(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "A handbook of sugar analysis: a practical and descriptive treatise for use in research, technical and control laboratories"

A HANDBOOK OF 

SUGAR ANALYSIS 



A PRACTICAL AND DESCRIPTIVE TREATISE 
FOR USE IN 

RESEARCH, TECHNICAL AND CONTROL 
LABORATORIES 



BY 

C. A. BROWNE, PH.D. 

Chemist in charge of the New York Sugar Trade Laboratory 

{Formerly Chief of the Sugar Laboratory, U. 8. Bureau of Chemistry, 

Washington, D.C., and Research Chemist of the Louisiana 

Sugar Experiment Station, New Orleans, La.) 



SECOND EDITION 

FIRST THOUSAND 



NEW YORK 
JOHN WILEY & SONS 

LONDON: CHAPMAN & HALL, LIMITED 
1912 



COPYRIGHT, 1912, 

BY 
C. A. BROWNE 

Copyright, 1912, in Great Britain 



Stanbepe iprcss 

F. H. GILSON COMPANY 
BOSTON, U.S.A. 



TP.3Z2 



AGRIC. 
LIBRARY 



DEDICATED 
TO HIS TEACHER, 

GEH.-RATH PROF. DR. B. TOLLENS, 

OF GOTTINGEN UNIVERSITY, 

AS A TOKEN OF GRATITUDE AND ESTEEM, 

BY THE AUTHOE 



PREFACE 



THE subject of sugar analysis, which a generation ago was limited 
to determinations of density, specific rotation and reducing power, 
has greatly expanded within the past twenty-five years. Instru- 
ments of greater accuracy have been devised, old methods have been 
improved and new methods have been discovered. In the present 
volume the purpose of the author has been to give a rather wide, 
but a by no means complete, selection of the more recent methods of 
sugar analysis and at the same time to retain the more important 
features of the older textbooks. 

The range of sugar analysis is so broad that in the selection of 
methods the author has been guided largely by his own experience in 
various research, technical and control laboratories. While the par- 
ticular methods chosen for description may not in all cases meet with 
general approval it is hoped that the underlying principles of sugar 
analysis have been covered sufficiently to enable the chemist to make 
his own applications and modifications. References to special works 
and original articles will assist the chemist in case he desires to follow 
some special line of investigation more fully. 

Next to the knowledge of a method the most important fact which 
the student of sugar analysis must acquire is the knowledge of this 
method's limitations. The great susceptibility of the sugars to 
chemical changes and to variations in specific rotation, reducing power 
and other "constants" is a factor which the sugar chemist must al- 
ways bear in mind. The prescribed methods of analysis are usually 
too silent upon these points, and the inexperienced chemist often pro- 
ceeds to make general use of a formula or method which has only a 
limited applicability. The author has endeavored to correct this 
tendency by including with the description of each method a brief 
account of its applicability and limitations. 

In the examination of sugar-containing materials the problems of 
analysis are much simplified by a knowledge of what one may expect 
to find. The author has felt that a work upon sugar analysis is not 
complete without some description of the sugars themselves. In 
Part II of the present volume, he has therefore included a brief 



Vi PREFACE 

account of the occurrence, methods of preparation, properties and 
reactions of the different sugars and their allied derivatives. Brief 
references are also made to methods of sugar synthesis; the latter play 
such an important part in the separation and isolation of the rarer 
sugars that the sugar analyst is not fully equipped without some knowl- 
edge of synthetic processes. 

The principal textbooks and journals which have been consulted 
in preparing the present volume are named in the Bibliography. 
The author's obligations to these are indicated in most cases by the 
footnotes. In reviewing original papers, the abstracts and references 
contained in Lippmann's "Chemie der Zuckerarten" and his "Berichte 
liber die wichtigsten Arbeiten aus dem Gebiete der reinen Zucker- 
chemie," published semiannually in "Die Deutsche Zuckerindustrie," 
have been of invaluable service. 

In concluding his task, which has extended with many interrup- 
tions over a period of five years, the author desires to thank the many 
friends and coworkers who, by their help and encouragement, have 
greatly lightened his labors. 

Special obligations are due to Dr. C. S. Hudson for reviewing the 
section upon mutarotation and to Prof. H. C. Sherman for suggestions 
upon methods for determining diastatic power. Acknowledgement is 
also made of courtesies extended by Mr. A. H. Bryan and by Mr. 
G. W. Rolfe. 

For the use of cuts contained in Dr. G. L. Spencer's "Handbook 
for Cane Sugar Manufacturers" and in A. E. Leach's "Food Inspec- 
tion and Analysis" the author owes an acknowledgment to the 
authors of these books and to his publishers Messrs. John Wiley & 
Sons. To the latter also he would express his appreciation of the 
hearty support which has been given and of the generous considera- 
tion which has been shown for the many delays incident to the com- 
pletion of the work. 

NEW YORK, N. Y., August, 1912. 



BIBLIOGRAPHY 



SPECIAL WORKS 



Author or Editor 



Abderhalden . 
Abderhalden . 

Allen 

Armstrong . . . 

Basset 

Browne . . 



Bryan .... 



Bujard and Baier 

Claassen (Hall and Rolfe) 

Cross and Bevan 

Cross and Bevan 

Czapek 

Deerr 

Emmerling 



Fischer, E . . . 
Fischer, F. . 
Fribourg 
Friihling.... 



Gange 



Geerligs . 
Geerligs . 



Gredinger 

Jago 

Konig.... 



Lafar. . . 
Landolt . 



Leach 

Lippmann 



Title 



Biochemisches Handlexikon, Vol. II (1911). 
Handbuch der Biochemischen Arbeitsmeth- 

oden, Vol II (1909).. 
Commercial Organic Analysis, Vol. I 

(1901). 
The Simple Carbohydrates and the Gluco- 

sides (1910). 
Guide pratique du Fabricant de Sucre 

(1872). 
Chemical Analysis and Composition of 

American Honeys (1908). Bull. 110, 

U. S. Bureau of Chemistry. 
Analyses of Sugar Beets, 1905 to 1910, 

together with Methods of Sugar De- 
termination (1911). Bull. 146, U. S. 

Bureau of Chemistry. 
Hilfsbuch fur Nahrungsmittelchemiker 

(1900). 

Beet Sugar Manufacture (1906). 
Cellulose (1895). 

Researches on Cellulose, 1895-1900 (1901). 
Biochemie der Pflanzen, Vol. I (1905). 
Cane Sugar (1911). 
Die Zersetzung stickstofffreier organischer 

Substanzen durch Bakterien (1902). 
Untersuchungen uber Kohlenhydrate und 

Fermente (1884-1908), (1909). 
Handbuch der chemischen Technologic, 

Vol. II (1902). 
L' Analyse chimique en Sucreries et Raf- 

fineries de Cannes et Betteraves (1907). 
Anleitung zur Untersuchung der fiir die 

Zuckerindustrie in Betracht kommen- 

den Rohmaterialien, Produkte, Neben- 

produkte und Hilfssubstanzen (1903). 
Lehrbuch der Angewandten Optik in der 

Chemie. Spectralanalyse, Mikroskopie, 

Polarisation (1886). 

Cane Sugar and its Manufacture (1909). 
Methods of Chemical Control in Cane 

Sugar Factories (1905). 
Die Raffination des Zuckers (1909). 
The Technology of Bread Making (1911). 
Die Untersuchung landwirtschaftlich und 

gewerblich wichtiger Stoffe (1898). 
Technische Mykologie (1897-1907). - 
Das optische Drehungsvermogen organ- 
ischer Substanzen und dessen prak- 

tische Anwendungen (1898). 
Food Inspection and Analysis (1911). 
Die Chemie der Zuckerarten (1904). i^ 



Vlll 



BIBLIOGRAPHY 



Author or Editor 
Maquenne 



Mittelstaedt (Bourbakis) 



Pavy 

Plimmer 



Preston 
Rolfe.. 



Riimpler. . . . 
Sherman .... 
Sidersky 



Sidersky , 

Sidersky, 
Spencer . 



Sykes and Ling. 
Tervooren . . 



Tollens. 



Tucker 

Van't Hoff (Marsh) . . . 

Walker 

Ware.. 



Wein (Frew) 



Wiechmann 

Wiedemann and Ebert , 



Wiley.... 
Wiley.. 



Title 

Les Sucres et leurs principaux Derives 
(1900). 

Technical Calculations for Sugar Works 
(1910). 

Physiology of the Carbohydrates (1894). 

The Chemical Changes and Products Re- 
sulting from Fermentations (1903). 

The Theory of Light (1901). 

The Polariscope in the Chemical Labora- 
tory (1905). 

Die Nichtzuckerstoffe der Ruben (1898). 

Methods of Organic Analysis (1912). 

Les Densites des Solutions Sucrees a dif- 
ferentes Temperatures (1908). 

Manuel du Chimiste de Sucrerie, de Raf- 
finerie et de Glucoserie (1909). 

Polarisation et Saccharimetrie (1908). 

A Handbook for Cane-Sugar Manufac- 
turers and their Chemists (1906). 

The Principles and Practice of Brewing 
(1907). 

Methoden van Onderzoek der bij de Java 
Rietsuiker-Industrie voorkomende Pro- 
ducten (1908). 

Kurzes Handbuch der Kohlenhydrate 
(1895-8). 

A Manual of Sugar Analysis (1905). 

Chemistry in Space (1891) 

Introduction to Physical Chemistry (1903). 

Beet Sugar Manufacture and Refining 
(1905-7). 

Tables for the Quantitative Estimation of 
the Sugars (1896). 

Sugar Analysis (1898). 

Physikalisches Praktikum mit besonderer 
Beriicksichtigung der Physikalisch- 
Chemischen Methoden (1899). 

The Principles and Practice of Agricultural 
Analysis, Vol. Ill (1897). 

Official and Provisional Methods of Analy- 
sis, Association of Official Agricultural 
Chemists. Bull. 107 (Revised) U. S. 
Bureau of Chemistry. 



PERIODICALS 



Abbreviation 

Am. Chem. Jour 

Am. Sugar Ind 

Analyst 

Ann 

Ann. chim. phys 

Archief Java Suiker Ind.. 
Archiv Pharm. . . 



Biochem. Zeitschrift 

Bull, assoc. chim. sucr. dist. 



Title 

American Chemical Journal. 
American Sugar Industry. 
Analyst. 

Annalen der Chemie (Liebig's). 
Annales de chimie et de physique. 
Archief voor de Java Suiker Industrie. 
Archiv der Pharmazie. 
Berichte der deutschen chemischen Gesell- 

schaft. 

Biochemische Zeitschrift. 
Bulletin de I'association des chimistes de 

sucrerie et de distillerie de France et 

des colonies. 



BIBLIOGRAPHY 



Abbreviation 

Bull. soc. chim 

Centralblatt 

Centrbl. Zuckerind 

Chem. News 



Chemiker-Ztg. 
Compt. rend.. 



Deut. Zuckerind 

Dingier 's Poly tech. Jour. 

Int. Sugar Jour 

J. Am. Chem. Soc 

J. Chem. Soc 

J. fabr. sucre 

Jour, f . Landwirtsch 
J. Ind. Eng. Chem 



J. pharm 

J. pharm. chim. . . 
J. prakt. Chem. . 
J. Soc. Chem. Ind. 

La. Planter 

Land. Vers.-Stat.. 



Monatshefte 

Mon. scient 

Neue Zeitschrif t 

Oest.-Ung. Z. Zuckerind. 

Pfliiger's Archiv 



Pogg. Ann 

Proceedings A. O. A. C 



Proceedings Int. Cong. App. Chem. 

Rec. trav. Pays-Bas 

Sitzungsber. Wiener Akad 

Stammer's Jahresbericht . . 



Sucrerie Beige 

West Indian Bull 

Wochenschr. f. Brauerei. . , 

Z. analyt. Chem 

Z. angew. Chem 

Z. Instrument 

Z. physik. Chem 

Z. physiol. Chem 

Z. Spiritusind 

Z. Unters. Nahr. Genussm. 



Z. Ver. Deut. Zuckerind. 
Z. Zuckerind. Bohmen.. 



Title 

Bulletin de la societe chimique de France. 

Chemisches Centralblatt. 

Centralblatt fur die Zuckerindustrie. 

Chemical News and Journal of Physical 
Science. 

Chemiker-Zeitung . 

Comptes rendus hebdomadaires des seances 
de 1'academie des sciences. 

Die Deutsche Zuckerindustrie 

Dingier 's Polytechniches Journal. 

The International Sugar Journal. 

Journal of the American Chemical Society. 

Journal of the Chemical Society (London). 

Journal des fabricants de sucre. 

Journal fur Landwirtschaft. 

Journal of Industrial and Engineering 
Chemistry. 

Journal de pharmacie. 

Journal de pharmacie et de chimie. 

Journal fur prakt ische Chemie. 

Journal of the Society of Chemical In- 
dustry. 

The Louisiana Planter- and Sugar Man- 
ufacturer. 

Die landwirthschaftlichen Versuchs-Sta- 
tionen. 

Monatshefte fiir Chemie. 

Moniteur scientifique. 

Neue Zeitschrif t fiir Riibenzuckerindustrie. 

Oesterreichisch-Ungarische Zeitschrift fiir 
Zuckerindustrie und Landwirthschaft. 

Pfliiger's Archiv fiir die gesammte Physiol- 
ogic der Menschen und der Thiere. 

Poggendorff's Annalen. 

Proceedings of the Association of Official 
Agricultural Chemists. 

Proceedings of the International Congress 
of Applied Chemistry. 

Recueil des travaux chimiques des Pays- 
Bas. 

Sitzungsberichte der kaiserlichen Akademie 
der Wissenschaften, Wien. 

Stammer's Jahresbericht iiber die Unter- 
suchungen und Fortschritte auf dem 
Gesamtgebiete der Zuckerfabrikation. 

La Sucrerie Beige. 

West Indian Bulletin. 

Wochenschrift fiir Brauerei. 

Zeitschrift fiir analytische Chemie. 

Zeitschrift fiir angewandte Chemie. 

Zeitschrift fiir Instrumentenkunde. 

Zeitschrift fiir physikalische Chemie. 

Zeitschrift fiir physiologische Chemie. 

Zeitschrift fiir Spirit usindustrie. 

Zeitschrift fur Untersuchung der Nahrungs- 
und Genussmittel. 

Zeitschrift des Vereins der Deutschen 
Zuckerindustrie. 

Zeitschrift fiir Zuckerindustrie in Bohmen. 



TABLE OF CONTENTS 



PAGE 

PREFACE v 

BIBLIOGRAPHY v jj 

PART I 

PHYSICAL AND CHEMICAL METHODS OF SUGAR ANALYSIS 1 

CHAP. 

I. SAMPLING OF SUGAR AND SUGAR PRODUCTS 3 

II. DETERMINATION OF MOISTURE IN SUGARS AND SUGAR PRODUCTS BY 

METHODS OF DRYING 15 

III. DENSIMETRIC METHODS OF ANALYSIS 27 

IV. PRINCIPLE AND USES OF THE REFRACTOMETER 50 

V. POLARIZED LIGHT, THEORY AND DESCRIPTION OF POLARIMETERS . . . 76 

VI. THEORY AND DESCRIPTION OF SACCHARIMETERS 108 

VII. POLARISCOPE ACCESSORIES 146 

VIII. SPECIFIC ROTATION OF SUGARS 172 

' IX. METHODS OF SIMPLE POLARIZATION 194 

X. METHODS OF INVERT OR DOUBLE POLARIZATION 263 

XI. SPECIAL METHODS OF SACCHARIMETRY 287 

XII. MISCELLANEOUS PHYSICAL METHODS AS APPLIED TO THE EXAMINA- 
TION OF SUGARS 307 

XIII. QUALITATIVE METHODS FOR THE IDENTIFICATION OF SUGARS 333 

XIV. REDUCTION METHODS FOR DETERMINING SUGARS 388 

XV. SPECIAL QUANTITATIVE METHODS 449 

XVI. COMBINED METHODS AND THE ANALYSIS OF SUGAR MIXTURES. ... 472 

XVII. MISCELLANEOUS APPLICATIONS 494 

PART II 

THE OCCURRENCE, METHODS OF PREPARATION, PROPERTIES AND PRINCIPAL 

REACTIONS OF THE SUGARS AND ALLIED DERIVATIVES 525 

XVIII. CLASSIFICATION OF THE SUGARS AND THEIR FORMATION IN NATURE. 527 

XIX. THE MONOSACCHARIDES 535 

XX. THE DISACCHARIDES 643 

XXI. THE TRISACCHARIDES AND TETRASACCHARIDES 731 

XXII. THE AMINO SUGARS AND THE CYCLOSES 751 

XXIII. THE SUGAR ALCOHOLS AND SUGAR ACIDS 764 

APPENDIX OF SUGAR TABLES 789 

INDEX.. xiii 



PAET I 

PHYSICAL AND CHEMICAL METHODS OF 
SUGAK ANALYSIS 



ANALYSIS 



CHAPTER I 

SAMPLING OF SUGAR AND SUGAR PRODUCTS 

IN the analysis of sugars and sugar products, special stress must be 
laid upon the correctness of sample. Accuracy in analytical details is 
of no value unless the portion of substance weighed out for examina- 
tion is an accurate sample of the entire lot of product in question. 
While the chemist is not always charged with the supervision of sam- 
pling, he should, nevertheless, acquaint himself so far as possible with 
the history of his product before it is received.. In this way he may 
often explain differences which might otherwise be attributed to mis- 
takes of analysis. A few introductory pages devoted to the general 
subject of sampling may, therefore, not be amiss. 

The best illustration of methods of sampling, and of the errors con- 
nected therewith, is furnished by raw cane sugar. The sampling of this 
commodity is selected first and discussed in somewhat fuller detail. 

SAMPLING OF RAW SUGARS 

The raw sugar imported from the various sugar-producing countries 
comes in a great variety of forms. Centrifugal sugar, from Cuba, 
Porto Rico, and most of the West Indian Islands, comes in 300-lb. 
jute bags; sugar from the Hawaiian Islands comes in 125-lb. bags; 
sugar from Java comes either in bags or large cylindrical baskets 
weighing from 500 to 700 Ibs.; sugar from the Philippines comes in 
small wicker mats weighing about 50 Ibs.; Muscovado sugars, which 
are purged by draining and contain much molasses, come usually in 
large hogsheads. In addition to the above forms of package, sugars 
come occasionally in boxes, barrels, grass mats, ceroons, and other 
receptacles. 

The need for carefully prescribed rules in sampling sugar becomes 
at once self-evident when we consider the different forms of the package 
and the exceedingly variable character of the sugar which may be con- 

3 



4 SUGAR ANALYSIS 

tained therein. The sugar, for example, may contain lumps of higher 
or lower polarization than the finer part of the product; the sugar may 
also retain considerable amounts of molasses, sometimes as high as 
30 per cent, which drain during transit or storage and form the " foots " 
at the bottom of the package. The difference in composition between 
the top and bottom layers of a hogshead of Muscovado sugar, which 
is a kind that " foots" easily, is very marked. In addition to the 
differences in composition of sugar within the single packages are 
the differences in composition between different packages of the same 
lot. These differences may be the result of manufacture; they may 
also result when no dunnage is used for covering the bottom of the 




Fig. 1. Trier for sampling sugar. 

holds of the ships used for transport, with the result that the bottom 
tiers of sugar may be damaged through absorption of bilge water. In 
many cases the top tiers of sugar suffer the damage, as when sugars 
sweat beneath the hatches; the vapors from the warm sugar rise, con- 
dense, and then drop back upon the upper layers of the cargo. If the 
packages of sugar run unevenly it is difficult to secure a representative 
fraction unless every container is sampled. The most approved method 
of sampling at present is to take a specimen of sugar so far as possible 
from every package.* 

Sugar is sampled in the same way as fertilizers and many other 
commodities, by means of a trier. This implement (Fig. 1) consists of 
a long pointed rod of steel with a groove or spoon upon one side. A 

* For a discussion of this and other points pertaining to methods of sampling 
raw sugar in different countries see paper by F. G. Wiechmann (Int. Sugar Journ., 
9, 18-28) read before the Fifth Meeting of the International Commission for Uni- 
form Methods of Sugar Analysis, Bern, 1906. 



SAMPLING OF SUGAR AND SUGAR PRODUCTS 



thrust of the trier into the package forces the sugar along its pathway 
tightly into the bowl of the spoon; the sugar thus adhering, after the 
trier is withdrawn, is removed by the thumb, or by means of a scraper, 
into a covered bucket, and the process is continued until a sufficient 
number of packages have been sampled to constitute a mix; this 
number may vary, according to the size of lot and kind of sugar, from 
one package to several thousand. The practice of the New York Sugar 
Trade is to mix twice daily, and in no case is a sample to remain un- 
mixed over night. 

It is of course important that the triers of the different workmen 
who are sampling a given lot of sugar should be exactly alike, es- 
pecially as regards the dimensions of the spoons. The specifications 
of the United States Treasury Department Regulations* are very ex- 
plicit upon this point and give the following dimensions of the short, 
long, and barrel triers. 

TABLE I 

Giving Dimensions of Triers for Sampling Sugar 





Short trier. 


Long trier. 


Barrel trier. 


Length over all 


Centimeters. 
40 6 


Centimeters. 

152 4 


Centimeters. 

104 


Length of spoon 


22 9 


132 1 


91 4 


Length of shank ... 


17 8 


20 3 


12 7 


Length of handle 


26 7 


38 1 


30 5 


Width of spoon 


2 7 


2 5 


2 5 


Depth of spoon 


8 


1.3 


1.1 


Diameter of handle 


3.8 


3.8 


3.8 











According to the United States Treasury Department Regulations,! 
"sugar in hogsheads and other wooden packages shall be sampled by 
putting the long trier diagonally through the package from chime to 
chime, one trierful to constitute a sample, except in small lots, when an 
equal number of trierfuls shall be taken from each package to furnish 
the required amount of sugar necessary to make a sufficient sample. 
In the sampling of baskets, bags, ceroons, and mats the short trier 
shall be used, care being exercised to have each sample represent the 
contents of the package." 

It is necessary in sampling to keep the triers always clean; the stick- 
ing of sugar to the bowl of the spoon is especially annoying with some 

* Regulations governing the weighing, taring, sampling, classification, and polari- 
zation of imported sugars and molasses. U. S. Treasury Department, Division of 
Customs, Document No. 2470, Art. 5. 

t Loc. cit., Art. 6. 



6 SUGAR ANALYSIS 

kinds of sugar under certain atmospheric conditions of humidity. 
The surface of the metal should be smooth and bright; the United 
States Treasury Regulations attach a penalty in case of samplers who 
neglect this precaution. When ready for making the composite sample, 
the contents of the sugar bucket are thoroughly mixed; the cans and 
bottles to receive the sample are compactly filled, labeled, and sealed, 
after which they are sent to the chemists who are to make the polariza- 
tions. 

The general rule in sampling sugar is that the package shall be 
stabbed at the middle to the center, and if this practice is conscien- 
tiously followed it will give no doubt as fair a sample as can be secured 
under the hurried conditions of discharging a cargo. There are times, 
however, when it is impossible to follow this rule. Sugar which has 
remained for a long time in storage will sometimes solidify upon the 
approach of cold weather to a hard mass of material resembling con- 
crete, a circumstance due to the evaporation of moisture and cement- 
ing together of the grain. A trier is almost useless under these con- 
ditions and such sugar is rarely sampled properly. The sugar broken, 
or chipped off, by the trier from the outside of the package is not a 
correct sample. A pickaxe is sometimes resorted to with hard sugar in 
order to open a passage for the trier; this is much better than just 
skimming the outside, but is far from satisfactory. 

To eliminate so far as possible the errors of personal equation in 
sampling, the practice of the New York Sugar Trade is for the samplers 
of buyer and seller to work alternately hour by hour; the one party in 
the interval of rest exercising a control upon the operations of the 
other. The tendencies to draw too high and too low from the package 
are thus counterbalanced and the personal errors equalized. This 
method seems as good as any that can be devised. 

The liability of change in composition of the product during sampling 
is an exceedingly important factor in the valuation of any commodity, 
and more important perhaps in the case of sugar than almost any 
other staple. Raw cane sugar upon exposure to the air may either 
absorb or lose moisture according to the conditions of atmospheric 
humidity. If the latter be very high or low, and the sugar be exposed 
to the air for any great length of time during drawing or mixing the 
sample, a considerable error may be introduced into the composition 
of the product. The buckets, which hold the samples for mixing, 
should always be kept tightly covered; this precaution will reduce the 
errors from absorption and evaporation to a large extent, although with 
present methods of sampling the errors from this source will never be 






SAMPLING OF SUGAR AND SUGAR PRODUCTS 7 

completely eliminated. On rainy days sugar is rarely sampled at the 
pier, and this is a wise precaution, considering the rapidity with which 
sugar absorbs moisture from a saturated atmosphere. No matter how 
pure the sugar, there will be absorption under such conditions, the 
amount of moisture taken up depending upon the initial dryness of the 
sugar, the fineness of the grain and the hygroscopic character of the im- 
purities present. 

If a layer of sugar be placed in a dish over water under a closed 
bell jar, it will soon absorb moisture enough to liquefy, and, according 
to the phase rule, this absorption of moisture will continue until the 
pressures of water vapor for solution and atmosphere are the same. 
Theoretically this limit is infinity, and if the dish under the bell jar be 
weighed from day to day it will be found that the liquefied sugar will con- 
tinue to attract moisture as long as one cares to follow the experiment. 

If the atmosphere is not completely saturated, the absorption of 
moisture by the sugar is less rapid, and with further decrease in humidity 
a point of equilibrium is soon reached where there is neither absorption 
nor evaporation. This point of equilibrium, which represents equality 
of vapor pressure between the moisture of the sugar and the air, is 
different for different sugars. With still further decrease in humidity 
the sugar begins to give up moisture, the rate of loss increasing as the 
percentage of saturation in the air becomes less and less. 

In the following table the percentages of moisture which different 
sugars gain or lose at 100 per cent relative humidity and at 60 per cent 
relative humidity are given, and the changes in moisture content at 
the point of equilibrium. Two grams of sugar were spread in a thin 
layer upon a watch glass and the change in weight noted after regular 
intervals of time in one case over water under a bell jar, and in the 
other case upon exposure to the open air. The temperature of experi- 
ments was 20 C. 

TABLE II 

Showing Variations in Moisture Content of Sugars 











Gain 


Change 
















first 


first 




W 'f\ 


Residual 


Kind of sugar. 


Grain. 


Polar- 
ization. 


Mois- 
ture in 
sugar. 


hour, 
100 per 
cent 


hour, 
60 per 
cent 


Total change at 
point of equilib- 
rium. 


ityat 
equilib- 


mois- 
ture at 
equilib- 










humid- 


humid- 






rium. 










ity. 


ity. 














Per cent. 


Per cent. 


Per cent. 


Per cent. 


Per cent. 


Per cent. 


Granulated 


Fine 


99.85 


0.10 


.78 


+0.03 


+0.01 (2 hours) 


56 


0.11 


Peruvian 
Porto Rico 


Large 
Medium 


98.40 
96.40 


0.35 
1.31 


.09 
.40 


-0.09 
-0.54 


-0.14 (4 hours) 
-0.73 (2 hours) 


56 
62 


0.21 
0.58 


Philippine mats. 
Cuban molasses 


Fine 
Large 


87.45 
82.75 


3.12 
4.85 


.80 
.12 


-0.68 
-1.00 


-1.25 (6 hours) 
-2.42 (24 hours) 


56 

59 


1.87 
2.43 



8 SUGAR ANALYSIS 

After the point of equilibrium was reached upon exposure of the 
above sugars to the air, no change in weight was noted as long as the 
temperature and relative humidity remained unchanged; with fluctua- 
tions in the latter corresponding gains and losses were always observed 
in the weight of the sugars. 

As to the absorption of moisture by sugars under excessive humidity, 
no relationship can be traced in the above table between composition 
and rate of absorption. The refined granulated sugar and the low- 
grade mats have equally high absorptive powers and the high-grade 
Peruvian crystals and the Cuban molasses sugar equally low absorptive 
powers. If the grain of these sugars is compared, however, it will be 
seen that the Peruvian crystals and molasses sugar of low absorptive 
power have the largest grain and that the granulated sugar and mat 
sugar of highest absorptive power have the smallest grain, so that the 
physical condition of the sugar is a very important factor in the in- 
fluences which bear upon absorption. 

As to the evaporation of moisture from sugars under diminished 
humidity, the table shows a very definite relationship between compo- 
sition and rate of evaporation, this rate being, as would be supposed, 
roughly proportional to the initial moisture content of the sugar. The 
percentage of residual moisture in a sugar at the point of equilibrium is a 
function of the hygroscopic power of the non-sugars, and is greatest 
with the sugars of lowest purity (highest molasses content). 

The point of greatest importance, in the bearing which these re- 
sults have upon the changes in composition of sugar during sampling, 
is that the gain or loss in weight through absorption or evaporation of 
moisture is most rapid at the beginning. A comparison recently made 
by the author of the changes in moisture content which sugars undergo 
upon exposure to the air shows that the relationship between time and 
loss or gain in moisture follows approximately the well-known equation 

for slow reactions, k = - log , in which a is the total change in 

t CL 3J 

moisture content at the point of equilibrium, x the loss or gain in weight 
at the end of any given time t, and k the coefficient of velocity, which 
is a constant quantity for each kind of sugar under fixed conditions of 
temperature and humidity. 

The assumption is frequently made by samplers of sugar that the 
errors from absorption and evaporation of moisture by the sample will 
equalize one another in the long run. This, however, is far from being 
the case. The percentage of moisture in the ordinary grades of raw 
cane sugar is considerably above the equilibrium point for the average 



SAMPLING OF SUGAR AND SUGAR PRODUCTS 9 

relative humidity at the port of New York. It should be stated, how- 
ever, that the loss from evaporation under the prescribed conditions of 
sampling is nowhere near as great as that in the above experiments, 
where the sugars were exposed to the open air in a thin layer. The 
error, however, does exist, and unless due care is exercised by the 
sampler there will be a very noticeable difference in the test. 

Another occasional source of error in the sampling of sugar is the 
introduction into the sample of particles of bag, basket, mat, shavings 
of barrels, etc., which are introduced from the package by the trier. 
The error from this cause is usually trifling; there are times, however, 
when it may be considerable. Such fragments of extraneous matter 
do not belong to the sugar, and it devolves upon the chemist to elimi- 
nate these as far as possible before weighing out the sugar for polariza- 
tion. In removing foreign material from sample sugar the chemist 
must carefully discriminate, however, between trash which belongs to 
the sugar and refuse which is introduced during sampling. 

In addition to removing trash, the chemist must complete the mix- 
ing of the sample. Lumps must be crushed and thoroughly incorporated 
with the rest of the sample. Even samples of sugar, which are well 
mixed at the point of sampling, must be mixed again at the laboratory 
owing to the segregation of foots at the bottom of the can or bottle. 
A neglect of such mixing of the sample in the laboratory is a cause of 
frequent differences between the results of different chemists. This 
mixing of the sample must be done with the utmost dispatch in order 
to avoid the errors due to absorption or evaporation already mentioned. 
Mixing of the sample upon paper or other porous substance which 
would absorb moisture is especially to be avoided. The method of 
mixing followed by the New York Sugar Trade Laboratory is as follows: 

When samples are brought into the laboratory during freezing 
weather, the cans or bottles are first allowed to come to approximately 
the room temperature before opening and mixing. This is done to 
guard against condensation of moisture upon the cold sugar, which 
would lower the polarization. The sugar is poured out from the can 
upon a clean sheet of plate glass, all pieces of bagging, baskets, mats, 
etc., are removed, and the sample is thoroughly mixed with a clean 
steel spatula. Lumps are reduced by means of a porcelain roller and 
incorporated with the rest of the sample. The plate glass and porce- 
lain roller are cleaned and wiped perfectly dry each time before using. 
The reduction of lumps is of greatest importance in securing uniformity 
of sample; the difference in polarization between the lumps and the 
fine portion of some sugars has been found to vary several per cent. 



10 SUGAR ANALYSIS 

The can from which the sugar was taken is then filled about three- 
fourths full, the excess of sugar upon the plate being discarded. By 
leaving a little empty space in the can, the weighing out of the sample 
by the chemist is facilitated. 

SAMPLING OF JUICES, SIRUPS, MOLASSES, AND LIQUID SUGAR PRODUCTS 

The sampling of juices, sirups, molasses, and other liquid sugar 
products involves no special difficulties provided the material be of even 
composition throughout the body of the container. A large glass or 
metal tube may serve for withdrawing samples of molasses, etc., from 
the bungholes of hogsheads, barrels, and casks, when other means are 
not available. Containers of different capacity should be sampled 
separately, and in making composite samples each individual fraction 
should be proportionate to the total amount of material from which it 
was drawn. 

The regulations of the United States Treasury Department* govern- 
ing the sampling of molasses are as follows: "In drawing samples of 
molasses, care shall be taken to secure a fair representation and an 
equal amount of the contents from each package. Packages of the 
same size shall be sampled in groups of not more than 25; samples 
from all of the packages of each group being put into a bucket. An 
accurate tally shall be kept and with each bucket shall be reported the 
number of packages the samples therein represent. The dock list 
accompanying the sample buckets shall convey the same information 
and account for every package of the mark. Packages of different size, 
although invoiced and permitted under the same mark, shall be sepa- 
rately sampled, tested, and returned for classification. Molasses dis- 
charged from tank vessels shall be sampled as it is pumped from the 
tanks, a sample of uniform quantity being drawn at either regular 
intervals of approximately fifteen minutes or for every 5000 gallons 
discharged." 

In sampling the juices from mills and diffusion batteries in sugar 
factories, various automatic sampling devices have been devised for the 
purpose of securing a sample of the main body of juice at each instant 
of tune. Coomb's drip sampler (Fig. 2) is an illustration of such a de- 
vice. A defect of such automatic contrivances is that they do not always 
give a flow of sample proportionate to the total amount of juice, f 

* Loc. cit., Art. 16. 

t A very efficient automatic liquid sampler is described by G. L. Spencer in the 
J. Ind. Eng. Chem., 2, 253; 3, 344. 



SAMPLING OF SUGAR AND SUGAR PRODUCTS 



11 



In grinding sugar cane, when it is desired to test the work of macera- 
tion or to determine the relative efficiency of each mill, the juices from 
the several sets of rollers are sampled and analyzed separately, the 
results of the work enabling the chemist to calculate the composition 
of the so-called "normal" juice or to determine the extracting power 




JUICE PIPE 




A. | TO | INCH VALVE. 
B, STRONG RUBBER TUBE CON- 
NECTING PIPE LEADING FROM"A"WITH 

C, A GLASS T-TUBEjTO CINCHES 
INSIDE DIAMETER. 

D, SHORT ARM OF T, FROM WHICH 
THE SAMPLE IS TO BE LED INTO AN 
APPROPRIATE RECEIVER. 



Fig. 2. Coomb's apparatus for sampling sugar juices. 

of each mill. This phase of sampling belongs, however, to the subject 
of sugar-house control, and the chemist is referred to the special treatises 
by Spencer, Prinsen Geerligs, Deerr, and others. 

ERRORS OF SAMPLING DUE TO SEGREGATION OF SUGAR CRYSTALS 

A serious error in the sampling of liquid sugar products is often 
occasioned by the crystallization and separation of sugar within the 
container. The deposition of sucrose crystals from molasses, and from 
maple, cane, and sorghum sirups, is an example of this; the granula- 
tion of strained honey through separation of crystallized glucose is 
another illustration. Containers of molasses, sirup, and honey fre- 
quently have a compact layer of crystals upon the bottom. Samples 
taken from the liquid surface and from the crystalline deposits of such 
products will show the greatest difference in composition. It is there- 
fore necessary to mix thoroughly the contents of a container before 
sampling. In the laboratory the crystallized sugar in a sample of 
sirup, molasses, or honey should be redissolved by gentle warming 



12 SUGAR ANALYSIS 

before beginning the analysis. This is impracticable, however, in 
sampling these products in bulk from casks or hogsheads, and the most 
that the sampler can do is to mix the contents as well as possible by 
shaking and stirring. 

The sampling of leaky containers, which allow the escape of liquid 
but retain all crystallized solids, is a fruitful cause of wide, and often 
puzzling, discrepancies in analytical results. 

ERRORS OF ANALYSIS DUE TO CHANGE IN COMPOSITION OF SAMPLES 

Owing to the liability of sugar products to change in composition 
through evaporation or absorption of moisture and through decomposi- 
tion by the action of enzymes or microorganisms, it is important that 
analyses be begun as soon as possible after samples are received. It 
happens, however, in many cases that samples must be sent for a long 
distance, or stored for a considerable time, before examination can be 
made; the long storage of products is often necessary, as in the case of 
reserve samples which are retained for the purpose of confirming an 
original analysis in the event of doubt or dispute. The sources of error 
from change in composition of samples will be briefly considered. 

Changes in Composition of Samples through Evaporation or 
Absorption of Moisture. Changes in composition due to this 
cause are prevented by hermetically sealing the samples in a perfectly 
tight container. If cans are employed all joints and connections should 
be soldered; cans of swaged metal, free from seams, are very desirable, 
but it has not been found possible as yet to manufacture these in large 
sizes. The covers should fit the cans closely and the space between 
the two should be sealed by means of melted paraffin or by a band of 
adhesive tape. In many respects wide-mouth glass bottles or jars are 
the best containers for samples; the stoppers or corks of these should 
be sealed by melted paraffin or wax. 

In a series of experiments by Stanek * upon the drying out of sam- 
ples of raw beet sugar in unsealed cans, the average daily evaporation 
of moisture for 1 month was 0.0115 per cent; when the covers of the 
cans were sealed with adhesive tape (leucoplast) the average daily evap- 
oration for 1 month was reduced to 0.0006 per cent. This loss from 
evaporation is of course not evenly distributed, but is greatest during 
the first few days. Samples of raw cane sugar kept in covered but 
unsealed cans frequently show a daily increase in polarization, through 
loss of moisture, of from 0.05 to 0.10 sugar degrees during the first days 
of storage. 

* Z. Zuckerind. Bohmen, 34, 155. 



SAMPLING OF SUGAR AND SUGAR PRODUCTS 13 

Changes in Composition of Samples through Action of Enzymes. 

Changes in composition due to this cause are frequently noted during 
the storage of plant substances, such as grains, seeds, fruits, tubers, etc. 
The change may consist in an inversion of sucrose by action of invertase, 
in a conversion of starch by action of diastase, in a modification of 
gums, hemicelluloses, etc., by action of other enzymes, or in a loss of 
sugars through respiration. It is impossible to preserve untreated 
plant materials of the above description for any length of time without 
change in composition, although the rate of change may be greatly 
retarded by cold storage. Heating the samples before storing will 
destroy enzymes, but has the disadvantage in some cases of causing 
inversion or of liquefying and saccharifying starch. Freezing the 
material may suspend enzyme action for the time, but may on the other 
hand incite changes of a different character, as in the production of 
sucrose from starch in frozen potatoes. 

When samples of fresh plant materials, which are liable to undergo 
enzymic decomposition, cannot be analyzed immediately, an effective 
method of preventing change is to weigh out a quantity of the finely 
reduced substance and preserve in a stoppered jar or bottle by the 
addition of alcohol. An excess of alcohol (over 50 per cent) destroys 
the action of enzymes, and samples thus preserved do not undergo any 
change in composition after many months' standing. 

Changes in composition through enzyme action may also occur in 
cold-strained honey. It has happened in the author's experience that 
a bottle of such honey, which contained over 20 per cent sucrose at the 
time of sampling, contained after 4 months' storage less then 10 per cent; 
in a second sample of the same honey, which was kept in a warm labora- 
tory during the same period, the sucrose was almost completely in- 
verted. The inversion was probably due to an invertase secreted by 
the bees. The action of enzymes in such products as honey may be 
destroyed by heating the sample to a temperature of 80 C. 

Changes in Composition of Samples through Action of Micro- 
organisms. The effect of yeasts, moulds, and bacteria in changing 
the composition of sugar products is well known. While the conditions 
for the development of microorganisms are most favorable in such 
dilute media as juices and musts, they may also cause deterioration in 
such concentrated products as molasses and sugar. The fermentation 
of such a thick menstruum as molasses, however, is confined entirely to 
the surface, which, through the attraction of hygroscopic moisture, be- 
comes dilute enough to favor microorganic growth. The same is true 
of raw sugars; the film of molasses coating the crystals undergoes a 



14 



SUGAR ANALYSIS 



gradual fermentation, with the result that the underlying sucrose is 
slowly dissolved and inverted. 

The changes which may occur as a result of fermentation in stored 
samples of raw cane sugar may be seen from the following polarizations 
made by Browne* at the Louisiana Sugar Experiment Station upon 
several samples of Cuban Centrifugal sugars after keeping 9 months in 
the can. 

TABLE III 
Showing Deterioration of Sugar Samples in Storage 



Number. 


April, 1904. 


January, 1905. 


Decrease. 




Polarization. 


Polarization. 




1 


96.50 


95.60 


0.90 


2 


96.05 


95.00 


1.05 


3 


95.50 


93.20 


2.30 


4 


94.20 


91.70 


2.50 


5 


97.15 


94.60 


2.55 


6 


93.95 


91.10 


2.85 


7 


94.70 


91.20 


3.50 


8 


95.00 


91.20 


3.80 


9 


95.90 


91.50 


4.40 


10 


96.80 


90.70 


6.10 


11 


96.20 


89.00 


7.20 


Average 


95.63 


92.25 


3.38 



The preservation of sugars and sugar products against micro- 
organisms by sterilization is not always desirable on account of the 
changes which the high temperature may produce in the physical and 
chemical properties of the sample. Sterilization of sugar products in 
order to be effective must be repeated upon several successive days 
owing to the extreme resistance of many spores to a single heating. 

The preservation of liquid sugar products such as juices, musts, 
sirups, etc., is sometimes effected by adding 0.05 per cent of formalde- 
hyde solution (40 per cent strength) or 0.02 per cent of mercuric chloride. 

The preservation of succulent plant substances, such as pulp of 
fruits, etc., is best accomplished by treating a weighed portion of the 
sample with alcohol in a stoppered jar or bottle, in the manner pre- 
viously described. 

Other essentials pertaining to the sampling of sugar-containing 
materials will be described elsewhere. 



* Bull. 91, Louisiana Sugar Expt. Station, p. 103. 



CHAPTER II 

DETERMINATION OF MOISTURE IN SUGARS AND SUGAR PRODUCTS 
BY METHODS OF DRYING 

THE accurate determination of moisture, in some respects the 
most simple of analytical operations, is frequently one of the most 
difficult determinations which the sugar chemist is called upon to 
make. Among the chief difficulties which confront the chemist in de- 
termining the moisture content of sugar products by the ordinary 
methods of drying, may be mentioned: (1) the very hygroscopic nature 
of many sugar-containing materials and the retention of water by ab- 
sorption or occlusion; (2) the extreme sensitiveness of some sugars, 
notably fructose, to decomposition at temperatures between 80 and 
100 C., with splitting off of water and other volatile products; (3) 
the liability of many impure sugar-containing substances to absorb 
oxygen during drying, with formation of acids and other decomposition 
products. The moisture determination is further complicated by the 
fact that many sugars, as maltose, lactose, and raffinose, retain variable 
amounts of water of crystallization under different conditions of drying, 
so that the chemist is not always certain even when no further loss 
of weight occurs in the oven as to the exact amount of moisture 
which may be retained in a hydrated form. 

In the following description of processes for determining moisture, 
methods will be given for a number of typical substances. The first 
class of methods to be described is intended only for products which 
are stable at 100 to 110 C. The determination of moisture in cane 
sugar is taken as an illustration. 

DETERMINATION OF MOISTURE IN CANE SUGAR 

Refined sugar, raw beet sugar, and the superior grades of raw cane 
sugar are dehydrated successfully by drying 2 to 5 gms. of the finely 
powered sample in a thin layer for 2 to 3 hours in a boiling-water oven 
and then heating in a special oven for 1 hour at 105 to 110 C. The 
sugar is cooled in a desiccator, and, after determining the loss in weight, 
reheated at 105 to 110 C. for another hour. The process is con- 
tinued until successive heatings cause no further loss. 

15 



16 SUGAR ANALYSIS 

For weighing out the sugar flat-bottomed aluminum, nickel, or 
platinum dishes may be used; clipped watch glasses are also con- 
venient. (See Figs. 3 and 4.) With lower-grade sugars, which con- 
tain hygroscopic salts and other impurities, the dish should be covered 
during weighing. For many purposes of dehydration low glass- 




Fig. 3 Fig. 4 Fig. 5 

Receptacles for drying sugar. 

stoppered weighing bottles (Fig. 5) are well suited, and prevent loss of 
moisture in weighing out the sample and absorption of moisture in 
weighing the dry residue. 

The official method* of the Association of Official Agricultural 
Chemists for determining moisture in sugars prescribes drying in a hot- 
water oven for 10 hours. With some sugars, more especially those of 
large grain, there is danger of occlusion and retention of water, and the 
last traces of moisture may not be expelled at 98 to 100 C. The 
method of the International Commission f upon Unification of Methods 
for Sugar Analysis prescribes in case of normal beet sugars drying at 
105 to 110 C.; this temperature is sufficient to expel the last traces 
of occluded water and is not attended with sufficient decomposition 
to affect the weight of product. The temperature of drying by this 
method should not exceed 110 C. 

For maintaining a uniform temperature of 105 to 110 C. a glycerin 
or salt-water bath may be used. The Soxhlet drying oven, shown 
in Fig. 6, is favored by many for rapid drying. The bath is filled 
with a salt solution of the desired boiling point, and closed with 
the condenser B. The material is placed in the oven and the door 
tightly clamped at A. Upon lighting a gas flame in the chimney C a 
current of air is generated through the flues at F, and, after being 
heated by the boiling salt solution, passes forward from the back of 
the drying chamber over the material to be dried. The thermometer 
T indicates the temperature of the drying chamber. By raising the 

* Bull. 107 (revised), U. S. Bureau of Chem., p. 64. 
t Proceedings, Paris Convention, 1900. 



DETERMINATION OF MOISTURE IN SUGARS 17 




Fig. 6. Soxhlet drying oven. 




Fig. 7. Wiesnegg hot-air oven with Reichert gas regulator. 



18 SUGAR ANALYSIS 

temperature gradually to 100 C. and then to 105 C. for the final dehy- 
dration, the time of -drying by the Soxhlet oven may be reduced hi 
many cases to less than an hour. A mixture of glycerine and water of 
the desired boiling point is less liable to corrode the metal of the oven 
than the salt solution, and is preferred by many for this reason. 

In case a hot-air oven is used for drying at 105 to 110 C., 
the temperature should be governed by means of a gas regulator. A 
Wiesnegg hot-air oven with porcelain inner chamber and glass door is 
a very suitable type. Illustration with Reichert gas regulator is shown 
in Fig. 7. In using hot-air ovens, where considerable variations in 
temperature are liable to occur through unequal distribution of heat, 
the exact temperature of drying should be determined by a thermom- 
eter placed near the material under examination. 

DETERMINATION OF MOISTURE IN SIRUPS, MOLASSES, MASSECUITES, 
ETC., WHEN FRUCTOSE is ABSENT OR PRESENT ONLY IN TRACES 

For dehydrating sirups, molasses, massecuites, and other sugar- 
containing substances, which contain but little or no fructose, the 
method of drying previously described may be used. The material, 
however, should first be absorbed upon dry sand, pumice stone, or 
asbestos in order to facilitate the removal of the large excess of water. 
The following provisional methods* of the Association of Official Agri- 
cultural Chemists are recommended for drying the semiliquid products 
of this class: 

Drying upon Pumice Stone. "Prepare pumice stone in two grades 
of fineness. One of these should pass through a 1-mm. sieve, while 
the other should be composed of particles too large for a millimeter 
sieve, but sufficiently small to pass through a sieve having meshes 
6 mm. in diameter. Make the determination in flat metallic dishes or 
in shallow, flat-bottom weighing bottles. Place a layer of the fine 
pumice stone 3 mm. in thickness over the bottom of the dish and upon 
this place a layer of the coarse pumice stone from 6 to 10 mm. in thick- 
ness. Dry the dish thus prepared and weigh. Dilute the sample with 
a weighed portion of water in such a manner that the diluted material 
shall contain from 20 to 30 per cent of dry matter. Weigh into the 
dish, prepared as described above, such a quantity of the diluted sample 
as will yield, approximately, 1 gm. of dry matter. Use a weighing 
bottle provided with a cork through which a pipette passes if this 
weighing cannot be made with extreme rapidity. Place the dish in. 

* Bull. 107 (revised), U. S. Bureau of Chem., p. 64. 



DETERMINATION OF MOISTURE IN SUGARS 19 

a water oven and dry to constant weight at the temperature of boil- 
ing water, making trial weighings at intervals of 2 hours. In case of 
materials containing much levulose or other readily decomposable 
substances, conduct the drying in vacuo at about 70 C." 

Drying upon Quartz Sand. " In a flat-bottom dish place 6 to 7 gms. 
of pure quartz sand and a short stirring rod. Dry thoroughly, cool in 
a desiccator, and weigh. Then add 3 or 4 gms. of the molasses, mix 
with the sand, and dry at the temperature of boiling water for from 
8 to 10 hours. Stir at intervals of an hour, then cool in a desiccator, 
and weigh. Stir, heat again in the water oven for an hour, cool, and 
weigh. Repeat heating and weighing until loss of water in one hour 
is not greater than 3 mgs. 

" Before using, digest the pure quartz sand with strong hydrochloric 
acid, wash, dry, ignite, and keep in a stoppered bottle." 

In order to prevent the occlusion or retention of water in the dried 
residue, an hour of drying at 105 to 110 C. is advisable as under 
the determination of moisture in sugar. 

Pellet's Method of Determining Moisture.* In a method of 
drying considerably employed in France, Pellet nickel capsules, 85 mm. 



3 




Fie;. 8 Fig. 9 

Pellet capsule for drying liquid sugar products. 

wide and 20 mm. deep, are used. The capsule has a circular depression 
in the center as shown in Fig. 8. Each capsule is provided with a 
cover having a small notch at the edge for the passage of a small stirring 
rod. 

The raised border of the capsule is filled with fine particles (about 
1 mm. diameter) of freshly ignited pumice stone, employing an inverted 
funnel as shown in Fig. 9. The funnel is then removed, the cover and 
stirring rod put in place, and the capsule weighed. Three grams of 
the substance to be dried are then weighed in the central depression of 
the capsule; 5 c.c. of hot distilled water are then added, and after 
* Fribourg's " Analyse .chimique" (1907), pp. 90-94. 



20 



SUGAR ANALYSIS 



stirring to dissolve all soluble matter, the capsule is slightly inclined 
on different sides to permit absorption of the solution by the pumice 
stone. The process is repeated with 3 c.c. more of hot water and then 
with 2 c.c. The contents of the capsule are then spread evenly over 
the entire bottom and dried in any suitable oven at a final temperature 
of 102 to 105 C. 

In case of products containing even traces of free acid, a drop or 
two of strong ammonia is added. The excess of ammonia is expelled 
and the amount retained in the combined form is usually too small 
to be regarded. If the free acid is not neutralized, inversion of sucrose 
may result, with the introduction of a considerable error in the deter- 
mination. 

DETERMINATION OF MOISTURE IN PRODUCTS WHICH CONTAIN FRUCTOSE 

Owing to the susceptibility of fructose to decomposition in presence 
of water at temperatures much above 70 C., the methods previously 
described are not applicable to the determination of moisture in such 
products as honey, sugar-cane molasses, jams, fruit products, and 
other similar substances. The error which may result from this source 
may be seen from the following experiment by Carr and Sanborn upon 
dehydrating a solution containing 17.75 per cent of fructose. The 
solution was dried upon pumice stone in flat-bottomed dishes at 100 C. 
in air. 



Hours of drying. 


Per cent of solids. 


1 


19.02 


2 


18.53 


3 


18.57 


4 


18.16 


5 


17.42 


6 


17.34 


8 


16 90 



It is seen that the per cent of solids after 5 hours' drying is lower 
than the actual amount of fructose taken. 

Methods of Drying in Vacuum. The susceptibility of many 
sugar products to decomposition at 100 C. in the air induced Scheibler 
in 1876 to propose drying in vacuum. Weisberg* in 1894, and Carr 
and Sanborn f in 1895, further emphasized the necessity of vacuum 
drying; and at present dehydration at low temperature under reduced 

* Bull, assoc. chim. sucr. dist., 11, 524. 

t Bull. 47, U. S. Bureau, of Chem., pp. 134-151. 



DETERMINATION OF MOISTURE IN SUGARS 21 

atmospheric pressure is the only recognized method for the accurate 
determination of moisture in fructose-containing materials. 

Carr and Sanborn's Method. Many methods have been devised 
for drying sugar solutions in vacuum. The following process is the 
one described by Carr and Sanborn,* who have employed their method 
successfully upon the widest range of materials, such as fructose solu- 
tions, honey, molasses, sorghum and maize juices, etc. 

" Select clean, fine-grained pumice stone and divide into fragments 
the size of No. 4 shot. Pass the dust through a 40-mesh sieve and 
treat separately from the larger particles. Digest hot with 2 per cent 
sulphuric acid and wash until the last trace of acid disappears from the 
wash water. Owing to the ready subsidence of the material, the wash- 
ing may be accomplished rapidly by decantation. After complete 
washing, place the material, wet, in a Hessian crucible, and bring to 
redness in a monitor or other convenient furnace. When complete 
expulsion of water is assured, place, hot, in a desiccator, or direct into 
the drying dishes if desired for use immediately. In loading the dishes 
place a thin layer of the dust over the bottom of the dish to prevent 
contact of the material to be dried with the metal; over this layer 
place the larger particles, nearly filling the dish. If the stone has been 
well washed with the acid, no harm may result from placing the dish 
and stone over the flame for a moment before placing in the desiccator 
preparatory to weighing. 

" If the material to be dried is dense, dilute until the specific gravity 
is in the neighborhood of 1.08 by dissolving a weighed quantity in a 
weighed quantity of water. (Alcohol may be substituted in material 
not precipitable thereby.) Of this, 2 to 3 gms. may be distributed 
over the stone in a dish, the area of which is in the neighborhood of 
3 sq. in., or 1 gm. for each square inch of area. Distribute this material 
uniformly over the stone by means of a pipette weighing bottle (weigh- 
ing direct upon the stone will not answer), ascertaining the weight 
taken by difference. 

" Place the dishes in a vacuum oven, in which may be maintained a 
pressure of not more than 5 in. mercury, absolute. The form of oven 
is not material so long as the moisture escapes freely by passing 
a slow current of air (dried) beneath the shelf supporting the dishes. 
The temperature must be maintained at 70 C. and the vacuum at 
25 in. 

" All weighings must be taken when the dish is covered by a ground 
plate, and the open dish must not be exposed to the air longer than 
* Bull. 47, U. S. Bureau of Chem., pp. 134-151. 



22 



SUGAR ANALYSIS 



absolutely necessary. Weighings should be made at intervals of 2 or 
3 hours." 

The following triplicate series of experiments were made by Carr 
and Sanborn upon a solution containing 17.10 per cent fructose. The 
solution was dried on pumice stone in flat-bottomed dishes at 70 C. 
under a vacuum of 25 in. 



Hours. 


Number 1. 


Number 2. 


Number 3. 


Means. 


4.. 

8 


Per cent. 
17.12 
17.11 


Per cent. 
17.09 
17.09 


Per cent. 
17.06 
17.08 


Per cent. 

17.09 
17.09 


12 


17 06 


17 05 


17 06 


17 06 


17 


17 09 


17 07 


17 07 


17.08 














Fig. 10. Carr vacuum oven. 

It is seen that constancy in weight is secured after 4 hours, and 
that no further appreciable loss takes place even after 17 hours' drying. 

An illustration of the Carr vacuum oven is shown in Fig. 10. The 
oven is provided with openings for attachment of manometer, insertion 



DETERMINATION OF MOISTURE IN SUGARS 



23 



of thermometer, and for inlet and exit of air. A gas drier contain- 
ing concentrated sulphuric acid may be used for removing moisture 
from the slow current of entering air. The detachable plate at the end 
of the oven is provided with a rubber gasket and is fastened into 
position by four screws which secure a perfectly air-tight joint. 

Browne's Method of Vacuum Drying. When one of the specially 
constructed types of vacuum drying oven is not available, the author 
has found the following arrangement (Fig. 11), which is easily con- 
structed from ordinary laboratory materials, to be perfectly efficient. 

I 




->To Vacuum Pump 




W 



Fig. 11. Browne's method of vacuum drying. 

The vacuum chamber consists of a large-mouth bottle (B) of heavy 
glass, which is supported by the shelf (S) of an ordinary water oven 
(0). The mouth of the bottle is closed by a tight-fitting rubber stopper 
(R) whose 3 holes permit the insertion, through the top opening of the 
oven, of the tubes I and E and the thermometer T. The bottle is 
easily fitted, and detached from the stopper by first withdrawing the 
shelf, the latter being shoved into position again when the bottle is in 
place. The current of air entering by tube / to the bottom of the 



24 SUGAR ANALYSIS 

vacuum bottle is controlled by a clamp pinchcock (C) and freed of 
moisture by a gas drier (D). The exit air from the vacuum bottle 
passes by the tube E to the vacuum pump or aspirator. 

For absorbing the sugar-containing liquid, asbestos in perforated 
brass or copper tubes is used. The tubes measure 9 cm. long by 2 cm. 
in diameter, and are nearly filled with freshly ignited asbestos, the 
latter being tightly packed with a rod against the sides in the upper 
half of the tube, so as to leave a central cavity. 

Each tube thus prepared is placed in a glass-stoppered weighing 
bottle of sufficient size, and the whole weighed. About 5 c.c. of the 
liquid to be analyzed are then delivered from a pipette into the cavity 
in the asbestos, the object of the cavity being to secure a rapid ab- 
sorption and even distribution of the liquid through the asbestos. 
The weighing bottle is then immediately stoppered and reweighed, the 
increase in weight being the amount of substance taken. After re- 
moving the stopper the weighing bottle with tube is placed in the 
vacuum bottle, as shown by W in the diagram, and the temperature 
raised to 70 C. During the first few hours of drying a brisk current 
of air is drawn through the vacuum bottle in order to remove the 
large excess of moisture first given off. In the last stages of the dry- 
ing the air current is decreased and the vacuum kept at about 25 in. 
At the end of a few hours the weighing bottle is removed, allowed to 
cool in a desiccator, and then restoppered and weighed. The bottle is 
then redried for a second short period to determine 
if all moisture has been expelled. 

In the weighing out of juices, sirups, sugar solu- 
tions, etc., for absorption upon pumice stone, sand, 
or asbestos, a small flask provided with a stopper 
and a rubber-bulbed pipette or medicine dropper 
will be found convenient (Fig. 12). The bottle is 
filled about two-thirds full with the sugar solution, 
which should not contain over 25 per cent solids, 
and then closed with the stopper and pipette. 
Fig. 12. Bottle for After weighing the bottle and contents, about 5 c.c. 

~ f Uquid are conve y ed b y means of the bulb P^ette 
to the absorbent material, and the flask restoppered 
and weighed. The difference in weight is the amount of sample taken. 
Honeys, molasses, jellies, and other water-soluble substances of high 
density should be diluted before this method is employed, by dissolv- 
ing a weighed amount of substance in a weighed amount of water. 
The above method of weighing samples is precluded, however, when 




DETERMINATION OF MOISTURE IN SUGARS 25 

insoluble matter is present, as with jams, sauces, and similar products. 
In such cases a weighed amount of the well-mixed sample is stirred 
with a little water until all soluble matter is dissolved and then com- 
pletely transferred to the absorbent material in the drying dish with 
help of a fine jet of water. The Pellet method of drying is especially 
convenient for products of this class. 

DETERMINATION OF MOISTURE IN SUGAR MATERIALS WHICH CON- 
TAIN WATER OF HYDRATION 

Difficulty is sometimes experienced in dehydrating sugars such as 
glucose, lactose, maltose, and raffinose, which crystallize with one or 
more molecules of water of crystallization. The principal precaution 
to be observed in drying such sugars is not to raise the temperature in 
the first stages of the process above the melting point of the hydrate, 
otherwise the sugar will liquefy to a thick viscous mass from which it 
is difficult to expel the last traces of water without decomposition. 

For drying glucose hydrate, C 6 Hi 2 O 6 + H 2 0, the sugar is spread in 
a thin layer and gently warmed at 50 to 60 C. for several hours, 
when most of the water will be removed without melting of the crystals. 
The sugar is then gradually heated to about 105 C., when the last 
traces of water will be expelled, with no evidence of liquefaction. 

For drying raffinose hydrate, Ci 8 H 32 Oi6 + 5 H 2 O, the finely powdered 
sugar is first warmed to 80 C. for several hours and then the tempera- 
ture gradually raised to about 105 C. The preliminary drying may 
be hastened greatly by heating the sugar in a vacuum oven. 

Maltose hydrate, Ci 2 H 22 On + H 2 O, gives off its water very incom- 
pletely at 100 C. under atmospheric pressure, and vacuum dehydra- 
tion is necessary. The sugar is gently heated under a strong vacuum 
at 90 to 95 C., and then after a few hours the temperature is raised 
to between 100 and 105 C. 

Lactose hydrate, Ci 2 H 22 On + H 2 O, retains its water of crystalliza- 
tion unchanged at 100 C. under atmospheric pressure. It is therefore 
customary in analytical work to estimate lactose as the hydrate. Lac- 
tose may be dehydrated, however, by gently heating the finely pulver- 
ized sugar in a strong vacuum to a temperature of 125 to 130 C. 

The method of drying devised by Lobry de Bruyn and van Laent,* 
and used by Brown, Morris, and Millar,f and also by Walker,t is to 
weigh the finely powdered sugar in a small flask and connect the latter 

* Rec. trav. chim. Pays-Bas, 13, 218. 
t J. Chem. Soc. Trans., 71, 76. 
j J. Am. Chem. Soc., 29, 541. 



26 SUGAR ANALYSIS 

by a T tube to a bottle containing phosphorus pentoxide, P 2 5 , as a 
dehydrating agent. The open branch of the T tube is connected with 
a strong vacuum; the flask containing the sugar is then placed in 
an oil bath and the temperature gently raised to the point desired. 
Walker found that lactose under these conditions, after heating 1 hour 
at 80 C. and then 1 hour at 130 C., remained perfectly white, but 
upon heating to 140 C. the sugar became tinged with brown, show- 
ing signs of decomposition. 

The method of Lobry de Bruyn and van Laent has also been suc- 
cessfully employed by Rolfe and Faxon * for determining the total car- 
bohydrates in acid-hydrolyzed starch products. In the modified appa- 
ratus of Rolfe and Faxon the T tube is provided with a three-way 
stop-cock, which allows the great excess of water first given off to be 
removed without coming in contact with the phosphorus pentoxide. 
* J. Am. Chem. Soc., 19, 698. 



CHAPTER III 

DENSIMETRIC METHODS OF ANALYSIS 

THE quantity of matter in a unit volume of substance is called the 
absolute density of that substance. If m be the mass and V the volume 

of a given substance, its absolute density D will be D = = The 

ratio between the masses of equal volumes of a substance and of some 
standard material is the relative density of that substance. Since, 
however, the masses of two bodies at any one place are proportional to 
their weights, the relative density S of a given substance may be ex- 

w 
pressed S = ^> where w and W are the weights respectively of equal 

volumes of the substance and standard material. Relative density is 
commonly known as specific gravity, and, since the standard substance 
of comparison is nearly always water, specific gravity is commonly 
defined as a number indicating how much heavier a substance or solu- 
tion is than an equal volume of water. 

The determination of specific gravity is one of greatest importance 
in the analysis of sugars; its great value consists in the fact that solu- 
tions of different sugars of equal concentration have very nearly the 
same specific gravity. The following specific gravities are given for 
10 per cent solutions of nine different sugars at 20 C. with reference 
to water at 4C.: Arabinose 1.0379, glucose 1.0381, fructose 1.0385, 
galactose 1.0379, sorbose 1.0381, sucrose 1.0381, maltose 1.0386, lactose 
1.0376, raffinose 1.0375. It will be noted that the specific gravity of 
each sugar solution is but little removed from the average 1.0380, which 
is almost the same as that of sucrose. It is possible, therefore, by means 
of specific gravity tables established for solutions of pure sucrose to 
determine very closely the percentage of dissolved substance for any 
sugar or mixture of sugars in aqueous solution. 

Units of Volume. The unit of volume universally employed in 
sugar analysis is the cubic centimeter. This unit is differently defined 
and the chemist must distinguish carefully between (1) the metric or 
true cubic centimeter, (2) the Mohr cubic centimeter, and (3) the 
reputed cubic centimeter. 

27 



28 SUGAR ANALYSIS 

The Metric Cubic Centimeter is defined as the volume occupied by 
one gram of water weighed in vacuo at 4 C., the temperature of maxi- 
mum density (D = 1.000000). At 20 C. the metric or true cubic centi- 
meter is equivalent to the volume occupied by 0.998234 gram of water 
weighed in vacuo, or 0.997174 gram of water weighed in air with brass 
weights. 

The Mohr Cubic Centimeter is defined as the volume occupied by one 
gram of water weighed in air with brass weights at 17.5 C. One Mohr 
cubic centimeter, as thus defined, is equivalent to 1.00234 metric cubic 
centimeters. 

The Reputed Cubic Centimeter, a term introduced by Brown, Morris, 
and Millar,* is defined as the volume at 15.5 C. of one gram of water 
weighed in air with brass weights. One reputed cubic centimeter, as 
thus defined, is equivalent to 1.00198 metric cubic centimeters. 

The true or metric cubic centimeter was adopted as the standard 
unit of volume by the International Commission for Uniform Methods 
of Sugar Analysis at its meeting in Paris, 1900. 

SPECIFIC GRAVITY TABLES FOR SUGAR SOLUTIONS 
Various tables have been established by different observers which 
give the specific gravity (sp. gr.) of cane-sugar solutions for different 
concentrations. These tables are expressed in several ways; they vary 
according to the temperature which is selected for the determination, 
15 C., 17.5 C., or 20 C. being usually taken, and also as to whether the 
weight of water at 4 C. (true specific gravity) is used for comparison, 
or water at 15C., 17.5C., and 20 C. (relative specific gravity). In 
expressing specific gravity it is customary to indicate the system em- 
ployed by writing the temperature of the solution above that of the 
water; thus, ^> ^ - , jf . f, etc. 

In Table IV the specific gravities of sucrose solutions at several 
concentrations are given according to the calculations of different 
authorities. 

Various formulae have been worked out for expressing the relation- 
ship between the specific gravity and percentage by weight of dissolved 
sucrose. Gerlach for specific gravity ]! has expressed the relation- 
ship by the equation 

y = 1+ 0.00386571327 Z + 0.00001414091906 z 2 

+ 0.0000000328794657176 z 3 , 

in which y is the specific gravity and x the per cent of sugar. 
* J. Chem. Soc., 71, 78 (1897). 



DENSIMETRIC METHODS OF ANALYSIS 



29 



Scheibler has recalculated Gerlach's equation for sugar solutions 
of different temperatures with the following results: 

Temperature. 

y=l + 0. 003976844 x + . 0000142764 x 2 + . 000000029120 x 3 

10 y = 1 + . 003915138 x + . 0000139524 x 2 + . 000000032728 z 3 

15 y = l + 0. 003884496 x + . 0000139399 z 2 + . 000000033806 z 3 

20 = 1 + 0. 003844136 x + . 0000144092 x 2 + . 000000030912 x 3 

30 = 1+0. 003796428 x + . 0000145456 x 2 + . 000000030664 x 3 

40 = 1 + 0. 003764028 x + . 0000143700 x 2 + . 000000035192 x 3 

50 0=1 + 0. 003722992 x + . 0000148088 x 2 + . 000000032440 x 3 

60 = 1 + 0. 0036831 12 x + . 0000155904 x 2 + . 000000026368 x 3 



TABLE IV 
Specific Gravity of Sucrose Solutions by Different Authorities 



Sucrose, per cent 
by weight. 


Balling-Brix, 
17.5 

d !7T C - 


Gerlach, 
,17.5 
d !7^ C - 


Gerlach- 
Scheibler, 

*&* 


German Imperial Commission. 


rf 15 r 
d l5~ c ' 


,20 

d To c. 





1.00000 


1.00000 


1.00000 


.00000 


0.99823 


5 


1.01970 


1.01969 


1.01978 


.01973 


.01785 


10 


1.04014 


1.04010 


1.04027 


.04016 


.03814 


15 


1.06133 


1.06128 


1.06152 


.06134 


.05917 


20 


1.08329 


1.08323 


1.08354 


.08328 


.08096 


25 


1 . 10607 


1 . 10600 


1 . 10635 


. 10604 


.10356 


30 


1 . 12967 


1 . 12959 


1 . 12999 


. 12962 


.12698 


35 


1.15411 


1 . 15403 


1 . 15448 


. 15407 


1.15128 


40 


1 . 17943 


1.17936 


1.17985 


.17940 


1 . 17645 


45 


.20565 


1.20559 


1.20611 


.20565 


1.20254 


50 


.23278 


1.23275 


1.23330 


.23281 


1.22957 


55 


.26086 


1.26086 


1.26144 


.26091 


1.25754 


60 


.28989 


1.28995 


1.29056 


.28997 


1.28646 


65 


.31989 


1.32005 


1.32067 


.31997 


1.31633 


70 


.35088 


1.35117 


1.35182 


.35094 


1.34717 


75 


1.38287 


1.38334 


1.38401 


.38286 


1.37897 



One of the best-known tables for the specific gravity of sugar solu- 
tions is that of Balling* (jfj), published in 1854, and which served 
as a basis for the better-known and more complete table of Brix, whose 
name is now almost universally given to the percentages of sugar or 
dissolved solids (degrees Brix) derived by densimetric means. Another 
well-known table is that of Gerlach f (})> published in 1863-64, and 
which served as a basis for Scheibler'st table calculated to jp- The 

* Z. Ver. Deut. Zuckerind., 4, 304. 
t Dingler's Polytech. Jour., 172, 31. 
| Neue Zeitschrift, 26, 37, 185. 



30 



SUGAR ANALYSIS 



most recent and most accurately established tables are those of the 
German Imperial Commission* upon Standards, based upon the deter- 
minations of Plato, and published in 1898 and 1900. These tables 
give the percentages of sucrose for specific gravities at y^' 15* > and 
^r. The ^? table, which was established according to the require- 
ments of the Fourth International Congress of Applied Chemistry 
(Paris, 1900), is given in the Appendix (Table 1). 

The specific gravity tables of the German Imperial Commission 
have since been enlarged by Sidersky,f so as to give the grams of sugar 
for 100 gms., and also for 100 c.c., of solution for ^ and ^ between 
10 and 30 C. and for concentrations between and 30 Brix. For 
their limited range Sidersky's tables are the most complete of any 
which have been compiled. 

Influence of Temperature upon the Specific Gravity of Sugar 
Solutions. With increase of temperature, sugar solutions expand in 
volume and the specific gravity becomes correspondingly less. The 
coefficient of cubical expansion of sugar solutions varies according to 
concentration. Josse and RemyJ give the following coefficients for 
different sugar solutions between 15 and 25 C.: 

TABLE V 

Coefficients of Cubical Expansion for Sugar Solutions 



d!5C. 


d25C. 


Concentration. 


Coefficient. 


1.02425 


.02211 


6.32 


0.0002052 


1.05100 


.04365 


12.75 


0.0002100 


1 . 10025 


.09744 


23.88 


0.0002250 


1.14782 


.14452 


33.71 


0.0002574 


1.19875 


.19500 


43.81 


0.0002896 


. 1.25110 


.24718 


5,3.37 


0.0003153 


1.30384 


1.29962 


62.39 


0.0003262 


1.33025 


1.32591 


. 66.74 


0.0003289 



The mean coefficient of expansion (7) of a solution containing p per 
cent of sucrose for temperatures between 10 and 27 C. can be found 
by Schonrock's formula with a probable error of only =b 0.000006. 

7 = 0.000291 + 0.0000037 (p - 23.7) + 0.0000066 (t - 20) 

- 0.00000019 (p - 23.7) (t - 20). 

* Z. ang. Chem. (1898), 774; Z. Ver. Deut. Zuckerind., 50, 982 to 1079. 

t " Les Densit6s des Solutions sucre"es & diflterentes Temperatures," Paris, 1908. 

% Bull, assoc. chim. sucr. dist., 19, 302. 

Z. Ver. Deut. Zuckerind., 60, 419. 



DENSIMETRIC METHODS OF ANALYSIS 31 

Knowing the value of 7, the specific gravity dt at temperature t 
can be calculated from the specific gravity dt at temperature to by the 
equation 

dt = 



In the employment of temperature corrections in densimetric 
methods of analysis, it is more customary to apply the correction to the 
percentage of sugar (degrees Brix) rather than to the specific gravity. 
The correction is to be added in case the temperature is above, and 
to be subtracted in case the temperature is below, the standard degree 
of the table (17.5 C. for the old Brix tables and 20 C. for the new 
tables of the German Commission). Lists of such corrections are 
affixed to the standard tables of specific gravities.* 

Determination of Dissolved Solids by Use of Solution Factors. 
In the investigation of starch-conversion products the percentage of 
solids in 100 c.c. of solution is frequently calculated from the specific 
gravity by means of a " solution factor." This method was introduced 
in 1876 by O'Sullivan, f who found that, when 10 gms. of maltose or 
dextrin were dissolved at 60 F. (15.5 C.) to 100 c.c., a solution of 
1.0385 sp. gr. (jf^) was obtained. Assuming that the percentage of 
dissolved substance is always proportional to the specific gravity of the 
solution (which is only approximately true), a solution containing 1 
gm. of maltose or dextrin in 100 c.c. should have a specific gravity 
of 1.00385 at 15.5 C. A solution of specific gravity d should contain 

KKOr , 1000 (d- 1.000) , r , 

at 15.5 C. - - gms. of solids. 

o.oO 

Brown, Morris, and Millar J determined the solution factors of a 

number of different sugars for a uniform specific gravity of 1.055 {575-0 
with the following results: 

TABLE VI 

Solution Factors of Sugars and Starch Conversions 

Anhydrous glucose ..................................... 3 . 825 

Anhydrous sucrose .................................... ' 3 859 

Anhydrous invert sugar ................................ 3 . 866 

Anhydrous fructose .................................... 3 . 907 

Anhydrous maltose .................................... 3 . 916 

Low starch conversion ([]/> = +149.7) .......... - ...... 3.947 

Medium starch conversion ([a] D = +173.9) .............. 3.985 

High starch conversion ([a] D = +188.6) ................ 4.000 

Dextrin ................................................ 4.206 

* Appendix, Tables 2 and 4. t J- Chem. Soc. (1876), 129. 

J J. Chem. Soc. (1897), 71, 72. 



32 



SUGAR ANALYSIS 



The solution factors of glucose, fructose, and maltose have recently 
been determined by Ling, Eynon, and Lane * with practically the same 
results as Brown, Morris, and Millar. 

For ordinary purposes Brown, Morris, and Millar recommend the 
use of the sucrose factor 3.86. A comparison of the actual grams of 
sucrose per 100 c.c. of solution with those calculated by means of this 
solution factor is given in the following table: 

TABLE VII. 



, 15.5 
15T ' 


Sucrose in 100 c.c. 
of solution. 


- Sucrose by formula, 
1000 (d- 1.0000) 






3.86 




Grams. 


Grams. 


1.0039 


1.00 


1.01 


1.0193 


5.00 


5.00 


1.0386 


10.00 


10.00 


1.0578 


15.00 


14.97 


1.0770 


20.00 


19.95 


1.0959 


25.00 


24.84 


1.1149 


30.00 


29.76 



It is seen that the employment of solution factors, while sufficiently 
accurate for dilute solutions, is attended with considerable error upon 
liquids of high concentration. The factor 3.86 is not exactly the same 
for all sugars, so that this method of estimating solids is only useful for 
approximate purposes. 

If the sugar solution be reduced to a uniform specific gravity of 
about 1.05 and a correction be made for the true density factor, the 
constant 3.86 can be employed without serious error. The correction 
is made by multiplying the results (percentages, specific rotation, re- 
ducing power, etc.) obtained by using the factor 3.86 by the value 

o o 

-W- 1 in which F is the true solution factor, according to Table VI, of 

the sugar in question. 

Contraction in Volume of Sucrose and Water Mixtures. A 

phenomenon, which has a most important bearing upon the specific 
gravity of solutions of sugars and other substances, is that of con- 
traction. If a definite quantity of sucrose, for example, be dissolved 
in a definite quantity of water, the volume of solution is always less 
than the sum of the volumes of sucrose and water taken. The same is 
also true, but to a less extent, of the mixture of sucrose solutions of 
different concentration and of sucrose solutions with water. The phe- 
* J. Soc. Chem. Ind., 28, 730. 



DENSIMETRIC METHODS OF ANALYSIS 33 

nomenon of contraction in volume during solution of sucrose and 
water has long been known. It was first observed by Reaumur and 
Petit le Medecin in 1733, and has been repeatedly studied by many 
subsequent observers.* The extent of this contraction has been vari- 
ously estimated. If x is the per cent of dissolved sucrose, the change 
in volume v according to Brixf is represented by the equation 

v = 0.0288747 x - 0.000083613 z 2 - 0.0000020513 x*. 
Scheibler J gives the equation 

v = 0.0273731 x - 0.000114939 x* - 0.00000158792 x 3 , 

according to which the maximum contraction is 0.8937 c.c. for 55.42 
gms. sucrose and 44.58 gms. water at 17.5 C. Gerlach gives the maxi- 
mum contraction as 0.9946 c.c. for 56.25 gms. sucrose and 43.75 gms. 
water, and Ziegler as 0.9958 c.c. for 56 gms. sucrose and 44 gms. water. 
According to Matthiessen and others, || the maximum contraction is 
reached at about 40 per cent sucrose; beyond this there is a decrease 
until at 60 per cent sucrose the contraction is 0; with concentrations 
above 60 per cent sucrose there is an expansion in volume. This view 
of the question is due, according to Plato, 1f to the mistaken idea that 
dissolved sucrose has the same specific gravity as the crystallized solid 
(1. 59103 |p for chemically pure powdered sucrose, 1.5892 1 for chemi- 
cally pure sucrose crystals). If we take Plato's calculated value for 
the specific gravity of dissolved sucrose in aqueous solution, 1.55626, the 
following results (Table VIII) are obtained which are in close concord- 
ance with those of Gerlach and Ziegler. The apparent change in 
specific gravity of dissolved sucrose is due to the phenomenon of con- 
traction, for which no satisfactory explanation has as yet been offered. 

* In contradiction to the results of all previous experimenters, Olizy (Bull. 
assoc. chim. sucr. dist., 27, 60) claims to have demonstrated by numerous experi- 
ments that absolutely no contraction takes place during the solution of sucrose in 
water. 

t Z. Ver. Deut. Zuckerind., 4, 308. 

| Neue Zeitschrift, 26, 37. 

Oest. Ung. Z. Zuckerind., 12, 760. 

II Lippmann, "Chemie der Zuckerarten," 1081. 

If Z. Ver. Deut. Zuckerind., 50, 1098. 



34 



SUGAR ANALYSIS 



TABLE VIII 

Showing Contraction in Volume of Sucrose- 
Water Mixtures 



Per cent 
sucrose. 


Contraction of mixture. 


For 1 kilo. 


For 1 liter. 





c.c. 

0.0 


c.c. 
0.0 


5 


1.5 


1.5 


10 


2.9 


3.0 


15 


4.2 


4.5 


20 


5.4 


6.0 


25 


6.5 


7.4 


30 ' 


7.5 


8.7 


35 


8.4 


9.9 


40 


9.1 


11.0 


45 


9.7 


12.0 


50 


10.1 


12.8 


55 


10.3 


13.4 


60 


10.3 


13.7 


65 


10.0 


13.7 


70 


9.6 


13.4 


75 


8.8 


12.6 


80 


7.7 


11.5 


85 


6.2 


9.8 


90 


4.6 


7.5 


95 


2.4 


4.3 


100 


0.0 


0.0 



The effect of mixing sucrose solutions and water is shown in the 
following table which gives the calculated contraction of mixtures of 
60 per cent sucrose solutions with water to make 100 gms. 

TABLE IX 

Showing Contraction in Volume of a 60 Per Cent Sucrose Solution and Water 



A 


B 


c 


D 


E 


F 




Solution 
taken. 


Volume of 
solution, 
17.5. 


Water 
taken. 


Volume of 
water, 
17.5. 


Volume before 
mixing, 
B+D. 


Volume after 
mixing. 


Contraction 

(E-F). 


Grams. 


c.c. 


Grams. 


c.c. 


c.c. 


c.c. 


c.c. 





0.000 


100 


100.126 


100.126 


100.126 


0.000 


5 


3.876 


95 


95.120 


98.996 


98.840 


0.156 


10 


7.752 


90 


90.113 


97.865 


97.682 


0.183 


20 


15.504 


80 


80.101 


95.605 


95.372 


0.233 


40 


31.008 


60 


60.076 


91.084 


90.789 


0.295 


50 


38 760 


50 


50.063 


88.823 


88.521 


0.301 


60 


46.512 


40 


40.050 


86.562 


86.273 


0.289 


80 


62 016 


20 


20.025 


82.041 


81.845 


0.196 


90 


69.768 


10 


10.013 


79.781 


79.670 


0.111 


95 


73.644 


5 


5.006 


78.650 


78.595 


0.055 


100 


72.526 





0.000 


72.526 


72.526 


0.000 



DENSIMETRIC METHODS OF ANALYSIS 



35 



The Specific Gravity of Impure Sugar Solutions. While the 

application of specific gravity tables established for sucrose to the esti- 
mation of dissolved substance in solutions of other sugars and car- 
bohydrates is fairly accurate, their use in the case of impure sugar 
solutions may lead to serious errors, owing to the fact that the per- 
centage of dissolved impurities for the same specific gravity differs 
from the corresponding percentage of sucrose. The errors resulting 
from this cause may be seen in Table X, which gives the concentrations 
of sucrose, tartaric acid, sodium potassium tartrate, and potassium 
carbonate for different specific gravities. When the specific gravity is 
determined after dilution with a definite amount of water, as is neces- 
sary with very thick sirups, the error in estimation of dissolved sub- 
stance is still further intensified, owing to the difference in contraction 

TABLE X 

Concentrations of Aqueous Solutions of Organic and Inorganic Com- 
pounds Compared with Those of Sucrose at 15 C. for 
the Same Specific Gravity 



Specific gravity. 


Sucrose. 


Tartaric acid. 


NaK tartrate. 


' K 2 C0 3 . 


1.0039 


Per cent. 
1 


Per cent. 
0.87 


Per cent. 
0.57 


Per cent. 
0.43 


1.0078 


2 


1.73 


1.14 


0.86 


1.0118 


3 


2.62 


1.71 


1.29 


1.0157 


4 


3.49 


2.28 


1.72 


1.0197 


5 


4.40 


2.87 


2.15 


1.0402 


10 


8.67 


5.87 


4.40 


1.0833 


20 


17.52 


12.16 


9.00 


1 . 1296 


30 


26.29 


18.38 


13.78 


1.1794 


40 


35.33 


24.73 


18.72 


1.2328 


50 


44.22 


31.10 


23.76 



TABLE XI 

Contraction on Diluting Mixtures of Solutions of Above Substances with Water to 

Reduce Degrees Brixfrom 50 to 10. Solution Taken, 100 gms., 1.2328 sp. gr., or 

81.49 c.c. Specific Gravity after Dilution, 1.0402. Temperature 15 C. 



Substance. 


Dissolved substance, 
per cent. 


Water added. 


Volume 
before mix- 
ing. 

tf =(+81.49) 


Actual 
volume 
after mixing. 

(100+C) 


Con- 
traction 

(E-F). 


Before 
dilution. 
A 


After 
dilution. 
B 


(*r)-* 

C 


D 


(1.0402) 


Sucrose 


50.00 

44.22 
31.10 
23.76 


10.00 
8.67 

5.87 
4.40 


Grams. 

400.00 
410.04 
429.81 
440.00 


c.c. 
400.34 
410.38 
430.17 
440.37 


c.c. 

481.83 
491.87 
511.66 
521.86 


c.c. 

480.67 
490.32 
509.34 
519.13 


c.c. 
1.16 
1.55 
2.32 
2.73 


Tartaric acid . 
NaK tartrate. 
K 2 CO 3 





36 



SUGAR ANALYSIS 



between sugar and dissolved impurities in aqueous solution. This can 
be seen by reference to Table X; it is also shown in Table XI, which 
gives the calculated differences in contraction obtained by diluting 
solutions of sucrose, tartaric acid, sodium potassium tartrate, and 
potassium carbonate with water to reduce degrees Brix from 50 to 10. 

Additional comparisons showing the 
differences between true dry substance 
and dry substance as calculated from 
specific gravity are given for a number 
of compounds in Table XVII. 

METHODS OF DETERMINING SPECIFIC 
GRAVITY OF SUGAR SOLUTIONS 

In the estimation of dissolved sugars 
by means of specific gravity, the tem- 
perature of the laboratory is not always 
the same as that prescribed by the table. 
It is then necessary either to bring the 
solution to the required temperature by 
artificial means or else to apply a fixed 
correction from a conversion table. The 
latter method is the more convenient 
and for ordinary purposes is sufficiently 
exact; in cases, however, where great 
accuracy is required the determination 
must be conducted under absolutely the 
same temperature conditions as speci- 
fied in the tables. 

Specific Gravity Bottle or Pycnom- 
eter. The most accurate method for 
the determination of specific gravity is 
the direct comparison of the weights of 
equal volume of water and sugar solu- 
tion. In this method some form of 
specific gravity bottle or pycnometer is used, various types of which 
are shown in Figs. 13 to 16. 

Before using the instrument the pycnometer is calibrated by de- 
termining the weight of distilled water which it contains at the tem- 
perature of comparison. The bottle is first thoroughly cleaned by 
means of dilute caustic soda and hydrochloric acid; it is then washed 
with distilled water and dried in an air bath. In case of pycnometers 



Fig. 13. Specific gravity bottle 
with thermometer. 



DENSIMETRIC METHODS OF ANALYSIS 



37 



constructed with a thermometer stem, the latter should never be 
warmed beyond the limit of graduation, which is frequently only 
40 C., otherwise the expansion of the mercury may break the in- 
strument. After drying and cooling the pycnometer is weighed. The 
bottle is next filled with distilled water, recently boiled and cooled to 
expel dissolved air. The temperature adjustment is best effected by 
filling the bottle with water a degree or so lower than the temperature 
desired; the stopper is then inserted, taking care to prevent the intro- 
duction of air bubbles, and the bottle placed in a bath of water kept 
exactly at the desired temperature. After about 10 minutes, or as 






Fig. 14 



Fig. 15 
Types of specific gravity bottles. 



Fig. 16 



soon as the thermometer of the instrument has risen to the right de- 
gree, the excess of water, exuding from the stem, or above the gradua- 
tion mark, is removed with a thin piece of filter paper, the cap is fitted, 
and the bottle wiped perfectly dry and reweighed. The increase in 
weight is the water capacity of the bottle at the desired temperature. 
The process is repeated and the average of several determinations used 
as a constant in all subsequent work. 

The pycnometer, after redrying or rinsing repeatedly with the liquid 
to be examined, is next filled with the sugar solution (observing the 
same precautions as to temperature as before) and reweighed. The 
weight of solution divided by the water capacity of the bottle gives the 
specific gravity. 

Since 20 C. has been adopted as the standard temperature* for 

* At the sixth session of the International Commission for Uniform Methods 
of Sugar Analysis (London, May 31, 1909) it was "voted unanimously to accept 
a single specific gravity table as standard, at the temperature of 20 C., which is to 
be based upon the official German table. From this, other tables may be calculated 
at other temperatures, for instance, at 15 C., 17.5 C., 30 C., etc." 



38 , SUGAR ANALYSIS 

all processes of sugar analysis, it is best to make the determination of 
specific gravity when possible at this temperature. For the specific 
gravity ^ the value for |j- must be multiplied by the density of water 
at 20 C., or 0.998234. 

For very exact work the calculation of specific gravity must be 
made upon the weights in vacuo, in which case a correction for the 
density of the air must be introduced. The method of making the cal- 
culation is as follows : Let A = apparent weight of pycnometer, B = ap- 
parent weight of pycnometer and water at t C., C = apparent weight 
of pycnometer and sugar solution at t C., d = density of water at 
t C., and s = density of air at t C. and the observed atmospheric 
pressure; then the corrected specific gravity S will be 

C ~ A B ~ C 



If the temperature of the laboratory is much above that of adjust- 
ment, the specific gravity bottle and contents must remain at rest 
until they acquire the surrounding atmospheric temperature, otherwise 
moisture will condense upon the instrument and interfere with the 
weighing. It is needless to add that the cap of the bottle must be suffi- 
ciently tight to prevent leakage of liquid displaced by expansion through 
increase of temperature. Pycnometers whose stems are to be filled 
to mark and hence allow room for expansion, as Fig. 13, are gener- 
ally to be preferred. For certain kinds of work (as for densities of 
very dilute sugar solutions) Sidersky* recommends Boot's pycnometer 
(Fig. 15), which, having a double wall with vacuum, keeps the tempera- 
ture of the solution constant for a long time. 

For highly concentrated sugar solutions, such as molasses, masse- 
cuites, or other viscous substances, the method must be somewhat 
modified, if the specific gravity of the undiluted material is desired. 
For this purpose a pycnometer with rather wide neck, of the form in 
Fig. 16, is chosen, and filled nearly to the mark with the hot material 
to be examined. To remove occluded air bubbles the bottle is placed 
for a short time in an oil or salt-water bath, the boiling point of which 
is sufficiently high to keep the material in a liquid condition. After 
cooling to 20 C. and weighing, the space between the substance and 
the graduation mark is filled with distilled water and the bottle re- 
weighed. The method of calculation is illustrated by the following 
example upon a molasses: 

* " Les Densit^s des Solutions sucrees," p. 17. 



DENSIMETRIC METHODS OF ANALYSIS 



39 



A, water capacity of pycnometer 

B, weight of molasses 

C, weight of molasses and water 
C B = weight of water added 
A (C B) = weight of water 

occupying space of molasses 
56.348 



= 50.124 gms. 
= 56.348 gms. 
= 66.536 gms. 
= 10.188 gms. 

= 39.936 gms. 



39.936 



= 1.411 sp. gr. of molasses. 



Reich* has modified the above method by filling the pycnometer 
to mark directly from a burette divided into 0.05 c.c. and noting the 




Fig. 17. Determining specific gravity by means of analytical balance. 

volume of water added. If the burette has 50 instead of as the top 
graduation, the actual cubic centimeters of molasses, etc., in the pyc- 
nometer is read off directly when the latter is calibrated to hold exactly 
50 c.c. This of course obviates a second weighing of the pycnometer, 
and, while not as accurate as the method of weighing, is sufficiently close 
for many purposes: 

A second method for determining the specific gravity of sugar 
solutions is based upon the well-known principle of Archimedes, that 
* Deut. Zuckerind., 34, 38. 



40 SUGAR ANALYSIS 

a body immersed in a liquid loses the same weight as that of the volume 
of liquid displaced. It is therefore only necessary to compare the 
losses in weight which the same body undergoes in water and in a given 
solution, in order to determine the specific gravity of the latter. The 
process may be carried out in a variety of ways; a common method 
is by means of the analytical balance. 

A sinker of heavy glass, or a bulb of glass containing mercury, is 
attached to a silk thread and weighed first in air, then in distilled water, 
and finally in the sugar solution. The method of conducting the 
weighing is shown in Fig. 17. 

The method of calculation is shown by the following example: 

A, weight of sinker in air = 25.345 gms. at 20 C. 

B, weight of sinker in water = 22.302 gms. at 20 C. 

C, weight of sinker in sugar solution, = 21.504 gms. at 20 C. 

Specific gravity of sugar solution, S = - = ^^ = 1.2622 ~ 

To convert to true density with reference to weights in vacuo, the 
above equation becomes S $> = (d s) -: ~ + s, in which d = den- 
sity of water at t, and s = density of air at t and the observed atmos- 
pheric pressure. 

Mohr's Specific Gravity Balance. The specific gravity balance 
of Mohr, as improved by Westphal, and hence frequently called the 
Westphal balance, makes use of the principle of the sinker described in 
the previous section. The construction and operation of the balance 
are best understood from Fig. 18. The beam (AC) of the balance is 
pivoted at B and between the pivot and point of suspension (C) is 
divided by notches into 10 equal parts. The distance between each 
division of the beam is ordinarily made exactly 1 cm. The balance, 
as usually supplied, has a specially constructed thermometer sinker . 
(Reimann's thermometer body) which by careful grinding of the lower 
end is made to displace exactly 5 gms. of distilled water at 15 C. 
The sinker is attached by means of a fine platinum wire to the brass 
hanger H, the combined weight of sinker, wire, and hanger being made 
to equal exactly 15 gms. Before using, the balance is first adjusted 
by hanging the sinker from the arm and regulating the screw S until, 
when the beam is at rest, the pointers of the arm and support at A 
exactly coincide. If the sinker be now submerged in distilled water 
at 15 C., it will require 5 gms. at the point of suspension C to re- 
store equilibrium. The standard weight for Reimann's thermometer 



DENSIMETRIC METHODS OF ANALYSIS 



41 



body is therefore 5 gms., and in determining the specific gravity of 
solutions heavier than water this weight must always be hung from 
the point C. To obtain the decimal figures of the specific gravity, 
weights are added to the notches on the beam until the pointers indicate 
equilibrium. The first decimal figure is obtained by means of a dup- 
licate 5-gm. weight, which is moved from notch to notch on the beam 





Fig. 18. Mohr's specific gravity balance (indicating 1.1267 sp. gr.). 

until the correct decimal is secured; the second decimal figure is ob- 
tained by means of a 0.5-gm. weight, the third decimal figure by a 
0.05-gm. weight, and the fourth decimal figure by a 0.005-gm. weight. 
The specific gravity is then read from the scale divisions of the beam 
in the order of the diminishing weights. The method of reading is 
easily understood from Fig. 19. 

In using the Westphal balance the temperature of the solution is 
read from the thermometer of the sinker. In case of turbid or dark- 



42 



SUGAR ANALYSIS 



colored solutions which render the reading of this thermometer difficult 
or impossible, the temperature is read either by carefully drawing up 
the thermometer body until the top of the mercury column is visible, 
or, better, by means of a larger thermometer immersed in the solution. 
Thermometers and cylinders of special form have been constructed for 
taking specific gravities, a type of which is shown in Fig. 20. 




0.9570 




1.2646 




1.4826 




Fig. 19. Method of reading West- 
phal balance. 



Fig. 20. Special cylinder and ther- 
mometer for Westphal balance. 



Hydrometers. A third method of determining the specific grav- 
ity of sugar solutions, and the one most commonly employed in technical 
operations, is by means of the hydrometer. In its usual form (Fig. 21), 
this instrument consists of a hollow glass body terminating at its lower 
extremity in a bulb (which can be weighted with mercury or shot) 
and at its upper extremity in a hollow slender stem, inside of which 
a paper scale is sealed. If this instrument is allowed to float in a 
solution, the weight of liquid displaced is equal to the weight of the 



DENSIMETRIC METHODS OF ANALYSIS 



43 



floating hydrometer. If placed in solutions of different concentration, 
the stem will sink to varying depths; that point upon the scale which 
is level with the surface of the liquid indicates the density or percentage 
for the given concentration and temperature. It is in this 
manner that hydrometers are calibrated and standardized. 

In actual practice a hydrometer scale is standardized 
at only a few of its points, the intermediary divisions 
being determined by interpolation. The method of inter- 
polation will depend upon whether the scale is to indicate 
specific gravity or direct percentages. 

The specific gravity D of a solution is equal to the 
weight W of the hydrometer divided by the volume V of 



Then V 



W 



-~ - If the scale is to be 



the part submerged. 

graduated for specific gravity the numerical divisions will 
proceed in arithmetical progression, such as 1.00; 1.05; 
1.10; 1.15; 1.20, etc. The difference between the volumes 
of the hydrometer for any two scale divisions will give 
the volume v between these divisions; letting r = half the 

diameter of the stem, then ^ = the distance between the 

two divisions. The relationship between the stem divi- 
sions of a hydrometer weighing 20 gms. and with a cross 
area of stem (irr 2 ) equal to 0.2 sq. cm. can be seen from 
the following table : 

TABLE XII 
Showing Hydrometer Scale Divided According to Specific Gravity 



Specific gravity 
(D). 


Volume of part 
submerged 

(-} 

\D) 


Volume between 
divisions 

(). 


Distance between 
divisions 

(o- 2 )' 




c.c. 


c.c. 


cm. 


1.00 


20.000 










0.952 


4.76 


1.05 


19.048 










0.866 


4.33 


1.10 


18.182 










0.791 


3.96 


1.15 


17.391 










0.725 


3.63 


1.20 


16.666 










0.666 . 


3.33 


1.25 


16.000 










0.615 


3.08 


1.30 


15.385 







Fig. 21. 
Hydrometer. 



44 



SUGAR ANALYSIS 



It will be noted that as the specific gravity increases the distance 
between the scale divisions decreases. Owing to the great labor in- 
volved in the making of calculations and measurements, the division 
of a hydrometer scale harmonically is accomplished in practice by means 
of a dividing engine. 

In the graduation of a hydrometer scale for indicating direct per- 
centages of sugar, the distance between the scale divisions is much more 
uniform. The relationship is best seen from the following table, where 
a hydrometer of 20 gm. weight and 0.2 sq. cm. cross area of stem (wr 2 ) 
was used as before. 

TABLE XIII 
Showing Hydrometer Scale Divided According to Sugar Percentage 



Percentage sugar. 


Specific gravity. 


Volume of part 
submerged 


Volume between 
divisions 

w. 


Distance between 
divisions 

C-V 

\Q.2j 


0.00 


1.00000 


c.c. 

20.000 


c.c. 


cm. 








0.772 


3.86 


10.00 


1.04014 


19.228 












0.766 


3.83 


20.00 


1.08329 


18.462 












0.758 


3.79 


30.00 


1 . 12967 


17.704 












0.747 


3.74 


40.00 


1 . 17943 


16.957 












0.733 


3.67 


50.00 


1.23278 


16.224 












0.719 


3.60 


60.00 


1.28989 


15.505 







The maximum difference between the length of the scale divisions 
in Table XII is 1.68 cm., while for the same range of specific gravity 
the maximum difference of Table XIII is only 0.26 cm. For a hydrom- 
eter graduated to read direct percentages of sugar, it is customary in 
practice to establish only a few points upon the scale by means of 
sugar solutions of known concentration, and then divide the intervals 
between these points into equal subdivisions. While this method is 
not absolutely accurate, the errors of division are less .than the probable 
errors of observation. 

The construction of a hydrometer to read direct percentages of 
sucrose is first due to Balling. The scale of this instrument, as after- 
wards recalculated by Brix, constitutes the form at present in most 
general use. The divisions of the scale are usually called degrees 
Balling or degrees Brix, as the case may be; the differences between 



DENSIMETRIC METHODS OF ANALYSIS 



45 



the two scales are so slight that they have no significance in practical 
work. 

