge of foils of a material of known thickness and of similar absorption characteristics as that of the sample.) In or- der to achieve this large angular field for the x-ray source without the use of excessively large, thin-foil window areas, a "line" source is employed as shown in Fig. 5, consisting of a water-cooled anode tube. The window is of slit geometry, the long dimension being normal to the direction of the anode so that effectively a "point" area of the anode is "seen" at any given position on the micro- radiographic plate. With the instrument described here, 2 by 3 inch glass microradio- graphic plates are exposed with a good uni- formity of field area for the exposure wedge and with four square inches of area for the mounting of multiple samples and controls (see Fig. 7). In order to gain high specific intensity from the target, the tube has been designed for large anode power — up to about 1000 ma beam currents for anode voltages in the range of 500 to 2000 volts. Almniniun, copper and iron anode tubes are used for Al-K (8.3A) Cu-L(13.3 A) and Fe-L(17.6 A) respectively, vising filter ma- terials of the same metal as the target. A copper anode is used, plated with chromium for Cr-L(21.7 A), oxidized for 0-K(23.7 A), vacuum evaporation coated with titanium for Ti-L(27.4 A), and painted with aquadag for C-K(44 A). In general, relatively heavy filtering and low-excitation voltages are used in order to insure monochromaticity. For example a typical transmission curve for a chromium filter is as shown in Fig. 8. For chromium radiation, one kilovolt excitation is used, placing the relatively sharp peak of the con- tinuous radiation well within the 12 to 18 angstrom absorption band which is illus- trated here. The LS radiation is repressed by the effect of its wavelength lying between those of the critical L II and L III absorption edges. To estimate the effectiveness of a given filter and to determine the maximum tube voltage permissible, it has been con- venient to observe the pulse height distribu- tion appropriately displayed on an oscillo- scope which is connected directly to the preamplifier of a flow-type proportional counter. A pulse height distribution curve for the 0-K radiation from a copper oxide tar- get at 1 kV potential and filtered by a chro- mium foil filter is illustrated in Fig. 9. 684 ULTRA SOFT X-RAY MICROSCOPY C/5 i= 750- I ' 2 r « 1 j^ 500 2 / A ^ Oh m " r^ 1 \ - 250 o / lo - f> i: , L -ba 0Ka(23.7 A) 0 10 20 30 " 40 Fig. 9. Pulse height distribution for 0-K(23.7 A) radiation obtained from an oxidized copper anode at one kilovolt excitation and with a chromium filter as described in Fig. 8. Fig. 10. Photograph of a "plug-in" x-ray source designed as schematically shown in Fig. 5. In Figs. 10 and 11 are shown photographs of the source and the camera, and in Fig. 12 is shown the complete instrument mounted with its power supply. Measurement of the Microradiographic Image The photographic method has been the most successful of those proposed to date for quantitative ultrasoft x-ray microradio- graphic analysis. It would be ideal if the transmitted intensity could be measured di- rectly, as, for example, by a proportional counter, by microphotometry from a fluores- cent screen image, or by the direct measure- ment of the photo and Auger electrons emit- ted by the sample by electron multiplier techniques. However, a very high image in- 685 X-RAY MICHOSCOPY Fig. 11. Showing the three sections, the baffle and trap, the x-ray source, and the camera. tensity is required for precise measurements to be made within one square micron of area or less — this fact is clearly evidenced by the requirement for very high-intensity light sources in ordinary microphotometry of such resolution. And, as indicated earlier, because of the very low efficiency of production of ultrasoft radiations, such high image intensi- ties are extremely difficult to realize. The amplification of the x-ray "signal," resulting from the development of the photographic image, has made this method highly efficient. The resolution of the finest grain photo- graphic emulsions which are currently avail- able, e.g., Eastman Spectroscopic-649, is be- tween 0.1 and 1.0 micron, depending upon the contrast and density of the image. The resolution for line gratings or grids (lines per mm resolution of black and white areas) is equal to or better than that for the light microscope itself. However the spatial