The Brix hydrometer* or spindle is supplied in a variety of forms. 
For approximate work spindles are used with graduation of 0-30, 
30-60, and 60^90, and divided either into 0.5 or 0.2 degree. The forms 
in most common use, however, have only a range of 10 degrees, 0-10, 
10-20, 20-30, 30-40, etc., graduated into 0.1 degree. For greater 
accuracy a third form of spindle has been made with a range of only 
5 degrees, 0-5, 5-10, 10-15, 15-20, etc., and graduated into 0.05 degree. 
With the help of a spindle for only approximate work, the choice of 





Fig. 22. Floating Brix 
spindle. 



Fig. 23. Winter's cylinder for taking 
specific gravity. 



the particular hydrometer for the finer reading will be facilitated. 
The accuracy of the spindle is of course the greater, the smaller the 
diameter of the stem and the consequently larger interval between the 
scale divisions. 

In determining specific gravity by means of the hydrometer, a tall, 
narrow cylinder is usually employed for holding the liquid to be ex- 
amined. The spindle is carefully lowered into the solution in such a 

* The term saccharometer, which is sometimes applied to a hydrometer indi- 
cating percentages of sucrose, is unfortunate, owing to the confusion with the word 
saccharimeter, of entirely different meaning. 



46 



SUGAR ANALYSIS 



10 



BRIX. 

10 



H 



15 



115 



17 



18 



way that the surface of the stem above the liquid is not moistened. 

Care should also be exercised that the instrument floats freely and does 
not touch the bottom or walls of the cylinder. The 
reading is made by bringing the eye upon a level with 
the surface of the solution and noting where the border 
line intersects the scale; the film of liquid drawn up 
around the stem by capillarity should be disregarded. 
The reading of the spindle, for example, in Fig. 22, is 20 
and not 17. The scale of the hydrometer is read with 
greater ease when the surface of the liquid is level with 
the brim of the cylinder. Cylinders of the form designed 
by Winter (Fig. 23) are convenient for this purpose; any 
overflow of liquid displaced by the spindle is caught in 
the circular trough. 

The same attention must be paid to temperature 
when the hydrometer is employed as in other methods 
of determining specific gravity. The Brix spindle is cal- 
ibrated at 17.5 C., and unless the solution be of this 
temperature a correction must be applied. A table of 
temperature corrections for degrees of the Brix scale is 
given in Table 4 of the Appendix; these corrections are 
to be added to readings made above 17.5 C. and sub- 
tracted from those made below. 

Brix hydrometers are sometimes fitted with ther- 
mometers, a form of which modification is shown in 
Fig. 24, The advantages of this construction disappear 
somewhat when working with turbid liquors, which ren- 
der the reading of the thermometer difficult or impos- 
sible. For general purposes the temperature of the 
solution is best taken by means of an accurately stand- 
ardized special thermometer. 

Volquartz* has constructed a Brix spindle with a 
correction scale, the mercury of the thermometer in the 
stem indicating, instead of temperature, the correction 
necessary to be added to the scale reading. The method 

Fig. 24. Brix of operation may be seen from Fig. 25. The spindle in 
spindle with the illustration indicates 10.0 Brix; the mercury of the 
thermometer, thermometer marks 2.7; the reading corrected to 17.5 C. 

is, then, 10.0 + 2.7 = 12.7 Brix. If the mercury is below the mark 

(17.5 C.), the correction must be subtracted. 

* Z. Ver. Deut. Zuckerind., 46, 392. 



\ 



23 



DENSIMETRIC METHODS OF ANALYSIS 



47 



80 0. 



Vos^tka* has constructed a Brix spindle with movable scale, which 
after adjustment to the temperature of the sugar solution gives the 
true reading directly. 

For determining the Brix of dilute sugar solutions, an operation of 
considerable importance in exhausting filter-press cake ("sweetening 
off"), a variety of spindles known as " sweet- water " ^-^ 

spindles has been constructed. These hydrometers 
have a large body with a thin stem, so that the read- 
ings can be easily made to 0.1 degree. The sweet 
water as it comes from the filters has usually a tem- 
perature of 60 to 80 C., and, to prevent the delay 
incident to cooling the solution to 17.5 C., sweet- 
water spindles are often calibrated at high tempera- 
tures. One form of such spindle is graduated to read 
degree Brix in water at 75 C., and 5 Brix in a 
5 per cent sugar solution of the same temperature; 
such a spindle cannot of course be employed at other 17 5 - 
temperatures, so that its usefulness is somewhat 
limited. 

Another form of sweet-water spindle (Fig. 26) is 
graduated from to 5 Brix in the normal way. Be- jj 
low the mark the divisions are continued in the 
same manner, the result being a double scale with 
the division in the middle. At 17.5 C. the read- 
ings of the upper scale give the true Brix; at temper- 
atures above 17.5 C., sweet waters will read less than 
the true Brix. At 70 C. a 5 per cent sugar solution 
reads on the spindle, a 4 per cent solution 1, a 



perature correc- 



3 per cent solution 2, a 2 per cent solution 3, a Fig. 25. Volquartz 
1 per cent solution -4, and pure water -5. The spindle with tem- 

. , 

true Brix can be determined for any temperature by 

means of a correction table; determinations by this 
instrument can always be controlled by cooling the solution to 17.5 C. 
Still another form of sweet-water spindle has been devised by 
Langen. This spindle (Fig. 27) contains within its body a thermom- 
eter graduated from 30 to 70 C. The graduated scale in the stem 
of Langen's spindle differs from other forms, however, in not giv- 
ing Brix degrees, but in simply indicating the thermometer reading 
for each division to which the hydrometer will sink in pure water. 
If placed, for example, in distilled water of 30 C., the instrument 
* Z. Zuckerind. Bohmen, 27, 689. 



48 



SUGAR ANALYSIS 



GO 



50 



40 



will sink to the division 30 on the stem, and in water of 70 C. to 
the division 70; in other words, the thermometer and scale of the 
spindle will give the same readings between 30 and 70 when the 
instrument is floated in distilled water. When the 
spindle is placed in a sweet water, the reading of ther- 
mometer and scale will no longer agree. The 
spindle necessarily sinks to a lesser depth than 
in water, and the scale of the stem gives a dif- 
ferent reading from that of the thermometer, 
the difference between the two being propor- 
tional to the concentration of solution. In 
sweetening off, it is only necessary to observe 
the readings of thermometer and scale; the 
differences between these decrease as the ex- 
traction proceeds, until with the coincidence 
of the two readings complete exhaustion is 
indicated. 

Another form of hydrometer which is fre- 
quently used in the sugar factory, but to a 
much less extent in the sugar laboratory, is 
that of Baume. This instrument is standard- 
ized by means of common salt; the point at 
the top of the stem is obtained by means of 
distilled water, and the 15-degree mark by 
means of a 15 per cent salt solution. The 
interval between these two divisions is then 
divided into 15 equal parts, this graduation 
being extended downwards on the scale as far 
as desired. Unfortunately, in the early instru- 
ments the temperature of the water and the 
specific gravity of the salt solution were not 
correctly obtained, so that the values of the 
Baume' scale divisions have been variously re- 
ported by different authorities. The so-called 
Fig. 26. "old" Baume degrees, as calculated by Brix, 
Sweet-water are s tiU used in European countries in the 
e> commercial analysis of molasses * notwith- 
standing the fact that Gerlach as long ago as 
1870 showed the incorrectness of the formulae employed by Brix in his 
calculations. 

* Friihling's " Anleitung," p. 74. 



Fig. 27. 

Langen's 

sweet-water 

spindle. 



I 



DENSIMETRIC METHODS OF ANALYSIS 49 



Gerlach found as the specific gravity of a 15 per cent salt solution 
at 17.5 C., 1.11383. The volume of a Baume" spindle up to the 
mark, in terms of the volume of a single scale division, is then equal 

1 11383 X 15 
to -T r = 146.78. The specific gravity S corresponding to any 

J..J.IGOO 1 

scale division N of the Baume scale can then be calculated by the 
formula S = ' ^ It is by use of this formula that the so- 

called " new " Baume degrees have been determined. The relationship 
between percentages of sugar, or degrees Brix, specific gravity and the 
new and old degrees Baume, is shown in Table 3 in the Appendix. 







CHAPTER IV 

PRINCIPLE AND USES OF THE REFRACTOMETER 

A SECOND method of estimating the percentage of sugars in solution 
is by means of the refractive index. The general applicability of this 
method, as in the case of specific gravity, depends upon the fact that 
solutions of all sugars of equal concentration have nearly the same 
index of refraction. 

Law of Refraction. If a beam of light from one medium, such as 
air, fall at an inclined angle upon the surface of a second medium, such 
as water, it will be found that the beam upon entering the second 
medium is bent or deflected from. its original course. A good example 
of this phenomenon, which is called refraction, is the bent appearance 
of the oar of a boat when seen obliquely under water. There is a 
general law of refraction for all transparent liquids and solids which 
may be stated as follows: For two given media and the same ray of 
light (same wave length), the ratio of the sine of the angle of incidence 
to the sine of the angle of refraction is always a constant quantity for 
the same temperature. 

In Fig. 28 m and m' are two media; PP f is drawn perpendicular to 
the dividing surface FF'. Let a beam of light pass through m in the 
direction LO; a part of the beam at the point of the surface is re- 
flected in the direction OL'; another part of the beam entering m' is 
refracted in the direction OS. The angle LOP which the falling ray 
makes with the perpendicular is the angle of incidence, or i; the angle 
SOP' which the refracted ray makes with the perpendicular is the 

angle of refraction, or r. The ratio = n is called the index of 

smr 

refraction. This ratio in Fig. 28 is represented by -r- 



line cd 



sin i 



The ratio - - is also that of the velocities of light in the two 
smr 

media. If v is the velocity of light in m and v' the velocity in m f , then 

S1T1 ? W 

n = - - - If the refracted ray is bent toward the perpendicular 
sin i v 

as in Fig. 28, the velocity v' is smaller than v, and the medium m' is 
called of greater optical density than m. Optical density must not be 

50 



. 

nnnfiiser 



PRINCIPLE AND USES OF THE REFRACTOMETER 51 

confused with material density, since the two expressions do not at 
all correspond. 

If the ray of light in Fig. 28 pass from a denser medium m! into a 
rarer medium m in the direction SO, it will be refracted in m in the 

direction OL. In this case the index of refraction is - .) which is the 

sin i 

reciprocal of the index for light passing in the opposite direction. The 
refractive index varies with the wave length of the light, increasing 







Fig. 28. Illustrating law of refraction. 

from the red towards the violet end of the spectrum. From this it 
follows that when ordinary light is refracted it is decomposed into light 
of the different prismatic colors; this unequal refraction for light of 
different wave lengths is called dispersion. 

Measurement of Refractive Index. The refractive index of a 
solution can be measured in a variety of ways. One of the simplest 
methods, which is of more value for demonstration than for accuracy, 
is by means of the refractometer trough. This apparatus, shown in 
Fig. 29, consists of a semicircular trough, the inner curved surface of 
which is divided into degrees. The side of the trough corresponding 
to the diameter of the circle consists of a plate of glass which is made 
nontransparent, excepting a narrow perpendicular slit at the center c. 
If the trough be filled partly with a solution and a beam of light fall 
upon the glass, that part of the beam passing through the slit above 



52 



SUGAR ANALYSIS 



the surface of the liquid will mark the angle of incidence and that part 
passing below the surface will mark the angle of refraction. In the 







Fig. 29. Measuring refractive index by refractometer trough. 

illustration, where water is used, these angles are 60 degrees and 40 
degrees respectively. 

sin 60 0.8660 



sin 40 0.6428 



= 1.34 or n, the approximate index of refraction. 




Fig. 30. Illustrating principle of total reflection. 

In the construction of refractometers for more accurate measure- 
ments, instrument makers generally employ the method of total re- 
flection. The principle of this method can be understood from Fig. 30. 

Let m and mi be two media, such as glass and water, of which m is 



PRINCIPLE AND USES OF THE REFRACTOMETER 53 

the more optically dense, the dividing surface being SF. The beams 
of light which fall from the source L upon SF at various angles are 
refracted, in mi in different directions. The beam LO J_ SF is not re- 
fracted and proceeds in the same direction; the beam Lo, making the 
angle of incidence i, is refracted in the direction ot, making the angle of 
refraction r; in the same way Loi is refracted to Oiti, and Lo z to O-&L. 
As the angle of incidence for the falling beam increases, there finally 
comes a point at o 3 where the refracted ray o 3 Z 3 coincides with the sur- 
face SF, and the angle of refraction r 3 = 90 degrees. If the angle of 
incidence be increased beyond i 3 to it, the beam which previously was 
only partly reflected is totally reflected in the direction 2 4 , and there is 

no refraction in m\. Since - , the index for the beam before total 

smr 3 

a .- i sin 1*2 sini , . . ^o 

reflection, equals-; > etc., = -: = n, and since sin r 3 = 90 = 1, it is 

sin r 2 sin r 

evident that for the border line of total reflection sin i = n. In other 
words, the sine of the angle of incidence for the border line of total re- 
flection is equal to the refractive index. It is seen from the diagram that 
total reflection can only take place when light passes into an optically 
rarer medium. 

For absolute measurements the refractive index of a substance is 
referred to a vacuum. Since, however, the absolute index of air is 
only 1.000294, refractive indices referred to air are sufficiently exact 
for most purposes. In the case of three media such as air, glass, and a 
liquid, if the index from air to glass be N ag and from glass to liquid N g i, 
then the index from air to liquid N a i N ag X N g i. The sine of the 
angle of incidence for the border line of total reflection between glass 
and a given liquid, multiplied by the index of refraction between air 
and glass, will give the index of refraction for the liquid with reference 
to air. 

ABBE REFRACTOMETER 

The best general instrument for determining the refractive index of 
sugar solutions is that of Abbe (Fig. 31). The essential part of the 
Abbe refractometer consists of two flint-glass prisms A and B of refrac- 
tive index n D = 1.75, each cemented into a metal mounting. To open 
the prisms the latter are rotated on their bearings to a horizontal posi- 
tion with the prism B uppermost; the clamp v is then released and prism 
B swung open on its hinge C. A few drops of the solution to be ex- 
amined are then placed upon the polished inner surface of the fixed 
prism A next to the telescope, and prism B, whose inner surface is 



54 



SUGAR ANALYSIS 







Fig. 31. Abbe refractometer. 



PRINCIPLE AND USES OF THE REFRACTOMETER 55 

ground, brought slowly back and clamped as before. The instrument 
is then swung into an upright position and light reflected from the 
mirror R upon the surface of the lower prism. 

In the following diagram (Fig. 32) FDE and ABC are longitudinal 
sections of the two prisms in an Abbe refractometer between whose 
hypotenuse surfaces FE and AB (separated by about 1.5 mm.) is the 



P' 




Fig. 32. Illustrating principle of Abbe refractometer. 

film of liquid to be examined. The beams of light passing from L 
through the lower prism to the surface of the solution AB are re- 
fracted or totally reflected, according to the refractive index of the 
liquid. As shown in the diagram the beams which fall upon the hypot- 
enuse surface AB at a less inclination than the line 10 undergo re- 
fraction in the liquid, and, passing through the upper prism, the sets 
of parallel rays s, s', s", . . . ,u,u', u", . . . , etc., are condensed by 
the objective K of the telescope upon the field XY. The beams in the 



56 SUGAR ANALYSIS 

prism parallel to 10 are refracted along the surface BA and the beams 
of greater inclination totally reflected; since these beams do not reach 
the surface of the upper prism, a part of the field XY remains in 
shadow. 

The telescope of the refractometer (F in Fig. 31) is attached to a 
sector S and the prisms to a movable arm J (the alidade) which carries a 
magnifying lens L. By moving the alidade until the intersection of the 
reticule in the telescope field (Fig. 32) cuts the dividing line between 
the bright and dark portions of the field, the refractive index can be 
read directly upon the scale of the sector by means of the lens. 

The relation between the angles of incidence and refraction of light 
between air and prism, and prism and liquid, in the Abbe refractometer 
may be understood from Fig. 32. Let PP' be drawn J_ to the end planes 
BC and DE of the double prism, and hh f be drawn J_ to the hypotenuse 
planes AB and EF. 

Let a = angle of incidence from air and 
b = angle of refraction in glass; then 

r = n for prism, which for the flint glass of the Abbe mstru- 
sm b 

ment is about 1.75. 
Let r = angle of prism. 

a! = angle of incidence in glass upon surface AB and 

b r = angle of refraction in liquid = 90 degrees for border line of 

total reflection. 

In A BOIZ. r + Z BOI + Z BIO = 2rt. Z's; 
Z BOI + Z a' + Z BIO + Z b = 2 rt. Z's; 
whence r = a' + b. 

By way of illustration the following values are given for a, b, and r, 
with water as the liquid between the prisms: 
a = 18 32'. 
b = 10 28'. 
r = 60 00'. 
sin a 0.3179 

E& = al8l7 = L75 = * for air to pnsm ' 

a' = 60 - 10 28' = 49 32'. 

sin a' 0.76 

T7 = = = 0.76 = n for glass of prism to water. 

1.75 X 0.76 = 1.33 = n for air to water. 



PRINCIPLE AND USES OF THE REFRACTOMETER 57 

Each division, therefore, upon the sector of the refractometer rep- 
resenting refractive index is equal to the sine of the angle of incidence 
in the prism for the border line of total reflection multiplied by the re- 
fractive index of the prism. Since total reflection can take place only 
when light passes from an optically denser to a rarer medium, the 
capacity of the refractometer is necessarily limited to solutions of 
smaller refractive index than 1.75. 

A second important feature of the Abbe refractometer is the com- 
pensator. The function of this is to correct the dispersion which white 
light undergoes in the double prism. Without the compensator the 
border line between the light and dark parts of the field, owing to the 
unequal refraction of light of different wave lengths, assumes the ap- 
pearance of a band of prismatic colors, which it is impossible to use for 
purposes of adjustment. 

The compensator of the refractometer is placed in the prolongation 
of the telescope tube between the objective and the double prism. It 
consists of two similar Amici prisms, such as are used in a direct-vision 
spectroscope, and which give no divergence for the yellow D line of the 
spectrum (i.e., the emergent D rays are parallel with the optical axis). 
The two prisms are rotated simultaneously in opposite directions by 
means of the screw head M (Fig. 31). 

Trapezoidal sections of the two Amici prisms are shown in Fig. 33. 
Each prism consists of a combination of two crown-glass prisms, with 




Fig. 33. Illustrating principles of compensator. 

a third right-angled flint-glass prism between them in the manner 
shown. If a beam of white light LT fall upon the surface of the first 
prism AB, it is decomposed into its colored constituents, as shown by 
the divergent broken lines. In their passage through the prism the red 
rays are refracted least and emerge at r, the yellow rays emerge at y, 
and the violet rays, which are refracted most, emerge at v. If the light 
emerging from the prism ABDE now enter a second prism A'B'D'E' 
similarly placed to the first prism (their refracting edges A and A' being 
parallel and on the same side of the optical axis LL'), the colored rays 
will emerge from the second prism at the points r', y', and v' respectively, 
the angle of dispersion for any two differently colored rays being twice 
that for the single prism ABDE. 



58 SUGAR ANALYSIS 

If the two Amici prisms be now rotated in opposite directions 
around the optical axis LL', the dispersion of the compensator will 
diminish until, when each prism has rotated 90 degrees (the difference 
from the previous position being 180 degrees), the dispersions of the 
two prisms neutralize one another and the dispersion of the compen- 
sator is zero. In this position the refracting edges A and A f of the 
two prisms will again be parallel, but on opposite sides of the optical 
axis LL'. If we now imagine the direction of the colored rays through 
the two prisms to be reversed, we have an exact representation of the 
work performed in the compensator. The band of colored light from 
the double prism of the refractometer, passing in the direction L'L, 
emerges at T as a colorless beam, and the bright and dark halves of the 
field are sharply divided. By rotating the screw head the compen- 
sator can be given an equal but opposite dispersion to that of the liquid 
examined for any value from zero up to twice the dispersion of a single 
Amici prism. 

After setting the compensator to the point where the colored bands 
disappear, the reading of the scale upon its drum (T, Fig. 31) enables 
one to calculate the dispersion of the liquid examined for the F and C 
rays of the spectrum, the mean dispersion np nc (difference in refrac- 
tive index for the F and C rays) being determined with the help of a 
special table supplied with the instrument. 

Duplicate readings upon the Abbe refractometer with a sharp defi- 
nition of the border line should agree within two places of the fourth 
decimal. After each determination the prisms should be cleaned with 
wet filter paper and then wiped dry with a piece of soft linen. 

Illumination of Abbe Refractometer. For illuminating the refrac- 
tometer ordinary daylight may be used, in which case the instrument 
should not be placed in the direct light of the sun. Since, however, 
daylight (especially in winter) is of variable intensity, and upon dark 
days not strong enough for the examination of deep-colored solutions, 
it is better on the whole to use artificial light of constant intensity. 
An incandescent electric lamp or Welsbach gas burner is a most con- 
venient method of illumination. A large sheet of cardboard, placed in 
front of the instrument so as to shield the light from the upper prism 
and from the eye of the observer, will protect the field of vision from 
the disturbing influences of extraneous light and increase to a marked 
extent the sensibility of adjustment. 

Regulation of Temperature in Abbe Refractometer. The re- 
fractive index of sugar solutions, as of all other substances, varies with 
the temperature, the index decreasing as the temperature rises. It is 



PRINCIPLE AND USES OF THE REFRACTOMETER 59 

therefore important in all refractometer work that the temperature be 
kept constant during the course of observation. In the Abbe refrac- 
tometer shown in Fig. 31 water of constant temperature is allowed to 
circulate in the direction of the arrow through the metal casings which 
surround the prisms; a thermometer screwed into the upper casing 
indicates the temperature. 

Zeiss Spiral Heater and Water-pressure Regulator. A conven- 
ient piece of apparatus for controlling the temperature of refractom- 
eters is the Zeiss spiral heater and water-pressure regulator. This 
apparatus shown in Fig. 34 consists of a constant-level reservoir A 
connected by rubber tubing to the water supply and attached to a 
sliding frame which can be adjusted to different heights. The water 
passes from the reservoir to the spiral heater, which is placed upon a 
level below the refractometer. The heater consists of about 12 feet 
of copper tubing wound in a spiral and inclosed in a metal jacket which 
is heated by a Bunsen burner. The water flows from the heater upward 
to the prisms of the refractometer and thence to a constant-level vessel 
B, from which the overflow escapes to a drain. The water, which 
should not flow too slowly, is first warmed to the approximate tem- 
perature by regulating the flame of the burner; the exact adjustment 
is then made by varying the speed of the flow, which is done by raising 
or lowering the pressure reservoir on its sliding frame. In this manner 
the temperature can be maintained for hours within 0.1 C., provided 
of course that no variations take place in the temperature of the main 
water supply. 

Instead of the Zeiss heater a large insulated heatable tank holding 
50 to 100 liters of water may be used. 

Testing the Adjustment of the Abbe Refractometer. The ad- 
justment of the Abbe refractometer can be tested by means of liquids 
or glass test plates of known refractive power. Freshly distilled water 
free from air (n = 1.33298) is convenient for testing the lower divi- 
sions of the sector scale; monobromonaphthalene (n= 1.658) is con- 
venient for testing the upper part of the scale; the latter substance 
unless freshly prepared usually requires to be redistilled (boiling point 
277 C.). The Abbe instrument is supplied with a glass test plate 
whose index is marked upon the upper ground surface. The method 
of using the plate, which can be applied to any transparent solid, is 
that of grazing incidence (explained in detail under the immersion 
refractometer). 

In using the test plate the instrument is reversed as shown in Fig. 35, 
the double prism spread open, and the polished surface of the plate 



60 



SUGAR ANALYSIS 



[51 




Befractometer X, 
Prisms 



Fig. 34. Zeiss spiral water-heater with pressure regulator. 



PRINCIPLE AND USES OF THE REFRACTOMETER 61 

attached to the upper prism by the capillary action of a drop of mono- 
bromonaphthalene; the polished end surface of the test plate is directed 
downwards to receive the reflected rays from the bright inner surface 
of the metal casing surrounding 
the lower prism. The average 
of several readings is taken, the 
prism being wiped clean and the 
plate reattached after each meas- 
urement. Care must be exer- 
cised not to confuse the reading 
in the reversed position of the 
sector scale. The average of the 
readings should not differ more 
than two points in the fourth 
decimal from the value marked 
upon the plate. Should greater 
differences than this occur, the 
refractometer should be adj usted. 
In some of the instruments the 

adjustment is made by moving 
, , . , . ,, , & Fig. 35. Verifying adjustment of refrac- 

the index of the sector scale tometer by test plate, 

with a setpin until it corre- 
sponds to the value marked upon the test plate. The border line of 
the field must remain meanwhile upon the intersection of the reticule, 
so that care must be exercised not to ^disturb the alidade while making 
the adjustment. 

In more recent forms of the Abbe refractometer the adjustment is 
made by moving the reticule instead of the index. The process is the 
reverse of that previously described. The alidade is first moved until 
the index of the scale corresponds to the reading of the test plate; then 
by means of a key the screw K (Fig. 31), which moves the reticule, is 
turned until the intersection of the cross threads coincides with the 
border line. 




REFRACTOMETER TABLES FOR SUGAR SOLUTIONS 

A number of tables have been constructed which give the refractive 
indices of sugar solutions for different concentrations. The first of 
such tables was published in 1883 by Strohmer,* who showed also that 
a fixed relation existed between the refractive index and specific gravity 



Oest. Ung. Z. Zuckerind., 12, 925; 13, 185. 



62 



SUGAR ANALYSIS 



of sugar solutions. Using the method of least squares, Strohmer cal- 
culated this relation to be n 5 = 1.00698 + 0.32717 d, in which d is the 
specific gravity of the solution at 17.5 C. 

In 1901 Stolle,* using a Pulfrich ref lactometer, constructed tables 
for sucrose, glucose, fructose, and lactose, a comparison of which showed 
that but very little variation existed in the refractive index of solutions 
of different sugars for the same concentration. The following table 
is made up from the observations of Stolle upon sucrose solutions of 
different concentrations. 

TABLE XIV 

Giving Index of Refraction of Sugar Solutions 



Concentration, 


Specific gravity (d) 


Per cent sucrose 


Refractive index (n) 


Refractive constant 








17 5 


n 1 




4 






(n*+2) d 


0.9979 


1.00241 


1.00 


1.33465 


0.20612 


4 0073 


1.01406 


3.95 


1.33889 


0.20615 


12.0052 


1.04484 


11.49 


1.35044 


0.20617 


17.9385 


1.06736 


16.81 


1.35891 


0.20621 


25.0120 


1.09420 


22.87 


1.36891 


0.20617 


35.0219 


1 . 13194 


30.94 


1.38306 


0.20610 


45.8381 


1 . 17246 


39.10 


1.39873 


0.20619 


55.0266 


1.20651 


45.61 


1.41150 


0.20602 



The average value for the refractive constant (calculated by the 
formula of Lorenz and Lorentz) is 0.20614; from this it follows that 
the specific gravity (d) of sugar solutions may be calculated from the 

n 2 1 
refractive index (n) by the equation d = 2 x 02Q61 4' 

In 1906 Tolman and Smith,f using an Abbe refractometer of latest 
construction, showed that "the refractometer is a satisfactory instru- 
ment for determining the soluble carbohydrates in solution under the 
same conditions as those under which specific gravity can be used, and 
in fact gives the same results; that it has many advantages over the 
specific gravity method in speed, ease of manipulation, and amount of 
sample required for the determination," and that the refractometer can 
be used for a great deal of work where quickness and approximate 
accuracy only are necessary. Tolman and Smith give the following 
table showing index of refraction at 20 C. and percentage of various 
carbohydrates in solution. 



* Z. Ver. Deut. Zuckerind., 51, 469. 
t J. Am. Chem. Soc., 28, 1476. 



PRINCIPLE AND USES OF THE REFRACTOMETER 



63 



TABLE XV 

Giving Index of Refraction of Various Sugar Solutions of Different Concentration 
(Dried in vacuum at 70 C. to constant weight.) 



Index of refraction, 
20 C. 


Sucrose. 


Maltose. 


Commercial 
glucose. 


Lactose. 


Dextrin. 


1.3343 
1.3357 
1.3402 
1.3477 
.3555 
.3637 


Per cent. 

1.00 
2.00 
5.00 
10.00 
15.00 
20.00 


Per cent. 

1.00 
2.07 
5.07 
10.07 
15.12 
20.17 


Per cent. 
1.00 

2.00 
5.00 
10.07 
15.06 
20.06 


Per cent. 
1.00 
2.00 
5.13 

10.13 
15.13 


Per cent. 
1.00 
1.93 

4.87 
9.60 
14.13 
18 94 


3722 


25 00 




25 00 




23 71 


3810 


30 00 




30 02 




28 78 


.3902 


35.00 




35.03 






.3997 


40.00 




40.05 






1.4096 


45.00 




45.04 






1 4200 


50 00 




50 03 






1 4306 


55 00 




55 02 






1 4419 


60 00 




60 01 






1 4534 


65 00 




65 01 






1 4653 


70 00 




70 00 






1 . 4776 


75 00 




75.00 






1.4903 


80.00 




80.00 






1 . 5034 


85.00 




85.00 






1.5170 


90.00 




90.00 



















It will be seen from the above table that dextrin alone of the carbo- 
hydrates examined differs appreciably from sucrose in its index of 
refraction. Comparing the specific gravity ^r of the above sucrose 
solutions with their refractive indices the method of least squares shows 
that nl= 0.9509 + 0.3818 d^. 

Tolman and Smith also studied the effects of temperature upon the 
refractive index of sugar solutions, and their results "show that the 
temperature correction for the specific gravity and the index of refrac- 
tion are practically the same, and the table as given for Brix can be 
used for the index of refraction. The manner of using the table is the 
same. The reading of index of refraction is made at room temperature 
and this reading calculated to per cent of sugar, then the proper correc- 
tion from the table calculated and applied." 

Following the work of Tolman and Smith was that of Main* in 
1907. Main was the first to demonstrate the practical utility of the 
Abbe refractometer in sugar-house work, and showed that the refractive 
index was an accurate measure of the moisture and total solids in all 

* Int. Sugar Jour., 9, 481. 



64 



SUGAR ANALYSIS 



refinery products except the very lowest. The table of Main (Table 5, 
Appendix), which agrees almost exactly with that of Tolman and Smith, 
is the one employed by most sugar chemists at present. The indices give 
the percentage of water to 0.1 per cent from 100 per cent to 15 per cent; 
the percentage of water subtracted from 100 gives the corresponding 
percentage of total solids. Stanek * has prepared a table of tempera- 
ture corrections for the table of Main, the figures of which show, as 
was indicated by Tolman and Smith, that the temperature corrections 
for specific gravity and refractive index are virtually the same (Table 6, 
Appendix) . 

Schonrock f of the Physikalisch-Technische Reichsanstalt in Berlin 
has made the most recent measurements of the refractive indices of 
sugar solutions. A preliminary report of Schonrock's determinations, 
which as regards attention to scientific detail are probably the most 
carefully conducted of any measurements thus far made, is given in 
Table XVI, in which n is the refractive index at 20 C. for the two D 
lines of sodium light (589.3 ///*) and w the water content of the solution. 

TABLE XVI 
Giving Refractive Index and Water Content of Sugar Solutions 



< 


W 


< 


W 


< 


W 


-v,20 

n D 


w 


1.3330 
1.3344 
.3359 
.3374 
.3388 
.3403 
.3418 
3433 


100 
99 
98 
97 
96 
95 
94 
93 


1.3639 
1.3655 
1.3672 
1.3689 
.3706 
.3723 
1.3740 
1 3758 


80 
79 
78 
77 
76 
75 
74 
73 


1.3997 
1.4016 
1.4036 
1.4056 
1.4076 
1.4096 
1.4117 
1 4137 


60 
59 
58 
57 
56 
55 
54 
53 


.4418 
.4441 
.4464 
.4486 
.4509 
.4532 
.4555 


40 
39 
38 
37 
36 
35 
34 


3448 


92 


1 3775 


72 


1 4158 


52 






3464 


91 


1 3793 


71 


1 4179 


51 






3479 


90 


1 3811 


70 


1 4200 


50 






3494 


89 


1 3829 


69 


1 4221 


49 






3510 


88 


1 3847 


68 


1 4242 


48 






3526 


87 


1 3865 


67 


1 4264 


47 






3541 


86 


1 3883 


66 


1 4285 


46 






1 3557 


85 


1 3902 


65 


1 4307 


45 






1 3573 


84 


1 3920 


64 


1 4329 


44 






1.3590 


83 


1 3939 


63 


1 4351 


43 






1.3606 


82 


1 3958 


62 


1 4373 


42 






1.3622 


81 


1.3978 


61 


1 4396 


41 























The above table shows no greater deviation at any reading than 
in the fourth decimal place from the previous work of Main. 

* Z. Zuckerind. Bohmen, 33, 153. 
t Z. Ver. Deut. Zuckerind., 61, 421. 



PRINCIPLE AND USES OF THE REFRACTOMETER 65 



The use of the Abbe refractometer was extended to raw sugar cane 
products by Prinsen Geerligs and van West* who made a special study 
of the effect of impurities upon the refractive index of sugar solutions. 
Their results, in connection with observations upon low-grade Java 
molasses, show that the refractive index of impure sugar solutions is a 
much truer measure of the actual amount of dry substance present than 
the specific gravity. The refractometer table (Table 7, Appendix) of 
Geerligs f is established at 28 C. and is the one best adapted for tropi- 
cal countries; the temperature corrections which accompany the table 
have a range from 20 to 35 C. When corrected to 20 C., Geerligs's 
results are identical with those of Tolman and Smith, and Main. 

The use of the refractometer in the examination of sugar-beet 
products has been studied by Lippmann, Htibener, Lange, and many 
others. As in the case of sugar-cane products, the refractometer gives 
values for solid matter much closer to the true dry substance than 
specific gravity. 

The percentage of moisture or dry matter in sugar products which 
have partly crystallized, such as massecuites, moist sugars, etc., can 
be made upon the refractometer after dissolving all soluble matter 
with a known amount of water. 

Example. 10 gms. of massecuite were dissolved in 10 c.c. of hot distilled 
water, the weight of mixture after cooling to 20 C. being brought to 20 gms. 
by addition of distilled water of 20 C. The refractive index of the mixture 
was 1.4107, which according to Main's table indicates 54.45 per cent water. 
54.45 per cent of 20 gms. = 10.89 gms. water in mixture. 10.89 - 10 (gms. 
water added) = 0.89 gm. water in original massecuite, or 8.90 per cent. 

Hardin has made comparative determinations of the moisture in 
different grades of sugar by drying and by the refractometer with the 
following results: 



Grade of sugar. 


Refractive index, 
20 C. (1 part sugar 
+1 part distilled 
water). 


Per cent of water. 


By refractometer. 


By drying to 
constant weight. 


Refined sugar 


1.4200 
1.4199 
1.4197 
.4190 
.4189 
.4181 
.4179 
.4172 
.4139 


Per cent. 
0.10 
0.20 
0.40 
1.00 

1.10 
1.90 
2.10 
2.70 
5.90 


Per cent. 
0.05 
0.45 

0.82 
0.82 
1.05 
1.93 
2.40 
2.83 
5.54 


Hawaiian centrifugal 


Philippine mats (dried out) 
Java centrifugal 


Louisiana centrifugal 


Cuban centrifugal 


Muscovado 


IVIolasses sugar 


Molasses sugar 





* Archief Java Suikerind. (1907), 15, 487. t Int. Sugar Jour., 10, 69-70. 



66 SUGAR ANALYSIS 

The variations in the results by the two methods are in both direc- 
tions, and may have been due either to the presence of trash in the 
sugar or to the influence of non-sugars. Since the refractometer only 
indicates the percentage of dissolved solids, any insoluble matter 
which is present in the weighed sample will introduce an error in the 
calculation. 

Insoluble suspended matter in sugar solutions, if present in large 
amounts, will darken the field of the refractometer and interfere with 
the adjustment of the border line. In such cases the solution must 
be filtered. 

Examination of Dark-colored Sugar Solutions with the Re- 
fractometer. In the examination of dark-colored sugar solutions, 
molasses, sirups, extracts, etc., by means of the refractometer, it is not 
always possible for the compensator to eliminate completely the effects 
of dispersion; the border line of the field is then more or less blurred 
and a sharp adjustment to the intersection of the reticule becomes a 
matter of some difficulty. In solutions which are not too strongly 
colored this trouble may be remedied by bringing the border line to 
the point of intersection alternately from each side of the field; the 
average of the readings thus obtained will correct to a large extent the 
errors of faulty adjustment. Some authorities have recommended 
with dark solutions to adjust the compensator to a colored border, 
selecting the color most sensitive to the observer's eye; this, however, 
is not very satisfactory, and if the blurring of the border line is excessive, 
the color of the solution must be reduced by some method of dilution 
or clarification. 

In the dilution of impure sugar products with water an error will be 
introduced in the refractometer reading in the same manner as in the 
determination of specific gravity, owing to the difference in contrac- 
tion between solutions of sugar and of the accompanying impurities 
(page 35). 

A study of the errors resulting from unequal contraction, when 
dilution is employed in densimetric and refractometric methods of 
analysis, has been made by Stanek.* Fifty per cent solutions of betaine 
and of various organic salts of sodium and potassium were prepared. 
These solutions were then diluted with known weights of water and the 
per cent of dry substance determined from the degrees Brix, from the 
refractive indices according to Main's table, and by drying on sand 
in a Soxhlet oven at 102 C. A few of the results are given in the 
following table: 

* Z. Zuckerind. Bohmen, 34, 5. 



PRINCIPLE AND USES OF THE REFRACTOMETER 



67 



TABLE XVII 
Comparative Determinations of Solids by Brix, Refractometer and Drying at 102 





True dry 


I 


)ry substance by 




SubststncG t&KGii. 


substance. 


Degrees Brix. 


Refractometer. 


Drying at 102. 


c 


Per cent. 
5 


Per cent. 
2.2 


Per cent. 
5.10 


Per cent. 
5.05 


Betaine (anhydrous) . . . . \ 


10 


4.3 


10.20 


10.01 




25 


10.8 


24.15 


25.03 


( 


5 


8.1 


4.60 


4.99 


Sodium formate \ 


10 


15.6 


8.85 


10.04 




25 


37.7 


20.55 


25.05 


( 


5 


7.3 


3.60 


5.00 


Potassium formate < 


10 


14.28 


7.20 


9.97 




25 


35.7 


17.20 


25.09 


( 


5 


6.7 


5.00 


4.97 


Sodium acetate \ 


10 


13.1 


9.70 


9.99 




25 


31.1 


22.70 


25.00 


, 


5 


6.6 


5.00 


5.00 


Potassium acetate ... . s 


10 


12.8 


8.25 


10.07 




25 


30.4 


19.75 


25.15 


( 


5 


4.75 


4.90 


4.90 


Sodium butyrate < 


10 


9.4 


10.25 


9.89 




25 


22.9 


24.35 


24.94 


( 


5 


6.3 


5.00 


5.10 


Sodium lactate ] 


10 


12.3 


10.00 


10.07 




25 


30.2 


24.05 


25.05 


jf 


5 


6.3 


4.85 


5.18 


Potassium lactate j 


10 


12.5 


9.10 


10.13 




25 


30.3 


21.65 


25.20 


, 


5 


6.8 


6.40 


5.05 


Sodium glutaminate < 


10 
25 


13.2 
31.1 


12.50 
30.05 


10.23 
26.41 


( 


5 


6.7 


5.90 


5.03 


Potassium glutaminate . . . . \ 


10 
25 


13.1 
30.65 


11.50 
27.70 


10.24 
25.27 



It will be noted from the above that the refractometer gives a 
much closer approximation to the true dry substance than the degrees 
Brix, the refractometer yielding usually lower results and the degrees 
Brix higher. It is also seen that the sodium salts of organic acids give 
higher results by both methods than potassium salts. Contraction 
upon dilution is noted in every case, the results corrected for dilution 



68 SUGAR ANALYSIS 

being higher according to the amount of water added. The usual 
effect of this contraction is to make the error in estimating non-sugars 
less by the refractometer and greater by degrees Brix. Neither of 
these methods for estimating non-sugars approaches in point of ac- 
curacy the method of actual drying. 

The errors in determining the refractive index of dark impure 
sugar solutions, resulting from dilution with water, may be largely 
eliminated by employing the method of Tischtschenko,* which con- 
sists in reducing the color of the product by means of a solution of pure 
sucrose of about the same density as the liquid to be examined. The 
disturbing influences of color dispersion in the refractometer field are 
in this way overcome without the errors of contraction. The method 
of operation is as follows: A known weight (a) of the molasses, sirup, 
etc., is intimately mixed with a known weight (6) of pure sugar solution, 
whose sugar content (p) has been previously determined by means of 
the refractometer. The refractive index of the mixed solution is then 
determined and the corresponding percentage (P) of dry substance found 
from the table. The percentage of dry substance (x) in the molasses, 
sirup, etc., is then calculated by the formula ax + bp = (a + 6)P, 

(a + b)P-bp 

whence x = - - 

a 

Example. Weight of beet molasses (a) = 14.1028 gms. 
Weight of sugar sirup (6) = 13.2438 gms. 
Sugar in sirup (p) = 51.3 per cent. 

ngof mixture = 1.4538 = 34.87 per cent water 

(Main's table). 
Solids of mixture (P) = 100-34.87 = 65.13 per cent. 

Substituting these values in the formula, x 78.12 per cent solids in molasses. 
The method by water dilution gave 79.11 per cent. Direct determination by 
drying gave 77.80 per cent. 

If a sugar sirup of greater density had been used for mixing, the value of x 
would have been more close to the result by direct determination. 

If equal weights of molasses and sugar solution are used in Tischt- 
schenko's method, then a = b .in the formula, whence x = 2P p; 
the labor of calculation is thus considerably reduced. In using the 
method, the mixture of molasses and sugar solution must be perfectly 
homogeneous. Care must also be exercised, as in all cases, that no 
air bubbles are inclosed with the liquid between the prisms. A com- 

* Z. Ver. Deut. Zuckerind., 69, 103. 



PRINCIPLE AND USES OF THE REFRACTOMETER 69 

parison of results in determining dry substance in different samples 
of beet molasses by various methods is given by Lippmann* in the 
following table: 

TABLE XVIII 

Comparative Determinations of Solids in Beet Molasses by Drying, Specific Gravity, and 

Refractometer 



Number. 


By direct 
determination. 


By degrees 
Brix. 


By refractometer. 


Water dilution. 


Tischtschenko's 
method. 


1 


76.78 
77.95 
76.22 
77.85 
77.05 
77.55 
77.97 
77.32 
77.50 
77.31 
76.58 
76.94 
77.43 
76.53 
77.82 
77.90 


78.90 

79.80 
78.60 
79.30 
79.40 
79.20 
79.90 
79.30 
79.30 
79.60 
78.90 
79.20 
79.60 
78.90 
80.00 
80.20 


77.90 

78.50 
77.00 
78.60 
78.20 
78.10 
78.60 
78.20 
78.60 
78.40 
77.70 
77.90 
78.50 
77.70 
79.00 
78.90 


76.80 
78.00 
76.10 
77.90 
77.30 
77.80 
78.30 
77.70 
77.88 
77.70 
77.00 
77.40 
77.90 
77.00 
78.30 
77.40 


2 


3 . . 


4 


5 


6 


7 


8 


9 


10 


11 


12 


13. .. . 


14 


15 


16 


Average 


77.29 


79.38 


78.24 


77.53 





It will be noted from the above that the average error of estimating 
dry substance in the 16 samples of beet molasses was, by degrees Brix, 
+2.09 per cent; by refractometer, using water dilution, +0.95 per cent; 
and by refractometer, using Tischtschenko's method, only +0.24 per 
cent. 

Another method of correcting the disturbances in refractometer 
work due to color of solution is by clarification. Lead subacetate is 
the reagent most generally employed for this purpose. The use of 
this and similar salts must be limited, however, to the greatest possible 
minimum, since the excess of salt remaining in the clarified solution 
causes an increase in the refractive index. In the following experiments 
made by Rosenkranzf at the Berlin Institute for Sugar Industry, the 
effect of increasing the quantity of subacetate is shown upon the re- 
fractive index of a molasses containing 78.59 per cent dry substance 
and diluted 1:1, inclusive of the lead solution added. 

* Deut. Zuckerind., 34, 402. 

t Z. Ver. Deut. Zuckerind., 68, 195. 



70 



SUGAR ANALYSIS 



Lead subacetate. 


Specific gravity, 
dilute solution, 
20. 


Calculated 
Brix of original 
molasses. 


Refractive 
index, dilute 
solution. 


Dry substance, 
dilute solution 
(Main's table). 


Calculated 
dry substance, 
original 
molasses. 


c.c. 

"5 
10 
12.5 


1.1813 

1.1865 
1.1912 
1.1951 


81.9 
84.0 

85.7 
87.2 


1.3994 
1.4003 
1.4009 
1.4022 


39.85 
40.3 
40.6 
41.3 


79.70 
80.60 
81.2 
82.6 



Another material recommended by Lippmann for decolorizing dark 
sirups, etc., for the refractometer is "Decrolin," the zinc salt of formal- 
dehyde sulphoxylic acid, CH 2 OH.O.SO.Zn.OH. One to two per cent of 
Decrolin is used and the liquid heated to about 55 C. to hasten solution 
and decolorization. 

For the refractometric examination of turbid beet juices, etc., 
Herzfeld* has recommended the addition of a few drops of 10 per 
cent acetic acid, heating for 2 minutes at 80 C. to coagulate albu- 
minoids, and filtering. With beet juices the effect of dilution (1 to 5 
per cent) is compensated by the greater refractive index of the 10 per 
cent acetic acid used, as shown in the following experiment: 



10 c.c. beet juice. 


Refractive 
index, rc*- 


Dry substance 
by Main's 
table. 


+0.5 c.c. water 


1.3583 


16.75 


H-0.5 c.c. acetic acid (10 per cent) 
+0.25 c.c. water 


1.3595 
1.3588 


17.45 
16.95 


+0.25 c.c. acetic acid (10 per cent) .... 
+0 10 c.c. water 


1.3591 
1 35905 


17.20 
17 15 


+0.10 c.c. acetic acid (10 per cent). . . . 


1.35905 


17.15 



THE IMMERSION REFRACTOMETER 

A second form of instrument which is used for determining the re- 
fractive power of sugar solutions is the immersion refractometer, the 
Zeiss model of which is shown in Fig. 36. While this instrument has a 
narrower range than the Abbe apparatus, the scale being adapted only 
for solutions containing from to 21.7 per cent sugar, it gives a much 
sharper border line, thus allowing a greater magnification in the tele- 
scope, with a corresponding increase in the accuracy of observation. 
In the immersion refractometer there is no sector; the scale is placed 
below the eyepiece of the telescope, the latter, unlike the Abbe re- 
fractometer, being rigidly connected with the prism. 

* Z. Ver. Deut. Zuckerind., 68, 197. 



PRINCIPLE AND USES OF THE REFRACTOMETER 71 

The principle of the immersion refractometer is the same as that of 
the Abbe instrument, being based upon an observation of the border 
line of total reflection. In Fig. 37, G is a cylindrical glass prism with 
its refracting surface DE immersed in the liquid W contained in the 
glass beaker V. If we suppose light to pass through the top of the 
prism from the surface A B, the parallel rays sP 5 s'P', s"P", etc., will 








Fig. 36. Zeiss immersion refractometer. 

be refracted in the liquid in the direction PM, P'M', P"M", etc. By 
increasing the angle of incidence for the parallel rays upon the surface 
DE, a point is reached where the parallel rays rP, r'P' , r"P", etc., are 
refracted along the surface of the prism towards D. This is the bor- 
der line of total reflection as explained under Fig. 30, where the angle 
of refraction is 90. In the use of the immersion refractometer the 
course of the light is in the reversed direction to that just described, 
being reflected from the mirror HK through the bottom of the beaker 
V so as to pass as nearly parallel as possible to the oblique surface of 
the prism. The rays of light which coincide with the surface DE form 



72 



SUGAR ANALYSIS 



the border line for total reflection and are refracted upward through the 
prism as the parallel rays Pr, PV, P"r", etc., which, being condensed 
by the objective of the refractometer telescope upon the point x of 

the scale S, form the border 
line for observation; the rays of 
light which may strike the prism 
surface obliquely, as MP, M'P', 
M"P" , etc., are refracted in the 
direction Ps, PY, P'Y', etc., 
and being condensed by the ob- 
jective between x and y cause 
this part of the scale to be illu- 
minated. There being no pos- 
sible angle of refraction for light 
in the prism greater than that 
for the border line of total re- 
flection, the part of the scale 
between x and z remains in 
shadow. 

As in the Abbe refractom- 
eter, the border line on account 
of differences in dispersion is 
fringed with color and must be 
corrected by a compensator in 
the manner described on p. 57. 
The compensator is placed at A 
(Fig. 38) between the objective 
and the prism P and is ro- 
tated by the milled ring R until 
the border line upon the scale 

becomes sharp and colorless. The position of the border line upon the 
scale marks the reading for the whole division; the fractional division 
is determined by rotating the micrometer screw Z, which controls the 
scale, until the whole division previously noted is brought into contact 
with the border line. The reading of the micrometer drum shows the 
fractional division which remains to be added. Readings can be made 
by careful observers to agree within 0.1 scale division, which cor- 
responds to 3.7 of the fifth decimal of the refractive index. This ex- 
ceeds considerably in accuracy the reading of the Abbe refractometer. 
The adjustment of the Zeiss immersion refractometer scale is made 
by means of distilled water, which should give a reading of 15 at 




Fig. 37. Illustrating principle of immer- 
sion refractometer. 



PRINCIPLE AND USES OF THE REFRACTOMETER 



73 



17.5 C. The adjustment, however, can be made at other tempera- 
tures according to the following table. 

The correctly adjusted refract ometer should show for distilled water: 



At a temperature of 


10 C. 


11 


12 


13 


14 


15 


16 


17 


17.5 


18 


19 C. 


























The scale division 


16.3 


16.15 


16.0 


15.85 


15.7 


15.5 


15.3 


15.1 


15.0 


14.9 


14.7 



























At a temperature of 


20 C. 


21 


22 


23 


24 


25 


26 


27 


28 


29 


30 C. 


The scale division 


14.5 


14.25 


14.0 


13.75 


13.5 


13.25 


13.0 


12.7 


12.4 


12.1 


11.8 




























Fig. 38. Showing inner construction of immersion refractometer. 

Should the average of several careful readings differ by more than 
0.1 division from the reading in the above table for the temperature of 
testing, the scale should be readjusted. This is done by first setting 
the micrometer at 10; then by inserting a setpin in the hole of an 



74 



SUGAR ANALYSIS 



adjusting screw inside the micrometer drum and turning anticlockwise, 
the border line of the field is made to agree with the whole scale division 
corresponding to the temperature of the water. The loosened microm- 
eter drum is now turned until its index marks the proper decimal; 
holding it firmly in this position, the nut which governs the micrometer 
is retightened. The new adjustment should be controlled by repeated 
readings. 

The readings of the Zeiss immersion scale extend from 5 to +105, 
and are converted into refractive indices or into percentages of sugar 




Fig. 39. Tempering bath for immersion refractometer. 

by means of special conversion tables which accompany the instrument. 
Sugar tables for the immersion refractometer have been prepared by 
Hiibener; * these give the sucrose values of the scale from 15 to 106 
with percentages of sucrose from 0.00 to 21.71. Each 0.1 division of 
the scale corresponds to about 0.02 per cent sucrose or other sugar, 
and readings can be made with this degree of exactness. (See Table 8, 
Appendix.) 

For controlling the temperature of the water bath, containing the 
* Dent. Zuckerind., 33, 108. 



PRINCIPLE AND USES OF THE REFRACTOMETER 75 

beakers of solution for the immersion refractometer, the spiral heater 
and water-pressure regulator previously described may be used. A 
tempering bath holding 10 liters of water and with a revolving frame 
for 12 beakers (shown in Fig. 39) is also recommended. When the 
proper temperature has been reached in the beakers the solutions are 
read in sequence, the refractometer prism being wiped dry after each 
immersion. When large numbers of solutions are to be tested, each 
solution as soon as read is replaced by]a beaker of fresh solution, thus 
giving sufficient time for regulation of temperature without interruption 
of work. 

When only a few cubic centimeters of solution are available or when 
the liquids to be examined consist of dark-colored sirups, molasses, 
extracts, etc., the immersion prism is fitted with an auxiliary prism 
held in position by means of a metal beaker and cover. The method 
of use is somewhat similar to that of the Abbe refractometer; the hy- 
potenuse surface of the auxiliary prism is covered with a few drops of 
solution and then inserted in the beaker against the face of the immer- 
sion prism so that a thin layer of liquid is spread between the two. 

The remarks upon illumination under the Abbe refractometer also 
apply to the immersion instrument. 

As to the particular choice of refractometer for the sugar laboratory, 
the chemist must be guided by his requirements. The Abbe refractom- 
eter has the widest range, is adapted to smaller quantities of solution, 
and is more convenient to operate. The immersion refractometer, 
however, is more accurate in adjustment and much less expensive. For 
general work the Abbe instrument will be found more useful; for more 
limited operations upon solutions below 20 Brix, such as beet and 
cane juices, sweet waters, etc., the immersion instrument possesses 
certain advantages. 



CHAPTER V 

POLARIZED LIGHT, THEORY AND DESCRIPTION OF POLARIMETERS 

IN order to arrive at a sufficiently clear understanding of the optical 
principles which underlie the construction and manipulation of polari- 
scopes, a brief reference must be made to the physical theories of light. 

According to the undulatory theory of Huyghens, light consists of 
vibrations or wave motions of the luminiferous ether, the imponder- 
able medium which pervades all space and penetrates all matter. 

Waves of light, contrary to those of sound, vibrate transversally 
instead of longitudinally. In Fig. 40 a graphic representation is given 



o 



D 
Fig. 40. Illustrating principle of a light wave. 

of a light wave vibrating transversally to the direction of motion LM. 
The plane of vibration of ordinary light takes all possible positions 
about this line of motion. The distance OB or O f D from the middle to 
the extremity of an oscillation is known as the amplitude of the wave. 
The distance from A to E (points in the same phase) is known as the 
wave length (X), which for light is expressed in millionths of a milli- 
meter GU/X). The number of waves per second is called the rate of 
vibration (N). If the velocity of light through a homogeneous medium 

be V, then N = ^- 

A 

According to Maxwell's electromagnetic theory, which has since 
been confirmed by the work of Hertz, there are two sets of transverse 
vibrations in the transmission of a ray of light, the one an electric dis- 
placement of the ether, and the other a magnetic displacement, the 
planes of these being perpendicular to each other. 

The intensity of a ray of light is proportional to the square of the 

76 



flTYIT 



THEORY AND DESCRIPTION OF POLARIMETERS 



77 



amplitude; the color depends upon the rate of vibration of the ether 
wave. The color of light may, therefore, be expressed mathematically 
in terms of the rate of vibration N or of its wave length X. The values 
of N and X for the average ray in each color of the spectrum are given 
in the following table : 

TABLE XIX. 



Color. 


Rate of vibration 
per second (N). 


Wave length (X) 
in millionths of a 
millimeter (jin). 


Red 


Billions. 
437 


683 


Orange 


485 


615 


Yellow 


534 


559 


Green 


582 


512 


Blue 


631 


473 


Indigo 


679 


439 


Violet 


728 


410 









The human eye is sensitive to light of vibration periods between 
about 366 and 804 billion per second, and of wave lengths between 
about 820 MM and 373 pp. Ether waves of greater length than 820 MM 
constitute the so-called infra-red or heat rays, and those of shorter 
length than 373 MM the so-called ultra-violet or chemical rays. 

Light of definite wave length is exceedingly important in making 
polariscopic measurements, and this is secured by using incandescent 
salts of certain metals, as sodium or lithium, which give bright spectral 
lines whose wave lengths are absolutely defined. The prominent lines 
of the different elements are usually designated by the letters of the 
alphabet, which have been adopted to mark their positions in the solar 
spectrum. For the sodium line * D, to which nearly all polariscopic 
measurements are referred, X = 589.3 MM- 

The vibrations of ordinary light proceed in an infinite number of 
planes. By means of various special contrivances it is possible, how- 
ever, to affect a beam of light so that the electric and magnetic vibrations 
will each proceed in a single plane. Such light is said to be plane- 
polarized; the plane to which the electric vibration of the waves is 
perpendicular is called the plane of polarization. 

The polarization of light was first noticed by Huygens in 1678, 
while studying the refraction of light in a crystal of Iceland spar. No 
satisfactory explanation of the phenomenon was made, however, until 

* The sodium line is double; the component Z>i has a wave length of 589.6 MM 
and the brighter component D 2 a wave length of 589.0 MM- The average wave length 
of the two lines, 589.3 /*/* (more exactly 589.25 MM), is the value taken for D. 



78 SUGAR ANALYSIS 

Malus, in 1808, discovered that the polarization noticed by Huygens in 
Iceland spar could also be produced by reflection. 

Polarization by Reflection. If a beam of light (as LO in Fig. 28) 
fall upon the smooth surface of a transparent substance, it is decomposed 
into reflected and refracted rays. The reflected rays at a definite 
angle of incidence are completely polarized, the plane of the lines of 
incidence and reflection being the plane of polarization.* These obser- 
vations, according to Fresnel and Arago, could be explained only by 
supposing that the vibrations in a light wave are tran verse to the direc- 
tion of motion, and that during reflection these vibrations are reduced to 
a single plane, which is perpendicular to the plane of polarization. 

The angle of incidence at which reflected light is completely polar- 
ized is called the polarizing angle, and varies according to the refractive 
power of the reflecting substance. This relationship is expressed by 
Brewster's law, viz. : The tangent of the polarizing angle is equal to the 
index of refraction for the reflecting substance, or tan i = n. The 
polarizing angle of glass (n = 1.54) is accordingly about 57 degrees. 

The Norrenberg Apparatus. A simple apparatus for producing 
and studying polarized light is that of Norrenberg, shown in Fig. 41. 
A and B are two mirrors of black glass; the upper mirror B can be 
rotated by the crank D around the vertical axis of the instrument, the 
angular displacement being indicated upon a divided circle S. The 
planes of the two mirrors are first placed parallel, at an angle of 45 
degrees to the vertical, and a beam of light is allowed to fall upon the 
mirror A at an angle of incidence of 57 degrees. The reflected beam is 
then completely polarized and, passing upward, is reflected from mirror 
B upon the screen C, where it appears as a bright spot. With the 
mirrors parallel, the planes of incidence and reflection, and hence of 
polarization, coincide for each surface. Without changing its inclina- 
tion, the mirror B with its screen C is rotated by the crank D around the 
vertical axis. The plane of incidence and reflection for the beams of 
polarized light at mirror B no longer coincide with that at mirror A; 
the intensity of the spot of light upon the screen accordingly begins to 
diminish until, after a revolution of 90 degrees, the screen is perfectly 
dark, all the light being refracted and absorbed in the mirror B. In 
the latter position the planes of incidence, and hence of polarization, for 
the light of the two mirrors are at right angles, and the mirrors are 
said to be crossed. By turning D in the same direction the spot of light 

* The refracted rays of light are also polarized, but not completely; most of the 
refracted rays, however, are polarized in one direction, their plane of polarization 
being perpendicular to that of the reflected rays. 



THEORY AND DESCRIPTION OF POLARIMETERS 



79 



reappears upon the screen, and after 180 degrees again reaches maxi- 
mum brilliancy, in which position the planes of incidence and of polar- 
ization again coincide in both mirrors; at 270 degrees, when these 
planes are again at right angles, the spot of light is reextinguished. 




Fig. 41. Norrenberg's polarizing apparatus. 

If at one of the points of extinguishment of light upon the screen 
the glass cylinder F containing a solution of sucrose or other optical 
active sugar be inserted in the path of the light rays reflected from A, 






80 SUGAR ANALYSIS 

the illumination upon the screen will reappear. The plane of polariza- 
tion of the light reflected from A must, therefore, have been rotated by 
the sugar solution through a certain angle in order that reflection could 
take place from B\ by turning D until the plane of polarization for the 
light upon B is again brought perpendicular to the plane of incidence, 
the point of maximum darkness is reestablished. By measuring upon 
S the positions of maximum darkness, before and after inserting the 
cylinder, the angle through which the sugar solution has rotated the 
plane of polarized light can be measured. In the Norrenberg ap- 
paratus the mirror A for polarizing the light is called the polarizer and 
the mirror B for measuring rotation the analyzer. 

Polarization by Double Refraction. Of the several contrivances 
available for producing plane polarized light, a modified crystal of Ice- 
land or calc spar is the only one used in the construction of polariscopes 
and saccharimeters. Calc spar is a clear, transparent mineral which 
cleaves readily into rhombohedra. If a small object be viewed through 
such a rhombohedron, the image will be doubled. Rays of light in? 
passing through the crystal undergo "double refraction." The phe- 
nomenon is noticeable in any position of the calc-spar rhombohedron 
except in a direction parallel to the diagonal joining the two opposite 






c D > 

Fig. 42. Calc spar rhom- Fig. 43. Illustrating double refraction of 

bohedron. light in calc spar. 



obtuse corners, known as the optical axis. Any plane including the 
optical axis and perpendicular to the face of the crystal is called an 
axial plane or principal section. 

In the rhombohedron of calc spar, in Fig. 42, the direction AH is 
the optical axis. The plane ABHG (or any parallel plane) perpendicu- 
lar to the face AFGD is an axial plane or principal section to that 
face. 

If a beam of light LA fall upon the surface of such a rhombohedron 
(Fig. 43), it is resolved into two rays, the ordinary ray ABO and the 
extraordinary ray ACE. Both of these rays emerging from the crystal 
are polarized, their planes of polarization being perpendicular to each 
other. 




THEORY AND DESCRIPTION OF POLARIMETERS 



81 



I 



The Nicol Prism. Before a crystal of calc spar can be utilized 
for polariscope construction it must be modified so as to eliminate one 
set of the component rays. The best known method (that of Nicol) 
is the following: A rhombohedron ABCD (Fig. 44) is selected whose 
length is about three times the width. At each 
end of the crystal, wedge-shaped sections BFC 
and ADE are removed so as to reduce the acute 
angles DAB and BCD of the axial plane from 
71 degrees to 68 degrees. The crystal is then 
divided by the plane FGEH perpendicular to 
the two modified end faces. The cut surfaces 
are then polished and reunited with Canada 
balsam.* The sides of the prism thus obtained 
are afterwards blackened and the whole is 
mounted by means of cork and wax in a metal 
tube. 

Let AFCE represent a principal section of 
the Nicol prism (Fig. 45). A beam of light LT 
entering parallel to the long sides of the prism is resolved into two 
component rays; the component most refracted (the ordinary ray) 
meets the film of balsam EF at such an angle that it is completely 
reflected to the side of the prism, where it is absorbed by the dark 
coating. The other component (the extraordinary ray), whose vibra- 
tions are in the plane of the principal section, is less refracted and, 
passing through the film of balsam, emerges in a polarized condition 




C 

Fig. 44. Illustrating 
construction of Nicol 
prism. 




Fig. 45. Illustrating polarization of light by a Nicol prism. 

from the end surface of the Nicol at the point e. With respect to the 
end surface of the Nicol FCLM (Fig. 44), the electric vibrations of the 
emergent light are in the plane of the principal section, i.e., in the direc- 
tion of the short diagonal FC; the plane of polarization is in the direc- 
tion of the long diagonal LM. 

"Iceland spar is rather friable, and in practice it is found easier to grind away 
half of the rhomb instead of cutting it, as generally described. The remaining halves 
of two rhombs thus ground are then cemented together." Preston, " Theory of 
Light," third edition, p. 319. 



82 



SUGAR ANALYSIS 



In the discussion of polarized light, it makes no difference which 
plane is taken for reference, provided it be always the same. In future 
pages the terms vibrate, vibration, plane of vibration, etc., refer entirely 
to the electric displacements in the transmission of light. With this 
understanding, the statement of Fresnel, which is followed in nearly 
all works upon polarimetry, that the plane of vibration of light is 
perpendicular to the plane of polarization, can be retained without 
confusion. 

The Glan Prism. The type of Nicol prism which is the most 
scientifically perfect and the one most used at present in constructing 

polariscopes and saccharimeters is that 
of Glan. In constructing this prism the 
opposite obtuse corners of a calc-spar 
rhombohedron (as ABCDEF, Fig. 46) 
c are cut off by planes PQR and STF 
perpendicular to the optical axis which 
passes through the point X. From this 
section a rectangular prism LMNO is 
sawed out, which is then cut in half 
along a plane through MN. After pol- 
ishing, the cut halves are cemented to- 
gether again by Canada balsam am 
mounted as in an ordinary Nicol. Th< 

great advantages of the Glan prism ovei 
Fie. 46. Illustrating construction ,. ,. XT . , ,, ,, 

of a Glan prism. the ordmarv Nlco1 are that the ra y s oi 

light enter the prism perpendicular 

the end surface and at right angles to the optical axis, thus securii 
the greatest amount of light capacity per unit of length. 

PRINCIPLE AND CONSTRUCTION OF POLARIMETERS 

Polarizer and Analyzer. A combination of two Nicol prisms, 
called the polarizer and analyzer, constitutes the essential feature of 
every polariscope. The function which these two parts play can b( 
be understood from the following diagram (Figs. 47 and 48). 

The polarizer, which is stationary, is represented by the pris 
ABCDEFGH, whose axial plane lies through ACEG. A beam of light 
entering from L at the point x is doubly refracted; the ordinary rays 
are eliminated at o, while the extraordinary rays emerge at e, vibratii 
in the axial planes of the prism, with the plane of polarization parallel 
with the plane BDFH. If the emergent polarized light now enter 
second prism A'B'C'D'E'F'G'H' (the analyzer), which can be rotat 





THEORY AND DESCRIPTION OF POLARIMETERS 



83 



about its long axis, its course will remain unimpeded only so long as it 
can continue to vibrate in the same axial plane. If the analyzer be 
rotated about its long axis, the light which enters from the polarizer 
is doubly refracted and only that component which vibrates in the 




Crossed 
Nicols 



Analyzer Fig.48 Polarizer 

Figs. 47 and 48. Illustrating principle of polarizer and analyzer. 

plane of the principal section emerges. As the analyzer is rotated the 
intensity of the emergent light diminishes until after a quarter revo- 
lution it is completely extinguished; in this position the axial planes 
of polarizer and analyzer are perpendicular to one another and the two 
prisms are said to be crossed (Fig. 48). If the rota- 
tion of the analyzer be continued, light will again 
begin to emerge, until after a half -re volution, when 
the axial planes are again parallel, the original in- 
tensity will be restored. 

The amount of light which will pass through the 
analyzer for any position of its axial plane with ref- 
erence to the polarizer may be readily calculated by 
referring to Fig. 49. 

Let AB be the axial plane of the polarizer (always 
stationary) and CD any given position of the axial 
plane of the analyzer, the two planes forming the 
angle DOB. From lay off any distance OP as the 
amplitude of the light emerging from the polarizer, 
From P erect PL perpendicular to CD; then the 
line OL represents the amplitude of the light emerg- 
ing from the analyzer and PL the amplitude of the 
light extinguished in the analyzer. As regards the relation in intensity, 
this is proportional to the squares of the amplitudes: OP 




extinguished by 
analyzer. 



OL + PL . 






84 SUGAR ANALYSIS 

If we erect LM perpendicular to AB and call the intensity of the light 
emerging from the polarizer OP, then the intensity of the light emerging 
from the analyzer will be represented by OM and the intensity of the 
light extinguished in the analyzer by MP (OM : MP :: OL 2 : PL 2 ). 
The intensities OM and OP are equal when the planes CD and AB 
coincide (parallel prisms) ; the intensity OM is zero when the planes CD 
and AB are perpendicular (crossed prisms). 

The construction and principle of the simplest form of polariscope 
can now be understood from the following diagram (Fig. 50). P is the 
polarizer consisting of a stationary Nicol and A is the analyzer con- 
sisting of a movable Nicol mounted in a revolving sleeve; the angular 




Fig. 50. Showing arrangement of parts in a simple polariscope. 

rotation of A is measured upon a graduated scale S. L is the source 
of monochromatic light which passes through the instrument to the 
eye of the observer at E. We will suppose the Nicol A to be crossed 
with reference to P, the point of light extinction marking the zero point 
on the scale S. If a tube T filled with a solution of some optically 
active substance, such as cane sugar, be now placed between P and A, 
the plane of polarized light emergent from P will be rotated from its 
original position and the light will no longer be entirely extinguished 
in A. By rotating the analyzer until its axial plane is perpendicular 
to the vibration plane of the light emergent from T, the point of ex- 
tinction is again reached. The angular rotation of the solution in T is 
then determined upon the graduated scale. By continuing the revo- 
lution of the analyzer, light will again emerge from the latter, to become 
reextinguished at a point 180 degrees from the first reading. Owing 
to the fact that light rays of different wave lengths are rotated to a 
different extent by optically active substances (a phenomenon known 
as rotation dispersion), it is necessary that the light used in this type of 
polariscope be monochromatic. 

Blot's Polariscope. The original polariscope of Biot* (Fig. 51), 
constructed in 1840, had an adjustable mirror (M) of black glass for 

* Ann. chim. phys. [2], 74, 401 (1840). 



THEORY AND DESCRIPTION OF POLARIMETERS 85 

the polarizer and a modified prism of calc spar for the analyzer (A). 
The essential features of this early instrument are still retained in 
modern polarimeters, although in a greatly modified form. 




Fig. 51. Biot's polariscope. 

Mitscherlich' s Polariscope. Mitscherlich* in 1844 modified the 
Biot apparatus by discarding the polarizing mirror and arranging the 
optical parts of his instrument as shown in Fig. 50. In the Biot polari- 
scope the end point was marked by total light extinction. But in the 
Mitscherlich apparatus a vertical black band with shaded margins 
marked .the zero point. By rotating the analyzer gently to and fro 
until the vertical band appears exactly in the center of the field, a zero- 
point adjustment can be secured with a probable error of 6 minutes. 
The Biot-Mitscherlich polariscope, with position of its optical parts, is 
shown in Fig. 52. 

Sections of the circular scales used upon the Mitscherlich and other 
polarimeters for measuring the angular rotation of the plane of polar- 
ized light are shown in Figs. 53 and 54. The scale in Fig. 53 for a 
small polariscope indicates 0.1 degree and is immovable, the rotation 
being indicated by the position of the zero mark of the movable ver- 
nier V. In the illustration the zero mark of the vernier lies between the 
* " Lehrbuch der Chemie" (1844), 1, 361. 



80 



SUGAR ANALYSIS 



2-degree and 3-degree division of the scale; to obtain the fractions of a 
degree, one proceeds from the zero mark of the vernier and, moving 
upward along the divisions of the main scale, comes finally to a divi- 
sion which exactly coincides with one 
of the divisions of the vernier. In 
the illustration this vernier division 
is 0.4, which, added to the reading 
on the- main scale, makes the angular 
rotation 2.4 degrees. For the larger 
polariscopes indicating 0.01 degree the 
main scale is movable, the circular 
rim divided into 0.25 degree rotating 
against the fixed vernier, which gives 
the readings to 0.01 degree. In the 
illustration (Fig. 54) the zero of the 
vernier falls between 13.50 degrees 
and 13.75 degrees; the 0.18 mark of 
the vernier is in coincidence with a 
division on the main scale. 13.50 -f 
0.18 = 13.68, which is the angular 
rotation indicated. 

Robiquet's Polariscope. Robi- 
quet increased the sensibility of the 
Biot-Mitscherlich polariscope by in- 
troducing a Soleil double quartz plate 
as the end-point device. The general 
appearance of this instrument, with 
position of optical parts, is shown in 
Fig. 55. 

Principle of the Soleil Double Quartz 
Plate. The Soleil double quartz plate 
consists of two plates of quartz of 
equal thickness, one of which rotates the plane of polarized light to the 
right and the other to the left. The plates, which are cut perpendicu- 
lar to the optical axis of the crystal, are cemented together at their 
edges and carefully ground and polished. If white polarized light pass 
through such a plate, the rays of different wave length and color will 
be rotated to a different degree (rotation dispersion), the rays of less 
wave length being rotated the most. For a piece of quartz 1 mm. thick, 
cut as above described, the rotation will be 15.75 degrees for the red B 
ray, 21.72 degrees for the yellow D ray of sodium, and 32.76 degrees for 




Fig. 52. The Biot-Mitscherlich 
polariscope. 

a = position of polarizer 
6 = position of analyzer 
c = lever for rotating analyzer 
I = condensing lens. 



THEORY AND DESCRIPTION OF POLARIMETERS 87 





Fig. 53 Fig. 54 

Sections of circular scales of polariscopes. 







Fig. 55. Robiquet's polariscope. 

d = polarizer 
e = condensing lens 
/ = Soleil double quartz plate 
g = analyzer 
h-i = telescope 
k = lever for rotating analyzer. 




88 SUGAR ANALYSIS 

the blue F ray. For the average ray in the middle of the yellow spec- 
trum the rotation is 24 degrees. The thickness of the Soleil plate is so 
chosen that this average yellow ray is extinguished in the analyzer. 
This corresponds to a rotation of 90 degrees, or to a thickness of 3.75 
mm. (90 -5- 24 = 3.75) for the double plate, when the end point is taken 

for parallel Nicols. If a plate of the 
above description be inserted between 
two parallel Nicols and examined with 
white light, the color of the two halves 
will be of a uniform rose color, the 
blending of the spectral colors minus 
the yellow. The relationship of the 
_j_ 60 angular rotations for red, yellow, and 

blue in the two halves of a 3.75 mm. 

CYellow-90J + 90) Extinguished P late at the transition point may be 

seen from Fig. 56. By rotating the 

Blue 120 j +120 analyzer to the right or left the uniform 

rose color of the plate will change, one- 
Fig. 56. Showing principle of half to blue and the other to red > or 
Soleil double quartz plate. vice versa. If a solution of an optical 

active substance be placed in the tube 

before the analyzer, the equilibrium in color of the transition tint will 
be destroyed and the two halves of the field will be differently colored. 
Rotating the analyzer to the point where the transition tint is again 
produced will give the angular rotation of the solution. 

The Robiquet polariscope, which has a sensibility of about db 4 min- 
utes, is of course only adapted for white light. The rotation angle (a) 
of a substance for extinction of the mean yellow ray was expressed 
by Biot as / (j = French, jaune; yellow). The fact that the point j 
corresponds to no well-defined line of the spectrum makes it a difficult 
one to verify, and some confusion has resulted from this cause. Landolt 
gives for 1 mm. quartz, a/ = 24.5 degrees instead of 24 degrees. The 
value a.j is always greater than a D (the rotation angle for the D ray of 

24 5 

sodium). The relationship given by Landolt is ,- = i ct a D = 1.128 a D ; 

Zi.iZ ' 

using the value 24 degrees / = l.W5a D . Many authorities employ 
the factor 1.111. 

In the examination of colored solutions, the transition tint of the 
Soleil double plate is affected to such a degree that a considerable error 
is introduced in the observation. The use of this end-point device is 
valueless for the color-blind. For these reasons the transition-tint 
polariscopes are at present but little used. 




THEORY AND DESCRIPTION OF POLARIMETERS 89 

Jellefs Half-shadow Polariscope. Efforts to obtain a polariza- 
tion apparatus which would be free from the defects of those previously 
named led Jellet* in 1860 to the construction of the first half -shadow 
polariscope. In this type of end-point adjustment, which can be 
secured in a variety of ways, the field of vision is divided into two or 
more parts, which at the zero position of the analyzer have a uniform 
shade. Rotating the analyzer to the right will cause one section of 
the field to become darker and the other lighter; rotation to the left 
will produce the opposite effect. 

The half-shadow device of Jellet consists of a rhombohedron of 
calc spar with its ends cut square and bisected lengthwise by a plane 
forming a small angle with the axial plane of the prism; the two halves 
are then cemented together in the reversed position, the result being 
that the axial planes of each part are no longer parallel but are tilted 
toward one another at a slight angle. This reunited prism, placed 
between the polarizer and analyzer with its line of union bisecting the 
field, causes the planes of vibration of light proceeding from the polar- 
izer to be slightly inclined towards one another in each half of the 
field. Rotating the analyzer until it is crossed with the polarizer will 
not produce extinction, but a uniform shadow or penumbra whose depth 
will depend upon the inclination of the axial planes in the two halves 
of the Jellet prism. 

Jellet-Cornu Prism. The Jellet polarizer was modified by 
Cornuf by taking an ordinary Nicol prism and dividing it length- 
wise by a plane passing through the shorter diagonal of the end. A 
small wedge-shaped section is then removed from each cut surface and 
the two halves reunited (see Figs. 57 and 58). This "split" or "twin" 
prism combines the effect of an ordinary Nicol and Jellet prism. 

The Jellet-Cornu prism was still further simplified by bisecting 
only one-half of the Nicol prism in the way described. The three 
pieces are then cemented together and the prism squared and mounted, 
with the split half turned toward the analyzer. This form of prism, 
sometimes called the Schmidt and Haensch polarizer, was formerly 
much used in the construction of half-shadow saccharimeters.t 

The principle of the half-shadow device of Jellet and its modifica- 
tions may be seen from Fig. 59. 

Let GO and HO represent the directions of the axial planes in each 
half of the Jellet prism, forming with each other the angle GOH (the 

* Rep. Brit. Assoc., 29, 13 (1860). 

t Bull. soc. chim. [2], 14, 140 (1870). 

t Landolt, " Das optische Drehungsvermogen " (1898), p. 307. 



90 



SUGAR ANALYSIS 



half-shadow angle designated by a and made usually not to exceed 
10 degrees). It will be seen that with the axial plane of the analyzer 
perpendicular to PO the light from the polarizer will not be completely 
extinguished in the analyzer; a small amount of light will emerge 





End of Nicol prism 

before and after 

splitting. 



Showing construction of a Jellet-Cornu prism. 

BDE and BDF, wedge sections removed. 

GE and H F, directions of axial plane before cutting. 

GK and HK, directions of axial planes after uniting cut surfaces. 

from each half of the field proportional to the amplitudes OM and ON 
(see Fig. 49). The equality of light in the two divisions of the field 
constitutes the end point. By rotating the analyzer to the position 
A'L' perpendicular to HO, the light in the right half of the field will be 




Fig. 59. Illustrating principle of Jellet's half-shadow polariscope. 

completely extinguished, and that in the left half will be increased 
from OM to OM 1 '; similarly, with A"L" perpendicular to GO the light 
in the left half of the field is extinguished and that in the right half in- 
creased from ON to ON'] it is evident from the above that the half- 
shadow angle GOH can be measured by the angle A'OA" through which 



THEORY AND DESCRIPTION OF POLARIMETERS 91 

the analyzer is rotated between the points of extinction in the two 
halves of the field. (For appearance of field at the several points see 
Fig. 61.) 

There are several types of polariscopes which use the Jellet-Cornu 
polarizer for an end point. All of these have the advantage that they 
can be used with either mixed or homogeneous light, but the disadvan- 
tage that the half-shadow angle is fixed and cannot be changed to suit 
the requirements demanded by different kinds of work. The sensi- 
bility of the instrument to slight changes of rotation becomes greater 
as the half-shadow angle of the polarizer is made smaller; but, on the 
other hand, the loss of light at the end point produced by decreasing 
the inclination of the planes in the two halves of the field lessens the 
usefulness of the instrument in polarizing dark-colored solutions. 

Laurent's Half-shadow Apparatus. To overcome the last- 
named defect of the Jellet-Cornu polarizer, Laurent * in 1877 con- 
trived an end-point device in which the half-shadow angle could be 
changed to suit varied requirements. The Laurent polariscope has 
the ordinary arrangement of Nicol prisms for polarizer and analyzer, 
the only difference being that the polarizer is attached to a small lever 
by which it can be rotated through a small angle to the right or left. 
The essential part of the end-point device is a thin plate of quartz cut 
perfectly plane and exactly parallel to its optical axis. This plate, 
which must be of specially prepared thickness, is mounted upon glass 
in such a way that it covers one-half of the field of vision. The rays 
of light from the polarizer on entering the plate are resolved into two 
components, one (the ordinary) vibrating in the plane of the optical 
axis, and the other (the extraordinary) in a plane perpendicular thereto. 
The extraordinary component, being less refracted, is transmitted more 
rapidly, and the thickness of the quartz plate is so regulated that when 
the two components emerge, the extraordinary one is in advance of the 
ordinary by half a wave length. The thickness of the plate depends 
upon the wave length X of the light, which must necessarily be homo- 
geneous. The component rays which emerge from the quartz plate 
with half a wave length's (or uneven multiple thereof) difference in vibra- 
tion are resolved by the analyzer into light which at the end point 
is of the same amplitude and intensity as that in the uncovered half 
of the field (the loss of light in the quartz plate by reflection and ab- 
sorption being negligible). The two planes of vibration, which are in- 
clined towards each other equally and symmetrically with reference to 
the optical axis of the plate, form the angle of the half shadow. The 
* Dingler's Polytech. Jour., 223, 608 (1877). 



92 



SUGAR ANALYSIS 



principle of the Laurent plate can be better understood from the 
following diagram (Fig. 60). 

Let LMNK represent the quartz plate with the edge MK bisecting 
the circular field, MK being assumed for convenience to coincide with 
the optical axis of the plate. Let A A' represent the plane of the 
analyzer at the end point and PP f the plane of the polarizer, the latter 
being set at the angle POM with the optical axis MK. Lay off OB as 




Fig. 60. Showing principle of Laurent's half-wave plate. 

the amplitude of the homogeneous light emerging from the polarizer 
and draw BC A.AA', then OC will represent the amplitude of the 
light emergent from the analyzer for the uncovered half of the field 
(Fig. 49). The light of amplitude OB upon entering the quartz plate 
is resolved into two components, one of which OF (the ordinary ray) 
vibrates in the plane of the optical axis MK, and the other OC (the ex- 
traordinary ray) vibrates in the plane OS _L MK. The quartz plate is 
of such thickness that the extraordinary component entering at the phase 
o> is accelerated in its passage one-half wave length and emerges at the 
opposite phase '. The amplitude OC' being equal to OC, the resultant 
OB', between OC' and OF, is equal to OB, and the angle B'OM equal 
to the angle BOM, the two together being the angle of the half-shadow. 
The light emergent from the analyzer in both halves of the field will 
therefore be equal in amplitude and intensity for any angle at which 
PP' may be set with reference to MK. On rotating the analyzer 
from its position, the equilibrium in shade between the two halves will 



THEORY AND DESCRIPTION OF POLARIMETERS 93 

be destroyed (Fig. 61),* the effect being the same as that described 
under Fig. 59. 

The Laurent polariscope, which is the standard instrument in 
France, has the great advantage, over other forms, of adjustable sen- 
sibility without change in zero point, but the great disadvantage of 
being adapted to only monochromatic light. It cannot be used with 




TI 



Fig. 61. Showing divisions of double field of a half -shadow polariscope. 

I, analyzer crossed with left half of field; 
II, analyzer crossed with right half of field; 
III, end point. 

white light except when adapted to bichromate filtered light for a 
quartz wedge saccharimeter. With intense illumination and a small 
half-shadow angle (the conditions of greatest sensibility for all half- 
shadow instruments), the average error of observation according to 
Landolt is less than 1 minute. 

Concentric Half -wave Plate. Pellin has modified the Laurent polari- 
scope by using a half-wave plate of quartz cut in circular or annular 
form. The field of vision is in this way divided concentrically as shown 
in Figs. 62 and 63.* While the concentric field may secure a more correct 





O 






Fig. 62 Fig. 63 

Concentric double field. Concentric triple field. 

alignment of the eye with the optical axis of the polariscope, it is much 
more fatiguing to the eye than the ordinary bisected field. The prin- 
ciple of the concentric half-wave plate is the same as that of the 
Laurent plate. 

* In Figs. 61, 62, 63, and 67b the dividing lines of the fields at end point are 
much intensified. With a properly adjusted instrument the dividing lines com- 
pletely disappear at end point leaving a plain disk of uniform shade. 



94 



SUGAR ANALYSIS 



Lippich's Half -shadow Polarimeter. In 1880 Lippich* davised a 
form of polarizer which combines the advantages of adjustable half- 
shadow and of adaptability to all kinds of light. 
The Lippich polarizer consists of two Nicol prisms, 
one large Nicol, which can be rotated about its 
long axis according to the needs of sensibility, 
and one smaller Nicol, known as the " half-prism," 
which is mounted in front of the large Nicol so 
as to cover one-half of the field. The half-prism 
is slightly tilted so that its inner vertical edge 
forms a sharply dividing line, which can easily 
be focused by the eyepiece of the instrument 
(Fig. 64). 

The principle of the Lippich polarizer can be 
understood by referring to the opposite diagram 
(Fig. 65) : 

Let OP be the plane of the large Nicol and OH 
the plane of the half-prism, the included angle 
POH being that of the half-shadow a. Let OB = 
the amplitude of the light emergent from the large 
Nicol. Draw BG _L OH. Then OG will represent 
the amplitude of the light emergent from the half- 
prism. It can readily be seen that with a loss of 
Fig. 64. Showing con- a part of the light in the half -prism the ampli- 
struction of a Lippich tudeg QQ> and Q D > in the two halves of the field 
polarizer for double -. *. , ^ A , , ,, 

do not agree when the perpendicular OA to the 



half 




N = large Nicol; 
n = small Nicol or 
prism; " 



plane of the analyzer bisects the half-shadow a. 

By rotating the analyzer slightly from L'M' to 

LM the amplitudes OC and OD are made equal, 

Z>= margin of diaphragm; m which position the perpendicular OA no longer 
F = projection of field. bisects a. The angle d which the perpendicular 

OA makes with the bisector OA ' will vary accord- 
ing to the size of the half-shadow angle a. The Lippich polarizer is 
therefore not symmetrical, which is a disadvantage, since by chang- 
ing the half-shadow a to vary the sensibility there is also a change 
in the zero point of the analyzer. The latter must accordingly be re- 
adjusted for each change in sensibility. 

The relation of intensities in the light emerging from the large 
and small prisms of the Lippich polarizer is found as follows: 



* Z. Instrument., 2, 167; 14, 326. 



THEORY AND DESCRIPTION OF POLARIMETERS 



95 



OG 

^ = cos Z BOG cos a. If / and /' are the intensities for the large 

and small prisms respectively, then 

T/ f)C^ 

-j = o = cos 2 a and I' I cos 2 a. (1) 

2 OB 




C O D r ~ 

Fig. 65. Illustrating principle of Lippich polarizer. 



-M' 



The relation between the angle of the half-shadow a and that of 
the change in zero point 5 may be calculated as follows : When the two 
halves of the field are matched the amplitudes OC = OD and the inten- 
sities OC = 

OC 
OB 



sin Z CBO = sin Z POA = sin ( - 



gg = sinOOD = sin Z #04 

s-> 



-sin(f + ). 



(2) 



OB 



(3) 



96 SUGAR ANALYSIS 

Substituting / and I' for OB and OG , we obtain 



OD 2 = sin 2 fe + A I'; since OC* = ~6T) for the matched field, we 
obtain 



(4) 

sin 2 (f ~ ^) = sin 2 (| + ) j = sin 2 (| + ) cos 2 a. (5) 

sin cos 5 cos - sin 8 = sin = cos 5 cos a + cos ^ sin 5 cos a. 
Dividing by cos ~ cos 8, we obtain 

tan = tan 5 = tan - cos a. + tan 5 cos a. 

- 2 

a 1 COS a , a 

tan 6 = tan ^ -=. = tan 3 ^ (6) 

2 1 + cos a 2 

In the above calculation only the light extinguished in the small 
Nicol has been considered. There are other factors, however, which 
must be taken into account in the calculation of the true zero-point 
correction. Schonrock* has shown that 7.5 per cent of the light is lost 
by reflection from the surface of the small Nicol, and that this amount 
is increased to 8 per cent or more by the loss through absorption. 
Equation 1 for intensity would then become 

WO Cl \J \J \ ' / 

The value of 8 thus modified would be expressed by 

1 - cos a A/O92 , a 

tan 6 = -. tan -. (8) 

l+cosaVo.92 2 

Bates f has shown, however, that a part of the light lost by reflection 
from the sides of the small Nicol is again restored in the analyzer, and 
that when all factors such as depolarization, size, shape, and inclination 
of the small prism, etc., are taken into account the true value of 8 is 
between those calculated by equations 6 and 8, the exact figure depend- 
ing upon the construction of each individual Lippich system. 

Apart from the disadvantage that the zero point must be corrected 

* Z. Ver. Deut. Zuckerind., 68, 111. 
t Ibid., 68, 821. 



THEORY AND DESCRIPTION OF POLARIMETERS 



97 




for changes in sensibility, the Lippich polarizer is the best for general 
use and the one most sensitive to minute changes in rotation. The 
average error of adjustment, according to Landolt, 
with bright illumination and a half-shadow angle of 
1 degree, is only about 15 seconds (0.004 degree). 

Lippich Polarizer with Triple Field. The sensibil- 
ity of the Lippich polarizer has been almost doubled 
by using two half-prisms in place of one, the system 
being so arranged that the field of vision is divided into 
three parts (Figs. 66 and 67). ^The principle of the 
triple field can be understood by referring to Fig. 67a. 

Let AC, ac, and a'c' represent planes of the large 
Nicol N, and ab and a'b' planes of the half-prisms n 
and n' respectively. It will be seen that ab and a'b' 
must be perfectly parallel in order that the half- 
shadow angles a and a' be equal for both half-prisms, 
an absolute essential if perfectly uniform illumination 
is to be obtained at the end point. It sometimes hap- 
pens that the two half-prisms get out of parallelism 
through jarring of the instrument or expansion and 

contraction of the mountings. There will then be F j g 6 g Showing 

two end points for the half-shadow, according to construction of 
which side the middle of the field is made to agree. Lippich polarizer 
The observer is then obliged either to take but one for tri P le fielch 
of these end points, which is equivalent to reducing A, large Nicol; 
the instrument to an imperfect double field, or else to # and C, small half- 
readjust the planes of the half-prisms to parallelism, p^ 8 '^ O f ^ia 
a most delicate as well as time-consuming operation, phragm; 
For instruments requiring constant use the increase E and F, inner edges 
in sensibility of the triple field can hardly be said to of half-prism 
offset the increased sensitiveness of the polarizer to 
disarrangement. The more simple double-field end- 
point device is much to be preferred for ordinary lab- 
oratory conditions.* 

Lippich Polarizer with Quadruple Field. Lummerf has constructed 
a polarizer with quadruple field (Fig. 68) by placing before the larger 

* Many chemists wrongly use the expressions half-shade and triple-shade in place 
of the terms double field and triple field. The term half-shade or half-shadow, (Ger- 
man, Halbschatten; French, penombre), refers to the depth of shade in the field at 
the end point and not to the division of the field. The expression triple shade is 
meaningless. 

t Z. Instrument., 16, 209. 



which form the 
divisions H, J, 
and K of the triple 
field. 




SUGAR ANALYSIS 




i n 

a b 

Fig. 67. Illustrating principle of Lippich polarizer for triple field. 
I, analyzer crossed with outer divisions of field; 
II, analyzer crossed with inner division of field; 
III, end point. 

Nicol A one large half-prism B, and before the latter two smaller half- 
prisms C and D. The increased complica- 
tion of this form of polarizer has prevented 
its general introduction. 

Wild's Polaristrobometer. Another 
form of polarizing apparatus, whose pecu- 
liarities of construction place it in a class 
by itself, is the jpolaristrobometer invented 
by Wild* in 1864. In this instrument, 
shown in Fig. 69, the polarizer (/) is at- 
tached to a divided circle, K, both being 
rotated by a rod and pinion from the screw 
C around the longitudinal axis of the Nicol 
prism. The end-point device placed at e 
consists of a Savart double plate made up 
of two sections of calc spar each 3 mm. 
thick, cut at an angle of 45 degrees to the 
optical axis of the crystal, and cemented 
together so that their principal sections 
cross at right angles. A diaphragm c with 
cross threads is placed in the focus of the 
objective lens d of the telescope. The an- 
alyzer at a is stationary, being usually 
mounted with its principal section hori- 
zontal and forming an angle of 45 degrees 
Fig. 68. Showing construction with the crossed sections of the Savart 

of Lippich polarizer for quad- plate 

ruple field. 

* " Ueber em neues Polaristrobometer," Bern, 1865. 




THEORY AND DESCRIPTION OF POLARIMETERS 



99 



To determine the zero point of the polaristrobometer, which is first 
illuminated at D with a sodium flame, a tube of water is placed in the 
instrument and the ocular of the telescope focused sharply upon the 
cross threads; the field, except near the end point, consists of a series of 
dark horizontal parallel bands, the so-called interference fringes, which 




Fig. 69. Wild's polaristrobometer. 

upon rotation of the polarizer increase and decrease in intensity; at 
certain points of rotation the bands gradually become paler until, at 
the maximum point of brightness, they are suddenly extinguished in 
the center of the field, leaving only a slightly shaded border at each 
edge (see Figs. 70 and 71). The point at which the shaded borders 
and the extinguished part of the field are symmetrically distributed 
with reference to the cross threads constitutes the end point. In this 



100 SUGAR ANALYSIS 

position the plane of the polarizer is parallel with one of the crossed 
planes of the Savart plate, so that the end point reoccurs every 90 de- 
grees. In case the extinguished part of the fringes is too wide for 
accurate adjustment, the intensity of the light should be diminished 
until the borders of the fringes are brought sufficiently close to the 
reticule. The fringes haye usually a different appearance at each of 
the end points, and also with colored solutions, so that a beginner must 
familiarize himself with the various characters of the field before making 





Fig. 70 Fig. 71 

Showing field of Wild's polaristrobometer. 

Fig. 70. Interference fringes before end point. 
Fig. 71. Interference fringes at end point. 

readings. In case the zero points of the scale and vernier do not coin- 
cide at the end point, the deviation may be noted and applied to the 
readings as a correction, or else they may be set at zero and the instru- 
ment brought into adjustment by gently turning the screw G until the 
proper end point is secured. 

If the polarizer is set at one of the four zero points and a tube of 
sucrose solution be placed in the trough, the interference fringes will 
reappear. The polarizer must then be rotated to the left (opposite to 
the rotation of the sugar solution) until the fringes again disappear. 
The angular displacement of the polarizer to the left gives the angular 
rotation of the sucrose solution to the right. The readings are made 
through a telescope P which is focused upon the fixed vernier J; the 
latter is illuminated by a flame at Q. The average error of adjustment 
according to Landolt is about 3 minutes. 

The divisions of the scale upon the Wild polaristrobometer are made 
usually in both circular degrees and in degrees of a sugar scale giving 
percentages of sucrose. The sugar scale is constructed by dividing 
53.134 circular degrees into 400 equal parts. Each of these sugar 
divisions corresponds to the rotation of 1 gm. of sucrose dissolved to 
1000 c.c. and polarized in a 200-mm. tube; 10 gms. of pure sucrose dis- 



THEORY AND DESCRIPTION OF POLARIMETERS 101 

solved to 100 c.c. will indicate the 100-degree point of Wild's scale, 
20 gms. sucrose dissolved to 100 c.c. will indicate the 200-degree point, 
30 gms. the 300-degree point, and 40 gms. the 400-degree point. The 
normal weight of the sugar scale of the Wild polaristrobometer can 
therefore be varied according to the concentration of the product to 
be examined, the readings obtained with the 20-gm., 30-gm., and 40-gm. 
normal weights being divided by 2 or 3 or 4, as the case may be. 

The Wild polaristrobometer, although formerly used in many 
European laboratories, finds at present but limited application in tech- 
nical sugar analysis. 

DESCRIPTION OF STANDARD MODERN POLARIMETERS 

The concluding parts of this chapter will be devoted to descriptions 
of a few standard forms of modern polarimeters. 

Laurent's Polarimeter. As a type of instrument of French manu- 
facture the Laurent polarimeter is shown in Fig. 72. 




Fig. 72. Laurent's polarimeter. 

A. A duplex Laurent sodium burner placed 200 mm. from B. 

B. Illuminating lens. 

C. Quadrant whose outer circle is divided into circular degrees and whose inner 

circle is divided into sugar degrees. 

D. Diaphragm containing half -wave plate of quartz. 

E. Light filter consisting of a crystal of potassium bichromate. 



102 SUGAR ANALYSIS 

F. Screw for adjustment of zero point. 

G. Geared screw for rotating the analyzer and the arm supporting the verniers. 

The upper vernier on the right is for reading circular degrees and the lower 
vernier upon the left for reading sugar degrees. 

L. Bronze trough 600 mm. long for holding observation tubes. 

M. Mirror for illuminating scale. 

N. Magnifying glass for reading scale. 

R. Tube section containing polarizer; the latter can be moved through a small 
angle by the arm K, which is moved by the crank J through the rod X by means 
of the lever U. If the solution to be examined is but little colored, the lever U 
is raised, which decreases the half-shadow angle. With colored solutions U is 
lowered until the half-shadow is increased to the point of greatest sensibility. 
The zero point should be redetermined after each change in the position of 
the polarizer. 



The 100-degree point of the sugar scale of the Laurent polarimeter 
corresponds to an angular rotation of 21.67 degrees (21 40'), which 
is the value given by French authorities to the angular rotation of the 
1 mm. thick plate of quartz cut perpendicular to the optical axis (see 
page 112). The normal weight of sucrose corresponding to this rota- 
tion is given as 16.29 gms. dissolved to 100 c.c. and polarized in the 
200-mm. tube. The sugar scale extends 400 divisions to the right and 
200 divisions to the left, thus giving ample range for polarizing all 
dextro- and levo-rotatory sugars. If desired, the sugar scale of the 
Laurent polarimeter is adjusted according to the so-called Interna- 
tional saccharimetric scale of 20 gms. The value of the 100-degree 
division of the International scale in circular degrees would equal 

21 67 X 20 

' = 26.605 degrees; this is a trifle more than the circular value 
lo.^y 

of the Wild 20-gm. scale, viz., 26.567 degrees, the difference being due 
presumably to the adoption of a slightly different standard value for 
the specific rotation of sucrose. 

Pellin's Polarimeter. Another type of French polariscope is the 
half-shadow polarimeter-saccharimeter made by Pellin, shown in Fig. 
73. The polarizer of this instrument consists of a modified Jellet-Cornu 
prism; the half-shadow angle is therefore fixed. The division of the 
quadrant into circular and sugar degrees is identical with that of the 
Laurent polarimeter. 

The Pellin polarimeter with variable half-shadow angle (Fig. 74) 
makes use of a half-wave plate of quartz for the end point, which is 
constructed for either divided or concentric fields. The arrangement 
of optical parts and method of manipulation are the same as in the 
Laurent polarimeter. 



THEORY AND DESCRIPTION OF POLARIMETERS 103 




Fig. 73. Pellin's polarimeter with Jellet-Cornu prism. 




Fig. 74. Pellin's polarimeter with half -wave plate. 



104 SUGAR ANALYSIS 

Lippich's Polarimeter. A simple form of Lippich's polarimeter 
adapted for general chemical use is shown in Fig. 75. Angular rotations 
can be measured with this instrument to about 0.015 degree. 