reso- lution needed, for example, in the measure- ment of the variation of mass thickness, as measured across an image of a biological cell, is limited to about 0.5 fx so that the recording material and not the light optical measure- ment becomes the limiting factor. The writer has made some preliminary measurements on very thin "grainless" phos- phors used as recording material, photomi- crographing the fluorescent image during the x-ray exposure. These were made by the evaporation of Mn-activated Zn2Si04 (ob- tained from General Electric) according to the method developed by Feldman and O'Hara (6). It was hoped that, at the ex- pense of sensitivity, higher resolution might be gained. Measurements from thin layers of this phosphor evaporated upon quartz cover-slips have indicated comparable granu- FiG. 12. The complete ultrasoft x-ray micro- radiographic analysis system with power supply and vacuum gage circuit. 686 ULTKASOFT X-RAY MICROSCOPY larity to that of Lippmann emulsions and lower contrast. Pattee has carried this work (7) much further and has developed an in- strument, called the microfluoroscope (p. 563), which permits direct viewing of the image from such a phosphor recording ma- terial, gaining the necessary maximum in- tensity by placing the sample within 0.1 mm or less of a microfocus source. He has re- ported that adequate intensity has been obtained for direct viewing, using a 3 mi- cron aluminum target foil, at 9 kilovolt excitation. At the Stockholm Symposium on X-Ray Microscopy in 1959, Auld and Mc- Neil reported that xerography (8). with liquid developers, will also yield resolution comparable to that of the concentrated Lippmann emulsions. To date none of the non-photographic recording materials has been demonstrated as being superior to the photographic method for ciuantitative meas- urements in ultrasoft microradiographic analysis. The following analysis is given in order to establish an optimimi procedure in photo- graphic measurement for microradiography. Let us consider first the determination of the absorption parameter /xm (n the mass ab- sorption coefficient and m the mass-per-unit- area-thickness) for a small object which has caused only a relatively small variation in the density of the microradiogram. The pho- tographic blackening or density will be de- fined as Ds in the region originally under the sample object, and as Db in the immediate surrounding. The corresponding x-ray ex- posures which resulted in these densities will be defined as Es and Eb . The photographic densities are related to the photometer read- ings on the microradiogram, (p2 in the sam- ple region, pi in the immediate surrounding, and po ill the clear emulsion) by the defining equations D, = log (P0/P2) and Db = log (po/pi) (S) The ratio of the x-ray exposures for the sam- ple and surroundings is equal to the ratio of the corresponding intensities (reciprocity law obtaining for these wavelengths) and there- fore given, from {1), as EJEi = e-*"" (9) SO that log Eb - log E, = y.m log e = A (log E) (10) Since the parameter jum is measured by the value A(log E), it is convenient to present the emulsion characteristics as D = /(log E) (11) as is conventional for light photography. Then from (8), (10) and (11) and for rela- tively small differences between Eb and Es , we may write ^D/A(\ogE) = df/d(logE) 7 = log (p2/pi)/fi»i log e giving finally y/im = In (P2/P1) (W where 7 is the slope of the D vs log E charac- teristic, conventionally designating emulsion contrast, and log(p2/pi)/log e has been re- placed by the logarithm to the natm'al base, e, viz. In(p2/pi). We may now determine the effect of the measurement errors upon the determination of fim. By differentiating (12) we obtain d(nm)/ixni = ^^(dpi/p-i) T (dpi/pi) yfitn V(dpi/pi)^ + (dp2/p2)^ (jixm) (IS) in which we have added (T) errors in the usual manner. In order to define the conditions required for minimvmi error it is necessary to define the type of photometric error, (dp/p), in- volved in a given measurement. For loiv mag- nification Avork in which emulsion granularity is not the dominant source of error, we may often have dpi ^ dp2 ~ a constant over the entire range of the photometer scale. For this case, (18) may be rewritten as 687 X-RAY MICROSCOPY d(Mm)/^m = |^/l + (P^ In (papi) (Low Magnification) (.dp2/p2) (W where (dp2/p2) is the relative error of the higher photometer reading. The low magni- fication error function, presented in brackets in (14) has a minimum value as illustrated in Fig. 13. In order to minimize such error, p2 should be set at full photometer scale reading and P2 should lie between about }^ and ^2 of this P2 reading. This requires a corresponding values of yfxm in the range 0.7 to 2.0. It is important to note the rapid increase of the error in nm for lower values of yfxni. To measure relatively small values of /xm, it is clearly evident that the exposure time be such as to gain a maximum value for the contrast, 7. As may be deduced from the characteristic curves, D vs log E, as plotted in Fig. 14, this requires exposures for rela- tively high densities, contrary to conven- tional x-ray practice. It is convenient that this high-7 region also produces optimum visual contrast — a fact which has been long recognized. And finally since 7 is constant in this high-density region, the relation given in (12) need not apply only for small values of jjLm, as assumed at the outset of this analy- sis. For high magnification work, the domi- nant source of photometric error is emulsion granularity. It has been found that for such errors, (dpi/pi) = (^^2/^2) = a constant for a given high magnification measurement (500 to lOOOX). For this case we rewrite (13) as d inyn) lixm = \/2 (dp/p) /yfim = \/2/m»i (y/N) (15) The value of (dp/p) , as due mainly to photo- graphic noise and designated by the symbol A'', has been measured as the average value of the relative variation in the photometer signal as read from microphotometer tracings at various microradiogram densities using a 0.5 by 1.0 micron slit. The results of such measurements are also illustrated in Fig. 14. It should be noted that nm increases more PHOTOMETRIC ERROR Ai./* = e(APi/P«) Y jJ-TI » 0 0.5 1.0 L5 2.0 8 .£. WHERE Aft/ft - RELATIVE ERROR OF Pt READING ASSUMING ■ y/in = LOGe Pt/Pi SAMPLE SIGNAL BACKGROUND — ZERO light- Fig. 13. The prediction of error for a measurement for the case of constant absolute photometric error, dpi ~ dp2 , as typical in low -magnification work. 688 ULTRASOFT X-RAY MICROSCOPY A, ALUMINUM K(,-8.34 A B. OXYGEN Ka-23.7 A J I I i_ I I I I I I I I I I I I I I NOISE- N -f^^^ ■•* H I I I I I I i I I I I I I 120 80 40 0 12% 8 4 0 3 y/N J I I I L. 3 0 LOG E- DENSITY-D- FiG. 14. The D vs log E characteristic for two wavelengths illustrating the longer linear region at the higher densities for the ultrasoft wavelength and the associated concontrast values, 7, and relative granularity noise, dp/p, measured with 0.5 X 1.0 micro slit microphotometry. rapidly with wavelength than does (y/N), indicating the value of the longer wave- lengths in reducing measurement error. For a given wavelength, the value of {y/N) may be considered a "figure of merit" for micro- radiographic emulsions. The value of (y/N) is a maximum in the linear high density re- gion is thus the optimum working region for both low and high magnification microradio- graphic measurement. (y/N) has been measured in a recent in- vestigation (9) as a function of the type and time of photographic development which has indicated that a "soft" development such as 2 to 3 minutes in D158 (Eastman) seems optimum. We may now summarize, from the fore- going discussion, the procedures for an opti- mum photographic measurement in mi- croradiographic analysis: The absorption parameter, /xm, may be simply determined for either high or low magnification through the relation fxm = (I/7) In (pi/pi) providing the exposure is for the linear high- density region (D > 1.0 for ultrasoft radia- tions and spectrographic-649 emulsion) for which 7 is a constant and of maximimi value. The minimiun wavelength permissible for low magnification (dpi ;^ dp2) is set by the requirement that yiiin should lie within the range 0.7 to 2.0. The measurement errors become very large in this case for yum < 0.5. The minimimi wavelength is set for high magnification (dpi/pi ^ dp2/p2) by the re- quirement that ij.m satisfies the relation (15) for a given required measurement precision, d(fj.m)/fjLm, (or for dm/w-uniform chemistry) and for a given emulsion and hence (y/N) value. The minimum wavelength and ex- posure time are thus fixed for each microra- diographic measurement. As shown, for example, in Fig. 14, the (y/N) value for Spectroscopic-649 plate for 23.7 A radiation is about 60. Therefore, from (15), for a minimum error 2%, the x-ray wavelength should be chosen so that juw is at least equal to