Fig. 75. Simple form of Lippich's polarimeter. 

h. Lever for moving large Nicol of polarizer and regulating sensibility. The half- 
shadow angle which is read by the scale can be varied from degrees to 20 de- 



K. Divided circle for measuring rotation. The circle with analyzer in A and 
telescope at F is rotated by the screw T. The readings of the scale are made 
on each side of the circle through the lenses I, which are focused upon the fixed 
verniers at n. 

P. Location of Lippich polarizer. 

S. Detachable end for holding light filter. 

A form of the Lippich apparatus devised by Landolt for more general 
use is shown in Fig. 76. This instrument presents an advantage in 
that any form of tube or container may be used for holding the solution 
or substance to be polarized. 

The trough D of the polariscope for holding ordinary tubes can be 
removed and the support T employed. The latter is raised or lowered 
by the screw q and moved laterally upon the tracks c. For polarizing 
materials in hot or cold condition, the apparatus G, consisting of a 



THEORY AND DESCRIPTION OF POLARIMETERS 105 




Fig. 76. Landolt's polarimeter for general use. 

g. Lever for rotating circle R', the final adjustment is made by means of the microm- 
eter screw m after fixing the clamp k. 
P. Position of Lippich polarizer with two half-prisms giving triple field. 




Fig. 77. Large model Landolt polarimeter. 



106 SUGAR ANALYSIS 

polariscope tube in an asbestos-jacketed bath, is employed. The plate 
T is then removed and the bath placed directly upon the tracks c. The 
burner for heating the bath is placed upon the adjustable stand under- 
neath. The center narrow tube projecting through the replaceable top 
of the bath receives the overflow from the observation tube; the other 
tubes serve for a thermometer and stirrer for the liquid of the bath. 
For polarizing at low temperature a cooling medium is used in the bath, 
in which case the ends of the observation tubes must be covered with 
desiccating caps to prevent condensation of moisture upon the cover 
glasses. 

A type of more elaborate polarimeter, which can be read to 0.01 
degree, is the large Landolt instrument shown in Fig ; 77. The divided 
circle (driven by the wheel T and micrometer screw m) is covered by a 
cap K. Small mirrors Si and S 2 reflect light from the observation 
lamp through openings in the cap to illuminate the scale. A feature of 
this instrument is the double trough by which different tubes of solu- 
tion can be brought into the field by movement of the large lever H. 

VERIFICATION OF SCALE READING OF POLARIMETERS 

The graduations of the divided circle upon a polarimeter should be 
verified by taking check readings at different points upon opposite 
sides of the disk. The division and mounting of the circle in the best 
instruments is made with great accuracy, and, unless the disk has been 
warped or bent, check readings on opposite sides of the circle will agree 
much closer than the observer can set the scale for a matched field. 

Polariscope readings should always be verified upon the opposite 
scale. It is also well to reverse the circle 180 degrees and repeat the 
readings each way from the other side. By so doing the observer will 
have 4 sets of readings, the mean of which will practically eliminate all 
errors due to faulty scale division or eccentricity. The example on page 

107 of readings made upon a sugar solution will illustrate the method. 
The adjustment of the halt-shadow angle is made to the point of 

greatest sensibility, the angle being small for light-colored solutions and 
larger for dark liquids. Since altering the half-shadow of the Lippich 
system produces a change in zero point (p. 95), the adjusting lever 
should never be disturbed during a set of observations. The analyzer, 
if desired, can be brought back to the of the scale for any change in 
the half-shadow angle by means of a small regulating screw (shown at 
a, Fig. 77). The better method, however, is to establish the zero point 
upon the scale, as in the following example, and subtract this from the 
scale reading. 



THEORY AND DESCRIPTION OF POLARIMETERS 107 





Zero point. 


Sugar solution. 






Right. 


Left. 


Right. 


Left. 






r 


3.07 


183.07 


29.30 


209.295 


>> 






3.09 


183.085 


29.28 


209.28 








3.11 


183.11 


29.295 


209.29 


* 






3.08 


183.075 


29.27 


209.28 




Half-shadow 

fln0 .1 p _ o^ 




3.10 


183.10 


29.285 


209.29 


Temperature 
" 20 C 


Average . 




3.09 


183.088 


29.286 


209.287 












3.090 


183.088 






[ 






26.196 


26.199 


, 




( 


183.075 


3.08 


209.270 


29.265 


1 






183.10 


3.10 


209.285 


29.28 








183.08 


3.085 


209.28 


29.28 




Reversing 
the circle < 




183.09 
183.09 


3.09 
3.095 


209.27 
209.285 


29.27 
29.285 


Temperature 

r 21 C 


180. 




183.087 


3.090 


209.278 


29.276 












183.087 


3.090 






^ 






26.191 


26.186 


J 



Average of 4 readings, 26.193 for 20.5 C. 



CHAPTER VI 

THEORY AND DESCRIPTION OF SACCHARIMETERS 

WHILE the instruments described in the previous chapter are 
adapted to the examination of all optically active substances, sac- 
charimeters are designed solely for polarizing sugars. For convenience 
the scale expressing angular rotation is replaced upon the saccharimeter 
by one graduated according to the decimal system indicating percent- 
ages. 

THE QUARTZ-WEDGE COMPENSATION 

Owing to the many difficulties and inconveniences connected with 
the use of sodium or other monochromatic light in practical work, the 
French physicist Soleil was led in 1848 to devise a means by which 
ordinary daylight or lamplight could be used for measuring the optical 
rotation of sugar solutions. This invention, known as the quartz- 
wedge compensation, is the characteristic feature of all saccharimeters. 

In the quartz-wedge saccharimeter the polarizer and analyzer are 
both stationary; the rotation of the sugar solution is measured by 
shifting a wedge of optically active quartz between the solution and 
analyzer until the rotation of the wedge system at a certain thickness 
exactly neutralizes or compensates the rotation of the sugar solution. 
By means of a scale attached to the quartz wedge the rotation of the 
sugar in solution is measured in percentage. 

The selection of quartz for compensation is based upon the fact 
that it has almost exactly the same rotation dispersion as cane sugar; 
i.e., a section of quartz and a cane-sugar solution of equal rotation for 
light of one wave length will have very nearly equal rotations for light 
of all other wave lengths (see Table XX). The small disturbances 
due to the slight difference in rotation dispersion between sugars and 
quartz are eliminated by a bichromate light filter. 

Single-wedge System. The quartz wedges used in the con- 
struction of saccharimeters are cut perpendicularly to the optical axis 
of the quartz crystal; they may be either of dextrorotatory or levo- 
rotatory quartz, the method of mounting the wedge depending upon 
the character of the rotation. This can be seen more clearly by in- 
specting the following diagrams (Fig. 78). 

108 



THEORY AND DESCRIPTION OF SACCHARIMETERS 109 



In diagram I, A is a fixed plate of levorotatory quartz, and R and 
C two wedges of dextrorotatory quartz, of which B is movable and C 
stationary. The two wedges, wljich though of different size must have 
equal angular dimensions, may be considered to form together a single 
section with sides parallel to the plate A and perpendicular to the 
axis of light through the instrument. The thickness of the two wedge 
sections can be increased or diminished by moving wedge B to the 
right or left. At the zero point of the instrument the right rotation of 



Dextro-rotatory wedge 
system 




Levo-rotatory wedge 
system 




u 



II 

Fig. 78. Showing construction of single wedge quartz compensation. 

the section Imno of the two-wedge system exactly neutralizes the left 
rotation of the quartz plate A. If a tube of dextrorotatory sugar 
solution be now placed in the instrument between the polarizer and the 
compensation plate A, the optical neutrality is destroyed, and it will 
be necessary to decrease the thickness of the two-wedge section by 
sliding B from o> towards ' until the excess of left rotation in A over B 
and C exactly neutralizes the right rotation of the sugar solution. If the 
solution of sugar is left-rotating, it will be necessary to slide B in the 
opposite direction until the excess of right rotation in B and C over A 
equals the left rotation of the sugar. In a levorotatory wedge system 
(diagram II) the compensation plate A is dextrorotatory and the wedges 
B and C levorotatory, the compensating motion of wedge B being the 
reverse of that in diagram I. 

Owing to the lateral refraction of light from the inclined surfaces 
of the wedges through the intervening air space (as shown by the 
dotted line efg), the planes of quartz are separated only just sufficiently 
to allow free movement of the parts without friction. The circum- 
stance that the field is not exactly at the end point, when the thickness 
of the two-wedge section agrees with that of the compensating plate, 
is due to this lateral refraction. The shifting of zero point due to re- 
fraction depends upon the wave length of light; the difference in zero 



110 



SUGAR ANALYSIS 



point between red light of 760 MM wave length, and violet light of 396.8 
WJL wave length was found by Schonrock to be 0.059 degree for the 
Ventzke sugar scale. 

The scale of the saccharin) eter is attached to the large or movable 
wedge, and is read by means of a vernier scale attached to a regu- 
lating screw. In case the zero marks of the two scales do not agree, 
when the two halves of the field correspond in shade, they can be 
brought into coincidence by shifting the vernier slightly to the right or 
left by means of a key which fits the regulating screw. The vernier 
is never to be moved except for making this adjustment, and when the 
two scales are once set has rarely to be disturbed. Owing to the in- 
evitable slight fluctuations in the zero point of saccharimeters, it is best 
to correct the reading by the zero-point error and not to adjust the 
scale unless there be a persistent difference of the zero point in one 
direction greater than 0.1 degree. The method of reading the sac- 
charimeter scale can be seen from Figs. 80 and 81. 

Double-wedge System. An elaboration of the quartz-wedge 
system just described is the double- wedge compensation introduced 

by Schmidt and Haensch. The 
arrangement of the parts in the 
double-wedge system is shown in 
A Fig. 79. 

In the double- wedge system the 
B compensation plate is lacking, this 
being supplied by one or the other 
c of the pair of wedges, which are 
of opposite rotation. The smaller 
D wedges A and D are stationary and 
the larger wedges B and C mova- 
ble. B and C are usually mounted 

p. - n f , , , with their points in the same direc- 

Fig. 79. Showing construction of double r 

wedge quartz compensation. tlon m order to equalize the refrac- 

tion of the light rays in the air 

spaces between the inclined surfaces of quartz (as indicated by the 
dotted line) ; for this reason also the corresponding wedges of each sys- 
tem are made as near alike as possible. Each of the large wedges is 
provided with a scale. These may be read through the same telescope 
as upon the Schmidt and Haensch saccharimeter (Fig. 80), or by sepa- 
rate telescopes as in the Fric instruments (Fig. 81). 

In using the double-wedge system for dextrorotatory substances, 
the scale K (Fig. 80) is set at zero with its vernier, and the optical rota- 




THEORY AND DESCRIPTION OF SACCHARI METERS 111 



tion measured upon the scale A; for levorotatory solutions, A is set 
at zero and the scale K employed. An additional advantage of the 
double-wedge system consists in the fact that any reading obtained upon 




PATENT 
JOSEF & JAN ERIC 




Fig. 80. Scale of double wedge 
Schmidt and Haensch saccharimeter. 
K, control scale; 

A, working scale indicating 85.5 de- 
grees Ventzke. 



Fig. 81. Scale of Fric saccharimeter 
with double vernier indicating 97.7 
degrees Ventzke . (The division be- 
tweenscale and vernier is intensified ; 
in reality no dividing line is seen.) 



the working wedge can be immediately verified by removing the tube 
of solution and moving the control wedge to the point of compensation. 
The control wedge under 'such conditions gives the true reading directly, 
even though the working wedge have a zero-point correction. 



Zero-point determination. 



Polarization of mat sugar. 



Control-wedge 

scale. 


Working-wedge 
scale. 


Difference. 


Control-wedge 
scale. 


Working-wedge 
scale. 


Difference. 


0.00 


0.10 


+0.10 


0.00 


89.40 


89.40 


11.55 


11.65 


+0.10 


0.75 


90.15 


89.40 


20.75 


20.80 


+0.05 


2.15 


91.50 


89.35 


32.20 


32.30 


+0.10 


2.90 


92.30 


89.40 


43.75 


43.80 


+0.05 


3.85 


93.25 


89.40 


52.50 


52.55 


+0.05 


5.45 


94.85 


89.40 


61.85 


61.95 


+0.10 


6.55 


96.00 


89.45 


70.50 


70.60 


+0.10 


7.95 


97.30 


89.35 


81.15 


81.30 


+0.15 


9.10 


98.45 


89.35 


91.15 


91.25 


+0.10 


10.15 


99.55 


89.40 


Average zero point +0.09 


Average polarization un- ) 89 . 39 




corrected ( 




Zero-point correction = 0.09 




Corrected polarization = 89.30 



112 SUGAR ANALYSIS 

Zero-point Determination. The zero-point correction of the work- 
ing wedge can be determined very accurately by taking check readings 
at different parts of the scale upon the control. By making polariza- 
tions in the same way, the local defects of scale or wedge will be almost 
wholly eliminated. The readings in this case are made without re- 
moving the tube, the difference between the two scales being the 
uncorrected polarization. The preceding table, giving the readings upon 
the working-wedge scale for various positions of the control, will illus- 
trate the method. 

THE SUGAR SCALE AND NORMAL WEIGHT OF SACCHARIMETERS 

The 100-degree point of a saccharimeter scale is usually based upon 
the rotation of a definite weight (the so-called normal weight) of chemi- 
cally pure sucrose dissolved in water to 100 c.c. at a specified temperature 
and polarized at the same temperature in a 200-mm. tube. The greatest 
confusion has prevailed in saccharimetry in the past, and unfortunately 
still prevails, not only as to the size of the normal weight of sugar to be 
taken for a specified scale, but also as to the conditions of volume and 
temperature under which this normal weight is to be polarized. 

French Sugar Scale. The 100-degree point of the sugar scale 
employed upon saccharimeters of French manufacture is based upon 
the rotation in sodium light of a plate of dextrorotatory quartz 1 mm. 
in thickness and cut exactly perpendicular to the optical axis. The 
choice of quartz as a standard proved to be unfortunate, for, owing 
either to mistakes of polarimetric measurement or to defects in the 
quartz (through natural imperfection or mistakes in cutting), the 
rotation of the 1-mm. plate has been given a different value from time 
to time, the results ranging from +20.98 degrees, the early figure of 
Biot, to +22.67 degrees. Most French authorities at present employ 
the value +21.67 degrees. The figure, regarded usually as the most 
exact, is that of Landolt, who, for spectral pure Na light of mean wave 
length 589.3 MM, found the value +21.723 degrees. The grams of 
sucrose necessary to give the same rotation in 100 c.c. as the 1-mm. 
quartz plate have also necessarily varied; over 20 different values have 
been assigned to this quantity, the amounts ranging from 16.000 gms. 
(Dubrunfaut) to 16.471 gms. (Clerget and Biot). The cause of these 
great differences is due partly to variations in the quartz standard and 
partly to variations in the purity of the light used for illumination. 

The old normal weight established for French instruments was 
16.35 gms., and this weight is still largely used in technical work with 
the Soleil-Duboscq saccharimeter. In 1875 the value of Girard and 




THEORY AND DESCRIPTION OF SACCHARI METERS 113 



de Luynes, 16.19 gms., was adopted as the official weight and remained 
such for more than 20 years, notwithstanding the severest criticism. 
In 1896 the International Congress of Applied Chemistry at Paris 
established the value of 16.29 gms. sucrose dissolved at 20 C. in 100 
metric c.c., and this is the official weight used at present by the 
French Ministry of Finance. 

Ventzke or German Sugar Scale. The sugar scale most generally 
used outside of France and the one employed upon all German sac- 
charimeters is that of Ventzke. This scale as originally devised by 
Ventzke * was based upon the rotation of a solution of pure sucrose of 
1.1 sp. gr. Y^ . It was soon found, however, inconvenient, as well as 
inaccurate, to make the specific gravity of solution a basis for saccha- 
rimetric work, and the grams of sugar in 100 c.c. of solution 1.1 sp. gr. 
was used for the -normal weight; this was determined to be 26.048 gms. 
weighed in air with brass weights and dissolved at 17.5 C. to 100 
metric c.c. 

Mohr Cubic Centimeter Standard. With the introduction in 1855 
of the Mohr f cubic centimeter (the volume of 1 gm. of water at 17.5 C. 
weighed in the air with brass weights), the original normal weight 
of 26.048 gms., designed for metric cubic centimeters, was strangely 
enough retained and used for determining the 100-degree point of the 
sugar scale. In this way the standard was established which up to 
1900 was the only one recognized for the Ventzke scale, and which at 
the present time is still the one most commonly used in commercial 
work. In accordance with this standard, the 100-degree point of the 
sugar scale is obtained by dissolving 26.048 gms. of chemically pure 
sucrose (weighed in air with brass weights) in 100 Mohr c.c. at 17.5 C. 
and polarizing the same in a 200-mm. tube at 17.5 C. in a saccharim- 
eter whose quartz-wedge compensation has also a temperature of 
17.5 C. This normal weight calculated to 100 metric c.c. (volume of 
100 gms. water at 4 C.) is equal to 26.048 gms. -=- 1.00234 = 25.9872 
gms. (1 Mohr c.c. = 1.00234 metric c.c.). 

Metric Cubic Centimeter Standard. On account of the confusion 
and mistakes resulting from two standards of volume, the International 
Sugar Commission, at its third meeting in Paris, 1900, advocated the 
abandonment of the Mohr for the metric cubic centimeter, and in so 
doing also recommended that the temperature of polarization be made 
20 C. The change in temperature from 17.5 C. to 20 C. necessitated 
a recalculation of the normal weight owing to the difference in specific 

* J. prakt. Chem., 26, 84 (1842) ; 28, 111 (1843). 

f "Chemisch-analytische Titrirmethode " (1886), pp. 44-50. 



114 SUGAR ANALYSIS 

rotation of cane sugar and quartz at these two temperatures. The 
calculation is made by the following equation, in which 0.000184 is the 
coefficient of decrease in specific rotation of sucrose at 20 C., 0.000148 
the coefficient of increase in rotation due to the effect of temperature 
upon wedge and scale, and 0.000008 the coefficient for expansion of the 
glass observation tube: 

2fi 048 

1 1 +(0.000184+0.000148 - 0.000008) (20 - 17.5)} = 26.0082 



gms. The International Commission decided, however, to make the new 
normal weight exactly 26 gms., and in accordance with its recommenda- 
tion the following definition for the 100-degree point of the Ventzke 
sugar scale has been universally adopted: "The 100-degree point of 
the saccharimeter scale is obtained by polarizing a solution containing 
26.000 gms. of pure sucrose (weighed in air with brass weights) in 100 
true c.c. at 20 C. in a 200-mm. tube in a saccharimeter whose quartz- 
wedge compensation must also have a temperature of 20 C." All sac- 
charimeters using the Ventzke scale are standardized at present in 
accordance with this definition. According to Bates and Jackson* a 
solution of chemically pure sucrose under the above conditions gives 
a reading of only 99.89 upon the German scale. 

United States Coast Survey Standard. The old original standard 
of the Ventzke scale was the one adopted by the Department of Weights 
and Measures of the United States Coast and Geodetic Survey, and 
was employed for many years by the United States Treasury Depart- 
ment in the Custom House laboratories. The 100-degree point of the 
scale was taken as the polarization of 26.048 gms. (in vacuo) of pure 
sucrose dissolved to 100 true c.c. of solution at 17.5 C. and polarized 
at this temperature in a 200-mm. tube. To avoid the labor of reducing 
this weight of sugar to vacuo, the flasks employed for the Coast Survey 
standard were graduated to contain 100.06 true c.c., the excess of 
0.06 c.c. being taken to correct the error of weighing the sugar in air 
against brass weights. These flasks contain 0.174 c.c. less than the 
old Mohr cubic centimeter flasks (100.234 true c.c.), which difference, 
unless compensated, would cause the normal weight of 26.048 of pure 
sucrose to polarize 0.17 V. too high. To save the operators the trouble 
of making this correction, the correction of 0.17 was applied to the 
quartz test plates used for controlling the instruments. The computed 
values of the Coast Survey test plates were thus 0.17 V. lower than 
the values marked by the instrument makers for the Mohr cubic centi- 
meter standard. 

* Scientific Paper, U. S. Bureau of Standards, No. 268 (1916). 



, 



THEORY AND DESCRIPTION OF SACCHARIMETERS 115 



The policy of the Department of Weights and Measures of the United 
States Coast Survey, in adopting a standard different from that in 
current use, was unfortunate. It gave rise to much confusion and mis- 
understanding, and traces of this confusion still exist, notwithstanding 
the fact that the United States Bureau of Standards, the Custom House, 
and all other United States Government laboratories have abandoned 
the old Coast Survey standard and now employ the standard of the 
International Commission of 26 gms. to 100 true c.c. at 20 C. 

According to the work of both Sawyer* and Rolfe,f who have made 
comparative readings of standard quartz plates upon various sac- 
charimeters, there are many instruments in the United States, even of 
recent manufacture, which are standardized for a normal weight of 
26.048 gms. in 100 true c.c. Whether this condition of affairs is due 
to a mistaken idea of some manufacturers that the old Coast Survey 
standard is still recognized officially in the United States, is difficult to 
say. It is evident, however, that chemists, in order to avoid the con- 
siderable errors due to confusion in standards, should state explicitly, 
in ordering saccharimeters from manufacturers, that their instruments 
be graduated according to the standard of the International Commis- 
sion. When purchasing second-hand saccharimeters, chemists should 
be particularly careful to subject the same to a thorough examination 
and verification before using. 

Value of the Ventzke in Circular Degrees. The rotation value of the 
100-degree point of the modern Ventzke scale has been very carefully 
determined by Schonrock,t who found it to equal 34.657 circular degrees 
for spectral pure sodium light. This is the value used at present by 
Schmidt and Haensch in the standardization of all their saccharimeters. 
According to Bates and Jackson (page 114) the rotation value of the 
normal quartz plate for pure sodium light is 34.620 circular degrees. 

Bichromate Light Filter. Schonrock|| has shown that in estab- 
lishing the 100-degree point of the Ventzke scale by means of sucrose 
the white light must be filtered through a 1.5-cm. layer of 6 per cent 
potassium-bichromate solution in order to eliminate the errors of rota- 
tion dispersion between cane sugar and quartz produced by the light 
of shorter wave length at the violet end of the spectrum. This light 
filter has been adopted by the Physikalisch-Technische Reichsanstalt 
of Germany and also by the United States Bureau of Standards If in 

* J. Am. Chem. Soc. 26, 990. According to statement in a letter to the author. 

t Technology Quarterly 18, 294. (1905) || Z. Ver. Deut. Zuckerind., 64, 521. 

t Z. Ver. Deut. Zuckerind., 64, 521. 

If Upon its certificates for standardization of quartz plates a sugar degree is thus de- 
fined by the United States Bureau of Standards : " A sugar degree is the one-hundredth 



116 



SUGAR ANALYSIS 



defining the 100-degree point of the saccharimeter scale, and its use is 
imperative for all accurate work. Many saccharimeters have a 3-cm. 
cell, and for this length of liquid a 3 per cent bichromate solution is 
sufficient (centimeter length of cell X per cent bichromate = 9). For 
carbohydrate materials of greater rotation dispersion than cane sugar, 
such as dextrin, commercial glucose, etc., the author has found it 
necessary to use a solution of double the above concentration (centi- 
meter length of cell X per cent bichromate = 18) in order to secure 
constancy of results between different observers for different sources of 
white light. 

In this connection it is important to note that the rotations of the 
normal weight of sucrose with bichromate-filtered white light and with 
sodium light, while very closely agreeing, are not absolutely identical 
owing to the slight differences in optical center of gravity. Measure- 
ments by Schonrock* show that, while a normal sugar solution at 
20 C. for bichromate filtered white light is exactly equal to the rota- 
tion of a quartz plate of 100 V. (34.657 angular degrees), by using 
sodium light a quartz plate of 100.03 V. (34.667 angular degrees) would 
be required. The relationship between Ventzke degrees for bichro- 
mate filtered white light and monochromatic light of different wave 
lengths is seen from the following table :f 

TABLE XX 
Showing Rotation of Quartz and Sucrose for Different Kinds of Light 



Source of light. 


Mean wave 
length /z M . 


Angular rotation, 20 C. 


Degrees 
Ventzke. 


Quartz plate 
(1.595 mm.). 


Sucrose solu- 
tion (26 gms. 
in 100 true cubic 
centimeters in 
200-mm. tube). 


White light filtered through 1.5cm. \ 
of bichromate solution, about . . j 
Spectral pure sodium light. . 


600 

589.3 
551 

546.1 
535 
460.7 
420.2 


34.65 
34.657 
39.82 

40.73 
42.49 

58.65 

71.78 


34.65 
34.667 
39.87 

40.81 
42.67 
59.18 

72.87 


100.00 
100.03 

100.12 

100.19 
100.42 
100.91 
101.52 


White light, Welsbach, unfiltered, ) 
about. ) 


Yellow-green mercury 
Green tantalum 


Blue strontium 


Violet rubidium . 





part of the rotation shown by 26 gms. of sucrose dissolved in water and the volume 
made up to 100 metric cubic centimeters, for light from an incandescent gas mantle 
passed through 1.5 centimeters of a 6 per cent potassium-bichromate solution, the 
temperature being 20 C. for graduation, preparation, and observation." 

* Z. Ver. Deut. Zuckerind., 54, 521. 

t Compiled from results by Landolt and by Schonrock. 



TI 

It is g 






THEORY AND DESCRIPTION OF SACCHARI METERS 117 

It is seen that while the quartz and sugar exactly agree for bichro- 
mate filtered light, the sugar is rotated to a continually greater extent 
than quartz for light of decreasing wave length. The normal sugar 
solution, reading 100 V. with filtered white light, was found to read 
100.12 degrees with unfiltered white light. The eyes of some observers 
are more sensitive than those of others to the disturbances of rotation 
dispersion when unfiltered light is used (owing perhaps to some differ- 
ence in the pigment of the eye), so that for accuracy and constancy of 
results in all saccharimetric measurements the bichromate filter should 
never be omitted.* 

Graduation of Saccharimeter Scales. Manufacturers of sac- 
charimeters in establishing the 100-degree point of their sugar scales 
employ a carefully standardized quartz plate instead of the normal 
weight of sucrose. The errors and inconveniences incident to the 
preparation of chemically pure sucrose and to making the solution up 
to exact volume are thus avoided; the plate, moreover, has the advan- 
tage of being a standard which at constant temperature is always un- 
changeable. Messrs. Schmidt and Haenschf thus describe the method 
of graduating the scales of their saccharimeters: 

" The establishment of the scale divisions of our saccharimeters is 
made at a temperature of 20 C. After fixing the zero point the linear 
distance of the 100-degree division is determined by means of a normal 
quartz plate reading exactly 100 degrees and standardized at the 
Physikalisch-Technische Reichsanstalt. This linear distance is then 
divided into 100 exactly equal parts, the intermediary divisions being 
also verified by means of corresponding normal standardized quartz 
plates. The surfaces of the quartz wedges are made perfectly plane so 
that a quartz stratum of half thickness corresponds to a half value in the 
division. Slight errors cannot be prevented, as it is impossible to 
obtain quartz wedges of the necessary length which are absolutely 
optically homogeneous throughout. The variableness in the specific 
rotation of sucrose with concentration of solution is not taken into con- 
sideration in the establishment of the scale division, and this must be 
corrected for by calculation. Aberrations in the scale division caused by 
impurities in the quartz can be detected by the control observation tube." 

The view that the Ventzke scale of modern saccharimeters is cor- 
rected for variations in specific rotation of sucrose with concentration, 

* At its New York Meeting (Sept. 10, 1912) the International Commission adopted 
the following resolution: " Wherever white light is used in polarimetric determina- 
tions, the same must be filtered through a solution of potassium bichromate of such 
a concentration that the percentage content of the solution multiplied by the length 
of the column of the solution in centimeters is equal to nine." 

t In a letter to the author. 



118 



SUGAR ANALYSIS 



either by curving the surface of the quartz wedges or by unequal spac- 
ing of the scale divisions, is not substantiated by the above statement. 
Effect of Concentration upon Scale Reading. A table has been cal- 
culated by Schmitz* to correct for the changes in specific rotation of 
sucrose through varying concentration, which gives the actual sucrose 
value of each scale division of the saccharimeter. These corrections, 
which were calculated by Schmitz's formula, [a] D = 66.514 0.0084153 c, 
would seem in light of more recent work to require considerable modi- 
fication. The formula of Landolt, 

[afS = 66.435 + 0.00870 c - 0.000235 c 2 , (c = to 65), 

calculated from the combined observations of Tollens, and of Nasini 
and Villavecchia, is regarded as the most accurate at present (see page 
176). In the following table the author has recalculated the sucrose 
values of the Ventzke scale for different concentrations, using Landolt's 
formula. The values of Schmitz are also given for comparison. 

TABLE XXI 
Showing Effect of Concentration of Sucrose upon Saccharimeter Readings 



Scale division. 


Concentration. 
Grams sucrose, 
100 true cubic centi- 
meters, 20 C. 


Specific rotation 
sucrose, 20 C. 


Actual sucrose value of scale division. 


By Landolt's 
formula. 


By Schmitz's 
formula. 


100.00 


26.00 


66.502 


100.00 


100.00 


96.00 


24.96 


66.506 


96.00 


95.98 


95.00 


24.70 


66.507 


94.99 


94.98 


90.00 


23.40 


66.510 


89.99 


89.97 


85.00 


22.10 


66.513 


84.99 


84.96 


80.00 


20.80 


66.514 


79.99 


79.95 


75.00 


19.50 


66.515 


74.99 


74.94 


70.00 


18.20 


66.516 


69.99 


69.93 


65.00 


16.90 


66.515 


64.99 


64.92 


60.00 


15.60 


66.514 


59.99 


59.92 


55.00 


14.30 


66.511 


54.99 


54.92 


51.00 


13.26 


66.509 


50.99 


50.92 


50.00 


13.00 


66.508 


50.00 


49.92 


45.00 


11.70 


66.505 


45.00 


44.92 


40.00 


10.40 


66.500 


40.00 


39.92 


35.00 


9.10 


66.495 


35.00 


34.92 


33.00 


8.58 


66.492 


33.00 


32.93 


32.00 


8.32 


66.491 


32.01 


31.93 


30.00 


7.80 


66.489 


30.01 


29.93 


25.00 


6.50 


66.481 


25.01 


24.94 


20.00 


5.20 


66.474 


20.01 


19.95 


15.00 


3.90 


66.465 


15.01 


14.96 


10.00 


2.60 


66.456 


10.01 


9.97 


6.00 


1.56 


66.443 


6.01 


5.98 


5.00 


1.30 


66.442 


5.00 


4.98 



Ber., 10, 1414; Z. Ver. Deut. Zuckerind., 28, 63, 887. 





It wi 



I 

II 



THEORY AND DESCRIPTION OF SACCHARI METERS 119 

It will be seen from the preceding table that the greatest deviation 
of the actual sucrose value from its scale division according to Landolt's 
equation is only 0.01 V., which is too small to be detected by the 
ordinary saccharimeter. The maximum error according to Schmitz 
is 0.08 V. 

As regards the concentration of sucrose employed in ordinary saccha- 
riinetric work, the variations due to changes in specific rotation may 
therefore be safely disregarded. The small extent of these variations, 
which are distributed both above and below the scale division, justifies 
the policy of the manufacturers in neglecting this factor when estab- 
lishing the divisions of the saccharimetric scale. 

VERIFICATION OF SCALES OF SACCHARIMETERS 

On account of the optical imperfections which quartz wedges occa- 
sionally possess, it is important that every user of a saccharimeter should 
verify the accuracy of his instrument. 

Owing to the fact that the quartz parts of the saccharimeter are 
mounted close to the objective of the telescope, the very local imper- 
fections of the wedge system are fortunately unnoticed, since, when the 
telescope is focused upon the polarizer, the cone of light rays emanating 
from the different parts of the field covers an area of the compensator 
equal to the aperture of the analyzer diaphragm (about 6 mm. diameter) 
and thus distributes and neutralizes any slight local errors due to defects 
of the quartz. Such defects in the fixed part of the system (small wedge 
and compensation plate) are of no account, since the rotatory power of 
this remains constant; the predominant optical defects of the large 
movable wedge are the only ones which vitiate the results of observation. 

Since local optical impurities in the large wedge are diffused over a 
considerable area, for the reason given above, the errors in the sac- 
charimeter scale never consist of sudden jumps, but only of gradual 
undulations. It is unnecessary, therefore, as Landolt has shown, to 
standardize every division of the scale. The errors at every fifth 
degree, if plotted upon coordinate paper, are sufficient to establish a 
correction curve from which the error of any division upon the scale 
can be accurately found (see Fig. 83). 

Verification by Quartz Plates. The simplest and easiest method 
of scale verification, as well as the most accurate, is by means of care- 
fully standardized quartz plates. The cost of a sufficient number of 
plates to standardize the entire scale is, however, prohibitive, so that 
the chemist is usually content with a few standard plates for that 
portion of the scale most used, as 80 to 100 for cane sugar. The pos- 



120 SUGAR ANALYSIS 

session of a few carefully standardized quartz plates is a necessity for 
accurate saccharimetric work, not so much for standardization (since 
the constant error of the scale need be determined but once), but for 
the determination of zero point, which is necessary with each set of 
observations. 

The standard quartz plates furnished by instrument makers are 
mounted in metal tubes upon which is stamped the reading that the 
plates should give upon the particular saccharimeter scale. It is im- 
portant that this reading be verified by some testing bureau, as slight 
errors in marking or faults in optical homogeneity of the plate are not 
uncommon. The surface of the plate when placed in the instrument 
must be perpendicular to the beams of polarized light which traverse 
it; for this reason the plates should never be loose in their mountings. 
On the other hand, the mounting must not press too tightly upon the 
plate, as optical errors might be produced in the quartz. Rotation of 
the plate about the axis of its tube should cause no change in the field 
at the end point. The plate when being used should be brought as 
close to the analyzer diaphragm as possible in order to give the greatest 
spread to the cone of light rays emanating from each part of the field. 
Care must be taken that the standard plate during polarization have 
exactly the same temperature as that of the quartz wedges of the 
instrument. If the plate have a temperature above that of the wedges, 
it will give a reading higher than its, true value. The temperature 
polarization coefficient of quartz is 0.000136, so that the polarization 
of a plate reading 100 V. at 20 C. would be for 30 C., 

100 f 1 + (0.000136) (30 - 20) j = 100.14 V. 

If plate and instrument are of different temperature, the plate should 
remain several hours in the trough of the saccharimeter before using, 
that sufficient time may be given for it to acquire the same temperature. 
While it is necessary that quartz plate and wedge system have the 
same temperature, it is not essential that this be the standard tem- 
perature for the instrument, since the variations due to temperature 
are practically the same for plate as for wedge. The slight differences 
due to effect of temperature upon shape of quartz wedge and upon 
expansion of nickeline scale are expressed by the formula (Schonrock), 
Vw = V t + V t 0.000005 (t - 20), in which F 20 and V t are the readings 
of the plate at 20 C. and t C. respectively. A standard plate polar- 
izing 100 V. at 20 C. would accordingly polarize 99.99 V. at 40 C. 
(plates and wedges in each case at same temperature), a variation of 
0.01 V. for 20 C. difference, which is negligible in practical work. 






THEORY AND DESCRIPTION OF S AC CHARI METERS 121 

Verification by Pure Sucrose. A second means of verifying the 
saccharimeter scale is with chemically pure sucrose. The preparation 
of sucrose of requisite purity is a matter of some difficulty; the method 
of the International Commission for Unifying Methods of Sugar Analy- 
sis * is as follows : 

"The purest commercial sugar is purified in the following manner: 
Prepare a hot saturated aqueous solution, precipitate the sugar with 
absolute ethyl alcohol, spin the sugar carefully in a small centrifugal 
machine, and wash in the latter with absolute alcohol. Redissolve 
the sugar obtained in water, again precipitate the saturated solution 
with alcohol, and wash as above. Dry the second crop of crystals 
between blotting paper, and preserve in glass vessels for use. Deter- 
mine the moisture still contained in the sugar and take this into account 
when weighing the sugar which is to be used." If a hand centrifugal 
is not available, the fine crystals of sugar may be filtered and washed 
free of sirup upon a Buchner funnel. In saturating the sugar solution 
before precipitation with alcohol, it is well not to heat above 80 C. 
The sugar solution thus prepared is filtered through a hot-water funnel 
into the alcohol, stirring vigorously. In this way the sugar is precipi- 
tated in the form of fine crystals which are easily dried in the air. 
Moisture is determined by drying at 105 C. 

In the selection of sugar for purification, the finest grades of small 
domino sugar (polarizing 99.90 to 99.95) have been found in the author's 
experience to give the best results. Rock-candy crystals, which are 
sometimes recommended, should never be used; they frequently con- 
tain perceptible quantities of acid, with the result that inversion takes 
place during purification. Complete absence of acidity in sugar and 
alcohol is necessary. 

To verify the 100-degree point of the saccharimeter scale, the 
normal weight of sugar is weighed into a 100-c.c. flask, dissolved in dis- 
tilled water, and the solution made up to volume, care being taken that 
the liquid is well mixed before making up the last few cubic centimeters. 
The solution, which must be perfectly clear, is then polarized in a 200- 
mm. tube. The conditions of weight, volume, and temperature required 
for the saccharimeter must be rigidly observed; the flasks and tubes em- 
ployed should have been previously calibrated. The average of 10 read- 
ings is taken and this result corrected for the moisture in the sugar, 
the amount of which must be determined in a separate portion with 
each set of observations. The sugar used for polarization should not 
be dried in a heated-air or water bath owing to the danger of slight 
* Proceedings of Paris Meeting, July 24, 1900. 



122 SUGAR ANALYSIS 

changes in composition. If the vernier of the scale is set at when 
the field is matched, the polarization of the sugar corrected for moisture 
should be exactly 100. In the same manner, other divisions of the 
saccharimeter scale can be verified by taking fractions of the normal 
weight (e.g., normal weight X 0.85 = 85-degree point of scale, etc.; see 
Table XXI). 

Verification by Control Tube. The most convenient means of 
verifying the scale divisions of a saccharimeter when using sucrose is 
by means of the Schmidt and Haensch control tube.* This method 
presents the advantage that perfectly pure sucrose does not need to be 
used; in addition to this, but very few solutions are necessary for 
verifying the entire scale. 

The control observation tube according to Landolt's latest form is 
shown in Fig. 82. It is telescopic in construction and can be adjusted 




Fig. 82. Control tube for verifying scales of saccharimeters. 

so as to give a column of solution for any length between 220 mm. and 
420 mm. The length of solution, which is regulated by the screw T, 
is read off upon the scale S by means of the vernier J to 0.1 mm. The 
tube is surmounted by a funnel E, which does not serve for filling, but 
simply receives the overflow of solution as the tube is shortened. 
For filling the tube, the funnel is removed and the opening closed by 
means of a plug (P) ; the tube is then drawn out its full length and 
filled from the end by unscrewing one of the caps. After rescrewing 
the cap, the tube is set in an upright position and the funnel replaced 
as before. After shortening the tube slightly, a few cubic centimeters of 
solution are poured in the funnel, which is then closed with a small cap 
to prevent evaporation. 

In using the control tube, it is best to begin at the 100-degree point 
(which is supposed to have been previously verified) of the saccharim- 

* Z. Instrument., 4, 169. 




THEORY AND DESCRIPTION OF SACCHARIMETERS 123 



eter scale and work downwards. A sugar solution is first made up of 
such concentration as to give a reading of 100 degrees at about 400 mm. 
length of tube. This will be sufficient to test the scale the few divisions 
above 100 and all divisions below 100 to 55. If the reading, for example, 
is 100 at 400 mm. upon the tube scale, it should read 105 at 420 mm., 
95 at 380 mm., etc. If a deviation be found at any division from the 
calculated value, other readings should be made at neighboring points 
of the scale to determine the position of maximum error. After test- 
ing the scale to the 55th division (220 mm.), another solution must be 
prepared which will give a reading of 55 at about 400 mm. and the 
scale tested down to 30. By proceeding in this way, always making 
the final point of one series the starting point of the next, the scale can 
be tested its entire length with 5 solutions. Landolt* has given the 
following table of concentration for solutions to be used with the 
control tube in testing the Ventzke scale: 



Number. 


Grams of su- 
crose in 100 c.c. 
of solution. 


Starting point 
for verifica- 
tion, V. 


Range of scale divisions for 
verification. 


1 


12.53 


100 


95, 90, 85, . . 60, 55 


2 


6.89 


55 


50, 45, 40, 35, 30 


3 


3.76 


30 


25, 20, 16 


4 


2.00 


16 


15, 10, 9 


5 


1.13 


9 


5 



In making the readings, the scale of the saccharimeter should first 
be set at the division which it is desired to verify and then the screw 
of the observation tube turned until the length of sugar solution gives 
a matched field. The reading upon the scale of the observation tube 
is then taken by means of a magnifying glass. The observed length of 
tube at any division in percentage of the observed length for the 100 V. 
point gives the actual value of the scale division. To distribute and 
equalize the errors due to changes in room temperature, warmth im- 
parted to the tube by the hand in making the adjustment, eye fatigue, 
and other causes, it is well to proceed forward and backward along the 
tube and not make all the observations for one point at one time. It 
is desirable to make several sets of readings upon different days and by 
different observers, and to take the average of the several series. The 
following results, obtained by the author upon one of the saccharim- 
eters belonging to the New York Sugar Trade Laboratory, will illus- 
trate the method: 

* " Das optische Drehungsvermogen " (1898), p. 341. 




124 



SUGAR ANALYSIS 



TABLE XXII 

Verification of S. & H. Saccharimeter, No. 7075 
Series No. I 



Scale division 
of saccha- 
rimeter. 


Reading of scale 
of control tube 
(average of 10 
readings). 


Value of scale 
division (in terms 
of 100-degree 
point). 


100 


mm. 

396.365 


100.000 


95 


376.495 


94.987 


90 


356.740 


90.003 


85 


336.930 


85.005 


80 


316.975 


79.972 


75 


297.120 


74.962 


70 


277.290 


69.957 


65 


257.465 


64.957 


60 


237.710 


59.972 



Average of Series 





Scale division of saccharimeter. 


Number 




of series. 






















100 


95 


90 


85 


80 


75 


70 


65 


60 


1 




94.987 


90.003 


85.005 


79.972 


74.962 


69.957 


64.957 


59.972 


2 




95.022 


90.028 


85.010 


80.033 


75.000 


69.990 


64.988 


59.960 


3 




95.008 


90.005 


85.005 


79.985 


74.998 


70.003 


65.012 


59.980 


4 




94.995 


90.023 


85.005 


79.990 


74.993 


69.980 


64.968 




5 




94.985 


90.015 


84.985 


79.985 


75.003 


69.995 


64.997 


59.990 


6 




95.037 


90.025 


85.038 


80.038 


75.028 


70.008 


64.990 


60.002 


Final 


100.000 


95.002 


90.017 


85.007 


80.001 


74.997 


69.989 


64.985 


59.981 


average 





















A similar average made upon another S. & H. saccharimeter (No. 6920) gave 





100.000 


95.004 


90.034 


85.041 


80.050 


75.028 


70.035 


65.031 


60.015 



The results show great exactness of graduation, the error in no in- 
stance exceeding 0.05 V. 

By marking the degrees of the saccharimeter scale upon a straight 
line and laying off the observed errors above or below this line for their 
respective scale divisions, the curve connecting the error points will 
give the correction for any degree of the scale. 

The following diagram (Fig. 83) for the observations of Table XXII 
will illustrate the method: 






THEORY AND DESCRIPTION OF SACCHARI METERS 125 



To verify the scales of a double-wedge saccharimeter, the scales of 
both wedges are first set at zero with their verniers for the matched 
field, any deviation of zero point being corrected by the regulating 
screw. The working-wedge scale is then verified and its curve of error 
determined by the control tube in the manner described. The control 
scale is then compared with the corrected readings of the working 
scale and its own error curve plotted. A still better direct method is 




100 



95 



90 



85 



80 



75 



70 



65 



















































































^" 


= 


1 - 


=^ 




.*- 








. .. 












= 


= 


=S-i 








*= 














































































































60 



Each division above line =0.01 V to be added to the scale reading 
n below =0.01 V " " subtracted from the scale reading 

Fig. 83. Example of diagram for correcting saccharimeter readings. 

set the working wedge at 100 and then verify the control scale from 
the division upwards by means of the control tube, using the same 
solutions as for verifying the working scale. If the tube, for example, 
with a length of 400 mm., gives a reading of 100 V. on the working- 
wedge scale with control-wedge scale at degrees, then with the work- 
ing-wedge scale at 100 V. the control-wedge scale should read 5 with a 
tube length of 380 mm., 10 with a length of 360 mm., etc. 

The millimeter scale of the control tube should be verified before 
the instrument is put to use. The control tube can be employed only 
upon the large-sized saccharimeters, which have a trough length of 420 



mm. 

v 



Verification by Scheibler's* Method of "Hundred Polariza- 
tion." Another means of verifying the scale readings of a saccha- 
rimeter is Scheibler's so-called method of " hundred polarization." In 
this process of verification the polarization of the raw sugar or other 
product is first determined and then the calculated amount of sub- 
stance weighed out which should give a polarization of exactly 100. 
Thus: if a normal weight of 26 grams of a sugar dissolved to 100 c.c. 



polarizes 82.5 then 



26 X 100 



= 31.515 grams, the weight of sugar dis- 



solved to 100 c.c. necessary to polarize exactly 100. If the polariza- 



* Z. Zuckerfabr. Deut. Reiches, 21, 320. 



126 SUGAR ANALYSIS 

tion obtained by the calculated weight of sugar is found to be 100, 
then the original scale reading of the saccharimeter is verified. 

EFFECT OF TEMPERATURE UPON THE READING OF SACCHARIMETER 

SCALES 

In the polarization of sugars and other materials upon quartz-wedge 
saccharimeters, the effect of temperature upon the scale reading is a 
most important factor. The saccharimeter is graduated to be used 
at a fixed temperature (17.5 C. or 20 C.), and in the most carefully 
regulated sugar laboratories this temperature is maintained through- 
out the year. But very few laboratories, however, are equipped with 
the necessary appliances for maintaining a temperature of 20 C. in 
summer, and the influence of temperature changes upon the saccha- 
rimetric readings and the methods for correcting the errors of the same 
should therefore be considered. 

Temperature Coefficient of Quartz. The changes in specific 
rotation of sugars with variation in temperature are considered on 
page 178. These changes apply to measurements made upon any 
kind of polariscope. But with the saccharimeter, as distinguished from 
the rotating polariscope, there must be considered an additional error 
due to the influence of temperature upon the quartz compensation of 
the instrument. This influence has been shown by Schonrock* to be 
threefold. There is (1) the change in shape of the wedge by expan- 
sion or contraction. The coefficient of expansion per 1 C. of quartz 
perpendicular to its axis (rj) is 0.000013, and parallel to its axis (Y) is 
0.000007. The polarization value of the 100 point of the scale through 
change in shape of the wedge decreases with increasing temperature 
by ??' -17, or by the coefficient -0.000006. There is (2) the change 
per millimeter thickness in the specific rotation of quartz itself, which 
for each degree increase in temperature increases by the coefficient 
0.000136. The combined temperature coefficient of the wedge system 
is therefore 0.000130. There is (3) the change due to the expansion 
and contraction of the material constituting the scale. The error due 
to this change, together with that resulting from atmospheric humidity, 
was so great with the old ivory scales that the latter have been replaced 
in most saccharimeters with the alloy nickeline which has an expansion 
coefficient per 1 C. of 0.000018. The total correction, therefore, for 
a quartz-wedge saccharimeter with nickeline scale is 0.000148. The 
polarization value w for any temperature t is then expressed by the 

* Z. Ver. Deut. Zuckerind., 54, 521. 






THEORY AND DESCRIPTION OF SACCHARIMETERS 127 

equation w* = w\l+ 0.000148 (t 20) J. With saccharimeters whose 
scale is etched directly upon the wedge itself, as is the case with Schmidt 
and Haensch instruments of recent construction, the coefficient remains 
0.000130. 

The above increase in polarization of quartz with increase in tempera- 
ture necessarily produces a lowering in the readings of the saccharimeter 
scale, since a smaller thickness of quartz is required for compensation. 
With sugars which undergo a decrease in specific rotation with increase 
in temperature, the combined influences are in one direction and the 
error thus introduced may be considerable. With sucrose, for example, 
the temperature coefficient of polarization becomes at 10 C. 0.000390 
(0.000148 + 0.000242), at 20 C. 0.000332 (0.000148 + 0.000184), and 
at 30 C. 0.000269 (0.000148 + 0.000121). 

Temperature Coefficient of Sucrose. The variation in the 
Ventzke reading of the normal weight of pure sucrose for 1 C. change 
in temperature has been found by different authorities to be as follows: 

Andrews* 0.0300 

The United States Coast and Geodetic Survey 0.0293 

Wiley t 0.0314 

Prinsen Geerligs{. 0.0300 

Watts & Tempany 0.0310 

T 



Average = 0.0303 



The average temperature coefficient of the above is therefore 
0.000303, which agrees with the figure of Schonrock for 25 C. (0.000148 
-f- 0.000152) = 0.000300. For temperatures between 20 and 30 C. the 
general equation F 20 =F'Jl-f 0.0003 (-20){ may be used for chang- 
ing the Ventzke reading (V 1 ) of pure sucrose at any temperature t to 
the reading- (F 20 ) at 20 C. 

Temperature Coefficients of Other Sugars. The temperature 
coefficients of other common sugars for readings upon the Ventzke scale 
are given in the following table. The temperature coefficient for fructose 
and invert sugar are for readings made upon the negative scale of the 
saccharimeter; while the coefficients of these sugars decrease the same 
as those of the dextrorotatory sugars, the direction of the decrease in 
both cases is towards the point and therefore opposite to each other 
(as indicated by the arrow points). 

* Technology Quarterly, Mass. Inst. Technology, May (1889), 367. 

t J. Am. Chem. Soc., 21, 568. 

t Archief Java Suikerind, July (1903). 

West Indian Bull., Vol. Ill, p. 140. 




128 



SUGAR ANALYSIS 



TABLE XXIII 
Giving Temperature Coefficients of Different Sugars for Ventzke Scale 



Sugar. 


A 

[< 


B 

Change in 
Mffer 

1 C. increase. 


C 

Temperature 
coefficient 
B 
A' 


Temperature coefficient of 
reading upon Ventzke scale for 
1 C. increase. 
C + coefficient for quartz 
(-0.000148). 


Fructose 


-92.50 
-20.00 
+52.53 
+ 138.04 
+53.23 


+0.625 
+0.312 
-0.070 
-0.095 
No change 


-0.006757 
-0.015600 
-0.001332 
-0.000688 
No change 


-0.006905 
-0.015748 
-0.001480 
-0.000836 
-0.000148 


I I 

O O O,O 

TTT 


Invert sugar . . . 
Lactose 


Maltose ... 
Glucose 



In case a mixture of sugar is polarized upon a saccharimeter, the 
combined influence of the temperature coefficients of each sugar must 
be considered. To arrive at a better understanding of the use of such 
coefficients the following special problem is considered: 

It is desired to find the amount of fructose and of invert sugar which, 
mixed with 26 gms. of pure sucrose, will give a constant saccharimeter reading 
at all temperatures. 

It has been shown that 26 gms. of pure sucrose, reading 100 V. at 20 C., 
undergo a decrease of 0.03 V. with 1 C. increase in temperature. Since a 
fructose solution reading 1 V. undergoes a decrease in polarization of 



0.0069 V. (Table XXIII), then 



0.03 



= -4.35 V., the scale reading of the 



0.0069 

required amount of fructose. Since 0.1869 gm. of fructose in 100 metric c.c. 
reads -1 V. at 20 C. in a 200-mm. tube, then 4.35 X 0.1869 = 0.813 gm., the 
required amount of fructose. 26 gms. sucrose and 0.813 gm. fructose (3.13 
per cent of the weight of sucrose) will give, therefore, a constant saccharimeter 
reading at all temperatures. 



In the same way for invert sugar, 



0.03 



= -1.90V., the scale reading 



0.01575 

of the required amount of invert sugar. Since 0.8645 gm. invert sugar in 
100 metric c.c. reads -1 V. at 20 C. in a 200-mm. tube, then 1.90 X 0.8645 
= 1.642 gms., the required amount of invert sugar. 26 gms. sucrose and 
1.642 gms. invert sugar (6.32 per cent of the weight of sucrose) will give, there- 
fore, a constant saccharimeter reading at all temperatures. 

The effect of 1 C. increase in temperature upon the reading of 
1 per cent each of sucrose, fructose, and invert sugar for a normal weight 
of 26 gms. in 100 metric c.c. is given in the following table: 



THEORY AND DESCRIPTION OF SACCHARIMETERS 129 

TABLE XXIV 

Showing Influence of Temperature upon Ventzke Reading of 1 per cent Sucrose, Fruc- 
tose, and Invert Sugar for a Normal Weight of 26 gms. Solutions made up to 
Volume at Temperature of Polarization 

1 per cent sucrose = -^ = -0.0003 V. for 1 C. increase. 

1 per cent fructose = ^ = +0.0096 V. for 1 C. increase. 
1 per cent invert sugar = ^^ = +0.0048 V. for 1 C. increase. 
( denotes change toward the left. + denotes change toward the right.) 

Since the influence of temperature upon the rotation of glucose is 
so small as to be negligible, the change in rotation for 1 per cent invert 
sugar should be the same as that for 0.5 per cent fructose, or +0.0048 V. 
This is the result actually obtained, so that the calculation is verified. 

SHALL SACCHARIMETERS BE ADJUSTED TO VARIABLE TEMPERATURES? 

The International Commission* has provided that "for laboratories 
in which temperatures are usually higher than 20 C., it is permissible 
to graduate saccharimeters at any suitable temperature, providing that 
the volume be completed and the polarization made at the same tem- 
perature." The Commission has neglected, however, to say how this 
graduation shall be made. It is evident that in order to have a normal 
weight of sucrose, under the conditions prescribed for a saccharimeter 
at 20 C., polarize 100 at 25 C. or 30 C., the compensating thickness 
of quartz in the wedge system must be made thinner for each part of 
the scale in order to counterbalance the decrease in specific rotation of 
sucrose. 

Owing, however, to the confusion and mistakes which would arise 
in the use of standard plates with saccharimeters of different compen- 
sating power, a better plan would be to make no change in the instru- 
ment itself, but to alter the conditions of polarization, such, for example, 
as increasing the normal weight of sugar, or increasing the length of the 
observation tube, or decreasing the volume of the flask, any one of which 
means will bring the polarization of pure sucrose to 100 for any desired 
temperature above the standard. Since odd lengths of tube or volume 
of flask are undesirable as well as confusing, a change in the normal 

* Proceedings of Paris Meeting, July 24, 1900. 



130 SUGAR ANALYSIS 

weight of sucrose is the simplest of all means of correction. The method 
of calculation can be understood from the following example. * 

What would be the normal weight at 25 C. for a quartz-wedge saccharim- 
eter standardized at 20 C. for 26 gms. sucrose dissolved to 100 true c.c. and 
polarized in a 200-mm. tube? 

The temperature coefficient of the specific rotation of sucrose at 22.5 C. 
is 0.000168 (Schonrock). The temperature coefficient of the nickeline scale 
and quartz wedge is 0.000148; the expansion coefficient for the glass observa- 
tion tube is 0.000008. The new normal weight would then be 

26,000 J 1 +(0.000148 + 0.000168 - 0.000008) (25 - 20) j = 26.040 gms. 
dissolved to 100 true c.c. in a flask standardized at 25 C. 

When saccharimeters are employed constantly in the investigation 
of pure sucrose solutions, it might be advisable to make a change such as 
the above in the normal weight. But for varied work with different 
classes and mixtures of sugars whose specific rotations are affected in 
opposite ways by changes in temperature, it is inaccurate to make al- 
terations based upon the change in properties of one single sugar. 
The results obtained upon saccharimeters differently standardized are 
then no longer comparable. The sucrose normal weight is frequently 
employed upon mixtures of sucrose with other sugars; in such cases 
changes in normal weight to correct for rotatory changes in the sucrose 
alone are wholly unwarranted. In view of the fact that the work of 
saccharimeters is usually of a varied character, it seems best to leave 
the scale and standard conditions of the instrument unchanged. The 
chemist should work wherever possible under the conditions of tem- 
perature prescribed for his saccharimeter, and when this cannot be done 
he should correct his readings as well as possible by a factor established 
for the particular product which is being examined. 

It must always be borne in mind that while the saccharimeter scale 
is established for the rotation of sucrose, its divisions indicate percent- 
ages only when pure sucrose is being polarized; in all other cases the 
scale division becomes a mere conventional number (degrees Ventzke, 
degrees polarization, degrees sugar scale, etc.) which the analyst must 
interpret according to his particular needs. 

* This example is from a calculation supplied by the Physikalisch-Technischc 
Reichsanstalt, in reply to a suggestion by the author to use the old Mohr c.c. normal 
weight 26.048 gms. (17.5 C.) for true c.c. at 25 C. The old normal weight would 
give a reading of 100.031 V. when dissolved in 100 true c.c. in a flask standardized 
at 25 C. If the true c.c. flask standardized at 20 C. be used at 25 C., this error 
would be reduced to 100.019 V., which is within the limits of error for observation. 



THEORY AND DESCRIPTION OF SAC CHARI METERS 131 

DESCRIPTION OF SACCHARIMETEBS 
Tint Saccharimeters 

The saccharimeter of Soleil as modified by Ventzke and Scheibler 
in Germany and by Duboscq in France consists of an adaptation of the 
quartz- wedge compensation to the polariscope of Robiquet (p. 86). 

The Soleil- Ventzke-Scheibler Saccharimeter. The construction 
and arrangement of the optical parts in the Soleil saccharimeter as 
modified by Ventzke and Scheibler are shown in Fig. 84. A is a Nicol 
prism and B a plate of left or right rotating quartz cut perpendicular 
to its optical axis; these constitute the tint producer and are mounted 




D C B A 

_, LJ 

F 
Fig. 84. Soleil- Ventzke-Scheibler tint saccharimeter. 

in a movable sleeve which can be rotated by a rod and pinion from J. 
C is a condensing lens, D the polarizer, and E a Soleil double quartz 
plate (p. 86). The quartz compensation is at F, the analyzer at G, 
and telescope at H. In using the instrument the telescope is focused 
upon the bi-quartz plate, so that the dividing line is sharply defined. 
The zero point of the scale is then determined by turning K until both 
sides of the field have the same tint (in the manner described on p. 88). 
By rotating the regulator or tint producer from /, the tint which is 
most sensitive to the eye of the observer is obtained. This tint, which 
is different for different eyes, is usually of a very delicate violet or 
pearl color; it will of course vary according to the angle with which 
the Nicol A is set with reference to the Nicol D of the polarizer. In 
order to remove the disturbances in transition tint due to colored 
solutions (which cannot be remedied in the Robiquet polariscope), 
the adjustment of the regulator is changed until the tint is again of 
greatest sensitiveness. With very dark solutions the transition tint is 
almost a shadow owing to the absorption of color. 



132 



SUGAR ANALYSIS 



The Soleil-Duboscq Saccharimeter. The Soleil saccharimeter 
as modified by Duboscq, the type of tint instrument used in France, 
differs from the form previously described in that the Nicol producing 
the sensitive tint is situated in the eyepiece of the telescope, as shown 
by N in Fig. 85. The latter is rotated by a milled ring B until the 
sensitive tint is produced with the quartz plate C, which in the Duboscq 
instrument is situated between the analyzer and the objective of the 
telescope. The telescope is focused upon the Soleil double plate at R 




Fig. 85. Soleil-Duboscq tint saccharimeter. 

by moving the eyepiece D in or out; longitudinal guides prevent any 
lateral rotation which might disturb the tint. In the Duboscq instru- 
ment the two wedges of the compensator are of equal size, and, being 
driven past each other by the pinion in opposite directions, give a 
stratum of quartz of variable thickness. A scale and vernier, which 
follow the wedges in their movement, indicate the reading. 

According to Landolt,* the average error of adjustment with the 
Soleil saccharimeter is 0.2 degree of the scale. The instrument has 
the same objection as the Robiquet polarimeter, in being unsuited to 
the color-blind. The adjustment of end point to color is also much 
more fatiguing to the eye than adjustment to uniformity of shade. 
Owing to these objections the color saccharimeter, although 20 years 
ago the standard instrument, is but little used at the present time. Its 
use is in fact condemned by the Imperial Testing Bureau of Germany. 

Half-shadow Saccharimeters 

The various types of half-shadow saccharimeter used at the present 
time consist simply of an adjustment of the quartz-wedge compensation 
to some one of the half-shade polarizers previously described. The 
principal forms are the double-field saccharimeter with Jellet-Cornu 

* " Das optische Drehungsvermogen" (1898), p. 348. 



THEORY AND DESCRIPTION OF SACCHARIMETERS 133 

polarizer; the double-, triple-, and concentric-field saccharimeters with 
Laurent plate; and the double- and triple-field instruments with 
Lippich polarizer. 

Saccharimeter with Jellet-Cornu Prism. A single-wedge half- 
shadow saccharimeter with Jellet-Cornu prism as polarizer is shown in 
Fig. 86. 




Fig. 86. Single-wedge saccharimeter with Jellet-Cornu prism. 

N. Sliding sleeve containing condensing lens. 

0. Modified Jellet-Cornu prism (Schmidt and Haensch prism). 

E, F. Parts of quartz- wedge compensation. 

H. Analyzer. 

J. Telescope, which is focused upon the dividing line of the split prism at 0. 

K. Microscope for reading scale. 

The above saccharimeter, which 15 years ago was the standard form 
of instrument employing the Ventzke scale, is at present almost en- 
tirely replaced with saccharimeters using the Lippich polarizer. 

Laurent's Saccharimeter. As a type of the saccharimeters con- 
structed by French instrument makers, the Laurent instrument shown 
in Fig. 87 is described. The arrangement of polarizer, half-wave plate, 
and device for regulating the half-shadow angle is identical with that of 
the Laurent polarimeter (Fig. 72). The divided circle and rotating 
analyzer of the latter, however, are replaced in the saccharimeter by 
the quartz-wedge compensation. 

The saccharimeter is adjusted to its zero point by first turning G 
until the two halves of the field agree in shade. If it should be found 
that one side of the field has more of a reddish tinge than the other at 
the end point, the screw F, which controls the analyzer, is turned so as 
to darken slightly the side of the field most colored. The screw G is 



134 



SUGAR ANALYSIS 



then turned again to equality of shade; if there is still a difference in 
color, F is moved slightly as before, and G again turned to equality of 
shade. By proceeding cautiously in this way the observer will at length 
reach the point where both sides of the field correspond in shade and 
color. When this point is reached the screw T is turned until the of 



M 




Fig. 87. Laurent's single-wedge sacchari meter. 

A. Lamp for producing white light (oil, gas, electricity, etc.), placed 200 mm. from B. 

B, E, R, K, J, X, U, D, L, the same as under Laurent polarimeter (Fig. 72). 

R. Saccharimeter scale, which with vernier V is illuminated by light reflected from 

A by the mirror M . 

N. Magnifying glass for reading scale and vernier. 
G. Screw for moving quartz wedges of the Soleil compensator. 

the scale coincides with the of the vernier. This adjustment should 
be verified by taking a number of check readings. 

The 100-degree point of the Laurent saccharimeter scale corre- 
sponds to a rotation of 21 40', the value given by French physicists to 
the rotation of the 1-mm. plate of quartz. The normal weight for this 
angular displacement, as previously noted, is 16.29 gms. sucrose for 
100 true c.c. polarized in the 200-mm. tube. The Laurent saccharim- 
eter is also manufactured with a scale adapted to the so-called Inter- 
national normal weight of 20 gms. The instrument is provided with 
double or triple field, as desired. The scale divisions extend from to 
110 to the right. 



THEORY AND DESCRIPTION OF SACCHARI METERS 135 

"Plaque Type." The 100-degree point of the Laurent saccha- 
rimeter is verified by a standard plate of quartz 1 mm. thick. This 
standard plate " plaque type" also serves for the polarization of levo- 
rotatory solutions. With the plate in the trough of the instrument, 
the zero point of the scale is transferred to 100; levorotatory solutions 
are then simply read backwards upon the scale, the reading being the 
difference between readings of plate and solution. A solution, for 
example, reading 67.4 with the 100-degree plate in position has a 
polarization of 32.6. This method of polarizing levorotatory solu- 
tions is of course applicable to all single-wedge saccharimeters. 

A 100-degree Laurent "plaque type " was remounted by the author 
and sent to the United States Bureau of Standards for a certification 
as to its angular rotation and its value in sugar degrees upon a sac- 
charimeter employing the Ventzke scale. The rotation of the plate 
for sodium light of 589.23 ^ wave length was given as +21.713 
+ 0.003 (T - 20) =b 0.004, and the rotation in sugar degrees as 
+62.65. The same plate read by the author upon a late-model 
Schmidt and Haensch saccharimeter gave a reading of +62.64, and 
upon a late-model Fric saccharimeter (Bates modification) a reading 
of +62.65. These readings of the " plaque type" not only prove 
the perfect identity of the Ventzke sugar scales employed by two 
different manufacturers, but also permit the establishment of the 
exact ratio between the French and German normal weights; for all 
other conditions as to the temperature and volume are the same in 
both these countries. The ratio 100 : 26 gms. :: 62.65 : X shows that 
the ratio of the German normal weight to the French normal weight 
is as 26 gms. to 16.289 gms., or, in even hundredths, 16.29 gms., which 
is identical with the official normal weight prescribed in France. 

Duboscq-Pellin Saccharimeter. The Duboscq-Pellin saccharim- 
eter for white light, as regards position of polarizer, half-wave plate, 
quartz-wedge compensation, etc., is the same as that of the Laurent. 
The concentric field of the Pellin saccharimeter requires a somewhat 
different cutting of the half-wave plate, but in other respects the two 
saccharimeters are very much alike. 

The saccharimeter with Lippich polarizer is the form most generally 
preferred at present. The half-shadow angle between the prisms of 
the polarizer is usually between 5 and 8 degrees; it can be measured 
approximately by noting the interval between the points of maximum 
light extinction each side of the zero point. The degrees Ventzke 
between the two points of maximum darkness multiplied by 0.34657 
gives the angle of the half shadow. 



136 



SUGAR ANALYSIS 



Schmidt and Haensch Saccharimeters. A single- wedge Schmidt 
and Haensch saccharimeter upon tripod support with electric attach- 
ment for illumination is shown in Fig. 88. 




Fig. 88. Single-wedge Schmidt and Haensch saccharimeter with electric attach- 
ment for illumination. 

V. Detachable end containing lamp and for inserting cell of bichromate solution. 
P. Position of Lippich polarizer for double or triple field. 
G. Casing of sheet brass for protecting wedges from dust. 

The method of scale illumination in Schmidt and Haensch saccharim- 
eters is shown in Fig. 89 which gives the arrangements of parts for a 
double- wedge instrument. The light from the lamp is focused upon the 
small window a in the wedge housing, and is reflected from the mirror 
b through the ground-glass plate c upon the scale from which it is re- 
flected through the prism p into the microscope whose objective is at d 
and eyepiece at / g. The working wedge is operated by the screw A 
and the control wedge by the screw K. The appearance of the scale of 
this instrument as viewed through the microscope is shown in Fig. 80. 

The latest and most improved type of Schmidt and Haensch sac- 
charimeter is the double-wedge apparatus shown in Fig. 90. The 
instrument is mounted upon a bock or trestle support, and for saccharim- 
eters which are in constant use this method of mounting is most satis- 
factory as it insures perfect rigidity and accurate alignment. The 
wedges are moved by milled screw heads at A and K which are so 



THEORY AND DESCRIPTION OF SACCHARIMETERS 137 

H 

m :| 




Fig. 89. Device for illuminating scale of Schmidt and Haensch saccharimeter. 




Fig. 90. Double-wedge Schmidt and Haensch saccharimeter upon bock support. 



138 



SUGAR ANALYSIS 



placed that the hand can rest upon the table during adjustment. The 
screw K moving the control wedge can be fastened with a clamp, and 
is placed at a slightly higher elevation to prevent liability of confusion. 
Peters's Saccharimeter. Very similar in construction to the 
above apparatus is the saccharimeter of Peters shown in Fig. 91. The 
long tube R prevents placing the light too close to the polarizer. 
The bichromate cell is placed within S; the cover C of the trough is not 
hinged but simply slides over or under the tube. The scale in the 




Fig. 91. Double-wedge Peters saccharimeter. 

sheet-metal housing is illuminated by light reflected from the mirror 
L; a black paper disc P protects the eye against the glare of the obser- 
vation lamp. 

Fric's Saccharimeter. The half-shadow saccharimeters of J. and 
J. Fric are very similar in construction to the instruments previously 
described except in the method of scale illumination. In the latest 
types of Fric saccharimeter a part of the light, as it passes from the 
source of illumination through the diaphragm at the end of the instru- 
ment, is reflected through a system of mirrors and lenses upon the 
scales. This illuminating attachment is shown in the Bates sac- 
charimeter (L in Fig. 94), but the distinctive feature of the Fric illumi- 



THEORY AND DESCRIPTION OF SACCHARIMETERS 139 



nating device is at the scale end of the instrument as shown in Fig. 92. 
The light from L is reflected from the mirror A (which in the instru- 
ments with enclosed wedges is stationary) through the milk-glass plate 
B upon the scale C, the latter in 
the latest Fric saccharimeters being 
made of glass. The light from L " 
C is reflected from the mirror D 
through the focusing lens E to the 
eye of the observer. The divisions 
of the scale illuminated in this 
manner appear with great distinct- 
ness. The Fric double-wedge in- 
struments are provided with sepa- 
rate focusing lenses for reading the 
working and control scales. The 
lens mountings and the milk-glass 
plates for the two wedge systems 
are usually of different colors in 
order to prevent confusion. 




SACCHARIMETERS WITH VARIABLE SENSIBILITY 

Of the instruments previously described, the French saccharimeters, 
using a Laurent half-wave plate and employing monochromatic or 
bichromate-filtered white light, are the only forms of apparatus which 
permit a variation of the half-shadow angle to suit the requirements of 
greatest sensibility. 

In all the Schmidt and Haensch saccharimeters the half-shadow angle 
is fixed. An attachment for shifting the large prism of the Lippich 
polarizer and regulating the half-shadow angle has been supplied by 
some manufacturers. While this regulating device presents certain 
advantages, it has been condemned by Landolt* on the ground that 
every change in the half shadow introduces a change in the zero point 
which has to be corrected by rotating the analyzer until the field is 
again evenly illuminated at the zero point an impossible remedy in 
a saccharimeter with fixed analyzer. 

Bates's Saccharimeter. To obviate the objection last named, 
Bates f has devised an attachment which rotates the analyzer automati- 
cally and makes it possible to correct the zero-point error for any change 



" Das optische Drehungsvermogen," 351. 
t U. S. Bur. Stand. Bull., Vol. 4, p. 461; Z. Ver. Deut. Zuckerind., 68, 105. 



140 



SUGAR ANALYSIS 



in the half-shadow angle without resetting the scale. The principle of 
the Bates saccharimeter can be understood from Fig. 93. 

Let OP be the direction of the plane of the large Nicol and ON that 
of the small Nicol .in a Lippich polarizer, let AZ be the plane of the 
analyzer at right angles to OB the bisection of the half-shadow angle 
PON or a. We will suppose for a moment that the intensities of light 
in OP and ON are equal and that the plane of the large Nicol be moved 
from OP to OP' forming with the plane of the small Nicol the new 




Fig. 93. Illustrating principle of Bates's saccharimeter. 

angle P'ON or a' . To obtain uniformity of field at the zero point for 
the new angle a' the bisection OB must be moved to OB'. It will be 



seen from the diagram that the angle BOB' = - = ~ 

Zi 2i 2i i ' 

To correct, therefore, for the displacement of zero point, assuming the 
intensities of light to be always the same for both Nicols, the plane of 
the analyzer must be moved through one half the angular displace- 
ment of the large Nicol of the polarizer. 

In the Lippich system, however, the intensities of light are not equal 
for the large and small prisms of the polarizer. A part of the light is 



THEORY AND DESCRIPTION OF SACCHARIMETERS 141 

extinguished in the small Nicol and there is also a loss from reflection 
and absorption. We will consider first the light lost by absorption. 

Let OK = amplitude of light from large Nicol. Draw KL _L ON; 
then OL = amplitude of light from small Nicol; the plane of the 
analyzer AZ must then be moved to A'Z' that the amplitudes OC and 
OF be equal in each half of the field. The angles AOA r and BOD, 
through which the plane of the analyzer and its perpendicular have 
moved, is 5 or the change from the true zero point when the intensities 
of light in OP and ON are equal, in which case a = 0. 

We will suppose in order to increase the intensity of light for the 
half shadow that the plane OP of the large Nicol be moved to OP' in- 
creasing a to a. The amplitude OK' remains the same as OK. Draw 
K'U J_ ON; then the amplitude in ON = OU. The plane of the 
analyzer must now be moved to A"Z" in order that the -Is K'C' and 
L'F f cut off the equal amplitudes OC' and OF' in the two halves of the 
field. OD' which is _L A"Z" will then form with OB', the bisection of 
a', the new angle d'. The angle = DOD' through which the analyzer 
has moved from its previous position is expressed by the equation 



In the polariscope of Bates (Fig. 94) the analyzing Nicol and the 
large Nicol of the polarizing system are mounted in bearings and are 
joined by gears with a connecting rod. The milled head, which oper- 
ates the driving mechanism, is shown at H. When the milled head is 
turned the two Nicols are rotated and the design of the gears is such 
that the analyzing Nicol always receives one half the angular dis- 
placement of the large Nicol of the polarizing system. Above the 
milled head is a circular scale which shows the polarizing angle for any 
position of the Nicols. In moving the plane of the large polarizing 
Nicol through the angle POP' (Fig. 93) the rotating device of Bates's 
polariscope moves the plane of the analyzer through the angle BOB'. 
In this way the zero-point error of the instrument will always be equal 
to the value of d for any angle of the half shadow, assuming that the 
zero had been previously adjusted for a = 0. If the zero point of the 
instrument be set for any value of the half shadow a, and a be then 

changed to a', the zero will have an error of 5' d ( the analyzer hav- 
ing rotated = , this value disappears from the equation 



142 



SUGAR ANALYSIS 



The calculated values of 5 in Ventzke degrees for different values of 
the half-shadow angle a according to the two equations, 

, a 1 COS a V0.92 a 

tan 5 = tan 3 s and tan 5 = - , tan 

2 1 + cos a V0.92 2 

(see p. 96), are given in the following table. 

TABLE XXV 

Giving Calculated Values of Error in Zero Point for Bates 's Saccharimeter 
VALUES OF a IN VENTZKE DEGREES. 





I 


II 


Values of 
a 
circular degrees. 


By formula 
tan = tans |. 


By formula 
_ l-cosaV / OL92^_ a 


ufln o tfin f. 
l+cosaV0.92 2 


1 


0.003 


0.033 


2 


0.004 


0.064 


3 


0.005 


0.09G 


4 


0.008 


0.129 


5 


0.014 


. 164 


6 


0.024 


0.205 


7 


0.038 


0.249 


8 


0.057 


0.299 


9 


0.080 


0.352 


10 


0.110 


0.412 


11 


0.150 


0.482 


12 


0.192 


0.554 



The values of 6 in the second column are greater than those in the 
first column by 0.03 a. The true values of 5 according to Bates lie 
between those calculated by the two equations and will vary according 
to the construction of the instrument. This true value of 8 will be the 
value by the first formula c a in which c is a constant for each in- 
dividual Lippich system. If a Bates saccharimeter be set, therefore, 
for a = 0, the calculated change in zero point for variations in a can 
be easily applied to the scale reading. If the instrument be set for any 
particular value of a, as 8 degrees, the half-shadow angle may be in- 
creased or diminished several degrees from this point without intro- 
ducing a change in zero greater than 0.1 V. 

The Bates saccharimeter, constructed by Josef and Jan Fric of 
Prague, is at present the standard instrument of the United States 
Customs Service. While the apparatus presents several advantages 
over the ordinary saccharimeter, the mechanical difficulties of con- 
struction make it expensive. In its present commercial form the 
instrument is not provided with a bichromate light filter. While this 



THEORY AND DESCRIPTION OF SACCHARIMETERS 143 

omission may occasion no serious error in the polarization of colored 
solutions (as of low-grade sugar-house products), a bichromate light 
filter is required in the examination of high-grade cane sugars, starch- 
conversion products, and many other substances. An absorption cell 
for this purpose should be placed just in front of the aperture between 
the saccharimeter and the source of light. A very commendable feature 
of the Bates instrument is the thermometer (T Fig. 94) which indicates 
the temperature of the quartz wedges. 




Fig. 94. Bates saccharimeter with variable sensibility. 

TF, milled head for operating working wedge. 

C, milled head for operating control wedge. 
w, microscope for reading working wedge scale. 

c, microscope for reading control wedge scale. 

S,_ scale indicating " degrees of brightness " or half-shadow angle. 

SACCHARIMETERS WITH MAGNIFIED SCALE. 

For special kinds of work involving the investigation of products 
with a narrow range in composition, saccharimeters have been con- 
structed with a limited magnified scale. The saccharimeter devised 
by Stammer,* shown in Fig. 95, for polarization of sugar beets is an 
* Z. Ver. Deut. Zuckerind., 37, 474. 



144 SUGAR ANALYSIS 

example of such an instrument. In this apparatus a magnified scale, 
reading from to 35, is attached to the side of the instrument at the 
observer's left and permits the reading of polarizations with the unaided 
eye. The pointer of the scale P is moved by the tension roller R, 
which is connected by a small steel chain with the movable quartz 
wedge. 

To adjust the saccharimeter the field is brought to a uniform shade 
by turning K when the of the wedge scale and vernier at Q should 
coincide. If the latter is not the case, coincidence is affected by turn- 




Fig. 95. Stammer's saccharimeter with magnified scale for polarizing sugar beets. 

ing the regulating key V. In this position the pointer P should mark 
exactly the zero division of the large scale S. Should there be any 
deviation the error is corrected by turning the adjusting lever L until 
the pointer is exactly at 0. Turning the screw K to any division upon 
the wedge scale Q should then give the same reading upon the scale 
S. If this is not the case the error is corrected by turning a small 
control screw upon R which increases or diminishes the diameter of the 
roller. The adjustment is one which requires considerable care and 
should be checked repeatedly. 

Saccharimeters of the above type are especially adapted for the 
polarization of mother beets for seed production; they are constructed 
for tubes of 200 mm., 400 mm., and 600 mm. length. 

Similar to the above instruments for sugar-beet analysis, saccharim- 



THEORY AND DESCRIPTION OF SACCHARIMETERS 145 

eters have been constructed with a magnified scale reading between 
80 and 100 for polarization of sugars. These are manufactured usually 
only for use with tubes 400 mm. long, and employ a normal weight of 
26 gms. to 100 c.c. solution. Doubling the length of observation tube 
necessitates of course doubling the interval between the scale divisions 
and thus facilitates the reading. 

Instruments with a magnified limited scale will be found to relieve 
eye fatigue, where large numbers of analyses of a single product have 
to be performed. With one person to prepare the tubes of sugar 
solutions, a second to manipulate the saccharimeter, and a third to note 
the readings, a large number of polarizations can be made in a very 
short period of time. 

CONVERSION FACTORS FOR POLARISCOPE AND SACCHARIMETER SCALES 

In the following table factors are given for converting 1 degree of 
the various polariscope scales into its equivalent in circular degrees, 
or in degrees of the different saccharimetric scales. The conversions 
are based so far as possible upon recent information supplied by the 
manufacturers of the several instruments. 

Scale. Equivalent. 

1 Ventzke sugar scale = 0.34657 angular rotation D. 
1 angular rotation D = 2.88542 Ventzke sugar scale. 
1 French sugar scale = 0.21666 angular rotation D. 
1 angular rotation D = 4.61553 French sugar scale. 
1 French sugar scale = 0.62516 Ventzke sugar scale. 
1 Ventzke sugar scale = 1.59960 French sugar scale. 
1 Wild sugar scale = 0. 13284 angular rotation D, 
1 angular rotation D = 7.52814 Wild sugar scale. 
1 Wild sugar scale = 0.38329 Ventzke sugar scale. 
1 Ventzke sugar scale = 2.60903 Wild sugar scale. 
1 Wild sugar scale = 0.61313 French sugar scale. 
1 French sugar scale = 1 . 63098 Wild sugar scale. 

(Normal weight = 26. 00 gms. Ventzke scale; 16. 29 gms. French scale; 10. 00 gms. 
Wild scale.) 

The Ventzke sugar scale is employed upon the Schmidt and Haensch, 
Peters, and Fric saccharimeters. The French sugar scale is employed 
upon the Laurent-Jobin and Duboscq-Pellin saccharimeters. 

The slight differences in ratio between normal weights and scale 
equivalents have already been discussed. 



CHAPTER VII 

POLARISCOPE ACCESSORIES 

ILLUMINATION OF POLARISCOPES 

FOR the illumination of polariscopes and saccharimeters numerous 
lamps have been devised and the chemist must be guided in his selec- 
tion by type of instrument, nature of substance to be polarized, and 
the kind of light supply available. Before describing the various 
types of lamps, a word should be said regarding the general subject 
of illumination. 

A much neglected point in the illumination of polariscopes and 
saQcharimeters is the placing of the light at the proper distance from 
the condensing lens. The light should never be placed so near as to 
over-heat the metal at the end of the instrument; neglect of this pre- 
caution may cause a softening of the balsam and wax mountings of 
the polarizer and lead to serious derangement of the optical parts. 

The proper rule in setting up the polariscope is to place the light in 
such a position that its image is clearly defined upon the analyzer 
diaphragm; this is best accomplished by fastening a needle or other 
sharp-pointed object just before the light and moving the instrument 
or light until a clear inverted image of the point is obtained upon a 
piece of white paper placed before the analyzer diaphragm. When 
the light is thus focused the polariscope is least susceptible to changes 
in zero point. The proper position of polariscope with reference to 
light can be seen from Fig. 96, which shows the arrangement of the 
optical parts in a double-wedge saccharimeter. When correctly placed 
an inverted magnified image of the light 7 is obtained at A. The 
reciprocal of the focal distance of the condensing lens will then equal 
the sum of the reciprocals of the distances of lens from light and of 
lens from image. 

Example. In the case of a Schmidt and Haensch saccharimeter the focal 
distance of the condensing lens was found to be 5 inches; the distance from 
lens to analyzer diaphragm was 20 inches; the distance for placing the light 

would then be--h^: = -or6f inches from the condensing lens. 

X ZO O 

146 



POLARISCOPE ACCESSORIES 147 

The telescope T (Fig. 96) is focused by the observer upon the 
dividing line of the field at C and the analyzer or compensator turned 
to the point of even illumination. The dividing line at C will then 
disappear and the entire field appear of equal intensity. This will be 
the case even with slight variations in intensity in different parts of 
the illumination, since at the point C, upon which the eye of the ob- 
server is focused, the light from any part p of the illumination will be 
dispersed through different parts of the field (as shown in the figure by 
the dotted lines); any slight uneveness in the source of illumination 
will thus be distributed and not noticed by the eye. Great irregular- 
ities in illumination, however, must be avoided, and for this reason it 





JP 
Fig. 96. Showing method of illuminating polariscopes. 

is important that the instrument be kept in perfect alignment with 
its longitudinal axis at right angle to the source of light. It is best to 
have instrument and light rigidly fixed. Polariscopes mounted upon 
trestle supports are preferable to those upon tripods since a slight 
knock may swing the latter out of alignment and cause a change in 
the zero point. 

Variations in the brightness of illumination are also undesirable 
and for accurate work the emission of light should be constant. The 
optical center of gravity of purified sodium light, for example, is 
589.22 HJJL for a certain average brightness of flame; variations in this 
brightness, however, may change the wave length by 0.11 ///* with 
corresponding differences in the rotation of polarized light (25" for a 
rotation angle of 20 degrees). With salts of the alkalies and alkaline 
earths, increasing the brightness of flame (increase of vaporized salt 
per unit volume of flame) produces an irregular broadening of the 
spectral lines with a shifting of the mean wave length toward the red 
nd of the spectrum. 

Lamps for Sodium Light. Of the various polariscope lamps for 
sodium light only a few of the more common forms will be described. 
The lamp shown in Fig. 97 illustrates the essential principles of most 
sodium lamps. This consists of a Bunsen burner with side entrance 
for gas at s to prevent stoppage of inlet through dropping of fused 
salt; the burner is surmounted by a chimney which can be adjusted 



148 



SUGAR ANALYSIS 



to the desired height by the screw h. The holder for the fused salt 
consists of a spoon-shaped bundle of fine platinum wires attached to 
an upright support and can be moved in and out the flame through a 
slot in the chimney by means of the screw p; 
the door k, which closes the front of the chimney, 
allows only the brightest section of the flame to 
shine through and excludes 
the greater part of the heat. 
The flame is adjusted so as to 
be colorless, with as strong an 
air blast as possible, that the 
light may be free from incan- 
descent carbon particles. 

In place of 
wire holders for 
the salt many 
sodium lamps use spoons 
or V-shaped boats of sheet 
platinum or nickel, which 
are in some cases perfo- 
rated with fine openings. 

The hot part of the 

flame impinges 

upon the spoon and 

produces a sheet of 

sodium light upon 

each side. The 





Fig. 97. Simple form of 
sodium lamp. 



Fig. 98. Pribram's 
sodium lamp. 



fused salt must be renewed as fast as vaporized; a convenient means 
of effecting this renewal is shown in Pribram's * sodium lamp, Fig. 98, 
which contains two boats; the empty one is drawn out for refilling and 
the one in reserve inserted in its place. 

The sodium lamp of Landolt f, Fig. 76, gives a more intense flame 
than either of the lamps just described. It consists of a powerful 
Muencke gas burner with cylindrical chimney L. Upon the latter are 
placed two heavy nickel wires supporting rolls of fine nickel wire net- 
ting which contains fused salt. The burner is surmounted by a second 
rectangular chimney of sheet iron with a movable brass door containing 
apertures of 20, 15, and 10 mm. diameter. 

The simplest and cleanest of sodium lamps and the one giving the 
most continuous flame is that of Zeiss, Fig. 99. This is composed of 
* Z. analyt. Chem., 34, 166. f Z. Instrument., 4, 390. 



POLARISCOPE ACCESSORIES 



149 




an upper part A, capping an ordinary Bunsen burner and secured to 
it by means of a screw. The casting A carries the 
diaphragm-screen K, out of which the rectangular 
opening L is cut, also the flat burner C producing a 
square flame, and a small support for the salt carrier 
E, which consists of a piece of pumice stone, measur- 
ing about 4 X 1 X J cm., saturated with salt. It is held 
upon the support by the spring clip F and can be 
regulated to the flame by means of the screw / 
operating on the spring GH. It is best to adjust 
the pumice stone so that it merely touches and 
tinges the flame. If E be too deeply inserted in the 
flame, the latter is over-cooled and a dark, rather 
sharply defined zone is produced. The flickering 
margins of the flame are cut off by the diaphragm K. 
A few minutes are needed for heating the pumice 
before the flame attains its maximum brilliancy, 
after which it will remain constant for hours together. 
The tablets of pumice stone saturated with salt are 
supplied by the trade at small cost. 

In place of common salt, sodium bromide is 
sometimes used for illumination. This gives a much 
stronger flame, but the vaporization is much more 
rapid than with salt and there is the additional Fig. 99. The Zeiss 
disadvantage of giving off bromine vapors which 
may attack the instrument unless the lamp is placed under a hood. 

Sodium carbonate, sodium phosphate, sodium nitrite, and mixtures 
of these with salt in various proportions are also used for sodium 
lamps. Sticks of fused sodium carbonate heated hi an oxygen blast 
lamp give a flame of great brilliancy, and this is the form of light 
recommended by Landolt * when intense illumination is desired. 

Purification of Sodium Light. For accurate polariscope measure- 
ments it is necessary to purify the sodium light from other rays. This 
can be done either by use of light filters or by spectral separation of 
the extraneous rays. 

Sodium light can be freed from most of the foreign rays at the 
violet end of the spectrum by means of bichromate solution, which 
has a strong absorption band in the green and blue. The rays at the 
other end of the spectrum can be removed by uranous sulphate solu- 
tion, which has a strong absorption band in the red. A combination of 
* "Das optische Drehungsvermogen " (1898), p. 359. 



150 



SUGAR ANALYSIS 



these two solutions, as in the Lippich light filter, constitutes the most 
effective absorbent means of sodium-light purification known. 

Lippich Light Filter. The Lippich light filter consists of a tubular 
cell closed at the ends by tightly fitting cover glasses and divided by a 
glass plate into two smaller cells of unequal size. The larger cell, 10 cm. 
long, is filled with a 6 per cent filtered solution of potassium bichromate, 
the smaller cell is filled with a solution of uranous sulphate, 11(804)2, 
prepared as follows: 5 gms. of purest uranyl sulphate, UO 2 S04 + 3 H 2 0, 
are dissolved in 100 c.c. of water, and 2 gms. of powdered chemically 
pure zinc added; 3 c.c. of concentrated sulphuric acid are then added 
in 1 c.c. portions, waiting each time until the evolution of hydrogen has 
nearly ceased; the flask is corked during the reaction, and is allowed to 
stand about six hours, when the solution is filtered and the cell imme- 
diately filled in such a way as to leave only the smallest possible bubble 
of air behind. After standing for a day the cell is ready for use; the 
uranous solution retains its stability for one to two months, or until 
its deep green color is changed by oxidation into the yellow of the 
uranyl compound, when the cell must be refilled with fresh solution. 
The weights and volumes prescribed for making up the absorbent solu- 
tions must be rigidly adhered to. 

The spectrum purification of sodium light by means of glass prisms 
is the most thorough of all methods of purification. The process, 
which is a somewhat complicated one, is required, however, for only the 
finest physical measurements. Landolt gives the following average 
wave lengths for sodium light from different sources in which the wave 
length of the D l line is placed at 589.62 w and the A line at 589.02 ^. 

TABLE XXVI 
Wave Length of Different Kinds of Sodium Light 



Number. 



Source of light. 



Purification. 



Wave length 
in /x/x. 



Bunsen flame with NaBr . . . 
Bunsen flame with NaCl.. . . I 
Burner with NaCl or NaBr. j 

Sodium light < 

Landolt lamp with NaCl . . . j 

Bunsen flame with NaCl . . . . < 
Landolt lamp with NaCl 



10 cm. layer of 9 per cent ) 

K 2 Cr 2 O 7 in water. 
10 cm. layer of 9 per cent 

K 2 Cr 2 O 7 in water. 
Lippich filter K 2 Cr 2 O 7 and 

U(S0 4 ) 2 . 
Perfectly spectral pure; 

light of only the two D 

lines. 
1.5 cm. layer of 6 per cent 

K 2 Cr 2 O 7 in water. 
10 cm. layer of 9 per cent] 

K 2 Cr 2 O 7 in water and 1 ( 

cm. layer of 13.6 per cent j 

CuCl 2 in water. J 

Unpurified 



592.04 
589.48 
589.32 

589.25 
588.94 

588.91 
588.06 







POLARISCOPE ACCESSORIES 



151 



The Lippich light filter gives a wave length exactly between the 
two D lines of sodium and agreeing very closely with that obtained by 
spectral purification. In all cases where light filters are used the solu- 
tions must be placed between lamp and condensing lens (see Fig. 96). 

Lamps for White Light. For illuminating polariscopes and sac- 
charimeters with white light, a large number of lamps have been devised 
for use with oil, alcohol, gas, acetylene, and electricity. 





100. Hinks's oil lamp 
with duplex burner. 



Fig. 101. Hinks's gas lamp 
with triplex burner. 



A convenient form of oil lamp with duplex burner and adjustable 
support is that of Hinks, Fig. 100. The Hinks gas lamp with triplex 
burner is shown in Fig. 101. The metal chimneys of these lamps are 
usually fitted on the inside with a porcelain reflector; the focusing lens 
which is sometimes placed in the aperture of the chimney should be 
removed as it may cause an incorrect passage of the beams of light 
through the polariscope. 



152 



SUGAR ANALYSIS 



The best forms of gas lamp for illuminating are those provided with 
an Auer or Welsbach mantle (Fig. 102). The outer cylinder of these 
lamps, composed of sheet metal or asbestos, contains an opening whose 
lower half is covered with a plate of ground glass for diffusing the 
light; the upper uncovered part of the opening serves for illuminating 
the polariscope scale. A form of lamp for burning alcohol somewhat 
similar in design to the above is shown in Fig. 103. Gas burners for 
producing lime or zircon light are also used for illuminating polari- 
scopes. Acetylene lamps of 25 to 50 candle power give a light of great 






Fig. 102. Gas lamp 
with Welsbach mantle. 



Fig. 103. Alcohol lamp 
with Welsbach mantle. 



Fig. 104. Stereopticon 
electric lamp. 



brilliancy and are especially valuable upon sugar plantations where gas 
or electricity is not available. The acetylene lamps should be fitted 
with cylinders similar to those in Figs. 100 or 102. 

For electrical illumination a stereopticon 32-candle-power incan- 
descent lamp is very suitable (Fig. 104); the closely wound filament 
concentrates the light within narrow compass, giving great intensity of 
illumination. These lamps are best mounted in cylinders similar to 
that in Fig. 102; a plate of ground, glass is necessary for diffusing the 



POLARISCOPE ACCESSORIES 



153 



light, otherwise the irregularities in source of emission will not be suffi- 
ciently equalized for obtaining a uniform field. 

A small electric attachment devised by Schmidt and Haensch for 
illuminating their saccharimeters is shown in Figs. 88 and 105. The 




Fig. 105. Schmidt and Haensch six-volt saccharimeter lamp. 

small lamps are adapted for a six-volt current which is supplied by a 
storage battery or from the main line after reducing the voltage. The 
apparatus which is controlled by the switch S (Fig. 88) is screwed on 
the polarizing end of the saccharimeter. The electric lamp is held in 
position by two spring clips which are in connection with the two 
terminals. The lenses K 2 and KI (Fig. 105) project the light upon the 
diaphragm of the analyzer. As the horizontal filament is not always 
quite concentric to the frame, a vertical adjustment is necessary. To 
work the adjustment, the ring Z), which carries the lens K 2 , is rotated 
by the screw and projecting arm 6. If the lamp is also to be used for 
illuminating the scale of the instrument, the mirror S' (Fig. 88) is set 
at an angle of 45 degrees, in which position the reflected light is con- 
centrated by the lens H upon the opening a (Fig. 89). 



POLARISCOPE TUBES 

For retaining sugar solutions during polarization there are a variety 
of tubes of different construction, form, and length. In the selection 
of these the chemist must be guided more or less by the nature of his 
work. All tubes, however, when accuracy of observation is desired, 
must conform to three general requirements: (1) the length of the 
tube must be accurately fixed; (2) the ends of the tube and the sur- 
faces of its cover glasses must be plane parallel; (3) the tube must be 
centered evenly in its mountings and, when fitted with its caps, should 
be free from eccentricity. There are other minor requirements of 



154 



SUGAR ANALYSIS 



tube construction which will be given under the description of the 
different forms. 

Fig. 106 shows the most common and simplest forms of glass polar- 
ization tubes. These and other forms of tube are usually supplied in 
lengths of 25 mm., 50 mm., 100 mm., 110 mm., 200 mm., 220 mm., 
400 mm., 500 mm., and GOO mm.; for special kinds of work tubes of 
several meters' length have been constructed. 

A tube of 200 mm. length is used for the normal weight of all sac- 
charimeters. If, on account of depth of color, a 100-mm. or 50-mm. 
tube is employed and the resultant reading is recalculated by mul- 
tiplying by 2 or 4, there is, of course, a corresponding doubling or quad- 
rupling of the errors of observation; short observation tubes are to be 
used therefore only in extreme cases. With very dilute sugar solutions 





25 mm. tube. 



100 mm. tube. 




200 mm. tube. 
Fig. 106. Forms of plain glass polariscope tubes. 

and with sugars or sugar mixtures of low specific rotation the 400-mm. 
or 600-mm. tube will increase the accuracy of the observation, provided 
the color be not too great to disturb the reading. Tubes of odd lengths, 
such as 55 mm., 110 mm., 220 mm., etc., should be distinctly marked lest 
they be confused with the 50-mm., 100-mm., and 200-mm. sizes. 

Mounting of Polariscope Tubes. The ends of the glass observa- 
tion tubes are cemented into metal mounts which are threaded for the 
purpose of receiving the screw cap. Litharge and glycerine make a 
much better cement than the waxy material employed by most manu- 
facturers. The latter substance, especially on warm days, softens 
readily and when in this condition there is danger in screwing on the 
cap of drawing the mount from its setting so that it projects slightly 
beyond the ends of the tube; the length of the column of liquid to be 
polarized may thus be increased and a considerable plus error intro- 
duced in the observation. The ends of the glass tubes should project 
only slightly beyond the threaded heads; if too much of the end is 
exposed there is danger of chipping or breakage. The chemist should 
not attempt to reset his tubes unless he has a small lathe in which they 



POLARISCOPE ACCESSORIES 155 

can be centered and revolved while the cement is hardening, otherwise 
the tubes may not be evenly mounted. 

A simple means of testing for eccentricity of mounting 
is to place the tube, with caps screwed on, in the trough 
of a polariscope and while giving it a rotatory motion to 
view the opening through the tube with reference to the 
polariscope field. If the tube has been properly centered 
and the caps are free from eccentricity the tube opening will 
remain in the center of the field and show no wabbling 
movement during rotation. To test for plane parallelism of 
the ends of the tube and of cover glasses, the experiment 
just described is repeated with the cover glasses in position 
and the tube filled with water. If the ends of the tube have 
not been ground squarely across or the cover glasses are not 
plane parallel, the opening of the tube will wabble perceptibly 
during rotation owing to the refraction of light through the 
water from the inclined surfaces of the cover glasses. A 
difference of several tenths of a Ventzke degree may be noted 
between the readings of a tube in different positions through 
lack of plane parallelism in ends or cover glasses. According 
to Landolt the angle between the opposite ground-end sur- 
faces of a polariscope tube should always be less than 10 
minutes and the angle between the two planes of a cover 
glass less than 5 minutes. The small angles of inclination 
between planes of cover glasses and between ends of tubes not 
exceeding 200 mm. in length is measured by a spectrometer. 

Calibration of Polariscope Tubes. A most convenient 
means of calibrating the length of polariscope tubes is the 
measuring gauge of Landolt, shown in Fig. 107. This gauge, 
which has an adjustable handle c, consists of a measuring 
rod A of steel graduated for a distance of 400 mm. and 
provided with a sliding vernier b which gives readings to 
0.1 mm. The lower end of the rod and the bottom of the 
vernier are provided with knife edges. When the knife edge 
of the rod is placed upon a smooth hard surface, such as glass, 
and the vernier brought down until its knife edges are in Fig. 107. 
close contact with the same surface, the zero point of scale gauge f or 
and vernier should agree. If there is lack of agreement, the calibrating 
zero point of the vernier may be either adjusted or the differ- polariscope 
ence noted and applied to all readings. To calibrate an tubes, 
observation tube, one end of the tube is closed with its cover glass and 



156 SUGAR ANALYSIS 

cap, and after placing in an upright position with the closed end down 
the measuring rod is inserted until its knife edgejtouches the cover 
glass; holding the rod perfectly upright the vernier is slipped down 
until its knife edges coincide with the upper end of the tube; the read- 
ing of the scale and vernier will then give the length of tube. Other 
readings are made, rotating the rod a little each time from its original 
position, and the average taken. Calibration of tubes should be made 
at the standard temperature 20 C.; if measurements are made at tem- 
peratures very different from this the changes in length of tube and 
gauge due to expansion or contraction must be taken into account (co- 
efficient of expansion in length 1 C. for steel = 0.000013 and for glass 
= 0.000008). Measuring gauges can be verified as to accuracy at the 
Government Bureau of Standards. 

The measuring gauge of Landolt will detect an error of 0.1 mm., 
which is equivalent to an error of 0.05 V. for a sugar solution polarizing 
100 V. in a 200-mm. tube. This is sufficiently close for ordinary 
saccharimetric measurements; if a finer determination of tube length 
is desired the measurement must be made upon a comparator; by means 
of this instrument measurements can be made to 0.01 mm. 

Cover Glasses. The cover glasses used upon polariscope tubes 
must be of strong, colorless, and optically inactive glass; their surfaces 
must be plane parallel and free from cracks or scratches. In screwing 
the caps upon observation tubes, care must be taken that no severe 
pressure is brought to bear upon the cover glasses; otherwise the strain 
will render the glass optically active and produce serious errors in the 
observation. If a cover glass is optically active turning the tube in the 
trough of the polariscope will usually show variations in the intensity 
of the field with considerable difference in the reading for various posi- 
tions of the tube. The practice of rotating the observation tube be- 
tween readings is always a good one; in this way errors due to defective 
cover glasses, bad washers, pressure of caps, eccentricity, etc., may be 
detected which would otherwise escape notice. Cover glasses which 
have been rendered optically active through pressure should not be 
used for a day, in order that sufficient time may elapse for readjustment 
to neutrality. 

Washers. Another common source of error in polariscopic work 
are badly fitting rubber washers in the screw caps of the tubes. The 
washers should be of soft rubber and lie evenly against the back of 
the cap without the slightest marginal elevation, otherwise the washer 
in tightening the cap may give the cover glass an inclined position and 
cause a considerable increase in the reading. 



POLARISCOPE ACCESSORIES 



157 



Special Forms of Polariscope Tubes 

Schmidt and Haensch Tube with Enlarged End. Another form of 
glass polarization tube which presents several advantages is the Schmidt 
and Haensch tube with one end enlarged (Fig. 108). The enlargement 
serves as a receptacle for any air bubbles which may be enclosed with 
the liquid; the retention of a small air bubble in the tube is in fact de- 
sirable since, by moving the bubble through the liquid from end to end 




Fig. 108. Schmidt and Haensch polariscope tube with enlarged end. 
(Air bubbles are collected at the point a, outside of the field of vision.) 

before reading slight differences in temperature are equalized, and no 
troublesome striations, due to currents of solution of different tem- 
perature, are present to distort the field. Tubes without enlargement 
must not retain air bubbles with the liquid; if striations are present 
the tube must remain at rest until the solution has reached equilibrium. 
The most frequent cause of a striated field is the warming of the solution 
in the tube by the hand; for this reason tubes should be handled only 
by the metal caps when placing in the instrument. 




Fig. 109. (a) 200 mm. Landolt polariscope tube with sliding cap and enlarged end; 
(6) 200 mm. metal polariscope tube. 

Landolt' s Tube. To prevent the liability of excessive pressure 
upon cover glasses, Landolt has devised a tube with sliding cap, 
which is shoved into position over the metal mount (Fig. 109a). The 
French manufacturers also provide a cap that is shoved on and 
fastened with a bayonet catch. Tubes with screw caps, however, are 
the ones most preferred and, if care be taken not to draw them up 
too tightly, will be found to answer all requirements. When observa- 
tion tubes are used in large numbers it is a great advantage to have 
all caps interchangeable. 

Metal Polarization Tubes. Polarization tubes of brass or nickel or 
other metal are preferred by many chemists. Such tubes, a form of 
which is shown in Fig. 109b, have the advantage of greater durability, 



158. 



SUGAR ANALYSIS 



but the disadvantage of being susceptible to the attack of acids (as 
in the method of inversion) or other corrosive liquids. Brass tubes 
have also more than twice the coefficient of expansion of glass tubes, 




Fig. 110. Pellet's tube for continuous polarization. 

the coefficient (/?) for 1 C. being 0.000008 for glass and 0.000019 for 
brass. For glass and brass tubes measuring exactly 200 mm. at 20 C., 
the length at 35 C. (Le = L 20 [l +0 (*-20)]) = 200.024 mm. for glass 




i 




Fig. 111. Glass polarization tube with metal jacket. 

and 200.057 mm. for brass, errors in length of no great significance. A 
more serious objection against metal tubes is the danger of their being 
bent out of alignment through hard or long usage. A knock or fall 
may cause a metal tube no apparent injury yet may bend it sufficiently 
to produce a considerable error in the polariscope reading. A number 
of brass polariscope tubes, recently submitted to the author for examina- 



POLARISCOPE ACCESSORIES 



159 



tion, were so badly out of alignment that rotating the tubes in the trough 
of the polariscope caused a difference of over 0.2 V. in the reading. 

Pellet's Tube for Continuous Polarization. In the polarization of a 
large number of solutions in succes- 
sion, as in the analysis of sugar beets, 
juices, etc., the Pellet tube for con- 
tinuous polarizations is often of great 
use. Sections of this tube, which is 
made of metal, are shown in Fig. 
110. The ends of the tube are closed 
and after placing in the instrument 
the solution to be polarized is poured 
through a small funnel into one of 
the nipples, a or 6, the excess escap- 
ing through an exit tube connected 
by rubber tubing to the nipple at 
the opposite end. As soon as the 
solution is polarized, the succeeding 
solution is poured into the tube; the 
disappearance of striations and the 
clearing of the field indicate when 
the previous solution has been com- 
pletely displaced. The Pellet tube 
will accomplish a valuable saving of 
time in certain kinds of work, but it 
is usually advisable to limit its use 
to sugar solutions of approximately 
the same density; to displace a con- 
centrated sugar solution with one 
that is exceedingly dilute, or vice 
versa, is attended with more or less 
risk of error. 

Polarization Tube with Metal 
Jacket. For polarizing sugar solu- 
tions, where the temperature must 
be measured or controlled, a jacketed 
observation tube such as shown in Fig. Ill is recommended. This con- 
sists of an inner tube of glass or metal with a central opening, c, which 
can be used for filling and for inserting a thermometer; an outer mantle 
of brass or nickel surrounds the inner tube and is provided with nipples 
for inlet and exit of hot or cold water as may be desired. 




Fig. 112. Reservoir for supplying 
water of constant temperature. 



160 



SUGAR ANALYSIS 



For supplying water of constant temperature for observation tubes, 
the Zeiss apparatus described on page 59 may be used. A form of 
water supply reservoir with stirrer, recommended by Landolt,* is shown 
in Fig. 112. The reservoir, which is insulated, is filled through the 
opening A with water to the desired level, indicated by the tube D. 
The water is heated by means of a burner to the desired temperature, 
shown by the thermometers at C, the heat being equalized by raising 
and lowering the stirrer B. 

A form of constant temperature bath designed by Hudson f is 
shown in Fig. 113. The mechanical stirrer not only secures an even 
temperature through the bath, but also acts as a rotary pump which 



Mercury 
Sealed Joint 



From.Polarimeter 



To Polarimeter 



\ 



Fig. 113. Hudson's constant temperature water-bath. 

creates a constant circulation of water as shown by the direction of 
the arrows. 

Wiley's Desiccating Caps. When solutions are polarized at tem- 
peratures below the dew point of the atmosphere, the cover glasses of 
the observation tube must be protected against condensation of moisture 
by means of desiccating caps such as designed by Wiley J (Fig. 114). 
These are generally made of some non-conducting material such as 
hard rubber: they are closed at the end with a tightly fitting cover 
glass and contain a tube for holding calcium chloride or other desiccat- 
ing substance. 

"Das optische Drehungsvermogen " (1898), pp. 397. 
t Hudson, J. Am. Chem. Soc. 30, 1572. J J. Am. Chem. Soc. 18, 81. 



POLARISCOPE ACCESSORIES 



161 



When solutions are polarized at very high temperatures as at 
87 C. (the point of inactivity for invert sugar) the use of glass, unless 
carefully annealed, for the inner tube of the water jacket is precluded. 
Polariscopic work at high temperature is generally performed in 





(I) 

Fig. 114. (I) Threaded cap of polariscope tube. (II) Dessicating cap which screws 
on over threads of (I) ; t, removable glass tube containing dessicating substance s; 
w, inner perforated metal tube; g, cover glass held in position by threaded disk r; 
the disk is unscrewed by inserting a spanner in the two holes marked in black. 

jacketed tubes constructed entirely of brass or nickel, the inner surface 
of which has been gold plated. The length of a 200-mm. tube (20 C.) 
at 87 C would be 200.107 mm. for glass and 200.255 mm. for brass, 
equivalent to a plus error of 0.054 V. and 0.128 V. respectively for 
solutions polarizing 100 V. in a 200-mm. tube. 




Fig. 115 Yoder's volumetric polariscope tube. 

Yoder's Volumetric Polariscope Tube. A volumetric polariscope 
tube is convenient for certain kinds of saccharimetric work. A tube of 
this description, designed by Yoder, is shown in Fig. 115. 

The capacity of the tube to the graduation mark upon the neck is 
10 c.c. By varying the length and diameter the tubes can be adjusted 
to any convenient volume. 



162 



SUGAR ANALYSIS 



BALANCES FOR POLARISCOPIC WORK 

For the operations of weighing in saccharimetric work three types 
of balances are required, an analytical balance, a so-called sugar balance, 
and a balance for coarse weighing. 

The analytical balance should have a capacity of 200 gms. and with 
this load be sensitive to 0.1 mg. Such a balance is required for all 
analytical processes, for determination of specific rotations, for cali- 
bration of flasks, weighing of pycnometers, and all other operations 
where accuracy is essential. A balance of the type shown in Fig. 17 
will answer for this purpose. With this balance a set of accurate 
analytical weights (including one 100 gms. weight) will be needed. 







Fig. 116. Sugar balance. 



In addition to the above a less delicate balance, sensitive to 2.5 mgs. 
with a load of 250 gms., will be required for the rapid weighing of definite 
amounts of sugar, molasses, and other products for ordinary sacchari- 
metric work. For saccharimeters employing a normal weight of 
26 gms., 0.01 degree Ventzke corresponds to 0.0026 gm. sucrose in 
100 true cubic centimeters. Since the majority of saccharimeters can 
be read only to 0.05 V it is evident that weighing within 5 mgs. is 
sufficiently accurate for ordinary purposes of saccharimetry. The 
weighing out of normal weights of sugar, etc., for saccharimeters should 
not be done upon an analytical balance; the errors due to evaporation 
from moist substances during the slower adjustment of the analytical 
balance will usually exceed any advantage in greater accuracy of 
weight. A so-called " sugar balance " of the type shown in Fig. 116 



POLARISCOPE ACCESSORIES 



163 



answers very well for this kind of work. This balance may also be 
used for the weighing out of chemicals for making up solutions of 
reagents. A set of second quality weights should be provided for ap- 
proximate weighing, and also the normal weights belonging to the 
saccharimeter. 

The Mohr cubic centimeter normal and half-normal weights (26.048 
gms. and 13.024 gms.) are usually furnished in a cylindrical form and 
the true cubic centimeter weights (26.000 gms. and 13.000 gms.) in a 
cubical form (Fig. 123), the shape of the weight thus guarding against 




Fig. 117. Metric solution scale. 

confusion. Normal weights, which are in constant use, should be tested 
frequently upon the analytical balance against losses in weight through 
wear. If a deficiency exceeding 1 mg. is noted, the stem of the weight 
should be unscrewed and a small piece of tin or aluminum foil be placed 
in the cavity sufficient to bring the weight up to its proper value. 

In addition to the two balances just described a heavy balance or 
scale for weighing out material in bulk, preparing large quantities of 
reagents, etc., will be required. A metric solution scale with sliding 
counterpoise such as is shown in Fig. 117 is very good for this purpose. 
A set of third quality weights up to 5 kgs. should also be provided for 
coarse weighings. 

FLASKS FOR POLARISCOPIC WORK 

For the preparation of sugar solutions in polarimetric and sac- 
charimetric work various flasks have been devised of different shape 
and construction. 



164 SUGAR ANALYSIS 

Flasks for Solution by Weight. When sugar solutions are made 
up according to percentage a glass-stoppered flask of the form shown 
as No. VI in Fig. 118 is recommended. The flask, which is supplied in 
many sizes, need not be graduated. Before using, it is thoroughly 
cleansed and dried, and then weighed. The approximate quantity of 
substance to be examined is then transferred to the flask and after 
stoppering the latter is reweighed. The approximate amount of dis- 
tilled water or other solvent is then added and the flask and contents 
reweighed as before. The percentage of substance in solution is 
then readily calculated from the weight of substance taken and the 
combined weights of substance and solvent. The flask should not be 
filled too full; sufficient space should be left for gentle rotation of the 
liquid while effecting solution. The flask should always be kept 
stoppered to prevent evaporation. 

Reduction of Solution Weights to Vacua. For very accurate physical 
measurements the weights taken in air must be reduced to vacuo, 
since a substance weighed in any medium loses in weight an amount 
equal to that of the medium displaced. If W is the true weight of a 

W 

substance of density D, in vacuo, then the volume of substance is -~> 

and if s is the density of the air at the time of weighing, the loss in 

sW 
weight of the substance in air will be -jr- - Similarly if P is the value 

of the weights in vacuo and d is the density of their material then the 

sP 

loss of the weights in air will be -7- The equilibrium upon the pans of 

the balance between substance and weights in air will then be repre- 
sented by the equation 



i-- 

whence W = P - - 



The mean value 0.0012 gm. may be taken as the weight of 1 c.c. of 
air without sensible error. When brass weights are used (d = 8.4), 
the weights in vacuo of glass, water and sugar are found as follows: 
for glass (D = 2.5) W = 1.000337 P, for water 20 C. (D = 0.998234) 
W = 1.001061 P, for cane sugar (D = 1.59), W = 1.000612 P. The 
following example will illustrate the method of application. 



POLARISCOPE ACCESSORIES 



165 



Weight of flask + sugar in air 35.2326 gms. 

Weight of flask alone in air 25.1240 gms. 

Weight of sugar in air 10.1086 gms. 

Weight of sugar in vacuo = 10.1086 X 1.000612 = 10.1148 gms. 

Weight of flask + sugar + water in air 95.3055 gms. 

Weight of flask + sugar in air 35.2326 gms. 

Weight of water in air 20 C 60.0729 gms. 

Weight of water in vacuo = 60.0729 X 1.001061 = 60.1366 gms. 

Weight of sugar + water in vacuo = 70.2514 gms. 

Per cent sugar in solution from weights in air = 14.403 per cent. 
Per cent sugar in solution from weights in vacuo = 14.398 per cent. 

It will be noted that the difference is exceedingly slight, so that 
weighing in air is sufficiently exact for all operations except those de- 
manding extreme accuracy. 

Volumetric Sugar Flasks. When solutions of dissolved sugars 
are made up to a definite volume before polarization, a carefully cali- 
brated volumetric flask must be used; such flasks are supplied in a 
variety of forms and sizes. If solutions are polarized immediately 
after making up to volume, as is usually the case, it is not essential 
that the flask be fitted with a glass stopper. 



'1 









III JV V 

Fig. 118. Types of flasks for polariscopic analysis. 

Volumetric flasks for sugar work are made in 10-c.c., 20-c.c., 25-c.c., 
50-c.c., 100-c.c., 200-c.c., and 250-c.c. sizes; 500-c.c. and 1000-c.c. flasks 
are also occasionally used. For certain kinds of work, where volume 
of insoluble matter is allowed for, flasks of irregular capacity are used, 
as 100.6-c.c., 201.2-c.c., etc., for polarization of sugar-beet pulp. 

A few of the more ordinary forms of sugar flask are shown in Fig. 
118. These may be obtained of any desired capacity. Small sized 
stoppered flasks similar to No. I are convenient for preparing solutions 
when small amounts of substance are available. Kohlrausch's sugar 



166 SUGAR ANALYSIS 

flask (No. IV) with enlarged top is convenient for transferring sub- 
stances and is in many ways a most desirable flask; it can be obtained 
in the small sizes and, if desired, with ground-glass stopper. Sugar 
flasks with double graduation (No. Ill) for one-tenth dilution are useful 
for the methods of inversion; they are supplied in 25-27.5-c.c., 50- 
55-c.c., 100-110-c.c., and 200-220-c.c. sizes. 

Specifications for Sugar Flasks. In the selection of sugar flasks 
the following requirements of the United States Bureau of Standards 
for volumetric flasks will be found useful. 

" The cross section of all flasks must be circular throughout and 
the neck must merge into the body of the flask so gradually that there 
will be no hindrance to uniform drainage. Flasks that are manifestly 
fragile or otherwise defective in construction will be rejected. The 
part on which the graduation mark is placed must be transparent, of 
uniform thickness, and free from striae. The graduation mark must be 
placed not less than 6 cm. from the upper end and not less than 2 cm. 
from the lower end of the neck of a flask larger than 100 c.c., and not 
less than 3 cm. from the upper end or 1 cm. from the lower end of the 
neck of a flask not larger than 100 c.c. The graduation mark must 
extend entirely around the neck. The bottom of the flask must be 
slightly reentrant, and the flask must be of such form that drainage 
can take place from the whole interior surface at the same time. The 
neck of a flask must be perpendicular to a plane tangent to the bottom. 
The flask must stand solidly when placed on a horizontal plane." 

A very desirable 100-c.c. flask for saccharimetric work is that shown 
in No. II, Fig. 118, and in Fig. 123 designed for use in the Custom-House 
Laboratories of the United States Treasury Department. The pear- 
shaped body with its low center of gravity gives the flask greater stability 
than a spherical form. According to the regulations of the Treasury 
Department "the flasks shall have a height of 130 mm.; the neck shall 
be 70 mm. in length and have an internal diameter of not less than 
11.5 mm. and not more than 12.5 mm. The upper end of the neck 
shall be flared, and the graduation marks shall be not less than 30 mm. 
from the upper end and 15 mm. from the lower end of the neck." 
With this size of flask the base of the thumb can cover the mouth and 
the fingers of the same hand easily enclose the bottom a feature 
of great convenience when mixing the contents after making up to 
volume. 

Calibration of Sugar Flasks. Sugar flasks are graduated to 
contain 100 true cubic centimeters at 20 C. or 100 Mohr cubic centi- 
meters at 17.5 C. and should be calibrated before using in the follow- 



POLARISCOPE ACCESSORIES 



167 



ing manner. The flask to be tested is first thoroughly cleaned and 

dried, then weighed empty at the temperature of standardization, and 

then again when filled to the mark with distilled water at the standard 

temperature. The distilled water should be boiled just before using, 

in order to expel dissolved air, and then cooled. Special 

care is necessary in adjusting the meniscus to the 

graduation mark; the lowest point of the curve when 

viewed against a white surface should just touch the 

level of the graduation mark, the latter appearing to 

the eye in proper position as a straight line and not 

as an ellipse. Fig. 119 indicates the proper method of 

adjustment. The inside of the neck above the meniscus 

should be wiped perfectly dry with filter paper before 

reweighing; air bubbles should not be allowed to adhere 

to the walls of the flask during calibration. Fi g< 119. Showing 

Volumetric 100-c.c. sugar flasks graduated according proper adjustment 
to the Mohr system should contain 100 gms. of distilled of meniscus, 
water at 17.5 C., when weighed in air against brass weights; 100-c.c. 
flasks graduated according to true cubic centimeters should contain 
100 gms. of distilled water at 4 C. when weighed in vacuo or 99.7174 
gms. at 20 C. when weighed in air with brass weights. (Weight in 
vacuo of 100 c.c. water at 20 C. is 99.8234 gms. and weight in air 
(p. 164) is 99.8234-^ 1.001061 = 99.7174 gms.) The grams of water 
contained by the flask at 20 C. plus the correction 0.282 will give the 
volume in true cubic centimeters. 

The limits of error allowed by the United States Bureau of Stand- 
ards for volumetric flasks are the following: 




Capacity. 


Limit of error. 


c.c. 


c.c. 


2000 


0.5 


1000 


.3 


500 


.15 


250 


.1 


200 


.1 


100 


.08 


50 


.05 


25 


.03 


10 


.01 



The limit of error allowed above for 100-c.c. flasks is, however, too 
high; the error of graduation should not exceed 0.05 c.c. and careful 
manufacturers can conform to this requirement without trouble. A 



168 



SUGAR ANALYSIS 



lot of 200 sugar flasks purchased by the New York Sugar Trade Labora- 
tory showed the following errors upon calibration. 



Error in volume. 


Number of 
flasks. 


Percentage. 


Between 0.00 c.c. and 0.01 c.c. . . 


65 


32.50 


Between 0.01 c.c. and 0.02 c.c. . . 


56 


28.00 


Between 0.02 c.c. and 0.03 c.c. . . 


43 


21.50 


Between 0.03 c.c. and 0.04 c.c. . . 


27 


13.50 


Between 0.04 c.c. and 0.05 c.c. . . 


7 


3.50 


Between 0.05 c.c. and 0.06 c.c. . . 


2 


1.00 




200 





It is seen that 99 per cent of the flasks were correct within 0.05 c.c. 
and that over 95 per cent were correct within 0.04 c.c. 

FUNNELS AND CYLINDERS 

In filtering sugar solutions for polarization short-stemmed funnels 
and cylinders of any of the forms shown in Fig. 120 will be found con- 




I II III IV 

Fig. 120. Types of cylinders for polariscopic analysis. 

venient. The funnels and filters should be of sufficient size to retain 
100 c.c. of solution; they should be covered with large watch glasses 
during filtration to prevent evaporation. Tall narrow filtering cylinders 
(Nos. I and II, Fig. 120) are preferred by some chemists for the reason 
that the least surface of filtered liquid is exposed to evaporation. The 
small-lipped filtering jars (No. IV, Fig. 120) are more convenient, how- 
ever, for filling tubes, and if covered by funnels and watch glasses will 



POLARISCOPE ACCESSORIES 



169 



not allow sufficient evaporation, during the necessary time of filtra- 
tion, to cause any appreciable error in the polariscope reading. 

MOUNTING OF POLARISCOPES AND CARE OF APPARATUS 
If the circumstances permit, polariscopes should always be mounted 
in a separate room or compartment, where there is no danger of cor- 
rosion from the action of fumes or vapors. The polarizing compart- 
ment should be well ventilated and easily darkened; lamps and burners 
for illumination should be placed upon the opposite side of a wall or 
partition. 




Fig. 121. Cabinet for constant temperature polarization (New York Sugar 
Trade Laboratory). 

In the New York Sugar Trade Laboratory the polariscope cabinet 
(Fig. 121) constitutes a section of the constant-temperature room. 
The roof of the cabinet is composed of shutters, for regulating the 
downward passage of cool air, and the sides of the cabinet are enclosed 
by dark curtains, which, when drawn, leave a space of one foot at the 
bottom. This arrangement allows free circulation of air, and the 
presence of several observers in the cabinet does not affect the tem- 
perature. 



170 



SUGAR ANALYSIS 



Where room is not available for a separate compartment, the 
polariscopes may be mounted in a large box in a dark corner of the 
laboratory as shown in Fig. 122. 

The table supporting polariscopes should be of solid construction. 
By placing the table upon rubber cushions and setting the polariscopes 
upon rubber mats, vibration of the instruments and consequent dis- 
turbance of the zero point will be largely obviated. 




Fig. 122. Portable polariscope cabinet with section of side removed. 

It is essential in saccharimetric work that all apparatus be kept 
scrupulously clean. The more delicate optical parts of polariscopes, 
such as polarizer, analyzer, and quartz compensation, are enclosed, in 
the most modern apparatus, in dust-proof housings, and very rarely 
require to be disturbed. The diaphragm glasses (A and P, Fig. 96) at 
each end of the polariscope trough are the parts which require most 
attention. Drops of solution, accidentally adhering to the polariscope 
tubes, are occasionally splashed against the diaphragm glasses. The 



POLARISCOPE ACCESSORIES 171 

diaphragms, which either screw or slide into position, should be ex- 
amined frequently and the glasses wiped free of dirt and dust particles. 
A paper napkin will be found very suitable for cleaning diaphragm 
glasses, eye pieces, and other exposed optical parts. 

The troughs of polariscopes in the hasty round of routine fre- 
quently become soiled from contact with wet tubes or spilled liquid. 
They should be wiped frequently with a damp cloth and the metal 
surface should be kept smooth and clean. 

The bichromate cell should be examined frequently, and the solu- 
tion replenished as soon as bubbles begin to form, otherwise their 
appearance may obscure the field. 

When the polariscope is not in use, the trough should be closed and 
the instrument kept covered. 

Strict cleanliness must also be observed in the use of polariscope 
tubes, flasks, and other accessories. In handling and carrying obser- 
vation tubes a portable rack of the form shown in Fig. 122 will be 
found convenient. 

Where sugar solutions are clarified with lead subacetate, the walls 
of flasks, cylinders, funnels, and tubes become coated in time with a 
thin white film of lead carbonate. A good solvent for this coating is a 
warm solution of sodium hydroxide and Rochelle salts, such as is used 
in preparing Fehling's solution. Hydrochloric or nitric acid may also 
be used for removing the deposit. After thorough rinsing in clean 
water, tubes, flasks, funnels, and cylinders should be allowed to drain 
and dry upon racks. 



CHAPTER VIII 

SPECIFIC ROTATION OF SUGARS 

IN the previous chapters the principles which underlie the con- 
struction and operation of polariscopes were described; it is now de- 
sired to study the application of these principles to some of the problems 
of sugar analysis. 

The polarizing power of a sugar is expressed as specific rotation, or 
specific rotatory power, by which is meant the angular rotation which 
a calculated 100-per cent solution, 1-dcm. long, gives to the plane of 
polarized light. The specific rotation, indicated by the expression [a] 
can easily be calculated from the angular rotation a of the solution of 

substance by means of the equation [a] = = > in which c is the con- 

c X l 

centration of substance (grams per 100-c.c. solution) and I the length 
of the observation tube in decimeters. Instead of the foregoing we 

may use the equation [u] = ^ 7 > in which p is the percentage of 

J) X CL X L 

substance in solution (parts by weight hi 100 parts by weight of solu- 
tion) and d is the specific gravity of solution, (p X d = c in previous 
equation.) 

The angular rotation, as shown below, depends upon the wave 
length of the light employed. Sodium light is the illumination most 
used for polariscopic measurements and as the bright yellow line of 
sodium is designated the D line of the solar spectrum, the expression [a] 
for sodium light is written [a]b. Specific rotation for the mean yellow 
ray j (now no longer used) is written [a]j. The temperature at which 
the specific rotation is taken is also usually affixed. Thus: the symbol 
for specific rotation using sodium light at 20 C. is written [a]. 

The method of calculating specific rotation may best be understood by 
an example; 20 gms: of cane sugar dissolved to 100 c.c. gives an angular 
rotation for sodium light of +53.2 degrees in a 400-mm. tube at 20 C. 

Substituting these values in the equation [a] = =-> we obtain 

c x l 

100 X 53 2 
MD = on x 4 = ~^~ ^'** *^ e specific rotation of sucrose for the given 

concentration. 

172 



SPECIFIC ROTATION OF SUGARS 



173 



To calculate specific rotation from the reading of a saccharimeter, 
the scale divisions of the latter must first be converted to angular 
degrees by means of the appropriate factor. Thus: 15 gms. of sucrose 
dissolved to 100 metric cubic centimeters gave a reading of +57.7 in 
a 200-mm. tube using a Ventzke scale quartz-wedge saccharimeter. 
Since 1 V = 0.34657 angular degrees (page 145) then 



r , 100 (0.34657 X 57.7) 
[a] D sucrose = 15 x 2 



+ 66.6. 



EFFECT OF KIND OF LIGHT UPON SPECIFIC ROTATION OF SUGARS. 

Mention has been made of the influence of wave length of light upon 
specific rotation. In Table XX a comparison was given of the rotations 
of quartz and sucrose for light of different wave lengths and it was 
shown that as the wave length decreases the polarizing power of sucrose 
increases. In the following table the specific rotations of nine different 
sugars are given for light of different wave lengths in the red, yellow, 
green, blue, indigo, and violet parts of the spectrum, according to 
recent measurements by Grossmann and Bloch.* The specific rota- 
tions for yellow sodium light, [O\D, the standard values of comparison, 
are printed in heavier type. 



Sugar. 


Concen- 
tration, 
gms. 
100 c.c. 


Red 
(r) 
656 MM- 


Yellow 
(f) 
589 'pp. 


Green 

to) 

535 MM- 


Blue 
(W 

508 MM- 


Indigo 
() 

479 MM- 


Violet 
() 
447 MM- 


Disper- 
sion co- 
efficient 

V 

r 


Xylose 
Rhamnose . . 
Galactose . 
Glucose 
Fructose. . . 
Sucrose .... 
Lactose. . . . 
Maltose. . . 
Raffinose. . 


0.866 

6.948 
5.603 
4.500 
4.500 
4.275 
2.000 
6.021 
3.713 


+ 13.28 
+ 7.08 
+ 60.80 
+ 41.89 
- 76.39 
+ 53.18 
+ 39.82 
+ 111.00 
+ 79.63 


+ 18.19 
+ 8.37 
+ 80.72 
+ 52.76 
- 90.46 
+ 66.50 
+ 52.42 
+137.04 
+105.20 


+ 21.08 

+ 10.27 
+ 99.63 
+ 65.35 
-107.21 
+ 82.25 
+ 62.09 
+166.11 
+131.71 


+ 24.50 
+ 11.11 
+116.76 
+ 73.61 
-136.85 
+ 91.53 
+ 72.25 
+176.26 
+150.75 


+ 27.70 
+ 12.84 
+131.84 
+ 83.88 
-151.11 
+104.24 
+ 83.25 
+227.12 
+163.77 


+ 31.94 
+ 14.38 
+ 152.90 
+ 96.62 
-166.55 
+ 121.63 
+ 98.17 
+233.36 
+ 188.55 


2.41 
2.03 
2.51 
2.30 
2.18 
2.29 
2.47 
2.10 
2.37 



Average 2.296 

It is seen that of the nine sugars galactose shows the greatest and 
rhamnose the smallest dispersion . coefficient, the average value 2.296 
being the same as that of sucrose and of glucose. 

Various formulae have been proposed for expressing the relationship 
between specific rotation and wave length of light. Stefan f gives for 

cy COQ 

sucrose the formula [a] = ^f- - 5.58, in which the wave length X is 

A 

* Z. Ver. Deut. Zuckerind., 62, 19. t Ber. Wiener Akad., 62, 486. 



174 



SUGAR ANALYSIS 



expressed in ten-thousandths of a millimeter 



The results as 



thus calculated have only an approximate value, as other factors, such 
as temperature, concentration, etc., are not considered. 

The specific rotations of the different sugars also vary according 
to the concentration of solution, the temperature of observation and 
the nature of the solvent. The following table gives the approximate 
values for the specific rotation of a number of sugars. The effect of 
concentration and temperature in increasing or lowering the specific 
rotation is indicated by the direction of the arrow in the respective 
columns. 

TABLE XXVII 

Showing Effect of Increase in Concentration and Temperature upon Specific Rotation 

of Sugars 



Sugar. 


wr- 


Increase in 
concentration 

-o+ 


Increase in 
temperature 

-0+ 


Arabinose 


+104.5 
+19.0 
+8.5 
+80.5 
+52.5 
-92.5 
-20.0 
+66.5 
+52.5 
+138.5 
+ 104.5 





? 

? 



? 

V 


5 


? 


? 


Xylose 


Rhamnose 


Galactose 


Glucose 


Fructose 


Invert sugar 


Sucrose 


Lactose -. 


Maltose .... 


Raffinose. . . . 





EFFECT OF CONCENTRATION UPON SPECIFIC ROTATION OF SUGARS 

The effect of varying concentration upon the specific rotation of 
sugars has been studied by many observers and the results of their ob- 
servations have been expressed in the form of equations. The method 
of deriving these equations, which is due to Biot,* is of considerable 
importance to the sugar chemist and deserves to be briefly considered. 
Concentration Equations. If the specific rotations of a substance 
for different concentrations be laid off upon a diagram, in which the 
specific rotations represent the ordinates and the percentages of sub- 
stance in solution the abscissae, the line which connects the several 
points, will be either a straight line, a section of a parabola, hyperbola, 
or other curve, or a combination of any two or more of these. Calling 
the percentage of sugar in solution p, the specific rotation can be rep- 
resented as follows: according to the well-known algebraic equations. 
* Ann. chim. phys. [3], 10, 385; 11, 96; 28, 215; 36, 257; 69, 219. 



SPECIFIC ROTATION OF SUGARS - . 175 

I. For the straight line [a] = a + bp. 
II. For the parabola [a] = a + bp + cp 2 . 

III. For the hyperbola [a] = a + -^ 

C + p 

Having plotted and determined the nature of the curve it remains 
to calculate the values of the constants a, b, and c in the above equa- 
tions. The method of doing this (the method of least squares) is simple, 
although the work of calculation is somewhat laborious. The following 
example is given as an illustration : 

From the average results of observations by Tollens, Thomson, Schmitz, 
Nasini, and Villavecchia, the following specific rotations of sucrose were found 
for different concentrations; 10 per cent + 66.56, 20 per cent + 66.52, 30 per 
cent + 66.41, 40 per cent + 66.27, 50 per cent + 66.06. An equation is desired 
for the specific rotation of sucrose for any concentration within these limits. 

By plotting the above observations a curved line is obtained presumably 
a parabola. (In calculating the concentration curves for the specific rotation of 
sugars the hyperbola is but little used.) Substituting the results in the previous 
equation II for the parabola we obtain the following : 

1. a + 10 b + 100 c = 66.56. 

2. a + 20 b + 400 c = 66.52. 

3. a + 30 b + 900 c = 66.41. 

4. a + 40 b + 1600 c = 66.27. 

5. a + 50 b + 2500 c = 66.06. 
Average: I. a + 30 6 + 1100 c = 66.364. 

From the above equations we obtain by subtraction the following: 

6. (5-1) 40 b + 2400 c = - 0.50. 

7. (5-2) 30 b + 2100 c = - 0.46. 

8. (5-3) 20 6 + 1600 c = - 0.35. 

9. (5-4) 106+ 900 c =-0.21. 

10. (4-1) 30 b + 1500 c = - 0.29. 

11. (4-2) 20 6 + 1200 c = - 0.25. 

12. (4-3) 10 b + 700 c = - 0.14. 

13. (3 - 1) 206+ 800 c =- 0.15. 

14. (3-2) 106+ 500 c =- 0.11. 

15. (2 - 1) 10 6 + 300 c = - 0.04. 
Average: II. 20 6 + 1200 c = - 0.25. 

By combining equations 6 to 15 into two series and subtracting we obtain 
the following: 

III. (7 + 8 + 10 + 12 + 14) 100 6 + 6400 c = - 1.35 
IV. (6 + 9 + 11 + 13 + 15) 100 6 + 5600 c =- 1.15 

800 c = - 0.20 
c = - 0.00025. 



176 SUGAR ANALYSIS 

Substituting the value for c in equation II we obtain b = 0.0025, and 
substituting these values for b and c in equation I we obtain a = 66.564. Sub- 
stituting these values in the original equation for the parabola we obtain: 

[a]g = 66.564 + 0.0025 p - 0.00025 p\ 

The calculated specific rotation of sucrose for various concentrations according 
to the above equation is as follows: 10 per cent 66.56, 20 per cent 66.51, 30 per 
cent 66.41, 40 per cent 66.26, 50 per cent 66.06, results which agree perfectly 
with the average observations taken. 

The above equation for the specific rotation of sucrose does not hold, 
however, for concentrations below 10 per cent or above 50 per cent. 
Tollens * from observations upon 19 solutions ranging from 3.8202 per 
cent to 69.2144 per cent sucrose calculated the following equations: 
For p = 4 to 18 per cent sucrose, 

[] = 66.810 - 0.015553 p - 0.000052462 p 2 . 
For p = 18 to 69 per cent sucrose, 

H 2 D = 66.386 + 0.015035 p - 0.0003986 p\ 

According to the above equations the maximum specific rotation 
of sucrose (66.53) is found at p = 18.86 per cent; for concentrations 
lower than this the specific rotation again decreases. 

Schmitz | from observations upon eight solutions for p = 5 to 65 
per cent gives the equation: 

[a] = 66.510 + 0.004508 p - 0.00028052 p\ 

Nasini and Villavecchia | for p = 3 to 65 give the equation 
[a] = 66.438 + 0.010312 p - 0.00035449 p 2 . The last named au- 
thorities found, however, for very dilute solutions (c = 0.335 gm. to 
1.2588 gms. sucrose per 100 c.c.) that the specific rotation of sucrose 
again increases, and for such dilute solutions give the equation 
[]g = 69.962 - 4.86958 p + 1.86415 p 2 . The variations noted in the 
above equations for the specific rotation of sucrose are no doubt partly 
due to the effect of rotation dispersion, as the result of using light of 
slightly different wave length for illumination. 

The equations of Tollens and of Nasini and Villavecchia are con- 
sidered to be the most accurate. The average of the two equations gives 
probably the most reliable expression for the specific rotation of sucrose. 
I- [a] 2 D = + 66.386 + 0.015035 p - 0.0003986 p 2 . (Tollens.) 
II. [a] = +66.438 + 0.010312 p - 0.0003545 p 2 . 

(Nasini and Villavecchia.) 
Average: III. [] = +66.412 + 0.012673 p- 0.0003766 p 2 . 

* Ber., 10, 1403. f Ber., 10, 1414. 

t Public, de lab. chim. delle gabelle. Rome, 1891, p. 47. 



SPECIFIC ROTATION OF SUGARS 



177 



Landolt * by recalculating this combined equation into terms of 
concentration (grams of sugar per 100 c.c.) gives the expression: 

IV. [] = + 66.435 + 0.00870 c - 0.000235 c 2 (c = to 65). 

The following table, which with the exception of column / is taken 
from Landolt,* gives a comparison of the specific rotation of sucrose 
for solutions of different percentage and concentration, according to 
each of the four preceding equations. 

TABLE XXVIII 
Giving Specific Rotation of Sucrose for Different Concentrations 



a 


6 


c 


d 


e 


/ 


g 




20 


Concentra- 


Specific rotation [af%. 


Percentage. 


Sp.gr.^. 


tion 
( c n j) 














(Tollens.) 


{C p.ttj. 

(Tollens.) 


By formula I 
calculated to 


By formula II 
calculated to 


By formula III 
calculated to 


By formula IV 
calculated to 


P 


d 


c 


P 


P 


P 


c 


5 


.01786 


5.0893 


+66.451 


+66.480 


+66.466 


+66.473 


10 


.03819 


10.3819 


66.496 


66.506 


66.501 


66.500 


15 


.05926 


15.8889 


66.522 


66.513 


66.517 


66.514 


20 


.08109 


21.6218 


66.527 


66.502 


66.515 


66.513 


25 


.10375 


27.5938 


66.513 


66.474 


66.493 


66.496 


30 


. 12721 


33.8163 


66.479 


66.428 


66.453 


66.460 


35 


.15153 


40.3036 


66.424 


66.365 


66.394 


66.404 


40 


.17676 


47.0704 


66.350 


66.283 


66.316 


66.324 


45 


.20288 


54.1296 


66.256 


66.184 


66.220 


66.217 


50 


.22995 


61.4975 


66.142 


66.067 


66.104 


66.081 



Concentration equations for the specific rotation of other sugars are 
given below : 



(c=3 to 34 gms. per 100 c.c.) [] =+18.095+0.06986 p. 

(c = 34 to 61 gms. per 100 c.c.) [a] =+23.089-0.1827 p+0.00312 p 2 . 



(P 



to 35 per cent) 
to 100 per cent) 

I to 30 per cent) 
) to 68 per cent) 



Xylose t 

Galactose I 
Glucose 

Fructose || 
Invert sugar 

Maltose ** 

Browne (J. Ind. Eng. Chem., 2, 526) has calculated the observations of Tollens 
to concentration and gives the equation for glucose [a]g = + 52.50 + 0.0227c + 
0.00022 c 2 . 

* " Das optische Drehungsvermogen" (1898), p. 420. t Schulze and Tollens, 
Ann., 271, 40. t Meissl, J. prakt. Chem. [2], 22, 97. Tollens, Ber., 17, 2238. 
II Ost. Ber., 24, 1636. 1[ Gubbe. Ber., 18, 2207. ** Meissl, J. prakt. Chem. [2], 
26, 114. 



(p = 5 to 35 per cent) 



[ a j2o = +79.703+0.0785 p. 
[ a ]2o _ +52.50+0.018796 p 

+0. 00051683 p 2 . 
[] = -(91. 90+0. Ill p). 
[] ==-19.447-0. 06068 p 

+0.000221 p 2 . 
[]> =+138.475-0.01837 p. 



178 SUGAR ANALYSIS 

EFFECT OF TEMPERATURE UPON SPECIFIC ROTATION OF SUGARS 

The effect of temperature upon the specific rotation of sugars is 
no less pronounced than that of concentration and, with a number of 
sugars such as fructose and galactose, the influence of temperature is 
the factor which has most to be considered in polarimetric measure- 
ments. The change in rotation of a sugar solution due to expansion 
or concentration in volume through temperature changes must not be 
confused with changes in specific rotation. In studying the latter 
phenomenon the sugar solutions must either be made up to volume at 
the same temperature at which they are to be examined or else a cor- 
rection be made for the changes in volume due to expansion or con- 
traction. 

The influence of temperature upon specific rotation is studied in 
the same way as that of concentration, by laying off the specific rota- 
tion for each temperature upon a diagram. The connecting points for 
the ordinary ranges of atmospheric temperature lie more nearly in a 
straight line than is the case with the concentration curves. For 
wider ranges of temperature, however, the increase or decrease in 
specific rotation is found to proceed unequally and the change must 
then be expressed by some curve equation. 

Effect of Temperature upon the Specific Rotation of Sucrose. 
The earlier investigators Mitscherlich, Hesse, and Tuchschmid re- 
garded the effect of temperature upon the specific rotation of sucrose 
as insignificant. Dubrunfaut* was the first to recognize the fact 
that increase of temperature caused a decrease in the value of this 
constant, the temperature coefficient of the specific rotation of su- 
crose having been found by him to be 0.000232 per 1C. increase. 
Andrews,! who reinvestigated the question in 1889, found a decrease 
of 0.0114 in the specific rotation of sucrose for 1 C. increase. The 
specific rotation of sucrose for any temperature t is then represented 
by the equation: 

MA = MS -0.01140 -20). 

SchonrockJ in 1896, as a result of observation upon 10 sugar solu- 
tions, showed that the decrease in specific rotation for 1 C. increase 
lay between 0.0132 and 0.0151 ; for temperatures between 12 C. and 
25 C. the change is expressed by the equation: 

MA = MS ~ - 0144 (* - 20). 

* Ann. chim. phys. [3], 18, 201. 

t Mass. Inst. Tech. Quarterly, May, 1889, p. 367. 

t Ber. phys.-techn. Reichsanstalt, 1896. 



SPECIFIC ROTATION OF SUGARS 179 

This equation is sometimes written 

Mb = [<*}D ~ MS 0.000217 (t - 20), 
in which the temperature coefficient of the specific rotation, 

- 0144 
66.5 



000217 -- 



Later experiments were made by Schonrock* at temperatures 
between 9 C. and 32 C. using light of three different wave lengths, 
the yellow sodium line 589.3 w, the yellow-green mercury line 546.1 "/*/*, 
and the blue mercury line 435.9 /*/*. These experiments showed that for 
the German normal sugar solution (p = 23.701 per cent) the rotation 
angle underwent a linear deviation with changes in temperature, this 
deviation being independent of the wave length of light employed. It 
was found, moreover, that the temperature coefficient of the specific 
rotation decreased with increase in temperature, the value being 0.000242 
at 10 C., 0.000184 at 20 C., and 0.000121 at 30 C. for Sodium light. 
This decrease proceeds in a straight line and the values of the tempera- 
ture coefficient for any intermediate temperature can be estimated by 
taking the proportionate difference. These later values of Schonrock 
are used by the Physikalisch-Technische Reichsanstalt of Germany 
and have therefore the highest sanction of authority. 

Effect of Temperature upon the Specific Rotation of Other Sugars. 
- The effect of temperature upon the specific rotations of a number 
of other sugars is given in Table XXIX. 

TABLE XXIX 

Rhamnosef ................ H = + 9.18-0.035 t (1= 6 to 20 C.) 

GalactoseJ (p = 10) .......... [] = + 84.67-0.200 t (< = 10 to 30 C.) 

Fructose (p=9) ............ H^ = - 103. 92+0. 671 t (/. = 13 to 40 C.) 

Fructose (p=23.5) ......... [a]^ = - 107. 65+0. 692 t (1= 9 to 45 C.) 

Invert sugar|| (c = 17.21) ..... W D = - 27.9 +0.32 t (t= 5 to 35 C.) 

Lactose^ ................... [! = + 52.53-0.07 i-20 (* = 15 to 25 C.) 

Maltose** (p=10) ........... [J^ = + 140. 19 -0.095 t (* = 15 to 35 C.) 

* Z. Ver. Deut. Zuckerind., 53, 650. 
t Schnelle and Tollens, Ann., 271, 62. 
j Meissl, J. prakt. Chem. [2], 22, 97. 

Honig and Jesser, Z. Ver. Deut. Zuckerind., 38 (1888), 1028. 
|| Tuchschmid, J. prakt. Chem. [2], 2, 235. 
i Schmoger, Ber., 13, 1922. 
** Meissl. J. prakt. Chem. [2], 25, 114. 



180 



SUGAR ANALYSIS 



While a linear equation is sufficiently exact for narrow ranges of tem- 
perature, the change in specific rotation for wider differences of temper- 
ature must usually be expressed by an equation of the order: 



or [<x}' D = [<*}% + a(t-20)+b(t- 20) 2 . 

Gernez,* for example, gives for rhamnose the equation 

\aY D = 9.22 - 0.03642 t + 0.0000123 t 2 
and Gubbef gives for invert sugar the following equations: 
For * = to 30 C., [a]^ = [ag+ 0.3041 (t - 20)+0.00165 (*-20) 2 . 
For * = 20 to 100 C., [*] = []%+ 0.3246 0-20) -0.00021 0-20) 2 . 

Sucrose and the different sugars mentioned in Table XXIX all show 
a decrease in specific rotation with increase in temperature. Of other 
sugars, which exhibit this property in marked degree, arabinose should 
be mentioned. TanretJ found for 1-arabinose [a]^ = +105.54 and 
[a]g= +88.61, or an average decrease of 0.394 for 1 C. increase in 
temperature, which is greater than that for any other sugar except 
fructose. 

Xylose presents an exception to the rule just noted, Schulze and 
Tollens having observed for temperatures above 20 C. an increase in 
specific rotation, as in the following example (p = 10.0829). 



t 


[a] D ]-xylose. 


15 


+ 18.898 


20 


18.909 


25 


19.248 


30 


19.628 



Glucose also seems to present an exception to the rule of dimin- 
ished rotation with increase in temperature. Observations by Dub- 
runfaut, Mategczek, and others show that the specific rotation of 
d-glucose undergoes no perceptible change between and 100 C. 

Equations giving the combined influence of concentration and tem- 
perature upon specific rotation have been worked out for many sugars. 
The following examples are given: 

* Compt. rend., 121, 1150. 
t Ber., 18, 2207. 
j Bull. soc. chim. [3], 15, 195. 
Ann., 271, 40. 



SPECIFIC ROTATION OF SUGARS 181 

Galactose * [] = + 83.883 + 0.0785 p - 0.209 1. (Meissl.)t 

Fructose [alp = [101.38 0.56 1 + 0. 108 (c 10)]. (Jungfleisch and 

Grimbert.)t 
Fructose [] = - 88.13 - 0.2583 p + 0.6714 (t - 20). (Honig and 

Jesser.) 
Sorbose []g = - [42.65 + 0.047 p + 0.00007 p 2 - (t- 20) 0.02]. (Tollens 

and Smith.) II 
Maltose []g = + 140.375 - 0.01837p - 0.095 1. (Meissl.) 1F 

EFFECT OF SOLVENT UPON THE SPECIFIC ROTATION OF SUGARS 

The constants of specific rotation for sugars are all expressed for 
aqueous solutions. It sometimes happens, however, that solutions of 
sugar in other solvents, such as alcohol, have to be examined; in such 
cases the changes in specific rotation due to the character of solvent 
must be taken into account. 

In the case of sucrose, Tollens** found the following values for [a]^ 
with different solvents for a 10 per cent solution: 

In water + 66.667. 

In 1 part water + 3 parts ethyl alcohol + 66.827. 
In 1 part water + 3 parts methyl alcohol + 68.628. 
In 1 part water + 3 parts acetone + 67.396. 

Methyl alcohol and acetone are thus seen to raise the specific rotation 
of sucrose perceptibly, but ethyl alcohol only slightly. Claassenft also 
found for 80 per cent alcohol a slight increase in the specific rotation 
of sucrose; the differences (0.1 to 0.15), however, are not sufficient to 
affect seriously the analytical results in such operations as the alcoholic 
extraction of sugar beet or cane pulp. 

In the case of fructose and invert sugar, ethyl alcohol produces a 
marked lowering of the specific rotation, and when these sugars are 
present the influence of ethyl alcohol as a solvent must be taken into 

* Tanret (Bull, societe" chimique [3], 16, 195) gives the change in specific 
rotation of galactose for 1 C. increase between 13 and 20 -0.39, between 20 
and 25 -0.226, and between 25 and 30 -0.180, a falling off in the temperature 
coefficient with increase in temperature similar to the one noted by Schonrock with 
sucrose. 

t J. prakt. Chem. [2], 22, 97. 
t Compt. rend., 107, 390. 
Z. Ver. Deut. Zuckerind., 38 (1888), 1028. 
II Ber., 33, 1289. 
1f J. prakt. Chem. [2], 25, 114. 
** Ber., 13, 2287. 
ft Z. Ver, Deut. Zuckerind., 40, 392. 



182 SUGAR ANALYSIS 

account. Fructose according to Landolt* has a specific rotation in 
alcohol which is only two-thirds that in water. Borntrsegert found for 
37.6 gms. invert sugar in 100 c.c. aqueous solution a rotation of 49.2 
at 20 C.; when the solution was made up with 10.45 c.c. alcohol 
the rotation decreased to 43.9 and with 20.60 c.c. alcohol to 38.3. 
According to Horsin-Deon t (whose conclusion, however, requires veri- 
fication) invert sugar in absolute alcohol is perfectly inactive and only 
becomes levorotatory upon the addition of water. It should also be 
noted that the rotation of alcoholic invert-sugar solutions is much 
more sensitive to changes in temperature than water solutions. 

With a number of sugars the specific rotations in aqueous and 
alcoholic solutions are almost the reverse of one another. The [O\D 
of rhamnose for example in water is +9-43 and in alcohol 9.0. 
The [O\D of sorbosej in water is 42.5 and in 85 per cent alcohol 
+41.8. The effect of pyridine and formic acid upon the specific rota- 
tions of several sugars is shown on page 190. 

Without giving detailed results of experiments upon all the various 
sugars it may be said that the effect of solvent upon specific rotation 
is too great to be disregarded; wherever possible the polarimetric ex- 
amination of sugars for purpose of analysis should be made in aqueous 
solution. 

EFFECT OF ACCOMPANYING SUBSTANCES UPON SPECIFIC ROTATION 

OF SUGARS 

Another factor of importance, especially in the polarimetric exami- 
nation of impure sugar solutions, is the effect which bases, acids, salts, 
and other substances exert upon the specific rotation of the sugars 
present. A very large amount of investigation has been done upon 
this subject and for complete details reference must be made to the 
original articles. Only brief mention will be made of the effects of a 
few substances upon the rotation of the more important sugars. 

The changes which foreign optically inactive substances may exert 
upon the rotation of sugars may be either chemical or physical. The 
hydroxides of the alkalies and alkaline earths, and all salts of alkaline 
reaction in general, cause a decrease in the specific rotation of most 
reducing sugars. Such changes in rotation are purely chemical, being 

* Ber., 13, 2335. 

t Z. ang. Chem. (1889), 507. 

J J. fabr. sucre, 20, 37. 

Rayman and Kruis, Bull. soc. chim. [2], 48, 632. 

II Adrian!, Rec. trav. chim. des Pays Bas., 19, 184. 



SPECIFIC ROTATION OF SUGARS 



183 



due either to a rearrangement of the sugar molecule or to the forma- 
tion of alkali-sugar compounds of lower specific rotation. The effect of 
acids and acid salts upon the rotation of sucrose by inversion is another 
example of purely chemical change. The avoidance of such chemical 
changes is imperative in accurate polarimetric work and to prevent 
these the solutions of sugar under examination should be, so far as 
possible, neutral in reaction. 

The influence of neutral salts upon the specific rotation of sugars, on 
the other hand, is largely physical, since the chemical properties of the 
dissolved sugars are not appreciably affected; the same is also true of 
the influence of acids upon the specific rotation of sugars which do not 
undergo inversion. 

Influence of Mineral Impurities upon the Rotation of Sucrose. 
The chlorides, nitrates, sulphates, phosphates, acetates, and citrates, 
of the alkalies, the chlorides of the alkaline earths, magnesium sulphate, 
and many other salts all produce a decrease in the specific rotation of 
sucrose, this decrease being generally greater with increased amount 
and smaller molecular weight of salt. 

The hydroxides of the alkalies and alkaline earths and the carbon- 
ates of the alkalies also lower the specific rotation of sucrose. The in- 
fluence of these substances, which is of especial importance technically, 
in view of the alkalinity of various sugar-house products, has been 
widely studied, the results being often expressed in parts of sugar whose 
rotation is obscured by one part of alkali. Pellet for example gives 
the following results: 



Substance. 


Concentration of sucrose solution. 


5.4 gms. 
100 c.c. 


17.3 gms. 
100 c.c. 


1 gm. caustic potash obscures rotation of 


Grams sucrose. 

0.170 
0.140 
0.044 
0.040 
0.7 
0.190 


Grams sucrose. 

0.500 
0.450 
0.065 
0.132 
1.0 
0.430 


1 gm. caustic soda obscures rotation of. 


1 gm. potassium carbonate obscures rotation of. . 
1 gm. sodium carbonate obscures rotation of 
1 gm. calcium oxide obscures rotation of 


1 gm. barium oxide obscures rotation of 








Strontium oxide also diminishes the specific rotation of sucrose. 
This lowering effect of alkalies upon the specific rotation of sucrose is 
largely due to the formation of soluble saccharates of lower specific 
rotation; the influence can be largely eliminated by neutralization with 
acetic acid. The original specific rotation is not entirely restored, 



184 



SUGAR ANALYSIS 



however, since the soluble acetates themselves lower the specific rota- 
tion of sucrose to a slight extent. 

The probable effect of a mixture of salts upon the polarization of 
sucrose, such for example as occurs in beet molasses, which contains 
about 50 per cent of sucrose and 10 per cent of soluble salts (mostly of 
potassium), may be judged from the following examples taken from 
experiments by Bodenbender and Steffens.* 

TABLE XXX 



Salt. 


Sucrose, 
parts. 


Salt, parts. 


Water, 
parts. 


Polarization, 
sugar degrees. 


Difference. 


Potassium chloride . . . . < 
Sodium chloride < 
Barium chloride \ 


5 
10 
20 
5 
10 
20 
5 
10 


1 

2 
4 

1 
2 
4 
1 
2 


94 
88 
76 
94 
88 
76 
94 
88 


4.987 
9.856 
19.869 
4.969 
9.853 
19.586 
4.952 
9.944 


0.013 
0.144 
0.131 
0,031 
0.147 
0.414 
0.048 
0.056 


Magnesium sulphate.. . < 
Sodium phosphate . < 


20 
5 
10 
20 
5 
10 


4 

1 
2 
4 
1 
2 


76 
94 

88 
76 
94 
88 


19.402 
4.995 
9.890 
19.880 
4.958 
9 933 


0.598 
0.005 
0.110 
0.120 
0.042 
0.067 


Potassium carbonate. . . < 
Sodium carbonate < 


20 
5 
10 
20 
5 
10 
20 


4 
1 
2 
4 
1 
2 
4 


76 

94 
88 
76 
94 
88 
76 


19.689 
4.927 
9.730 
19.300 
4.910 
9.711 
19.173 


0.311 
0.073 
0.270 
0.700 
0.090 
0.289 
0.827 



The effect of four-fold concentration is seen to depress the difference 
in rotation about ten times, so that an apparent loss of sucrose may 
seem to take place in the evaporation of sugar solutions rich in mineral 
salts, when such solutions are examined by the polariscope before and 
after evaporation. 

The effect which the various salts, used for clarifying impure sugar 
solutions for optical analysis, exercise upon the specific rotation of 
sucrose and other sugars is also of great importance. Lead subacetate 
is the salt most used for this purpose; its effect upon the rotation of 
sucrose is considered elsewhere (page 216). 

Influence of Mineral Impurities upon the Rotation of Reducing 
Sugars. The action of salts of alkaline reaction in depressing the 
rotation of reducing sugars has already been mentioned. In sacchari- 

* Z. Ver. Deut. Zuckerind., 31, 808. 



SPECIFIC ROTATION OF SUGARS 



185 



metric analysis the influence of lead subacetate, as a clarifying agent, 
upon the rotations of fructose and invert sugar, is f great importance. 
As was first observed by Gill* in 1871 when solutions containing invert 
sugar are treated with lead-subacetate solution in excess, the formation 
of soluble lead fructosate of low specific rotation is so pronounced that 
the rotatory power of fructose sinks below that of glucose and the 
invert sugar becomes dextrorotatory. Similar observations have been 
made by Pellet, Bittmann, Koydl, Borntraeger, and many others. In 
the following experiments by Bittmann f 50 c.c. of invert-sugar solution 
were treated with 50 c.c. of a mixture of water and lead subacetate in 
different proportions. 






Water. 


Lead-subacetate 
solution. 


Polarization. 


c.c. 


c.c. 




50 





-2.3 


40 


10 


-1.0 


30 


20 


+3.7 


10 


40 


+7.5 



The influence of neutral salts upon the specific rotation of reducing 
sugars is variable. Some salts produce an increase, others a decrease 
and some no change whatever in rotation; no general rule can be given. 

Of particular importance in this connection is the influence of differ- 
ent neutral salts upon the rotation of invert sugar; the occurrence of 
such salts in molasses and other low-grade sugar-house products may 
increase the levorotation of the invert sugar several degrees, with the 
result that erroneous conclusions are sometimes drawn from the polari- 
scopic examination of such products. 

Influence of Acids upon the Specific Rotation of Sugars. The 
presence of free mineral acids exerts a very pronounced influence upon 
the specific rotation of certain sugars. This influence is very slight in 
case of glucose, but is most pronounced with fructose and hence also 
with invert sugar. O 'Sullivan, for example, found for invert sugar, 
prepared by inverting sucrose with invertase, [a]J$ = 24.5, and for 
invert sugar, prepared by inverting sucrose with sulphuric acid in the 
cold, [] = 27.7, an increase of 3.2, which if referred entirely to 
fructose would mean an increase of 6.4 in the specific rotation of that 
sugar. The increase in rotation increases with the amount of acid, as 
is seen from the following results by Hammerschmidt t which the author 

* Z. Ver. Deut. Zuckerind., 21, 257. 

t Ibid., 30, 875. 

} Ibid., 40, 465; 41, 157. 



186 



SUGAR ANALYSIS 



has calculated to the [a]g of invert sugar and fructose. The results 
Were obtained by inverting a half-normal weight of sucrose with vary- 
ing amounts of concentrated hydrochloric acid and then completing 

the volume to 100 c.c. 

TABLE XXXI 

Showing Influence of Varying Quantities of Hydrochloric Acid upon the Rotation 
of Invert Sugar and Fructose. 



Volume of 
HC1 added. 


Observed saccha- 
rimeter reading, 
20 C. 
(13.6842 gms. invert 
sugar to 100 c.c.) 


Calculated []g 


Invert sugar. 


Fructose. 


c.c. 

5 
10 
15 
20 


V. 

"-IG'SO" 

-17.06 
-17.58 
-18.02 


-20.00 
-20.89 
-21.60 
-22.26 

-22.82 


-92.50 
-94.28 
-95.70 
-97.02 
-98.14 



The influence of the change in specific rotation of fructose upon the 
determination of sucrose by the methods of acid inversion is discussed 
on page 270. The action of organic acids upon the rotation of fructose 
and invert sugar is much less pronounced than that of mineral acids, 
and can usually be disregarded in polariscopic analysis. 

Influence of Foreign Optically Active Substances upon the Specific 
Rotation of Sugars. The effect of other optically active ingredients 
upon the rotation of a sugar is of importance especially in determining 
the polarizing power of several sugars in solution or of mixtures of 
sugars with organic non-sugars which are optically active. The difficul- 
ties in conducting studies of this kind seem to have deterred investigation 
somewhat; the results upon the polarizing power of sugar mixtures, 
so far as they have been carried, show, however, no change in the 
rotation of the individual sugars. 

The polarizing power, for example, of solutions of sucrose and glucose 
in different proportions was found by Hammerschmidt * to agree with 
the sum of the values calculated by the concentration formulae of 
Tollens (page 177) within experimental limits of error. Similar results 
were also obtained by Creydtf in case of cane sugar and raffinose. 
Results by Brownet upon the polarization of mixtures of glucose and 
fructose, glucose and galactose, fructose and galactose, fructose and 

" Das specifishe Drehungsvermogen von Gemengen optisch activer Substan- 
zen," Dissertation, Rostock University, 1889. 
t Z. Ver. Deut. Zuckerind. (1887), 37, 153. 
J J. Am. Chem. Soc., 28 (1906), 339. 



SPECIFIC ROTATION OF SUGARS 



187 



arabinose, arabinose and xylose also show that it is safe to assume in 
analytical work that the specific rotation of these sugars is not per- 
ceptibly affected by other sugars in solution. 

MUTAROTATION 

A phenomenon observed in the polarization of all optically active 
reducing sugars is that of mutarotation (also called birotation or multi- 
rotation). The polarizing power of such sugars undergoes after solu- 
tion at first a rapid change which slowly becomes more gradual until 
after a few hours the polariscope reading remains constant. This phe- 
nomenon was first observed upon glucose in 1846, by Dubrunfaut * and 
the fact that the initial rotation of this sugar was about twice the con- 
stant value caused the introduction of the name birotation. The re- 
lation 2 : 1 was found, however, to be different in the case of other 
sugars; Wheeler and Tollens,f for example, found the ratio in case of 
xylose to be about 4.5:1 and accordingly suggested the name multiro- 
tation. This term, however, in recent years has given place to the 
more expressive word mutarotation (Latin mutare = to change) intro- 
duced by Lowry 1 in 1899. 

The effect of mutarotation upon the rotatory power of sugars is 
shown in the following table, in which results are quoted from the work 
of Tollens and his coworkers, giving the specific rotation of a number 
of sugars directly after solution and after standing until no further 
change was noted. The time after solution is given after each value 
for []. 

TABLE XXXII 
Showing Mutarotation of Different Sugars 



Sugar. 


Grams 
per 
100 c.c. 


[a]5> initial. 


[a]JJ constant. 


Difference. 


Velocity 
constant 
(Osaka). 








min. 




hours 






1- Arabinose 


9.73 


+ 156.7 


6.5 


+ 104.6 


1.5 


-52.1 


0.031 


1-Xylose 


10.235 


+ 85.9 


5. 


+ 18.6 


2.0 


-67.3 


0.022 


d-Glucose 


9.097 


+ 105.2 


5.5 


+ 52.5 


4.5 


-52.7 


0.0104 


d-Galactose.. . . 


10.000 


+117.4 


7. 


+ 80.3 


4.5 


-37.1 


0.0102 


d-Fructose 


10.000 


-104.0 


6. 


- 92.3 


0.5 


-11.7 


0.096 


Rhamnose 


10.000 


- 5.0 


5.5 


+ 9.4 


1.0 


+14.4 


0.039 


Fucose 


6.916 


-111.8 


11. 


- 77.0 


2.0 


-34.8 


0.022 


Lactose 


4.841 


+ 87.3 


8. 


+ 55.3 


10.0 


-32.0 


0.0046 


Maltose 


9.2 


+ 118.8 


6. 


+136.8 


6.5 


+17.0 


0.0072 



* Compt. rend., 23, 38. 

t Ann., 254, 312. 

t J. Chem., Soc., 75, 212. 



188 SUGAR ANALYSIS 

It is noted that in case of rhamnose there is a decrease in rotation 
from 5.0 to and then an increase from to + 9.4. Maltose also 
differs from the other sugars in showing a less rotation at time of solu- 
tion than after standing. 

Effect of Temperature on Mutarotation. The speed of muta- 
rotation is influenced by a large number of factors. It is accelerated 
by increase in temperature, the change proceeding very slowly at 
C., and almost instantly at 100 C. Dilute sugar solutions show the 
same velocity of change for all concentrations. Highly concentrated 
solutions, however, do not always give the true end rotation; such 
solutions must first be diluted and then allowed to stand for the change 
in rotation to be completed. This fact must be borne in mind in the 
polariscopic examination of concentrated sugar solutions, such, for ex- 
ample, as liquid honey, otherwise a considerable error may be intro- 
duced in the work of analysis. 

Velocity of Mutarotation. The velocity of the change from 
initial to constant rotation is different for different sugars, and also 
varies according to temperature, solvent, and other conditions. Urech * 
was the first to show that the speed of mutarotation followed the same 
law as that noted by Wilhelmy in the inversion of sucrose (page 660), 
and which is expressed by the following general formula for a reaction 
of the first order, 

- = k (a - x), 

in which k is the coefficient of velocity, a the total change between the 
beginning and end point, and x the change at the end of any time t. 
The above equation by integration gives 

1, a 

k = - log 
t & a x 

Owing to the impossibility of measuring the specific rotation of a 
sugar at the exact moment of solution, the velocity of mutarotation is 
generally determined by the modified formula 



in which ft and ft are the rotations at the end of the corresponding 
times h and ^, and <f> the constant end rotation. 

The method of calculation is shown by the following example, 
which is taken from the work of Levy,f 

* Ber., 16, 2270; 17, 1547; 18, 3059. 
t Z. physik. Chem., 17, 301. 



SPECIFIC ROTATION OF SUGARS 



189 



TABLE XXXIII 
Showing Velocity of Mutarotation for a Glucose Solution 

Per cent, C 6 H 12 O 6 = 3.502. d = 1.0114. Temperature.= 20.5 to 20.9 C. 



Time after solution. 


Angular rotation 
(8 dm. tube). 


<I-1 


Temperature. 


*=rV log >'(!H) 

Tj fj \P </ 


i = 25 min. 


0! = 27.865 





20.9 C. 




J 2 = 30 min. 


2 =27.060 


5 


20.9 


0.00649 


< 2 = 35 min. 


02 = 26.159 


10 


20.9 


0.00719 


Z 2 = 40 min. 


02 = 25.637 


15 


20.8 


0.00644 


2 = 45 min. 


2 =24.927 


20 


20.7 


0.00662 


* 2 = 50 min. 


02=24.369 


25 


20.6 


0.00652 


2 = 55 min. 


02 = 23.895 


30 


20.5 


0.00636 


< 2 = 60min. 


02 = 23.166 


35 


20.5 


0.00677 


<2 = 65 min. 


02 = 22.797 


40 


20.5 


0.00656 


* 2 = 70 min. 


02=22.171 


45 


20.5 


0.00687 


2 = 75 min. 


2 = 21.837 


50 


20.5 


0.00674 


< 2 = 80 min. 


2 = 21.470 


55 


20.5 


0.00671 


< 2 = 85 min. 


2 =21.088 


60 


20.5 


0.00675 


24 hours 


= 16.692 




Average 


0.00662 



The velocity constants by Osaka * given in Table XXXII were cal- 
culated by this method. It is seen that the change to constant rota- 
tion is most rapid for fructose and slowest for lactose. 

Effect of Acids, Bases, and Salts on Mutarotation. The action of 
acids, bases, and salts upon the velocity of mutarotation has been a 
subject of considerable study. Acids accelerate the change according 
to their degree of dissociation, or electric conductivity, preserving ap- 
proximately the same order as that noted in the inversion of sucrose. 
Levy,t for example, gives the following constants for the speed of 
mutarotation of glucose in presence of different acids dV normal) and 
the relative acceleration of each acid in terms of hydrochloric acid = 100. 

TABLE XXXIV 

Showing Acceleration of Different Adds upon Mutarotation 



In presence of. 


Velocity con- 
stant of muta- 
rotation. 


Temperature. 


Relative 
acceleration. 


Water.. 


0.00610 


20.1C. 




Water 


0.00637 


20.25 












Hydrochloric acid 


0.02300 


20.25 


100.00 


Nitric acid. ... . ... 


0.02283 


20.1 


98.99 


Trichloracetic acid. .... 
Sulphuric acid 


0.02325 
0.01886 


20.25 
20.0 


96.67 
71.95 


Dichloracetic acid 
Monochloracetic acid . . 
Acetic acid. . . ... 


0.01670 
0.01004 
0.00716 


20.2 
20.25 
20.2 


62.41 
17.25 
4.70 


Propionic acid 


0.00636 


19.8 


1.63 



Z. physik. Chem., 35, 661. 



t Z. physik. Chem., 17, 301. 



190 



SUGAR ANALYSIS 



The values for relative acceleration of the different acids preserve 
the same order as those noted for the inversion constants in Table XCV 
(page 663). 

It is scarcely necessary to state that the speed of mutarotation 
increases with the strength of acid employed. Thus Levy found for 
ft/10 hydrochloric acid, k = 0.02300 and for n/50 hydrochloric acid, 
k = 0.00971; for n/10 acetic acid, k = 0.00716 and for n/50 acetic 
acid, k = 0.00654. 

Alkalies also accelerate the speed of mutarotation, the change to 
constant rotation being almost instantaneous. Schulze and Tollens* 
using 0.1 per cent ammonia obtained the normal constant rotation 
with arabinose, xylose, rhamnose, galactose, glucose, fructose, and lac- 
tose within 9 minutes; n/200 alkali (KOH) gives the end rotation of 
glucose almost instantly. The use of much stronger alkali, however, 
induces chemical change with a decrease of the rotation below the normal 
value. Treyf for example using 0.2 gm. sodium hydroxide per 100 c.c. 
obtained as the [O\D for glucose after 15 minutes + 52.7 (normal), after 
24 hours + 36.7, after 48 hours + 26.0, after 34 days + 15.1, and after 
65 days - 0.4. 

The different salts nearly all accelerate the speed of mutarotation, 
those of alkaline reaction standing first in this respect. Sodium chlor- 
ide, however, presents an exception to this rule, having been found 
by Levy I and also by Trey to cause the mutarotation of glucose to 
proceed slower than in pure aqueous solution. 

Mutarotation of sugars takes place not only in water but also in 
o^her solvents such as absolute methyl alcohol, ethyl alcohol, acetone, 
etc. The change in rotation proceeds much more slowly, however, in 
organic solvents than in aqueous solution. This is shown in the follow- 
ing results by Grossmann and Bloch || which give the mutarotation of 
several sugars in pyridine and formic acid. 



Sugar. 


fa]jT) in pyridine. 


[O\D in formic acid. 




After solution. 


Constant. 


After solution. 


Constant. 


Xylose 


+ 117.39 


Min. 

8 


+ 40.63 


Days. 

4 


+ 40.34 


Min. 
4 


+ 66.60 


Days. 


Rhamnose .... 
Galactose. . . . 
Glucose 
Fructose. 


- 41.39 

+ 154.28 
+ 149.60 
-174.13 
+ 103.48 


5 
23 
10 
10 
15 


- 32.77 
+ 59.83 
+ 74.79 
- 34.83 
+ 123.80 


4 
3 
4 
1 
11 


+ 10.20 
+ 89.11 
+ 72.16 
- 94.32 
+ 129.11 


5 
5 
5 
5 

10 


- 35.76 
+ 127.35 
+ 122.51 
-47.83 
+ 172.15 


6 
5 
4 
8 
3 


Maltose 



' Ann., 271, 49. f Z. physik. Chem., 22, 439. J Ibid., 17, 320. 

Z. physik. Chem., 22, 424. || Z. Ver. Deut. Zuckerind., 62, 19. 



SPECIFIC ROTATION OF SUGARS 191 

A peculiarity of xylose and rhamnose in pyridine is an increase in 
the rotation after solution. Grossmann and Bloch observed a maximum 
of + 122.07 in case of xylose 15 minutes after solution and a maximum 
of 45.92 in case of rhamnose 30 minutes after solution. It is seen 
that mutarotation in the two solvents proceeds in many cases in opposite 
directions and that there is no relation between the constant rotations 
and those observed in aqueous solution. The addition of water to 
solutions of sugar in organic solvents accelerates, and conversely the 
addition of alcohol, acetone, etc., to aqueous solutions retards, the speed 
of mutarotation. As a general rule the presence of any soluble non- 
electrolyte, such, for example, as sucrose, will increase the time necessary 
for a mutarotating sugar to reach constant polarization. 

Mutarotation takes place not only after dissolving reducing sugars, 
but also occurs upon the liberation of these sugars from higher saccharides 
by the action of enzymes. The phenomenon is one which the sugar 
chemist has always to bear in mind. Polariscopic measurements are 
always referred to the normal constant rotation. The latter condition 
may be produced almost instantly by heating the solution or by adding 
a little free alkali, but when such means are employed care must be 
taken to prevent the liability of chemical change. The safest course 
is to allow the solution to stand until the rotation has come to equili- 
brium in the natural way. 

Theories of Mutarotation. Many theories have been proposed 
to explain mutarotation. According to the views of Landolt* and other 
authorities it was thought that the phenomenon might be due to the 
formation of molecular aggregates immediately after solution, which 
afterwards decompose into simple molecules of lower rotation. These 
earlier theories were largely disproved, however, by the experiments of 
Arrhenius,t and of Brown and Morris,! who showed that no change 
occurred in the molecular weight of a sugar during mutarotation. 
Tollens and others of his school have supposed that mutarotation 
might be caused by the formation of unstable hydrates which, by the 
splitting off of water, cause a change in rotation. 

Much additional light was thrown upon the subject in 1895 by 
Tanret, || who discovered that sugars could exist in both a high- and a 
low-mutarotating form. The relationship of these several modifica- 
tions, according to Tanret's classification, is shown for four different 
sugars in the following table. 

* "Das optische Drehungsvermogen " (1879), 58. 

t Z. physik. Chem., 2, 500. t Chem. News, 67, 196. 

Ber., 26, 1799. Compt. rend., 120, 1060. 



192 



SUGAR ANALYSIS 



Sugar. 


a 
Metastable. 




Stable. 


y 
Metastable. 


d-Glucose 


+105 


+52.5 


+22.5 


d-Galactose 


+135 


+81 


+52 


Lactose 


+ 88 


+55 


+36 


Rhamnose 


- 6 


+ 9 


+23 











Tanret's a. modification represents the ordinary sugar as obtained 
by crystallization from aqueous solution. The /3 modification, or form 
of constant rotation, was usually obtained by precipitating a saturated 
aqueous solution of the a sugar with several volumes of absolute alcohol. 
The 7 modification was usually prepared by evaporating a concentrated 
solution of the a sugar to dryness and then heating for several hours to 
about 100 C. Repeating the process several times increases the purity 
of the various modifications. In the case of rhamnose the a modifica- 
tion is the lower, and the y modification the higher rotating form. 

Previous to Tanret's work, Lippmann* had expressed the view that 
mutarotation might be due to a stereochemical change between two 
forms of the same sugar, and showed, how by adopting a form of struc- 
ture first proposed by Tollens, that one of the terminal carbon atoms 
of the sugar molecule became asymmetric (i.e. connected to four dis- 
similar atoms or groups), thus permitting the existence of two con- 
figurations for the same sugar. The theory of mutarotation most 
generally accepted at the present time assigns one of these configura- 
tions to the high-rotating, and the other configuration to the low-rotat- 
ing form. The mutarotation reaction according to Lowryf is thus 
regarded as a balanced reaction between two molecular forms of the 
same sugar, as for example: 




CH 2 OH 
HOCH 

HCOH 
HOCH 




CH 2 OH 
HOCH 

C^H 

CH > 

HCOH < 

HOCH 

f^C OH 

a glucose 
MD = + 105 

Which of the above configurations belongs to the a or sugar has 
not been determined. 

* Ber., 29, 203. f J. Chem. Soc., 76, 212. 




glucose 
=+22.5. 



SPECIFIC ROTATION OF SUGARS 193 

Lowry's view was supported by Hudson,* who showed by quan- 
titative experiments that the change between the high- and low-rotating 
forms of lactose was a balanced reaction. According to this view, 
Tanret's solid (3 sugars of constant rotation are simply equilibrated 
mixtures of the high- and low-rotating forms. The designation /3 is 
applied at present to Tanret's y modification. 

While mutarotation is most generally regarded at present as a 
balanced reaction between high- and low-rotating forms, the intermediate 
steps of the process have not been definitely established. The change 
in polarization of a sugar solution to constant rotation is regarded by 
some chemists as simply a conversion of the a or /3 oxygen ring com- 
pound into the ordinary aldehyde or ketone form. Other chemists 
regard the solution at constant rotation as containing simply a mixture 
of the a and /3 sugars in equilibrium, while still others believe it to 
contain the a and # sugars with variable amounts of the aldehyde or 
ketone form. For a review of the different hypotheses, which have 
been proposed in this connection, the chemist is referred to the various 
special works, f 

* Z. physik. Chem., 44, 487. See also page 711. 

f Lippmann, "Chemie der Zuckerarten" (1904), 293. 

Hudson (J. Am. Chem. Soc., 32, 889) in a paper entitled "A Review of Discov- 
eries on the Mutarotation of Sugars," gives a very complete review and bibliography 
of the subject. 



CHAPTER IX 



METHODS OF SIMPLE POLARIZATION 

DETERMINATION OF SUGARS FROM ANGULAR ROTATION 

THE amount of a single optically active sugar, in presence of opti- 
cally inactive substances or in presence of substances without effect 
upon its specific rotation, may be calculated by means of either formula 
for specific rotation (page 172). 

100 a , 100 a 



, 

whence 



100 a 



100 a 



lXdX[a] D 






As to which of the above methods of calculation is to be used, the 
first or concentration formula is the better where a definite weight of 
substance is made up to volume before polarization, the usual method 
of procedure; in case, however, a sugar solution of known specific 
gravity is polarized directly, then the second or percentage formula 
is to be employed. 

The following formulae are given for calculating the concentration 
(grams per 100 c.c.) of different sugars from the angular rotation (a) 
in a 2-dm. tube. 



Arabinose c = 



x 



= - 4785 a - 



3. Glucose c = 



4. Fructose c = 



5. Galactose c = 



10 



X 



& X -f- ol.U 

6. Sucrose c = ~ = 0.7519 a. 



= 0.9470 a. 

= 0.5405 a (left degrees). 

= 0.6173 a. 



194 



METHODS OF SIMPLE POLARIZATION 195 

8. Lactose c = =0.9524 a. 



9. Raffinose (+ 5 H 2 0) c = 2 x 5 = 0.4785 a. 

10. Raffinose (anhydride) c = J* = 0.4060 a. 

z x T~ iZo.io 

The percentage p of a sugar in solution is equal to the value of c, 
as expressed above, divided by the specific gravity of the solution. 

Such formulae, as the above, are sufficiently accurate for most pur- 
poses of analysis. In cases, however, where the specific rotation of the 
sugar is affected by changes in concentration or temperature, the results 
as obtained above can be considered only approximate; to obtain the 
correct concentration or percentage, it is necessary to calculate the 
specific rotation corresponding to the approximate value of c or p at 
the temperature of polarization and substitute this corrected specific 
rotation in formulae (1) or (2) for the final calculation of c or p. 

Example. 50 gms. of a dextrose sirup were dissolved to 100 cc.; the 
constant rotation of the solution thus obtained was -f 34.55 circular degrees 
in the 200-mm. tube. Required the percentage of dextrose in the sirup. 

From formula 3 we obtain by substitution c = 0.9470 X 34.55 = 32.72 gms. 
dextrose in the 100 cc. of solution or for the 50 gms. of sirup, 65.44 per cent 
approximately. The specific rotation of dextrose for c = 32.72 is found from 
the formula []g = + 52.50 + 0.0227 c + 0.00022 c 2 (p. 177) to be +53.48; 
substituting this in the general formula for c we obtain 

< = 5 =32.30 gm , 



in the 100 cc. of solution or for the 50 gms. of sirup the true percentage 64.60, 
0.84 per cent less than the value by the uncorrected formula. 

By modifying the formula for c, so as to correct for the variations 
in specific rotation, the labor of the second calculation in the above ex- 
ample may be eliminated. In the case of glucose, by calculating the 
angular rotation, (a) for the 2-dm. tube, corresponding to concentra- 
tions ranging from 10 to 60, we obtain, using the method of least 
squares (p. 165), the formula c* = 0.958 a - 0.00067 a 2 . 

Example. Applying the last formula to the previous example, we obtain 
for c, 32.299 gms. dextrose in the 100 cc. of solution or for the 50 gms. sirup 
64.60 per cent. 

* For p Landolt gives the formula p = 0.948 a - 0.0032 a 2 . (" Optisches Dre- 
hungsvermogen," p. 447.) 



196 SUGAR ANALYSIS 

DETERMINATION OF SUGARS FROM SACCHARIMETER READINGS 

Conversion of Saccharimeter Readings into Angular Rotation. 

The general methods of optical analysis just described are more es- 
pecially applicable to polarimeters, where readings are taken in angular 
degrees; the formulae given are equally applicable, however, to saccharim- 
eters in which case the scale reading of the latter must be converted 
into angular degrees by means of the proper conversion factor. For 
general purposes the factor established for sucrose may be applied to 
other sugars. In the case of the Ventzke scale, sugar degrees 
X 0.34657 = angular rotation. Since, however, the rotation disper- 
sion of the various sugars, with reference to the quartz compensation of 
the saccharimeter, may differ somewhat from that of sucrose, it is 
always better, where exact data are available (which is unfortunately 
not always the case), to use the conversion factor established for the 
particular sugar. In the case of a few sugars Landolt * has established 
the following factors for converting divisions of the Ventzke scale into 
circular degrees. 

Sucrose 0.3465 

Lactose . 3452 

Glucose 0.3448 

Invert sugar . 3432 

Raffinose 0.3450 

Brown, Morris, and Millar f give the following: 

Sucrose, 10 per cent solution 0.3469 

Maltose, 10 per cent solution . 3449 

Maltose, 5 per cent solution . 3457 

Glucose, 10 per cent solution ' 0.3442 

Glucose, 5 per cent solution . 3454 

Starch products, 10 per cent solution 0.3458 

Starch products, 5 per cent solution 0.3454 

Herzfeld,J with a solution containing 11.29 per cent anhydrous 
maltose, obtained upon a Peters saccharimeter, using a Welsbach light 
with chromate filter, a reading of 93.88 Ventzke degrees at 20 C., and 
with the same solution upon a Lippich polarimeter a reading of 32.60 
circular degrees at 20 C. The value of a Ventzke-scale division for 

maltose under these conditions is therefore H^ = 0.3471 circular 

9o.oo 

degree, a figure perceptibly greater than the values of Brown, Morris, 
and Millar. Differences in concentration of the sugar solutions ex- 
amined but more especially differences in the optical center of gravity 
of the light employed for illuminating the saccharimeter are the chief 
* Ber., 21, 194. f J- Chem. Soc. Trans., 71, 92. J Ber., 28, 441. 



METHODS OF SIMPLE POLARIZATION 



197 



causes of such discrepancies. The chemist should, therefore, employ 
any prescribed conversion factor with caution and use it only under the 
same conditions for which it was established. It is also well to verify 
a conversion factor wherever possible, by comparative readings of the 
same sugar solution upon a polarimeter. The latter instrument does 
away with the errors of rotation dispersion and, aside from the objection 
of using monochromatic light, is always to be preferred in methods 
where the concentration or percentage of sugar is calculated from the 
angular rotation. If a quartz-wedge saccharimeter is the only instru- 
ment available, the average factor 0.346 may be used for most pur- 
poses without serious error. 

Normal Weights of Sugars. If a normal weight of each particular 
sugar be taken for polarization, (i.e. the weight of pure sugar which 
dissolved to 100 c.c. will give a scale reading of 100), the percentage 
(uncorrected) of sugar may be read directly upon the saccharimeter. 

There are a number of methods of calculating the normal weight 
for different sugars. If we assume in case of the Ventzke scale that 
the angular rotation of each division is 0.34657 circular degree for all 
sugars, then the normal weight (20 C., 100 true c.c.) of any sugar, for 
the 2-dm. observation tube, as compared with 26.00 gms., will be in- 
versely proportional to the specific rotations of this sugar and of sucrose, 

that is: 

1729 
[a]g: 66.5 : : 26 gms. : X, whence X (the normal weight) = f-W' 

\. a \D 

The normal weights of several sugars calculated by this method are given 

in the following table: 

TABLE XXXV 
Giving Normal Weights of Different Sugars for Ventzke Scale 



Sugar. 


Specific rotation [a]%- 


Normal weight. 


Glucose > - 


+53.46 c=32.5gms. 
-93.00 c=18.5gms. 
-20.00 c= 10.0 gms. 
+52.53 
+138.25 c= 12. 5 gms. 
+104.5 
+123.17 


^ 6 =32.342 gms. 
^ = 18. 592 gms. 
l -86. 450 gins. 


Fructose . 




I,aptr>9f (4-TToO^ 


1729 -32. 914 cms. 


IVIaltose 


52.53 
I- 12. 506 gms. 


TJnflRnrxsA (4-^ TTO^ 


138.25 

179Q 

1/zy -IQ 545 gms 




104.5 
1729 _j. _ 




123.17 B 



198 SUGAR ANALYSIS 

While the normal weights calculated in this manner are sufficiently 
exact for most purposes of analysis they must not be regarded as 
absolute. Owing to the differences, previously mentioned, in rota- 
tion dispersion for the different sugars the angular rotation of each 
Ventzke-scale division will vary slightly from 0.34657 circular degree 
with a corresponding change in the value of the normal weight. 

If the value of the 100-degree saccharimetric reading of each sugar 
has been established in circular degrees, for the same conditions under 
which analyses are made, it is always better to base the calculation of 
the normal weight upon this. The method of calculation for the 
Ventzke scale, using as illustrations four of the sugars previously taken, 
is as follows: 

From the general formula c = , f-r we obtain for 

IX [a\D 

Glucose 

(1 V. = 0.3448 circular degree, Landolt), c = l ***' 4 * = 32.248 gms. 

Z X oo.4o 

Lactose 

(1 V. = 0.3452 circular degree, Landolt) , c = 1 *J?* 2 = 32.857 gms. 

Z X 

Maltose 



l- V. . 0.3449 e ircular degrees, c - , 12 , 74 gms . 



Raffinose + 5 H 2 O 

(1 V. = 0.3450 circular degree, Landolt) , c = 1 ' = 16.507 gms. 

& x\ 104.0 

The conversion factors to be employed, and hence the values of the 
normal weights, will necessarily depend upon the quality of the light 
used for illuminating the saccharimeter. The value of a saccharimeter 
division in circular degrees for a solution of the sugar of the approximate 
concentration, should, therefore, be established by the chemist himself 
wherever possible. 

Correction for Concentration and Temperature. When normal 
weights of the different sugars are used, the observed saccharimeter 
readings require correction for changes in concentration and tempera- 
ture as described on page 195. Where much work is done with a single 
sugar a table of corrections should be prepared, giving the actual sugar 
value corresponding to each scale division of the saccharimeter. The 
correction table for sucrose (page 118) or the following results calcu- 
lated by Browne* for glucose upon the basis of the normal weight of 
32.25 gms. will illustrate the method. 

* J. Ind. Eng. Chem., 2, 526. 



METHODS OF SIMPLE POLARIZATION 



199 



Scale division. 


Concentration. 
Grams glucose 
100 true c.c. 
20 C. 


Specific rota- 
tion, glucose 

[]. 


Actual glucose 
value of scale 
division. 


Correction 
to be added. 


100 V. 


32.250 


53.46 


100.00 


0.00 


90 


29.025 


53.34 


90.20 


0.20 


80 


25.800 


53.23 


80.35 


0.35 


70 


22.575 


53.12 


70.45 


0.45 


60 


19.350 


53.02 


60.50 


0.50 


50 


16.125 


52.92 


50.51 


0.51 


40 


12.900 


52.83 


40.48 


0.48 


30 


9.675 


52.74 


30.41 


0.41 


20 


6.450 


52.66 


20.30 


0.30 


10 


3.225 


52.58 


10.17 


0.17 


1 


0.323 


52.51 


1.02 


0.02 



The correction necessary to be added to any reading (s) of the 
saccharimeter scale, as formulated from the above table, is equal very 
closely to + 0.02 s - 0.0002 s 2 . The percentage of glucose (G) corre- 
sponding to any scale reading (s) of the saccharimeter is, therefore, 
expressed by the formula 

G = s + 0.02 s - 0.0002 s 2 . 

Some authorities have established the normal weights of sugars for 
5, 10, 15, 20, and 25 per cent solutions. Landolt* gives as the normal 
weight of glucose for a 5 per cent solution 32.91 gms., for a 15 per cent 
solution 32.75 gms., and for a 25 per cent solution 32.50 gms., in which 
connection he states that, in weighing out the glucose-containing ma- 
terial for polarization, the chemist must select his normal weight ac- 
cording to the amount of glucose present. This, of course, involves a 
preliminary assay of the material under examination, which means prac- 
tically doubling the work of analysis. A variable normal weight is, 
moreover, confusing, and a source of error. Wherever possible one 
fixed value should be given to the normal weight, the value to be selected 
(as in the case of sucrose) being that weight of chemically pure sugar, 
which dissolved to 100 true c.c. and polarized at 20 C. in a 200-mm. 
tube will give a constant reading of exactly 100 upon the saccharimeter. 
If in the use of such a normal weight with impure products, readings of 
less than 100 are obtained, the latter are corrected by a table or formula 
similar to that just given for glucose. 

Conversion of Saccharimeter Readings into Weight of Sugars. 
It is often desirable to express the equivalent of a saccharimeter read- 
ing, for a 200-mm. tube, in grams of a particular sugar in 100 c.c. This 
equivalent can be found by multiplying the values of the formulae 

' * Landolt, "Das optische Drehungsvermogen " (1898), p. 448. 



200 SUGAR ANALYSIS 

on page 194 by the angular rotation of 1 degree of the saccharimeter 
scale (page 145), thus: 

1 angular rotation D = 0.4785 gm. arabinose. 

1 Ventzke sugar scale = 0.4785 X 0.34657 = 0.1658 gm. arabinose. 
1 French sugar scale = 0.4785 X 0.21666 = 0. 1037 gm. arabinose. 
1 Wild sugar scale = . 4785 X . 13284 = . 0635 gm. arabinose. 

Owing to the lack of absolute agreement in the value of each sac- 
charimeter scale in circular degrees, due to rotation dispersion, varia- 
tion in quality of light, etc., the equivalent of 1 degree of a saccharimeter 
scale is best expressed as T ^ of the weight of sugar, which will give a 
reading of 100 degrees under the prescribed conditions of analysis (i.e. T ^ 
of its normal weight). The correction for concentration is afterwards 
applied as indicated above. 

The approximate value of 1 V. for the more common sugars is 

given below. 

Weight of Sugar in 100 metric ex. 

1 V. at 20 C. = 0.2600 gm. sucrose. 

1 V. at 20 C. = 0.3225 gm. glucose. 

1 V. at 20 C. = 0.1859 gm. fructose. 

1 V. at 20 C. = 0.3286 gm. lactose hydrate. 

1 V. at 20 C. = 0.1247 gm. maltose. 

1 V. at 20 C. = 0.1655 gm. arabinose. 

1 V. at 20 C. = 0.9100 gm. xylose. 

1 V. at 20 C. = 0.2135 gm. galactose. 

1 V. at 20 C. = 0.8645 gm. invert sugar. 

1 V. at 20 C. = 0.1651 gm. raffinose hydrate. 

Use of One Normal Weight for All Sugars. For many laboratory 
purposes it is convenient to employ but one fixed normal weight for all 
saccharimetric work. In such cases the normal weight of sucrose is 
usually taken, the percentage of each particular sugar being calculated 
from the scale reading by means of an appropriate factor. 

The constant polarizations in degrees Ventzke of a normal weight 
of 26.00 gms. of different sugars, when dissolved to 100 metric c.c. and 
polarized in a 200-mm. tube, are given in table XXXVI. The values 
are calculated only to the nearest 0.5 degree, which is sufficiently exact 
when the variations due to change in concentration are considered. 

If no other optically active substances are present, the scale reading 
(V.) of 26.00 gms. of the sugar-containing substance multiplied by 
100 and divided by the corresponding polarizing power of the pure 
sugar will give the percentage. of sugar present. Owing to the changes 
in specific rotation with varying concentration, the percentages thus 
calculated will not be absolutely exact. 



METHODS OF SIMPLE POLARIZATION 



201 



TABLE XXXVI 
Giving Ventzke Reading of 26.00 gms. of Different Sugars in 100 c.c. 



Sugar. 


wf 

26.00 gms. in 100 
metric c.c. 


Calculated read- 
ing v. 

M *D X 100. 
66.5 


Sucrose 
Arabinose 


+ 66.5 
+104 5 


+ 100 

+157 


Xylose 


+19 6 


+29 5 


Glucose. 


+53 2 


4-80 


Fructose 


93 2 


140 


Invert sugar 


-20 


30 


Galactose 


+81 8 


+123 


Maltose . . . 


+138 


+207 5 


Lactose (H 2 O). .' 
Raffinose (5 H 2 O) 
Raffinose (anhydride) . . 


+52.5 
+104.5 
+123.2 


+79 
+157 
+ 185 



TECHNICAL METHODS OF SACCHARIMETRY 

The saccharimeter is most generally employed in the analysis of 
products of the cane- and beet-sugar industry. It must be borne in 
mind, however, that the readings of the saccharimeter scale indicate 
percentages of sucrose only in cases where other constituents have no 
effect upon the scale reading; the results obtained with impure products 
are, therefore, more correctly expressed as degrees polarization or degrees 
sugar scale. For a more accurate determination of sucrose by the 
saccharimeter, the method of inversion must be used which will be 
described in the following chapter. 

Methods for Polarizing Raw Sugars 

Rules of the International Commission. The rules of the Inter- 
national Commission for Unifying Methods of Sugar Analysis* are as 
follows : 

"In general all polarizations are to be made at 20 C. 

"The verification of the saccharimeter must also be made at 20 C. 
For instruments using the Ventzke scale 26 grams of pure dry sucrose, 
weighed in air with brass weights, dissolved to 100 metric c.c. at 20 C. 
and polarized in a room, the temperature of which is also 20 C., must 
give a saccharimeter reading of exactly 100.00. The temperature of the 
sugar solution during polarization must be kept constant at 20 C. 

"For countries where the mean temperature is higher than 20 C., 
saccharimeters may be adjusted at 30 C. or any other suitable tem- 
* Proceedings of Paris Meeting, July 24, 1900. 



202 SUGAR ANALYSIS 

perature, under the conditions specified above, provided that the sugar 
solution be made up to volume and polarized at this same temperature. 

"In effecting the polarization of substances containing sugar employ 
only half-shade instruments. 

"During the observation keep the apparatus in a fixed position 
and so far removed from the source of light that the polarizing Nicol 
is not warmed. 

"As sources of light employ lamps which give a strong illumination 
such as triple gas burner with metallic cylinder, lens and reflector; gas 
lamps with Auer (Welsbach) burner; electric lamp; petroleum duplex 
lamp; sodium light. 

"Before and after each set of observations the chemist must satisfy 
himself of the correct adjustment of his saccharimeter by means of 
standardized quartz plates. He must also previously satisfy himself of 
the accuracy of his weights, polarization flasks, observation tubes and 
cover-glasses. (Scratched cover-glasses must not be used.) Make 
several readings and take the mean thereof , but no one reading may be 
neglected. 

"In making a polarization use the whole normal weight for 100 c.c., 
or a multiple thereof, for any corresponding volume. 

"As clarifying and decolorizing agents use either subacetate of 
lead, alumina cream, or concentrated solution of alum. Boneblack and 
decolorizing powders are to be excluded. 

"After bringing the solution exactly to the mark at the proper 
temperature, and after wiping out the neck of the flask with filter 
paper, pour all of the well-shaken clarified sugar solution on a rapidly 
acting filter. Reject the first portions of the filtrate and use the rest, 
which must be perfectly clear for polarization." 

Methods of the New York Sugar Trade Laboratory., Details of 
manipulation for the above rules are left largely to individual prefer- 
ence or requirement. The course of operations pursued by the New 
York Sugar Trade Laboratory, where rapidity as well as accuracy is 
required, is as follows: 

Weighing. Twenty-six grams of sugar are weighed out in a nickel 
sugar dish provided with a counterpoise (Figs. 116 and 123). The 
sugar is stirred with a horn spoon and, approximately, the normal 
weight transferred to the dish. The final adjustment is then made with 
the dish upon the scale pan of the balance, a little sugar being added 
or removed until the exact weight is secured. The danger of spilling 
sugar upon the scale pan during the weighing is thus largely avoided. 
The weighing is performed as rapidly as possible to avoid loss from 



METHODS OF SIMPLE POLARIZATION 203 

evaporation of moisture and does not usually consume more than a 
minute of time. 

Transferring. The 26 gms. of sugar in the nickel dish are poured 
into a large funnel placed in a sugar flask; any sugar adhering to the 
dish and funnel is then washed into the flask with distilled water, the 
funnel being thoroughly rinsed inside and outside around the bottom 
to insure the complete removal of all sugar to the flask. From 50 to 60 
c.c. of water are sufficient to effect the transference. 

The funnels employed in transferring the sugar are of German 
silver, and have a mouth 4 in. (ll cm.) in width and 3 in. (9 cm.) in 
depth, and a stem 3 in. (9 cm.) in length. The inner diameter of the 




(a) (b) (c) 

Fig. 123. (a) Nickel weighing dish and counterpoise. (6) Funnel for transferring 
sugar, (c) Normal and half-normal metric c.c. sugar weights. 

stem (8J mm.) is sufficiently large to allow a free passage of the sugar 
into the flask and the outer diameter (10 mm.) sufficiently small to 
allow the escape of air from the flask (see Fig. 123). 

Dissolving. The solution of the sugar in the flasks is performed by 
means of a mechanical shaker. The machine employed in the New 
York Sugar Trade Laboratory is a modification by the author of the 
Camp shaker used in iron and steel laboratories. (Fig. -124.) 