unity, corresponding to ap- proximately 60 % absorption. To measure high densities on microradio- gram areas of about one square micron or 689 X-RAY MICROSCOPY Fig. 15. The utilization of standard light mi- croscope optics for high resolution microphotom- etry. less presents a relatively difficult problem in microdensitometry which has been solved in the following manner: A very intense, con- centrated light source is necessary, such as the high pressure mercury arc (AH-6 of Gen- eral Electric for example). A trinocular mi- croscope with a Leitz photomicrographic attachment has been found to be most con- venient for the optical system, as shown in Fig. 15. The light is diverted to the binocular section only when surveying the microradio- gram and centering the object area to be measured. For densitometric measurement the first image, as formed by the objective just below the projector lens in the photo-at- tachment, falls upon a stop with a 0.5 mm hole at its center. Only light from this cen- tral portion of the image proceeds directly to the top of a camera chamber which has been added to the Leitz attachment. Here 2" x 2'' lantern slide plates can be exposed for con- venient photomicrography, at 500 X or less. A low-power eyepiece is placed here for critical, direct focusing on the enlarged field. Also, at this final image plane, a "light pipe" may be slid into place consisting of a thin piece of glass with a 45° polished end and masked so that light from a 0.25 X 0.5 mm central field area may be internally and to- tally reflected at its tip, and channeled down the long glass section to a 1P21 (RCA) pho- tomultiplier tube. The precise position of the equivalent 0.5 by 1.0 micron object field which is being measured may be simultane- ously viewed by the upper eyepiece. The phototube output is "matched" to a ten millivolt chart recorder by connecting it di- rectly to the 1.5 volt input of a standard vacuum tube voltmeter. The recorder is at- tached to the meter circuit of the VTVM. The sensitivity of the system is set by vary- ing the multiplier tube voltage in the range 450-900 volts. When a profile rather than point meas- urement is desired, the microradiogram is translated by means of a simple mechanical linkage between the stage screw and a syn- chronous motor. The latter is geared down to effect a 40 micron motion of the stage in about 3 minutes. A 95X oil immersion apochromat objec- tive is used. Monochromatic light, effectively the mercury 4360 A blue light, is obtained by a filter — it has been found that the (y/N) ratio for blue light is 30 % higher than that for the green Hght (5461 A) because of the higher scattering efficiency of the Lippmann grains for the shorter wavelengths. The mon- ochromatic light yields a significant improve- ment in image quality also. Once the image field to be measured is lo- cated in the central hole at the stop in the projector lens, the prisms for side viewing are thrown out of the beam so that there are a minimimi number of glass svirfaces in- volved, and only the light fonning the cen- tral image area which closely surrounds that 690 ULTRASOFT X-RAY MICROSCOPY being measured is allowed to pass through the projector lens to the phototube window. In this manner stray light scatter is effec- tively eliminated. In Fig. 16, the resolution of this system is demonstrated by the resolv- ing of the Fresnel fringes associated with the imaging of a diffraction grating of 10 micron spacing ; the freedom from stray light scatter- ing or "spill-over" is shown by the relatively sharp corners of the shadow produced by a 25 micron wire. Application of Contact Microradiog- raphy to Microniass Measurement A standard test object which has been used in this study for quantitative microradio- graphic analysis has been the human red blood cell (erythrocyte) which was chosen not only because of its importance to certain medical research, but in particular here be- cause it is of such material and size as to make its precise measurement approach the MICROPHOTOMETER RESOLUTION ( 0.5/1 x|0/i SLIT) IQS 1 . ' i ' • - - - |: - — ^1 — V- Hrz r:M: IO/< - ZEISS CALIBRATION GRATING 7FRn 1 ir;nT-^- ^t ^LrxU Llofl 1 1 ' "" 1 1 j ( r 1 _l t-~.A ^ '^ . 1 ■;: 1 1 ZERO LIGHT — o.oor— WIRE Fig. 16. Tests of the resolution and freedom from stray light scatter for the microphotometer shown in Fig. 14. 