The metal disk of this shaker is replaced by a circular piece of oak 
1 in. thick, of the same diameter and of about the same weight, and 
containing 12 holes 2J in. in diameter, each large enough to accommodate 
the bottom of a sugar flask. Six extra gripping devices are inserted in 
the collar of the shaker, thus giving 12 grips in all to hold the necks 
of the flasks. The collar is adjusted so as to bring the grips at the 
right height and exactly over the centers of the circular holes in the 
wooden disk. The bottom of the flasks are inserted in the holes, and, 



204 



SUGAR ANALYSIS 



by pressing the necks against the springs of the grips, the flasks are 
snapped quickly and securely into position. The shaker is connected 
with a small J horse-power electric motor, provided with a rheostat, 
and the speed of its driving wheel gradually brought up to 120 to 130 
revolutions per minute. At this speed, solution of sugar in the flasks, 
using 50 to 60 c.c. of water, is effected in from 5 to 10 minutes, according 
to the size of grain, stickiness of sample, etc. If too much water is 
used in transferring the sugar, less motion is given to the body of the 
liquid, and a longer time is required to effect solution. 




Fig. 124. Mechanical shaker for dissolving sugars. 

Clarifying. The solution is then clarified with the requisite 
amount of lead subacetate solution (sp. gr. 1.25), but no more than 
the amount necessary to secure a clear polariscope reading is ever 
employed. As a rule not over 1 c.c. of the lead subacetate solution 
is used for Java, Peruvian, and high-grade centrifugal sugars, not over 
1 to 2 c.c. for muscovado sugars, from 2 to 6 c.c. for molasses sugars, 
and 3, 4, and 5 c.c. for Philippine mat sugars according to grade. Ex- 
cess of lead solution increases the polarization very markedly and 
strict observance is paid to the rule of minimum quantity necessary 
for clarification. After the lead solution 2 c.c. of alumina cream are 
added, the contents of the flask are well mixed and the volume of liquid 
made up to 100 c.c., after allowing sufficient time for any air bubbles 
to arise which may have been occluded in the lead precipitate. Foam 
and air bubbles, adhering to the surface of the liquid in the neck of 
the flask, are broken up with a fine spray of ether before adjusting the 



METHODS OF SIMPLE POLARIZATION 



205 




I 



volume to the graduation mark. A small bulb atomizer (Fig. 125) 

is convenient for removing foam. 

The distilled water used in all the work is supplied through 

rubber tubing from a large bottle placed at an elevation above 

the laboratory table. The 
outlets of the rubber tubes 
are fitted with pinch cocks 
and glass tips of large and 
fine opening, the former 
being used for transferring 
the sugar and the latter for 
setting the meniscus. The 
adjustment of the meniscus 
to the graduation mark is 
the same as that used in 
calibration (Fig. 119). The /\ 
distilled water used for solu- 
tion is kept as nearly as 
Fig. 125.-Ether atomizer. ^^ ^ 2Q o c> ^ ^ 

completion of the volume of sugar solution to 100 c.c. is always 
made with the contents of the flask at this temperature. 

Filtering. The contents of the flasks after thorough mixing 
are poured upon plaited filters in stemless funnels resting in 
J-pint jars or cylinders (Fig. 120). All glassware is thoroughly 
cleaned and dried before using. The plaited filters, which are 
large enough to hold the entire contents of the flask, are kept 
in a large desiccator until ready for use. The funnels are 
covered with watch glasses during the filtration to prevent 
concentration of liquid through evaporation. The first run- 
nings (10 to 15 c.c.) of the filtrate are rejected and the re- 
mainder used for polarization. 

Methods for Polarizing Juices, Sirups, Molasses, 

Massecuites, etc. 

The method of polarization just described for raw sugars 
may be applied with minor modifications to the examination ^8- 126. 
of sugar-cane, sugar-beet, sorghum, and other plant juices, 
sirups, molasses, massecuites, and all other products which are p i pe tte. 
mostly soluble in water. 

Sucrose Pipette. In the analysis of sugar-containing juices the 
work of analysis may be lightened considerably by the use of Spencer's 



\ 



206 SUGAR ANALYSIS 

or Crampton's sucrose pipette shown in Fig. 126. This pipette is grad- 
uated upon the stem with divisions, divided into tenths, reading from 
5 to 25. The pipette is so calibrated that the volume of juice de- 
livered from the division upon the stem, which corresponds to its 
degrees Brix, is exactly a double normal weight. The pipette is con- 
structed either for Mohr cubic-centimeter or true cubic-centimeter 
flasks, delivering 52.096 gms. and 52.000 gms. of juice respectively. 
The method of employing the pipette is thus described by Spencer.* 

" Determine the density of the juice with a Brix hydrometer, 
noting the degree Brix without temperature correction. Fill the 
pipette with juice to the mark corresponding with its observed degree 
Brix, and discharge it into a 100-c.c. flask. Add 3 to 5 c.c. of diluted 
lead-subacetate solution, complete the volume to 100 c.c. with water, 
mix thoroughly and filter the contents of the flask. Polarize the 
filtrate, using a 200-mm. tube, and divide the polariscope reading by 
2 to obtain the percentage of sucrose. The juice should not be .expelled 
from the pipette by blowing, and sufficient time should be allowed for 
thorough drainage. Each pipette should be tested when received from 
the maker, and in regular work should be used under the conditions of 
the test. The pipette may be conveniently checked against a balance 
by delivering a measured quantity of juice into a tared capsule and 
weighing it. The uncorrected degree Brix and juice of the temperature 
of the Brix observation must be used. If the hydrometer and pipette 
are correct at the parts used, the juice delivered should weigh 52.096 
gms. (or 52.00 gms. for true cubic centimeters). 

" It is not advisable to use these pipettes with liquids of a higher 
density than 25 degrees Brix or of greater viscosity than cane juice. 
These pipettes are usually used in the analysis of miscellaneous samples 
of juice and in the rapid testing of diluted massecuites and molasses for 
guidance in the vacuum-pan work. They should be frequently cleaned 
with a strong solution of chromic acid in sulphuric acid." 

For the analysis of highly concentrated sugar products, such as 
sirups, molasses, massecuites, etc., the normal weight of substance is 
weighed out as with raw sugar. In case of very dark-colored molasses 
and massecuites, it is often necessary to make the normal weight of 
substance after clarification up to 200 c.c. instead of 100 Q.C. in order 
to reduce the depth of color sufficiently to polarize in a 200-mm. or, 
even at times, in a 100-mm. tube. The reading thus obtained is mul- 
tiplied by 2 (or if polarization is made in a 100-mm. tube by 4) to 
obtain the true direct polarization. 

* Spencer's "Handbook for Cane Sugar Manufacturers " (4th Ed.), p. 122. 









METHODS OF SIMPLE POLARIZATION 207 

CLARIFYING AGENTS AND ERRORS ATTENDING THEIR USE 

In the clarification of dark-colored molasses and other sugar-house 
products a much larger amount of clarifying agent must be used than 
is necessary with raw sugars, juices, and other substances of high 
purity. The employment of excessive quantities of clarifying agent 
introduces, however, serious errors in the work of polarization. These 
errors for convenience will be considered under the following heads: 
I. Errors due to the volume of precipitated impurities. 

II. Errors due to precipitation of sugars from solution. 
III. Errors due to change in specific rotation of sugars. 

The influence of these errors will first be considered in connection 
with the different acetates of lead which are the salts most generally 
used for clarification. 

Acetates of Lead. Three well characterized acetates of lead* 
have been isolated in the crystalline form. These are (1) the normal 
or neutral acetate of lead Pb(C 2 H 3 2 ) 2 ,3 H 2 0; (2) the basic acetate 

3 Pb(C 2 H 3 2 )2,PbO,3 H 2 0; (3) the basic acetate Pb(C 2 H 3 O 2 ) 2 ,2 PbO, 

4 H 2 0. The clarifying power of solutions of these acetates is in general 
proportionate to the content of basic PbO. The normal acetate, although 
deficient in decolorizing power and unsuited for the clarification of dark- 
colored products for polariscopic readings, has certain advantages in 
that it does not precipitate reducing sugars from solution and does not 
form soluble lead-sugar compounds of different specific rotation. For 
these reasons the neutral acetate of lead should be employed for clarify- 
ing wherever possible in preference to the basic salt. 

Neutral Lead-acetate Solution. In preparing the neutral acetate 
of lead reagent, a concentrated solution of commercial lead acetate 
(sugar of lead) is made, any free alkali or acid neutralized with acetic 
acid or sodium hydroxide, and the liquid diluted to a density of 30 
degrees Be\ (54.3 degrees Brix or 1.2536 sp. gr. ^). The solution is 
filtered and kept in a stock bottle ready for use. 

Lead-subacetate Solution. Upon digesting litharge with normal 
acetate of lead solution varying amounts of lead oxide are dissolved ac- 
cording to the time and temperature of digestion. Numerous methods 
are employed for preparing lead-subacetate reagent. The following 
examples are given: 

I. Concentrated Solution.] Heat, nearly to boiling, for about half 
an hour, 860 gms. of neutral lead acetate, 260 gms. of litharge, and 

* R. F. Jackson: Scientific Paper, U. S. Bureau of Standards, No. 232 (1914). 
t Spencer's " Handbook for Cane Sugar Manufacturers," p. 229. 



208 



SUGAR ANALYSIS 




500 c.c. of water. Add water to compensate for the loss by evaporation. 
Cool, settle, and decant the clear solution. The solution may be pre- 
pared without heat, provided the mixture is set aside several hours 
with frequent shaking. 

Dilute Solution. Proceed as described above, using, however, 
1000 c.c. of water. The solution should be diluted with cold, re- 
cently boiled distilled water to 54.3 
degrees Brix (30 degrees Be., or 
1.2536 sp.gr. ^). 

II.* Boil 430 gms. of normal lead 
acetate, 130 gms. of litharge, and 
1000 c.c. of water for half an hour. 
Allow the mixture to cool and settle 
and dilute the supernatant liquid 
to 1.25 sp. gr. with recently boiled 
distilled water. 

Ill.t Treat 600 gms. of neutral 
lead acetate and 200 gms. of litharge 
with 2000 c.c. of water. After stand- 
ing 12 hours in a warm place with 
occasional shaking, the solution is 
filtered and the nitrate stored in 
tightly stoppered bottles. The solu- 
tion thus prepared must show a 
strongly alkaline reaction and have 
a specific gravity of 1.20 to 1.25 (at 
17.5 C.) with a content of about 20 
per cent PbO. 

IV. Lead - subacetate solution 
may also be prepared by dissolving 
the solid basic salt (see page 214). 
The concentrated solution is diluted 
with distilled water to a specific 
gravity of 1.25. 

Stock solutions of lead subace- 
tate, both in bottle and burette, 




Fig. 127. Stock bottle and burette for 
lead subacetate solution. 



should be protected by a soda-lime tube from the carbon dioxide of the 
air to prevent deposition of lead carbonate (see Fig. 127). 

* " Methods of Analysis A. O. A. C.," Bull. 107 (revised), U. S. Bur. of Chem., 
p. 40. 

t Fruhling's "Anleitung," p. 457. 



METHODS OF SIMPLE POLARIZATION 209 

I. Errors of Clarification Due to Volume of Precipitated Impurities 

Since all sugar solutions after clarification with lead subacetate, or 
other means, are made up to a definite volume, the space occupied by 
the precipitated impurities will cause the sugar solution to occupy a 
somewhat smaller volume than that of the flask in which the solution 
was made up. An increase in concentration and also in polarization 
is the result. 

Scheibler's Method of Double Dilution. Several methods have 
been devised for estimating the extent of this error. The first to be 
described is Scheibler's* method of double dilution. In this method a 
normal weight of product is dissolved in water, clarified with a meas- 
ured volume of lead subacetate, the volume completed, and solution 
filtered and read in the usual way. A second normal weight of product 
is then weighed out, clarified with the same volume of reagent as be- 
fore and the solution made up to twice the volume of the previous 
experiment. The second solution is filtered and polarized as before. 
The true polarization (P) is then calculated as follows: 

Let PI be the polarization of the first solution made up to volume 
V, and P% the polarization of the second solution made up to volume 
2 V. Let v be the volume of the precipitated impurities which is 
assumed to be the same in both experiments. The normal weight in 
the second solution may be considered to be divided as follows: one 
half dissolved in volume V free from precipitate, the reading of which 

p 

would be j and one half dissolved in volume V containing precipitate, 
A 

p 

the reading of which would be -^ The sum of these quantities divided 

2 

by 2 is the value of P 2 , or 



2 

whence P = 4 P 2 PI. In other words the true polarization is equal 
to four times the polarization of the diluted solution less the polariza- 
tion of the undiluted solution, f 

* Z. Ver. Deut. Zuckerind., 25, 1054. 

f The true polarization is also expressed in other ways as: multiply reading 
of dilute solution by 2, subtract the product from reading of undiluted solution; 
twice the remainder subtracted from reading of undiluted solution will give the true 
polarization: or the difference between the reading of the undiluted solution, and 
twice the reading of diluted solution subtracted from twice the reading of the diluted 
solution will give the true polarization. 



210 SUGAR ANALYSIS 

Example. Polarization of 26 gms. raw sugar, dissolved in water, clarified 
with 2 c.c. lead subacetate and made to 100 c.c. = 94.2 (Pi). 

Polarization of 26 gms. same sugar, dissolved in water, clarified with 2 c.c. 
lead subacetate and made to 200 c.c. = 47.0 (P 2 ). 

True polarization (P) = (47.0 X 4) - 94.2 = 93.8. 

The volume v occupied by the precipitated impurities is calculated 

V X P 
as follows. The reading PI of the undiluted solution is equal to y _ > 

V (Pi - P) 
whence v = 

r\ 

Example. Required the volume of the lead precipitate hi the previous 
example. 

Substituting the values for V, P and Pi,- we obtain 

9 - 10 (94 'V 2 93 ' 8) = - 42 c - c - 

The method of Scheibler owing to its rapidity and ease of execution 
has been very widely used for correcting polarizations for the error due 
to volume of the lead precipitate. The method is open to several 
objections. It is not probable that the volume of the precipitate is 
exactly the same in the dilute as in the undiluted solution, but the prin- 
cipal objection against the method is the very large multiplication of 
any error made in reading the diluted solution. 

Sachs's Method of Correcting Precipitate Error. The method 
devised by Sachs* in 1880 for determining the error due to volume of 
precipitate was intended to obviate the errors of Scheibler's method. 
In the Sachs method the precipitate of impurities obtained in the 
clarification of the sugar solution is washed with cold and hot water 
until all sugar is removed. The precipitate is then transferred to a 
100-c.c. flas'k, a one-half normal weight of sucrose added, the latter 
dissolved and the volume completed to 100 c.c. The solution is mixed, 
filtered, and polarized in a 400-mm. tube. The volume of precipitate 
is then calculated as follows: Let P = the true polarization of the 
sucrose used and PI = the polarization of the sucrose with precipitate. 
The volume (v) of precipitate is then found by the equation 

100 (Pi - P) 



Pi 

Example. A normal weight of granulated sugar dissolved to 100 c.c. 
polarized 99.8 in a 200-mm. tube. 

A one-half normal weight of the same sugar -+- lead precipitate dissolved 
to 100 c.c. polarized 100.25 in a 400-mm. tube. Volume of precipitate (v) = 



* Z. Ver. Deut. Zuckerind., 30, 229. 



METHODS OF SIMPLE POLARIZATION 



211 



Knowing the volume (v) of lead precipitate, the true polarization (P) 
of a product may be determined by the equation P = 1 ~ V S or 

when V= 100, P = - ) ^ 1 Q ~* ;Pl . 

Example. The polarization of a raw sugar (26 gms. to 100 c.c.) was 
96.20 (Pi). The volume of the lead precipitate bySachs's method was 0.22 c.c. (v). 



The true polarization (P) of the sugar = 



100 X 96.2 - 0.22 X 96.2 
100 



= 95.99. 



The method of Sachs has been modified as follows. Instead of 
making a polarization with the washed precipitate the latter is first 
dried. From the weight and specific gravity of the dried lead precipi- 



tate the volume is calculated [v = 



sp. gr. 



and from the volume the 



true polarization is determined by means of the preceding formula. 

The specific gravity of the dried lead precipitates of raw cane sugars 
was determined by Wiechmann* by weighing in a pycnometer with 
benzine. The results of Wiechmann are given in Table XXXVII. 

TABLE XXXVII 

Giving Specific Gravity and Volume of Lead Precipitates from 26 gms. of Different 

Raw Cane Sugars 



Sugar. 


(Weight 
of precipitate 
in grams. 


Specific gravity 
H 2 O = 1.00. 


Volume in 
cu. centimeters 


Jamaica Muscovado 


4559 


1.88 


0.24 


Maceio Muscovado . . 


8112 


1.65 


0.49 


San Domingo centrifugal 


0.2525 


2.91 


0.09 


Sandwich Island centrifugal. 


0.1378 


2.84 


0.05 


San Domingo concrete 


1 0139 


3 80 


27 


Porto Rico molasses sugar 


0.8959 


4.35 


0.21 


Sandwich Islands 


1 0195 


4.38 


0.23 


Cebu mats 


1 5400 


2.17 


0.71 


Manila mats 


1.3350 


2.22 


0.60 











Similar results by Home are given in Table XXXVIII. The 
method employed by Hornet consists in weighing the freshly washed 
precipitate in a calibrated pycnometer filled to the mark with distilled 
water; the precipitate is then washed upon a weighed filter, dried and 
weighed. 

The methods, which are based upon the separation and examina- 
tion of the washed lead precipitate, throw much light upon the errors 

* Proc. Fifth Int. Cong. Applied Chem. (Berlin, 1904) III, 118. 
t J. Am. Chem. Soc., 26, 186. 



212 



SUGAR ANALYSIS 



of clarification; they are not adapted, however, to practical work 
owing to the large amount of time and labor involved. 

Home's Method of Dry Defecation. A third method of eliminat- 
ing the volume of precipitate error is Home's* process of dry defeca- 
tion. The method is thus described by its author: 

" The normal weight of sugar is dissolved in water in a 100-c.c. 
flask and made up to the mark without defecation. The concentra- 
tion is thus at exactly the proper degree. It now remains to defecate 
the solution properly by precipitating the impurities in such a way as to 
produce the minimum change in the concentration of the solution of 
sucrose. This is accomplished by adding to the 100 c.c. of liquid 
small quantities of powdered anhydrous lead subacetate until the im- 
purities are nearly all precipitated. This point is as easily determined 
as in the defecation by a solution of the same salt. The organic and 
mineral-acid radicals in the solution combine with and precipitate the 
lead and lead oxide of the dry salt, while the acetic-acid radical of the 
lead subacetate passes into solution to combine with the bases originally 
united to the other acid radicals." 

Results obtained by Home upon 12 raw cane sugars are given in 
Table XXXVIII, and show a very close agreement between the cor- 
rected polarization by Sachs's method and the polarization by dry 
defecation. 

TABLE XXXVIII 





Grade, country. 


Ordinary 
polariza- 
tion. 


Specific 
gravity of 
precipitate. 


Volume of 
precipitate. 


Corrected 
polariza- 
tion. 


Dry lead 
polariza- 
tion. 


1 

2 

3 
4 
5 

6 
7 
8 
9 
10 
11 
12 


Centrifugal 
Centrifugal (mixed ( 


95.0 
94.5 

96.95 
97.425 
85.8 
89.4 
89.225 
86.45 
90.675 
89.35 
89.4 
88.4 


2.98 


c.c. 

0.10 
0.0765 

0.0378 
0.0884 
0.4118 
0.39 
0.4204 
0.7108 
0.3204 
0.8500 
0.4554 
0.4924 


94.9 
94.43 

96.91 
97.33 
85.45 
89.05 
88.85 
85.84 
90.39 
88.59 
88.99 
87.97 


94.9 
94.4 

96.95 
97.375 

85.5 
89.0 
88.85 
85.95 
90.45 
88.775 
89.0 
88.0 


Centrifugal, Trinidad . . 
Centrifugal, Java 
Muscovado, St. Croix . . 
Molasses, Cuba 
Molasses. . . . 


2.91 
2.30 
1.91 
3.20 

2.85 
1.96 
3.20 


Molasses 


Molasses 


Molasses . 


Molasses . . 


3.01 
2.64 


Molasses, Cuba 





Home's method has been tested by a number of chemists upon 
raw cane sugars with results very similar to the above. Pellet, f how- 
* J. Am. Chem. Soc., 26, 186. 
t Bull, assoc. chim. sucr. dist., 23, 285. 



METHODS OF SIMPLE POLARIZATION 213 

ever, has criticized the method principally upon the ground that the 
increase in polarization due to the volume of precipitate is not as 
great as calculated, owing to the decrease in polarization caused by the 
retention of sucrose in the precipitate, this retention error frequently 
more than counterbalancing the error due to volume of precipitate. 
Subsequent results by Home* and other chemists show, however, that 
there is no appreciable retention of sucrose when the dry lead reagent 
is used in minimum amounts. Another objection by Pellet, that only 
part of the lead salt acts and that the rest passes into solution, thus 
increasing the volume and diminishing the polarization, deserves con- 
sideration. 

With the higher grade of sugar-house products there is no difficulty 
in securing a satisfactory clarification with a minimum amount of the 
dry lead salt, the lead dissolved being immediately precipitated and 
but very little remaining in solution. With low-grade sugars, molasses, 
etc., the case is otherwise. If dry lead subacetate, or subacetate solu- 
tion, be added to a solution of such products to the point of satis- 
factory clarification a considerable amount of lead salt will usually 
remain dissolved. The rule of adding the powdered salt until no more 
precipitate forms is not always a criterion of the absence of lead in the 
filtrate. When subacetate is added to solutions of low purity the first 
portions of lead are completely precipitated; then comes a point where 
with the formation of additional precipitate a small amount of lead 
remains in solution; the amount of the latter continues to increase 
until at the point where no more precipitate is formed nearly all of the 
lead added remains dissolved. (See Table XXXIX.) With very low 
grade products there is therefore a danger of the dry lead salt increasing 
the volume of solution; whether this increase will cause a lowering of 
the polarization or not will depend upon the character of the product. 
With low-grade sugar-cane products the error due to increase in volume 
of solution may be more than counterbalanced by the precipitation 
of levorotatory fructose. 

In the following experiments by Hall f in the New York Sugar Trade 
Laboratory the effect of increasing amounts of dry lead subacetate upon 
the polarization of a Philippine mat sugar was studied. The quantity 
of lead in the clarified filtrates was determined and the dilution calcu- 
lated by allowing an increase of 0.22 c.c. in volume for 1 gm. of dry 
subacetate dissolved in 100 c.c. of solution. 

* J. Am. Chem. Soc., 29, 926. 

t Bull. 122, U. S. Bur. of Chem., p. 225. 



214 



SUGAR ANALYSIS 



TABLE XXXIX 

Showing Estimated Dilution of a Sugar Solution by Dry Lead Subacetate 





Amount of 


In 100 c.c 


-. filtrate. 


Estimated 




Clarifying agent. 


clarifying 
agent used. 


PbO. 


Pb sub- 
acetate. 


dilution. 


Polarization. 


Subacetate solution. . 


3 c.c. 


grams. 

0.2678 


grams. 


c.c. 


86 70 


Dry subacetate . 


0.5 gm. 


Trace 




Trace 


Too dark to read. 


Dry subacetate 
Dry subacetate 
Dry subacetate 


1.0 gm. 
2.0gms. 
4.0 gms. 


0.1530 
0.7203 
2.1078 


(0.20) 
(0.94) 
(2.73) 


0.05 
0.20 
0.60 


86.50 
86.60 
86.50 















It is noted that with an estimated dilution of 0.2 c.c. instead of a 
decrease in polarization, as would be expected, there is an increase. 
With an estimated dilution of 0.6 c.c. the reading is the same as that 
first obtained, so that the combined effect of the dry lead upon the 
precipitation of fructose and upon the lowering of the rotation of the 
fructose in solution is seen to be most pronounced. With sugar-cane 
products the use of dry lead subacetate to the point of satisfactory 
clarification would seem to involve no decrease in polarization. With 
low-grade sugar-beet and other products, which are comparatively 
free from fructose, there is however a danger of too low polarization 
since there is no compensating influence for the dilution caused by the 
excess of lead subacetate dissolved. 

In using dry lead subacetate for defecation the chemist must be 
certain of the composition of his preparation. The powdered salt must 
be dry and should contain the requisite amount of basic lead. Some 
samples of dry lead subacetate sold by the trade have been found by 
the author to consist almost entirely of the normal acetate. A very 
pure anhydrous lead subacetate is manufactured at present having 
closely the formula, 3 Pb(C 2 H 3 2 )2, 2 PbO. A sample of such a prepa- 
ration analyzed at the New York Sugar Trade Laboratory gave the 
following results: 





Total Pb. 


Basic Pb. 


Found 


Per cent. 
7Q 00 


Per cent. 
Qn n^ 


Theory for 3 Pb(C 2 H 3 O 2 j 2 , 2 PbO.. . . 


72.84 


29.14 



The above formula would correspond to a mixture of four parts 



METHODS OF SIMPLE POLARIZATION 215 

of the basic acetate 3 Pb(C 2 H 3 2 )2,PbO and three parts of the basic 
acetate Pb(C 2 H 3 2 ) 2 ,2 PbO.* 

A solution of lead subacetate of 1.259 sp. gr., as employed for clari- 
fication in the wet way, was found to contain 0.2426 gm. total Pb per 
Ic.c. One-third gram dry salt is therefore equivalent to 1 c.c. subace- 
tate solution in clarifying power. A low-grade sugar requiring 6 c.c. 
of subacetate solution of the above strength for clarification would 
accordingly need 2 gms. of salt for dry defecation. 

The dry subacetate of lead employed in sugar analysis should be 
finely ground in order that it may be acted upon quickly and com- 
pletely by the dissolved impurities. The tendency to form insoluble 
crusts upon the powdered grains of dry salt has been noted by Home, 
especially in refinery products subjected to the influence of bone black. 
In such cases Home recommends the addition of a little dry sand with 
the powdered lead salt; the particles of sand in shaking will grind off 
the crusts of insoluble matter and allow the lead to be acted upon. 

II. Errors of Clarification due to Precipitation of Sugars from Solution 

In the absence of free alkalies sucrose is not precipitated from solu- 
tion by lead subacetate. Reducing sugars, however, are precipitated 
by solutions of basic lead salts. This precipitation does not occur 
with the amounts of lead used in ordinary clarification except in pres- 
ence of those salts or acids which form insoluble lead compounds! (as 
chlorides, sulphates, phosphates, carbonates, oxalates, tartrates, 
malates, etc.). Whether this precipitation of reducing sugars is due 
to simple occlusion or to the formation of insoluble sugar-lead com- 
plexes is not definitely known. 

The extent to which the common reducing sugars, glucose and 
fructose, are precipitated by different lead clarifying agents, has been 
investigated by Bryan. J Separate solutions of glucose and fructose 
were prepared, using 5 gms. of sugar with 1 gm. each of magnesium 
sulphate and ammonium tartrate. To 50 c.c. of this solution the 
clarifying agent was added and the volume made up to 100 c.c. After 
filtering, the excess of lead was removed with potassium oxalate, and 
the sugar in solution determined by Allihn's method. The results of 
Bryan's experiments are given in the following table. 

* Jackson in an unpublished experiment communicated to the author shows that 
Home's dry subacetate is in fact a mixture of these two basic acetates. 
t Prinsen Geerligs, Deut. Zuckerind., 23, 1753. 
t Bull. 116, U. S. Bur. of Chem., p. 73. 



216 



SUGAR ANALYSIS 



TABLE XL 

Showing Precipitation of Glucose and Fructose by Basic Lead Salts 



Clarifying agent. 


Amount per 
100 c.c. of 
solution. 


Glucose pre- 
cipitated. 


Fructose pre- 
cipitated. 


Neutral lead acetate solution 
Neutral lead acetate solution 
Lead subacetate solution 


3. 5 c.c. 
7.0 c.c. 
3 5 c.c. 


Per cent of total. 

0.93 
0.84 
3 35 


Per cent of total. 

0.00 
0.00 
8 03 


Lead subacetate solution 


7.0 c.c. 


8.34 


19 91 


Dry lead subacetate 


1.0 gm. 


3.85 


14 93 


Dry lead subacetate 


2.5 gms. 


17.48 


35.33 


Basic lead nitrate solution 
Basic lead nitrate solution 


4.0 c.c. 
8.0 c.c. 


6.27 
5.61 


13.84 
25.12 











It is seen that neutral lead acetate precipitates but very little reduc- 
ing sugar, whereas the basic lead salts remove a large percentage of 
both glucose and fructose, the latter sugar, however, in more than 
double the amount. This precipitation of reducing sugars during 
clarification has a most marked effect upon the polarization, the re- 
moval of glucose from solution diminishing the dextrorotation, and that 
of fructose the levorotation. The greater precipitation of fructose in 
mixtures with sucrose and glucose, as in the clarification of sugar-cane 
products, jellies, jams, etc., causes an increase in the dextrorotation, 
frequently exceeding 1 Ventzke. The precipitation of reducing sugars, 
while of no consequence as regards the saccharimetric or gravimetric 
determination of sucrose, is of the greatest importance when the valua- 
tion of a product is based upon the polarization alone, or upon a deter- 
mination of reducing sugars. 

III. Errors of Clarification due to Change in Specific Rotation of Sugars 

Action of Lead Subacetate on Rotation of Sucrose. The results 
of Miintz,* Weisberg,| Svoboda,t Groger and other investigators show 
no perceptible influence of basic lead acetate upon the specific rotation 
of sucrose in aqueous solution. Recent experiments by Bates and 
Blake || indicate, however, a very perceptible influence in case the lead 
reagent is used in large excess. The following table, showing the loss 
and gain in polarization for a normal weight of pure sucrose, is taken 
from the work of Bates and Blake. 

* J. fabr. sucre., 17, 25. 

t Sucrerie Beige, 16, 407. 

t Z. Ver. Deut. Zuckerind., 46, 107. 

Oest. Ung. Z. Zuckerind., 30, 429. 

|| Bull. U. S. Bur. of Standards, 3 (1), p. 105. 



METHODS OF SIMPLE POLARIZATION 



217 



TABLE XLI 



Number of cubic 
centimeters of 
basic lead solution 
(1.25 sp. gr.) added. 


Difference in 
degrees Ventzke 
between similar 
solutions, one with 
the other without, 
basic lead acetate. 


0.5 


-0.09 


1.0 


-0.13 


2.0 


-0.13 


3.0 


-0.08 


4.0 


-0.06 


5.0 


-0.03 


6.0 


0.00 


7.0 


+0.05 


8.0 


+0.09 


10.0 


+0.19 


15.0 


+0.29 


20.0 


+0.45 


25.0 


+0.58 


30.0 


+0.62 


35.0 


+0.77 


40.0 


+0.77 


63.0 


+0.95 



The + sign indicates that the solution containing the lead sub- 
acetate gives the higher polarization, and conversely for the sign. 

Action of Lead Subacetate on Rotation of Fructose. While the 
specific rotation of sucrose under the ordinary conditions of analysis 
is not modified sufficiently by subacetate of lead to introduce serious 
errors, the case is otherwise with fructose. Gill* first showed, in 1871, 
that the specific rotation of fructose was greatly diminished by the 
presence of lead subacetate, this decrease being so great that in presence 
of sufficient basic lead the rotation of invert sugar ([a]= 20) was 
changed to the right. This change in rotation is due to the formation 
of soluble dextrorotatory lead fructosate, the presence of which, even 
in small amounts, is sufficient to reduce the figure for the rotation of 
fructose (Wg=-92) below that of glucose (Ms = + 52.5). Gill f 
showed that the error due to formation of soluble lead fructosate could 
be entirely avoided by adding acetic acid to the point of acidity, thus 
decomposing the soluble lead fructosate into lead acetate and free 
fructose of normal specific rotation. In case the soluble lead fructosate 
is not decomposed by some precipitating agent of lead, acetic acid 

* Z. Ver. Deut. Zuckerind., 21 (1871), 257. 

t LOG. cit. See also "Spencer's Handbook for Cane Sugar Manufacturers" 
(4th Ed.), p. 88; Edson, Z. Ver. Deut. Zuckerind., 40, 1037; Pellet, Bull, assoc. 
chim. sucr. dist., 14, 28, 141. 



218 SUGAR ANALYSIS 

should be added to weak acidity before making up the volume of the 
clarified solution to 100 c.c. for the direct polarization of low-grade 
fructose containing products. 

Miscellaneous Methods of Clarification 

Numerous modifications of the lead process of clarification have 
been proposed as a means of reducing or eliminating the several sources 
of error just mentioned. Freshly precipitated lead carbonate, lead 
chloride, and lead nitrate have been employed as clarifying agents, but 
with only indifferent success. Two methods of lead clarification, 
which have found considerable favor in France and Austria, should, 
however, be mentioned in addition to the processes previously de- 
scribed. These are Zamaron's method by means of hypochlorite of 
lime and neutral lead acetate, and Herles's method by means of basic 
lead nitrate. 

Zamaron's * Method of Clarification with Hypochlorite. 625 
grams of dry commercial bleaching powder are thoroughly ground up 
in a large mortar with 1000 c.c. of water. The mass is squeezed out 
in a sack and the extract filtered through paper. The solution thus 
obtained (700 c.c. to 800 c.c. of about 18 Be.), is preserved in a stop- 
pered bottle of dark glass away from the light. 

The solution to be clarified is treated with a few cubic centimeters 
of the hypochlorite solution, sufficient to effect decolorization, and 
then a few cubic centimeters of neutral lead acetate solution are added. 
There is usually a slight rise in temperature after addition of the clarify- 
ing agents so that the solution must be recooled before making to 
volume. 

The Zamaron process secures usually a good clarification, does not 
precipitate reducing sugars, and forms no objectionable lead sugar 
compounds. The chief fault of the method is the volume of precipitate 
error, which in this case is augmented by the formation of considerable 
lead chloride. 

Herles's f Method of Clarification with Basic Lead Nitrate. Dis- 
solve 100 grams of solid sodium hydroxide in 2000 c.c. of water; a second 
solution is prepared by dissolving 1000 gms. of neutral lead nitrate in 
2000 c.c. of water. Upon mixing equal volumes of the two solutions 
basic lead nitrate is precipitated according to the equation 

2 Pb(N0 3 ) 2 + 2 NaOH = Pb(NO 3 ) 2 .Pb(OH) 2 + 2 NaN0 3 

Lead nitrate Basic lead nitrate 

* Fribourg's "Analyse chimique," p. 129. 

t Z. Zuckerind., Bohmen, 13, 559; 14, 343; 21, 189. 



METHODS OF SIMPLE POLARIZATION 219 

The precipitated basic lead nitrate is washed free from sodium com- 
pounds and then mixed with water to a cream, in which form it may be 
used for clarification. 

The clarification is performed more commonly by forming the basic 
nitrate within the solution to be clarified. This is done by first adding 
a measured quantity of the lead-nitrate solution (1 c.c. to 15 c.c. accord- 
ing to depth of color) and then, after mixing, an equal volume of the so- 
dium hydroxide solution. After shaking, the solution is made to volume, 
well mixed, and filtered. Care must be taken that the reaction of the 
solution is not alkaline after mixing; this is best provided for by testing 
the two solutions against one another before using. 

Formation of the basic lead nitrate within the solution gives usually 
a much better clarification than addition of the washed cream, but has 
the disadvantage of introducing considerable sodium nitrate, which, if 
present in large quantity, will affect the rotation of the sugars. 

The basic lead nitrate method gives an exceedingly brilliant clari- 
fication. The process is open, however, to the same errors as basic 
lead acetate. There is first the volume of precipitate error, which is 
further augmented by the copious bulk of the basic lead nitrate itself; 
and secondly there is a precipitation of reducing sugars as shown by 
the results of Bryan in Table XL. 

The numerous errors incident to the use of basic lead compounds 
in clarification have led chemists to seek other means of decolorizing 
solutions for polarization. It is impossible, as well as unnecessary, to 
take up all the processes which have been devised to accomplish this 
end. Two of these methods, however, should be described: (1) De- 
colorization by means of bone black or blood charcoal; (2) Decoloriza- 
tion by means of hydrosulphites, sulphoxylates, etc. 

Decolorization of Sugar Solutions by means of Bone Black. The 
use of bone black as a decolorizing agent in sugar refineries is well 
known. The same substance in a more finely divided specially pre- 
pared form is employed at times as a decolorizer in sugar analysis. 

Purification of Bone Black. If purified animal charcoal (preferably 
blood charcoal) has not been obtained from the dealer the chemist may 
purify the commercial product as follows: The char is finely ground in 
a mortar and then digested several hours in the cold with dilute hydro- 
chloric acid. The acid is then decanted, the char brought upon a 
filter and washed with distilled water until all traces of hydrochloric acid 
are removed. After drying in a hot-air oven, the char is heated to dull 
redness in a covered porcelain crucible, and then, after cooling suffi- 
ciently, placed while still warm in a dry stoppered bottle. 






220 SUGAR ANALYSIS 

Several methods are followed in the employment of animal charcoal 
for decolorizing. One very common practice is to make up the solu- 
tion to volume and shake thoroughly with a small quantity of charcoal, 
using from 0.5 to 3 gms. according to depth of color. The contents of 
the flask are then poured upon a dry filter and the filtrate taken for 
polarization. 

Absorption Error of Bone Black. In the above method of decolor- 
izing, a certain error is introduced owing to the absorption and reten- 
tion of sugar by the char. Sugars differ markedly in the extent to 
which they are absorbed by animal charcoal. In the case of the simple 
reducing sugars, glucose, fructose, etc., the error through absorption is 
so small as to be almost negligible, but in the case of sucrose and other 
higher saccharides the absorption is so great that an error of several 
degrees Ventzke may be occasioned in the polarization. 

One method of eliminating the error through absorption of sucrose 
consists in adding a correction previously established by experiment 
upon pure sugar solutions. If, for example, a sucrose solution polariz- 
ing 95.0 V. gives, after shaking 50 c.c. with 2 gms. of charcoal for 5 
minutes, a polarization of only 94.7 V., then a correction of 0.3 V. must 
be added to all polarizations of about 95 V. for sugars decolorized in 
this same way. A correction table is thus made for sugar solutions of 
different concentrations, but in applying these corrections care must be 
taken that the quality and quantity of the char are alike in both in- 
stances and that the time of shaking is always the same. With impure 
products of variable composition the employment of absorption factors 
is attended with considerable uncertainty. 

Spencer* has recommended a different method of employing animal 
charcoal for the purpose of reducing the absorption error to a minimum. 
The process is thus described: 

" Place a small quantity of bone black, about 3 gms., in a small 
plain filter, selecting a rather slow filtering paper. Add a volume of 
the solution equal to that of the char, or just completely moisten the 
latter, and let this liquid filter off. After four or five similar nitrations, 
the filtrates from which are rejected, test the filtrates by a polariscopic 
observation and note whether the reading varies. Solutions must be pro- 
tected from evaporation during the filtration. As soon as the reading 
is constant, showing no further absorption, record it as the required 
number." 

The method just described, while largely eliminating, does not 
completely remove, the errors of absorption, for while the retention of 

* Spencer's " Handbook for Cane Sugar Manufacturers" (4th Ed.), p. 89. 



METHODS OF SIMPLE POLARIZATION 



221 



sucrose by the char rapidly diminishes with each successive portion of 
solution, it soon becomes only a gradually receding quantity. This is 
shown by the following experiments upon a sucrose solution polarizing 
49.9 V. 



Fraction of filtrate. 


Polarization. 


Absorption 

error. 


First running. . . . 


48 9 


1 


Second running. . . 


49 4 


5 


Third running 
Fourth running 
Fifth running 


49.75 
49.80 
49.80 


0.15 
0.10 
0.10 



With dark-colored solutions it also happens that with each suc- 
ceeding portion of the nitrate, the charcoal loses its absorptive power 
for coloring matter as well as for sucrose, so that the final running least 
free from the error of absorption is too dark for satisfactory polariza- 
tion. 

The general consensus of opinion regarding the use of animal char- 
coal in sugar analysis is that it should be used as a decolorizing agent 
only as a last resort. Its employment in the polarization of raw cane 
sugars has been condemned by the International Commission upon 
Unification of Methods.* In the polarization of low-grade sugar 
products its use, however, seems at times justified by necessity; in 
all such cases efforts should be made to reduce the absorption error to 
a minimum. 

Decolorization of Sugar Solutions by Means of Hydrosulphites. 
Attempts have been made to employ various decolorizing agents for 
the purpose of avoiding the precipitate errors of basic lead salts and 
the absorption error of bone black. The most promising of the numer- 
ous substances which have been tried in this connection are the salts 
and derivatives of hydrosulphurous acid.f 

The employment of commercial hydrosulphite preparations, such 
as " Blankit," " Redo," etc., has been common in the sugar factory, 

* See page 202. 

t The dry sodium hydrosulphite is prepared by allowing zinc, sodium bisulphite, 
and sulphuric acid to react in the following molecular proportions: 

2 NaHSOa + Zn + H 2 SO 4 = ZnS 2 O 4 + Na 2 SO 4 + 2 H 2 O. 
The zinc hydrosulphite is then decomposed with sodium carbonate, 

ZnS 2 O 4 + Na 2 CO 3 = Na 2 S 2 4 + ZnCO 3 . 

The sodium hydrosulphite is salted out from solution by means of sodium chloride 
and dehydrated by warming with strong alcohol. The compound is then dried in 
vacuo at 50 to 60 C. 



222 SUGAR ANALYSIS 

where they have been used for bleaching dark-colored massecuites and 
also, in solution, as a wash for whitening sugars in the centrifugal. 
They have also been employed by unscrupulous manufacturers for 
bleaching low-grade molasses in the preparation of table sirups. 

For their use in sugar analysis the solution to be decolorized is 
treated with a lew cubic centimeters of alumina cream and a few 
crystals of sodium hydrosulphite (0.1 gm. to 1.0 gm., according to the 
depth of color) ; after mixing and dissolving, the volume is made up to 
the mark, and the solution filtered. The filtrate should be polarized 
immediately. 

In many cases tjiere is a rapid redarkening of solutions decolorized 
with hydrosulphites. Weisberg,* from his study of the action of 
hydrosulphites, concludes that the bleaching action is a double one, 
first, by means of the free sulphurous acid when decolorization is per- 
manent, and secondly by means of the nascent hydrogen which is 
evolved, when there is a redarkening of the solution through oxidation- 
Afterdarkening may be prevented by the use of another hydrosulphite 
derivative, sodium sulphoxylate-formaldehyde, sold commercially as 
" Rongalite." The latter, however, is much slower in its bleaching 
action than hydrosulphite and is not always an effective decolorizing 
agent. 

A serious objection against hydrosulphite is its action upon the 
polarizing power of certain reducing sugars. Bryan f has found that 
the polarizing power of glucose was decidedly lowered after the ad- 
dition of hydrosulphite, owing to the formation of a levorotatory oxy- 
sulphonate. Rongalite did not produce this effect. Neither rongalite 
nor hydrosulphite caused any immediate change in the polarization of 
fructose or sucrose. Numerous cases of inversion of sucrose by the 
prolonged action of hydrosulphites have been reported, however, in the 
literature. 

The experience of chemists, in the use of hydrosulphites as a de- 
colorizing agent for sugar analysis, has been upon the whole unfavor- 
able. In many cases the decolorized solution becomes turbid through 
separation of sulphur, thus rendering polarization impossible. The 
bleaching action of hydrosulphite is also limited, and has but little 
decolorizing effect upon caramel substances, which are among the 
chief causes of discoloration in sugar-house products. 

Aluminum Hydroxide as a Clarifying Agent. A common prepa- 
ration, used in connection with other clarifying agents, yet having but 

* Centrbl. Zuckerind, 15, 975. 

t Bull. 116, U. S. Bur. of Chem., p. 76. 



METHODS OF SIMPLE POLARIZATION 223 

little decolorizing power in itself, is aluminum hydroxide, or, as it is 
more generally termed, "alumina cream." The method of preparing 
alumina cream, as prescribed by the Association of Official Agricultural 
Chemists, is as follows:* 

"Prepare a cold saturated solution of alum in water and divide 
into two unequal portions. Add a slight excess of ammonium hydrox- 
ide to the larger portion and then add by degrees the remaining alum 
solution until a faintly acid reaction is secured." 

The reagent as above prepared consists of aluminum hydroxide 
suspended in a solution of ammonium and potassium sulphates. The 
salts have a certain advantage, when alumina cream is used as an 
adjunct with lead salts, in helping to precipitate any excess of lead 
from solution. In certain cases, however, the presence of ammonium 
and potassium sulphates is detrimental, so that for many purposes it is 
better to employ a salt-free cream. For the preparation of the latter, 
concentrated alum solution is precipitated with a slight excess of am- 
monia and then washed by decantation with water until the solution 
is free from sulphates. The excess of water is then poured off and 
the residual cream stored in a stoppered bottle. 

The clarifying effect of alumina cream is chiefly mechanical; its 
action consists largely in carrying down finely suspended or colloidal 
impurities which would otherwise escape filtration. When used in 
connection with lead subacetate it promotes the coagulation of the 
precipitated impurities and renders filtration more perfect and rapid. 

For the polarization of very high grade sugars, sirups, honeys, etc., 
alumina cream is the only clarifying agent required. In all such cases 
only the salt-free reagent should be used. About 2 c.c. of the cream 
are sufficient for clarification and the volume of aluminum hydroxide 
in this amount is too insignificant to affect the polarization. 

Concentrated alum solution is sometimes used with lead subacetate 
for clarifying. The precipitate, formed between the lead salt and alum, 
helps to remove coloring matter, but the increase in precipitate and 
other errors tend to nullify any advantages of the method. 

Comparisons of Different Clarifying Agents. 

A few examples, taken from the reports of Referees upon Sugar for 
the Association of Official Agricultural Chemists, are given in order to 
show the probable error of different clarifying agents in polarization. 

* Methods of Analysis A. O. A. C. Bull. 107 (revised), U. S. Bur. of Chem., 
p. 40. 



224 



SUGAR ANALYSIS 



TABLE XLII 

Polarization of Mixtures of Sucrose, Glucose, and Fructose with 
0.5 gm. Ammonium Oxalate and 0.5 gm. Sodium Sulphate, 
using Different Clarifying Agents (Bryan) * 



Clarifying agent. 


Amount of clari- 
fying agent used. 


Direct polari- 
zation. 


Alumina cream . . . 


5 c.c. 


89 00 V. 


Lead subacetate solution 


3.5 c.c. 


89 50 


Lead subacetate solution 
Neutral lead acetate solution. . . 
Neutral lead acetate solution. . . 
Basic lead nitrate solution 
Dry lead subacetate 


7 c.c. 
3 c.c. 
6 c.c. 
4 c.c. 
1 5 gms. 


89.55 
89.20 
89.20 
89.00 
89 05 


Sodium hydrosulphite . 


1 cm. 


88.60 









Taking the experiment with alumina cream as the true polarization, 
it is seen that the lead subacetate solution gives a reading 0.5 V. too 
high and the normal lead acetate 0.2 V. too high. The excess reading 
in the second case is due to the volume of precipitate and in the former 
to both volume of precipitate and precipitation of fructose. The dry 
lead subacetate and basic lead nitrate clarifications give readings 
practically identical with the true polarization. This might seem to 
indicate no precipitation of optically active reducing sugars; such a 
precipitation does take place, however, and the experiment only shows 
that in this particular instance the various errors of clarification happen 
to neutralize one another. Treatment with hydrosulphite gives a 
polarization below the true value owing to the change in rotation of the 
glucose. 

TABLE XLIII 

Polarizations of Raw Cane Sugar and Cane Molasses, using Different Clarifying 
Agents (Average Results of Several Collaborators) 





Sugar. 


Molasses. 


Alumina cream and hydrosulphite 


+92 75 


+41 09 


Neutral lead acetate solution 


92 92 


42 46 


Basic lead acetate solution. . 


93 05 


42 82 


Basic lead nitrate solution 


92 98 


43 23 


Dry lead subacetate 


92 90 


42 63 









Direct polarization. 



The experiments show a lower polarization using hydrosulphite, a 
result due in large part to the change in rotation of glucose. Basic lead 
* Bull. 116, U. S. Bur. of Chem., p. 71. 



METHODS OF SIMPLE POLARIZATION 225 

acetate and nitrate solutions give much higher polarizations owing to both 
the volume of precipitate error and the precipitation of fructose. Neutral 
lead acetate solution and dry lead subacetate give polarizations between 
these two extremes, there being, however, in case of the former, a volume 
of precipitate error and in case of the dry lead an error due to precipita- 
tion of reducing sugars. The true polarization would be somewhere be- 
tween the results obtained with hydrosulphite and neutral lead acetate. 
The selection of an appropriate clarifying agent is one of the most 
important operations of saccharimetry, and in making his selection the 
chemist must be governed by the requirements of each particular case. 
Rapid nitration and brightness of clarification are factors which must 
be considered as well as minimum degree of error. Beginning with 
products of highest purity alumina cream alone should be used wherever 
possible. With products of slight discoloration, when alumina cream 
is insufficient, neutral lead acetate solution should be tried. When 
alumina cream and neutral lead solution fail, lead subacetate, or basic 
lead nitrate, or neutral lead acetate with hypochlorite may be employed; 
dry lead subacetate will usually give more accurate results with sugar- 
cane and other products containing fructose. Animal charcoal or hydro- 
sulphites should be used only as a last resort, when other means of 
clarification have failed. The smallest possible quantity of clarifying 
agent should be used in all cases. 

POLARIZATION OF SUGAR PRODUCTS CONTAINING INSOLUBLE MATTER 

In the analysis of juices, sirups, molasses, massecuites, and sugars, 
the chemist has to deal with substances which are entirely soluble in 
water. The work of polarization becomes more complicated when 
considerable insoluble matter is present, as happens in the analysis of 
fruits, tubers, stalks, and other vegetable substances or in the examina- 
tion of filter-press cake, scums, and other sugar-house residues. 

The methods for polarization of succulent plant materials may be 
divided into three general classes: (1) Methods of Expression; (2) 
Methods of Extraction, and (3) Methods of Digestion. As an illus- 
tration of these several methods the polarization of sugar beets offers 
a good and classic example. 

Sampling Sugar Beets, Etc. In preparing sugar-beets, sugar 
cane, fruits, etc., for analysis the material must first be reduced to a 
finely divided condition. For this purpose any of the numerous 
mechanical rasps, shredders, graters, etc., may be employed, provided 
that the cellular tissue be thoroughly disintegrated and that no losses 
occur through leakage of juice or evaporation. 



226 



SUGAR ANALYSIS 



Keil's Beet Sampler. The Keil boring machine (Fig. 128) is very 
frequently used for taking samples of individual sugar beets. The 
essential feature of the apparatus consists of a hollow detachable bit, 
the construction of which is shown in Fig. 129. The conical rasp at 




Fig. 128. Keil's boring rasp for sampling sugar beets. 

the end, revolving at a speed of about 3000 revolutions per minute, re- 
duces the substance of the beet to an extreme degree of fineness and at 
the same time forces the pulp through a small opening into the cavity 




Fig. 129. Detachable bit of Keil's boring rasp. 

within. Each beet is bored in an inclined direction, as shown in Fig. 
130, in order to secure the best representative sample. When only 
single beets are examined (as in the selection of " mother beets " for 
seed production) the bit is detached after each boring and a new one 
screwed on. The bits are numbered, and to obtain the sample the 
conical rasp is removed and the pulp (from 8 to 14 gms., according to 
the size of beet and length of boring) forced out with a rod. In samp- 
ling large numbers of beets the bit is kept in constant use, the pulp 



METHODS OF SIMPLE POLARIZATION 



227 



being discharged in a continuous stream into a covered container at 
the end of the apparatus. 

/. Determination of Sugar in Sugar Beets by Expression of Juice 

The determination of the sugar in sugar beets by polarization of the 
expressed juice was formerly quite common, but has now 
given place to more accurate methods of analysis. 

Assuming (as is incorrect) that the sugar, amides, 
albuminoids, salts, gums, and other water-soluble solids 
of the beet are in the same condition of solution within 
the beet as in the expressed juice, and letting M = the 
per cent of water-insoluble matter or " marc " and 100 
- M = the per cent of juice, then the sugar content (S) 
of the beet can be calculated from the polarization (P) 
of the expressed juice by the formula 

P(100 - M) 
100 

Example. The expressed juice of a sugar beet gave a 
polarization of 16.2 V. for the normal weight: the beet con- 
tained 4.6 per cent of marc. Required the per cent of sugar in 
the beet. 

o = 16.2 (100 - 4.6) 
100 



15.45 per cent. 




rection of 
boring in 
sampling 
sugar beets. 



The above method is, of course, equally applicable to ?' 1 
the analysis of sugar cane, fruits, and other succulent 
plant substances. 

Method of Expressing Juice. For expressing the 
juice from the pulp of sugar beets, sugar cane, etc., any 
suitable form of hand press may be used. The small hydraulic press 
shown in Fig. 131 is one of great efficiency and is a piece of apparatus 
almost indispensable in a sugar laboratory. 

The pulp to be pressed is placed in a strong sack inside the per- 
forated container C, and covered evenly with a heavy metal disk. By 
turning the wheel W the screw A is driven downward as far as possible 
upon the disk, thus squeezing out through the openings of C a con- 
siderable part of the juice, which escapes by the spout D into a can or 
other receptacle. The horizontal hydraulic screw B is then turned in- 
wards. This screw, operating by means of glycerol which fills the 
hollow base H, forces the piston E upwards and removes by vertical 
pressure a second fraction of juice. The final pressure, indicated by 



228 



SUGAR ANALYSIS 



the manometer M, can be raised to 300 atmospheres. The juice, as the 
pressure increases, is of gradually diminishing purity; it is important 
therefore that all the runnings should be well mixed before taking the 
sample for polarization. 



W 




Fig. 131. Laboratory hydraulic press for expressing juices. 

Determination of Marc. A determination of the insoluble cellular 
matter, or marc, is necessary before the per cent of sugar in plant sub- 
stances can be calculated from the polarization of the expressed juice. 
For rough purposes of estimation a constant percentage of 5 per cent 
or 4.75 per cent marc is sometimes assumed for the sugar beet and 10 
per cent or 12 per cent for the sugar cane. Such figures, however, 
have no exact value, as the percentage of cellular matter varies con- 
siderably according to the age of the plant, dryness of the season, and 
many other conditions. 

For the determination of marc 20 to 50 gms. of the finely divided 
pulp are digested with 200 to 500 c.c. of cold water for 30 minutes, 
and then filtered as dry as possible upon a piece of finely woven linen, 
using suction. The washing is repeated with successive portions of 
cold water until the filtrate, from color and taste, is judged to be free 
of extractive matter. The residue is then washed several times with 
hot distilled water, then, after pressing together, with 2 to 3 portions of 



METHODS OF SIMPLE POLARIZATION 229 

90 per cent alcohol, and finally with a little ether. After the ether has 
volatilized the marc is dried in an oven, gradually raising the tem- 
perature after a few hours to between 100 and 110 C. After cooling 
in a desiccator the residue, which is very hygroscopic, is rapidly weighed 
(preferably in a stoppered weighing bottle) and the weight taken as 
the amount of cellular matter or marc. For a determination of the 
organic cellular matter, the marc is incinerated and the percentage of 
ash deducted. 

The percentage of marc subtracted from 100 gives the percentage 
of juice. 

Where many determinations of marc have to be performed, a 
battery of small continuously operating percolators will effect a con- 
siderable saving of time. 

Errors of Expression Method. Several sources of error are 
involved in the determination of sugar in plant substances by analysis 
of the expressed juice. In the first place a considerable amount of 
juice, varying from 10 per cent to 30 per cent, according to the effi- 
ciency of the press, is not eliminated and this residual juice, containing 
a larger amount of albuminoids, pectin, etc., is of much lower purity 
than the part first expressed. This excess of impurities in the unex- 
pressed juice is washed out, however, in the marc determination. 
The polarization of the expressed juice is thus higher than that of the 
composite juice of the entire plant. (See under Distribution of Water, 
page 230.) 

A second source of error is the extraction during the marc determi- 
nation by the excessive amounts of cold water, but more especially 
by the hot water, alcohol, and ether of variable amounts of hemi- 
celluloses, wax, oil, and other substances which are, strictly speaking, 
not juice constituents and should therefore be included in the marc. 
The percentage of juice is thus estimated too high, and a plus error 
introduced in the calculation. Except for the disadvantage of loss of 
time in drying, the use of alcohol and ether as dehydrating agents 
should be omitted in the marc determination, and cold water alone be 
used for extracting. 

" Colloidal " or " Imbibition " Water. A third source of error to 
be mentioned is the much-debated question of " colloidal " or " imbibi- 
tion" water, by which is meant water, in a more or less hydrated 
form, in combination with hemicelluloses and other plant constituents. 
This imbibed water contains no sugar in solution, and, being expelled 
from the pulp upon drying, the percentage of sugar-containing juice is 
overestimated. v ^ , 



230 SUGAR ANALYSIS 

Heintz* showed, in 1874, when the air-dried and sugar-free marc of 
beets was placed in sugar solutions, that water was imbibed, thus leav- 
ing the sugar more concentrated and increasing the polarization. In 
the following experiments by Heintz air-dried beet marc, which had 
been washed completely free from sucrose, was treated 16 hours in a 
cool place with solutions containing a normal and half-normal weight 
of sucrose, in the proportion of 1 gm. marc to 20 c.c. of solution. 





Half normal 
weight. 


Normal weight 


PolH.riza.tion before marc treatment 


49 8 


99 6 


Polarization after marc treatment 


53 9 


104 6 









The observations of Heintz were verified in a different way by 
Scheibler.f The latter found that samples of sugar beets, whose ex- 
pressed juice polarized 14.5 had a marc "content of 4.71 per cent. The 
percentage of sugar in the beets according to the formula 

pqoo - 3Q 
~ 



would be 13.82. Scheibler found, however, by his method of alcoholic 
extraction a percentage of only 13.1 or a difference of 0.72 per cent. 
The percentage of sugar-containing juice in the beets, assuming that 
this juice is of the same polarization as the part expressed, is found by 

the formula, per cent juice = 100-^ = 100 -^-= 90.34 per cent, in 

Jr 14. o 

which p is the polarization of the beets by the extraction method and 
P the polarization of the expressed juice. The percentages of juice and 
marc being respectively 90.34 and 4.71, there is left a remainder of 
4.95 per cent, which Scheibler termed " colloidal " water. This method 
of estimation is based, however/ upon the assumption that the juice 
expressed is of the same composition as the combined juices of the 
beet, which is not exactly true.J 

Distribution of Water in Plant Tissues. The distribution of the 
water in plant tissues has such an important bearing upon certain 
problems of sugar analysis that a short discussion of the question may 
be introduced with profit at this point. 

* Z. analyt. Chem. (1874), 262. 
t Ibid. (1879), 176, 256. 

| For a very full discussion with bibliography of the subject of "colloidal" 
water see Rumpler, " Die Nichtzuckerstoffe der Ruben " (1898), pp. 1-13. 



METHODS OF SIMPLE POLARIZATION 231 

Fig. 132 shows a magnified cross section of a part of a sugar-cane 
stalk. The sugar-containing juice proper, represented by S (the 
vacuoles), constitutes the principal part of the cell contents in the 
thin-walled parenchyma or fundamental tissue, and includes the great- 
est part of the water in the cane. Lining the walls and permeating 




Fig. 132. Magnified cross-section of sugar-cane (protoplasmic 
lining P much intensified) . 

through these cells are thin layers and threads of protoplasmic matter 
P which contains a considerable amount of water, but is deficient in 
sugar. Running longitudinally through the stalk are large numbers of 
fibro vascular bundles whose ducts, D, are filled with water taken up 
from the soil. The water of these ducts may often be seen spurting 
from the end of a cane stalk as it passes between the rollers of a mill, 
and is found upon analysis to be almost free of sugar. Running parallel 
with the ducts are the sieve tubes T which carry in solution the prod- 
ucts of assimilation from the leaf to the stalk. The water of these 
tubes contains reducing sugars but is deficient in sucrose. The cellular 
walls of the parenchyma and fibrovascular tissues contain about 50 
per cent cellulose, 20 per cent xylan, 5 per cent araban and a remainder 
of lignin substances, all of which may hold a certain amount of water 
in the imbibed or colloidal form. 



232 



SUGAR ANALYSIS 



Variation in Composition of Juice from Different Mills. The press- 
ings from the first rollers or crusher of a cane mill consist mostly of 
the sugar-containing juice S (Fig. 132). The pressings from succeed- 
ing rollers, where the pressure is greater, contain more and more of 
the protoplasmic juice P and the juice from the ducts and tubes. The 
colloidal water of the cellular substance is of course not affected by 
the milling. 

The composition of the pressings from the different rollers of a 
cane mill is given in Table XLIV. 



TABLE XLIV 





First rollers. 


Second rollers. 


Third rollers. 


Water 


Per cent. 

84.64 


Per cent. 

85.40 


Per cent. 

85.35 


Sucrose 


12.93 


11.41 


11.30 


Reducing sugars 


1 54 


1 29 


1 23 


Ash . 


37 


0.58 


77 


Albuminoids. 


0.18 


0.50 


58 


Gums, acids, etc . . . 


0.34 


0.82 


0.77 










Total 


100.00 


100.00 


100.00 


Per cent extraction of cane 


64.50 


5.50 


2.13 



The pressed cane (bagasse) from the third rollers still contained 
over 60 per cent of water, corresponding to about 20 per cent of the 
total juice in the cane. If this residual juice could all be squeezed 
out by some inconceivable pressure, its sugar content would be much 
inferior to that of the pressings from the third rollers. It would of 
course be inaccurate to estimate the sugar content of the cane from the 
polarization of the first pressings; the same is also true, but to a much 
less degree, of the composite pressings of several mills. 

The impossibility of obtaining by pressure a true composite sample 
of the different juices of a plant, the difficulty of estimating the true 
content of marc, and the uncertain influence of the colloidal or imbibed 
water are the chief objections to the expression methods of sugar de- 
termination. 

II. Determination of Sugar in Sugar Beets by Extraction with Alcohol 

The method most accurate in principle for determining sugar in 
beets and other plant substances, is that of extraction. In this pro- 
cess the sugar is washed out from the pulp and the extract made up 
to volume and polarized. The errors due to uneven composition of 



METHODS OF SIMPLE POLARIZATION 233 

juices, faulty marc estimation, and colloidal water are thus com- 
pletely eliminated. 




Fig. 133. Apparatus for Scheibler's alcohol-extraction method. 

Scheibler's Alcohol-extraction Method. The solvent most gener- 
ally used for the extraction of sugar from beet pulp is 90 per cent ethyl 
alcohol. The original method of Scheibler* as modified by Sickelf is 
as follows: 

* Neue Zeitschrift, 2, 1, 17, 287; 3, 242. 

t Ibid. 2, 692. 



234 SUGAR ANALYSIS 

A normal (or double normal) weight of finely prepared pulp is 
weighed rapidly in a weighing dish, 3 c.c. of lead subacetate (6 c.c. for 
the double normal weight) are then added and thoroughly mixed with 
the pulp by means of a glass rod, adding at the same time 5 to 10 c.c. 
of 90 per cent alcohol. The pulp is then transferred to the extraction 
cylinder B of a Soxhlet extractor, of which Fig. 133 shows six in the 
form of a battery. The bottom of the extraction cylinder is covered 
with a clean wad D of felt or cotton; the pulp is washed in with 90 per 
cent alcohol, and pressed down so that its upper surface is below the 
upper bend of the siphon tube S. The top of the extraction vessel is 
then connected by means of a tight-fitting cork with the condensing 
tube C, and the bottom with the 100 c.c. flask F, which should contain 
about 75 c.c. of 90 per cent alcohol. 

The water in the bath is heated until the alcohol in the flask begins 
to boil vigorously, when the heat is regulated to this constant temper- 
ature. The vapor from the boiling alcohol passes upward through the 
side tube A and condensing in C drops back upon the pulp in B. As 
soon as the level of alcohol in B rises above the bend of the tube S, 
the alcoholic solution of sugar siphons mechanically into the flask F. 
The distilling and siphoning are continued until all the sugar is ex- 
tracted, which, according to the fineness of the pulp, 
usually requires from 1 to 2 hours. Immediately after 
the last siphoning the flask F is disconnected, cooled to 
room temperature, the volume completed to 100 c.c., 
and the solution mixed, filtered, and polarized. 

A form of extraction vessel devised by Miiller (Fig. 
134) permits the withdrawal of a small sample of liquid 
from the siphon tube for determining the completion of 
extraction. The opening at a is closed during operation 
with a stopper. To obtain the sample this stopper is 
removed, a few cubic centimeters of liquid are sucked 
up with a pipette and subjected to the a-naphthol test 
(page 341 ). If the test is positive, the stopper is replaced 
Fig. 134. Miil- and the extraction continued until the reagent gives no 

tlon Tsfxh" coloration - 

let's extractor" ^ n determining sugar by the Scheibler process of 
extraction special care must be exercised to prevent 
evaporation of alcohol during filtration. The funnel should be covered 
with a watch glass and the filtrate received in a cylinder or flask with 
narrow neck. The first 20 to 30 c.c. of the runnings should be dis- 
carded. The greater susceptibility of alcoholic sugar solutions to 




METHODS OF SIMPLE POLARIZATION 



235 



expansion and contraction with changes in heat and cold necessitates 
the maintenance of uniform temperature conditions during the polar- 
ization. The specific rotation of sucrose in ethyl alcohol is slightly 
higher (0.1 degree to 0.2 degree) than in water; but the difference is so 
small that it falls within the limits of experimental error. 

The method of alcoholic extraction gives results considerably lower 
than those calculated from the polarization of the expressed juice. The 
results of Scheibler previously quoted (page 230) show a difference of 
about 0.75 for the polarization of sugar beets. 

Some authorities prefer adding the lead subacetate to the alcoholic 
extract rather than to the pulp previous to extraction. This practice 
is attended, however, with some danger. One main object of adding 
the basic lead to the pulp is to neutralize any free acid which would 
otherwise invert some of the sucrose in the hot solution. In presence 
of alcohol, lead subacetate solution must be used in lowest possible 
amount owing to the danger of precipitating sucrose or of changing its 
specific rotation through formation of lead saccbarate. 

The alcoholic extraction method can be applied to the polarization 
of fruits and all other sugar-containing plant substances. With very 
dry materials the strength of the alcohol should 
be correspondingly reduced. With substances 
containing reducing sugars in large amount, it 
is desirable to omit the addition of lead sub- 
acetate, but when this is done the substance 
should be well mixed with powdered calcium 
carbonate to neutralize any free acid that might 
cause inversion. 

II. Determination of Sugar in Plant Substances 
by Extraction with Water 

Water is sometimes used instead of alcohol 
in extracting sugar for the polarization of plant 
substances. In such cases a process of per- 
colation must be used in place of distillation Fig. 135. Section of Zam- 
owing to the danger of decomposition through 
the prolonged boiling of aqueous extracts. As 
an example of the water extraction process the Zamaron* method for 
determining sugar in sugar cane is given. 

Zamaron's Water-extraction Apparatus. The Zamaron extraction 
apparatus (Figs. 135 and 136) consists of a cylindrical copper vessel 
* Sidersky's " Manuel," p. 261. 




aron's hot-water extrac- 
tion apparatus. 



236 



SUGAR ANALYSIS 



V provided at the bottom with a small cock C. A basket B of per- 
forated copper, provided with a tripod support, fits loosely within this 
copper vessel; 100 gms. of the finely divided pulp are transferred to 
the basket, and 200 c.c. of hot water poured in, the pulp being pressed 




Fig. 136. Battery of Zamaron's hot-water extractors. 



down beneath the surface of the liquid by means of the plunger P. 
The contents of the vessel are then boiled for 10 minutes, after which 
the flame is turned down, the cock opened, and the hot solution drawn 
off into the 1000-c.c. graduated flask F, as much as possible of the 
liquid being pressed out by means of the plunger. The cock is then 
closed and the process repeated with a second portion of 150 c.c. water. 
The process is continued 6 times, making altogether about 950 to 975 c.c. 
of extract. After cooling and adding a few cubic centimeters of lead 
subacetate, the contents of the flask are made to 1000 c.c., shaken, 
filtered, and polarized in a 400-mm. tube. The reading multiplied by 
1.3 gives the polarization (degrees Ventzke) of the sugar cane. 

The principal objection, which has been brought against the Zam- 
aron process, is the danger of incomplete extraction. Some idea of the 
probable magnitude of this error may be formed from the following 
consideration : 



METHODS OF SIMPLE POLARIZATION 



237 



Suppose a sugar cane to contain 18 per cent of sucrose; suppose 
also that 6 extractions of the pulp are made and that one- third of the 
liquid is retained by the fiber after each extraction. If the sugar is 
evenly diffused through all parts of the liquid at the end of each 10 
minutes boiling, as is no doubt very nearly true, there would be the fol- 
lowing percentages of sugar removed at each extraction. 





Percentage 
removed of 
total sugar. 


Percentage of 
sugar removed 
per 100 of cane. 


First extraction 


66 66 


12 00 


Second extraction 


22 22 


4 00 


Third extraction 


7 41 


1 33 


Fourth extraction 


2 47 


44 


Fifth extraction. . . 


82 


15 


Sixth extraction. . . . 


27 


05 








Amount extracted 
Amount unextracted 


99.85 
0.15 


17.97 
0.03 



It is seen that the residual sugar left after 6 extractions can be 
only very slight. In order to reduce the possibility of error through 
incomplete extraction Fribourg* recommends that only 50 gms. of 
pulp be taken for analysis. This, however, while halving the errors of 
extraction, necessitates a doubling of any error in the polariscope 
reading. 

Another source of error, in the method of hot water extraction as 
described, is the danger of inversion of sucrose through the natural 
acidity of the pulp. One method of preventing this is to mix with 
the pulp previous to extraction finely powdered calcium carbonate. 
Another method* is to employ very dilute milk of lime water for the ex- 
traction. The presence of minute quantities of free alkali does not 
affect the determination of sucrose; a danger exists, however, in the 
action of hot alkaline solutions (even where very dilute) in modifying 
or destroying reducing sugars. Careful neutralization of the free acid 
in the pulp with lime water, or dilute sodium hydroxide, would eliminate 
the risk of inversion without serious danger of affecting the reducing 
sugars. 

Another objection to the method of hot-water extraction is the 

solution of optically active dextrins, gums, and hemicelluloses. These 

substances introduce at times a considerable error in the polarimetric 

determination of sugars in aqueous plant extracts. The error does 

* Fribourg's "Analyse chimique," p. 223. 



238 SUGAR ANALYSIS 

not exist in the alcohol-extraction method, owing to the insolubility of 
dextrinoid substances in ethyl alcohol. 

///. Determination of Sugar in Sugar Beets by Methods of Digestion 

The method of alcoholic extraction, although the most accurate 
and scientifically perfect, is not the best from a practical standpoint 
on account of the long period of time necessary for extraction, and also 
because of the rather fragile nature of the extraction apparatus. For 
the rapid determination of sucrose in sugar beets some one of the num- 
erous digestion processes is usually followed. 

The digestion method may be regarded in principle as a combination 
of the extraction and juice-expression methods. A weighed amount of 
pulp is digested with 5 to 6 times its volume of alcohol or water. After 
the complete diffusion of the sugar through the liquid, f the solution 
is made up to volume, allowing for the space occupied by insoluble 
matter, and then filtered and polarized. 

Rapp-Degener Alcohol-digestion Method. The first process of 
digestion employed alcohol, and is known as the Rapp-Degener * method. 
The double normal weight of fine beet pulp is transferred to a 201.2-c.c. 
flask (the extra 1.2 c.c. being the estimated volume of the insoluble 
cellular matter in 52. gms. of pulp). The forms of flask shown in 
Fig. 137 are convenient for the purpose. Three to four c.c. of lead- 
subacetate solution are mixed with the pulp and then about 150 c.c. of 
90 per cent alcohol added. The flask is closed 
with a stopper containing a condensing tube 
and placed in a hot- water bath. The alcohol 
is gently boiled for 20 minutes, when diffusion 
of the sugar through the solution may be con- 
sidered complete. The tube and stopper are 
rinsed into the flask and the volume completed 
nearly to the mark with 90 per cent alcohol. 
Fig. 137. -Flasks for alco- The flagk ig in laced in the hot-water bath 
hohc .digestion of beet - - , . . 

pul for 1 to 2 minutes, to secure even mixing of 

the contents and expulsion of air bubbles, and 

then allowed to cool slowly in the air for J hour. The liquid is then 
brought to room temperature and the volume completed to 201.2 c.c. 
with 90 per cent alcohol. The solution is then mixed, filtered and 
polarized in a 200-mm. tube, using the necessary precautions to prevent 
evaporation and changes in temperature. 

* Z. Ver. Deut. Zuckerind., 32, 514, 786. 





METHODS OF SIMPLE POLARIZATION 



239 



The employment of alcohol in analytical work is expensive; it was 
also found that with any coarse particles of pulp the diffusion of sugar 
through the alcohol was considerably retarded. Pellet* was accord- 
ingly induced in 1887 to devise a method for determining sugar in beets 
in which water was used for digesting instead of alcohol. The Pellet 
method may be carried out with either hot or cold water. 

Pellet's Cold- water-digestion Process. Twenty six gms. of finely 
divided pulp are transferred by means of a jet of water into a 200.6-c.c. 
flask (the extra 0.6 c.c. being the estimated volume of the insoluble 
marc in 26 gms. of pulp); 5 to 6 c.c. of lead-subacetate solution are 
then added and sufficient water to fill the flask about two-thirds. After 
mixing, the flask is allowed to stand for 20 to 30 minutes to permit 






Fig. 138. " Sans-Pareille " press for preparing finely divided pulp. The substance, 
which is placed in the cell C, is forced in a semiliquid condition by the piston P 
through the fine openings at the bottom into a container underneath; the latter 
also receives any overflow of juice which escapes by the outlet T. 

diffusion of sugar and allow enclosed air bubbles to escape. Water is 
then added nearly to the mark, any foam destroyed with a drop of 
ether, and the volume completed to 200.6 c.c. The solution is well 
mixed, filtered, and polarized in a 400-mm. tube; the scale reading 
gives without correction the polarization of the beet. 

With pulp of extreme fineness, such as is obtained with the "Sans- 
Pareille" press (Fig. 138), the diffusion of sugar from pulp to water 
becomes almost instantaneous, and the solution can be completed to 
volume as soon as air bubbles have arisen. The time of analysis is 
thus considerably lessened. 

* Deut. Zuckerind. (1888), 1229; (1889), 531. 



240 SUGAR ANALYSIS 

Pellet's Hot- water-digestion Process. If apparatus is not avail- 
able for obtaining pulp of suitable fineness, hot water should be used to 
promote the diffusion of sugar from the coarser particles of pulp. 
Twenty-six grams of pulp, mixed with 6 c.c. of lead-subacetate solution, 
are washed into a 200.6 c.c. flask, water is added with shaking until 
the volume is almost up to the mark, and the flask heated in a boiling 
water bath for J to 1 hour, according to the fineness of the pulp. The 
flask is then immersed in cold water; as soon as the contents are of 
room temperature, the volume is completed to the mark. The remain- 
der of the process follows as under cold-water digestion. 

Kriiger's Cold- water-digestion Process. Kriiger,* in 1896, de- 
vised a water-digestion process, an interesting feature of which is that 
the use of normal weights and of volumetric flasks is entirely dis- 
pensed with. The principle of the method may be understood from 
the following: 

The weight of juice per 26 gms. in an average sugar beet of 5 per 
cent marc content is 26 X 0.95 = 24.7 gms. The specific gravity of 
the average beet juice is very nearly 1.07, so that the volume of juice 
in a normal weight (26 gms.) of pulp is 24.7 gms. -r- 1.07 = 23.08 c.c. 
The amount of water necessary to complete this volume of juice to 
100 c.c. is therefore 100 - 23.08 = 76.92 c.c. The ratio of normal 
weight to volume of added water is then 26 gms. : 76.92 c.c. = 1 gm. : 
2.958 c.c., or in round numbers 1 gm. : 3 c.c. The addition, therefore, 
of water in the proportion of 3 c.c. to every 1 gm. of pulp yields a 
solution whose polarization in a 200-mm. tube will give the approximate 
sugar content of the beet. 

The automatic pipette (Figs. 139, 140) for rapidly measuring water 
and lead solution is an essential feature of the Kriiger process. The 
pipette is prepared in several sizes for approximate double-normal, 
normal, half-normal, and quarter-normal weights of pulp (i.e., approxi- 
mately 50, 25, 12, and 6 gms.), the smaller sizes being used in polarizing 
mother beets, where the quantities of pulp obtained by the Keil sam- 
pler (p. 226) are small (8 to 14 gms.). The pipette, which is fastened 
to a fixed support S (Fig. 140), is provided at opposite ends with the 
three-way cocks C and C', the movements of which are controlled by 
the double lever L. The lower inlet of the pipette is connected by the 
tube A to the vessel V which contains the "lead water" (9 vols. of 
water to 1 vol. of lead-subacetate solution). The upper outlet which 
permits the escape of air is connected with the upright tube B. By 
raising L to the stop c (Fig. 139) the pipette is filled with "lead water," 
* Deut. Zuckerind. (1896), 2434. 



METHODS OF SIMPLE POLARIZATION 



241 



any overflow passing into the tube B. Upon dropping L to the stop d, 
the cocks are both. reversed, air entering through/, and the contents 
of the pipette being discharged through e into the metal weighing dish 
D, which contains the weighed sample of pulp. 





d 

Fig. 139 Fig. 140 

Kriiger's automatic pipette for sugar beet analysis. 

The weight of pulp corresponding to each pipette is determined by 
calibration with water, as in the following example. The weight of 
distilled water discharged by a Kriiger pipette at 20 C. was found to 
be 78.38 gms. The volume of the pipette in true cubic centimeters is 
then 78.38 -v- 0.9972 = 78.6 c.c. 78.6 ^ 3 = 26.2 gms., the weight of 
beet pulp corresponding to the pipette. 

After mixing the pulp and " lead water " the weighing dish is 
covered and the contents allowed to remain for 20 to 30 minutes. 
The solution is then well stirred, filtered, and polarized in a 200-mm. 
tube. 



242 



SUGAR ANALYSIS 



The Kriiger method, while not claiming extreme accuracy, is suffi- 
ciently exact for many purposes of analysis. On account of its sim- 
plicity and rapidity the method has been widely used in such places as 
beet-seed nurseries, depots for purchase of beets, etc., where large num- 
bers of samples have to be polarized with the least possible loss of time. 

Sachs-Le Docte Process of Water Digestion. The occlusion of 
air bubbles by pulp and the uncertainty of knowing whether such 
bubbles are completely absent before making up to volume have been 
the principal objections against the original Pellet process of digestion. 
This error does not occur in the Kriiger method, where the volume of 
solution is established independent of any occluded air. The necessity 
of employing irregular weights for each individual pipette and the use 
of insufficient water for the complete diffusion of the sugar during the 
cold digestion have been raised on the other hand as objections against 
the Krtiger method. Sachs * and Le Docte f have met these difficulties 
by always taking the regular normal weight (26 gms.) of pulp for 
analysis and adding a constant volume (177 c.c.) of water and lead 
subacetate so that the final estimated volume of solution, regardless 
of insoluble marc or occluded air, is always 200 c.c. 

The constant-volume figure 177 c.c. in the Sachs-Le Docte process 
is derived from the following consideration. Sachs assumes as the 
average marc and juice content of the sugar beet 4.75 per cent and 
95.25 per cent respectively. For the normal weight (26 gms.) of pulp 
there would then be 26 gms. X .9525 = 24.765 gms. juice. The aver- 
age sugar content and density of juices from beets of different richness 
are given in the following table together with the calculated volume of 
juice (24.765 -f- sp. gr.), the volume of lead-water solution (200 c.c. less 
the volume of juice) and the polarization error resulting from use of the 
constant volume 177 c.c. 

TABLE XLV 



Sugar in beet. 


Sugar in 
juice. 


Brix of 
juice. 


Specific 
gravity of 
juice. 


Volume of 
juice. 


Volume of 
lead-water 
solution. 


Calculated 
polariza- 
tion.* 


Polariza- 
tion error. 


Per cent. 
12 

13 
14 
15 
16 
17 


Per cent. 
12.59 
13.65 

14.70 
15.75 
16.80 
17.85 


14.86 
15.82 
16.82 
17.86 
18.92 
20.00 


1.0609 
1.0651 
1.0694 
1.0740 
1.0787 
1.0835 


c.c. 

23.34 
23.25 
23.16 
23.06 
22.96 
22.86 


176.66 
176.75 
176.84 
176.94 
177.04 
177.14 


11.979 
12.984 
13.988 
14.995 
16.003 
17.012 


-0.021 
-0.016 
-0.012 
-0.005 
+0.003 
+0.012 



* Calculated polarization 



sugar in beet X 200 
volume of juice -f- 177 



* Z. Ver. Deut. Zuckerind. (1906), 56, 918. f Ibid. (1906), 66, 924. 



METHODS OF SIMPLE POLARIZATION 



243 



It is seen that by use of the constant 
volume 177 c.c. the calculated polariza- 
tion error is too small to be detected 
upon the saccharimeter. 

The constant-volume pipette em- 
ployed in the Sachs-Le Docte process is 
shown in Fig. 141. A three-way cock 
K at the bottom serves for the inlet of 
lead reagent and water at B and C and 
for the delivery of the 177 c.c. of mixed 
solution through D. The cap A at the 
top, which receives the overflow, is con- 
nected with a waste bottle. Instead of 
drawing in the lead reagent and water 
separately, a single " lead- water " solu- 
tion of proper dilution may be used. 
One of the cock connections may thus 
be dispensed with. By raising or lower- 
ing the capillary tube h upon its support 
at H the capacity of the pipette is easily 
adjusted to exactly 177 c.c. 

The method of operation is similar 
to that in the Kriiger process. Weigh 
26 gms. of pulp in one of the tared metal 
beakers; the latter are of about 250-c.c. 
capacity and are provided with a tight- 
ficting cover of rubber; add 177 c.c. of 
water containing 5 to 6 c.c. of lead sub- 
acetate solution (of about 30 Be.) and 
shake thoroughly. Filter, add a drop 
of glacial acetic acid to the filtrate, and 
polarize in a 400-mm. tube. The scale 
reading gives the polarization of the 
beet. Where many analyses have to 
be performed a large number of metal 
beakers are used, all of which are 
counterpoised against the same weight. 

If the particles of pulp are coarse Fig 141i _ Sach8 . Le Docte 
the Sachs-Le Docte process should be mat j c p i p ette for sugar 
carried out by hot digestion.* The analysis. 

* Sucrerie Beige, Oct. 15, 1908. Bull, assoc. chim. sucr. dist., 27, 180. 




auto- 
beet 



244 SUGAR ANALYSIS 

method of operation is similar to that just described, except that the 
metal beakers, after addition of the 177 c.c. of lead- water solution to the 
pulp, are each covered with a special pneumatic cap of rubber which 
prevents any loss by evaporation. Fig. 142 shows a water bath for 
the Sachs-Le Docte hot-digestion process. The metal beakers are 
placed for 30 minutes in a water bath heated to 80 C. After cooling 
the beakers are well shaken, when the contents are filtered and polarized 
in the usual way. 

Herzfeld* has slightly modified the Sachs-Le Docte process for hot 
digestion. The pulp is weighed into small copper cans, 11 cm. high, 




Fig. 142. Sachs-Le Docte bath for hot-water digestion. 

6 cm. body diameter, and 4 cm. mouth diameter. The cans are closed 
during digestion with rubber stoppers or with good corks covered with 
tinfoil. The blowing out of stoppers during digestion has been raised 
as an objection against the Herzfeld modification. Stanek and Urban f 
recommend the use of cans provided with a spring cap and rubber 
gasket.J 

A comparison of sugar determinations in beets by the Sachs- 
Le Docte cold- and hot-digestion methods and by the Kruger method is 
given in the following table. The results are the average of many de- 
terminations reported by Herzfeld.* 

* Z. Ver. Deut. Zuckerind., 69, 627. 

t Z. Zuckerind. Bohmen, 34, 625. 

t A very full description of methods for analyzing sugar beets and a complete bib- 
liography of the subject from 1839 to 1907 has been compiled by Bryan (Bull. 146, 
U. S. Bur. of Chem.) 



METHODS OF SIMPLE POLARIZATION 



245 





Sachs- Le Docte method. 


Kriiger method, 
cold digestion. 


Cold digestion. 


Hot digestion. 


Average 14 analyses 
Average 19 analyses. . . . 


Per cent. 
16.66 
15.91 


Per cent. 
16.87 
16.28 


Per cent. 
16.56 

16.12 



Errors of Digestion Methods 

Solution of Dextrorotatory Gums. It is noted in the preceding table 
that the hot-digestion gives from 0.2 to 0.3 higher than the cold-digestion 
methods. This excess is no doubt due in large part to a higher extrac- 
tion of sucrose from the coarser particles of pulp. Some chemists, 
however, attribute a part of the excess to a solution of dextrorotatory 
hemicelluloses (parapectin, metapectin, etc.) which are dissolved by the 
hot water from the pulp. According to Pellet these substances are 
completely precipitated by the lead-subacetate solution, when this 
reagent is of proper strength (about 30 degrees Be.) and used in proper 
amount (5 to 6 c.c. per 26 gms. of pulp). To insure complete precipi- 
tation of all dextrorotatory gums some authorities advise using 7 
or 8 c.c. of. basic-lead solution. Herzfeld,* however, has shown that 
lead subacetate in hot solution forms a levorotatory combination 
with certain constituents of beet pulp and is opposed to the use of 
more than 5 c.c. of the reagent per 26 gms. pulp for hot-water 
digestion. 

The extraction of high polarizing dextrorotatory gums is very liable 
to occur, even with cold-water digestion, in the case of sugar beets 
which are unripe, frost-bitten, diseased, or otherwise abnormal. Under 
such circumstances the method of extraction with alcohol, in which 
the gums are insoluble, should be employed. 

Solution of Asparagine. Another constituent of sugar beets which 
may introduce an error in the polarization is asparagine. Degenerf 
has shown that asparagine, which in neutral solutions is slightly levo- 
rotatory ([a] D = 5.2), becomes strongly dextrorotatory ([O\D = +61.76 
to +69.10) in presence of 10 per cent lead-subacetate solution, every 
0.1 per cent asparagine polarizing about the same as every 0.1 per cent 
sucrose. To obviate this error the French chemists add a drop of 
glacial acetic acid to the filtered solution from the aqueous digestion 
before polarizing. Asparagine is dissolved only 1 part in 290 parts of 

* Z. Ver. Deut. Zuckerind., 59, 627. 
f Deut. Zuckertnd. (1897), 65. 



246 SUGAR ANALYSIS 

80 per cent alcohol and this solubility is diminished by the addition of 
lead subacetate. The asparagine error is therefore negligible in the 
methods of alcoholic extraction or digestion. 

Variation in Marc Content. Among other sources of error peculiar 
to the digestion methods may be mentioned the difference in quantity 
and volume of insoluble cellular matter in the normal weight of pulp. 
This volume is in fact variously given by different authorities as 0.6 c.c.,* 
0.75 c.c.,f 1.35 c.c.,{ and the digestion flasks have been correspondingly 
graduated at 200.6 c.c., 200.75 c.c., and 201.35 c.c. Pellet has devised 
a special digestion flask with 5 graduations at 200.0 c.c., 200.5 c.c., 
200.75 c.c., 201.0 c.c., and 201.5 c.c., so that the chemist may vary the 
volume according to the weight and character of pulp. While the 
volume most generally prescribed is 200.6 c.c. for 26 gms. of pulp, it 
is evident that this figure must be greater for wilted beets and less for 
unripe beets. In the same way the volume of lead-water solution in 
the Sachs-Le Docte process would be greater or less than 177 c.c. The 
polarization errors due to normal variations from the average of 4.75 per 
cent marc are considerably less than 0.1, but in extreme cases of wilted 
or watery beets the alcoholic extraction method should be used as a 
control. 

The error due to imbibition or colloidal water (p. 229) has also 
been raised against the digestion methods. The average difference 
between the expression and extraction methods was found by Scheibler 
to be about 0.75 per cent, which difference represents the combined 
influence of unequal composition of juice and of the colloidal water. 
In the digestion methods the 23 c.c. of juice is diluted to 200 c.c. or 
nearly ninefold, so that the combined errors of the juice methods are 
reduced to less than 0.1. In the digestion methods the error due 
to unequal composition of juice is largely eliminated; the residual 
error due to the so-called colloidal water must therefore be very 
small. 

The agreement between the aqueous digestion and alcoholic extrac- 
tion methods upon normal sugar beets is usually very close. As to 
which of the water-digestion methods is preferable it may be said that 
if apparatus is available for securing pulp of extreme fineness the cold- 
water digestion is upon the whole less open to error. But for pulp of 
coarse or uneven character hot-water digestion should be used to insure 
complete extraction. 

* Friihling's " Anleitung," 209. 

t Fribourg's " Analyse chimique," 253. 

t Sidersky's "Manuel," 241. 



METHODS OF SIMPLE POLARIZATION 247 

POLARIZATION OF PLANT SUBSTANCES CONTAINING BUT Low PER- 
CENTAGES OF SUGAR 

The methods previously described may be applied with minor 
modifications to the polarization of plant substances containing but 
low percentages of sugar. The polarization of spent sugar-beet chips 
and sugar-cane bagasse may serve as illustrations of the methods. 

Polarization of Spent Beet Chips by the Expression Method. 
While the water circulating through the diffusion battery removes 
most of the sugar from the beet chips, a small amount of sugar always 
remains unextracted; this residual sugar occurs for the most part 
within the uncrushed cells of the beet. It is necessary, therefore, in 
squeezing out the water from diffusion chips to apply extreme pressure, 
in order to secure the maximum quantity of residual sugar. A polari- 
zation of the expressed diffusion water and a determination of its 
amount are sufficient for the calculation. 

Example. 100 c.c. of the diffusion water pressed from a sample of spent 
beet chips were clarified with 2 c.c. of lead-subacetate solution and the volume 
completed to 110 c.c. The filtered solution gave a polarization of 2.0 V. in a 
400-mm. tube. The water content of the chips, upon drying 10 gms. at 100 to 
110 C. to constant weight, was 90.5 per cent. 

The polarization corrected for the dilution is 2.0 X 1.1 = 2.2 V. Calling 
the sp. gr. of the waste diffusion water 1.000 (which can be done without serious 
error) the polarization of a normal weight would be (26.00 X 2.2) -5- 100 = 
0.572 V., or for a 200-mm. tube 0.29 V. The polarization of the spent chips 
would then be (90.5 X 0.29) ^ 100 = 0.26. 

Polarization of Dried Beet Chips by the Alcoholic Digestion and 
Extraction Method. Dried sugar-beet chips have frequently under- 
gone a change in composition through formation of water-soluble 
optically active gums at the high temperature of drying. The aqueous 
digestion method may then give a polarization different from the true 
sucrose content. In such cases it is recommended to use the alcoholic 
digestion and extraction method of Herzfeld.* 

A half normal weight of the finely ground dry chips is digested in 
a hot-water bath with 50 to 60 c.c. of 60 per cent alcohol, adding 
3 to 5 c.c. of lead-subacetate solution, for 30 minutes. The contents 
of the digestion flask are then transferred by means of a little 60 per 
cent alcohol to a Soxhlet extractor and extracted under reduced pres- 
sure for 5 to 6 hours (see Fig. 143). The alcoholic extract is then made 
up to 100 c.c., filtered, and polarized in a 400-mm. tube. 
* Z. Ver. Deut. Zuckerind., 69, 627. 



248 



SUGAR ANALYSIS 



Polarization of Sugar-cane Bagasse by Hot-water Extraction. 

The hot-water-extraction method of Zamaron may be employed upon 
bagasse in the same manner as described for sugar cane. Owing, how- 
ever, to the much larger amount of cellular matter in bagasse only 

50 gms. are taken for extraction. The ex- 
tract is made up to 1000 c.c. and polarized 
in a 400-mm. tube. The reading multiplied 
by 2.6 gives the polarization of the bagasse. 
Extraction waters of very low sugar con- 
tent are sometimes concentrated before 
polarization. Five hundred cubic centi- 
meters of the neutralized solution are evapo- 
rated to somewhat less than the desired 
volume, and then made up to 100 c.c. or 
250 c.c. for polarization. The saccharimeter 
reading is divided by 5 or 2 to obtain the 
polarization of the extract. 

Polarization of Sugar-cane Bagasse by 
Hot-water Digestion. Bagasse is also 
polarized by the method of hot- water diges- 
tion, in which case, however, it is necessary 
to know the percentage of fiber. The de- 
termination may be made by the methods 
Fig. 143.-Her Z feld's appa- of the Hawaiian chemists.* 
ratus for alcoholic extraction . t . . 

under reduced pressure. Determination of Fiber in Bagasse. - One 

hundred grams of bagasse are placed in a 

strong linen bag, and the juice pressed out with an hydraulic press. 
The sample is then treated with cold running water for two minutes, 
and again pressed, the two operations being repeated alternately five 
times. The bag is then placed in an air bath at 125 C. for half an hour, 
after which the fiber is removed from the bag and dried in a shallow 
dish for four hours at the same temperature. When an hydraulic press 
is not available, the sample may be treated in cold running water for 
12 hours and dried as above described. 

Digestion of Bagasse. Fifty grams of bagasse are weighed in a 
tared flask; 500 c.c. of water containing 2 c.c. of 5 per cent sodium 
carbonate are added, and the flask connected with a vertical condenser. 
The solution is boiled gently for one hour, the flask being shaken 
thoroughly every 15 minutes. After cooling the flask is re weighed, and 
the weight of contents determined. The weight of contents multi- 
* Hawaiian Planters' Record, 3, 317. 




METHODS OF SIMPLE POLARIZATION 249 

plied by 2 gives the weight (W) of fiber and solution corresponding to 
100 gms. of bagasse. Letting F = the per cent fiber in the bagasse, 
W F = the weight of solution corresponding to 100 gms. of bagasse. 
The aqueous extract obtained by the hot digestion is squeezed out; 
99 c.c. of the solution are made up to 100 c.c. with lead-subacetate 
reagent, filtered, and polarized in a 400-mm. tube. The polarization 

100 jP 

(P) corrected for dilution is . , and this reduced to a normal weight 

26 100 P 26 P 
of extract is -r X -7^ = ~7^r , which value for a 200-mm. tube be- 

iuu yy yy 
13 P 

comes QQ The polarization of the bagasse is then found by the 
yy 

13 P (W- F) P(W - F) 
formula -99- - . ^ - 






II 



POLARIZATION OF SUBSTANCES CONTAINING INSOLUBLE MINERAL 

MATTER 

The polarization of substances containing insoluble mineral matter 
can in general be carried out by the methods of extraction or digestion 
previously described. Certain classes of products, however, such as 
carbonatation filter-press cake may contain sugar in the form of in- 
soluble saccharates, and in such cases special methods of treatment are 
required. As examples of methods to be employed several processes 
for the polarization of filter-press cake will be described. 

Polarization of Filter-press Cake Free from Saccharate. If 
saccharate-free press cake be triturated with a known quantity of 
water and the filtered extract polarized, the polarization of the cake 
may be calculated very closely, provided its moisture content has been 
determined. 

Example. 50 gms. of press cake were ground in a mortar with 200 c.c. of 
water. The solution (which should not be alkaline) was then clarified with a 
little dry lead subacetate and polarized in a 400-mm. tube. A reading of 5.2 V. 
was obtained. The moisture content of the cake, determined by drying 10 gms. 
in a hot-water bath to constant weight, was 45.6 per cent. It is desired to 
know the polarization of the cake. 

The weight of water in the 50 gms. of cake is 50 X 0.456 = 22.8 gms. 
The total volume of liquid (disregarding the slight increase in volume through 
solution of sugar) is then 200 + 22.8 = 222.8 c.c. The polarization of the 
solution reduced to a normal weight of 26 gms. to 100 c.c. (calling the sp. gr. 
1.000, which may be done without serious error) is (5.2 X 26) -s- 100 = 1.35 V., 
which for a 200-mm. tube is 0.68 V., or 0.68 gms. of sucrose in 100 c.c. of 



250 SUGAR ANALYSIS 

solution. This corrected to 222.8 c.c. = 0.68 X 2.228 = 1.52, the grams of 
sucrose in 50 gms. of cake; 1.52 X 2 = 3.04, the polarization or percentage 
of sucrose in the cake, if no other optically active substances are present. 

The above method of calculation is sufficiently exact for substances 
of low polarization. When the polarization is high, however, neglect 
of the increase in volume through solution of sugar and of the change 
in specific gravity introduces a considerable error. In such cases the 
polarization should be determined by some method of extraction. 

In sugar-house practice the determination of moisture in the press 
cake is usually dispensed with, it being assumed that the volume of in- 
soluble matter in 26 gms. of cake is 4 c.c. The normal weight of cake 
is then made up to 104 c.c.; or, if a 100-c.c. flask be used, 25 gms. of 
cake, when triturated, clarified with lead solution, and the liquid made 
up to volume, will give the polarization (104 : 26 : : 100 : 25). In 
practice 50 gms. of cake are generally weighed out and the volume 
made up to 200 c.c. 

In the previous example if the 50 gms. of cake had been made up with 
water to 200 c.c., there would be 192.3 c.c. of solution (allowing 4 c.c. for volume 
of insoluble matter in 26 gms.). The polarization for 222.8 c.c. of solution was 
5.2 V., therefore 192.3 : 5.2 : : 222.8 : 6.02, the calculated polarization of the 
cake for a 400-mm. tube. This for a 200-mm. tube would be 3.01, which is 
only 0.03 V. lower than the result previously found. 

Polarization of Filter-^press Cake Containing Saccharate. When 
filter-press cake contains insoluble saccharates, the sugar must be 
liberated from combination before the solution to be polarized is made 
up to volume. Several methods have been followed for accomplishing 
this result. 

Decomposition of Saccharate by Means of Acetic Acid. The 50 gms. 
of press cake, after transferring with water to a 200-c.c. flask, are heated 
to boiling, and acetic acid added drop by drop until all free alkali is 
neutralized. The solution is then cooled, clarified, made up to volume, 
filtered, and polarized as previously described. 

Decomposition of Saccharate by Means of Carbon Dioxide. The 
method is practically the same as that just described, except that a 
stream of carbon dioxide led into the solution is used for decomposing 
the saccharate, instead of acetic acid. 

The frothing, caused by evolution of carbon dioxide, is the principal 
objection against the acetic-acid method, and the decomposition by 
means^of carbon dioxide usually requires considerable time. Methods 
have been devised, therefore, to decompose insoluble saccharates in other 
ways. One of the most common of such methods is the following: 



METHODS OF SIMPLE POLARIZATION 251 

Decomposition of Saccharate by Means of Ammonium Nitrate. 
The saccharates of calcium are quickly decomposed by ammonium 
nitrate with formation of free sugar, calcium nitrate, and ammonia. 
The reaction for monocalcium saccharate is 

Ci 2 H 22 OnCaO + 2 NH 4 N0 3 + H 2 = dsH^On + Ca(N0 3 ) 2 + 2 NH 4 OH. 

Saccharate Sucrose 

In carrying out the process 50 gms. of press cake are ground up 
with 15 gms. of ammonium nitrate and 100 c.c. of cold distilled water. 
The mixture is then washed into a 200-c.c. flask, clarified with a little 
lead-acetate solution, made up to volume, and polarized in the usual 
way. 

An objection against the ammonium-nitrate method is the libera- 
tion of free ammonia, which in presence of the lead-clarifying agent 
may precipitate a part of the sucrose as lead saccharate. The free 
ammonia in some cases causes a darkening of the solution; contact 
with the brass fittings of polariscope tubes may also color the ammo- 
niacal solution blue. Care should be exercised, therefore, to prevent 
contact of the solution with copper or brass during the analysis. 

Decomposition of Saccharate by Means of Zinc Nitrate. In order to 
eliminate the formation of free alkali Stanek* has proposed the em- 
ployment of zinc nitrate for decomposing the saccharate. The reaction 
proceeds as follows: 

CisH^OnCaO + Zn(N0 3 ) 2 + H 2 = Ci 2 H 22 On + Ca(N0 3 ) 2 + Zn(OH) 2 

Monosaccharate Zinc nitrate Sucrose Calcium nitrate Zinc hydroxide. 

The precipitated zinc hydroxide is removed with the insoluble mineral 
matter of the cake and a perfectly neutral filtrate is obtained. 

In carrying out the process a double normal weight (52 gms.) of 
press cake is thoroughly triturated with 100 c.c. of water; a few drops 
of phenolphthalein indicator are then added, and a neutral solution of 
zinc nitrate run in until the red color is just discharged. The volume 
is then completed to 210 c.c. (10 c.c. being allowed for the volume of 
insoluble cake and zinc hydroxide), and the solution filtered and polarized. 

The methods, which have been described for polarizing products of 
the cane- and beet-sugar industry, may be applied equally well to the 
polarization of other sucrose-containing substances, such as maple and 
sorghum products, jellies, preserves, confections, etc. The same 
methods may also be applied to the polarization of substances which 
contain other sugars than sucrose, the only change necessary to make 

* Z. Zuckerind. Bohmen, 34, 161. 



252 



SUGAR ANALYSIS 



being in the constant for the normal weight. As an example of the 
application of saccharimetric methods to other sugars besides sucrose, 
the determination of milk sugar in milk is selected. 

SACCHARIMETRIC DETERMINATION OF LACTOSE 

Polarization of Milk.* The normal weight of lactose for a saccha- 
rimeter with the Ventzke sugar scale may be taken ^is 32.9 gms. (see 
p. 197). Owing to the low percentage of lactose in milk (2 to 8 per 
cent) it is best to employ double the normal weight, and, as it is more 
convenient to measure the milk, tables have been prepared which give 
the volumes of milk corresponding to multiples of the normal weights 
for different saccharimeters. The following table gives the volumes 
of milk for 65.8 gms. which correspond to different specific gravities. 

TABLE XLVI 
Giving the Volumes of Milk Corresponding to a Lactose Double Normal Weight 



Specific gravitv of 
milk. 


Volume of milk for a 
Lactose double normal 
weight (Ventzke scale). 


.024 


c.c. 

64.25 


.025 


64.20 


.026 


64.15 


.027 


64.05 


.028 


64.00 


.029 


63.95 


.030 


63.90 


.031 


63.80 


.032 


63.75 


.033 


63.70 


.034 


63.65 


1.035 


63.55 


1.036 


63.50 



For ordinary purposes a pipette graduated to deliver 64 metric c.c. 
is sufficiently exact. 

Acid Nitrate of Mercury Solution. In clarifying milk for polariza- 
tion acid nitrate of mercury is generally used. The reagent is prepared 
as follows: Dissolve metallic mercury in twice its weight of nitric acid 
of 1.42 sp. gr., and dilute with an equal volume of water. 

Mercuric-iodide Solution. Mercuric-iodide solution may also be 
used for clarification. The reagent is prepared by adding 33.2 gms. of 
potassium iodide to a solution of 13.5 gms. mercuric chloride in 20 c.c. 
of glacial acetic acid and 640 c.c. of water. 

* Methods of Analysis A. O. A. C. Bull. 107 (revised), U. S. Bur. of Chem., p. 118. 



METHODS OF SIMPLE POLARIZATION 253 

In carrying out the process, the volume of milk corresponding to 
the lactose double normal weight is measured into a 102.6-c.c. flask. 
For clarification either 1 c.c. of the acid mercuric nitrate, or 30 c.c. of 
the mercuric- iodide solution may be used (an excess of either -reagent 
does no harm). The liquid is shaken and then made up to a volume 
of 102.6 c.c., the extra 2.6 c.c. being the estimated volume of the pre- 
cipitated casein, albumin, and fat. After mixing, the liquid is filtered 
and polarized in a 400-mm. tube; the scale reading divided by 4 gives 
the approximate percentage of lactose in the milk. 

Wiley and Swell's * Double-dilution Method. The volume of precipi- 
tate in the preceding method varies according to the content of protein 
and fat so that the fixed estimate of 2.6 c.c. is not always accurate. 
For more exact purposes of analysis the double-dilution method of 
Wiley and Ewell may be used. The general principle of double dilu- 
tion, due to Scheibler, has been considered on page 209. 

Two separate double lactose-normal-weight portions of milk are 
introduced into a 100-c.c. and 200-c.c. flask respectively. The same 
volume of clarifying agent is then added to each flask and the volume 
completed to the mark. The solutions are shaken, filtered, and read 
in a 400-mm. tube. The reading of the 100-c.c. solution subtracted 
from 4 times the reading of the 200-c.c. solution gives the reading cor- 
rected for volume of precipitate, and this reading divided by 4 gives the 
percentage of lactose in the milk. 

Example. The saccharimeter readings (400-mm. tube) of a milk analyzed 
by the above method were 20.00 for the 100-c.c. flask and 9.80 for the 200-c.c. 
flask. 

The reading corrected for volume of precipitate is then (4 X 9.80) 20.00 
= 19.20, and the percentage of lactose is 19.20-=- 4 = 4.80. 

The volume of precipitate according to the above observations would be 

100 (2o.o- 19.2), 4e-e . (seep . 210) . 

Leffman and Beam's Method. When the percentages of fat and 
protein are known in a milk, the volume of precipitate formed during 
clarification can be calculated according to Leffman and Beam f by the 
following method. 

Calling the specific gravity of milk fat 0.93 the volume of precipi- 
tated fat is found by multiplying the grams of fat in the weight of 

sample by ^r-^ = 1.075. In the same way the volume of the precipi- 



Analyst, 21, 182. t " Analysis of Milk and Milk Products " (1896), p. 39. 



254 



SUGAR ANALYSIS 



tated protein-mercury compound is found by multiplying the grams of 
protein in the weight of sample by -^= = - 8 - The sum of the volumes 

of fat and protein is the volume in cubic centimeters of the precipitate. 

For the polarization of evaporated or condensed milks the single 
lactose-normal-weight of substance is taken. The method of analysis 
in other respects is the same as described for ordinary milk. 

The determination of lactose in milk by the saccharimeter is not 
considered upon the whole to be as accurate as by the gravimetric 
method of copper reduction. A considerable variation is frequently 
found in the determinations by the two methods. In ten comparative 
determinations of lactose in condensed milk by different collaborators 
of the Association of Official Agricultural Chemists* an average varia- 
tion of 0.30 was found between the results by the optical and by 
the gravimetric method, the differences ranging from 0.03 to 0.90. In 
a series of comparative determinations by Patrick and Boylef upon 
unsweetened condensed milks, the following results were obtained: 





Lactose. 


Sample. 


By polariscope, 
clarification with 


By copper reduc- 
tion, 




acid Hg(NO 3 ) 2 . 


Soxhlet's method. 


1 


10.07 


10.04 


2 


10.19 


10.51 


3 


10.57 


10.69 


4 


9.97 


10.15 


5 


8.71 


9.20 


6 


9.00 


9.37 



The correction for volume of mercury precipitate in the above 
samples was made by the method of Leffman and Beam. It is seen 
that there is an average difference of about 0.25 between the two 
methods. 

The cause of the occasional wide deviations between the results of 
the optical and gravimetric methods for determining lactose has been 
variously explained. The difference has been attributed by some to 
the presence of foreign optically active substances, such as unpre- 
cipitated proteids, organic acids, " animal gum," etc., but this has not 
been conclusively established. Differences due to variation in volume 

* Proceedings A. O. A. C., 1906, 1907, Bulls. 105 and 116, U. S. Bur. of Chem. 
t Bull. 105, U. S. Bur. of Chem., p. 109. 



METHODS OF SIMPLE POLARIZATION 255 

of precipitated fat and proteids are of course greater in case of con- 
densed or evaporated milks. 

Polarization of Milk Sugar. The optical method for determining 
lactose is easily applied to the analysis of commercial milk-sugar, 
when other optically active compounds are absent. The lactose- 
normal-weight of sugar is made up to 100 c.c. with the addition of 
a little alumina cream; with dark-colored products containing milk 
sugar the solution of substance must be clarified, following the same 
methods and precautions as in the polarization of :*aw cane sugars. 
In polarizing milk sugar the saccharimeter reading must not be taken 
until mutarotation has disappeared; the solution of sugar is either al- 
lowed to remain in the tube until a constant reading is obtained or 
the mutarotation is destroyed by adding a few cubic centimeters of 
N/ 10 sodium carbonate solution at the time of making up to volume. 

The methods of simple polarization described in the present chapter 
may obviously be applied to the polarization of products containing 
glucose, maltose, and other sugars. But in practical work it is found 
that such sugars generally occur in mixtures with other carbohydrates, 
and the methods for their determination are accordingly given elsewhere. 

INFLUENCE OF TEMPERATURE UPON SACCHARIMETRIC OBSERVATIONS* 

Before concluding this chapter upon methods of simple polarization, 
the influence of changes in temperature upon the accuracy of sac- 
charimetric observations should be considered. 

It has been shown (p. 127) that with an increase in temperature 
the specific rotation of sucrose undergoes a decrease and the rotatory 
power of the quartz compensation an increase, the combined effect of 
all influences producing a decrease in the saccharimeter reading of a 
normal weight of pure sucrose of 0.03 V. for 1 C. increase in temper- 
ature, and that for temperatures between 20 and 30 C. the general 
equation F 20 = V*\ 1 + 0.0003 (t - 20) J may be used for changing the 
Ventzke reading (V) of pure sucrose at any temperature t to the read- 
ing at 20. 

Saccharimeter Temperature Corrections. The employment of 
a temperature correction, similar to the above, was made by the 

* For a full discussion of this question with bibliographic references see paper 
by Browne, " The Use of Temperature Corrections in the Polarization of Raw Sugars 
and Other Products upon Quartz Wedge Saccharimeters," read before Section V, 
Seventh International Congress of Applied Chem., London, 1909, also in J. Ind. 
and Eng. Chem. I, 567, and Z. Ver. Deut. Zuckerind., 69, 404. 



256 SUGAR ANALYSIS 

United States Treasury Department in 1897, in its polarization of 
sugars assessed for duty. The right of the Treasury Department to 
make such corrections in the observed saccharimeter readings was con- 
tested in the courts by several importers of sugar, who founded their 
case largely upon the claim that the rotation of pure sucrose is not ap- 
preciably affected by changes in temperature. The chemists repre- 
senting the government were successful, however, in showing that the 
specific rotation of sucrose is thus affected, and after a final appeal to 
the United States Supreme Court the case of the importers was dis- 
missed for want of jurisdiction.* 

The decision of the courts, which apparently justified the use of 
temperature corrections established for pure sucrose in correcting the 
polarization of all grades of raw sugars, has unfortunately seemed to 
many chemists sufficient authorization to use such corrections indis- 
criminately in the polarization of any and every kind of sugar-contain- 
ing material. Since the saccharimetric reading of a raw sugar or other 
impure product is simply an expression of the sum of the optical ac- 
tivities of the various constituents, sucrose, glucose, fructose, organic 
acids, gums, etc., it is evident that a system of temperature corrections 
which shall give the saccharimeter reading that would be obtained at 
20 C., must correct for the variations produced by temperature in the 
specific rotation of all the optically active ingredients and not of the 
sucrose alone. 

Wiley's Temperature Correction Table. Wiley f has prepared a 
temperature table for correcting the readings of quartz wedge sac- 
charimeters which is based upon the variations in the Ventzke scale 
reading of normal and fractional normal weights of pure sucrose. 
This table has a range from 75 V. to 100 V. for temperatures be- 
tween 4C. and 40 C.; the corrections are to be subtracted from 
the observed readings, when the temperature of polarization is be- 
low and to be added when the temperature is above that of stand- 
ardization. 

United States Treasury Department Method of Temperature Cor- 
rections. The method of temperature corrections devised by the 
Office of Weights and Measures of the United States Coast and Geodetic 
Survey and adopted by the United States Treasury Department for 
use in the Custom-House laboratories, consists in increasing or dimin- 
ishing the saccharimeter reading of each sugar solution by the variation 

* For testimony in this case see "Transcript of Record," U. S. Supreme Court, 
the American Sugar Refining Company, vs. The United States, 
t J. Am. Chem. Soc., 21, 568. 



METHODS OF SIMPLE POLARIZATION 



257 



in reading which a standard quartz plate shows from the computed 
sugar value of this plate for the temperature of observation. 

The following report gives the temperature corrections in sugar 
degrees for a quartz control plate tested by the United States Bureau 
Standards. 

DEPARTMENT OF COMMERCE AND LABOR, BUREAU OF STANDARDS, 

WASHINGTON 

ACCOMPANYING REPORT OF TEMPERATURE CORRECTIONS IN SUGAR DEGREES FOR 
QUARTZ CONTROL PLATE 233-B.S. 1910 



Degrees 
centigrade. 


Sugar 
value. 


Degrees 
centigrade. 


Sugar 
value. 


Degrees 
centigrade. 


Sugar 
value. 


Degrees 
centigrade. 


Sugar 
value. 


13.0 


90.04 


20.0 


90.25 


25.0 


90.40 


30.0 


90.55 


14.0 


90.07 


20.5 


90.27 


25.5 


90.42 


30.5 


90.57 


15.0 


90.10 


21.0 


90.28 


26.0 


90.43 


31.0 


90.58 


16.0 


90.13 


21.5 


90.30 


26.5 


90.45 


31.5 


90.60 


17.0 


90.16 


22.0 


90.31 


27.0 


90.46 


32.0 


90.61 


17.5 


90.18 


22.5 


90.33 


27.5 


90.48 


32.5 


90.63 


18.0 


90.19 


23.0 


90.34 


28.0 


90.49 


33.0 


90.64 


18.5 


90.21 


23.5 


90.36 


28.5 


90.51 


34.0 


90.67 


19.0 


90.22 


24.0 


90.37 


29.0 


90.52 


35.0 


90.70 


19.5 


90.24 


24.5 


90.39 


29.5 


90.54 


36.0 


90.73 



If the polarization temperature is above 20C., add to the reading the difference 
between the reading of the plate and the sugar value of the plate at the polariza- 
tion temperature shown by the above table. If the polarization temperature is 
below 20C., subtract the correction. 

It will be noted from this table that the variation of 0.030 V. per 
1 C., for the reading of a normal weight of pure sucrose, is applied 
without change to a plate testing 90.25 V. at 20 C. The true tem- 
perature correction for a sucrose solution reading 90.25 V. upon the 
saccharimeter would of course be 0.030 X 0.9025 = 0.027 per 1 C. 
The correction table is strictly true therefore only for sugar solutions 
polarizing 100 V. at 20 C. It would be wrong in principle to apply 
such corrections to sucrose solutions testing 80 V. or 50 V. or 20 V. 
since in the latter instances the corrections are only 80 per cent, 50 per 
cent, and 20 per cent, respectively, of the correction for a 100 V. sucrose 
solution. The correction formula 7 20 = V 1 \1 +0.0003 (t - 20)\ or 
the equivalent corrections of Wiley's table are, therefore, to be pre- 
ferred to the method used by the United States Treasury Department, 
when it is desired to correct the polarizations of pure sucrose solutions 
for change in temperature. 

Errors Involved in Use of Saccharimeter Temperature Corrections. 
The probable errors involved in the use of the above methods for cor- 



258 



SUGAR ANALYSIS 



recting polarizations may be seen from the following diagram (Fig. 
144), which gives the correction for pure sucrose solutions, and the ap- 
proximate corrections for solutions of sugar-beet and sugar-cane prod- 
ucts (according to results obtained by Browne*), to be applied to the 
readings of the Ventzke scale for 1 C. increase in temperature. 

It will be seen that the correction for beet products is much nearer 



+ U.U3U 

+ 0.025 

-1-0.020 
|f0.015 
g+ 0.010 
P.+ 0.005 
S n noo 


s 


^S 


eo( . on 


















\ 


i 


^ 


' ^ 


-iiu 


s^ 


QUO* 


07 












\ 


^ 


- 


^s^j 


. 






^ 


^^ 






a- 0.005 


9 


8t 


\ 


6 
Vent; 


5 
keRe 


ading^ 


J 


2 


1 


i 






g -0.015 


- 




\ 


9 








% 


u 








0-0.025 

rH 








% 












^? 


./ 




g O.OaO 
"-0.035 


- 








\3* 












\ 


x 


,g U.U4U 

|- 0.045 


-:'. 








^ 


^ 














o U.UoU 
|-0.055 












\ 















g- 0.065 


- 












\ 












'-g -0.075 


- 












^ 


'? 

\ 










8-0.085 
000 
















\ 










-0.095 


















\ 









Fig. 144. Diagram for correcting polarizations of sugar products for 
changes in temperature. 

the correction for pure sucrose than that for cane products. This is 
due to the fact that raw cane products contain a larger amount of 
fructose, the change in specific rotation of which towards the right, as 
the temperature increases, compensates to a greater or less degree the 
change in specific rotation of sucrose towards the left. This is made 
more evident in Table XL VII, which gives the polarization and com- 
position of various grades of raw cane sugar. 

* J. Ind. Eng. Chem., 1, 567. 



METHODS OF SIMPLE POLARIZATION 



259 



TABLE XLVII 

Showing Effect of Increase in Temperature upon the Polarization of Sugar-cane 

Products, Browne f 



No. 

1 
2 
3 
4 
5 
6 
7 
8 
9 
10 
11 


Description of sugar. 


Polari- 
zation. 


Sucrose 


Invert 
sugar. 


Water. 


Ash. 


Organic 
non- 
sugar 
by dif- 
ference. 

cent. 
0.22 
0.96 
0.50 
0.53 
0.48 
0.89 
1.40 
0.86 
2.57 
2.46 
7.32 

8.47 


Change in polarization 
for 1 C. increase. 


Found. 


By formula 
0.0003 P. 


Java 


98.55 
97.45 
97.15 
96.15 
94.50 
93.75 
89.20 
87.60 
82.40 
79.65 
67.70 

20.06 


Per 
cent. 

98.74 
97.61 
97.38 
96.61 
95.05 
94.44 
90.59 
89.00 
84.64 
81.69 
71.05 

29.58 


Per 
cent. 

0.64 
0.52 
0.78 
1.53 
1.83 
2.29 
4.63 
4.67 
7.45 
6.80 
11.18 

30.09 


Per 
cent. 

0.19 
0.45 
1.03 
0.85 
1.97 
1.83 
2.11 
2.30 
3.49 
4.84 
6.70 

23.62 


Per 
cent. 

0.21 
0.46 
0.31 
0.48 
0.67 
0.55 
1.27 
3.17 
1.85 
4.21 
3.75 

8.24 


-0.0311 
-0.0301 
-0.0276 
-0.0230 
-0.0212 
-0.0160 
-0.0110 
-0.0106 
0.0000 
+0.0068 
+0.0286 

+0.1120 


-0.0296 
-0.0292 
-0.0291 
-0.0288 
-0.0287 
-0.0281 
-0.0268 
-0.0263 
-0.0247 
-0.0239 
-0.0203 

-0.0060 


Peru 


Cuba 


San Domingo. . 
Cuba 


Cuba 
Philippine 
Louisiana 
Philippine 


Louisiana 


Cuba 


Louisiana ) 
molasses*.. ) 



Calculated mixtures of sucrose and cane molasses. 



Sucrose, 
per cent . 


Molasses, 
per cent. 


















95 


5 


96.00 


96.50 


1.50 


1.10 


0.40 


0.50 


-0.0229 


-0.0288 


90 


10 


92.00 


93.00 


3.00 


2.20 


0.80 


1.00 


-0.0158 


-0.0276 


85 


15 


88.00 


89.50 


4.50 


3.30 


1.20 


1.50 


-0.0087 


-0.0264 


80 


20 


84.00 


86.00 


6.00 


4.40 


1.60 


2.00 


-0.0016 


-0.0252 


75 


25 


80.00 


82.50 


7.50 


5.50 


2.00 


2.50 


+0.0055 


-0.0240 


70 


30 


76.00 


79.00 


9.00 


6.60 


2.40 


3.00 


+0.0126 


-0.0228 



* Average of 4 samples. 

Raw sugars can be regarded as simple mixtures of sucrose crystals 
and molasses, and the results in the second part of the table calculated 
for various theoretical mixtures of sucrose and exhausted cane molasses 
agree closely with those observed for the different raw sugars. 

The observations by Browne in Table XL VII have also been con- 
firmed by Wiley and Bryan J who obtained very similar figures upon 
different grades of raw cane sugar. 

The effect of temperature upon the polarization of American beet 
sugar and molasses is shown in Table XL VIII. 



t J. Ind. Eng. Chem., 1, 567. 



Z. Ver. Deut. Zuckerind., 59, 916. 



260 



SUGAR ANALYSIS 



TABLE XLVIII 

Showing Effect of Increase in Temperature upon the Polarization of Sugar-beet 

Products, Browne f 





















Change in polari- 


















Organic 


zation for 1C. 


d 
K 


Product. 


Polari- 
zation. 


Su- 
crose. 


Raffi- 
nose. 


Invert 
sugar. 


Water. 


Ash. 


non- 
sugar 
by dif- 


increase. 






















ference. 




Formula 






















0.0003 P. 








Per 
cent. 


Per 
cent. 


Per 

cent. 


Per 
cent. 


Per 

cent. 


Per 
cent. 






1 


Beet sugar . . 


91.25 














-0.0276 


-0.0274 


2 


Beet sugar. . 


86.60 














-0.0263 


-0.0260 


3 


Beet sugar 


85 50 














-0.0214 


-0.0257 


4 


Beet \ 
molasses* ) 


51.22 


48.13 


1.72 


0.94 


19.86 


7.62 


21.74 


-0.0053 


-0.0154 



Calculated mixtures of sucrose and beet molasses. 



Sucrose, 
per cent 


Molasses, 
per cent. 




















90 
80 
70 
60 


10 
20 

30 
40 


95.00 
90.00 
85.00 
80.00 


94.80 
89.60 
84.40 
79.20 


0.15 
0.30 
0.45 
0.60 


0.10 
0.20 
0.30 
0.40 


2.0 
4.0 
6.0 
8.0 


0.75 
1.50 
2.25 
3.00 


2.20 
4.40 
6.60 
8.80 


-0.0275 
-0.0250 
-0.0225 
-0.0200 


-0.0285 
-0.0270 
-0.0255 
-0.0240 



* Average of 3 samples. 

It will be seen from the above that the temperature formula 
P 20 = P i [1 + 0.0003 (t - 20)], or the corresponding corrections of the 
Wiley table, can be applied without serious error to practically all 
grades of beet sugar and to those grades of cane sugar polarizing over 
96. As the polarization of raw cane sugars falls below 96, and the 
percentage of invert sugar (or fructose) increases, the effect of change 
in temperature upon the rotation of the latter begins to lower appre- 
ciably the temperature coefficient for the rotation of sucrose until, at a 
point about 80 V., the two influences that of the temperature upon 
the fructose and other impurities and that of the temperature upon 
the sucrose and quartz wedges of the instrument exactly counter- 
balance one another.t Under these conditions a sugar will polarize the 
same at all temperatures. Below 80 V. the temperature coefficient for 
the rotation of the sucrose in raw cane sugars is usually more than 

t J. Ind. Eng. Chem., 1, 567. 

I The calculation upon page 128 shows that the proportion of fructose to sucrose 
for equilibrium between their temperature coefficients is 3.13 to 100.0. 



METHODS OF SIMPLE POLARIZATION 261 

counterbalanced, the result being that the polarization of these sugars 
increases with elevation of temperature. This increase continues, as 
the polarization diminishes (the percentage of fructose and other im- 
purities being greater), until, at a polarization of about + 20 for ex- 
hausted cane molasses, an increase of 1 C. in temperature causes an 
increase of over 0.1 V. in the saccharimeter reading. 

Correction of Polarizations for the Combined Influence of Temperature 
upon the Rotation of Sucrose and Invert Sugar. Since the ingredient of 
sugar products, whose polarization is most susceptible to the influence 
of temperature, is invert sugar, a more accurate method of correcting 
saccharimeter readings is to combine the temperature coefficients of 
sucrose and invert sugar as by the formula: P 20 = P t + 0.0003 S (t - 20) 
- 0.0045 I (t 20) in which P t is the polarization at t C., S the per- 
centage of sucrose and / the percentage of invert sugar. 

If the percentage of invert sugar is unknown the temperature correc- 
tion for converting polarizations to 20 C. may be determined approxi- 
mately by the following empirical equations: 

For cane products, P 20 = P l + 0.0015 (P< - 80) (t - 20), 

For beet products, P 20 = P t + 0.0006 (P t - 50) (t - 20). 

Such formulae as the above while more accurate than corrections 
which are based upon the temperature coefficients of pure sucrose, fail 
to give accurate results upon many individual products whose com- 
position differs from that of the average type. 

Polarization at Constant Temperature. It is evident from the 
foregoing that the method of applying temperature corrections es- 
tablished for pure sucrose to the polarization of sugar products in 
general is faulty. Since it is impossible to devise a simple reliable 
method of temperature corrections that can be applied to the polari- 
zation of all kinds of substances, the one means of securing uniformity 
and accuracy in saccharimetric work is to make all polarizations at the 
temperature at which the instruments are standardized. Custom- 
house laboratories, arbitration laboratories, and all other laboratories, 
upon the results of which great interests are involved, should be 
equipped with cooling and warming apparatus for maintaining a uni- 
form standard temperature throughout the year. 

The New York Sugar Trade Laboratory was the first testing labo- 
ratory in the United States to follow out the requirements of the In- 
ternational Commission for Uniform Methods of Sugar Analysis and 
make all polarizations at 20 C. The laboratory room and polarizing 
cabinet used for this purpose are insulated. In warm weather the air 



262 



SUGAR ANALYSIS 



is circulated by an electric fan through ducts over cooling coils, fresh 
air being introduced from outside according to the needs of ventilation. 
A small ammonia compressor driven by an electric motor serves for 
the work of refrigeration. The temperature can be controlled either 




Fig. 145. Refrigerating machine for constant temperature polarization 
(New York Sugar Trade Laboratory) . 

automatically by means of a thermostat which operates dampers regu- 
lating the passage of air to and from the cooling box, or directly by 
means of the rheostats controlling the speed of compressor and venti- 
lating fan. The general arrangement of the equipment is shown in 
Figs. 145 and 121. 



CHAPTER X 

METHODS OF INVERT OR DOUBLE POLARIZATION 

THE methods of direct polarization, as previously explained, give 
percentage of sucrose only in the absence of other optically active sub- 
stances. To determine the percentage of sucrose when other optically 
active substances are present, the method of inversion or double polari- 
zation is used, the principle of which may be understood from the 
following. 

Law of Inversion. When a solution of sucrose is acted upon by 
some inverting agent, such as an acid or the enzyme invertase, the 
sucrose molecule is broken up or inverted, giving rise, by the addition 
of one molecule of water, to one molecule each of glucose and fructose, 
the mixture of these two sugars in equal amounts being termed invert 
sugar. This reaction, known as hydrolysis or inversion, is expressed by 
the following equation: 

C 12 H 22 O n + H 2 = C 6 H 12 6 + C 6 H 12 6 . 

Sucrose (342) Water (18) Glucose (180) __ Fructose (180) 

Invert Sugar (360) 

It is seen from the above that one part of sucrose is converted into 

360 

n = 1.05263 parts of invert sugar. Calling the specific rotation at 



20 C. +66.5 for sucrose, and - 20.00 for invert sugar (p. 174), the 
relation of the optical activity of one part sucrose before and after in- 
version will be + 66.5 : 1.05263 (- 20.00) = 66.5 : - 21.0526 or a de- 
crease of 87.5526 in specific rotation. This decrease for one degree of 

87 5526 

the saccharimeter scale would therefore be - - = 1.3166. The gen- 

bo. o 

eral law of inversion* as applied to the determination of sucrose may 
then be stated as follows: 

The total decrease in the saccharimeter reading at 20 C. of the 
normal weight of product after inversion divided by 1.3166 gives the 
percentage of sucrose when no other optically active ingredient is 
hydrolyzed and when the inverting agent produces no change in the 
specific rotation of the other optically active constituents present. 

* For a fuller discussion of the laws of inversion see page 659. 
263 



264 SUGAR ANALYSIS 

The enzyme invertase fulfills most perfectly the conditions above 
named, and when this is used as the inverting agent the percentage of 
sucrose in mixtures with glucose, fructose, invert sugar, maltose, milk 
sugar, etc., may be determined very closely by use of the factor 1.3166. 
The inverting agent most commonly used in optical analysis is not in- 
vertase, however, but hydrochloric acid, the presence of which, as shown 
on page 185, has a most pronounced influence in increasing the specific 
rotation of fructose. When hydrochloric acid is used for inverting, the 
factor 1.3166 must be modified according to the amount of acid used for 
inverting, the concentration of the sugar solution, and the manner of 
conducting the inversion. The extreme susceptibility of fructose to 
changes in specific rotation and composition makes it necessary in 
employing any method of inversion to adhere most rigidly to the rules 
of procedure prescribed. 

THE CLERGET METHOD OF INVERSION 

The method of inversion for determining sucrose was devised in 
1849, by Clerget,* who found that a solution of the French normal 
weight of pure sucrose in 100 c.c., reading + 100 degrees upon the 
saccharimeter, gave after inversion with hydrochloric acid a reading of 
- 44 degrees at C. or - 34 degrees at 20 C. The total difference 
between the readings before and after inversion, correcting for the in- 
fluence of temperature, is expressed by the quantity 

100- (-44)-| = 144-|, 

t being the temperature of the inverted solution at polarization. 

If D represents the algebraic difference (P P f ) between the direct 
polarization (P) and the invert polarization (P'} of a given product, 
then the percentage (S) of sucrose by Clerget's formula is expressed by 

the equation S - - If the invert polarization is made at 20 C. 



the equation becomes S = Q . or ^-~oT' The factor 1.34 is considerably 



greater than the factor 1.3166 for pure aqueous solutions of invert sugar. 
Tuchschmidf who subjected the Clerget process to an exhaustive 
analysis, arrived at the following formula, 

100 D 



S 



144.16035- 0.50578 1 



* Compt. rend., 16, 1000; 22, 1138; 23, 256; 26, 240; Ann. chim. phys. [3], 26, 175. 
t Z. Ver. Deut. Zuckerind., 20, 649. 



METHODS OF INVERT OR DOUBLE POLARIZATION 265 

The original Clerget formula does not differ sufficiently from this 
to warrant the greater labor of calculation involved in the use of the 
long decimals. 

If the direct and invert readings are made upon a polarimeter with 
circular degrees the Clerget formula would be, for the German normal 
weight (1 sugar scale = 0.34657 circular degrees), 

100 D 100 D 

.34657 (144 -.50"" 49.906 - 0.173 V 

for the French normal weight (1 sugar scale = 0.21719 circular 
degrees), 

100 D 100 D 



.21719 (144 - .5 31.275 - 0.109 1 

One gram of sucrose dissolved to 100 metric cubic centimeters gives 
a direct reading of ^ = 1.333 circular degrees and an invert read- 

15 249 

ing of ^Q~ = 0.5865 circular degrees at 0C; the grams of su- 
crose (C) in 100 c.c. of any solution may be found from the polarimeter 
reading before and after inversion by the equation 

P-P' P-P' 

(49.906- 0.173 Q 1.9195 - 0.0067 1 
26 

The Clerget formulae, given above, are to be employed only when 
the following method of inversion prescribed by Clerget is followed. 
After. taking the direct polarization (p. 202), the clarified solution re- 
maining is filled up to the 50-c.c. graduation mark of a flask graduated 
at 50 and 55 c.c.; concentrated hydrochloric acid is then added to the 
55-c.c. mark, a thermometer is inserted, and the flask slowly warmed 
until the temperature reaches 68 C., 15 minutes being taken in the heat- 
ing.* The solution is then quickly cooled, filtered if necessary, and 
polarized as nearly as possible at the original temperature of making up 
to volume. The polariscope reading for a 200-mm. tube of solution must 
be increased by T V to correct for the dilution with acid. The reading 
of the inverted solution is sometimes made in a 220-mm. tube, when 
no correction for dilution is needed. 

* The addition of the acid causes an elevation of 2 to 3 C. in temperature; 
there is also a slight loss from evaporation during the inversion. It is, therefore, 
better to control the temperature by inserting the thermometer in a 50-55 c.c. flask 
filled with water and placed in the bath with the solutions undergoing inversion. 
After cooling to room temperature, the volumes are readjusted to 55 c.c. 



266 SUGAR ANALYSIS 

In carrying out the inversion special attention must be paid to all 
details. If the temperature of 68 C., or the time of 15 minutes, is ex- 
ceeded, a partial destruction of fructose may result; if the temperature 
of 68 C. is not reached, or if the time of heating is less than 15 min- 
utes, some of the sucrose may escape inversion. Care must also be 
taken to maintain a constant temperature in the polarization tube 
during the reading. Even a slight warming of the tube, as from han- 
dling, will affect the observation. A polarization tube provided with a 
jacket for circulation of water at the desired temperature is very de- 
sirable for polarizing inverted solutions. (See Fig. 111.) 

Herzf eld's Modification of the Clerget Method. The original 
method of Clerget has been variously modified from time to time in 
order to diminish the danger of destroying fructose and to secure better 
uniformity of conditions. The inversion method of Herzfeld,* which is 
the one most generally employed at present, is as follows: 

The half normal weight (13.00 gms.) of product is transferred with 
75 c.c. of water into a 100-c.c. flask; after solution of soluble matter, 
5 c.c. of hydrochloric acid of sp. gr. 1.188 are added, a thermometer 
is introduced and the flask placed in a water bath heated to between 
72 and 73 C. As soon as the thermometer in the flask indicates 
69 C. (2.5 to 5 minutes) the solution is kept at this temperature for 
exactly 5 minutes, rotating the flask gently at frequent intervals to se- 
cure even distribution of the heat. The entire time of heating, accord- 
ing to the length of the preliminary period, will vary thus from 7 J to 
10 minutes, and should never exceed 10 minutes. When the 5-minute 
heating at 69 C. is completed the flask is cooled as quickly as possible 
to 20 C., the thermometer is rinsed from adhering sugar solution and 
the volume made to 100 c.c. After mixing and filtering, the solution is 
polarized with all the precautions previously mentioned. The polari- 
scope reading is doubled to obtain the correct invert reading for a nor- 
mal weight of substance. 

The invert reading for 26 gms. of chemically pure sucrose under the 
above conditions is -42.66 V. at 0, or -32.66 V. at 20 C. The 
Clerget formula, according to Herzfeld's modification, is then expressed 
by the equation ~ _ 100 D 

~ 142.66 -0.5*; 
or, if the polarization be made always at 20 C., by 

100 D 7 _ Qft 
= 132^6 ==07538Z> - 

* Z. Ver. Deut. Zuckerind. (1888), 38, 699. 



METHODS OF INVERT OR DOUBLE POLARIZATION 267 



Effect of Concentration on the Clerget Factor. The factor 132.66 in 
the preceding equation is correct only for a solution containing the 
half normal weight of sugar to 100 c.c. For other concentrations than 
this the value of the invert reading will vary according to the general 
formula P /20 = - (31.78 + 0.0676 c), or P' = - (41.78 + 0.0676 c), in 
which P f is the invert reading upon the Ventzke scale and c the grams 
of sucrose in 100 c.c. The following table gives the value of the 

factor 142.66 in the equation S = 
trations of sucrose. 



142.66-0.5* 



for different concen- 



TABLE XLIX 

Giving Clerget Factors at Different Concentrations of Sucrose 
for Herzfeld's modified Method 



Grams sucrose 
in 100 c.c. 


Factor. 


Grams sucrose 
in 100 c.c. 


Factor. 


1 


141.85 


14 


142.73 


2 


141.91 


15 


142.79 


3 


141.98 


16 


142.86 


4 


142.05 


17 


142.93 


5 


142.12 


18 


143.00 


6 


142.18 


19 


143.07 


7 


142.25 


20 


143.13 


8 


142.32 


21 


143.20 


9 


142.39 


22 


143.27 


10 


142.46 


23 


143.33 


11 


142.52 


24 


143.40 


12 


142.59 


25 


143.47 


13 


142.66 


26 


143.54 



Instead of the above correction table the following general formula 
has been proposed by Herzfeld:* 

100 (P - P') 
141.84 + 0.05 N- 0.5*' 



S 



in which P and P' are the direct and invert polarizations for a normal 
weight of substance and N the scale reading of the inverted solution. 
This formula assumes that the value N always bears a constant ratio 
to the concentration of sucrose, which is of course only true when 
other optically active substances are absent. 

Example. The application of the above Herzfeld formula is best illus- 
trated by an example: 26.00 gms. of a sugar sirup dissolved to 100 true c.c. at 
20 C. gave a direct reading in a 200-mm. tube of + 60.00 (P). 13.00 gms. of 

* Z. Ver. Deut. Zuckerind., 40, 194. 



268 



SUGAR ANALYSIS 



this same sirup inverted according to Herzfeld's method gave a reading of 
9.7 (N) at 20 (t) upon the negative scale, - 9.7 X 2 = - 19.4 (P'). Substi- 
tuting these values in the formula, we obtain 

S - 100 [+60 -(-19.4)] _ 

~ 141.84 +(0.05X9.7)- (0.5 X 20) ~ 
the amount of sucrose present in the sirup. 

If the direct and invert polarizations be made at 20 C. the Clerget 
and Herzfeld formulae become simplified as follows: 

100 (P - P') 



Clerget formula = 
Herzfeld formula = 



144-2 
30 (P - 



= 0.7463 (P-P'); 
= 0.7538 (P - P'). 



66 - 

The values of the Herzfeld factor in the simplified formula for tempera- 
tures between 10 and 40 9 C. are given in Table L. 

TABLE L 

Giving the Inversion Factors for Herzfeld's Modification of Clerget' s Method 
at Different Temperatures 



Temper- 
ature. 


Factor. 


Temper- 
ature. 


Factor. 


Temper- 
ature. 


Factor. 


10 C. 


0.7264 


20 C. 


0.7538 


30 C. 


0.7833 


11 


0.7290 


21 


0.7566 


31 


0.7864 


12 


7317 


22 


0.7595 


32 


0.7895 


13 


0.7344 


23 


0.7624 


33 


0.7926 


14 


0.7371 


24 


0.7653 


34 


0.7957 


15 


0.7398 


25 


0.7682 


35 


0.7989 


16 


0.7426 


26 


0.7712 


36 


0.8021 


17 


0.7454 


27 


0.7742 


37 


0.8053 


18 


0.7482 


28 


0.7772 


38 


0.8086 


19 


0.7510 


29 


0.7802 


39 


0.8119 










40 


0.8152 



The inversion method of Herzfeld gives correct results only when 
the prescribed conditions of concentration, amount of acid, volume, 
temperature and time of inversion are carefully followed. The tem- 
perature of inversion for the 5-minute period should be maintained at 
exactly 69 C. if possible; a variation of 1 C. from this temperature is 
found to produce a difference of over 0.1 in the calculated percentage 
of sucrose. The extreme sensibility of fructose to decomposition during 
inversion and its wide fluctuation in optical rotation with slight changes 
of temperature necessitate the greatest care in manipulation. Neglect of 
this precaution is a frequent cause of variation between the results of 
different analysts. 



METHODS OF INVERT OR DOUBLE POLARIZATION 269 

Inversion at Ordinary Temperature. The dangers of too high or 
too prolonged heating in the Clerget determination may be avoided by 
inverting at the ordinary laboratory temperature. The time neces- 
sary to invert a half-normal weight (13 gms.) of sucrose in 100 c.c. of 
solution employing hydrochloric acid of 1.18 sp. gr. was found by Ham- 
merschmidt * to be as follows : 



Temperature. 


5 c.c. HC1. 


10 c.c. HC1. 


C. 
10 


Hours. 

225 


Hours. 

94 


15 


101 


44 


20 


47 


20 


25 


23 


10 


30 


11.6 


5 



The method of Tolmanf for cold acid inversion is to place 50 c.c. 
of a solution containing the half-normal, or normal, weight of substance 
in a 100-c.c. graduated flask, add 5 c.c. of strong hydrochloric acid, 
allow to stand at room temperature (above 20 C.) for 20 to 24 hours, 
make up to 100 c.c. and polarize. At 25 C. the inversion is complete 
in about 10 hours and at 20 C. in about 20 hours. The Clerget factor 
for a half-normal weight (13 gms.) of sucrose inverted in the cold was 
found by Tolman to be 142.88. 

Effect of Amount of Acid on the Clerget Factor. The effect of vary- 
ing the quantity of hydrochloric acid used for inversion upon the 
Clerget factor was studied by Hammerschmidt,* who obtained the fol- 
lowing invert readings at 20 C. for a normal weight of pure sucrose, 
using 5 c.c., 10 c.c., 15 c.c., and 20 c.c. of hydrochloric acid per 100 c.o. 

5 c.c. 10 c.c. 15 c.c. 20 c.c. 

Reading of normal weight, (Degrees Ventzke) -34.00 -35.04 -35.95 -36.80 
Reading of \ normal weight X 2, (Degrees Ventzke) -33.00 -34.12 -35.15 -36.03 

It will be noted that there is a pronounced but diminishing increase 
in the invert reading with the addition of each 5 c.c. of acid. 

Results similar to those of Hammerschmidt were obtained by 
Tolman, f who found for a solution of invert sugar made up to volume 
with no hydrochloric acid a reading of 23.0 V., for the same amount 
of invert sugar solution made up with 5 c.c. hydrochloric acid a reading 
of 24.2, and for a third similar portion made up with 10 c.c. hydro- 
chloric acid a reading of 25.0. 



* Z. Ver. Deut. Zuckerind, 40, 465. t Bull. 73, U. S. Bur. Chem., p. 



270 



SUGAR ANALYSIS 



Effect of Fructose on the Clerget Factor. Owing to the influence of 
hydrochloric acid upon the polarization of fructose a Clerget formula 
based upon the inversion of pure sucrose by means of this acid is not 
absolutely correct when applied to the analysis of impure products 
containing invert sugar, since the specific rotation of fructose is differ- 
ent in the neutral and acid solutions before and after inversion. A 
considerable error is introduced, in fact, if the Clerget formula estab- 
lished for pure sucrose be employed in the examination of molasses, 
honey, jam, jelly, and other materials containing considerable fructose. 

Effect of Amino Compounds on the Clerget Factor. The hydrochloric 
acid used for inversion may also affect the polarization of other ingredi- 
ents than fructose. Low-grade molasses, plant extracts, and other 
sugar-containing materials frequently contain considerable quantities of 
optically active ammo compounds such as asparagine, aspartic acid, 
glutaminic acid, leucine, isoleucine, etc., the optical activity of which 
varies with the alkalinity and acidity of the solution. This may be 
seen from the following table which gives the approximate specific rota- 
tions of several amino derivatives in alkaline solution, in water, and in 
hydrochloric acid. 

TABLE LI. 
Approximate Value for [a\D. 





In presence of 
NaOH. 


In water. 


In presence of 
HC1. 


Asparagine 


- 8 


- 6 


+34 


Aspartic acid 


- 9 


+ 4 


+34 


Glutaminic acid 


-68 


+10 


+20 


Leucine 


+ 7 




+17 


Isoleucine 


+ 11 


+10 


+37 



The influence of such variations upon the Clerget calculation is 
illustrated in the work of Andrlik and Stanek * who showed that a 1 per 
cent solution of glutaminic acid gave a reading of 1 .45 V. in presence 
of lead subacetate, 0.35 V. in water alone, and +-1.77 V. in dilute 
hydrochloric acid. In the case of an osmose water from a beet-sugar 
factory the direct polarization was 14.75 V. in alkaline, 14.85 V. in 
neutral, and 15.80 V. in acid solution. Ehrlichf had previously also 
called attention to the large errors in the Clerget method due to the 
presence of amino compounds. 



* Z. Zuckerind. Bohmen, 31, 417. 
t Z. Ver. Deut. Zuckerind., 63, 809. 



METHODS OF INVERT OR DOUBLE POLARIZATION 271 

Clerget Modifications for Impure Sugar Products. It is evident 
that to overcome the variations in specific rotation of fructose, amino 
compounds, etc., which occur in the presence and absence of hydro- 
chloric acid, the original method of Clerget must be considerably modi- 
fied in the case of impure products. Several such modifications of the 
method have in fact been devised and these for convenience may be 
grouped into two general classes. I. Clerget modifications which at- 
tempt to equalize the conditions before and after inversion with hydro- 
chloric acid. II. Clerget modifications which employ an inverting agent 
free from the objections of hydrochloric acid. 

Among the modifications of Class I may be mentioned the following. 

(1) Neutralizing the Free Acid after Inversion before Making the In- 
vert Polarization. This modification is best carried out in the Herzfeld 
process of inversion. After cooling the solution the free hydrochloric 
acid is carefully neutralized by means of sodium hydroxide, using phe- 
nolphthalein as indicator, and avoiding any excess of alkali. After 
neutralizing, the volume is completed to 100 c.c. at 20 C. and the in- 
vert polarization made in the usual way. In order that the direct 
polarization may be made under similar conditions Saillard* recom- 
mends that sodium chloride, equivalent to the amount present after 
neutralizing the hydrochloric acid, be added to a separate solution before 
making up to the 100 c.c. for the direct polarization. The fructose, 
amino compounds, etc., are thus polarized under similar conditions 
before and after inversion. The Clerget constant for this method is 
determined by making a parallel analysis upon pure sucrose. 

(2) Making the Direct Polarization in Presence of Hydrochloric Acid 
and Urea. This modification, due to Andrlik and Stanek,f is based 
upon the retarding influence which urea (or betaine) exercises upon the 
inversion of sucrose with hydrochloric acid in the cold. Fifty cubic 
centimeters of the solution for the direct polarization are made up to 
100 c.c. with a solution containing 5 gms. urea and 5 c.c. strong hydro- 
chloric acid per 50 c.c. of reagent. After mixing, the solution is filtered 
and polarized as quickly as possible. It is claimed by the authors of the 
method that a sufficient interval (7 to 10 minutes) elapses before inver- 
sion is noticeable to make the direct polarization. While this claim 
may be true for certain classes of products, it is certainly not the case 
with substances rich in sucrose. The following experiment shows a 
comparison of the rate of inversion of 13 gms. of sucrose at 20 C. in 
presence of 5 c.c. strong hydrochloric acid and in presence of 5 c.c. strong 
hydrochloric acid plus 5 gms. urea in 100 c.c. of solution. 

* Eighth Int. Cong. Applied Chem., Communications Vol. XXV, p. 541. 
t Z. Zuckerind. Bohmen, 31, 417. 



272 



SUGAR ANALYSIS 



TABLE LII 
Showing Influence of Urea upon the Rate of Inversion of Sucrose 





Inversion with 5 c.c. HC1. 


Inversion with 5 c.c. HC1 + 5 gins, 
urea. 


rftt 








Reading V. 


Velocity constant. 


Reading V. 


Velocity constant. 


min 


+49 9 




+49.9 




2 min. 


49.4 


0.0016 


49.6 


0.0009 


5 min. 


48.9 


0.0013 


49.4 


0.0007 


7 min. 


48.6 


0.0012 


49.3 


0.0005 


10 min. 


48.0 


0.0012 


49.1 


0.0005 


30 min. 


44.3 


0.0013 


47.2 


0.0006 


60 min. 


39.7 


0.0012 


44.8 


0.0006 


120 min. 


31.4 


0.0012 


40.1 


0.0006 


180 min. 


24.7 


0.0012 


35.8 


0.0006 


2 days 


16 5 




17.2 




4 days 


16 5 




-21.3 














Average 




0.00128 




0.00063 







Taking the reading before inversion as + 49.9 and the reading at 
completion of inversion as 16.5 it is seen that the velocity of inver- 
sion (k = - log > see p. 660), is diminished one-half by the addition 

of 5 gms. urea. There is no suspension of the inversion at the beginning, 
there being a decrease of 0.3 in the reading at the end of 2 minutes, 
and of 0.5 after 5 minutes. Under such circumstances it is impossible 
to take the true direct polarization. 

A second objection to the Andrlik-Stanek modification is that the 
method cannot be used when reducing sugars are present owing to the 
change which the urea causes in their specific rotation. The extent of 
this change can be seen from the following experiments upon solutions 
of fructose, glucose, and invert sugar. The same volume of sugar solu- 
tion was taken in each case and, after addition of substance, was 
completed to 100 c.c. The readings were taken immediately except as 
otherwise stated. 





Fructose. 


Glucose. 


Invert sugar. 


Volume completed with water alone 
Volume completed with water + 5 gms. I 
urea ] 


-26.2 V. 

-27.0 


+56.5V. 
+56.1 


-10.2V. 
-10.6 


Volume completed with water + 5 c.c. HC1 
Volume completed with water +5 gms. / 
urea +5 c.c. HC1 ] 


-26.9 
-27.3 


+56.7 
+56.5 


-10.5 
-10.7 


Volume completed with water + 5 gms. ) 
urea + 5 c.c. HC1 after 2 days f 


-27.3 


+48.0 


-11.9 



METHODS OF INVERT OR DOUBLE POLARIZATION 273 

It is seen that the 5 gms. urea + 5 c.c. hydrochloric acid produce a 
different rotation than the 5 c.c. hydrochloric acid alone, this difference 
being greater for fructose. On long standing, glucose in presence of hy- 
drochloric acid and urea shows a loss in rotation owing to the formation 
of glucose ureide ([O\D = 23.5). This explains the high levorotation of 
invert sugar solutions prepared in presence of urea. (See Table LII.) 

The Andrlik-Stanek method is a dangerous one for it may intro- 
duce greater errors than those which it was designed to correct. The 
process, notwithstanding several favorable notices in the literature, is 
not to be generally recommended. 

Among the modified methods belonging to Class II, which employ 
for the Clerget determination inverting agents less open to the objec- 
tions of hydrochloric acid, may be mentioned the following: 

(1) Inversion by Means of Organic Acids. A number of organic 
acids, especially such as have no pronounced influence upon the optical 
activity of fructose, have been employed in place of hydrochloric acid 
for the determination of sucrose by the Clerget method. Weber* showed 
that in presence of acetic acid invert sugar had the same rotatory power 
as in aqueous solution. Acetic acid, however, is an unsatisfactory re- 
agent for the Clerget determination on account of its very weak invert- 
ing action (%%Q that of hydrochloric acid, see p. 663). Tolmanf has 
tested the use of citric acid for the Clerget process and found that with 
2 gms. of this acid to 100 c.c. complete inversion of sucrose could be ac- 
complished in 30 minutes at the temperature of boiling water. Under 
these conditions the Clerget factor] for the normal weight of sucrose 
was 141.95 and for the half-normal weight 141.49. Tolman noted, 
however, that the presence of soluble acetates greatly retarded the in- 
verting action of citric acid and that the latter was consequently of no 
value as an inverting agent with products which required previous 
clarification with lead subacetate. This same objection would apply to 
many other organic acids. Another serious objection, as with hydro- 
chloric acid, against the use of organic acids as inverting agents is the 
difference in optical activity of contaminating amino compounds in 
the solutions used for direct and invert polarization asparagine, for 
example, being levorotatory in aqueous solution, but dextrorotatory in 
presence of strong acetic acid. 

Oxalic acid | has also been recommended as an inverting agent, 
2 gms. of the acid being used for 100 c.c. of solution. This acid has 

* J. Am. Chem. Soc., 17, 321. 

t Bull. 73, U. S. Bur. of Chem., p. 69. 

} Kulisch. Z. ang. chem. (1897), 45. 



274 SUGAR ANALYSIS 

a much stronger inverting power than either acetic or citric acid, but 
is open to the same objections previously stated. 

The employment of organic acids as inverting agents in the ex- 
amination of impure sugar products has not been found upon the whole 
to be satisfactory. 

(2) Inversion by Means of Invertase. The employment of yeast as 
an inverting agent in the Clerget determination of sucrose was first in- 
dicated by Kjeldahl* in 1881. O'Sullivan and Tompson,f in 1891, and 
Ling and Baker { in 1898, extended the use of the method and more 
recently Ogilvie has applied it to the analysis of sugar-factory products. 
The yeast method of O'Sullivan and Tompson, as modified by Ogilvie, 
is as follows : 

" Four times the normal sugar weight of the sample are trans- 
ferred to a standardized 200-c.c. flask, defecated with the minimum 
amount of basic lead-acetate solution (sp. gr., 1.26), a little alumina 
cream added, then the liquid adjusted to bulk at standard temperature, 
well shaken, and filtered; 100 c.c. of the filtrate are measured by a 
standard pipette into a small beaker, sulphur dioxide passed in from a 
siphon of the liquefied gas till a faint smell is perceptible (all the lead 
thus being indicated to be precipitated), then the liquid transferred to 
a 200-c.c. flask, made up to the mark, and well mixed. Now sufficient 
calcium carbonate (dried) in fine powder to neutralize the excess of 
acidity, and a little recently ignited kieselguhr (to promote filtration) 
are added, after which filtration follows. In this way a normal solu- 
tion is obtained, which is sufficiently clarified to give a distinct polari- 
metric reading, is free from lead and excess of acidity, and is therefore 
well suited for the invertase inversion. 

" Fifty cubic centimeters of the solution, prepared in the manner 
just described, contained in a 100-c.c. flask, are raised in a constant- 
temperature bath to between 50 and 55 C., after which 0.5 gm. of washed 
brewery yeast and 2 drops of acetic acid are added and the tempera- 
ture maintained as near 55 C. as possible for 4J to 5 hours. At the end 
of this time the liquid is cooled, and a little alumina cream or kieselguhr 
added to assist filtration, and made up to bulk at standard tempera- 
ture. The clear filtrate is then polarized in a lateral-branched water- 
jacketed tube at exactly 20.0 C." 

The Clerget factor determined by Ogilvie for the above process 
from experiments upon pure sucrose is 141.6. 

* Compt. rend. Lab. Carlsberg (1881), 1, 192. f J- Chem. Soc. Trans., 59, 46. 
t J. Soc. Chem. Ind., 17, 111. Int. Sugar Jour., 13, 145. 



METHODS OF INVERT OR DOUBLE POLARIZATION 275 

Instead of employing yeast, a solution of invertase prepared there- 
from may be used to advantage. Hudson* has developed a method 
upon this principle, which is described as follows: 

" Dissolve 26 gms. of the substance to be analyzed for cane sugar 
in water, clarify with the usual substances (neutral or basic lead acetate 
or alumina cream or kaolin) and make up to 100 c.c. volume at 20 C. 
Filter and read the polarization of the nitrate in a 200-mm. tube. 
Remove the excess of lead from the filtrate, if lead has been used 
as clarifying agent, with sodium carbonate or potassium oxalate, and 
filter. To 50 c.c. of the filtrate add acetic acid by drops until the 
reaction is acid to litmus, add 5 c.c. of the stock invertase solution 
(p. 669), and make up the volume to 100 c.c. Add a few drops of 
toluene to the solution to prevent the growth of microorganisms, shak- 
ing so as to saturate, and allow to stand at any temperature between 
20 and 40 C. over night. Under usual conditions about six hours' 
time is required to accomplish complete hydrolysis." When the inver- 
sion is finished, the solution is read at 20 C. and the invert reading cal- 
culated to the normal weight of substance. The Clerget factor for the 
above method as determined by Hudson from experiments upon pure 
sucrose is 141.7. 

The invertase method is unquestionably the most ideally perfect of 
the numerous Clerget modifications. No disturbances are produced in 
the specific rotations of fructose, amino acids, or other optically active 
substances which may accompany sucrose and no other substances 
than sucrose are hydrolyzed except in the few special cases where 
raffinose, stachyose or gentianose may be present. 

The complications involved in the preparation of the invertase 
reagent, the uncertainty of knowing whether a given preparation is 
always of constant strength, and the long period of time frequently 
necessary to accomplish inversion are the chief drawbacks against the 
use of the method in practical analytical work. 

The inverting power of the stock invertase solution should be care- 
fully determined from time to time by experiments upon pure sucrose 
and with any decrease in activity the quantity of reagent used for in- 
version must be correspondingly increased. The time of inversion can 
be shortened considerably by conducting the inversion at a tempera- 
ture of about 55 C. To determine whether or not inversion is com- 
plete the closed flask or tube of solution may be warmed again to 
55 C. for an hour and then, after cooling to 20 C., reread. If no change 
in polarization is noted, the inversion is complete. 
* J. Ind. Eng, Chem., 2, 143. 



276 SUGAR ANALYSIS 

The invertase method will be found of especial value in research 
work and in controlling the results of other methods. In this con- 
nection, however, it should be noted that the influence of salts and 
other impurities upon the rotation of the accompanying sugars intro- 
duces the same error as in other Clerget modifications. 

CLARIFICATION OF SOLUTIONS FOR THE DETERMINATION OF SUCROSE 
BY THE CLERGET METHOD 

In the analysis of sucrose-containing products by the Clerget method, 
clarification by means of basic lead compounds must precede and not 
follow the process of inversion. This precaution is necessary, owing to 
the occlusion of a part of the invert sugar in the basic lead precipitate 
and the consequent diminution of the invert polarization. In so far as 
the work of analysis will permit, the solution for the direct polarization 
and that used for inversion should both be taken from the same clarified 
filtrate after deleading. The following method of procedure is given as 
an example. 

Method of Deleading. Transfer 57.20 gins, of product with 
about 100 c.c. of water to a graduated 200-c.c. flask. After solution, 
lead-subacetate reagent (1.26 sp. gr.) is added to the necessary point of 
clarification and the volume completed to 200 c.c. After mixing well, 
the solution is filtered and 100 c.c. of the filtrate (28.6 gms. substance) 
treated in a 110-c.c. flask with successive amounts of finely powdered 
potassium oxalate, or sodium carbonate, or sodium sulphate, etc., until 
no more lead is precipitated. If the deleaded solution is alkaline to 
litmus paper or phenolphthalein it is exactly neutralized with acetic 
acid and the volume completed to 110 c.c. The solution is mixed, 
filtered, and the filtrate (26 gms. substance to 100 c.c.) used for the 
direct polarization. Fifty cubic centimeters of the same filtrate are 
then inverted in a 100-c.c. flask, according to the method desired, and, 
after completing the volume to 100 c.c., polarized for the invert reading. 
The latter multiplied by 2 gives the invert polarization. 

In this connection it should be remarked that with substances re- 
quiring large amounts of basic lead for clarification the 5 c.c. of hydro- 
chloric acid prescribed for the Clerget or Herzfeld inversion may be 
insufficient on account of the formation of chlorides and the liberation 
of the weakly inverting acetic acid. In such cases it is usual to em- 
ploy 6 c.c. of hydrochloric acid for making the inversion. 

Instead of the powdered salts above mentioned, concentrated sulphur- 
ous acid (prepared by saturating water with sulphur dioxide) has been 
proposed by Pellet for deleading. This reagent has certain advantages, 



METHODS OF INVERT OR DOUBLE POLARIZATION 277 

for, in addition to precipitating excess of lead, it neutralizes any free 
alkalinity and at the same time acts as a bleach upon any coloring 
matter which might darken the solution for reading. The sulphur 
dioxide has even been added to excess for deleading, sufficient quan- 
tity (10 c.c.) of the solution being taken to complete the volume from 
100 to 110 c.c. This excess does no harm, as the acid in the cold is a 
very weak inverting agent and has no immediate depressing influence 
upon the direct polarization. This excess of sulphurous acid has also 
the advantage of preventing the troublesome afterdarkening which fre- 
quently results from the inverting action of hydrochloric acid. Ogilvie* 
claims as another advantage an equalizing effect in the conditions before 
and after inversion in that both direct and invert polarizations are made 
in acid solution. It is evident, however, that the total quantity of acid 
is not the same in both cases and that these different amounts of acid 
will exercise a variable influence upon the rotation of fructose, amino 
compounds, etc. 

An objection against sulphur dioxide as a deleading agent is the very 
troublesome character of the lead-sulphite precipitate which, on ac- 
count of its finely divided colloidal condition, is very apt to pass through 
the filter. Agitating the solution with paper pulp, infusorial earth 
(kieselguhr), or kaolin previous to filtration has been recommended as 
a means of securing a clear filtrate. 

Decolorization of Inverted Solutions. The afterdarkening which 
results from the action of the hydrochloric acid upon coloring sub- 
stances, caramel, or other organic impurities, is frequently so great as 
to cause difficulty in reading the solution for the invert polarization. 
In such cases a number of expedients may be followed. 

(1) Use of a 100-wra. or 50-mm. Tube. Since shortening the 
length of the observation tube always necessitates a corresponding 
multiplication of any errors of observation this method is to be used 
only as a last resort. 

(2) Decolorization by Means of Bone Black. Animal charcoal or 
bone black should never be used upon solutions for direct polarization 
on account of its great absorptive power for sucrose. It may, how- 
ever, be employed with comparative safety upon solutions of invert 
sugar, provided the char be previously purified by washing with dilute 
hydrochloric acid and water and then dried. Two to five grams (de- 
pending upon the coloration of the solution) of the finely ground bone 
black are placed in the apex of a folded filter and the solution to be 
treated poured through in successive portions of about 10 c.c. The 

* Int. Sugar Jour. 13, 145. 



278 SUGAR ANALYSIS 

first 25 to 30 c.c. of filtrate are discarded and the remainder used for 
the invert polarization. 

(3) Decolorization by Means of Reducing Agents, Zinc Dust, 
Sodium Sulphite, Etc. A large number of reducing agents have been 
used for decolorizing acid solutions of invert sugar. Zinc dust has 
been frequently employed for this purpose, the destruction of coloring 
matter being due to the nascent hydrogen generated by the action of 
the hydrochloric acid upon zinc. The powdered metal is added to 
the solution to be decolorized in successive small amounts, thus pre- 
venting a too violent evolution of gas with loss of solution. 

Sodium sulphite and bisulphite have also been employed for decol- 
orizing acid invert sugar solutions. In this case the bleaching agent is 
the sulphur dioxide liberated by the action of the hydrochloric acid. 

The use of zinc and sodium sulphite as decolorizing agents is not 
attended with serious danger, provided only the minimum amounts be 
employed. 

General Reliability of the Clerget Method 

While the method of double, or invert, polarization gives perfectly 
reliable results upon pure sucrose, it is evident that the method has 
serious limitations when applied to the investigation of impure prod- 
ucts. The influence of mineral and organic impurities upon the 
specific rotations of sucrose and other sugars, and the lead-precipitate 
error affect all modifications of the Clerget process. The influence of 
hydrochloric acid upon the specific rotations of fructose and ammo 
compounds is an additional source of error in all modifications where 
the invert polarization is made in hydrochloric acid solution. Under 
such circumstances the chemist need not expect, under the most favor- 
able conditions, to obtain upon products containing a mixture of sucrose 
with reducing sugars, salts, and organic impurities an accuracy much 
greater than 0.5 per cent; in certain cases the error may exceed 1 per 
cent. The Clerget method gives therefore at best only an approximation, 
the degree of exactness depending not only upon the care and skill of 
the chemist, but also upon the nature of the substance being analyzed. 
The introduction of excessive refinements in the method has usually 
proved a thankless labor and is not to be recommended. The employ- 
ment, for example, of a Clerget factor elaborated to the fifth decimal 
(as in Tuchschmid's formula, p. 264) is of no possible value in practical 
work. 

In employing any of the numerous Clerget modifications it is 
always advisable for the chemist to establish his own factor for the 



METHODS OF INVERT OR DOUBLE POLARIZATION 279 

particular conditions of the analysis. This is best done by making a 
blank determination upon pure sucrose, or, better still, upon a mixture 
of pure sucrose with approximate amounts of the accompanying sub- 
stances which are known to occur in the product undergoing examina- 
tion. By so doing the chemist will gain an idea of the reliability of 
his method, such as can be secured in no other way. 

APPLICATION OF THE CLERGET METHOD TO THE DETERMINATION OF 
SUGARS IN PRESENCE OF SUCROSE 

When sucrose occurs in presence of another sugar, whose specific 
rotation is not affected by the inverting agent, and no other optically 
active substances are present, the percentage (Z) of the accompanying 
sugar may be determined as follows: 

If P is the direct polarization for the sucrose normal weight of sub- 
stance, and S the percentage of sucrose by the Clerget method, then 
P S is the polarizing power of the accompanying sugar. The per- 
centage Z may then be determined as upon page 200, by dividing the 
value 100 (P S) by the polarizing power of the accompanying sugar 
(Table XXXVI). The calculation may also be expressed in general 
terms by the equation 

_ 66.5 (P - S) 

~wT 

in which 66.5 is the specific rotation of sucrose and [ag that of the 
accompanying sugar. The method of calculation may be illustrated 
by several examples. 

Example I. A sirup containing sucrose and dextrose gave a direct 
polarization of -+ 58.0 and an invert polarization of 8.33 at 20 C. Required 
the percentages of sucrose and dextrose. 

Pe.ee. = 



Per cent dextrose = 66 - 5 ^ ~ 50) = 10 per cent. 

O-6.O 

Example II. A sirup containing sucrose and invert sugar gave a direct 
polarization of + 52 and an invert polarization of 21 at 20 C. Required 
the percentages of sucrose and invert sugar. 



Per cent _e = z - P cent. 

(EO 

( 
2,0 



Afi ^ (EO _ ^i^ 

Per cent invert sugar = : ( - * =10 per cent. 





280 SUGAR ANALYSIS 

Example III. A sweetened condensed milk (26 gms. in 100 c.c.) gave a 
direct polarization of -f 51.50 and after inversion in the cold a polarization of 
4.20 at 20 C. Required the percentages of sucrose and lactose. 

100 [51.50- (-4.20)] 5570 
Per cent sucrose = - I = = 41.99 per cent. 



66.5(51.50-41.99) 10n . 

Per cent lactose = - s - - - = 12.05 per cent. 

o2.o 

The percentages of sugars calculated in this manner have of course 
no greater degree of accuracy than the Clerget sucrose determination. 
With impure products clarified by means of basic lead compounds there 
may be an appreciable error due to the occlusion of reducing sugars in 
the lead precipitate. 

Method of Dubois for Determining Sucrose and Lactose in Milk 
Chocolate. Dubois* has applied the Clerget method to the deter- 
mination of sucrose and lactose in milk chocolate. The usual procedure 
is somewhat modified in that 100 c.c. of water are added to the 26 gms. 
of substance, a correction being afterwards applied for the increase in 
volume through solution of sugars. A preliminary extraction of the 
chocolate with ether to remove fat secures a more rapid solution of 
sugars. The following method of solution may also be used. 

Transfer 26 gms. of the finely ground chocolate to a flask, add 100 c.c. 
of water, cork and heat in a steam bath for 20 minutes, releasing the 
pressure occasionally during the first 5 minutes. Shake thoroughly 
twice during the heating so as to emulsify completely. Cool to room 
temperature, add 10 c.c. of lead-subacetate solution, mix and filter. 
After taking the direct polarization (a), delead the solution with dry 
potassium oxalate. Invert the deleaded solution according to Herz- 
f eld's method and take the invert polarization (6), correcting for 
dilution. Calculate the approximate percentages of sucrose (S) and 
lactose (L) by the following formulae: 

(a - 6) X 110 r _ (o X 1.10) - S 

~ " 



The approximate grams (G) of total sugar in the normal weight of 
chocolate are calculated from S and L, and the volume (X) of solution 
estimated by the formula X = 110 + (G X 0.62), in which 0.62 is the 
increase in volume caused by dissolving 1 gm. of sugar in water. The 
corrected percentages of sucrose and lactose are then found as follows: 

Q-rr T y 

True per cent sucrose = True per cent lactose = ^77)' 
* Or. 66, U. S. Bur. of Chem., p. 15. 



METHODS OF INVERT OR DOUBLE POLARIZATION 281 

The employment of an expansion factor, as in the above method, is 
permissible only in case of water-free substances and where no other in- 
gredients than sugars are dissolved. The factor 0.62 is not absolutely 
correct for all concentrations, as is seen from the following table: 



Sucrose dis- 




Increase in vol- 


Sucrose dis- 




Increase in vol- 


solved in 100 
c.c. water at 


Volume of resulting 
solution. 


ume through 
solution of 1 


solved in 100 
c.c. water at 


Volume of re- 
sulting solution. 


ume through 
solution of t 


20 C. 




gram sucrose. 


20 C. 




gram sucrose. 


Grams. 


c.c. 


c.c. 


Grams. 


c.c. 


c.c. 


1 


100.51 


0.506 


26 


115.98 


0.614 


2 


101 . 12 


0.560 


50 


130.94 


0.619 


5 


102.96 


0.592 


100 


162.37 


0.624 


10 


106.07 


0.607 


200 


225.82 


0.629 



The error attending the use of the factor 0.62 upon dilute solutions 
is so small as to be negligible. 



APPLICATION OF THE CLERGET PRINCIPLE TO THE DETERMINATION OP 

RAFFINOSE 

The principle of the Clerget inversion method may be applied to the 
analysis of any optically active substance whose specific rotation un- 
dergoes a known change with a special method of treatment. The 
most common application of the principle, outside of sucrose, is in the 
determination of the trisaccharide raffinose, the occurrence of which in 
sugar-house products, plant substances, etc., is referred to on page 732. 

The hydrolysis of raffinose with hydrochloric acid, under the condi- 
tions prescribed for the Clerget inversion, proceeds very closely according 
to the equation: 



Raffinose 

[a]g=+ 104.5 (for 
raffinose hydrate 
before hydrolysis) 



d-Fructose 

MS 



Melibiose 

- +143 



[ a ]= + 53.5 (for raffinose hydrate after 
hydrolysis) . 

The specific rotation of raffinose decreases during the hydrolysis 
from +104.5 for the hydrate to +53.5, which corresponds to that of 
a molecular mixture of fructose and melibiose (see note p. 737). The 
normal weight of raffinose for the Ventzke scale, using metric cubic 
centimeters, is 16.545 gms. for the hydrate and 14.037 gms. for the 
anhydride (see p. 197) . These amounts of raffinose, polarizing + 100 V., 
show after hydrolysis, following exactly the procedure of Herzfeld, a 
polarization of + 51.24 V. at 20 C., or a decrease of 48.76 V. This 
decrease for the weight of raffinose reading 1 V. (0.16545 gm. hydrate 



282 SUGAR ANALYSIS 

or 0.14037 gm. anhydride) is 0.4876 V. The calculation of raffinose 
by the hydrolysis method may then be expressed as follows : 

P-P r 



R = 



0.4876 



in which R is the percentage of raffinose, P the polarization of the 
normal weight of product before hydrolysis and P' the polarization of 
this normal weight after hydrolysis. 

APPLICATION OF THE INVERSION METHOD TO MIXTURES OF SUCROSE 

AND RAFFINOSE 

Raffinose is almost always associated in nature with sucrose, and 
since sucrose undergoes inversion simultaneously with the hydrolysis of 
raffinose, the formula previously given for the calculation of raffinose 
has but little practical value. Creydt,* however, showed that it was 
possible to combine the equations for the calculation of raffinose and 
sucrose, and in this way obtain formulae which can serve for the esti- 
mation of the two sugars in mixtures. The original formulae of Creydt 
were based upon the old Clerget process of inversion and have now 
been largely replaced by formulae worked out for the Herzfeldf modi- 
fication (p. 266). The method of establishing these formulae may be 
understood from the following: 

If the sucrose normal weight (26.00 gms.) of a substance contain- 
ing S per cent of sucrose and R per cent of raffinose (anhydride) be 
dissolved to 100 metric cubic centimeters and polarized in a 200-mm. 
tube, the polarization of the sucrose in degrees Ventzke will be repre- 
sented by S and the polarization of the raffinose by 1 .852 R (the value 

nn (\f\f\ 

1.852 being the ratio r of the normal weight for raffinose anhydride 
14.Uo7 

to that for sucrose). The direct polarization P (the sum of the sucrose 
and raffinose polarizations) is represented then by the formula 

P = S + 1.852 R, whence R = ^-^ and S = P - 1.852/2. (1) 

l.ooz 

If the sucrose normal weight of the above substance be inverted 
according to the Herzfeld method and polarized at 20 C., the invert 
polarization of the sucrose will be represented by -0.3266 S (since 1 V. 
sucrose before inversion reads 0.3266 V. at 20 C. after inversion). 
In the same manner the polarization of the raffinose after hydrolysis will 
be 1.852 R x 0.5124 = 0.9490 R (since 1 V. raffinose before hydrolysis 
reads +0.5124 V. at 20 C. after hydrolysis by Herzfeld's method). 

* Z. Ver. Deut. Zuckerind., 37, 153. t Ibid., 40, 194. 



METHODS OF INVERT OR DOUBLE POLARIZATION 283 

The invert polarization P r (the sum of the sucrose and raffinose invert 
polarizations) is represented then by the formula 

P'=- 0.3266 S + 0.9490 R. (2) 

p _ o 

By substituting the quantity -pcHo" f equation (1) for R in equa- 
tion (2), we obtain the formula 



whence _ 0.5124 P - P' 

0.839 
Having calculated S from P and P f , the value of R is obtained from 

r> _ o 

equation (1), R = 



By substituting the quantity P 1.852/2 of equation (1) for $ in 
equation (2), we obtain the formula 

P' = -0.3266 (P - 1.852/2) + 0.9490 R, 
whence 

p _ 0.3266 P + P' , . 

1.554 

By formula (4) the raffinose may be calculated at once, from the 
direct and invert polarizations. 

The method of employing the formulae may be understood from 
the following: 

A beet-molasses, free of reducing sugar, gave a direct polarization of 
+50 V. and an invert polarization of 12 V. Required the percentages of 
sucrose and raffinose. 

By formula (3), per cent sucrose = - 5124 * ^~ ( ~ 12) = 44.84 per cent. 



50 44 84 
By formula (1), per cent raffinose = ' - = 2.79 per cent, or 

L.OO& 

By formula (4), per cent raffinose = 0-3266 X 50 + (- 12) = 3.79 per cent. 

I.OO4 

Correction of Raffinose Formula for Changes in Temperature. - 

The determinations of sucrose and raffinose by the preceding formulae 
must be carried out at exactly 20 C. In case the analysis is made at 
other temperatures the formulae require to be modified. Several 
formulae have been worked out for correcting the invert polarizations 
of sucrose and raffinose for changes in temperature. Among the 
simplest of these are the formulae of Herles,* which are derived as 

follows: 

* Z. Zuckerind. Bohmen, 13, 559; 16, 528. 



284 



SUGAR ANALYSIS 



26.000 gms. of sucrose and 14.037 gms. of raffinose anhydride which 
read 100 per cent upon the saccharimeter before inversion, give after 
inverting by Herzf eld's method the following: 



Temperature. 


Inverted sucrose 
solution. 


Inverted raffi- 
nose solution. 


20 C 


-32 66 V. 


+51 24 


C 


42.66 


+47 24 








Difference for 20 C 
Difference for 1 C 


10.00 
0.50 


4.00 
0.20 



For sucrose and raffinose reading 1 per cent upon the saccharimeter 
before inversion, the reading after inversion is: 

For 1 per cent sucrose = -0.4266, at C. 

For 1 per cent sucrose = -0.4266 + 0.005 t, at t C. 

For 1 per cent raffinose = +0.4724, at C. 

For 1 per cent raffinose = +0.4724 + 0.002 t, at t C. 

The invert polarization for S per cent sucrose 

= (- 0.4266 + 0.0050- 
The invert polarization for R per cent raffinose 

= #(+0.4724 + 0.0020, 

or for the sucrose normal weight (26 gms.) 1.852 R (+0.4724 + 0.002 t). 
The invert polarization P' for S per cent sucrose and R per cent 
raffinose for 26.000 gms. to 100 c.c. would be 

P' = S (-0.4266 + 0.005 + 1.852 R (+0.4724 + 0.002 1). 

p o 

Substituting for R the value in equation (1), R = 



1.852 
P' = S (-0.4266 + 0.005 + (P - S) (0.4724 + 0.002 1). 



Whence 



and, as before, R 



S = 



P (0.4724 + 0.002 -' P' 



P-S 



0.899 - 0.003 1 
Equation (5) at 20 C. becomes necessaril 



(5) 



1.852 
the same as equation (3). 

Bone-black Error in Raffinose Determinations. A source o 
error peculiar to certain applications of the inversion method for d 
termining raffinose is the increase in levorotation after decolorizi 
inverted solutions by means of bone black. This error was fi 
studied by Reinhardt,* who attributed the phenomenon to the ab- 
* Z. Ver. Deut. Zuckerind., 62, 114. 






METHODS OF INVERT OR DOUBLE POLARIZATION 285 

sorption of the highly dextrorotatory melibiose. Reinhardt's expla- 
nation is no doubt correct as bone black shows a similar absorptive 
power for other disaccharides, such as sucrose. Davoll,* who has made 
a detailed study of methods for estimating raffinose, gives the follow- 
ing results upon a mixture containing 94.98 per cent pure cane sugar 
and 5.02 per cent raffinose hydrate (4.26 per cent raffinose anhydride). 
The direct polarization for a normal weight of this mixture was +102.48. 
The invert polarizations for different methods of treatment were as 
follows : 



Method of treatment. 


Invert polari- 
zation. 


Calculated sugars. 


Raffinose. 


Sucrose. 


Without char. 


-27.00 
-27.14 
-27.40 
-28.00 


Per cent. 
4.16 
4.11 

3.95 
3.56 


Per cent. 
94.77 

94.87 
95.16 
95.89 


Blood charcoal (purified with acid) 
Animal charcoal (highest purity) .... 


Animal charcoal (reagent) 








I 

II 



In the above experiments the solutions were shaken 5 minutes with 
3 gms. of char before filtering. Pouring the solutions in successive 
portions through the char with rejection of the first runnings (as de- 
scribed on p. 220) would no doubt reduce the error due to absorption 
considerably. 

As a remedy for the error due to the use of bone black Davoll 
proposes the employment of zinc dust as a decolorizing agent. At the 
end of the Clerget inversion 1 gm. of powdered zinc was allowed to act 
upon the acid solution at 69 C. for 3 to 4 minutes. Under these con- 
ditions the zinc was not found to affect the polarization of the inverted 
solution. 

General Reliability of the Optical Method for Estimating Raffinose 

The remarks (p. 278) made upon the limitations of the Clerget 
method apply with even greater force to the optical determination of 
raffinose. The method does not give accurate results, when optically 
active substances other than sucrose and raffinose are present. In 
cases where sucrose occurs with caramelization products, gums, and 
organic acids, application of the formula may indicate the presence of 
raffinose when in reality none is present. The formula should only be 
used in the investigation of substances in which raffinose is liable to 
occur (as sugar-beet products, cotton seed, etc.) and should never be 
* Proc. Fifth Int. Congr. Applied Chem., Ill, p. 135. 



286 SUGAR ANALYSIS 

employed, as is sometimes done, as a test for the presence of raffinose 
in unknown mixtures. 

As in the Clerget determination of sucrose the chemist need not 
expect in the analysis of commercial products for raffinose an accuracy 
much exceeding 0.5 per cent. The indication of a smaller amount of 
raffinose than 0.5 per cent is, in fact, not regarded by the best author- 
ities as sufficient to justify reporting its presence (as in raw beet sugars). 

Before applying the method to the analysis of unknown products 
the chemist should first satisfy himself of the presence of raffinose by 
suitable tests (see p. 740); he should also confirm the results of his 
analysis so far as possible by making blank determinations upon known 
mixtures. A practical test of this kind is the best means for testing the 
reliability of the method in particular cases. 



CHAPTER XI 

SPECIAL METHODS OF SACCHARIMETRY 

THE methods of inversion, described in the previous chapter, are 
only special instances of a more general course of procedure. It is pos- 
sible to calculate the percentage of any sugar, provided its rotatory 
power, in distinction from that of associated sugars, can be given a 
definite alteration by some special method of treatment. The changes 
produced in the rotation of sucrose and raffinose by the action of in- 
vertase or acids are but single illustrations of such special methods of 
treatment. As other examples may be mentioned (1) the determina- 
tion of sugars by noting the change produced in polarization under 
different conditions of temperature. (2) The determination of sugars, 
by noting the change in polarization after fermenting with yeast. 
(3) The determination of sugars by noting the change in polarization 
after destroying the optical activity of reducing sugars. Numerous 
other examples might be given but the three cases cited are sufficient 
to illustrate the general application of the principle to special problems 
of saccharimetry. 

DETERMINATION OF SUGARS BY POLARIZATION AT HIGH TEMPERATURE 
DETERMINATION OF INVERT SUGAR BY HIGH-TEMPERATURE 

POLARIZATION 

The principle of this method is based upon the fact that solutions 
of pure invert sugar, when heated to a temperature between 85 and 
90 C., become optically inactive. This inactivity is due to the lowering 
in specific rotation of fructose with increase in temperature (page 179) ; 
the specific rotation of glucose being unaffected by temperature, the 
point of optical inactivity will be the degree at which the polarizing 
powers of glucose and fructose exactly neutralize each other. 

Temperature of Optical Inactivity of Invert Sugar. The temper- 
ature of optical inactivity of invert sugar has been variously estimated. 
Dubrunfaut,* who made the earliest measurements of this constant, 
set the figure at 90 C. Casamajorf and Wiley J have given 88 C., 

* Compt. rend., 42, 901. 
t Chem. News, 44, 219. 
j J. Am. Chem. Soc., 18, 81. 

287 



288 



SUGAR ANALYSIS 



Lippmann,* 87.8 C., Wolf,f 87.6 C. and Tuchschmid,t 87.2 C. 
These variations may be due in part to slight experimental errors 
(such as incipient destruction of sugar at the high temperature) and in 
part to the influence of concentration. Inasmuch as the [O\D of glucose 
varies from + 52.5 for a 1 per cent solution to + 54.0 for a 40 per cent 
solution it is evident that the temperatures at which these different 
polarizations are neutralized must vary somewhat. 

The effect of concentration upon the temperature of optical in- 
activity for invert sugar may be determined by means of the carefully 
established formulae of Gubbe. || 

I Concentration [a] = - 19.657 - 0.0361 c. 
II Temperature 

(20 to 100 C.) [aY D = [a] + 0.3246 (t - 20) - 0.00021 (t - 20) 2 . 

In Table LIII, column B gives the [a] of invert sugar, as calcu- 
lated by formula I, for different concentrations; column C gives the 
grams of invert sugar in 100 c.c. necessary to produce a reading of 

1729 



1 V., as calculated by the expression 



1UU 



(page 197); column D 



gives the temperature of optical inactivity, as determined by formula II 
of Gubbe; column E gives the variation in degrees Ventzke, produced 
by 1 gm. of invert sugar in 100 c.c. for 1 C. difference in temperature 

and is calculated by the expression n , - p^- 

C (D 20) 

TABLE LIII. 



A 


B 


C 


D 


B 


Concentration, 
grams invert sugar 
in 100 c.c. 


ll 


Invert sugar in 100 
c.c. corresponding 
to -1 V. at20 9 C. 


Temperature of op- 
tical inactivity. 


Variation for 1 
gram invert sugar 
for 1C. 


Grams. 




Grams. 


Deg. C. 


De ? . V. 


2 


-19.72 


0.8768 


83.2 


0.01805 


10 


-20.02 


0.8636 


84.2 


0.01804 


20 


-20.38 


0.8484 


85.4 


0.01802 


30 


-20.74 


0.8336 


86.6 


0.01801 


40 


-21.10 


0.8194 


87.8 


0.01800 


50 


-21.46 


0.8057 


89.0 


0.01799 


60 


-21.82 


0.7924 


90.2 


0.01798 



For general purposes 87 C. is usually taken as the temperature of 
optical inactivity for invert sugar. 



* Ber., 13, 1823. 

t Oest. Ung. Z. Zuckerind., 16, 331. 

t J. prakt. Chem. [2], 2, 235. 

|| Ber., 18, 2207. 



SPECIAL METHODS OF SACCHARIMETRY 289 

The application of the method to the determination of invert sugar 
is easily understood. Since a change of 1 C. produces a constant 
variation of 0.018 V. for 1 gm. of invert sugar in 100 c.c., regardless of 
the concentration, then the grams of invert sugar in 100 c.c. of a given 
solution is found by the formula 

Invert sugar = ~ 



in which P f = Ventzke-scale reading at higher temperature t f , 
and P Ventzke-scale reading at lower temperature t. 

The method of applying the formula may best be understood by 
taking a typical example. 

Example. 50 gms. of a solution, containing a mixture of glucose and 
fructose in unequal amounts, were made up to 100 c.c. at 20 C. The polariza- 
tion was + 10.20 V. at 20 C. in a 200-mm. tube. 

50 gms. of the same solution were made up to 100 c.c. at 87 C. The 
polarization was + 20.75 V. at 87 C. in a 200-mm. tube. Required the per- 
centage of sugars in the original solution. 



Invert sugar -- 8.75 gn.. 



Q 7 P\ 

-^ X 100 = 17.50 per cent invert sugar. 
50 

The dextrorotation at 87 C. shows an excess of glucose over the amount 
necessary to be paired with the fructose for invert sugar. This excess of glucose 
may be estimated as follows : 

Since 1 V. = 0.3225 gm. glucose (page 200) then the grams of glucose cor- 
responding to the dextrorotation at the inactivity of invert sugar is 20.75 X 
0.3225 = 6.69 gms. (uncorrected for concentration), or 13.38 per cent. To 
correct for the influence of concentration, the true glucose value of the Ventzke- 
scale reading -f 20.75, according to the formula G = s -f 0.02 s 0.0002 s 2 , 
(page 199) is 21.08. 21.08 X 0.3225 = 6.80 gms. glucose or 13.60 per cent in 
the original solution. 

The percentage of glucose determined by this method of calculation can, of 
course, be considered as only approximate, for, as shown in Table LIII, the 
temperature of optical inactivity, according to concentration, may be above or 
below 87 C. 

DETERMINATION OF COMMERCIAL GLUCOSE BY HIGH-TEMPERATURE 

POLARIZATION 

Method of Chandler and Ricketts. The method of high-tem- 

perature polarization as first developed in 1880 by Chandler and Rick- 

etts * was not employed for determining invert sugar but for detecting 

the presence and estimating the amount of commercial glucose in cane 

* J. Am. Chem. Soc., 2, 428. 



290 



SUGAR ANALYSIS 



sugar, molasses, honey and other products whose sugars, after inversion, 
consist almost wholly of invert sugar. The material under examina- 
tion was first inverted to convert any sucrose to invert sugar and then 
polarized at the temperature of optical inactivity for invert sugar. Any 
dextrorotation observed at this temperature was attributed to com- 
mercial glucose and its percentage estimated by means of an empirical 
factor. 

The factor for converting the readings of the Ventzke sugar scale 
into grams of commercial glucose depends entirely upon the nature of 
the product. Commercial glucose, as manufactured in the United 
States, varies in density from 41 Be. to 45 Be. (sp. gr. 1.388 to 1.442) 
and in specific rotation from about [a:]^ + 100 to + 125 for the liquid 
product. The grams of commercial glucose corresponding to 1 V. for 
products of different specific rotation are given in Table LIV. 

TABLE LIV. 



M D (for 
liquid prod- 
uct). 


Polarization 
(deg. V. of 
26 grams to 
100 truec.c.). 


Grams of liquid 
product in 100 c.c. 
corresponding to a 
polarization of 1 V. 


Mo (for 

liquid prod- 
uct). 


Polarization 
(deg. V. of 
26 grams to 
100 true c.c.). 


Grams of liquid 
product in 100 c.c. 
corresponding to a 
polarization of 1 V. 


+ 125 

+ 120 
+ 115 
+ 110 


+ 188.0 

+180.5 
+172.9 
+ 165.4 


0.1383 
0.1440 
1503 
0.1572 


+ 108 
+ 105 
+ 100 


+ 162.5 

+157.9 
+150.4 


0.1600 
0.1647 
0.1729 



For purposes of analysis the products of [oj^ +108 may be taken as 
the grade of commercial glucose most commonly used. The chemist 
should always state the polarizing power of the commercial glucose in 
terms of which his results are expressed. 

The form of polariscope devised by Chandler and Ricketts for high- 
temperature polarization is shown in Fig. 146. The instrument con- 
sists of an ordinary saccharimeter, with trough removed and replaced 
by a water bath which is heated from below by means of gas or spirit 
lamps. The ends of the water bath, before the diaphragms of the ana- 
lyzer and polarizer, are provided with metallic caps containing small 
windows of plate glass. The polarization tube, which in its earliest 
form was constructed of platinum, is completely immersed in the 
water of the bath, and rests upon supports opposite the windows and in 
perfect alignment with the axis of the instrument. The tube is pro- 
vided with an upright tubule for inserting a thermometer and for re- 
ceiving any excess of liquid displaced by expansion. The cover of the 
bath, which fits over the tubule, contains an opening for a thermometer 
to determine the temperature of the bath. 



SPECIAL METHODS OF SACCHARIMETRY 



291 



The use of a special type of saccharimeter for high-temperature 
polarization has been largely discontinued. At present it is customary 
to make the polarizations upon an ordinary type of saccharimeter, em- 
ploying a metal- jacketed tube; the latter may be insulated to advan- 
tage by a mantle of asbestos or other non-conducting material. The 




FIG. 146. Chandler and Ricketts's polariscope for high-temperature polarization. 

hot water for heating the tube is conveyed by rubber tubing from a 
water-heater, which should be placed at a distance sufficient to prevent 
heating the polariscope. A convenient arrangement for this purpose, 
described by Leach,* is shown in Fig. 147. 

Method of Leach. The following description of a method for 
determining commercial glucose in molasses, sirups, honey, etc., is given 
by Leach. f 

* "Food Inspection and Analysis" (1911), p. 644. 

t Bull. 81, U.S. Bur. of Chem., p. 73. Bull. 107 (revised), U.S. Bur. of Chem., 
p. 74. 



292 



SUGAR ANALYSIS 



"Invert a half-normal portion in the usual manner in a 100-c.c. 
flask; after inversion, cool, add a few drops of phenolphthalein and 
enough sodium hydroxide to neutralize; discharge the pink color with a 
few drops of dilute hydrochloric acid, add from 5 to 10 c.c. of alumina 
cream, and make up to the mark and filter. Multiply by 2 the read- 
ing at 87 C. in the 200-mm. tube; multiply this result by 100 and 




Fig. 147. Apparatus for polarizing at high temperatures. 

divide by the factor 163 to express the commercial glucose in terms of 
glucose polarizing + 175 V." * 

In the above method the solution is made up at room temperature 
and polarized at 87 C. When this is done a correction must be made 
for the expansion of the solution and consequent lowering of the 
reading. The best method of making this correction is by means of 
an empirical test. Thus Lythgoe,f following the above course of 

* Provisional Method of the Association of Official Agricultural Chemists, 
Bull. 107 (revised), U.S. Bur. of Chem., p. 71. 
t Bull. 81, U.S. Bur. of Chem., p. 74. 



SPECIAL METHODS OF SACCHARIMETRY 



293 



procedure, obtained the following results upon five samples of commer- 
cial glucose. 



Sample. 


Density. 


Polarization (26 gms. in 100 c.c.). 


Ratio f. 


f 

Ratio 2> 


A 


B 


C 


Direct. 


Invert at 22 C. 


Invert at 87 C. 


1 
2 
3 
4 
5 


Deg. Be. 
42 

42 
42 
43 
45 


Deg. V. 
156.6 

158.6 
169.6 
167.4 
174.0 


Deg. V. 
153.4 

154.6 
165.4 
162.8 
171.0 


Deg. V. 
146.6 
149.0 
159.4 
155.0 
161.2 

Average 


0.956 
.964 
.964 
.952 
.943 


0.936 
.940 
.940 
.926 
.927 


.956 


.933 



It is seen that the polarization of commercial glucose is slightly 
lowered by the action of the acid during inversion, as well as by the 
expansion of the solution upon heating to 87 C. To correct for both 
of these influences, the polarization value of the glucose is multiplied 
by the factor 0.933. The Association of Official Agricultural Chemists 
expresses glucose in terms of a product polarizing 175 V. for a weight 
of 26 gms. in 100 c.c. and this polarization corrected gives 175 X 0.933 = 
163 which is the factor employed in the calculation. 

Example. 13 gms. of a sample of table sirup inverted according to Herz- 
feld's method and made up to 100 c.c. at 20 C. polarized +65.2 V.at 87 C. 
Required the percentage of commercial glucose in terms of a product polarizing 

+ 162.5 V.for 26 gms. in 100 c.c. 

65.2 



The factor for 162.5 is 162.5 X 0.933 
per cent commercial glucose. 



151.6. Then 



151.6 



X 100 = 43.0 



Dextrorotation of Inverted Honey at 87 C. The method of 
estimating commercial glucose in honeys, sirups, molasses, etc., by 
polarizing at 87 C., can be regarded only as an approximate one. 
The chief limitation of the method is the fact that pure honeys, mo- 
lasses, sirups, etc., are more or less dextrorotatory, after inversion, at 
87 C., owing to the presence of gums, dextrins, or other similar com- 
pounds. 

Table LV, which is taken from the work of Browne,* gives the 
polarization of various samples of American honey at 20 and 87 C., 
before and after inversion. 

* " Chemical Analysis and Composition of American Honeys," Bui. 110; U. S. 
Bur. of Chem. 



294 



SUGAR ANALYSIS 



TABLE LV 



Kind of honey. 


Num- 
ber 
samples 
aver- 
aged. 


Direct polarization. 


Invert polarization. 


20 C. 


87 C. 


20 C. 


87 C. 


Difference. 


Levorotatory Class: 
Mangrove 


1 

3 
4 
8 
2 
2 
15 
3 
2 
3 
2 
6 
1 

1 
1 
1 
1 

92 

7 


Deg. V. 

-24.80 
-20.93 
-17.61 
-15.10 
-16.80 
-17.50 
-13.01 
-12.33 
-12.40 
-10.47 
-8.55 
-8.90 
-4.90 

+3.60 
+7.80 
+ 11.00 

+ 17.75 

-14.73 
+9.43 


Deg. V. 

+0.50 

+4.45 
+6.80 
+9.63 
+8.20 
+6.80 
+11.65 
+10.87 
+ 13.00 
+ 12.53 
+ 17.00 
+ 15.05 
+ 17.80 


Deg. V. 

-27.94 
-25.01 
-22.85 
-22.99 
-20.41 
-21.01 
-17.77 
-16.43 
-18.92 
-14.01 
-13.73 
-12.25 
-9.68 

-2.53 
+3.4?. 
+5.17 
+ 13.53 

-19.16 

+5.47 


Deg. V. 

-0.66 
+2.83 
+4.70 
+5.00 
+5.94 
+6.05 
+9.25 
+9.35 
+9.51 
+ 11.51 
+ 12.76 
+13.62 
+ 15.40 

+20.90 
+26.62 
+28.60 
+34.76 

+7.91 
+27.56 


27.28 
27.84 
27.55 
27.99 
26.35 
27.06 
27.02 
25.78 
28.43 
25.52 
26.49 
25.87 
25.08 

23.43 
23.21 
23.43 
21.23 

27.07 
22.09 


JVlesouit 


Sweet clover 


Alfalfa 


Buckwheat 


Cotton 


White clover 


Goldenrod 


Dandelion 


Sumac 


Apple 


Basswood 


Whitewood 


Dextrorotatory Class: 
Poplar 


Hickory . 


+28.50 
+32.30 


White oak . 


Sugar-cane honey dew. 

Levorotatory honeys 
Dextrorotatory honeys . . 


+ 10.15 
+32.20 


Average of 50 varieties 


99 


-13.02 


+10.81 


-17.41 


+9.30 


26.71 



100 P 
Application of the formula fi< ^ to the invert polarizations at 

87 C. would indicate nearly 10 per cent commercial glucose in some of 
the levorotatory and nearly 20 per cent in several of the dextrorotatory 
honeys. 

Browne's Method for Estimating Commercial Glucose in Honey. - 
Browne* has modified the application of the high-temperature polariza- 
tion, for estimating commercial glucose in honeys, by taking the differ- 
ence between the invert polarization at 20 and 87 C. as a basis of 
calculation. It is seen from Table LV that while the invert readings 
at either 20 or 87 C. are subject to the widest variations, the differ- 
ence between the polarizations at these two temperatures is a fairly 
constant quantity for nearly all honeys. The average value of this 
constant for the 99 samples of honey examined by Browne was 26.7. 
Since this difference in polarization is due entirely to the percentage of 
invert sugar in the honey, the addition of any commercial glucose will 

* "Chemical Analysis and Composition of American Honeys," p. 60, Bui. 110; 
U. S. Bur. of Chem. 



SPECIAL METHODS OF SACCHARIMETRY 295 

cause a depression in the polarization difference, which will be pro- 
portional to the amount of commercial glucose used but irrespective of 
its specific rotation. In order to correct for the variations in moisture 
and non-sugars of pure honey it is better to express the polarization 
difference in terms of a uniform basis of 77 per cent reducing sugars, 
which is the average percentage of invert sugar after inversion for pure 
honey. The formulae for making the calculation are then: 

100 (P f - P) X 77 288.4 (P' - P) 
Per cent pure honey = 257 XI ~V~ 

Per cent commercial glucose = 100 ^j 

in which P' = the Ventzke polarization of the inverted honey at 87 C. 
P = the Ventzke polarization of the inverted honey at 20 C. 
7 = the per cent of invert sugar in the honey after inversion. 
Another method, used in European countries, for estimating the 
amount of commercial glucose in honey is based upon the variation in 
the invert polarization of the sample from that of pure honey. Calling 
the average invert polarization of pure honey 17.5 at 20 C. (Table 
LV) and employing the official figure + 175 V. for the polarization of 
commercial glucose, then if 

x = per cent of honey in sample, 
y = per cent of commercial glucose in sample, 
P = invert polarization of sample in degrees Ventzke, 
x + y = 100. 
- 0.175 x + 1.75 y=P 

P + 17.5 
y= -T93T 

This method of calculation, the same as that based upon the polari- 
zation at 87 C. , makes no allowance for the wide range in the invert 
polarization of individual honeys (30 to + 15), so that a considerable 
error may be introduced in the final result. 

In Table LVI the polarizations of 5 honeys and of mixtures of the 
same, with 20 per cent commercial glucose, are given together with the 
percentage of commercial glucose as calculated by the three methods 
described. 

It will be seen from the results in the table that with admixtures of 
low-purity honeys and commercial glucose there is a considerable error 
in the calculation of the percentage of added adulterant. The results 
obtained by any method for estimating commercial glucose have only 
an approximate value, and in no case ought such analytical results as 



296 



SUGAR ANALYSIS 



those obtained for the pure basswood or white-oak honey to condemn a 
sample as being adulterated. In all suspicious or doubtful cases con- 
firmatory qualitative tests such as that with iodine should be employed." 

TABLE LVI * 

Polarization of Honeys and Commercial Glucose Mixtures, with Calculated Percent- 
ages of Glucose by Different Formulce. 