10^ 10' — — ™ A A v" / — — — ^^ h — / y- / / / / / / / / ^ t ' 1 1 1 — — VAGUU M DRIED - 20° C C-53.5 % % 0-20.5% ^GANIC - 2 % H- T\ « N-16.5 IN0( 8 10 50 100 X{A)- FiG. 17. Absorption characteristics for ultra- soft x-radiations of essentially the hemoglobin composition of human red blood cells (as deter- mined for one test sample). limit of present day x-ray microscopy and appreciably beyond the capability of other types of microscopy, e.g., interferometric microscopy. Sample smears of these cells are easy to prepare in a standard, reproducible manner. The single vacumn dried cell is ap- proxmiately 8 microns in diameter, 2 microns in depth, biconcave, of total mass 30 X V ^ n DATA RCOUCHON GRID 1/2 » I /I SLIT MEASUREMENT Fig. 19. Photometer readings, p2 and pi , are taken from an average profile, assuming radial symmetry, at sixteen points at radial positions of such interval as to yield equal effect upon the total mass measurement. The reduction of such data to mass per unit area, m, and subsequent numerical integration for total mass was carried out with a digital computer. .75 1.00 Fig. 20. The calculated values of yum for eight hundred single cell measurements were divided by the average value to obtain a corresponding num- ber of m/fFi values which are thus independent of 7 and M and their respective errors . These data were counted in intervals of 0.1. The precision (average deviation) of a single measurement is of the order 0.03. 692 ULTRASOFT X-RAY MICROSCOPY crophotometer tracing of an individual cell. Using the relation {12) the mass thicl<;ness, m, was calculated for each radial position and the total mass was obtained from these values b}^ a nimierical integration (using Simpson's rule). Each experimental point is of equal effect in the calculation for total mass because these were taken at radial posi- tion of equal steps in the T^^-value. {R the radial position) as shown by the relation for total mass, M M 2irRm{R) dR Jo 7rm d{R^) {16) A digital computer Avas programmed to do this numerical integration so as to allow a large number of cells to be measured viz. 800. The average cell mass for the particular sam- ple used in this study was found to be 29.7 X jO-12 gram. Twenty-five independent meas- urements were taken on one typical cell in order to determine a precision measure. For this cell, the yiim value was 45.8 and the average deviation was 1.78 or about 2.6% which is consistent with that predicted above from {15). This corresponds to a precision of about 1 micro-microgram for this cell measurement. With such precision, it was possible to pre- sent the 800 single-cell measurements as a histogram in order to indicate the manner with which the cell mass is distributed about an average value. This single cell mass distri- bution curve is shown in Fig. 20. This distri- bution is presented in mass units m relative to average mass in and was based directly upon the yum data so it is independent of the accuracy of measurement of both 7 and the mass absorption coefficient, /x. REFERENCES 1. a. "Principles of Microradiography and Bib- liography," prepared by the Eastman Ko- dak Research Laboratory and Published by the Philips Electronics Corporation, Mt. Vernon, N. Y. b. George L. Clark, "Applied X-Rays," McGraw-Hill, N. Y., Chapter 10, 1955. c. "X-Ray Microscopy and Microradiogra- phy," Academic Press, N. Y., 1957. d. A. Engstrom, "Historadiography," Physi- cal Techniques in Biological Research, Aca- demic Press, N. Y., Vol. Ill, 1956. e. LiNDSTROM, B., (Thesis), Acta Radiol., Supp. 125, 1955. f. Ong Sing Poen, "Microprojection with X-Rays," Martinus Nijhoff, The Hague, 1959. g. Henke, B. L., "Microstructure, Mass and Chemical Analysis with 8 to 44 A X-Radia- tions," Proceedings of the Seventh Annual Conference on Industrial Applications of X-Ray Analysis, University of Denver, 1958. 2. Ong Sing Poen and Le Poole, J. B., Applied Scientific Research B7, 265 (1958) . 3. Nixon, W. C, Proc. Rorj. Soc. A232, 475 (1955). 4. Henke, B. L. and Miller, J. C, Technical Report— AFOSR TN 59-895. 5. Henke, B. L., White, R., and Lundberg, B., J. Appl. Phijs. 28, 98 (1957). 6. Feldman, C. and O'Hara, M., J. Opt. Soc. Am., 47, 300 (1957). 7. Pattee, H. H., Science, 128, 977 (1958). 8. See also Schaefert, R. N. and Oughton, C. D., /. Opt. Soc. Am. 38, 991 (1948). 9. Henke, B. L. and Ong Sing Poen (in prepara- tion) . 10. Henke, B. L. and Maley, C. (in preparation). Burton L. Henke VASCULAR AND DENTAL APPLICATIONS OF PROJECTION MICROSCOPY. See MICRO- ANGIOGRAPHY, p. 627. 693 m