.1 


Invert polariza- 
tion. 


I 
1 


a 
1 


hi 


Calculated glucose. 










--.' 






'^u 


P 


P' 


'3 


"3*- 


; 3*- a 






I 




Kind of sample. 


~0 






Jl 


k 


S-* 


. 


W5 


fl 






* 






^OH 


If 


1> 

.2 C -S 


Si 


+ 2 








t-i 






c3 




c3 *"* ^ 




a. 


Csl 






(5 


20 C. 


87 C. 


1 


> 


2" 






| 






Deg. 


Deg. 


Deg. 


Deg. 


Per 


Deg. 


Per 


Per 


Per 




V. 


V. 


V. 


V. 


cent. 


V. 


cent. 


cent. 


cent. 


Alfalfa 


19 5 


22 66 


+ 3 52 


26 18 


77 84 


25 90 


2 16 


00 


^ 


on 


Alfalfa+20 per cent glucose 


+19.4 


+16.88 


+35.82 


18.94 


70.01 


20.83 


21.97 


17.82 


21 


!)8 


Hop vine 


12 6 


16 83 


+ 9 68 


26 51 


75 83 


26 92 


5 94 


35 




00 


Hop vine+20 per cent glucose 


+24.9 


+21 54 


+40 74 


19 20 


68 14 


21 70 


25 00 


20 28 


18 


72 


Whitewood 


- 4.9 


- 9.68 


+15.40 


25.08 


71.88 


26.87 


9.45 


4.06 




00 


Whitewood+20 per cent ^lucose 


+31 1 


+27 26 


+45 32 


18 06 


64 99 


21 40 


27 80 


23 25 


U 


^"i 


Basswood 


- .3 


- 1.32 


+23.21 


24.53 


70.60 


26.75 


14.24 


8.40 




00 


Basswood +20 per cent glucose 


+ 3 48 


+33 94 


+51 57 


17 63 


63 97 


21 22 


31 64 


26 72 


80 


53 


White oak 


+11.0 


+ 5 17 


+28 60 


23 43 


70 44 


25 61 


17 56 


11 23 


4 


08 


White oak+20 per cent glucose 


+43.8 


+39.14 


+55.88 


16.74 


63.84 


20.20 


34.28 


29.35 


24 


.35 



1 " Chemical Analysis and Composition of American Honeys," Bui. 110, U. S. Bur. of Chem., p. 61. 

Dextrorotation of Inverted Molasses at 87 C. The observa- 
tions made upon the dextrorotation of inverted honey at 87 C. also 
pertain to sugar-cane molasses and sirups, but to a much less degree. 
Eighteen samples of Louisiana sugar-cane molasses, of known purity, 
examined by Bryan,f gave an average direct polarization at 20 C. of 
+ 40.6 V., an average invert polarization at 20 C. of 17.8 V. and an 
average invert polarization at 87 C. of + 2.53, the range of the latter 
being from 0.0 to + 4.18, or an equivalent of to 2.5 per cent commer- 
cial glucose. 

DETERMINATION OF FRUCTOSE BY POLARIZATION AT LOW AND 
HIGH TEMPERATURES 

Method of Wiley. A second illustration of the methods of high- 
temperature polarization is afforded by Wiley's t method for estimating 
fructose. In his description of this method Wiley shows that 1 gm. of 
fructose in 100 c.c. of solution gives a variation of 0.0357 V. for each 
I C. difference in temperature. The grams of fructose present in 
100 c.c. of any solution can be calculated, therefore, from the polariza- 

t Bull. 122, U. S. Bur. of Chem, p. 182. 

t Wiley's "Agricultural Analysis" (1897), 3, 267. 



SPECIAL METHODS OF SACCHARIMETRY 



297 



tions made at two widely separated temperatures by means of the 
formula. 

F= P '~ P 

0.0357 (Z'-O' 

in which F = grams of fructose in 100 c.c. of solution. 

P' = Ventzke polarization at high temperature t'. 
P = Ventzke polarization at low temperature t. 

The factor 0.0357 employed by Wiley is confirmed by the observa- 
tions of other investigators as shown in Table LVII. 

TABLE LVII. 
Showing Change of Polarization of Fructose for 1 C. Change of Temperature 





A 


B 


C 


Observer. 


Change in [a} D 
of fructose per 
1C. 


Change in rotation 
for a fructose solu- 
tion reading 100V. 
per 1 C. 
100 A 


Change in rotation 
for 1 gram fructose 
in 100 c.c. per 1 C. 
B 






92.5 


18.692 


Dubrunfaut* . . 


62 


6702 


03586 


Honig and Jesserf . . . 


68 


7351 


03933 


Jungfleisch and Grimberti 


56 


6054 


03239 


Gubbe 


63 


6811 


03644 


Tuchschmid || 


64 


6919 


03702 










Average 


626 


6767 


03621 











The average value 0.0362 is practically identical with that of Wiley. 

Another method of determining the variation in the Ventzke 
polarization of fructose for changes in temperature is by means of 
Gubbe's equations (page 288). Since the specific rotation of glucose is 
not affected by changes in temperature, the results of Table LIII are 
converted into terms of fructose by dividing the values of columns A 
and C, and by multiplying those of column E, by two. The variation 
in polarization of 1 gm. of fructose in 100 c.c. for 1 C. change in tem- 
perature, as thus determined, is 0.0360 V., which value is constant for 
all concentrations. This quantity, which is also the average of Wiley's 
figure and that of Table LVII, may be accepted as the most probable 
value. 

* Compt. rend., 42, 901. 

t Z. Ver. Deut. Zuckerind., (1888), 1028. 

t Compt. rend., 107, 390. 

Z. Ver. Deut. Zuckerind., 34, 1345 ; calculated from results for invert sugar. 

II J. prakt. Chem. [2], 2, 235; calculated from results for invert sugar. 



298 SUGAR ANALYSIS 

If 26 gms. of product are made up to 100 c.c. and polarized (P) at a 
low temperature t, and a second 26 gms. are made up to 100 c.c. and 
polarized (P') at a high temperature t', then the percentage of fructose 
F is determined by the equation 

IQOCP'-P) _100(P / -P) 
" 26 X 0.036 (t f -t) = 0.936 (' - t) ' 

Example. 26 gms. of honey made up to 100 c.c. and polarized at 20 C. 
gave a reading of 14.8 V. 26 gms. of the same honey made up to 100 c.c. 
and polarized at 87 C. gave a reading of -f- 10.50 V. Required the percent- 
age of fructose. 



In making polarizations at high temperatures it is desirable to make 
the readings as soon as the solution in the tube has reached tempera- 
ture equilibrium, as indicated by the thermometer placed in the solution 
and by the disappearance of striations from the field. After noting the 
polarization the temperature is again taken and the average thermom- 
eter reading used in the calculation. Prolonged heating at high tem- 
peratures causes a destruction of fructose. A difficulty is sometimes 
experienced in obtaining a clear unobscured field of vision when using 
the hot-water polariscope tube. Too slow a circulation of hot water 
through the jacket of the tube, with production of currents of unequally 
heated solution, is the usual cause of the trouble. The hot water should 
be several degrees above the desired temperature and the circulation 
must be rapid enough to prevent loss of heat by radiation. 

Limitations of Methods of High-temperature Polarization. The 
method of determining invert sugar or fructose by polarization at 
widely-separated temperatures, while giving good results upon dilute 
solutions of the pure sugars, gives only an approximation in case of 
many sugar mixtures. The method is strictly applicable only when the 
specific rotations of the accompanying sugars are unaffected by changes 
in temperature; in all other cases there will be a certain error in the 
determination depending upon the temperature coefficient and the per- 
centage of other sugars present. While no other sugars are affected to 
the same extent as fructose, yet it must be remembered that 1.5 gms. 
arabinose, or 3.0 gms. galactose, or 7.0 gms. maltose, or 9.0 gms. lactose, 
or 50 gms. sucrose produce approximately the same alteration in the 
Ventzke reading with 1 C. variation in temperature as 1 gm. of fructose, 
or 2 gms. of invert sugar. 

But notwithstanding this limitation the method of high-temperaturo 
polarization has a distinctive value, and, when employed with due 



SPECIAL METHODS OF SACCHARIMETRY 299 

caution, will be found of great service in many problems of analysis 
and research. 

DETERMINATION OF SUGARS BY POLARIZATION BEFORE AND AFTER 

FERMENTATION 

By employing pure cultures of specially selected organisms, it is 
sometimes possible to ferment one or more sugars of a given mixture, 
and from the variation in polarization thus produced to calculate the 
percentage of one or more of the members present. 

Action of Pure Yeast Cultures upon Different Sugars. The 
fermentative action of various yeasts upon different sugars has been 
studied by Tollens and Stone,* Hansen,f Fischer and Thierf elder, J and 
many others. The results of their experiments show a pronounced 
selective action on the part of different yeasts. While pure cultures of 
such well-known yeasts, as Saccharomyces cerevisice, or Saccharomyces 
Pastorianus, ferment completely d-glucose, d-fructose, d-mannose, 
d-galactose, sucrose, and maltose, these cultures are without action 
upon 1-xylose, 1-arabinose, rhamnose, sorbose and lactose. A " milk- 
sugar yeast," employed by Fischer and Thierf elder, fermented lactose 
and sucrose completely but did not attack maltose. Saccharomyces 
apiculatus ferments d-glucose, d-mannose and d-fructose but not 
galactose, sucrose, maltose or lactose. (See also Table CII, page 714.) 

Method of Fermentation. In carrying out experiments for the 
separation of sugars by fermentation it is very essential that the culture 
of particular yeast be pure. The presence of foreign yeasts, moulds or 
bacteria may produce changes in sugars, which a pure culture would 
leave unattacked. The solution to be fermented should be sterilized 
before inoculating. 

The most favorable conditions for the action of the yeast are obtained 
with a solution containing about 10 per cent sugar and kept at a tem- 
perature of about 30 C. It is also necessary, in order to secure a 
rapid and complete fermentation, to have a suitable supply of nutritive 
matter present for the growth and sustenance of the yeast. A food 
supply for yeast in fermentation experiments is generally furnished by 
means of a nutritive salt solution or by means of yeast extract. 

Hayduck's Nutritive Salt Solution. Dissolve 25 gms. potassium 
phosphate, 8 gms. crystallized magnesium sulphate and 20 gms. aspara- 
gine in 1000 c.c. of spring water. 

One cubic centimeter of the above solution to each 25 c.c. of liquid 
to be fermented insures a favorable development of yeast. 

* Ann., 249, 257. t Centralblatt, 88, 1208, 1390. J Ber., 27, 2031. 



300 SUGAR ANALYSIS 

Yeast Extract. Wash 100 gms. of pure yeast (starch-free) re- 
peatedly with cold water and repress. The residue of yeast is then 
heated to boiling for one-fourth hour with 500 c.c. of water; the 
liquid is then filtered through a folded filter, the filtrate, in case of 
turbidity, being returned to the filter until the extract runs through per- 
fectly clear. The extract is then made faintly acid with 
citric acid, when it is sterilized and preserved in flasks closed 
by cotton wadding. 

The liquid to be fermented is diluted with an equal volume 
of the above extract. 

Fermentation experiments are best carried out in flasks 
closed with a washing tube for the escape of carbon dioxide. 
The apparatus shown in Fig. 148 answers very well for the 
purpose. The fermentation is continued until bubbles of gas 
cease to pass through the water in the washing tube, when 
the process is considered to be finished. The washing tube 
is then removed, the solution heated to expel all carbon 
dioxide, and, after cooling, clarified, and the volume com- 




. 

tion flask. The polarization of the filtered solution is calculated 

to unfermented sugar, and the difference in polarization, 
before and after fermentation, calculated to fermented sugar. The 
application of the method is best understood from a special case. 

Example. By hydrolyzing a sample of sawdust with sulphuric acid, 
treating the resultant liquid with an excess of powdered calcium carbonate, 
filtering and evaporating, a sirup resulted which contained the two sugars, 
glucose and xylose. 

50 gms. of the sirup, made up to 100 c.c., gave a polarization of + 43.5 V. 
in a 200-mm. tube. 

50 gms. of the sirup were then diluted in a 200-c.c. flask with 100 c.c. of 
water and 5 c.c. of nutritive salt solution. After sterilizing, cooling and in- 
oculating with pure-yeast culture, the flask was closed with a washing tube and 
fermented for 5 days in an incubator at 30 C. The evolution of gas having 
ceased, the solution was heated to expel C0 2 , cooled, clarified with a little 
normal acetate of lead solution, made up to 200 c.c., and filtered. The polariza- 
tion of the filtrate in a 400-mm. tube was + 5.2 V. Required the percentages 
of glucose and xylose in the sirup. 

The loss in polarization by fermenting was 43.5 - 5.2 = 38.3 V. Since 
1 V. = 0.3225 gms. glucose in 100 c.c. then the grams of glucose fermented 
were 38.3 X 0.3225 = 12.35 gms. or 24.7 per cent glucose (unconnected) in 
the sirup. 

Since 1V.= 0.91 gms. xylose in 100 c.c., then, calling the residual 



SPECIAL METHODS OF SACCHARIMETRY 301 

polarization of + 5.2 as due entirely to xylose, 5.2 X 0.91 = 4.73 gms. or 9.46 
per cent xylose (uncorrected) in the sirup. 

Corrections for concentration are made as indicated on page 198. 

Determination of Dextrin in Fruit Products. The fermentation 
method is sometimes employed for the determination of dextrin in 
jams, jellies and other products, which might be adulterated with com- 
mercial glucose. The provisional method of the Association of Official 
Agricultural Chemists is as follows : * 

" Dissolve 10 gms. of the sample in a 100-c.c. flask, add 20 mgs. of 
potassium fluoride, and then about one-quarter of a cake of compressed 
yeast. Allow the fermentation to proceed below 25 C. for two or 
three hours to prevent excessive foaming, and then place in an incuba- 
tor at a temperature of from 27 to 30 C. for five days. At the end 
of that time, clarify with lead subacetate and alumina cream, make up 
to 100 c.c. and polarize in a 200-mm. tube. A pure fruit jelly will 
show a rotation of not more than a few tenths of a degree either to the 
right or to the left. If a polariscope having the Ventzke scale be used 
and a 10 per cent solution be polarized in a 200-mm. tube, the number 
of degrees read on the sugar scale of the instrument multiplied by 
0.875 will give the percentage of dextrin, or the following formula may 
be used: 

Percentage of dextrin = 198 xLXW 

in which 

C = degrees of circular rotation. 
L = length of tube in decimeters. 
W = weight of sample in 1 cubic centimeter." 

The factor 0.875 is found as follows: Calling + 198 the [a] D of dex- 
trin, then the grams of dextrin (D) in 100 c.c. of solution are found from 
the Ventzke reading (7) in a 200-mm. tube by the formula: 



If 10 gms. of product are made up to 100 c.c. then the percentage of 

087^ V 
dextrin in the sample = ^ X 100 = 0.875 V. 

The use of potassium fluoride in the method just described is to 
prevent the development of bacteria. Its employment is not necessary 
when pure-yeast cultures are used and the solution to be fermented has 
been previously sterilized. 

* Bull., 107 (revised) U. S. Bur. of Chem., p. 80. 



302 SUGAR ANALYSIS 

The work of Brown and Morris * shows that the dextrins and malto- 
dextrins of starch conversion are not fermented by Saccharomyces cerevi- 
sice; their experiments prove, however, that other yeasts, such as Sac- 
charomyces elUpsoideus and Saccharomyces Pastorianus, strongly ferment 
these dextrins. In carrying out the fermentation method for the estima- 
tion of dextrin, it is best to work with a pure culture of Saccharomyces 
cerevisice. 

Limitations of Fermentation Methods. The methods of estimat- 
ing sugars by difference in polarization, before and after fermentation, 
give at best only a fair approximation. Several dangers attend the 
employment of the method, chief among which are the attack of sugars, 
or carbohydrates, supposed to be unfermented, and the incomplete de- 
struction of sugars supposed to be completely fermented. Careful 
attention to the details of pure culture, sterilization and nutrition will, 
however, largely eliminate these dangers. The formation of optically 
active fermentation by-products may introduce a disturbing factor 
under certain irregular conditions, but with a normal alcoholic fermen- 
tation the error from this cause is insignificant. The optical activity 
of the nutritive solution used in the experiments should of course be 
determined, and its value, if significant, should be considered in the 
calculation. 

The length of time required for completing a determination has 
been a strong objection against the use of fermentation methods in 
general sugar analysis. The more rapid, and generally more accurate, 
methods based upon polarizing and copper-reducing power have, for 
this reason, been given the preference. 

POLARISCOPIC METHODS BASED ON DESTROYING THE OPTICAL 
ACTIVITY OF REDUCING SUGARS 

The determination of sugars by methods of this class is based upon 
the fact that solutions of reducing sugars, when heated with alkalies or 
alkalies and hydrogen peroxide, or with alkalies and metallic oxides or 
salts, lose more or less completely their optical activity. These methods 
have been applied not so much to the determination of reducing 
sugars themselves, as to the determination of sucrose, dextrin and 
other non-reducing carbohydrates in presence of reducing sugars. 

DESTRUCTION OF OPTICAL ACTIVITY OF REDUCING SUGARS BY MEANS 

OF ALKALIES 

Method of Dubrunfaut. The first efforts to establish a quan- 
titative method in this direction were made by Dubrunfaut f in 1850. 
* J. Chem. Soc. Trans., 47, 527. f Compt. rend., 32, 439. 



SPECIAL METHODS OF SACCHARIMETRY 



303 



Later investigators found, however, that the end-products in Dubrun- 
faut's method, obtained by the action of different alkalies upon reducing 
sugars, were not completely inactive, so that the polariscopic reading 
always required a certain correction. Efforts to establish a constant 
correction factor for modifications of Dubrunfaut's method have been 
made by Pellet,* Jesser,| Koydl,} Bardach and Silberstein and others, 
but the results, on account of the variability in conditions, have not been 
wholly satisfactory. 

Method of Lobry de Bruyn and van Ekenstein. The rate of de- 
struction of optical activity upon heating solutions of reducing sugars 
with dilute alkalies is illustrated by the following experiment taken 
from the work of Lobry de Bruyn and van Ekenstein; II 20 gms. of 
anhydrous glucose were heated with 10 c.c. of normal potassium hy- 
droxide in 500 c.c. of solution at 63 C. The following decrease in 
rotation was noted: 



Time. 


Angular rotation. 


Specific rotation. 


Time. 


Angular rotation. 


Specific rotation. 


Minutes. 
10 


+5 30' 


[]/> = + 48 


Minutes. 
50 


1 50' 




20 


4 20' 




85 


43' 




30 


3 10' 




135 


10' 


[ a ] n = -t 1 


40 


2 20' 











At the end of the experiment the solution had not darkened per- 
ceptibly and the original reducing power had only slightly diminished. 

Explanation of Optical Inactivity Produced by Alkalies. The ex- 
planation of the change of an optically active into an optically inactive 
solution of reducing sugar by action of alkalies was first given by Lo- 
bry de Bruyn and van Ekenstein. In the experiment just quoted the op- 
tical inactivity of the solution is due not to a destruction of glucose, but 
to its partial conversion into mannose and fructose, the combined rota- 
tions of the mixture of sugars producing optical neutrality. In one ex- 
periment the authorities, just named, noted after heating with alkali a 
loss of 18 per cent in reducing power; the residue was estimated to con- 
sist of 49 per cent unchanged glucose, 5 per cent mannose and 28 per 
cent fructose; the calculated rotation of such a mixture would in fact 
be very nearly zero. 

* Bull, assoc. chem. sucr. dist., 8, 623. 

t Oest. Ung. Z. Zuckerind., 27, 35. 

t Ibid., 29, 381. 

Z. Unters. Nahr. Genussm., 21, 540. 

|| Rec. Trav. Pays-Bas, 14, 156, 203; 16, 262. 



304 



SUGAR ANALYSIS 



Method of Jolles. Recent experiments by Jolles * upon arabinose, 
glucose, fructose, invert sugar, lactose and maltose show that these 
sugars in 1 to 2 per cent solution are rendered optically inactive by heat- 
ing for 24 hours at 37 C. with T o normal sodium hydroxide while 
sucrose is completely unchanged by this treatment. Stronger solutions 
of reducing sugars than 2 per cent show usually a residual activity after 
the alkaline treatment; it is necessary, therefore, in Jolles's method to 
dilute solutions to 2 per cent reducing sugar before making the deter- 
mination. With substances containing much reducing sugar such dilu- 
tion necessarily involves a considerable multiplication of any errors in 
the polariscope reading. 

Method of Bardach and Silberstein. Bardach and Silbersteinj 
have modified Jolles's method so as to include solutions of reducing 
sugar up to 5 per cent concentration. Their method of procedure is as 
follows : 

Take 45 c.c. of the neutralized sugar solution and make up to 50 c.c. 
with normal sodium hydroxide, thus making the solution T V normal 
alkaline. The solution is then polarized and a measured volume placed 
in a small beaker (8 to 10 cm. high and 5 cm. diameter) and kept 
at 36 to 39 C. for 20 hours by means of a thermostat, the beaker 
remaining uncovered. The solution is then cooled, made up to the 
original volume and repolarized. The final polarization is corrected for 
residual activity by means of an empirical factor, which in case of 
glucose was found to be as follows: 

TABLE LVIII 
Showing Change in Polarization of Glucose upon Warming ivith Dilute Alkali 



Approximate 


Polarization value. 


Approximate 


Polarization value. 


glucose in 
solution. 


Before 
treatment. 


After treat- 
ment. 


glucose in 
solution. 


Before treat- 
ment. 


After treat- 
ment. 


0.5 


+0.51 


-0.09 


2.5 


+2.54 


-0.36 


1 


+ 1.02 


-0.19 


3 


+3.05 


-0.26 


1 


+ 1.02 


-0.15 


3 


+3.06 


-0.27 


1.5 


+ 1.53 


-0.26 


4 


+4.10 


-0.32 


2 


+2.04 


-0.25 


4 


+4.07 


-0.25 


2 


+2.05 


-0.26 


5 


+5.12 


-0.21 



The loss in polarization, after treatment with alkali under the pre- 
scribed conditions, must be diminished, therefore, by about 0.25 to 
give the correct polarization value of glucose. So also the residual 
* Z. Unters. Nahr. Genussm., 20, 631. t Loc.cit. 



SPECIAL METHODS OF SACCHARIMETRY 305 

polarization must be increased by 0.25 to give the correct polarization 
equivalent of the residual sucrose, or other non-reducing carbohydrate 
present. 

It is evident that the chemist in employing such methods as the 
above must establish his own correction factor for the particular re- 
ducing sugar with which he is working. The lack of absolute uni- 
formity of conditions in the analysis of impure sugar products, leaves 
the general reliability of such correction factors more or less in doubt. 

DESTRUCTION OF OPTICAL ACTIVITY OF REDUCING SUGARS BY MEANS OF 
ALKALI AND HYDROGEN PEROXIDE 

Other chemicals have been used in connection with alkalies to pro- 
mote the destruction of reducing sugars. Lemeland,* for example, has 
devised a method for destroying the optical activity of reducing sugars 
in presence of sucrose by means of alkali, manganese dioxide and hy- 
drogen peroxide. 

Method of Pellet and Lemeland. Pellet and Lemeland f have 
recently proposed a method for the analysis of sugar-cane molasses, 
which is based upon destroying the optical activity of reducing sugars 
by means of alkali and hydrogen peroxide. The details of the method 
are as follows: 

" Make a solution of the cane molasses that will contain at most 
5 per cent of reducing sugars. Measure 50 c.c. of this solution into a 
300-c.c. flask, add 7.5 c.c. of sodium hydroxide (36 Be.), then 75 c.c. of 
hydrogen peroxide (12 vols.), and 60 c.c. of water. Mix and place the 
flask in a boiling water-bath for 20 minutes, cool, neutralize the re- 
maining alkalinity fairly exactly with acetic acid, and defecate with 
basic lead-acetate solution (36 Be.), the necessary amount of which 
will be found to vary from 15 to 40 c.c., according to the weight of the 
material taken, the amount of reducing sugars destroyed and the im- 
purities initially contained in the liquid. Complete the volume to 
300-c.c., mix well and filter. First polarize directly in the 200-mm. or 
400-mm. tube. Then 50 c.c. of the filtered liquid may be taken, 1 c.c. 
of glacial acetic acid added to it, the volume completed to 55 c.c., and 
after mixing a second polarization made, account being taken of the 
dilution. This is done because the second polarization is often a little 
different from the first, in which the liquid is alkaline. If a difference 
is observed, then the second, or acid polarization, should be used. The 
percentage of sucrose is calculated on the solution, and then on the 

sample." 

* J. Pharm. Chim., 2, 298. 
t Int. Sugar J., 13, 616. 



306 SUGAR ANALYSIS 

The authors state that the results by this method agree very closely 
with those obtained by the method of inversion, when special pre- 
cautions are observed to insure the utmost accuracy. 

DESTRUCTION OF OPTICAL ACTIVITY OF REDUCING SUGARS BY MEANS OF 
ALKALI AND MERCURIC CYANIDE 

Method of Wiley. The destruction of the optical activity of re- 
ducing sugars by means of Knapp's alkali-mercuric-cyanide solution 
was first employed by Wiley* in the determination of dextrin in com- 
mercial glucose. The reagent is prepared as follows: 

Alkali-mercuric-cyanide Solution. Dissolve 120 gms. sodium hy- 
droxide and 120 gms. mercuric cyanide in separate portions of water; 
the two solutions are then mixed and made up to 1000 c.c. Any 
precipitate which forms is removed by filtration. 

In making the determination 10 gms. of the commercial glucose are 
dissolved in water and made up to 100 c.c.; 10 c.c. of this solution are 
transferred to a 50-c.c. graduated flask, 20 to 25 c.c. of the alkali- 
mercuric-cyanide solution are added, and the mixture boiled 3 minutes 
under a well-ventilated hood. The solution is cooled, and neutralized 
with concentrated hydrochloric acid, the latter being added until the 
brown color of the liquid is just discharged. The solution is then clari- 
fied, made up to volume, filtered and polarized. The optical activity 
of the maltose and dextrose being destroyed, the residual polarization 
is that of the dextrin. 

In Wiley's experiments, the specific rotation of the dextrin was 
taken as + 193. Adopting this figure, and taking the reading of a 
Ventzke-scale saccharimeter, the grams of dextrin in 100 c.c. of solu- 
tion = 66 - 5 *.26 y0 = Q Q896 yo Since the golution p i ar i ze d con- 

17O 

tained 1 gm. of original sample in 50 c.c. (or 2 gms. in 100 c.c.), then 

0896 V 
2 X 100 = per cent dextrin in the commercial glucose. 

In concluding this chapter upon special methods of saccharimetry 
the chemist is advised, as in case of the methods of inversion, to test the 
reliability of any untried process by means of check analyses upon mix- 
tures of known sugars. It is only in this way that an idea can be 
formed of the errors which are due to defect of method or to personal 
equation. 

* Wiley's " Agricultural Analysis" (1897), 3, 290. 



CHAPTER XII 



MISCELLANEOUS PHYSICAL METHODS AS APPLIED TO THE EXAMINA- 
TION OF SUGARS 

IN addition to specific gravity, refractive index and specific rota- 
tion there are a number of other physical constants, which, though of 
lesser analytical importance, have nevertheless a considerable value in 
certain investigations of sugars and sugar solutions. Among the con- 
stants of this class may be mentioned viscosity, heat of combustion, 
osmotic pressure, rate of diffusion, surface tension, heat of solution, 
thermal conductivity, specific heat and magnetic rotation. It is be- 
yond the scope of the present volume to discuss the methods of making 
each one of these physical measurements. Viscosity, heat of combus- 
tion and the constants connected with osmotic pressure have acquired, 
however, a certain importance in general laboratory practice and the 
present chapter will discuss their use in the investigation of sugars. 

VISCOSITY OF SUGAR SOLUTIONS 

The determination of viscosity is a measurement which is frequently 
applied to solutions of sugars and other carbohydrates for 
special purposes of technology, analysis or research. The 
viscosity of a liquid as ordinarily determined is an arbitrary 
constant and is usually taken as the ratio between times of flow, 
through a narrow tubular opening, of the same volumes of 
water and liquid, all conditions of temperature, etc., being the 
same. 

Viscosity Pipette. The simplest example of this method 
of measurement is afforded by the viscosity pipette. (Fig. 149.) 

The pipette is first filled with water so that its meniscus 
coincides with the upper mark A; after holding in a perfectly 
upright position the water is released and the interval of time 
noted for the passage of the meniscus from A to the lower 
mark B. The process is repeated a number of times and the Fl s- 149 -~ 
average result taken as the water constant of the pipette at . 
the temperature of the experiment. The pipette is dried 
and the process repeated in exactly the same manner with a sugar 
solution. If the average time of flow at 20 C. for water be 20.2 

307 



308 



SUGAR ANALYSIS 



seconds and that of a sugar solution at 20 C. 105.1 seconds, then 
105.1 



20.2 



= 5.2, the relative viscosity of the sugar solution at 20 C, as 



compared with water of the same temperature. 

Engler's Viscosimeter. The apparatus of Engler* (Fig. 150) is 
used very generally for determining viscosity. The instrument con- 
sists of a bath B, which is filled with water or oil of the desired tempera- 
ture. The container A is gold plated, the conical bottom terminating 




Fig. 150. Engler's viscosimeter. 

in a narrow tube a, 3 mm. wide and 20 mm. long, which serves as the 
outlet; the latter is closed by the valve rod b. The container holds at 
the marks c exactly 240 c.c. of solution. After filling to c with water 
or solution, the cover A', holding a thermometer t, is placed in position 
and the temperature brought to the desired point. The valve rod is 
then withdrawn and the time noted for the delivery of exactly 200 c.c. 
of liquid in the flask C. The calculation of viscosity is made as pre- 
viously described. 

* Konig's "Untersuchung" (1898), p. 432. 



MISCELLANEOUS PHYSICAL METHODS 



309 



Coefficient of Viscosity. While the viscosity, as calculated by 
the above method, is sufficiently exact for many purposes, it is necessary 
in comparing liquids of different densities to employ the more exactly 
defined coefficient of viscosity. 

In Fig. 151 the volume V-oi liquid which is discharged in a time t 
through a given capillary tube A-B of the length I and radius r under a 
pressure p is found by the equation 

T7 TT X p X r 4 X t 



SpXl 

in which p is the coefficient 
interior friction of the liquid, 
lows from the foregoing that 

_irpr A t 



(1) 

of the 
It fol- 



(2) 




JL 



Fig. 151. Showing principle of 

viscosimeter. 

When Vj r and I are unchanged, as 

happens in the use of the same viscosity apparatus, p under constant 
pressure p becomes 

P = Kt, (3) 

in which K is a single constant peculiar to each individual viscosim- 
eter. 

In the previous figure the pressure p, with which a given volume of 
liquid M-M' is discharged at the beginning of flow, is equal to its density 
d multiplied by the height h of its surface above the outlet B, and at the 
end of the flow to its density 8 multiplied by the height h'. In the dis- 
charge of a constant volume V of different liquids, between the marks 
M and M' ', h and h' are unchanged, so that for the mean pressure of 
flow, p = C X <5, in which C is a constant. The coefficient of interior 
friction for different liquids using the same viscosimeter is then repre- 
sented by the formula 

P =KXCX8Xt, 

in which K and C are two constants. 

For water (8 = 1), p = K XC Xt. For any liquid of density 8 and 
time of flow T, the viscosity coefficient YJ, or ratio between the internal 
friction of water and liquid, is 

KXCX8Xr _8r 
KXCXt "V 

The viscosity coefficients of liquids are, therefore, always proportional 
to the products of their densities and times of flow. 



310 



SUGAR ANALYSIS 



Viscosity Coefficients of Pure Sucrose Solutions. The viscosity 
coefficients of pure sucrose solutions, as determined by Orth* for differ- 
ent concentrations and temperatures, are given in Table LIX. 

TABLE LIX 
Viscosity Coefficients of Pure Sucrose Solutions 





Temperatures. 


Grams sucrose 


















in 100 grams 


20 C. 


30 C. 


40 C. 


50 C. 


60 C. 


70 C. 


80 C. 


90 C. 


solution. 


















60 


6.29 


4.33 


3.22 


2.54 


2.10 


1.81 


1.61 


1.46 


62 


8.57 


5.54 


3.92 


2.98 


2.39 


2.00 


1.74 


1.55 


64 


12.31 


7.41 


4.94 


3.58 


2.76 


2.25 


1.91 


1.67 


66 


18.80 


10.14 


6.47 


4.43 


3.28 


2.58 


2.13 


1.83 


68 


30.82 


15.40 


8.86 


5.70 


4.01 


3.02 


2.42 


2.02 


70 


54.91 


24.42 


12.79 


7.64 


5.06 


3.65 


2.81 


2.28 


72 


107.85 


41.84 


19.65 


10.76 


6.65 


4.53 


3.34 


2.62 


74 


237.49 


78.50 


32.47 


16.05 


9.15 


5.85 


4.09 


3.08 


76 


596.76 


163.74 


64.16 


25.63 


13.30 


7.88 


5.19 


3.72 



It is seen that at low temperatures the viscosity is much higher and 
that at certain concentrations it begins to undergo a most marked 
change in value. This relationship is made more plain in the opposite 
diagram (Fig. 152) which is taken from the work of Orth. 

Attempts have been made to express the relationship between the 
viscosity and concentration of sugar solutions by means of a general 
equation. For dilute solutions the relationship according to Arrhenius f 
may be expressed by the equation 



Or logeYJ = loge A(x), 

in which A is a constant and x the concentration. According to this 
equation the natural logarithm of the viscosity coefficient is propor- 
tional to the concentration. 

But for concentrated sugar solutions the above relationship does 
not hold. The law for solutions of high sucrose content, according to 
Orth, is expressed by the equation: 



r loge (log e Y)) = log e (fog. A) + 

in which A and B are constants. 



* Bull, assoc. chim. sucr. dist., 29, 137. 
t Z. physik. Chem., 1, 285. 



MISCELLANEOUS PHYSICAL METHODS 
For changes in temperature Orth gives the equation 



311 



or log, (loge ij) = log e (log. A ) + log* B (x) + log e C (t) ; 

in which x and t are the concentration and temperature of the sugar 
solution, and A, B and C constants. 



300 



250 



200 



150 




Temperature 

Fig. 152. Diagram showing viscosity curves of four sugar solutions 
at different temperatures. 



Viscosity Coefficients of Impure Sucrose Solutions. From the 
viscosity coefficients of solutions of different sugar-house products Orth 
has made a compilation, the results of which are shown in Table LX. 



312 SUGAR ANALYSIS 

TABLE LX 
Viscosity Coefficients of Sucrose Solutions of Different Purities. 



Tempera- 
ture. 


Purity (per 
-cent sucrose in 
solids). 


Grams of solids in 100 grams of solution. 


65 


70 


75 


80 


85 


20 

40 
60 
80 


100 
90 

80 
70 
60 

100 
90 
80 
70 
60 

100 
90 
80 
70 
60 

100 
90 
80 
70 
60 


15.09 

15.18 
15.31 
15.41 
15.51 

. 5.62 
5.49 
5.35 
5.23 
5.10 

3.00 
2.90 
2.81 
2.72 
2.64 

2.01 
1.95 
1.89 
1.83 

1.78 


54.91 
52.91 
50.99 
49.16 
47.41 

12.79 
12.25 
11.74 
11.24 
10.78 

5.06 
4.86 
4.67 
4.49 
4.33 

2.81 
2.71 
2.63 
2.54 
2.47 


369.67 
324.0 
283.5 
249.9 
221.1 

43.03 
39.91 
36.96 
34.38 
32.03 

10.95 
10.50 
10.05 
9.65 
9.27 

4.59 
4.48 
4.37 
4.27 
4.17 






4450 
3251 
2400 
1808 


196,600 
102,960 
55,360 
80,770 


225.9 
199.4 
175.7 
155.3 


2,892 
2,334 
1,884 
1,538 


33.03 
31.97 
30.87 
29.90 


184.0 
183.2 
183.2 
183.2 


9.55 
9.65 
9.76 
9.84 


30.41 
33.24 
36.62 
40.33 



The relation between viscosity and concentration of impure sugar- 
factory solutions is represented according to Orth by the equation 



in which t is the temperature and K a linear function of t, x the percent- 
age of sucrose and n the percentage of non-sugar, and A, B and C 
constants. 

The viscosity of the non-sugars of sugar-house products was cal- 
culated by Orth not to differ greatly from that of pure sucrose; it was 
somewhat greater for the cold dilute and hot concentrated solutions 
and a little less for the other solutions, the average value for solutions 
of the same concentration being about 96 per cent that of sucrose. 

The above conclusions of Orth pertain, however, only to the ordi- 
nary impurities of sugar-house products, such as reducing sugars, salts 
of mineral and organic acids, ammo compounds, etc. The observation 
does not hold for dextran, levan and other gums which may occur in 
abnormal products and which greatly increase the viscosity of sugar 
solutions with consequent disturbance in the work of evaporating and 
boiling. 



MISCELLANEOUS PHYSICAL METHODS 313 

Excessive viscosities may also occur in sugar-house practice from 
supersaturation of sucrose, the result of careless sugar boiling. The 
successful sugar boiler aims to prevent supersaturation and to keep the 
viscosity of the pan contents as low as possible, in order that the maxi- 
mum yield of sugar crystals may be obtained. 

The determination of viscosity is of great value in certain branches 
of analytical work, as, for example, the examination of commercial dex- 
trins, for which see page 508. 

SPECIFIC HEAT OF COMBUSTION 

Units Employed in Calorimetery. The number of calories or 
heat units which a substance gives off, when burned in oxygen under 
specified conditions, is a constant which has been extensively used in the 
investigation of sugars. The determination has been especially em- 
ployed in studying the calorific value of the different carbohydrates 
which are used in foods. 

The Small, or Gram, Calorie (cal.) is defined as the quantity of heat 
necessary to raise 1 gm. of water through 1 C. The quantity of heat 
necessary to raise 1 gm. of water from to 1 C. is not, however, ex- 
actly the same as that necessary to raise 1 gm. of water from 99 to 
100 C., so that the measurement has been defined more precisely as one 
one-hundredth of the heat required to raise 1 gm. of water from to 
100 C. 

The Large, or Kilogram, Calorie (Cal.) contains 1000 small calories, 
and may be defined, with the limitations previously noted, as the 
quantity of heat necessary to raise 1000 gms. of water through 1 C. 

The Centuple Calorie (K) is defined as the quantity of heat necessary 
to raise 1 gm. of water from to 100 C. 

For ordinary purposes the ratio of the several units may be ex- 
pressed as: 

1 Cal = 10 K = 1000 cal. 

THE BOMB CALORIMETER 

The determination of calories of combustion is made in an atmos- 
phere of compressed oxygen by means of a bomb calorimeter", the in- 
vention and extensive application of which to heat measurements are 
due to Berthelot.* The original bomb of Berthelot, on account of the 
large amount of platinum which it contains, is exceedingly expensive, 
and has been variously modified by Mahler, Hempel, Atwater and 
others for the purpose of reducing the cost. The Berthelot calorimeter, 
* "Trait6 pratique de Calorimetrie chimique ; " also Ann. chim. phys,, [6] 6, 546. 



314 



SUGAR ANALYSIS 



as modified by Hempel and Atwater* and improved by Blakeslee, is 

shown in Fig. 153. 

Description of Calorimeter. The most important feature of the 

calorimeter is the steel bomb, the cup (A) and cover (B) of which are 

lined with platinum, or heavily 
plated with gold. The cover is pro- 
vided with a sunken lead gasket K, 
which rests upon the rim of the cup, 
and is held in place by the steel 
collar C, which is screwed tightly 
into position by means of a clamp 
and heavy spanner. The cover of 
the bomb is provided with a neck 
having an opening leading from G 
to the interior of the bomb for the 
entrance of oxygen; the inlet is 
opened and closed by a valve screw 
F. The cover is also provided, on 
its inner surface, with two stiff 
platinum rods I and H, between 
which passes a small spiral of iron 
wire for igniting the charge; the 
latter, consisting of 1 to 2 gms. of 
the sugar or carbohydrate to be 
burned, is placed in a platinum 
capsule, with a small piece of 
naphthalene to act as a kindler, 




Fig. 153. Bomb calorimeter. 



directly under the spiral. The rod / is connected through the cover 
with the electric wire /' and the rod H, insulated from the cover, with 
the electric wire H' '. 

Operation of Calorimeter. The bomb, after introducing the 
charge, is filled with pure oxygen under 20 atmospheres pressure and 
then placed in the brittania-metal vessel M, which contains a weighed 
quantity of water, sufficient to cover all parts of the bomb. The vessel 
M rests within two buckets, N and 0, which, with their covers, form 
two dead-air spaces, and insulate the bomb system from the room at- 
mosphere. The temperature of the water in M should be 2 to 3 C. 
below that of the inner air-chamber. A Beckmann thermometer, P, 
passes through the covers of the pails, and is fastened so that its 

* See article by Atwater and Snell, J. Am. Chem. Soc., 25, 659, for a very com- 
plete description of this instrument and its use. 



MISCELLANEOUS PHYSICAL METHODS 315 

bulb is immersed in the water about opposite the middle of the bomb. 
The thermometer can be read, by means of a magnifying lens, to 
the thousandth of a degree; it should be provided with a certificate 
for correcting errors of construction and for converting readings to 
true centigrade degrees. The mercury thread of the thermometer is 
adjusted at the desired point by partly filling or emptying the upper 
reservoir. 

When the apparatus is in readiness the mechanical stirrer L is set 
in motion and the thermometer read at intervals of one minute, tapping 
the top gently with an electric hammer before each reading to prevent 
lagging of the mercury thread. When five successive readings show a 
uniform rise in temperature, the electric switch is closed exactly at the 
end of the fifth minute. As soon as the extinction of the lamp in a re- 
sistance circuit indicates the fusion of the iron wire, the switch is re- 
opened to avoid heating the water by the current. The readings of the 
thermometer should be noted at the end of each minute, until the 
maximum elevation of mercury is reached and the rate of fall has be- 
come regular. With the stirring mechanism making 40 revolutions per 
minute equilibrium is obtained usually within 5 minutes. After stirring 
5 more minutes a final reading is taken, when the calculation may be 
made. 

Hydrothermal Value. The calories of combustion are calculated 
from the observations of a calorimeter experiment by multiplying the 
hydrothermal value (in grams) of the calorimeter system by the cor- 
rected rise in temperature and dividing the product (after subtracting 
the heat units due to accessory combustions) by the weight in grams of 
substance taken. 

The accuracy of all calorimetric experiments is dependent upon the 
exactness with which the hydrothermal value of the calorimeter is 
known. The most common method for computing the water equivalent 
of the calorimeter system is to multiply the weight of each part by its 
specific heat and take the sum of these water equivalents as the hydro- 
thermal value of the entire system. An example of the method is 
given by Fries, in Table LXI. 

The hydrothermal value may also be determined by measuring the 
rise in temperature of the calorimeter system from burning a substance 
of known calorific value, as benzoic acid (1 gm. = 6322 cals.). For a 
description of this and other methods reference should be made to the 
work of Fries.* 

* Fries, " Methods and Standards in Bomb Galorimetry," Bull. 124, Bur. of 
Animal Ind., U. S. Dept. of Agr., p. 9. 



316 



SUGAR ANALYSIS 



TABLE LXI 
Computed Water Value of Bomb Calorimeter 



Material. 


Weight. 


Specific heats. 


Water equiva- 
lent. 


Steel 


Grams. 

3236.0 


0'.1114 


Grams. 

360 49 


Platinum 


196.0 


0.0320 


6 27 


Lead 


66 


0300 


1 98 


German silver (approximate) 


4 


0940 


38 


Rubber (approximate) 


4 


3310 


1 32 


Iron (approximate) 


10 


1114 


1 11 


Mercury (approximate) 


50 


0330 


1 65 


Glass (approximate) 


10 


1900 


1 90 


Britannia metal 


855 


0548 


46 85 


Oxygen (constant volume) 


11 4 


1570 


1 79 


Water at 22 C 


2000.0 


0.9975 


1995 00 










Total 






2418 74 











Correction for Radiation. When the conditions of the experi- 
ment are properly controlled the calorimeter system at the beginning of 
combustion is slightly cooler, and at the end of combustion slightly 
warmer, than the surrounding air. During the first period the calorim- 
eter gains heat, and in the second loses heat to the surrounding air; 
the thermometer readings must be corrected, therefore, for the errors 
of radiation. This correction is made by the Regnault-Pf aundler * 
formula 



) 



where n = number of time units (minutes) in combustion period. 
V = rate of fall of temperature of calorimeter during initial period. 
(The change is actually a rise but for convenience is expressed as a fall, 
the value of V thus being negative.) 

V = rate of fall of temperature of calorimeter during final period. 
6 = mean temperature of calorimeter during initial period. 

0' = mean temperature of calorimeter during final period. 

0i, Oz, . . . B n = temperature at end of first, second, . . . nth min- 
utes of combustion period. 

0o = temperature at moment of ignition. 

Illustration of Method. The application of "the formula is best 
understood from a special case and the following example of the com- 
bustion of sucrose is taken from a paper by Atwater and Snell.f The 
calorimeter employed had a water equivalent of 2100 gms. The data 



* Pfaundler. Pogg. Ann., 129, 113. 



t J. Am. Chem. Soc., 25, 659. 



MISCELLANEOUS PHYSICAL METHODS 



317 



of the experiment are given in the following record, which is a convenient 
form for determinations of this kind. 



Sample No. Description Cane Sugar. Date, July 13, 1901. 
Bomb No. 3 Observer, J. F. Snell. Thermometer, No. 733. 


Capsule No. 1 
Wt. caps. H- subs. 
Wt. capsule 


= 4.2501 

= 2.8783 




Correction foi 
Wt. Fe 13.0 
Wt. naphth 
HN0 3 
Correction foi 


Accessory Combustions. 
-1.1 = 11.9 mgs. = 19.0 cal. 
alene= 6.4 mgs. =61.6 cal. 


Wt. substance, W= 1.3718 


accessories =87.2 cal. 


Final period. Main period. Initial period. 


' 
' 


Readings. 

1 1.018 
2 1.021 
3 1.025 
4 1.027 
5 1.030 
60o 1.032 


Corrected 
readings. 

1.015 
1.029 


Initial period. 

Fall =- .014 
Rate V =- .0028 
Meanf, = 1.022 

Corrected reading. 

5 = 3.646 
= 1.029 


Thermometer correction. 

7 70 air = 25.2 
T water =23.8 
1st reading = 1.0 


T of zero =22.8 
Corr. for 1 =+ .001 
Rise (degrees) = 2.6 


Thpr r*nrr -1- 002fi 


70! 2.300 
80 2 3.650 
90 3 3.678 
1004 3.662 
110 5 3.653 


2.3 
3.7 
3.7 
3.7 




Final calculations. 

05 = 3.646 
0o = 1.029 


05 + 00 
\ 

Fi 

Fall 
Rate T 
I 
V V 


= 4.675 
= 2.3 

nal period. 

= + .013 
rf = + .0026 
r = - .0028 


13.4 
= 2.3 

= 15.7 
= 5.1 


05 + 00 


05-0o = 2.617 
Th. corr. =+ .0026 
Rad. corr. =+ .0079 


2 
Sum 
58 


Corr. rise = 2.6275 
Corr. rise ) 
X2100 [= 5517.8 
= total heat ) 
Accessories = 87.2 


Diff. =10.6 
Log. diff. = 0253 
Log. V'-V = 7324 
Colog.0 / -0 = 5820 


= + .0054 

,0' = 3.640 
= 1.022 


Antilog. 
+5F 
Radiation ) 
correction) 
16 3.640 
Time 3.. 30 


3397 
= + .0219 
= -.014 


Mean 2 
0'-0 


Corrected heat = 5430.6 
Log. corr. heat = 73485 
Log. W =13729 


= + .0079 
3.633 


= 2.618 


59756 
Heat of com- ) 
bustion per > = 3959 
gram ) 



Applying the formula to the above example, where the number of time 
units, n, is 5, we obtain for the several expressions, F= .0028 and 



The combination of these values in the formula gives a radiation correction of 
C=+ 0.0079. 

The corrected rise of the Beckmann scale was 2.617 degrees and this cor- 
rected to true degrees C. and for radiation gives 2.6275 C. as the corrected 
rise in temperature, which, multiplied by 2100, the water equivalent of the 
calorimeter, gives 5517.8 calories. 



318 SUGAR ANALYSIS 

Correction for Accessory Combustions. The weight of the iron wire was 
13 mgs. The quantity unburned was 1 .1 nig. The quantity burned was there- 
fore 11.9 mgs. The specific heat of combustion of iron being 1601 calories, the 
heat of combustion of 11.9 mgs. is 11.9 X 1.6 = 19 calories. The quantity of 
naphthalene burned was 6.4 mgs., which yields 6.4 X 9.63 = 61.6 calories, the 
specific heat of combustion of naphthalene being 9628 calories. The heat of 
combustion of nitrogen in the bomb as determined by titration of the nitric 
acid is 6.6 calories. (N 2 + 5 + H 2 = 2 HN0 3 . .004406 gm. HNO S = 1 
cal.) The total heat from accessory combustions is, therefore, 19 -f- 61.6 + 6.6 = 
87.2 calories. 

Deducting this quantity from the total heat set free in the apparatus, we 
have 5517.8 87.2 = 5403.6 calories as the heat due to the combustion of the 
sugar. The quantity of sugar burned was 1.3718 gms. The specific heat of 
combustion according to this determination is, therefore, 5430.6 *- 1.3718 = 
3959 calories. 

Gram-molecular Heat of Combustion. The gram-molecular heat 
of combustion is found by multiplying the calories per gram by the 
molecular weight (M ) . To avoid large figures it is customary to express 
this unit in terms of large calories. 

cals. X M 



Gm. mol. Cals. = 



1000 



CALORIFIC CONSTANTS OF DIFFERENT SUGARS 

In Table LXII, compiled by Tollens,* the calorific constants are 
given for the principal sugars, polysaccharides and sugar alcohols. 

It is seen from the table that the molecular heat of combustion is 
always higher for the anhydride than for the hydrate of the same sugar. 
The molecular heat of combustion of the higher saccharides is also 
greater than the sum of the values of their components. Thus : 
Sucrose = 1352.7 Gm. mol. Cals. " 

Glucose =673.7] 

an* n r = 1349.6 Gm. mol. Cals. 
Fructose = 675.9 J 

Difference 3.1 Gm. mol. Cals. 

This difference may be taken as the equivalent of heat which is liber- 
ated during inversion. 

In the same way Rafnnose = 2026.1 Gm. mol. Cals. 
Glucose = 673.71 

Fructose = 675.9 V = 2019.5 Gm. mol. Cals. 
Galactose = 669.9 J 



Difference = 6.6 Gm. mol. Cals. 
* Tollens's "Handbuch der Kohlenhydrate, " II, p. 45. 






MISCELLANEOUS PHYSICAL METHODS 



319 



TABLE LXII. 

Giving Heats of Combustion of Sugars, Poly saccharifies and Sugar Alcohols. 



cal. 1 gram. 



Cal. (1 Cal. = 1000 cal.) 
for 1 gram-molecule. 



Sugars 
Arabinose, C 5 H 10 5 

Xylose, C 5 H 10 5 | 3740 (IV 

Rhamnose, C 6 H 12 O5 4379.3 (St. 

Rhamnose (cryst.), C 6 H 12 O 5 +H 2 O .... 3909.2 (St. 

Fucose, C 6 H 12 O 5 4340.9 (St. 

Glucose, C 6 H 12 O 6 3742.6 (St.) 

Galactose, C 6 H 12 O 6 3721 .5 (St.) 

Fructose, C 6 H 12 O 6 3755 (St.) 

Sorbose, C 6 H 12 O 6 3714.5 (St. 

Sucrose, Ci ? H 22 O u 3955.2 (St. 

Lactose, Ci 2 H 22 O n 3951 .5 (St. 

Lactose, Ci 2 H 22 O u +H 2 O 3736.8 (St.) 

Maltose, Ci 2 H 22 O u 3949.3 (St. 

Maltose, Ci 2 H 22 O u +H 2 O 3721 . 8 (St. 

Trehalose (anhydr.), Ci 2 H 22 O u 3947.0 (St. 

Trehalose (cryst.), Ci 2 H 22 On+2H 2 O . . 3550.3 (St. 

Raffinose (anhydr.), C 18 H 32 Oi 6 j ^'(B^' 

Raffinose (cryst.), Ci 8 H 32 Oi 6 +5 H 2 O. . . 3400.2 (St.) 

Melezitose, Ci 8 H 32 Oi 6 +H 2 O 3913.7 (St.) 

Poly saccharifies: 

Cellulose, (C 6 H 10 O 5 ) n 4185.4 (St.) 

Starch, (C 6 H 10 O 5 )n 4182.5 (St.) 

Dextran, (C 6 H 10 O 5 )n.. . 4112.3 (St.) 

Inulin, C 3 6H 62 O 3 i 4133.5 (St.) 

Glycogen, (C 6 H 10 O 6 )n 4190.6 (St.) 

Sugar Alcohols: 

Erythrite, C 4 H 10 O 4 4132.3 (St.) 

Arabite, C 6 H 12 O 5 4024.6 (St.) 

Mannite, C 6 H H O6 3997.8 (St.) 

Dulcite, C 6 H 14 O 6 3975.9 (St.) 

Perseite, C 7 H 16 O 7 3942.5 (St.) 

Quercite, C 6 H 12 O 5 4293.6 (St.) 

Inosite, C 6 H 12 O C . . 3679.6 (St.) 



558.3 (St.) 

557.1 (B.) 
561.9 (St.) 

560.7 (B.) 

718.5 (St.) 

711.8 (St.) 

712.2 (St.) 
673.7 (St.) 
677.2 (B.) 

669.9 (St.) 
675.9 (St.) 

668.6 (St.) 
1352.7 (St.) 

1345.2 (St.) 

1340.6 (Gibson) 

1350.7 (St.) 

1339.8 (St.) 

1349.9 (St.) 

1345.3 (St.) 
2026.5 (St.) 
2026.1 (B.) 
2019.7 (St.) 
2043.0 (St.) 



678.0 (St.) 

673.1 (Gottlieb) 

680.4 (B.) 

677.5 (St.) 

675.6 (Gibson) 

666.2 (St.) 
4092.1 (St.) 

678.9 (St.) 



504.1 (St.) 

502 (Louguinine) 

502.6 (B.) 

612.0 (St.) 
729.9 (St.) 
720.5 (Gibson) 
723.9 (St.) 

836.1 (St.) 
704. 4 (St.) 

710.4 (B.) 
662.3 (St.) 

665.5 (St.) 



St. = Stohmann and Langbein, J. prakt. chem. [2], 45, 305. 

B. = Berthelot and coworkers, from results in the Ann. chim. phys. [6], 6, 552; 10, 455; 13, 304, 
341; 21, 409. 



320 SUGAR ANALYSIS 

The hydrolysis of sugars may be regarded, therefore, as an exother- 
mic reaction. 

Calculation of Calories from Chemical Formulae. Various 
methods have been proposed for calculating the molecular heat of com- 
bustion from the chemical formula of sugars. 

The calorific value for the combustion of the elements carbon (dia- 
mond) and hydrogen have been determined as follows: 
C + O 2 = C0 2 + 94.3 Cals. 
H 2 + O = H 2 + 68.3 Cals. 

Welter's * rule for computing the molecular heat of combustion is to 
subtract as much O and H 2 as will unite to form water from the 
molecular formula, and multiply the number of remaining atoms by 
their respective heat values. The sum of the products is taken as the 
molecular heat of combustion. 

Example. Glucose C 6 Hi 2 6 . The 6 atoms of unite with 12 atoms of H 
to form 6 H 2 0. The Cals. of the 6 remaining C atoms = 6 X 94.3 = 565.8 Cals. 
This value is 16 per cent less than the value found experimentally by Stoh- 
mann, viz. 673.7 Cals. 

A second method of calculating heat of combustion is to combine all 
the and C that will unite to form C0 2 , and calculate the heat of the 
remaining atoms in the manner just described. 

To take again the example of glucose : The 6 atoms of unite with 3 atoms 
of C to form 3 C0 2 . The remaining C 3 and Hi 2 then give 
For C, 3 X 94.3 = 282.9 Cals. 
For H 2 , 6 X 68.3 = 409.8 Cals. 



692.7 Cals. 

The results by this method are much closer than those obtained by 
Welter's rule, being about 3 per cent higher than the value found ex- 
perimentally by Stohmann. 

A third method of calculating heat of combustion is to distribute 
the O of the molecule among its C and H atoms according to the pro- 
portionate number and combining powers of the latter. Since the 
necessary to form C0 2 is represented by 2 C and the O to form H 2 by 

TT 

-g > the uncombined equivalents of C and H, after deducting CO 2 and 

TT 

H 2 O, would equal 2 C + -^ - O. The ratio of total to uncombined 
* Walker's "Introduction to Physical Chemistry," (3rd Ed.), p. 129. 



MISCELLANEOUS PHYSICAL METHODS 321 

H 



\ 
calculation is then: 



equivalents is then ( 2 C + -= - o) -r- ( 2 C + 5\ - The formula for the 
\ . / \ 2> 

is then: 
Gm. mol. Cals. = ^94.3 C + 68.3 ^ 



-O 



Applying this formula to glucose, we obtain, 

Gm. mol. Cals. = (94.3 X 6 + 68.3 X ) = 650.4, 

v ' 12 +l | 

a result a little over 3 per cent below the value found experimentally by 
Stohmann. 

The true molecular heat of combustion is about midway between the 
values calculated by the last two methods. It is evident, however, 
that absolute agreement cannot be attained by any method- of calcu- 
lation, since the experimental results are different for different isomers. 
The gram-molecule Calories for the C 6 Hi 2 6 sugars were found by 
Stohmann to vary from 668.6 for sorbose to 675.9 for fructose. 



OSMOTIC PRESSURE AND RELATED PHYSICAL CONSTANTS, AND THEIR 
APPLICATION IN DETERMINING MOLECULAR WEIGHTS OF SUGARS 

The determination of the molecular weights of sugars and sugar 
derivatives is a problem which may confront the chemist in his examina- 
tion of unknown carbohydrates of plant or animal origin. 

In the case of a reducing sugar an elementary analysis of one of its 
osazones or hydrazones (p. 370) will serve to fix the class to which the 
sugar belongs and thus indicate the molecular weight. In the case, 
however, of non-reducing sugars, such as sucrose, raffinose, etc., and 
of the sugar derivatives, which do not form osazones and hydrazones, a 
determination of the molecular weight by some physical method is 
usually required. 

The molecular weights of sugar derivatives, which can be distilled 
without decomposition or dissociation, are best determined by the well- 
known vapor-density method of Victor Meyer. All the sugars, how- 
ever, and most of their compounds undergo decomposition at or below 
the melting point so that the vapor-density method is excluded. Re- 
course is, therefore, usually made to some one of the methods which in- 
volve the principle of osmotic pressure. 



322 



SUGAR ANALYSIS 



OSMOTIC PRESSURE OF SUGAR SOLUTIONS 

Pfeffer,* the plant physiologist, in 1877, during his classical studies 
upon osmosis in vegetable cells, discovered that the osmotic pressure of 
dilute sugar solutions was proportional to the concentration. Pfeffer's 
experiments were performed by placing the sugar solutions in a porous 
bulb, which had deposited within its walls a semipermeable membrane 
of copper ferrocyanide. The bulb, which was connected with an up- 
right tube, was then immersed in distilled water. The membrane, 
which is permeable to water but not to sugar, allows water to enter the 
bulb; the sugar solution begins to rise in the tube and the elevation 
continues until, after many hours, a maximum is reached; at this point 
the difference between the level of liquids within and without the bulb 
gives a pressure corresponding to the osmotic pressure of the sugar solu- 
tion. This maximum pressure, expressed in centimeters or millimeters 
of mercury, was called by Pfeffer the osmotic pressure. 

The following results by Pfeffer give the osmotic pressure of sucrose 
solutions at different concentrations. 



Concentration 


Pressure (P) in 




(C) of sucrose 


centimeters of 


Ratio J- 


solution. 


mercury. 


C 


Per cent. 

1 


53.5 


53.5 


2 


101.6 


50.8 


4 


208.2 


52.1 


6 


307.5 


51.3 



p 

The ratio -^ is a constant, the slight differences noted being due to 

variations in temperature and other experimental errors. 

Pfeffer also showed that the osmotic pressure of sugar solutions un- 
derwent a regular increase with elevation of temperature. The follow- 
ing experiment was made upon a 1 per cent sucrose solution. 



Temperature 
C. 


Absolute tempera- 
ture (T). 


Osmotic pres- 
sure (P). 


Ratio ~. 


14.15 


287.15 


51.0 


.1776 


15.5 


288.5 


52.05 


.1804 


32.0 


305.0 


54.4 


.1784 


36.0 


309.0 


56.7 


.1835 



* Pfeffer's "Osmotische Untersuchungen," Leipzig, 1877. 



MISCELLANEOUS PHYSICAL METHODS 323 

p 
The ratio -^ is thus also found to be constant, the slight variations 

being due as before to experimental errors. 

Relation of Osmotic to Gas Pressure. In 1887 van't Hoff* 
showed that Pfeffer's osmotic pressures were identical in value with 
those obtained by gas pressure; in other words that the osmotic pres- 
sure per gram-molecule of substance is the same as the gas pressure per 
gram molecule at the same temperature and volume. This identity is 
expressed by the equation 

pv = RT, 

in which p is the pressure and v the volume, T the absolute temperature 
and R a constant. Van't Hoff showed that the constant R is the same 
for substances in dilute solution as well as in the gaseous state. 

The molecular weight of a substance is equal to the weight of its 
vapor in grams which would occupy the same volume, under equal 
temperature and pressure, as 2 grams of hydrogen (2 being the weight 
of the hydrogen molecule). This volume, called the gram-molecular 
volume, is 22,380 c.c. at C. (273 abs.) and 76.0 cm. of mercury pres- 
sure (1 atmosphere). 

Calling V the volume occupied by a gram-molecule of gas we obtain 
from the previous equation, 

*-* 

The pressure p, per square centimeter of mercury (sp. gr. = 13.59), is 
equal to 76 cm. X 13.59 = 1033 gms. We obtain, therefore, for the 
constant R, 

1033 X 22,380 

~^73~ * 4 ' 683 ' 

To prove the identity of this constant for the osmotic pressure of 
sucrose one of the experiments of Pfeffer may be selected. A 1 per 
cent solution of sucrose at C. (273 abs.) gave an osmotic pressure of 
49.3 cm. of mercury. The latter corresponds to a pressure per square 
centimeter of 49.3 X 13.59 = 670 gms. Since the molecular weight of 
sucrose is 342, the volume (V) of a 1 per cent solution containing a 
gram-molecule would be very closely 34,200 c.c. Substituting these 
volumes in the equation, we obtain, 



which value is in substantial agreement with that derived by the other 
method. 

* Ostwald's " Grundriss " (2nd Ed.), p. 131. 



324 SUGAR ANALYSIS 

Application of the Method. If we accept now the identity of the 
laws for gaseous and osmotic pressure, the molecular weight of a sugar 
can be determined from its osmotic pressure in a manner analogous to 
that followed by the vapor-density method. 

Example. In one of the experiments previously cited Pfeffer found at 
15.5 C. (288.5 abs.) for a 1 per cent sucrose solution an osmotic pressure of 
52.05 cm. mercury. 

If 1 gm. of sucrose occupies 100 c.c. at 52.05 cm. pressure and 15.5 C., then 
the number of grams which would occupy 22,380 c.c. at C. (273 abs.) and 
76 cm. pressure would be: 

1 gm. X 22,380 c.c. X 288.5 X 76 cm. = _ 
100 c.c. X 273 X 52.05 cm. 

345 the number of grams in the gram-molecular volume is the molecular 
weight of sucrose. This agrees closely with the actual value 342 calculated from 
the formula Ci2H 22 On. 

It follows from the previous discussion that the sugars of lowest 
molecular weight will show for equal concentration and temperature 
the highest osmotic pressure. 

Measurement of Osmotic Pressure by Plasmolysis. A second 
method of applying the principle just described is due to the Dutch 
botanist de Vries,* who discovered that the plasmolysis, or loosening of 
the protoplasmic lining of plant cells, offered a simple and reliable means 
of measuring osmotic pressure. Fig. 154 shows the miscroscopic ap- 
pearance of a plant cell in sugar solutions of different concentration. In 
such a cell the thin layer p of protoplasm (the protoplast) acts as a 
semipermeable membrane. So long as the osmotic pressure of the cell 
liquid I exceeds or equals that of the surrounding sugar solution s, the 
protoplast is not affected. When, however, the osmotic pressure of 
the sugar solution becomes greater than that of the cell liquid there is a 
diffusion of water outward through the protoplasmic membrane. The 
latter, in consequence of the loss of a part of the cell water, is loosened 
from the cell wall and contracts, as shown in the figure. 

The application of the method may be understood from the follow- 
ing: de Vries found that the hair roots of the frogbit (Hydrocharis 
Morsus-rance) showed no plasmolysis in a 7 per cent, but a very pro- 
nounced loosening of the protoplast in a 7.1 per cent, sucrose solution. 
For these particular root hairs under the conditions of the experiment, 
plasmolysis was produced by a solution containing 0.208 gm. mol. of 
sucrose to 1000 gms. of solution (71 gms. -f- 342, the molecular weight 
of sucrose). 

* Bot. Ztg., 46, 229, 393. 



MISCELLANEOUS PHYSICAL METHODS 



325 



Suppose that, using these same root hairs, a solution containing 3.7 
per cent of glucose just produced plasmolysis. Then 37 (the grams of 
glucose per 1000 gms. of solution) divided by 0.208 = 178, the molecular 
weight of glucose, which corresponds to the formula C 6 Hi2O 6 (molecular 
weight =180). 






Fig. 154. Illustrating plasmolysis. 

I. Condition of plant cell before plasmolysis; II. Beginning of plasmolysis; 
III. Advanced stage of plasmolysis. 

It was by this means that de Vries,* in 1888, established the mo- 
lecular weight of raffinose. The following formulae had been proposed 
for the constitution of this sugar. 

I. Ci2H 22 Oii + 3 H 2 = 396, molecular weight. 
II. Ci 8 H 3 20 16 + 5 H 2 = 594, molecular weight. 
III. C36H 64 O 32 + 10H 2 = 1188, molecular weight. 
De Vries found by his method of plasmolysis that, when standardized 
against a sucrose solution for the same plant cell, 595.7 parts of raffinose 
were equimolecular with 342 parts of sucrose. This figure agrees with 
the molecular weight of formula II; the correctness of de Vries's con- 
clusion was afterwards verified by chemical means. 

Owing to the variation in composition of cell liquids, it is evident 
that the particular plant cells chosen for this method of examination 
must always be standardized before using. 

FREEZING AND BOILING POINTS OF SUGAR SOLUTIONS 

On account of the difficulty of preparing a perfect semipermeable 
membrane and owing to the extreme liability of such membranes to 
rupture, the determination of molecular weights by direct measurement 
of osmotic pressure, although most sound in principle, is not generally 
followed. Use is accordingly made of the measurement of some re- 
lated constant, such as that of vapor pressure, depression of freezing 
* Compt. rend., 106, 751. 



326 



SUGAR ANALYSIS 



M 



M' 



point or elevation of boiling point. The freezing and boiling points of 
sugar solutions vary in fact according to their vapor pressure, the 
value of which, it can be shown, is directly proportional to the osmotic 
pressure. 

Isotonic Solutions. In Fig. 155 suppose the closed vessel V to be 
divided by a semipermeable membrane M-M' into two equal compart- 
ments, which open into one another above M. Suppose, next, equal 

volumes of sucrose and glucose solutions 
of the same concentration to be placed 
in each of the compartments. Then 
water will diffuse from the sucrose solu- 
tion Sj where the osmotic pressure is 
lower, into the glucose solution G, where 
the osmotic pressure is higher, until at 
the point of equilibrium the osmotic 
pressures upon both sides of the mem- 
brane are equal. The two sugar solu- 
tions are then said to be isotonic and 
Fig. 155.-Illustrating principle of isotonic solutions 'must have the same 
isotonic sugar solutions. va P or pressure. For if the vapor pres- 

sures were unequal, water vapor would 

pass from the solution of higher to that of lower vapor pressure, the 
concentration of the sugar solutions would thus be changed, and water 
must again diffuse to the compartment of higher osmotic pressure. 
There would thus be established a perpetual motion which is con- 
trary to law. Consequently isotonic solutions must have the same 
vapor pressure. 

Suppose next a piece of ice / to be placed in the closed compart- 
ment above the partition M, and suppose this ice to be of the same 
temperature as the freezing point of the isotonic sucrose solution S. 
Then the vapor pressure between 7 and S must be equal, otherwise 
water vapor would pass between the two and change the freezing point 
of S. But since S and G are both isotonic and have the same vapor 
pressure, both must also have the same freezing point. 

In the same way the two isotonic solutions S and G must have the 
same boiling point, the vapor tension of the aqueous vapor at the boil- 
ing point being the same for both solutions. 

The proportionality between changes in vapor pressure and between 
changes in freezing or boiling point is easily illustrated by means of a 
diagram. In Fig. 156, let OW be the pressure curve of water for 
change in temperature and 01 the pressure curve of ice, the projection of 



MISCELLANEOUS PHYSICAL METHODS 



327 



at T being the freezing point of water. Let Ss be the corresponding 
curve of a 1 per cent sucrose solution and Gg of a 1 per cent glucose 
solution, the projection of the points s and g at t and t f being the re- 
spective freezing points of the two solutions. For comparatively 
small areas the lines gO, ss' and gg r may be regarded as straight and ss' 




t' t T 

Temperature 

Fig. 156. Showing relation of vapor pressure of sugar solutions to depression in 

freezing points. 

and gg' as parallel. In the A Ogg', Os f : Og f : : Os : Og and so also 
Os : Og : : Tt : Tt' . Therefore the lowerings in vapor pressure (and 
hence osmotic pressure) Os' and Og' of the two sugar solutions as com- 
pared with the solvent water are directly proportional to the correspond- 
ing depressions in freezing point Tt and Tt' . 

Raoult's Method for Determining Depression of Freezing Point. 
For determining the depression of freezing points by Raoult's * method 
the apparatus of Beckmann f (Fig. 157) is generally used. This con- 
sists of a large tube A (2.5 cm. X 21 cm.) provided with a side tube A'. 
The main opening is provided with a stopper through which pass the 
Beckmann thermometer D and a small stirrer, provided with a cork 
handle r. The thermometer has a range of about 6 degrees and the 
scale is divided into hundredths, the thousandths of a degree being 
estimated by aid of a magnifying glass. The tube A fits through a 
cork into the larger tube B, which serves as an air-jacket, and the 
whole sets in the cover of a large glass cylinder which is filled with a 
freezing mixture a few degrees lower than the freezing point of the 
solution to be examined. 



* Compt. rend., 94, 1517; 101, 1056; 103, 1125. 
t Z. physik. Chem., 2, 638. 



328 



SUGAR ANALYSIS 



In making an experiment, using water as the solvent, the freezing 
bath is set at about 5 C. and the mercury of the Beckmann ther- 

mometer adjusted by means of 
its regulating device c, so that 
the top of the column falls within 
the proper range of the scale. 
A weighed quantity of water, 
sufficient to cover the bulb of 
the Beckmann thermometer, is 
placed in A, the thermometer 
and stirrer are inserted and the 
tube plunged through the small 
opening b into the freezing mix- 
ture. When signs of freezing 
begin to appear, the tube is 
withdrawn from the freezing 
mixture, wiped dry and then 
inserted in the air-jacket B. 
The water and forming ice are 
now stirred vigorously by r; the 
temperature after reaching a 
certain minimum begins to in- 
crease suddenly with the lib- 
eration of latent .heat. The 
mercury soon ceases to rise and 
the point at which it stops, after 
tapping to prevent any lag, is 
taken as the freezing point of 
the water. The operation is 
repeated several times and the 
average of the observations 
taken as the final value. The 
same operations are now re- 
peated after introducing through 

A' known weights of the sugar 
be examined (1 to 5 

" . 

100 gms. of water), the maxi- 




Fig. 157. Beckmann's apparatus for de- 
termining depression of freezing point. 



mum point to which the mercury rises after overcooling being taken as 
the freezing point of the solution. The corrected difference between 
the freezing point of water and that of water + sugar is the depression 
of freezing point. 



MISCELLANEOUS PHYSICAL METHODS 



329 



Molecular Depression of Freezing Point. According to what 
was said under osmotic and vapor pressure, solutions of undissociated 
substances (non-conducting solutions), which contain the same num- 
ber of gram-molecules per liter, should show the same depression of 
freezing point. The depression for 1 gm. mol. of undissociated sub- 
stance per 1000 gms. of solvent, according to van't Hoff,* is expressed by 

002 T 2 
the formula w > in which T is the absolute temperature of melting, 

and W the latent heat of melting for the solvent. This expression in 
case of water, whose latent heat of melting is 80 calories and temper- 



ature of melting 273 abs., would give 



0.002 X 273 2 
80 



= 1.86. Loomis, as 



a matter of fact, in the examination of solutions of some 25 different sub- 
stances obtained a depression in freezing point for 1 gm. mol. to 1000 
gms. of water of almost exactly 1.86 C. The following experiments by 
Loomis f give the results of 6 tests upon maltose. (M, the molecular 
weight of maltose anhydride C^H^Ou = 342.) 



Grams maltose to 1000 
grams water (P). 


Gram-molecules of mal- 
tose to 1000 grams water 

( P V 

WA 


Depression of freezing 
point (A), 
degrees C. 


Molecular depression 
of freezing point 
f./P AM\ 

(*/M = -p-y 


3.431 


0.0100 


0.0193 


1.86 


6.879 


0.0201 


0.0378 


1.88 


10.350 


0.0302 


0.0560 


1.85 


17.316 


0.0506 


0.0946 


1.87 


35.004 


0.1023 


0.1919 


1.876 


71.548 


0.2091 


0.3946 


1.887 



Applications of Freezing-point Method. The application of the 
freezing-point method to the determination of molecular weights may 
be understood from the following example: 



20 gms. of water in the apparatus gave 
20 gms. of water + 0.3647 gms. fructose gave 
Depression of freezing-point ( A) = 



Corrected freezing point 
upon Beckmann scale. 

4.320 
4.131 



0.189 G. 

The grams of fructose calculated to 1000 gms. of water would be 
0.3647 X 1000 



20 



Since 



= 18.235 gms. = P. 

1.86P 



~ = the constant 1.86, M = 

P a 

* Ostwald's "Grundriss" (2nd Ed.), p. 142. 
t Z. physik. Chem., 37, 407. 



330 SUGAR ANALYSIS 

Substituting the values obtained for the A and P of fructose we obtain 



which agrees closely with the value 180, required by the formula C 6 Hi 2 6 . 

If w is the weight of sugar taken and W the weight of water, the 
various steps of the calculation are represented by the general equation : 

w X 1000 X 1.86 
M= JFXA 

The method of determining molecular weight by the depression of 
freezing point is one that requires considerable care in manipulation, 
and the inexperienced chemist should thoroughly test the method upon 
substances of known molecular weight before applying it to the exami- 
nation of unknown compounds. The method is open to a large number 
of experimental errors, such as too low a temperature of freezing bath, 
too high a room temperature, radiation of heat from the observer, 
faulty thermometer or error in reading, solution of air by the water, 
careless handling of the instrument, etc. For a thorough discussion 
of these various points the chemist is referred to the original papers 
by Raoult, Beckmann, Loomis and others.* Owing to the small value 
of A any slight error in its determination becomes greatly magnified in 
the final calculation. 

The freezing-point method has been successfully employed by Tollens 
and Mayer, Brown arid Morris, and others in determining the molecular 
weights of many sugars. The following examples of determinations for 
nine sugars are selected from a compilation of results by Tollens. f 



Sugar. 


Formula. 


Molecular weight. 


Authority. 


Calculated. 


Found. 


Arabinose 


C E H 10 5 
CsHioOa 
CeH^Os 
C 6 H 12 6 
C 6 H 12 6 

Cl2H22Oll 

C^HaaOn 

(^12il 22^11 , H.2V-) 

CisH 32 Oi6,5H 2 O 


150.08 
150.08 
' 180.10 
180.10 
180.10 
342.18 
342.18 
360.19 
594.32 


150.3 
154.1 
179 
174.3 
177 
352 
322 
353 
594 


Brown and Morris 
Tollens and Mayer 
Tollens and Mayer 
Brown and Morris 
Brown and Morris 
Raoult 
Brown and Morris 
Tollens and Mayer 
Tollens and Mayer 


Xylose 


Glucose 


Invert sugar 


Galactose.. ... 


Sucrose 
Maltose 


Lactose 


Raffinose . . . 





The freezing-point method can be applied to the examination of 
sugar solutions for other purposes than those of molecular weight de- 

* For a complete review and bibliography of the subject see Lippmann's 
"Chemie der Zuckerarten," 1126. f "Handbuch der Kohlenhydrate," II, p. 26. 



MISCELLANEOUS PHYSICAL METHODS 



331 



termination. Kahlenberg, Davis and Fowler,* for example, have em- 
ployed it in measuring the speed of inversion of sucrose. Table LXIII, 
by the above authorities, gives a comparison of the inversion coefficient 
of sucrose as determined by the polariscope and freezing-point methods. 
One-half gram molecule of sucrose to 1000 c.c. was inverted at 55.5 C. 
by T&<J gni. mol. of hydrochloric acid. 

TABLE LXIII 

Giving Rate of Inversion of Sucrose as Determined by Bolariscope and by Depression 

in Freezing Point 



Time. 


Polariscope read- 
ing. 


Inversion coeffi- 
cient K by polar- 
iscope. 


Depression in 
freezing point. 


Inversion coefficient 
K by freezing point. 


Hours. 



22 62 




Degrees C. 
175 




1.0 
2.0 

2.5 
3.0 
4.0 
4.5 
7.0 
17.5 
26.5 


16.58 
9.92 
7.68 
5.94 
2.54 
1.42 
-2.40 
-6.90 
-7.20 


0.0983 
0.1205 
0.1208 
0.1186 
0.1215 
0.1198 
0.1130 
0.1142 


.393 
.635 
.705 
.809 
.912 
.954 
2.105 
2.230 
2.247 


0.0977 
0.1217 
0.1185 
0.1296 
0.1263 
0.1252 
0.1254 
0.1028 














Average 


0.1158 




0.1147 



It is seen that the value of the constant K, as determined by the 



Wilhelmy equation K = - log 

t 



(p. 660), is identical by the two 



methods of measurement. 

Beckmann's Method for Determining Elevation of Boiling Point 

Beckmann'sf method of determining molecular weights by the eleva- 
tion of boiling point is the same in principle as that by depression of 
freezing point. A gram-molecule solution of an undissociated sub- 

002 T 2 

stance should show according to van't HofFs formula '- yy O n which 

T = 373 degrees, the absolute boiling point of water and W = 536 cals., 
the latent heat of evaporation), an elevation in boiling point of 
.002 X 373 2 



536 



0.519 = A. 



Beckmannt found in one experiment an elevation in boiling point 

* J. Am. Chem. Soc., 21, 1. 

t Z. physik. Chem., 3, 603; 4, 532; 6, 76; 6, 437; 8, 223. 

t Ibid., 6, 459. 



332 SUGAR ANALYSIS 

of 0.315 C. for a solution containing 216.8 gms. of sucrose to 1000 

216 8 
gms. of water, or ' = 0.634 gm. mols. The elevation in boiling 



point for a 1 gm. mol. solution would then be 'OA = 0.497 C., which 



is slightly lower than the value calculated by van't Hoff's formula. 

The general formula for calculating molecular weights from the 
elevation in boiling point (A) is similar to the formula for the freezing 
point method (p. 330) and is 

w X 1000 X 0.52 

M - JFXA 

The boiling-point method, upon the whole, is open to more sources of 
error than the freezing-point method and has proved much less satis- 
factory as a means of establishing the molecular weights of sugars. 



CHAPTER XIII 

QUALITATIVE METHODS FOR THE IDENTIFICATION OF SUGARS 

PEOBABLY no other class of organic compounds gives such a variety 
of reactions, or forms so large a number of chemical derivatives as the 
sugars. Owing to the great extent of the field it will be possible to 
describe only a few of the more general tests and reactions. 

In describing the various chemical tests, the sugars will be classified 
for convenience under two general groups: I. The reducing sugars. 
II. The non-reducing sugars. The reducing sugars are distinguished 
by the fact that they cause a marked precipitation of cuprous oxide 
when warmed with Fehling's alkaline copper solution, whereas the non- 
reducing sugars do not exhibit this property, or only to a very slight 
extent after prolonged boiling. The reducing sugars constitute by far 
the larger group; of the some one hundred known natural or synthetic 
sugars, about ninety are reducing and only about ten non-reducing. 

Reactions of the Reducing Sugars 

The characteristic chemical properties of the reducing sugars are 
due for the most part to the occurrence of a common carbonyl-alcohol 

H-C-OH 

group I . The reducing sugars, as aldoses or ketoses, give in 
C = O 
I 

fact nearly all the reactions peculiar to aldehydes and ketones. The 
chemist must, therefore, first of all, guard against deciding as to the 
presence oi^a sugar from a reaction which would also be given by 
formaldehyde, acetaldehyde or acetone. A number of confirmatory 
tests must usually be applied, before it can be stated definitely whether 
a sugar is or is not present. 

The qualitative reactions for reducing sugars are divided for con- 
venience into I. General tests; II. Special tests; III. Individual 
tests. After it has been determined from general tests that a sugar is 
present, special tests must be applied in order to determine what classes 
or groups of sugars are present, whether hexoses or pentoses, aldoses or 
ketoses, monosaccharides or disaccharides. After the class or group of 
sugars has been ascertained, individual tests must be applied in order 

333 



334 SUGAR ANALYSIS 

to determine what particular sugars are present. Only the general 
and special tests are taken up in the present chapter. The individual 
tests are given under the description of the different sugars in Part II. 

GENERAL TESTS FOR REDUCING SUGARS 

Among the general tests which are sometimes given for sugars may 
be mentioned the familiar property which all carbohydrates have of 
giving off a characteristic sweetish odor upon heating over a flame in a 
closed tube. This odor, which is usually designated as caramel-like, is 
given off, however, by many polyatomic alcohols and acids (as by tar- 
taric acid) so that the test is not characteristic of sugars alone. Among 
the decomposition products obtained by heating sugars in a closed tube 
may be mentioned (besides water and the gaseous products carbon di- 
oxide and carbon monoxide) formic acid, acetic acid, acetone, furfural 
and various products of an aldehyde nature. It is to the furfural 
and aldehyde products that the characteristic odor of burnt sugar is 
largely due. 

The general tests for reducing sugars may be divided for conven- 
ience into four general groups of reactions. 
I. Reducing reactions with alkaline solutions of metallic salts. 
II. Color reactions with alkalies, acids and phenols. 

III. Hydrazone and osazone reactions with phenylhydrazine and its 

substituted derivatives. 

IV. Miscellaneous reactions. 

I. REDUCING REACTIONS OF SUGARS WITH ALKALINE SOLUTIONS OF 
METALLIC SALTS 

The simple sugars and certain of the disaccharides, as maltose and 
lactose, have the property of reducing alkaline solutions of many 
metallic salts, such as those of copper, silver, mercury and bismuth. 
'This reaction, which is common to most aldehydes, is due to the with- 
drawal of oxygen from the metallic base, the latter being precipitated 
either as a suboxide or in the metallic form. The aldehyde group of 
the sugar molecule is oxidized by the oxygen withdrawn from the 
metallic base to the acid carboxyl group, as indicated by the following 
general equation: 

H-C:O + 2CuO = H-O-C:O + Cu 2 O. 

Aldehyde Copper Oxide Acid Copper Suboxide. 

The above, however, marks only the beginning of the reaction, for, upon 
heating, the oxidation of the sugar molecule usually proceeds with the 



METHODS FOR THE IDENTIFICATION OF SUGARS 335 

conversion of alcohol into carboxyl groups as in the following reaction 
for glycol aldehyde : 
H 

H-C-O-H + 3Ag 2 = 0:C-0-H + 6Ag + H 2 O. 
H-C:O O:C-O-H 

Glycol Silver oxide Oxalic Metallic Water, 

aldehyde acid silver 

This oxidation in the case of the higher monosaccharides is usually at- 
tended by a breaking down of the carbon chain as by the oxidation of 
glucose in ammoniacal silver solution : 

C 6 H 12 6 + 9 Ag 2 O = 3(COOH) 2 + 18 Ag + 3 H 2 0. 

The reaction between sugars and alkaline salts of metals, as ordinarily 
carried out, gives rise to a number of monobasic and dibasic acids 
(formic, oxalic, etc.) in varying proportions according to the conditions 
of the experiment. It is not possible, therefore, to express the reaction 
.by chemical equations except in a very general way. 

The most common of the alkaline salt solutions employed in test- 
ing sugars are those of copper. The sulphate and acetate of copper 
are the salts most generally used and sugar literature is filled with 
descriptions of modifications for making the test. Only a few of these 
will be described. 

Fehling's Copper Solution. This is the most common chemical 
reagent employed in testing sugars. As ordinarily prepared the reagent 
consists of two solutions: solution A containing 34.64 gms. crystallized 
copper sulphate to 500 c.c. and solution B containing 173 gms. Rochelle 
salts and 51.6 gms. sodium hydroxide to 500 c.c. The solutions are the 
same as those used in quantitative analysis and are to be kept separate 
until just before using. By mixing 5 c.c. each of solutions A and B in 
a test tube, adding a few c.c. of the solution to be examined and heat- 
ing to boiling for 2 minutes, a brick-colored precipitate of cuprous oxide, 
Cu 2 0, will form, if reducing sugars are present, the intensity of coloration 
and amount of precipitate being proportional to the amount of sugar 
present. The test is sensitive to about 0.01 mg. of glucose to 1 c.c. 

Products Obtained by Heating Reducing Sugars with Fehling's Solu- 
tions. The chemical reactions which take place in the oxidation of 
sugars by means of Fehling's solution are exceedingly complex. Nef,* 
who has made the most complete studies in this field, found that in 
case of 1-arabinose, the oxidation proceeds along three separate lines. 

* Ann., 357, 214-312. 



336 SUGAR ANALYSIS 

I. From 10 to 25 per cent of sugar are oxidized to form pentonic 
acids. 

C 5 Hi 5 + = C 5 Hi 6 . 

II. From 35 to 45 per cent of sugar are oxidized to form formic and tri- 

oxybutyric acids. 

C 5 H 10 5 + 2O = HCOOH + C 4 H 8 5 . 

III. From 30 to 38 per cent of sugar are oxidized to form formic and 

glycollic acids. 

C 5 Hio0 5 + 30 = HCOOH + 2 C 2 H 4 3 . 

In case of the hexose sugars, d-glucose, d-mannose and d-fructose, 
Nef obtained analogous reactions with formation of carbonic, formic, 
glycollic, glyceric, trioxybutyric and hexonic acids. The amount of the 
different acids was found to vary according to the amount of alkali 
present. 

In testing solutions containing much foreign organic matter such as 
urine, the reaction with Fehling's solution may be interfered with. 
Uric acid, creatine, creatinine, albumin, peptones and other substances 
may either check the precipitation of cuprous oxide, when reducing sugars 
are present, or in some cases cause a precipitate of copper in the com- 
plete absence of sugars. Solutions containing xanthine bases, such as 
low-grade molasses, distillery waste, etc., when heated with Fehling's 
solution may precipitate greenish-yellow copper compounds, which may 
be mistaken for cuprous oxide. In all such cases the impure solution 
should be clarified with a little normal acetate of lead and filtered ; any 
excess of lead is removed from the filtrate with sodium carbonate and 
the clear solution tested with Fehling's reagent in the usual way. 
Filtering the- impure solution through animal charcoal is also of ad- 
vantage when foreign coloring matter masks the reaction. 

Barfoed's Copper Solution. Instead of the sulphate, solutions 
of other copper salts have been employed in testing for sugars. Bar- 
foed * has prepared a solution containing one part crystallized neutral 
copper acetate in 15 parts of water; 5 c.c. of 38 per cent acetic acid are 
added to 200 c.c. of the copper-acetate solution before use. On boiling 
the solution a basic acetate of copper is formed, the liberated cupric 
oxide being reduced in presence of monosaccharides. Barfoed's reagent 
is not reduced to any great extent by the disaccharides, lactose and 
maltose, and is, therefore, of value in distinguishing these sugars from 
monosaccharides. 

* Z. analyt. Chem., 12, 27. 



METHODS FOR THE IDENTIFICATION OF SUGARS 337 

Soldaini's Copper Solution. Carbonate of copper solution has 
also been used in testing for sugars. Soldaini * has prepared a solution 
containing 15 gms. precipitated copper carbonate, CuGO 3 , and 416 gms. 
potassium bicarbonate, KHCO 3 , dissolved to 1400 c.c. Instead of start- 
ing with copper carbonate, copper sulphate may be used; a solution of 
the latter is added to the KHCOs solution, the precipitate of CuCOs 
first formed being dissolved in the excess of bicarbonate. A solution 
containing 3.464 gms. copper sulphate and 297 gms. potassium bicar- 
bonate to 1000 c.c. is especially adapted for detecting small amounts of 
reducing sugars. 

Among other copper solutions recommended for testing sugars may 
be mentioned copper ammonium tartrate and ammoniacal copper sul- 
phate or acetate. None of these preparations has been found, how- 
ever, to equal Fehling's reagent for general usefulness in practical 
sugar analysis. 

Tollens's Silver Solution. The most sensitive of metallic-salt solu- 
tions for detecting sugars is ammoniacal silver solution, first employed 
by Tollens f and hence usually known as Tollens's reagent. This is pre- 
pared by dissolving one part silver nitrate in 10 parts of water; a second 
solution is then made containing one part sodium hydroxide in 10 parts 
of water. Before making the test equal parts of the two solutions are 
mixed and then ammonia added drop by drop until the precipitate of 
silver oxide is completely dissolved. A solution containing one part of 
glucose in 1000 parts of water will cause a strong reduction of Tollens's 
reagent in the cold, a mirror of silver being deposited within 15 min- 
utes. A solution containing one part glucose to 100,000 parts of water 
will also produce a perceptible reduction. in the cold, but the solution 
must stand one to two days. The reduction takes place more rapidly 
upon warming, but warming or heating the solution is to be avoided 
owing to the danger of forming explosive silver compounds. For the 
latter reason the reagent should be prepared only just before using. 
Tests should be carried out in the dark and solutions containing the 
reagent should not be kept for any length of time. 

Tollens's silver reagent is also reduced by all aldehyde substances; 
it is affected not only by the sugars which reduce Fehling's solution 
but also by sucrose, raffinose and all other soluble carbohydrates. 
Even the alcohol derivatives of the sugars produce reduction, glycerol, 
for example, causing the formation of a silver mirror. The readiness 
with which ammoniacal silver solution is reduced by soluble organic 

* Z. Ver. Deut. Zuckerind., 39, 933; 40, 792. 
t Ber., 15, 1635; 16, 921. 



338 SUGAR ANALYSIS 

non-sugars has proved a serious objection against the use of this reagent 
in ordinary analytical work. 

Knapp's Mercury Solution. A third reagent which has been used 
for testing sugars is Knapp's* alkaline mercuric-cyanide solution. The 
latter contains 10 gms. of mercuric cyanide dissolved in 100-c.c. sodium 
hydroxide solution of 1.145 specific gravity. Similar alkaline solutions 
have been prepared by Sachsse f from mercuric iodide and by Bauer J 
from mercuric chloride. These solutions are reduced upon warming 
with sugar solutions giving grayish deposits of metallic mercury. 
The mercury solutions have the same objection, however, as those of 
silver in being reduced by different organic non-sugars, such as creatine, 
creatinine and glycerol and even under certain conditions by alcohol. 
Alkaline solutions of mercury salts are, therefore, of but little value in 
detecting sugar in urine and other liquids rich in organic non-sugars. 

Nylander's Bismuth Solution. A fourth reagent, which has been 
used considerably for detecting reducing sugars in urine, is an alkaline 
solution of bismuth sub-nitrate, known as Nylander's (or Almen's) 
reagent. This solution as prepared by Nylander is made by dissolving 
2 gms. of bismuth sub-nitrate and 4 gms. of Rochelle salts in 100 gms. 
of 8 per cent sodium hydroxide solution. After standing for a few days 
the solution is filtered through glass wool and the clear filtrate preserved 
in a stoppered bottle. The solution will keep indefinitely. When 
Nylander's reagent is heated with a solution containing reducing sugars 
a precipitate of dark metallic bismuth is produced. Heating with T V its 
volume of 0.01 per cent glucose solution will cause a perceptible darken- 
ing. In testing urine 1 c.c. of the reagent and 10 c.c. of urine are 
heated in a test tube 2 to 5 minutes over the flame; after standing 
for 5 minutes the solution is examined for the appearance of a dark- 
colored sediment. 

Nylander's reagent, however, is open to the same objections noted 
for the alkaline silver and mercury solutions. The presence of albu- 
min, nuclein, glucuronic acid and other organic non-sugars in urine 
will also cause a precipitation of bismuth, even when glucose is com- 
pletely absent. While the failure of a precipitate with Nylander's re- 
agent may indicate the absence of reducing sugars, the occurrence of a 
precipitate may be said to indicate the presence of sugar only when re- 
ducing non-sugars are proved to be absent. 

* Z. analyt. Chem., 9, 395. 
t Z. Ver. Deut. Zuckerind., 26, 872. 
t Landw. Vers.-Stat., 36, 304. 
Z. physiol. Chem., 8, 175. 



METHODS FOR THE IDENTIFICATION OF SUGARS 339 

Miscellaneous Solutions. Of other alkaline solutions of metallic 
salts proposed for sugar testing may be mentioned alkaline nickel sul- 
phate and tartaric acid which gives a dark-red precipitate of nickel sub- 
oxide in presence of reducing sugars, and alkaline ferric chloride and 
sodium tartrate which gives a brown-colored precipitate on heating 
with reducing sugars. None of these reagents, however, or any of the 
other alkaline solutions of metallic salts previously mentioned, has been 
found to equal Fehling's copper reagent for all-around usefulness and 
reliability. 

II. COLOR REACTIONS OF SUGARS WITH ALKALIES, ACIDS AND PHENOLS 

As a second general reaction of reducing sugars may be mentioned 
certain color effects which nearly all soluble carbohydrates give when 
brought into contact with different reagents. The reagents employed 
may be divided into three groups: 

I. Alkalies. 

II. Concentrated mineral acids. 
III. Phenols. 

Color Reactions of Sugars with Alkalies. All reducing sugars 
have the property of coloring solutions of the alkalies and alkaline 
earths yellow, the application of heat turning the color a dark brown. 
This reaction is common to all aldehydes. The exact nature of the 
coloring matter formed by the action of alkalies upon sugars in solution 
is not understood. Considerable oxygen is absorbed from the air dur- 
ing the reaction and a variety of products of an acid nature are among 
the substances formed. 

Products Obtained by Heating Reducing Sugars with Alkali. Lactic 
acid is produced in considerable amount by the action of alkalies upon 
many reducing sugars such as xylose, arabinose, glucose and fructose. 
The presence of calcium lactate in certain sugar-cane molasses is ex- 
plained by the action of an excess of lime during clarification upon the 
reducing sugars of the juice. Formic, acetic and oxalic acids have also 
been found among the products resulting from the action of alkalies 
upon sugars in solution. Certain phenol bodies such as pyrocatechin 
and protocatechuic acid have also been detected among the oxidation 
products of sugars resulting from treatment with alkalies. 

Nef * has studied the action of J normal sodium hydroxide upon dif- 
ferent sugars and obtained in case of d-glucose, d-mannose, and d-fruc- 

* Ann., 376, 1-119. 



340 SUGAR ANALYSIS 

tose a yield of from 40 to 45 per cent d,l-lactic acid, from 10 to 15 per 
cent d,l-l-hydroxybutyrolactone, about 25 per cent of saccharin, meta- 
saccharin and isosaccharin and a small quantity of tarry decomposition 
products. 

The action of dilute alkalies in causing transformations of sugars into 
one another by molecular rearrangement is referred to elsewhere. 

Color Reactions of Sugars with Mineral Acids. Treatment of 
solutions of sugars and carbohydrates with concentrated mineral acids 
gives rise to a number of decomposition products, the color of which 
frequently throws some light upon the nature of the sugars present. 
The acids most commonly used for this purpose .are sulphuric and 
hydrochloric. The character of the color generated will depend partly 
upon the kind of sugar, partly upon the strength of acid used and 
partly upon the temperature of the reaction. 

Products Obtained by Heating Sugars with Acids. The darkening 
produced in all sugar solutions upon warming with concentrated sul- 
phuric or hydrochloric acid is due largely to the formation of insoluble 
so-called "humus" substances of relatively high carbon content (C = 62 
to 67 per cent and H = 3.5 to 4.5 per cent), the percentage of carbon and 
depth of color increasing with the strength of acid used. Attempts have 
been made to classify the humus substances formed by the action of 
acid upon sugars into ulmin and humin and ulmic and humic acids, 
to which various formulae have been assigned by different authorities. 
The constitution of the humus substances has not been definitely 
settled, however, and until considerable more work has been done the 
formulae of these must remain more or less a matter of conjecture. 

In addition to the insoluble humus substances a number of soluble 
and volatile products are formed by the action of sulphuric and hydro- 
chloric acids upon sugars. Among such products may be mentioned 
formic acid, levulinic acid, furfural, methylfurfural, oxymethylfurfural 
and a number of dextrin-like condensation or reversion products of high 
specific rotation. The nature and amount of these various products 
depend largely upon the kind of sugar, and a number of methods of 
group distinction are based upon the separation of characteristic de- 
compositfon products. Further reference will be made to these under 
the special reactions. 

The ketoses are much more easily decomposed by strong mineral 
acids than the aldoses and their solutions give rise to color reactions 
with corresponding greater facility. This offers one means of dis- 
tinguishing between a ketose and aldose or of detecting a ketose sugar 
in presence of an aldose. If a cold sugar solution be treated in a test 



METHODS FOR THE IDENTIFICATION OF SUGARS 341 

tube with a few cubic centimeters of concentrated sulphuric acid, allow- 
ing the latter to flow down the walls of the tube to the bottom without 
shaking, a brown ring will quickly form at the junction of the acid and 
sugar solution if fructose, sucrose or a sugar containing the ketone 
group is present; with glucose, lactose, maltose and the aldoses in 
general no such coloration will develop. 

Color Reactions of Sugars with Phenols. The most distinctive 
color reactions of the sugars are those obtained by treatment with 
different phenols in presence of concentrated hydrochloric or sulphuric 
acid. The development of a color in this case is due to the formation 
of condensation products between the phenol derivatives and the de- 
composition products obtained from the sugar (humus substances, 
furfural, aldehydes, etc.). a-Naphthol, thymol, resorcin, orcin, naph- 
thoresorcin and phloroglucin are among the more important phenol de- 
rivatives used for making color reaction with sugars. 

The color reactions with the phenols are performed in various 
ways. The test with a-naphthol, for example, which is perhaps used 
more frequently than any of the others, is made as follows: 1 to 2 
cubic centimeters of the sugar solution are treated in a test tube with 
1 to 2 drops of a 10 to 20 per cent alcoholic solution of a-naphthol. A 
few cubic centimeters of concentrated sulphuric acid (must be free from 
nitric acid) are then carefully added so as to flow down the walls of the 
tube to the bottom. If sugars containing a ketone group are present 
a violet ring will form instantly at the junction of the two liquids; in 
presence of aldoses a gentle warming of the test tube is usually neces- 
sary in order to bring out the full intensity of color. The a-naphthol 
test, which is of extreme delicacy, is frequently employed in sugar 
houses and refineries in testing the condensation water from the vacuum 
pan for presence of sucrose lost by entrainment. 

If the reaction described for a-naphthol is carried out with thymol, 
menthol, resorcin and other phenols similar colorations are produced, 
the tints varying from cherry red to deep purple. 

The tests with phenols and hydrochloric acid are usually made by 
warming a few cubic centimeters of the sugar solution with a solution 
of the phenol (resorcin, orcin, phloroglucin, etc.) in concentrated hydro- 
chloric acid. The colorations thus obtained are usually very brilliant, 
varying in tint from a bright red to a bluish violet. The colors formed 
are not permanent, however; they rapidly darken and the clear-colored 
solution soon becomes turbid with the precipitation of a dark-colored 
condensation product. 



342 



SUGAR ANALYSIS 



USE OF THE SPECTROSCOPE IN STUDYING COLOR REACTIONS FOR SUGARS 

The spectroscope has been used with great success by Tollens and 
his coworkers in studying the colors obtained by treating sugars with 
different reagents. The appearance of characteristic absorption bands 
in different parts of the spectrum, when the colored solution is viewed 
through the spectroscope against white light, is peculiar of many sugars. 

Description of Direct-vision Spectroscope. A simple type of 
spectroscope for studying absorption spectra is the direct vision in- 




Fig. 158. 




3- 



Fig. 159. 
Showing outer and inner construction of a direct-vision spectroscope. 

strument illustrated in Fig. 158, the interior construction of which is 
shown in Fig. 159. 

The essential parts of the apparatus consist of a telescopic tube con- 
taining an Amici prism P and an achromatic objective 0. At one end 
of the tube, protected by the screw cap K, a diaphragm is situated con- 



METHODS FOR THE IDENTIFICATION OF SUGARS 343 

taining a narrow slit S, the width of which can be adjusted by turning 
the milled ring B. The upper half of the slit is covered with a small 
prism V; a mirror D, which can be rotated through a small angle about 
the axis of the tube, is also attached to the slit end of the instrument. 

At the prism end of the spectroscope there is fixed a small lateral 
tube T containing a graduated scale E. The latter is attached to a 
small prism b to which is fixed a converging lens a. At R is a right 
angle prism, from the hypotenuse surface of which the image of the 
scale E is reflected through the achromatic objective 0' upon the cut 
surface cc of the Amici prism. 

If the slit end of the spectroscope be pointed towards a sodium 
flame the rays of light will pass into the spectroscope along the paths 1, 
2 and 3. The telescope is first focused by turning the milled ring G 
until a sharply defined image of the lower uncovered half of the slit 
is obtained by the light passing along 2 upon the surface cc. The image 
of the scale E is reflected at the same time, by the light passing along 
3 ; also upon cc. The position of the sodium line is noted upon the 
graduated scale, the latter being in this way standardized. If the 
spectroscope be now directed towards the sky a continuous spectrum 
is obtained upon the surface cc; the mirror D is next turned until the 
light passing along 1 is reflected through an opening in the cap K upon 
the small prism V and thence through the upper half of the slit S', in 
this way a continuous spectrum is obtained upon cc the width of which 
is equal to the total length of the slit S. 

If the slit has been sufficiently reduced in width, the spectrum of 
sunlight is seen to be crossed by a number of dark lines, the so-called 
Fraunhofer lines, which are due to the absorption of certain rays of 
light from the incandescent mass of the sun by the vaporized elements 
of the solar atmosphere. A dark line (the D line of Fraunhofer's scale), 
for example, corresponds to the position of the bright-yellow line 
obtained with the sodium flame and so of the other elements. The 
position and wave-length of the more important Fraunhofer lines is 
shown in Fig. 165 (p. 384) ; their presence is very helpful in defining the 
position of absorption spectra. 

For studying absorption spectra the spectroscope is mounted upon 
a stand as shown in Fig. 160, a screen L being attached to the tube to 
shade the eye of the observer. The solution to be examined is placed 
in a small cell T, before the front opening in the screw cap and viewed 
against white light. The rays of light absorbed by the solution will 
cause characteristic dark-colored bands to appear upon that part of the 
spectrum corresponding to the lower half of the slit. The part of the 



344 



SUGAR ANALYSIS 



spectrum corresponding to the half of the slit covered by the prism V 
meanwhile remains continuous and together with the scale, or Fraun- 
hofer lines, serves for the exact location of the absorption bands. 





_... - :: 

Fig. 160. Fig. 161. 

Methods of mounting apparatus for study of absorption spectra. 

Solutions which arc only weakly absorptive are best examined through 
a large tube H, in the manner shown in Fig. 161. The spectroscope is 
turned and clamped in a vertical position and the light reflected upward 
from the mirror F through the glass bottom of the support G. 

Tollens's Method of Studying Absorption Spectra. In preparing 
color tests of sugar solutions for spectroscopic examination it is im- 
portant that the color remain permanently in solution and that no tur- 
bidity develop which would obscure the visible parts of the spectrum. 
This is sometimes accomplished by carrying out the reaction in presence 
of alcohol or some other solvent to hold the color compound in solu- 
tion. A better way is by use of Tollens's* deposit method ("Absatz- 
methode ") In this method the deposit of insoluble condensation prod- 

* Ber., 29, 1202. 



METHODS FOR THE IDENTIFICATION OF SUGARS 345 

ucts obtained by treating the sugar solution with hydrochloric acid 
and the phenol (orcin, phloroglucin, naphthoresorcin, etc.) is filtered off, 
washed several times with water and then dissolved in alcohol. Bright- 
colored solutions are thus obtained which can be brought by dilution with 
alcohol to the degree of intensity suitable for spectroscopic examination. 
Descriptions of characteristic absorption spectra will be given under the 
reactions for groups and individual sugars. 

. Of less importance than the color reactions with phenols are the 
color tests obtained by treating sugars with aromatic amines (aniline, 
xylidine, diphenylamine, etc.) in presence of concentrated hydrochloric 
acid. The colors in this instance are due to a combination between the 
aromatic amine and the furfural, methylfurfural, and oxymethyl- 
furfural derived from the decomposition of the sugar. 

III. HYDRAZONE AND OSAZONE REACTIONS OF REDUCING SUGARS WITH 
PHENYLHYDRAZINE AND ITS SUBSTITUTED DERIVATIVES 

In many respects the most important of the qualitative tests for 
sugars are those obtained with phenylhydrazine and its substituted 
derivatives. Phenylhydrazine was introduced as a reagent in sugar 
chemistry by Emil Fischer* in 1884; it has been of immense service 
not only as a means of separation and identification but also in first 
opening a way to a thorough understanding of the molecular constitu- 
tion of sugars. 

Hydrazone Reaction. The reaction with phenylhydrazine is 
limited to such sugars as contain a free carbonyl group and proceeds in 
two phases with production of two entirely different classes of com- 
pounds. The first phase of the reaction is common to all aldehydes 
and ketones, the of the carbonyl group combining with H 2 of the 
amino group in the phenylhydrazine with formation of a group of 
compounds called hydrazones. With formaldehyde, for example, the 
reaction proceeds as follows: 

H 2 C:0 + H 2 N-NHC 6 H 6 = H 2 C:N-NHC 6 H 5 + H 2 O 

Formaldehyde Phenylhydrazine Formaldehyde- Water 

phenylhydrazone 

With the carbonyl group of a sugar the reaction would be for a 
diose : 

CH 2 OH CH 2 OH 

HC:0 + H 2 N-NHC 6 H 6 = HC:N-NHC 6 H 6 + H 2 O 

Diose Phenylhydrazine Dioae-phenylhydrazone Water 

* Ber., 17, 579. 



346 SUGAR ANALYSIS 

The hydrazone reaction is carried out by treating the sugar solution 
in the cold with a solution containing one volume of phenylhydrazine, 
one volume of 50 per cent acetic acid, and three volumes of water. A 
little more of the phenylhydrazine is used in making the test than the 
theoretical quantity corresponding to the supposed amount of sugar 
present. In place of the above solution the crystalline chloride of 
phenylhydrazine may be used to advantage, a few grams of sodium 
acetate being also added to promote the reaction. After the above 
treatment the hydrazones of the sugars will separate sooner or later 
as well-defined crystalline compounds, the length of time for separation 
depending upon the solubility of the hydrazones formed. The phenyl- 
hydrazone of mannose, for example, being very insoluble, will separate 
almost immediately; those of the methylpentoses, fucose, rhamnose 
and rhodeose also deposit readily; the phenylhydrazone of glucose, on 
the other hand, which is quite soluble in water, may require one or two 
days for its precipitation. By filtering off the hydrazones as they are 
formed a separation of sugars in mixtures may often be accomplished. 

After separation of the hydrazones the latter are filtered off and re- 
crystallized either from water or, in case of difficultly soluble hydrazones, 
from alcohol or pyridine. 

Use of Substituted Derivatives of Phenylhydrazine. In place of 
phenylhydrazine any of its substituted derivatives may be used for the 
purpose of precipitating sugars. The substituted phenylhydrazines 
yield in many cases characteristic hydrazones with sugars and their use 
in sugar chemistry in recent years has been of the greatest service. Of 
the various substituted phenylhydrazines the following are among the 
most important. 



1. Methylphenylhydrazine H 2 N N 

X C 6 H 5 

2. Ethylphenylhydrazine H 2 N N x 

N C 6 H 5 



3. Amylphenylhydrazine H 2 N N 

X C 6 H 5 

C H 

4. Allylphenylhydrazine H 2 N N 

N C 6 H 5 



X C 6 H 5 



5. Diphenylhydrazine H 2 N N 

\ 

C* TT 

6. Benzylphenylhydrazine H 2 N N / 

x CeH 6 



TT 



METHODS FOR THE IDENTIFICATION OF SUGARS 347 

7. Parabromophenylhydrazine H 2 N N 

8. Paranitrophenylhydrazine H 2 N N 



C 6 H 4 Br 

TT 

C 6 H 4 N0 



Other hydrazines than those of the phenyl group are also employed 
as, for example, 

TT 

9. Naphthylhydrazine H 2 N-N^ 

CloHj 

The reactions with the substituted hydrazines are usually best 
carried out in alcoholic solution, the hydrazones formed being for the 
most part much less soluble than those of ordinary phenylhydrazine. 

In the examination of the hydrazones obtained from sugar solutions 
a melting point of the product is taken before and after recrystalliza- 
tion. If the melting point remains unchanged the hydrazone is pure. 
Should a difference in the temperature of melting be obtained the 
hydrazone should be recrystallized until successive determinations show 
no change in melting point. A table of melting points will then usually 
identify the hydrazone of the sugar. (See Table 24, Appendix.) 

Separation of Sugars from Hydrazones. When a sufficient quantity 
of hydrazone is available it is always well to decompose the compound 
and make a direct examination of the separated sugar. For the separa- 
tion of sugars from their hydrazones two processes are available: 
First, by means of concentrated hydrochloric acid as originally used by 
Fischer. Second, by means of benzaldehyde and formaldehyde as rec- 
ommended by Herzfeld* and by Ruff.f 

When the hydrazone of a sugar is treated with concentrated hydro- 

chloric acid the chloride of the hydrazine and free sugar are formed: 

C 6 H 12 5 N-NHC 6 H5 + HC1 + H 2 = C 6 H 12 O 6 + HC1 H 2 N-NHC 6 H 5 

Hexose-pheny I hydrazone Hexose Phenylhydrazine 

chloride 

The phenylhydrazine chloride is almost insoluble in concentrated hydro- 
chloric acid and is removed by filtration. The filtrate is neutralized with 
lead carbonate; the lead chloride is filtered off and the filtrate evapo- 
rated to a syrup. The latter is shaken with 95 per cent alcohol, any 
remaining lead chloride filtered off and the alcoholic filtrate evaporated 
to a sirup which is set aside for the sugar to crystallize. 

* Ber., 28, 442. 
t Ber., 32, 3234. 



348 SUGAR ANALYSIS 

The separation of sugars from their hydrazones by means of alde- 
hydes is much simpler than by use of hydrochloric acid and this is the 
process most generally used at present. For this purpose benzaldehyde 
is usually employed for the hydrazones of phenylhydrazine and formal- 
dehyde for the hydrazones of the substituted hydrazines. The reaction 
between the aldehyde and hydrazone is a simple one, the aldehyde dis- 
placing the sugar with formation of aldehyde hydrazone. 



+ CeHsCHO = C 6 H 12 6 + C 6 H 6 CHN-NHC 6 H5 

Hexose-phenylhydrazone Benzaldehyde Hexose Benzaldehyde- 

phenylhydrazone 

C 6 Hi 2 O 5 N-N(C 6 H 5 ) 2 + CH 2 O = C 6 Hi 2 O 6 + CH 2 N-N(C 6 H 5 ) 2 

Hexose-diphenylhydrazone Formaldehyde Hexose Formaldehyde- 

diphenylhydrazone 

The reaction is best carried out by treating a solution of the hydra- 
zone in 50 per cent alcohol in a flask with an amount of the aldehyde 
slightly in excess of the theoretical quantity necessary to effect de- 
composition. The flask is then attached to a reflux condenser and 
the solution gently boiled for an hour. After cooling, the solution is 
filtered from the aldehyde hydrazone, the filtrate shaken out several 
times with ether in a separatory funnel, the sugar solution, after de- 
colorizing with animal charcoal, evaporated to a sirup and set aside for 
crystallization. Should crystallization not take place immediately, 
the process may be promoted by priming the sirup with a minute 
crystal of the sugar suspected to be present. After crystallization the 
sugar crystals are filtered off, washed with alcohol and ether (using 
suction) and dried between filter paper in a desiccator over concen- 
trated sulphuric acid. The identity of the sugar thus obtained is then 
established by determination of its specific rotation. 

If the filtrate obtained from filtration of a hydrazone be shaken out 
with ether to remove excess of hydrazine, the solution can be treated a 
second time with a different hydrazine. In this manner a qualitative 
separation of several mixed sugars may be accomplished. 

Osazone Reaction. While the hydrazone reaction is of pre- 
eminent value in the isolation of sugars, the osazone test with phenyl- 
hydrazine is usually of more qualitative significance owing to the 
greater insolubility of the osazones in water and the consequent 
greater rapidity and ease of their separation as compared with hydra- 
zones. 

If a solution of a reducing sugar be treated with an excess of phenyl- 
hydrazine and then warmed, two molecules of phenylhydrazine unite 
with the sugar molecule forming an osazone. The aldehyde or ketorffe 



METHODS FOR THE IDENTIFICATION OF SUGARS 349 

group of the sugar and the adjacent alcohol group are the ones which 
always participate in this reaction. 

CH 2 OH CH 2 OH 

(HCOH)s_ (HCOH) 3 

HCOH + ^N-NHCeHs = C:N-NHC 6 H 6 + 2H 2 O + . H, 
HC:O HC:N-NHC 6 H 6 

Hexose Phenylhydrazine Hexose-phenylosazone Water Hydrogen 

The free hydrogen liberated in the above reaction acts upon a part 
of the excess of phenylhydrazine reducing this to aniline with liberation 
of ammonia. 

H 2 N-NHC 6 H 5 + H 2 NH 2 C 6 H 5 + NH 3 

Phenylhydrazine Hydrogen Aniline Ammonia 

Since the first stage in the reaction with phenylhydrazine is the for- 
mation of a hydrazone, it follows that all phenylhydrazones when 
treated with phenylhydrazine in excess are changed to the correspond- 
ing osazones. 
C 6 H 12 5 N-NHC 6 H 5 + H 2 N-NHC 6 H 5 = C 6 H 10 O4(N-NHC 6 H 5 ) 2 + H 2 O + H 2 

Hexose-phenylhydrazone Phenylhydrazine Hexose-phenylosazone Water Hydrogen 

In conducting the reaction for osazones the original method of 
Fischer* is usually followed. For 1 gm. of sugar, 2 gms. of phenyl- 
hydrazine chloride and 3 gms. crystallized sodium acetate (CH 3 COONa + 
3 H 2 O) and 20 c.c. of water are heated together for f to 1J hours in a 
large test tube of about 50 c.c. capacity placed in a boiling-water bath. 
The contents of the tube are stirred occasionally to promote crystalliza- 
tion. Instead of the chloride one may employ a solution of phenyl- 
hydrazine acetate, prepared by adding concentrated acetic acid drop by 
drop to phenylhydrazine until the turbid emulsion clears. The osa- 
zone reaction with the substituted hydrazines is conducted in the same 
way as with phenylhydrazine. 

The osazones of the sugars are yellowish-colored crystalline com- 
pounds of variable solubility. The osazones of the monosaccharides 
crystallize out from the hot solutions; those of the disaccharides, maltose 
and lactose, however, separate only after cooling. A separation of the 
osazones of the mono- and disaccharides can be accomplished in this 
manner, a second crystallization usually rendering the separation com- 
plete. While the osazones of the monosaccharides are nearly all of 
much lower solubility than the corresponding hydrazones, the osazone . 
separation is never complete. 

* Ber., 17, 579. 



350 



SUGAR ANALYSIS 



Yield and Time for Formation of Osazones. Sugars differ greatly in 
the amount of osazone which is formed under a definite method of 
treatment, and this property has been utilized as a means of identifica- 
tion. Maquenne,* for example, has determined the yield of osazones 
obtained by heating 1 gm. of different sugars in 100 c.c. of water with 
5 c.c. of a solution, containing 40 gms. phenylhydrazine and 40 gms. 
glacial acetic acid in 100 c.c., for 1 hour in a boiling- water bath. The 
sugars studied by Maquenne are arranged in Table LXV in the order 
of yield of osazone. 

TABLE LXV 

Showing Yield of Osazones and Time of Precipitation for Different Sugars 



Sugar. 


Phenylosazone 
from 1 gram 
sugar. 


Time for precipitation. 


Sorbose 
Fructose 


Gram. 

0.82 
0.70 


Turbid in 12 min. 
Precipitate in 5 min. 


Xylose 
Glucose 


0.40 
0.32 


Precipitate in 13 min. 
Precipitate in 8 min. 


Arabinose 


0.27 


Turbid in 30 min 


Galactose 


0.23 


Precipitate in 30 min 


Rhamnose 


0.15 


Precipitate in 25 min 


Lactose 


0.11 


Precipitate only on cooling. 


Maltose 


0.11 


Precipitate only on cooling. 



It is noted that the ketoses, sorbose and fructose are characterized 
by a much greater yield of osazone. The theoretical yield of osazone 
from 1 gm. of sugar is 2.19 gms. for pentoses, 1.99 gms. for hexoses and 
1.53 gms. for disaccharides. This shows how large a part of even the 
more insoluble osazones were unprecipitated in Maquenne's experi- 
ments. The latter, however, were not intended to give the conditions 
of maximum yield and were designed simply for purposes of comparison. 

Fischer by heating one part glucose with two parts phenylhydrazine 
chloride, three parts sodium acetate and 20 parts of water for 1J hours 
upon the water bath obtained 85 to 90 per cent of the weight of sugar as 
osazone. This is nearly three times the amount obtained by Maquenne, 
but is still less than 50 per cent of the theoretical yield. 

Mulliken f has based a scheme for the identification of pure sugars 
upon the time of separation of the osazones. Fischer's method of mak- 
ing the test is followed, 0.1 gm. sugar, 0.2 gm. pure phenylhydrazine 
chloride, 0.3 gm. sodium acetate and 2 c.c. water being mixed in a 

* Maquenne's "Les Sucres," p. 266; Compt. rend., 112, 799. 
t Mulliken's "Identification of pure Organic Compounds." 



METHODS FOR THE IDENTIFICATION OF SUGARS 351 

small test tube, corked loosely to prevent evaporation and heated in 
boiling water. The tube is shaken occasionally without removing from 
the bath and the time noted for the separation of a precipitate. Under 
the above conditions Mulliken noted the following: 



Sugar. 


Time for 
osazone 
separation. 


Sugar. 


Time for osazone separation. 


Fructose 
Sorbose 


Minutes. 

2 

01 


Arabinose 
Galactose 


Minutes. 
10 
1 1; iq 


Glucose 


2 
4K 


Sucrose - 


30 (due to sliffht inversion) 


Xylose . . 


7 


Maltose 




Rhamnose 




Lactose 













The relation of the sugars as regards time of osazone formation agrees 
closely with that noted by Maquenne. 

Sherman and Williams * give the following time of osazone forma- 
tion for different quantities of sugar under the conditions followed by 
Mulliken, but with double the quantity of reagents and water. 

Time for Precipitation of Osazones 



Weight of sugar 
taken. 


Glucose. 


Fructose. 


Invert sugar. 


Sucrose. 


Gram. 


Minutes. 


Minutes. 


Minutes. 


Minutes. 


0.2 


4-5 


H-H 


1MI 


31 


0.1 


5 


lf-2 


2 


35 


0.05 


6 


2 


3 


78 


0.01 


17 


5| 


6-6| 


No ppt. 


0.005 


34 


10 


14 




0.0025 


65 


17 







Sherman and Williams found that with mixtures of different sugars 
the time of osazone formation was greatly modified. The following re- 
sults were noted. 



Influence of Maltose on Glucose 



Weight of 
glucose. 


Weight of maltose. 


In absence of 
maltose. 


0.2 gram. 


0.1 gram. 


0.05 gram. 


0.01 gram. 


Gram. 
0.01 
0.02 


Minutes. 
No. ppt. 
26-28 


Minutes. 
40 


Minutes. 
30 


Minutes. 
22 


Minutes. 
17 . 
12-13 



* J. Am. Chem. Soc., 28, 629. 



352 



SUGAR ANALYSIS 



Influence of Lactose on Glucose 



Weight of 
glucose. 


Weight of lactose. 


In absence of 
lactose. 


0.2 gram. 


0.1 gram. 


0.05 gram. 


0.01 gram. 


Gram. 

0.01 
0.02 


Minutes. 

No ppt. 
45-48 


Minutes. 

50 


Minutes. 

32 


Minutes. 

25 


Minutes. 
17 
12-13 



Influence of Sucrose on Glucose 





Weight of sucrose. 




Weight of 




In absence of 


glucose. 










sucrose. 




0.2 gram. 


0.1 gram. 


0.05 gram. 


0.01 gram. 




Gram. 
0.005 


Minutes. 
15-17 


Minutes. 

15-17 


Minutes. 
22 


Minutes. 
30 


Minutes. 

33-39 


0.01 


14-16 


16 


17 


17 


17 


0.2 


9 








12-13 



Influence of Raffinose on Glucose 



Weight of 
glucose. 


Weight of raffinose. 


In absence of 
raffinose. 


0.2 gram. 


0.1 gram. 


0.05 gram. 


0.01 gram. 


Gram. 

0.005 


Minutes. 

27-30 


Minutes. 

33-37 


Minutes. 

36-38 


Minutes. 

37-39 


Minutes. 

33-39 



Influence of Maltose on Fructose 



Weight of 
fructose. 


Weight of maltose. 


In absence of 
maltose. 


0.2 gram. 


0.1 gram. 


0.05 gram. 


0.01 gram. 


Gram. 
0.01 


Minutes. 

7-8 


Minutes. 
5-6 


Minutes. 

5H* 


Minutes. 

5^ 


Minutes. 

5 



Influence of Lactose on Fructose 



Weight of 
fructose. 


Weight of lactose. 


In absence of 
lactose. 


0.2 gram. 


0.1 gram. 


0.05 gram. 


0.01 gram. 


Gram. 

0.01 


Minutes. 
9i-10 


Minutes. 

7| 


Minutes. 

ef 


Minutes. 
6 


Minutes. 

5| 



METHODS FOR THE IDENTIFICATION OF SUGARS 353 



Influence of Sucrose on Fructose 



Weight of 
fructose. 


Weight of sucrose. 


In absence of 
sucrose. 


0.2 gram. 


0.1 gram. 


0.05 gram. 


0.01 gram. 


Gram. 

0.005 


Minutes. 

8^ 


Minutes. 

8f 


Minutes. 

9* 


Minutes. 

9i 


Minutes. 

9 



The results show that sucrose accelerates, while maltose and lactose 
retard the separation of osazone from solutions containing glucose and 
fructose. 

A scheme of identification, based upon yield, or time of formation of 
osazone under a prescribed method of treatment, is of value only in 
working with a known quantity of pure sugar. In case of products 
containing foreign organic and mineral matter, or a mixture of several 
sugars, the presence of impurities or of other osazones influences crys- 
tallization to a very marked degree. This fact prevents the employ- 
ment of .the osazone reaction for exact quantitative purposes. 

The osazones of sugars after precipitation require to be purified. 
The crystalline precipitate is filtered off, well washed with cold water, 
and then pressed as dry as possible between filter paper. The product 
is then recrystallized from boiling 50 per cent alcohol to which a few 
drops of pyridine may be added, in case of very insoluble osazones, to 
promote solubility. Recrystallization may also be effected from ace- 
tone and other organic solvents and in case of easily soluble osazones, 
as of maltose and lactose, from hot water. After dissolving the osa- 
zones, the hot solution is filtered and set aside in the cold until crystalli- 
zation is complete. The purified osazone is then filtered off and dried 
at a gentle heat. A melting point is then taken which, if the osazone 
is pure, will remain unchanged after further crystallization. A table of 
melting points is then consulted and this in many cases is sufficient to 
identify the osazone. (See Table 24, Appendix.) 

Limitations of the Osazone Reaction. The osazone reaction with 
phenylhydrazine, while invaluable, is not always an absolute test of the 
identity of a sugar, owing to the fact that a number of isomeric sugars 
give the same osazone. The pentose sugars d-lyxose and 1-xylose, for 
example, yield the same phenylosazone of melting point 160-161 C. 
Similarly the hexose sugars, d-glucose, d-mannose and d-fructose, yield 
the same phenylosazone of melting point 206 C. In fact any of the 
isomeric sugars which are mutually transformable (as in contact with 



HO-C-H 
H-C-OH 
H-C-OH 



This circumstance, although nullifying the use of phenylosazones in 
certain cases as a means of identification, has yet thrown a flood of 
light upon the molecular constitution of sugars. 

Test for Ketoses with Methylphenylhydrazine. In distinction from 
phenylhydrazine the substituted hydrazines do not always give the 
same osazone reaction with sugars which are mutually transformed. 
The osazone reaction with substituted hydrazines has, therefore, a dis- 
tinct qualitative value. Methylphenylhydrazine, for example, forms 
very readily a characteristic osazone with d-fructose, but does not form 
an osazone with d-glucose or d-mannose or any of the other aldose 
sugars. The osazone reaction with methylphenylhydrazine is, there- 
fore, serviceable in distinguishing aldoses from ketoses. 

Decomposition of Osazones into Osones. While hydrazones, upon 
decomposition with strong hydrochloric acid or with benzaldehyde or 
formaldehyde, yield the component sugar, the osazones cannot be re- 
solved in this manner. The osazone reaction is consequently of value 



354 SUGAR ANALYSIS 

alkalies) give the same osazone. This is made more clear from the fol- 
lowing stereoformulse of glucose, mannose and fructose. 

H-C=0 H-C = CH 2 OH 

H-C-OH HO-C-H C=O 

I I ! 

HO-C-H'" "HO-C-H" HO-C-H 

H-C-OH H-C-OH H-C-OH 

H-C-OH H-C-OH H-C-OH 

CH 2 OH CH 2 OH CH 2 OH 

d-Glucose d-Mannose d-Fructose 

The part of the molecule below the dotted line has the same spatial 
arrangement in all three sugars. The part of the molecule above the 
dotted line is the only part of the molecule affected in the osazone re- 
action, this in all three sugars giving rise to an osazone which has the 
same structural formula: 

H-C=N-NH-C 6 H 5 



METHODS FOR THE IDENTIFICATION OF SUGARS 355 

only as a means of identifying and not of separating sugars. The de- 
composition of osazones with acids and aldehydes has, however, a con- 
siderable theoretical interest which may be considered briefly in this 
connection. 

Treatment of osazones with concentrated hydrochloric acid or with 
certain aldehydes causes, as in the case of hydrazones, a separation of 
the phenylhydrazine ; the product remaining behind, however, is not the 
original sugar, but a compound with two adjacent carbonyl groups called 
an osone. The reaction of glucosazone with hydrochloric acid, for ex- 
ample, is: 

CH 2 OH CH 2 OH 

(CHOH) 3 + 2 HC1 + 2 H 2 O = (CHOH) 3 + 2 CeHsNH - NH 2 HC1 

C : N - NHC 6 H 5 C : O 



HC : N - 



NHC 6 H 5 HC : O 

Glucosazone Hydrochloric acid. Glucosone Phenylhydrazine chloride. 

In case of osazones soluble in hot water the conversion into osones 
can be easily effected with benzaldehyde in presence of sufficient al- 
cohol, the phenylhydrazine being separated as benzaldehyde-phenyl- 
hydrazone and the osone remaining behind in solution. 

Osones upon treatment with zinc dust and acetic acid are reduced 
by the nascent hydrogen to a sugar, the end carbonyl group being con- 
verted always to an alcohol group, as shown in the following equation 
for glucosone. 

CH 2 OH CH 2 OH 

(CHOH) 3 + H 2 (CHOH) 3 

C:O C:0 

HC : O CH 2 OH 

Glucosone Hydrogen Fructose 

It will be seen from the above reaction that the sugar obtained by re- 
duction of an osone is always a ketose. By this means glucose and 
mannose can be transformed into fructose, and this type of reaction is 
true for the conversion of any aldose into the corresponding ketose, the 
steps of the transformation being always 

Aldose Osazo'ne > Osone > Ketose. 

The osones, while of great service in establishing the relationship of 
different sugars to one another, have no value either in qualitative or 
quantitative sugar analysis. 



356 



SUGAR ANALYSIS 




THE IDENTIFICATION OF HYDRAZONES AND OSAZONES 

The identification of hydrazones and osazones, by examination of 
their physical properties, although belonging strictly to the tests for 
individual sugars, is introduced for convenience at this point. 

Determination of Melting Point of Hydrazones and Osazones. 
The determination of melting point is the principal physical method 
for identification of hydrazones and 
osazones. 

Capillary-tube Method. The capil- 
lary-tube method is the one most gener- 
ally employed for determining melting 
points. The essential requirements in 
way of apparatus are shown in Fig. 162. 
A long-neck flask with a small body 
of about 20-c.c. capacity is filled about 
two-thirds with pure concentrated sul- 
phuric acid; to prevent discoloration of 
the acid through accidental contamina- 
tion with organic matter a small crystal 
of potassium nitrate, the size of a pin- 
Fig. 162. Ap- head, is dropped in. The flask is clamped 
paratus for de- ^ o a lamp-stand in the manner shown. 
The opening of the flask is fitted with a 
perforated cork containing a groove upon 
the side to allow escape of expanding air. The perfora- 
tion in the cork should be of such a size as to hold a 
thermometer, graduated to 300 C., tightly in position; 
the bulb of the latter should be above the bottom of 
the flask and yet be submerged entirely in the acid. 

The capillary tubes for holding the hydrazone or 
osazone are best prepared by thoroughly softening a 
piece of glass tubing by turning it in the flame and then 
drawing it out to about 1 to 1.5-mm. diameter. By 
continuing this process backwards along the tube a Fi g- 163 - Show- 
number of sections are obtained similar to Fig. 163a; ing 
the sections are then filed off at the points indicated 
and the smaller ends melted together in the flame, melting points. 
Small tubes of the size and shape shown in Fig. 1636 are thus obtained. 
A small amount of finely powdered hydrazone or osazone is then in- 
troduced into the open end of the tube and the latter gently tapped until 



METHODS FOR THE IDENTIFICATION OF SUGARS 357 

the substance has settled to the bottom. To prevent the powdered 
material from forming too loose a layer it is usually well to push it 
tightly down by means of a platinum-wire or thin-glass rod. The depth 
of substance in the tube should not exceed 2 mm. The capillary tube 
containing the substance is then attached to the thermometer either by 
binding it with a piece of fine platinum wire or by dipping it first in con- 
centrated sulphuric acid and allowing it to stick to the thermometer 
bulb by adhesion. The tube is placed so that the layer of substance is 
even with the center of the mercury bulb. 

After placing the thermometer and tube in position, as shown in Fig. 
162, a small flame is placed beneath the flask and the temperature raised 
until the liquefaction of the powdered crystals indicates the tempera- 
ture of melting. Hydrazones and osazones at the point of melting de- 
compose with darkening of color, the evolution of gas causing the 
liquefied substance to foam upwards in the stem of the tube. The first 
determination of melting point is only preliminary and a second and 
third trial should always be made with fresh tubes and material. The 
acid in the subsequent tests is heated rapidly to about 5 C. below 
the melting point first observed and then the temperature raised 
gradually so that the thread of mercury in the thermometer comes to 
rest just at the point of liquefaction. The entire operation for glucosa^ 
zone, for example, melting at 204 to 205 C., should not consume over 
4 minutes. Undue protraction of the time of heating affects the result 
of the determination very markedly and the wide discrepancies noted 
in the literature between melting-point determinations of the same 
osazone by different authorities are due largely to this cause. 

Maquenne's Block. A second method for determining melting 
points of hydrazones and osazones is employed considerably by French 
chemists. This method involves the use of the Maquenne Block, an 
apparatus invented by Maquenne in 1887, the essential features of 
which are shown in Fig. 164. 

The important part of Maquenne's apparatus consists of a prismatic 
block (A) of brass, weighing about 2 kilos, which is placed in a frame 
with one of its edges resting above the openings of a long gas burner (B). 
In one end of the block about 5 mm. below the upper surface a hole is 
bored, extending nearly the length of the block, into which a thermom- 
eter (T) can be inserted. In the upper level surface of the block are a 
number of small, round cavities. In conducting a determination a small 
amount of substance is placed in one of the cavities, which, to prevent 
disturbances from air drafts, is covered with a small glass; the thermom- 
eter is then inserted so that its bulb is about underneath the cavity and 



358 SUGAR ANALYSIS 

the burner started with a low, uniform flame. The temperature is 
slowly elevated until the substance begins to melt when the thermometer 
is drawn out or pushed in until just the end of the mercury thread pro- 
jects and the temperature noted. The block is now cooled slightly and 
a second determination made more slowly than before, using a cavity 
above the bulb of the thermometer in its second position. Owing to 
the fact that the block has nearly the same temperature, the entire 
column of mercury is brought to the same temperature as that of the 
melting substance and no correction due to contraction of the thread 
outside the unheated portion of the thermometer is necessary as by the 
method of melting-point determination previously described. 




Fig. 164. Maquenne's block for determining melting points. 

A comparison of melting points of glucose-phenylosazone by the 
two methods shows the following: capillary tube 205 C. (Fischer), 
Maquenne Block 230 to 232 C. (Bertrand). From this it would 
appear that the Maquenne Block gives considerably higher melting 
points than the capillary-tube method. A critical comparison of the 
two methods by Miither * (see Table LXVI, opposite page) shows, how- 
ever, that this is not always the case. 

It will be seen that Miither obtained for glucosazone results by the 
block agreeing very closely with those by the tube, the range found by 
the block being 200 to 206 C. and by the tube 203.5 to 205 C. 
The greater variation by the block is attributed by Miither to the 
unequal distribution of heat through the brass, the outer surface being 
more quickly warmed than the center; differences from 3 to 6 C. 
were also noted for different positions of the thermometer inside the 
block. The slowness with which the block is heated and cooled and 
the difficulty with which the cavities are cleaned are also serious objec- 
tions. With substances which sublime, the Maquenne Block cannot 
* Dissertation, Gottingen, 1903. 



METHODS FOR THE IDENTIFICATION OF SUGARS 359 



be used on account of the rapid condensation of material from the 
cavity upon the cover glass. These objections together with the high 
cost of the apparatus (about $15.00, duty free) render it much less 
desirable for determining melting points than the simpler capillary-tube 
method. 

TABLE LXVI 

Showing Melting Points of Hydrazones and Osazones by Different Methods. (Miither.) 



Compound. 


Method of melting point. 


Capillary tube. 


Maquenne block. 


Arabinose-methylphenylhydrazone 


Deg. C. 
164 ' 

203-204 
177 

172-173 

188-189 
188-189 
203.5 
204-205 
203.5 
203-204 
203.5 


Deg. C. 
158-160 
159-160 
159-160 
162 
198 
199-200 
174-175 
172-173 
170-171 
165-167 
173-174 
187 
191-192 
202-203 
* 200-201 
204-205 
205-206 
205 


Arabinose-diphenylhydrazone 


Fucose-methylphenylhydrazone. 


Fucose-benzylphenylhydrazone . . . 


Mannose-phenylhydrazone. 


Fructose-osazone (glucosazone) 





Isomerism and Variability in Melting Points of Hydrazones. A 
peculiarity of a number of hydrazones is the existence of two isomers of 
different crystalline form, melting point and specific rotation. Thus 
in case of d-glucose-phenylhydrazone the following properties were 
noted by Fischer and Tafel,* and by Simon and Benard.f 





I. 


II. 


Crystalline form 


Fine needles. 


Long needles. 


Melting point 


144M46 


115-116^ 


Specific rotation after solution 


-66.57 


-15.3 


Specific rotation after standing 


-52.00 


-52.9 



It is seen that the isomeric hydrazones each possess mutarotation, 
and in solution undergo transformation into the same compound. 

* Ber., 20, 2566. 

t Compt. rend., 132, 564. 



360 SUGAR ANALYSIS 

The isomerism has been attributed to the existence of hydrazones of a- 
and jS-glucose, but the conditions for their separate formation have not 
been definitely established. 

Similar differences have been noted in the case of other hydrazones, 
but whether the variation in properties is due to isomerism or to a 
difference in purity is not always certain. 

Optical Activity of Hydrazones and Osazones as a Means of Iden- 
tification. In addition to melting point the optical activity of hydra- 
zones and osazones is sometimes employed as a means of identification. 

Owing to the low solubility of some of the compounds and the high 
color of some of the solutions the polarization of hydrazones and osa- 
zones can not always be measured with exactness. In the case of 
hydrazones the existence of different isomers, as in the case of glucose- 
phenylhydrazone just cited, may cause wide differences in polarization. 
Mutarotation, which was noted in the case of glucose-phenylhydrazone, 
has also been observed with some of the osazones. Thus Allen and 
Tollens* found for 1-arabinose-phenylosazone [a]o =+18.9 after dis- 
solving in alcohol, but after standing a short time the solution became 
optically inactive. 

The rotatory power of hydrazones and osazones also varies greatly 
for different solvents. Thus Lobry de Bruyn and van Ekenstein f 
found the following rotations for different 0-naphthylhydrazones in 
methyl alcohol and glacial acetic acid. 





Methyl alcohol. 


Glacial acetic acid. 


Rhamnose-/3-naphthylhydrazone . . 


+ 8.4 


-11.8 


Glucose-/8-naphthylhydrazone. 


+40.2 





Mannose-/3-naphthylhydrazone 
Galactose-/8-naphthylhydrazone 


+ 16.8 
+24.8 



+ 2 



For purposes of comparison and identification the rotations of 
hydrazones and osazones must be measured, therefore, under exactly 
similar conditions as to quantity of material and nature of solvent. 
Neubergt recommends dissolving 0.2 gm. of osazone in a mixture of 
4 gms. pyridine and 6 gms. absolute alcohol, and reading the solution 
in a 200-mm. tube in a polarimeter. The following rotations were 
obtained by Neuberg for different osazones when working under the 
above conditions: 

* Z. Ver. Deut. Zuckerind., 40, 1033. 
t Rec. Trav. Pays Bas, 16, 226. 
t Ber., 32, 3384. 



METHODS FOR THE IDENTIFICATION OF SUGARS 361 



TABLE LXVII 
Giving Polarization of Different Osazones 



1-Arabinose-phenylosazone 

1-Arabinose-p-bromophenylosazone 

Xylose-phenylosazone 

Xylose-p-bromophenylosazone 

Rhamnose-phenylosazone 

d-Glucose-phenylosazone 

d-Glucose-p-bromophenylosazone. 

d-Galactose-phenylosazone 

Sorbose-phenylosazone 

Maltose-phenylosazone 

Lactose-phenylosazone 



+028' 
-015' 




+048' 
-015' 
+130' 




The rotations are small and in some cases uncertain so that this 
method of identification upon the whole is less satisfactory than a melt- 
ing-point determination. 

In case of the hydrazones and osazones of optically opposite isomeric 
sugars (which, as regards melting point and solubility, behave alike ex- 
cept in the special case where optically active hydrazines are used), a 
determination of the optical activity of the compound is the only ready 
means of identification. Thus Fischer * gives for the phenylhydra- 
zones of d- and 1-galactose the following constants. 

Melting point. [a]^ 

d-Galactose-phenylhydrazone 158 21.6 

1-Galactose-phenylhydrazone 158 +21.6 

Fischer also gives for the phenylhydrazones of d- and 1-mannose 

M 'Sf Station. 

d-Mannose-phenylhydrazone 195 - 1.2 

1-Mannose-phenylhydrazone 195 +1.2 

The rotations in the latter case were the angular readings obtained 
in a 100-mm. tube upon a solution of 0.1 gm. hydrazone in 1 c.c. cold 
concentrated hydrochloric acid and diluted with 5 c.c. of water. 

Employment of Optically Active Hydrazines for Separating 
Sugars from Racemic Mixtures. Neuberg f has recently employed 
optically active hydrazines for analyzing racemic mixtures of sugars. 

If two optically opposite isomeric sugars (" antipodes ") + S and S 
form hydrazones with an optically inactive hydrazine H, the result- 
ing compounds, which may be represented by the symbols +SH and 
SH are also antipodes, and, although of exactly opposite rotations, 

* Fischer's " Untersuchungen uber Kohlenhydrate." 
t Ber., 36, 1192; 38, 866, 868. 



362 SUGAR ANALYSIS 

have in other respects, such as specific gravity, melting point, solubility, 
etc., the same physical properties. A separation of two such hydra- 
zones is consequently not possible by the ordinary methods of analysis. 

If, however, the two sugars +S and S combine with an optically 
active hydrazine as +H, the resulting hydrazones + S + H and 
S + H are not optical antipodes and show well-defined differences 
in solubility, melting point and other properties. A separation of the 
two hydrazones is thus made possible by the ordinary methods of 
fractional crystallization. 

The hydrazines, which have been used by Neuberg and his co- 
workers for this method of separating sugars, are 1-menthylhydrazine 
and d-amylphenylhydrazine, the structural formulae of which are as 

follows : 

CH 3 CH 3 



i 



n )cH-CH 2 



CH 2 CH-NH-NH 2 

\ / 

CH 

CH 3 -CH-CH 3 

l-Menthylhydrazine. d-Amylphenylhydrazine 

The method has been employed successfully by Neuberg in resolv- 
ing the racemic sugar d,l-arabinose, which occurs in the urine of many 
persons suffering from pentosuria; d,l-arabinose gives with 1-menthyl- 
hydrazine an easily soluble 1-arabinose-l-menthylhydrazone and a very 
insoluble d-arabinose-1-menthylhydrazone. The latter is filtered off and 
upon treatment with formaldehyde (p. 348) is easily decomposed with 
liberation of the free sugar d-arabinose. 

IV. MISCELLANEOUS REACTIONS OF SUGARS 

Reactions of Sugars with Reducing Agents. The simple reducing 

sugars, in their character of aldehydes or ketones, are easily transformed 

by reducing agents into the corresponding alcohols. The sugar man- 

nose, for example, is reduced by sodium amalgam to the alcohol mannite. 

CH 2 OH CH 2 OH 

(CHOH) 4 + H 2 (CHOH) 4 

CHO CH 2 OH 

Mannose Hydrogen Mannite 

A more general type of equation would be: 

CnH 2n O n + H 2 C n H 2n+2 O n 

Sugar Hydrogen Sugar alcohol 



METHODS FOR THE IDENTIFICATION OF SUGARS 363 



The reactions of the different sugars with reducing agents are of 
comparatively minor importance as regards use in sugar analysis. 

A description of the different sugar alcohols, with reactions and 
methods of identification, is given in Chapter XXIII. 

Reactions of Sugars with Weak Oxidizing Agents. Reducing 
sugars belonging to the aldoses are changed by means of the less power- 
ful oxidizing agents, such as bromine water, into the corresponding 
monobasic acids. Thus: 



CH 2 OH 

I 
(CHOH)< 



4- 2Br 



H 2 



CH 2 OH 
(CHOH) 4 



Aldo-hexose 



Bromine water 



Hexonic acid 



2HBr 



Hydrobromic acid 



In carrying out the reaction 1 part sugar is treated with 5 parts 
of water and 2 parts of bromine, and the solution kept at room tempera- 
ture for 1 to 3 days. 

Ketose sugars, upon treatment with bromine water, undergo but 
little oxidation during the first few days. Prolonged action, or eleva- 
tion of temperature, will, however, oxidize ketoses with a breaking up 
of the molecule into several acids of fewer carbon atoms. 

Rate of Oxidation with Bromine as a Test for Aldoses and Ketoses. 
The rate of oxidation of several aldose sugars with bromine water, as 
compared with fructose, is shown in the following experiments by 
Votocek and Nemecek;* 0.5 gm. of pure sugar was dissolved in a 50-c.c. 
flask in 9 c.c. of water, 40 c.c. of bromine water (saturated at room 
temperature) were then added and the volume made up to 50 c.c. After 
standing at room temperature (21 C.) for 24 hours, the unoxidized 
sugar was determined in each flask with the following results: 



Sugar. 


Per cent sugar 
unoxidized. 


Sugar. 


Per cent sugar 
unoxidized. 


d-Galactose .. 


5 10 


1-Xylose ..,,.. vv 


25.68 


1-Arabinose . 


7 56 


Rhamnose 


39.19 


d-Glucose. . . 


22 20 


d-Fructose 


100.00 











Votocek and Nemecek propose their method as a means for dis- 
tinguishing aldoses from ketoses and also as a method for examining 
sugar mixtures. In case of the latter the aldoses are oxidized away with 
bromine water, leaving the ketoses in better condition for isolation. 

Reactions of Sugars with Strong Oxidizing Agents. Reducing 
sugars belonging to the normal unsubstituted aldoses are changed upon 
* Z. Zuckerind. Bohmen, 34, 399. 



364 SUGAR ANALYSIS 

warming with stronger oxidizing agents, as 30 per cent nitric acid, into 
the corresponding dibasic acids. Thus 

CH 2 OH COOH 

(CHOH) 4 + 2HN0 3 = (CHOH) 4 + 2H 2 0+2NO 
CHO COOH 

Galactose Nitric acid Mucic acid 

In carrying out the reaction one part of sugar is heated with 2| parts 
nitric acid of 1.2 sp. gr. and gently warmed at 40 to 50 C. until no 
more nitrous fumes are evolved. The solution is then heated upon the 
water bath until all nitric acid is expelled and then evaporated, when 
the acid or its lactone will in many cases crystallize; when crystalliza- 
tion does not occur separation from impurities is effected by forming an 
insoluble salt or other derivative from which the acid can afterward be 
liberated in the pure condition. 

Ketose sugars, upon oxidation with nitric acid, are degraded into 
lower oxidation products, of which oxalic acid is usually formed in 
largest amount. 

The substituted aldose sugars, as the methyltetroses, methylpen- 
toses, methylhexoses, etc., lose the methyl group upon oxidation with 
nitric acid and are degraded into dibasic acids of one less carbon atom. 
CH 3 

CHOH COOH 

(CHOH) 2 + 5 O = (CHOH) 2 + HCOOH + H 2 O 
CHO COOH 

Methyltetrose Tartaric acid Formic acid Water 

In the same way the methylpentoses, rhamnose, rhodeose and 
fucose are oxidized into trioxyglutaric acids, the methylhexoses into 
tetraoxyadipic acids, etc. 

Oxime Reaction of Sugars. Many of the reducing sugars react 
with hydroxylamine, after the manner of all aldehydes and ketones, 
with formation of oximes. The following combination of glucose with 
hydroxylamine is an illustration of this type of reactions. 
CH 2 OH CH 2 OH 

(CHOH) 4 (CHOH) 4 +' H 2 

H-C:O + H 2 N-OH = H-C:N-OH 

Glucose Hydroxylamine Glucose-oxime Water 

The oximes of the sugars are often difficult to isolate and the reac- 
tion, for this reason, has but little value in sugar analysis. In sugar 
synthesis, however, the oxime reaction has considerable importance, for 
by its means a monosaccharide may be changed into another sugar con- 



METHODS FOR THE IDENTIFICATION OF SUGARS 365 

taining one less carbon atom. This is done by first making the oxime of 
the sugar and then heating the latter with acetic anhydride; the result- 
ing acetyl-nitrile derivative is then heated with an ammoniacal solu- 
tion of silver oxide which splits off the acetic acid and hydrocyanic 
acid groups with formation of a lower sugar (Wohl's* synthesis). The 
reaction in its simplest phase is represented as follows: 
CH 2 OH CH 2 OH CH 2 OH 

(CHOH) 4 (CHOH) 3 = (CHOH) 3 + HCN 

HC : NOH CHOH CHO 

i-N 

d-Glucose-oxime d-Gluconic acid nitrile d-Arabinose . Hydrocyanic acid 

The hexose sugar d-glucose is thus converted into the pentose sugar 
d-arabinose. In the same manner d-arabinose can be converted into 
the tetrose sugar d-erythrose. 

Cyanhydrine Reaction of Sugars. The reducing sugars, similar 
to all aldehydes and ketones, react with hydrocyanic acid forming a 
characteristic group of compounds known as cyanhydrines. 
CH 2 OH CH 2 OH 

(CHOH) 4 + HCN (CHOH) 4 

CHOH 



C : O 



d-Glucose Hydrocyanic acid d-Glucose-cyanhydrine (d-glucoheptonic acid nitrile) 

The Cyanhydrine reaction, as that of the oximes, while having but 
little value in sugar analysis, has very great importance in sugar synthesis 
for by its means a monosaccharide may be 'built up into another sugar 
having one more carbon atom. This is done by first making the 
cyanhydrine, saponifying this to form the corresponding acid, and then 
reducing the latter with sodium amalgam which produces the corres- 
ponding sugar. The formation of glucoheptose from glucose is given 
as an illustration of this type of reaction. 

CH 2 OH CH 2 OH 

(CHOH) 6 + 3H/) (CHOH) 6 + NH 4 OH 

C = N COOH 

d-Glucose-cyanhydrine d-Glucoheptonic acid Ammonia 

CH 2 OH CH 2 OH 

(CHOH) 5 + H 2 = (CHOH) 5 + H 2 O 

COOH HC : O 

d-Glucoheptonic acid (lactone) Hydrogen d-Glucoheptose Water 

* Ber., 26, 730. 



366 SUGAR ANALYSIS 

In the same manner, starting from the hexoses, mannose and galac- 
tose, mannoheptose and galaheptose can be derived. The heptoses by 
the same cyanhydrine synthesis have been built up into the correspond- 
ing octoses CsHieOg and the latter in turn into the corresponding nonoses 
CgHisOg. For details as to this method of forming sugars the work of 
Fischer * should be consulted. 

Ureide Reaction of Sugars. Nearly all reducing sugars, with ex- 
ception of the ketoses, react at moderately warm temperatures with 
urea in presence of dilute sulphuric or hydrochloric acid to form a 
group of compounds called ureides. The reaction is analogous to that 
with phenylhydrazine, the hydrogen of the amino group withdrawing 
the oxygen from the aldehyde group of the sugar. The reaction with 
glucose and urea is given by way of example. 

CH 2 OH CH 2 OH 

(CHOH) 4 (CHOH) 4 

HC:0 + H 2 N-CO-NH 2 = HC:N-CO-NH 2 + H 2 O 

Glucose Urea Glucose-ureide Water 

The ureides are partly crystalline and partly amorphous bodies. In 
aqueous solution they are decomposed upon heating with evolution of 
ammonia and liberation of the free sugar. 

Semicarbazone Reaction of Sugars. Very similar to the reaction 
of sugars with urea is that with semicarbazide; the latter in alcoholic 
solution combines with the aldoses to form a group of substances called 
semicarbazones. The reaction with glucose is given as illustration. 
CH 2 OH CH 2 OH 

(CHOH) 4 (CHOH) 4 

I H I H 

HC:O + H 2 N-N-CONH 2 = HC : N - N - CONH 2 + H 2 O 

Glucose Semicarbazide Glucose-semicarbazone Water 

The semicarbazones are well-defined crystalline compounds; when 
warmed with benzaldehyde in alcohol solution they are decomposed into 
free sugar with formation of benzaldehyde semicarbazone. 

Thiosemicarbazone Reaction of Sugars. Exactly similar to the 
previous reaction is the behavior of aldose sugars with thiosemicarbazide. 
The reaction with glucose proceeds as follows: 

CH 2 OH CH 2 OH 

( CHOH) 4 (CHOH) 4 

H H 

HC:O + H 2 N-N-CSNH 2 = HC:N-N-CSNH 2 + H 2 O 

Glucose Thiosemicarbazide Glucose-thiosemicarbazone Water 

* Ann., 270, 64; 288, 139. 



METHODS FOR THE IDENTIFICATION OF SUGARS 367 

The thiosemicarbazones are well-defined crystalline compounds simi- 
lar in many properties to the semicarbazones. 

Reactions of Sugars with Aromatic Amines. The ease with 
which reducing sugars unite with compounds containing an amino 
group, as shown in the case of the hydrazones, oximes, ureides, semi- 
carbazones, etc., is further exemplified by the reactions of sugars with 
different aromatic amines, such as aniline, toluidine, etc. Glucose, for 
example, reacts with aniline in alcoholic solution as follows: 

CH 2 OH CH 2 OH 

(HCOH) 4 (HCOH) 4 

H-C:0 + H 2 NC 6 H 5 = H-C:N-C 6 H 5 + H 2 

Glucose Aniline Glucose anilide Water 

Reactions of Sugars with Alcohols. By leading dry hydro- 
chloric-acid gas into the solution of a reducing sugar in an alcohol the 
corresponding alcohol derivative of the sugar is formed. The com- 
pounds thus prepared are called glucosides from their resemblance to 
the group of plant substances known under this name. The reaction 
of glucose with methyl alcohol is given as illustration. 
CH 2 OH CH 2 OH 

CHOH CHOH 
7 CHO FH"" 7 CH , 



/3 CHOH 



HOH 



H-C : !O + H 



/SCHOHJ) 



a. CHOH 



OCH 3 



= H-C=0~- 



-CH 3 + H 2 



Glucose Methyl alcohol Methyl glucoside Water 

In the same manner glucosides of the other sugars have been made 
as methyl arabinoside, methyl xyloside, methyl rhamnoside, methyl 
fructoside, also of the other alcohols as ethyl glucoside, etc. The com- 
pounds thus prepared are well-defined crystalline substances, easily 
soluble in water, do not reduce Fehling's solution and do not react with 
phenylhy drazine . 

The reactions of the reducing sugars with alcohols are but little 
used as a means of identification. The synthetic glucosides have, 
however, a great interest for the sugar chemist in other ways. 

Mercaptal Reaction of Sugars. Nearly all reducing sugars, ex- 
cept ketoses, react with the mercaptans in presence of concentrated 
hydrochloric acid to form mercaptals. The reaction with glucose and 
ethyl-mercaptan is given as illustration. 



368 SUGAR ANALYSIS 

CH 2 OH CH 2 OH 

(CHOH) 4 (CHOH) 4 

H-C:O 4- H-S-C 2 H 5 TT i/S-C 2 H s 

H - S - C 2 H 5 \ S - C 2 H 6 

Glucose Ethyl-mercaptan LGlucose-mercaptal Water 

The mercaptals of the sugars are well-defined crystalline compounds, 
soluble in hot water; they do not reduce Fehling's solution and do not 
react with phenylhydrazine. 

Reactions of Sugars with Aldehydes. The simple reducing 
sugars react with a large number of aldehydes (formaldehyde, acetal- 
dehyde, benzaldehyde, salicylaldehyde, furfural, etc.) to form a variety 
of condensation products. The latter, for the most part, are of a 
gummy or sirupy nature and do not crystallize readily. The combi- 
nation of glucose with acetaldehyde is given as an illustration of this 
type of reaction. 

CH 2 OH CH 2 OH 

(CHOH) 4 (CHOH) 4 



M | XK u 

H-C:O + O:C-CH 3 = H-C( > C - CH 3 



Glucose Acetaldehyde Glucose-acetaldehyde 

Reactions of Sugars with Polyvalent Phenols. The simple re- 
ducing sugars unite with different polyvalent phenols (resorcin, orcin, 
hydroquinone, phloroglucin, pyrogallol, etc.) to form a series of amor- 
phous ill-defined condensation products. The reaction is carried out in 
the cold in presence of hydrochloric acid. The following combination 
of arabinose with resorcin is given as an illustration of this type of 
reaction. 

CsHioOs -j- CeHeC^ CnHwOe ~h H^O. 

Arabinose Resorcin Arabinose-resorcin Water 

The condensation products of the sugars with polyvalent phenols 
when heated with concentrated hydrochloric acid are decomposed, show- 
ing the color and spectral reactions characteristic for each class of sugar 
(see p. 341). 

Reactions of Sugars with Acid Radicals. In the many different 
reactions previously described the aldehyde or ketone group of the 
sugar molecule is the one mostly involved. In the reactions of sugars 
with acid radicals, as acetic and benzoic, the alcohol groups of the mole- 
cule are affected; the aldehydic characteristics of the sugar are also 
usually modified in the higher derivatives. The number of acid de- 
rivatives obtainable with a sugar is dependent upon the number of 



METHODS FOR THE IDENTIFICATION OF SUGARS 369 

alcohol groups. In the case of hexoses having five such groups there 
are mono-, di-, tri-, tetra- and penta- acetates and benzoates; with 
sugars of fewer alcohol groups the number of these combinations is 
correspondingly less. 

Reaction of Sugars with Acetic Anhydride. Acetates of the sugars 
are formed by heating with acetic anhydride. A mixture of different 
acetates usually results during the reaction, the separation of these 
being effected by fractional crystallization or by the use of different 
solvents. To obtain the highest acetates, the reaction must be carried 
out in presence of zinc chloride or some other condensing agent. The 
formation of glucose pentacetate is given as illustration of this type of 
reaction : 

CH 2 OH CH 3 - CO CH 2 OCOCH 3 

1 \ I 

(CHOH) 4 +5 O (CHOCOCH 3 ) 4 + 5CH 3 COOH 

HC:O CH 3 -CO HC : O 

Glucose Acetic anhydride Glucose-pen tacetate Acetic acid 

The lower acetates of the sugars are amorphous, easily soluble sub- 
stances; the higher acetates are crystalline and less soluble in water. 
By warming with alcoholic potassium or sodium hydroxide, the acetates 
are all easily saponified with regeneration of the sugar. The lower ace- 
tates of the sugars are copper reducing and exhibit other aldehydic prop- 
erties; the higher acetates, as glucose-pentacetate, lack, however, many 
aldehyde characteristics, such as formation of hydrazones and oximes. 
This is probably due to a stable lactonic rearrangement of the molecule 
as shown by the following formula of Erwig and Konigs * for glucose 
pentacetate. 

,CHOCOCH 3 



/CHOCOC] 
CHOCOCH 3 



HOCOCH 3 



Reaction of Sugars with Benzoyl Chloride. The acetates of the 
sugars owing to their solubility are not well adapted for the identifi- 
cation of sugars; the sugar benzoates, however, are marked by a high 
insolubility in water and their formation is sometimes used as a quali- 
tative test for sugars. 

* Ber., 22, 1464, 2209. 



370 SUGAR ANALYSIS 

The test, according to the method of Baumann,* is carried out by 
treating a solution of the sugar with benzoyl chloride in presence of 
sodium hydroxide; the benzoic radical displaces the H of the hydroxyl 
groups with formation of sodium chloride and water. A number of 
benzoates are usually formed in the reaction. In the case of glucose- 
pentabenzoate the formation proceeds as follows: 

CH 2 OH CH 2 OCOC 6 H 5 

(CHOH) 4 + 5 C 6 H 5 COC1 + 5 NaOH = (CHOCOC 6 H 5 )4 + 5 NaCl + 5 H 2 O 

CHO CHO 

Glucose Benzoyl chloride Sodium hydroxide Glucose-pentabenzoate Salt Water 



The Baumann reaction is sufficiently delicate to detect 1 to 2 mgs. 
glucose in 100 c.c. of water and is sometimes employed for testing urine; 
100 c.c. of solution are well shaken with 2 c.c. of benzoyl chloride. 

SPECIAL TESTS FOR REDUCING SUGARS 

To the second class of reactions for examining sugars belong the 
special tests pertaining to group identification; the reactions chosen for 
description may be divided for convenience into three general classes. 
I. Analysis of hydrazones and osazones. 

II. Separation of products obtained by decomposition with concen- 
trated hydrochloric acid. 

III. Color reactions with phenols in presence of concentrated mineral 
acids. 

I. ANALYSIS OF HYDRAZONES AND OSAZONES AS A MEANS OF IDENTIFYING 

SUGAR GROUPS 

If the hydrazone or osazone of a sugar has been separated in a pure 
condition, an elementary analysis of the compound will serve to identify 
the group to which the sugar belongs. The osazones, owing to their 
greater insolubility and ease of preparation, are best adapted for this 
purpose. The determinations necessary for the identification of an 
osazone are those of the elements nitrogen and carbon; a determina- 
tion of hydrogen is also usually included since this element can be 
determined with little extra trouble at the same time as the carbon 
determination. 

The elementary analysis of osazones and hydrazones is carried out by 
burning about 0.2 gin. of the substance over cupric oxide in a com- 

* Ber., 19, 3220. 



METHODS FOR THE IDENTIFICATION OF SUGARS 371 



bustion tube. For nitrogen the combustion is carried out by Dumas 's 
method in a current of carbon dioxide after complete displacement of 
the air. The evolved nitrogen is received in a eudiometer over strong 
potassium hydroxide solution and its volume measured. From the vol- 
ume of gas the weight of nitrogen is calculated, making the necessary 
corrections for atmospheric pressure and temperature. 

For carbon and hydrogen the combustion is carried out by Liebig's 
method in a current of air or oxygen which must be perfectly dry and 
free from carbon dioxide. The evolved water is collected in weighed 
tubes, or spirals, containing concentrated sulphuric acid, and the evolved 
carbon dioxide absorbed in weighed Liebig bulbs containing concentrated 
potassium hydroxide solution, or in U-tubes filled with soda lime (NaOH 
+ CaO). From the weights of water and carbon dioxide obtained the 
percentages of carbon and hydrogen are calculated. The percentage of 
oxygen in osazones and hydrazones is determined by subtracting the 
sum of the percentages of the other elements from 100. 

In the elementary analysis of osazones and hydrazones, as of all 
other nitrogen compounds, a spiral of copper should be placed in the 
combustion tube at the exit end in order to effect the reduction of 
oxides of nitrogen. For complete details as to methods of combustion 
the chemist is referred to the standard textbooks upon organic analysis. 

Having determined the elementary composition of an osazone or 
hydrazone, reference to a table of percentage composition will usually 
locate the class of sugar to which the compound belongs. In the fol- 
lowing table the formula and percentage composition of phenylosazones 
are given for various groups of sugars. 



Phenylosazone. 


Formula. 


Composition. 


C 
per cent. 


H 

per cent. 


N 
per cent. 


O 

per cent. 


Diose 
Triose 


C 14 H 14 N4 
C 15 H 16 N 4 
C 16 H 18 N 4 2 
Ci 7 H 2 oN 4 3 
C 18 H 22 N 4 3 
Ci 8 H 22 N 4 4 
Ci 9 H 24 N 4 6 
C 20 H 26 N 4 6 
C 2 iH 28 N 4 7 
C 24 H 32 N 4 0, 


70.54 

67.12 
64.39 
62.16 
63.12 
60.30 
58.73 
57.38 
56.22 
55.35 


5.93 
6.01 
6.08 
6.14 
6.48 
6.19 
6.23 
6.27 
6.29 
6.20 


23.53 
20.90 
18.80 
17.08 
16.38 
15.64 
14.43 
13.40 
12.50 
10.77 


'5.'97 
10.73 
14.62 
14.02 
17.87 
20.61 
22.95 
24.99 
27.68 


Tetrose 


Pentose. 


Methylpentose 
Hexose 
Heptose 


Octose 


Nonose 


Disaccharide 



372 SUGAR ANALYSIS 

II. SEPARATION OF PRODUCTS OBTAINED BY DECOMPOSITION WITH CON- 
CENTRATED HYDROCHLORIC ACID AS A MEANS OF IDENTIFYING 
SUGAR GROUPS 

While an elementary analysis of osazones is one of the best means 
of determining the class to which a sugar belongs there are a number 
of other special group reactions which are of great value. The most 
important of these is the separation and identification of some char- 
acteristic decomposition product obtained by treating the sugar with 
concentrated sulphuric or hydrochloric acid. The latter acid is less 
drastic in its action and is the one most commonly used. 

The varied nature of the decomposition products humus sub- 
stances, aldehydes, acids, etc. obtained upon heating sugars with 
concentrated hydrochloric acid has already been mentioned. It is 
found, however, that when this treatment is carefully controlled some 
one characteristic decomposition product will predominate for each 
particular group of sugar. The following equations, representing ideal 
types of reaction, are given as illustrations : 

I. C 6 H 12 6 = C 5 H 8 3 + HCOOH + H 2 O 

Hexose Levulinic acid Formic acid Water 

II. C 5 H 10 5 = C 5 H 4 O 2 + 3H 2 O 

Pentose Furfural Water 

III. C 5 H 9 (CH 3 )0 5 = C 5 H 3 (CH 3 )0 2 + 3 H 2 O 

Methylpentose Methylfurfural Water 

The above types of reaction hold true not only of the simple sugars 
above named, but also of the higher saccharides which yield these 
sugars upon hydrolysis. In fact the initial phase of the reaction in 
case of the polysaccharides (sucrose, maltose, lactose, raffinose, starch, 
pentosans, methylpentosans, etc.), is purely hydrolytic, the simple 
sugars formed being subsequently decomposed after the manner just 
indicated. 

Levulinic Acid Reaction for Hexose Groups. This reaction, 
which is due to Tollens * and has been extensively studied by his co- 
workers, has been employed with great success in detecting hexose 
groups in a large variety of plant and animal substances (cellular tis- 
sues of plants, nucleic acids of animal origin, etc.) Owing to the much 
greater predominance of hexose-producing substances in nature the 
levulinic acid reaction is usually among the first tests applied in in- 
vestigating materials of unknown composition. 

Description of Test. In carrying out the reaction 5 to 10 gms. of 

* Ann., 206, 207, 226; 243, 314; Ber., 33, 1286. 



METHODS FOR THE IDENTIFICATION OF SUGARS 373 



material are treated with 20 to 50 c.c. of hydrochloric acid of 1.09 to 
1.10 sp. gr. (18 to 20 per cent) in a flask provided with a rubber stopper 
and condensing tube, and heated in a boiling-water bath for 5 to 20 
hours. The brownish-colored liquid is then cooled and filtered from the 
precipitate of humus substances; the filtrate is shaken out in a sep- 
aratory funnel four times with ether, and the ether extract, after pouring 
through a dry filter, evaporated. The sirupy residue is then gently heated 
in an open dish to expel the formic acid (see previous equation I). If 
levulinic acid is present a drop of the sirup dissolved in water in pres- 
ence of sodium carbonate and iodine will give a precipitate of iodoform, 
which can also be recognized by its characteristic odor. 

The main portion of the sirup is dissolved in water, boiled with an 
excess of zinc oxide (ZnO), and then, after decolorizing with animal 
charcoal, filtered and evaporated. The zinc salt of levulinic acid will 
soon crystallize; the crystals are filtered off, washed with absolute 
alcohol and ether, and then converted into the silver compound. This 
is done by dissolving the zinc salt in 5 to 10 c.c. of water, adding silver 
nitrate slightly in excess of the equivalent amount and heating nearly 
to boiling, with addition of a little water until the precipitated silver 
salt has completely dissolved. A little animal charcoal is then added 
and the solution filtered. The levulinate of silver, CsHyOsAg, which 
crystallizes will show under the microscope, in case the compound is 
pure, hexagonal crystals or plates; if the compound is less pure the 
crystals will be feather-like in appearance. The silver salt is filtered 
off, washed with cold water, pressed between filter paper and dried in a 
dark place over concentrated sulphuric acid. The per cent of silver in 
the salt is determined by strongly igniting a weighed portion in a por- 
celain crucible. The theoretical amount is 48.39 per cent Ag. 

The yield of levulinic acid obtained by treating hexose sugars with 
hydrochloric acid will vary greatly according to the time of heating and 
other conditions of the experiment. Conrad and Guthzeit * obtained 
upon heating 10.5 gms. each of fructose, glucose, and galactose, with 50 
c.c. of acid (containing 4.87 gms. HC1 gas) for 17 hours the following yield 
of products. 



Sugar. 


Humua. 


Levulinic acid. 


Formic acid. 


Fructose 
Glucose 


Grams. 
2.12 
1.00 

1.77 


Per cent. 
20.19 

9.52 
16.86 


Grams. 
4.09 
3.12 

2.85 


Per cent. 
38.95 

29.71 
27.14 


Grams. 
1.73 
1.35 
1.11 


Per cent. 
16.48 
12.86 

10.57 


Galactose .... 





* Ber., 19, 2575. 



374 SUGAR ANALYSIS 

From these results it appears that of the three hexose sugars fruc- 
tose gives the largest yield of levulinic acid and galactose the least. 
That this is due largely to the greater resistance of glucose and galac- 
tose toward the acid was shown by the fact that at the end of the above 
experiments considerable quantities of these sugars were still unde- 
composed (in case of glucose 26 per cent). The yield of levulinic acid 
is too variable for the method to be of any quantitative value. 

Furfural Reaction for Pentose Groups. This reaction, which is 
also due to Tollens,* has been of the greatest value not only as a means 
of detecting the presence of pentose carbohydrates but also as a means 
of their quantitative estimation. 

The reaction of the pentose sugars with hydrochloric acid proceeds 
much more nearly according to the equation (II, p. 372) than the 
reaction of the hexoses, the formation of humus substances being cor- 
respondingly less. The following graphic equation shows the decom- 
position of a pentose sugar into furfural. 



I CH:CH 

CH-CH-iOHi - v 

I xOjH ! in-p' + 3H2 

CH-C V , ........ CH ' C \C-0 

J ...... :l x c:o M 

JQH...H! * 

Pentose (150 parts) Furfural (96 parts) Water (54 parts) 

The theoretical yield of furfural, according to the above equation, 
is 64 per cent; actual determinations of the furfural, obtained by dis- 
tilling weighed amounts of the pentose sugars, arabinose and xylose, 
with hydrochloric acid, give about 47 per cent in case of arabinose and 
about 57 per cent in case of xylose yields which are about 75 per cent 
and 90 per cent respectively of the theoretical. 

Description of Test. In carrying out the qualitative test about 
5 gms. of substance are heated in a distillation flask with 100 c.c. of 
hydrochloric acid of 1.06 sp. gr. and successive portions of about 30 c.c. 
distilled into a receiver, new portions of acid being added to the flask for 
each quantity distilled. The distillates are then tested for the presence 
of furfural; the latter in large amounts can usually be detected by its 
pleasant aromatic odor somewhat resembling that of bitter almond oil. 
The presence of very small amounts of furfural is best indicated by 
Schiff's reaction with aniline or xylidine acetate. Aniline acetate re- 
agent is best prepared according to Tollens by mixing in a test tube 
equal volumes of aniline and water and then adding with constant shak- 
* Landw. Vers.-Stat., 39, 425. 



METHODS FOR THE IDENTIFICATION OF SUGARS 375 

ing glacial acetic acid drop by drop until the milky solution becomes 
clear. Test paper is prepared by moistening strips of filter paper with 
the aniline-acetate solution. Application of a drop of distillate con- 
taining furfural, even in minute traces, will cause the aniline-acetate 
paper to turn a bright cherry red. 

The presence of furfural in the distillate may also be indicated by 
first neutralizing the acid solution with sodium carbonate and then add- 
ing a solution of phenylhydrazine acetate and stirring. Furfural if pres- 
ent is precipitated as furfural-phenylhydrazone, C 4 H 3 OCHN 2 HC6H5, 
which melts at 97 to 98 C. 

A better precipitating agent for furfural than phenylhydrazine is 
phloroglucin. A solution of this compound in hydrochloric acid when 
added to a distillate containing furfural will cause an immediate dark- 
ening of the solution with final precipitation of furfural-phloroglucide, 
according to equation: 

C 5 H 4 2 + C 6 H 6 03 = CnH 8 O 4 + H 2 O 

Furfural Phloroglucin Furfural-phloroglucide Water 

Limitations of Furfural Reaction for Pentoses. While all carbohy- 
drates containing a pentose group yield large amounts of furfural upon 
distillation with hydrochloric acid, it must also be borne in mind that 
other substances have the same property. All hexose carbohydrates such 
as starch, cellulose, sucrose, glucose, etc., give small amounts of furfural 
upon distillation with hydrochloric acid but the yield is too small to in- 
terfere seriously with the test for pentoses. Two substances, however, 
of a non-pentose nature are especially marked by their property of 
yielding furfural upon distilling with acids and hence require brief 
mention. These are glucuronic acid and oxycellulose. 

Glucuronic acid is an aldehyde-acid derivative of glucose and has 
the formula COH(CHOH) 4 COOH. By the action of putrefactive 
bacteria it is converted into the pentose sugar 1-xylose. 

C 6 H 10 O 7 C 5 H 10 5 + C0 2 

Glucuronic acid Xylose Carbon dioxide 

The intimate relationship of glucuronic acid to the pentoses is also 
shown by the reaction upon distilling with hydrochloric acid. 

C 6 H 10 O 7 = C 5 H 4 2 + 3H 2 O + C0 2 

Glucuronic acid Furfural Water Carbon dioxide 

Glucuronic acid is sometimes found in the urine, especially after the 
ingestion of chloral, menthol, camphor, turpentine, acetanilide, alka- 
loids and many other compounds. Under such conditions a combina- 
tion takes place in the animal organism between the ingested compound 



376 



SUGAR ANALYSIS 



and the glucuronic acid, the latter apparently being formed as an oxi- 
dation product of glycogen. The glucuronic-acid derivative, which is 
excreted in the urine, may be mistaken for a pentose sugar if the chemist 
relies solely upon such tests as the furfural reaction and reduction of 
metallic salt solutions. 

One means of determining the presence of glucuronic acid is by 
means of p-bromophenylhydrazine, which was found by Neuberg * to 
give a characteristic glucuronic-acid derivative, C^HnOr^Br. The 
exact nature of the compound, whether hydrazone or hydrazide, was 
not determined. The solution to be tested is heated in a water bath at 
60 C. with 5 gms. of p-bromophenylhydrazine chloride and 6 gms. of 
sodium acetate. If glucuronic acid is present yellowish needle-like crys- 
tals will separate in 5 to 10 minutes. The solution is cooled, the crystals 
filtered off and the filtrate again heated as before; a second crop of 
crystals may thus be obtained which are filtered off again and the 
process continued until no more crystals form. The combined precipi- 
tates are thoroughly washed with warm water and then with absolute 
alcohol. Recrystallized from 60 per cent alcohol the crystals melt at 
236 C. The crystals dissolved in a mixture of 6 c.c. absolute alcohol 
and 4 c.c. pyridine have a strong levorotation, [O\D = 369. 

Spectroscopic methods for distinguishing between pentoses and glu- 
curonic acid will be described under the color reactions for sugar groups. 

Cellulose, when treated with different oxidizing agents, such as nitric 
acid, chromic acid, hypochlorous acid and permanganate, undergoes a 
partial oxidation. The oxycellulose derivatives formed under such 
conditions have the property of yielding furfural upon distillation with 
hydrochloric acid. 

According to the researches of Tollens and Faber f oxycelluloses 
consist of mixtures of cellulose (CeHioOs^ in different porportions with 
an oxy-derivative celloxin (C 6 H 8 6 )n. The greater the amount of cel- 
loxin in the oxycellulose the greater the yield of furfural upon distilla- 
tion with hydrochloric acid. Cotton, for example, upon treatment with 
nitric acid at 100 C. for different periods of time, gave the following 
results : 



Time of treatment. 


Composition. 


Yield of furfural. 


Hours. 

2* 

4 


4 C 6 HioO 5 , C 6 H 8 O 6 
3 CeHioOs, C 6 H 8 O 6 


Per cent. 
2.3 
3.2 



* Ber., 32, 2395. 
t Ber., 32, 2589. 



METHODS FOR THE IDENTIFICATION OF SUGARS 377 

The yield of furfural calculated to pure celloxin (which has not as 
yet been isolated) is about 12 per cent. 

The oxycelluloses are widely distributed in nature and if reliance 
is based exclusively upon the furfural reaction erroneous conclusions 
may be formed as to the occurrence of pentose carbohydrates in plant 
materials. The oxycelluloses may be easily distinguished, however, 
from pentosans by the fact that they yield glucose exclusively upon 
hydrolysis with acids, the hydrolytic products giving none of the re- 
actions (osazone, color tests, etc.) characteristic of the pentoses. 

Methylfurfural Reactions for Methylpentose Groups. In the 
same way that all substances containing pentose groups yield furfural 
upon distilling with hydrochloric acid, those materials containing methyl- 
pentose groups yield methylfurfural. The reaction is perfectly anal- 
ogous to that described upon page 374. 



iOH H; 
| >CH, 

-C-iOHi = y H -% 



| | X C:O H 

jpH"H|H 

Methylpentose (164 parts) Methylfurfural (110 parts) Water (54 parts) 

The theoretical yield of methylfurfural from methylpentose accord- 
ing to the above reaction is 67.07 per cent. In actual distillation ex- 
periments with the methylpentoses, fucose and rhamnose, only from 
35 to 40 per cent methylfurfural is obtained or 50 to 60 per cent of the 
theoretical amount. 

In testing natural products for the presence of methylpentose 
groups, the material is distilled with hydrochloric acid of 1.06 sp. gr. in 
exactly the same manner as described for pentoses and the distillate 
tested for methylfurfural. If no furfural is present in the distillate the 
presence of methylfurfural will be indicated by aniline-acetate paper, 
which in this instance is colored yellow. If pentosans are also present 
in the plant material being examined, as is nearly always the case, the 
presence of furfural in the distillate will color the aniline-acetate paper 
red and completely mask the yellow color of the methylfurfural re- 
action. Other tests must, therefore, be employed to detect the presence 
of methylfurfural. 

Maquenne * has devised a reaction by which 1 part methylfur- 
fural can be detected in presence of 9 parts furfural. A small amount 
of the solution to be tested is added to a mixture containing 3 volumes 

* Compt. rend., 109, 573. 



378 SUGAR ANALYSIS 

95 per cent alcohol and 1 volume concentrated sulphuric acid and the 
whole gently warmed. The development of a bright grass-green color 
throughout the body of the solution indicates the presence of methyl- 
furfural. 

Spectral reactions for methylfurfural will be described in a succeed- 
ing section. 

Reactions for Tetrose and Triose Groups. Excepting the hexoses, 
pentoses and methylpentoses, but few experiments have been made 
concerning the reactions of other sugar groups with hydrochloric acid. 

Experiments of Tollens and Ellett * show that 1-erythrose is de- 
composed upon heating with hydrochloric acid into lactic acid. The 
reaction may proceed as follows: 

C 4 H 8 4 = C 3 H 6 3 + CH 2 

Tetrose Lactic acid Formaldehyde 

Tollens and Ellett suggest that the above may be a general reaction 
for tetrose groups, just as levulinic acid is formed from hexoses, fur- 
fural from pentoses, and methylfurfural from methylpentoses. 

The formation of considerable methylglyoxal CH 3 CO COH by 
heating dioxyacetone, CaHeOa, with sulphuric acid has been observed 
by Pinkus.f This may perhaps be a group reaction of trioses. 

Further investigations require to be made upon the tetroses and 
trioses before any results from the above observations can be applied to 
sugar analysis. 

III. COLOR AND SPECTRAL REACTIONS AS A MEANS OF IDENTIFYING 

SUGARS 

A study of the color reactions and absorption spectra which solu- 
tions of different sugars give with various phenols as a-naphthol, orcin, 
resorcin, naphthoresorcin and phloroglucin, in presence of concentrated 
sulphuric or hydrochloric acids offers frequently a most rapid as well 
as most reliable method for detecting sugar groups. 

Color Reactions of Ketoses. Reference has already been made 
(p. 340) to the greater ease with which solutions of ketoses show colora- 
tion phenomena in contact with concentrated sulphuric acid. The same 
fact has been noted with the colorations produced with sugars and 
a-naphthol and sulphuric acid, and this has been utilized as one means 
of detecting the presence of ketose sugars in mixtures. 

a-Naphthol Test. Pinoff { has modified the a-naphthol test for sugars 
by using a mixture of 750 c.c. 96 per cent alcohol and 200 gms. con- 
centrated sulphuric acid as the condensing agent. By treating in a test 

* Ber., 38, 499. f Ber., 31, 31. t Ber., 38, 3314. 



METHODS FOR THE IDENTIFICATION OF SUGARS 379 



tube 0.05 gm. of sugar with 10 c.c. of the alcohol-acid mixture and 
0.2 c.c. of alcoholic a-naphthol (5 gms. a-naphthol dissolved in 100 c.c. 
96 per cent alcohol) and heating in boiling water, Pinoff obtained red 
colorations which in case of sugars containing ketone groups appeared 
almost immediately; with the aldose sugars 20 minutes or more elapsed 
before coloration developed. The following table for 1 1 different sugars 
by Pinoff gives the time of heating before coloration, the number of 
absorption bands shown by the solution before the spectroscope and 
the position of the bands with reference to the wave length of the light 
absorbed. 

TABLE LXVIII 

Giving Absorption Spectra of Sugars with a-Naphthol and Sulphuric Acid in Alcohol 



Sugar. 


Time for 
develop- 
ment of 
color. 


Number of 
absorption 
bands. 


Wave length in nn and position of bands. 


Arabinose . . . 


Minutes. 
20 






Rhamnose . 


20 


1 


562.5 (in yellow) 


Glucose 


35 




532 5 (between yellow and green) 


Mannose 


31 


1 


532 . 5 (between yellow and green) 


Galactose 


31 


1 


532.5 (between yellow and green) 


Fructose 


1 


2 


573.6 (in yellow), 508.8 (in green) 


Sorbose 


1 


2 


573.6 (in yellow), 508.8 (in green) 


Sucrose 


1 


2 


573.6 (in yellow), 508.8 (in green) 


Lactose 


31 


1 


532 5 (between yellow and green) 


Maltose . . 


31 


1 


532 5 (between yellow and green) 


Raffinose 


1 


o 


573.6 (m yellow), 508.8 (in green) 











It will be noted that for the ketose sugars fructose and sorbose and 
for the di- and tri-saccharides sucrose and raffinose, which give the 
ketose sugar fructose upon hydrolysis, a red coloration is obtained in 
1 minute, while for the other sugars 20 to 35 minutes must elapse 
before coloration. By diluting the 10 c.c. of sulphuric-acid alcohol 
mixture with 10 c.c. of 96 per cent alcohol before making the test, 
Pinoff obtained no coloration sufficient to show absorption bands with 
any of the aldose sugars. For the ketose sugars he obtained the fol- 
lowing results: 



Sugar. 


Time for de- 
velopment of 
color. 


Number of 
bands. 


Wave length in MM and 
position of bands. 


Fructose . . . 


Minutes. 
13 


1 


508.8 (in green) 


Sorbose 


30 


1 


508.8 (in green) 


Sucrose 


15 


1 


508.8 (in green) 


Raffinose 


19 


1 


508.8 (in green) 











380 SUGAR ANALYSIS 

While diluting the acid-alcohol mixture has practically eliminated 
the aldoses from the reaction, it has also materially lessened the sen- 
sibility of the test for the ketoses. 

Resorcin Test. The most convenient color test for distinguishing 
ketose from aldose sugars is the color reaction with resorcin and hydro- 
chloric acid generally known as Seliwanoff's * test. The test was 
originally regarded as peculiar to fructose, but later experiments have 
shown that it is given by sorbose, tagatose, the keto-pentoses and all 
other sugars having a ketone group. 

The reaction is carried out by mixing in a test tube 10 c.c. of the 
clarified solution to be tested with 10 c.c. of 25 per cent hydrochloric 
acid, then adding a little resorcin (about the tip of a knifebladeful), and 
heating gently over a small flame. If fructose or other ketose is present 
a fiery eosin-red color will develop, which upon cooling and standing 
will deposit as an amorphous powder mixed with humus decomposition 
products. 

If the acid solution is made alkaline with soda and then shaken with 
amyl alcohol, the red coloring matter is dissolved with a greenish 
fluorescence. If a few drops of absolute alcohol be now added the 
color becomes a beautiful rose red. 

If the red-colored solutions obtained by Seliwanoff's reaction be ex- 
amined before the spectroscope a distinct absorption band will be noted 
in the blue near the F line. (See Fig. 165.) 

It is important in making the test with resorcin that an excess of 
hydrochloric acid be avoided. The percentage of acid in the final mix- 
ture should be about 12 J per cent. If too much strong acid is present, 
glucose and other aldoses will also react with resorcin and form pink- 
colored solutions; the latter, while lacking the intensity of color obtained 
with the ketoses, may nevertheless lead to erroneous conclusions. The 
resorcin reaction obtained with glucose may be due to a slight trans- 
formation of this sugar into fructose. Ost, as a matter of fact, has 
succeeded in effecting such a transformation by treating glucose in 
the cold with strong sulphuric acid. 

Pinoff f has modified the resorcin test for ketoses by using the 
alcohol-sulphuric-acid mixture previously described as the condensing 
agent. In making the test 0.05 gm. of sugar was treated in a test 
tube with 5 c.c. of the alcohol-sulphuric-acid reagent, 5 c.c. alcohol and 
0.2 c.c. of a 5 per cent resorcin solution and the mixture placed in boil- 
ing water. The following table for 11 different sugars by Pinoff gives 
the length of time required for development of color, the number of 
* Ber., 20, 181. t Ber., 38, 3314. 



METHODS FOR THE IDENTIFICATION OF SUGARS 381 



absorption bands and the position of the bands with reference to wave 
length of light absorbed. 

TABLE LXIX 

Giving Absorption Spectra of Sugars with Resorcin and Sulphuric Acid in Alcohol 



Sugar. 


Time for de- 
velopment of 
color. 


Number of ab- 
sorption bands. 


Wave lengths in nn and 
position of bands. 


Arabinose ... 


Minutes. 
35 






Rhamnose 


35 






Glucose 


32 


1 


487.5 (in blue) 


Mannose 


35 






Galactose 


35 






Fructose 


1 


1 


487.5 (in blue) 


Sorbose 


1 


1 


487.5 (in blue) 


Sucrose 


1 


1 


487 5 (in blue) 


Lactose 


32 


1 


487 5 (in blue) 


Maltose 


32 


1 


487 5 (in blue) 


Raffinose 


1 


1 


487 5 (in blue) 











Naphthoresorcin Test. Tollens and Rorive * have employed in 
place of resorcin naphthoresorcin or 1 : 3 dioxynaphthalin. The ketose 
sugars fructose and sorbose and the di- and trisaccharides sucrose 
and raffinose show upon heating with a little naphthoresorcin in pres- 
ence of hydrochloric acid (1 vol. acid 1.19 sp. gr. and 1 vol. water) beauti- 
ful red-colored solutions which show a weak absorption band in the 
green. The sensibility of this test is about the same as that obtained in 
Seliwanoff's reaction, but the color has more of a violet tinge than the 
fiery red obtained with resorcin. The red-colored solutions obtained 
with naphthoresorcin soon become turbid with formation of a deposit. 
If the latter is filtered off and dissolved in alcohol a yellowish-brown 
solution with green fluorescence is obtained which shows a weak absorp- 
tion band in the green. 

Color Reactions of Pentoses (and Glucuronic Acid) . The pentoses 
are distinguished above all other sugar groups for the depth and variety 
of the color reactions obtained with different polyvalent phenols in 
presence of concentrated hydrochloric acid. Phloroglucin, orcin and 
naphthoresorcin are the three compounds most used for this purpose, 
and the reactions for each of these will be described in the order 
named. 

Phloroglucin Test. Ihl f discovered that solutions of the pentose 
sugars, or of hydrolytic products derived from substances containing 

* Ber., 41, 1783. t Chemiker Ztg. (1885), 231. 






382 SUGAR ANALYSIS 

pentosans, gave, upon heating with an equal volume of concentrated 
hydrochloric acid and a little phloroglucin, a beautiful violet-red color. 
The colored solution thus obtained when viewed before the spectro- 
scope was found by Tollens and Allen * to show a sharp black absorp- 
tion band in the yellow of the spectrum between the D and E lines. 

The violet-red solution obtained in the phloroglucin reaction for 
pentoses soon becomes turbid with deposition of a dark-colored precipi- 
tate. If the turbid solution is allowed to stand 3 to 5 minutes, then 
cooled, filtered and the precipitate washed with cold water on a small 
rapid filter and then dissolved in 95 per cent alcohol, a permanent red 
solution is obtained which is perfectly adapted to the study of ab- 
sorption spectra. If the color is too deep it can be reduced by careful 
dilution with 96 per cent alcohol. (Tollens's " absatz " method.) 

The same color reaction of the pentoses with phloroglucin and 
hydrochloric acid is given by glucuronic acid and its derivatives, but 
not by oxycellulose. The test, therefore, while enabling the chemist to 
distinguish between such furfural-yielding substances as pentosans and 
oxycellulose, does not permit the distinction between glucuronic acid 
and pentoses (as for example in urine). 

Orcin Test. If the reaction for the pentoses just described be 
carried out with orcin in place of phloroglucin a violet-blue coloration 
is obtained. The solution, however, becomes rapidly turbid with de- 
position of a bluish-colored flaky precipitate. If the latter is filtered 
off and dissolved in alcohol by Tollens's " absatz " method a blue- 
colored solution is obtained which shows before the spectroscope a very 
sharp dark band almost exactly over the D line of the spectrum. The 
same reaction is also obtained with glucuronic acid. 

Bial f has made the orcin reaction more sensitive by carrying out 
the test in presence of a little ferric chloride. In this manner it is 
found possible to distinguish between pentoses and glucuronic acid. 

Bial's orcin reagent is prepared by dissolving 1 gm. orcin in 500 c.c. 
hydrochloric acid of 1.15 sp. gr. (30 per cent) to which 20 drops of 
an officinal solution of ferric chloride (liquor ferri sesquichloridi) are 
added. 

In making the test 4 to 5 c.c. of the reagent are heated in a test 
tube to boiling; the solution is removed from the flame and a few drops 
(never over 1 c.c.) of the solution to be tested added. If pentoses are 
present a vivid green color will develop almost immediately; the re- 
action is not given under the above conditions with glucuronic acid. 

* Ann., 260, 289. 

t Biochem. Zeitschrift., 3, 323. 



METHODS FOR THE IDENTIFICATION OF SUGARS 383 

BiaFs test has been studied and generally confirmed by Sachs,* and 
also by Tollens and Lefevre.f The last-named authorities found that 
a dilute solution of glucuronic acid produced no perceptible coloration 
under the conditions prescribed by Bial, but that if the solution was 
heated for any length of time a green color speedily developed. The 
cause of the retardation is explained by the slower decomposition of 
glucuronic acid by hydrochloric acid as compared with the pentoses; 
such a difference in the rate of decomposition is also noted between the 
pentose sugars themselves, xylose, for example, giving a coloration with 
Dial's reagent in a shorter time than arabinose. 

The green solution obtained by BiaFs reaction shows before the 
spectroscope a dark absorption band in the red between the lines B and 
C and a second band in the yellow covering the position of the D line 
of the spectrum. 

Naphthoresorcin Test for Pentoses and Glucuronic Acid. Tollens and 
Rorive J have found that when solutions of different sugars are heated 
with a little naphthoresorcin in presence of an equal volume of concen- 
trated hydrochloric acid (1.19 sp. gr.) characteristic colored solutions 
and deposits are formed. 

With the pentoses arabinose and xylose a red color develops on 
heating followed by a bluish turbidity. The deposit dissolves in alco- 
hol to a reddish-brown solution with beautiful green fluorescence, show- 
ing a weakly-defined absorption band in the green. 

Glucuronic acid gives with naphthoresorcin and hydrochloric acid 
a bluish turbid solution with blue deposit. The alcoholic solution of 
the latter is a beautiful blue, only slightly fluorescent, and shows a 
dark absorption band in the yellow covering the D line of the 
spectrum. 

The naphthoresorcin test for glucuronic acid has been improved by 
Tollens in the following way. The deposit of coloring matter is 
treated with ether instead of alcohol; if glucuronic acid is present the 
ether is colored a violet blue and shows before the spectroscope an 
absorption band in the yellow, its center lying a little to the right of the 
D sodium line (i.e., toward the green). 

The naphthoresorcin deposits obtained with sugars (pentoses, 
hexoses, etc.) in presence of hydrochloric acid are insoluble in ether 
and so do not appear in the reaction. The presence of sugar and also 
of foreign organic matter, as in urine, may change the color of the ether 
solution from the violet blue characteristic of pure glucuronic acid to a 

* Biochem. Zeitschrift., 1, 384. { Ber., 41, 1783. 

t Ber., 40, 4520. Ber., 41, 1788. 




384 



SUGAR ANALYSIS 



violet, red, or reddish brown. The characteristic absorption band in 
the yellow part of the spectrum will not, however, be interfered with. 



B C 




Indigo 



Violet 



Fructose, resorcin and hydro- 
chloric acid. 



Sucrose, a-naphthol and sulphuric 
acid. 



Arabinose, phloroglucin and hy- 
drochloric acid. 



Methylfurfural, phloroglucin and 
hydrochloric acid. 



Methylfurfural and hydrochloric 
acid. 



Fig. 165. Absorption spectra given by different sugars. 

The naphthoresorcin test as prescribed by Tollens is made as fol- 
lows: 5 to 6 c.c. of the solution (urine, etc.) to be tested are treated in 



METHODS FOR THE IDENTIFICATION OF SUGARS 385 

a 16 mm. wide test tube with \ to 1 c.c. of a 1 per cent solution of 
naphthoresorcin in alcohol and an equal volume of hydrochloric acid 
of 1.19 sp. gr. added. The solution is carefully heated to boiling and 
then kept for 1 minute over a small flame. The dark-colored solution 
is set aside for 4 minutes and then cooled under a stream of cold water; 
an equal volume of ether is then added and the whole thoroughly 
shaken. After the acid solution has settled the ether layer will be 
colored blue or bluish violet to red, in case glucuronic acid is present, 
and, if the tube is held before the spectroscope, will show the character- 
istic absorption band near the D line. In case the ether does not sepa- 
rate readily a drop or two of alcohol will hasten the process. If the 
ether solution is too deeply colored for spectroscopic examination more 
ether is added until the color is reduced and the unabsorbed part of the 
spectrum made visible. 

The naphthoresorcin deposits of the pentoses and other sugars 
being insoluble in ether separate as a layer between the colored ether 
and the lower acid solution. 

Color Reactions of Methylpentoses. The color reactions for 
detection of methylpentoses may be divided into two classes: (1) color 
reactions made upon the distillate obtained by distilling methylpentoses 
or methylpentosans with hydrochloric acid; (2) color reactions made 
directly upon these substances without distillation. The color reactions 
of the first class are in reality color reactions of methylfurfural to 
which reference has already been made. It remains, however, to describe 
some of the spectral reactions of methylfurfural. 

Spectral Reactions of Methylfurfural. Tollens and Widtsoe * have 
detected the presence of methylfurfural in the hydrochloric acid dis- 
tillate from various plant materials by mixing a few cubic centimeters 
of the solution with an equal volume of concentrated hydrochloric acid 
and gently warming. If the solution is colored yellow methylfurfural is 
present. The yellow solution viewed before the spectroscope will show 
a dark absorption band between the green and blue of the spectrum 
near the F line. If much methylfurfural is present the band will grad- 
ually darken and broaden, the increase in width extending toward the 
violet and leaving the green unaffected. With considerable methyl- 
furfural the violet end of the spectrum is completely extinguished, the 
green, however, always remaining clear and transparent. Furfural does 
not give this reaction although it may affect the delicacy of the test if 
present in large amount. The reaction, however, will indicate 1 part of 
methylfurfural in presence of 64 parts furfural (^V drop methylfurfural 

* Ber., 33, 146. 



386 SUGAR ANALYSIS 

in presence of 2 drops furfural in 10 c.c. of hydrochloric acid). By use 
of this test Tollens and Widtsoe were able to detect methylpentosans 
in different gums, sea weed, leaves of different kinds of trees and a 
large variety of other plant substances. 

Tollens and Oshima * have rendered the spectral reaction for 
methylfurfural more sensitive by carrying out the test in presence of 
phloroglucin; 5 c.c. of the hydrochloric acid distillate are treated with 
5 c.c. of concentrated hydrochloric acid and a few cubic centimeters of 
a solution of phloroglucin (in hydrochloric acid of 1.06 sp. gr.) added. 
After 5 minutes the solution is filtered from the greenish-black precipi- 
tate of furfural phloroglucide; if the filtrate is colored yellow or reddish 
yellow methylfurfural is present. The solution gives before the spectro- 
scope a dark absorption band in the blue. On long standing the solu- 
tion deposits a red precipitate of methylfurfural phloroglucide which 
is readily distinguished from the dark-green furfural compound. Ab- 
sorption spectra of methylfurfural are shown in Fig. 165. 

The vivid color reactions of the pentoses with orcin and phloroglucin 
are not obtained with the methylpentoses. Naphthoresorcin, however, 
was found by Tollens and Rorive to give a deposit of coloring sub- 
stance with the methylpentoses, rhamnose and fucose, when heated in 
presence of hydrochloric acid. The alcoholic solution of the deposits 
showed a violet blue color with an exceedingly brilliant green fluores- 
cence, which showed before the spectroscope an absorption band in the 
yellow over the D line and a second band in the green. 

There are a number of other color spectral reactions which have not 
been described; these belong, however, more to the reactions of individ- 
ual sugars and will be given under the description of these. 

A few characteristic absorption spectra, useful in testing for sugars, 
are shown in Fig. 165. 

Reactions of the Non-reducing Sugars 

The comparatively small number of sugars, which do not reduce 
Fehling's solution, all belong to the higher di-, tri- and tetrasaccharides 
and include sucrose, trehalose, raffinose, melezitose, gentianose, lacto- 
sinose, secalose, lupeose and stachyose. The soluble polysaccharides, 
such as dextrin, inulin, glycogen, etc., although not classified as sugars, 
are sometimes included for convenience in the group of non-reducing 
saccharides. 

A free aldehyde, or ketone group, to which the reducing sugars owe 
their peculiar reactivity in the formation of hydrazones, oximes, ureides, 

* Ber., 34, 1425. 



METHODS FOR THE IDENTIFICATION OF SUGARS 387 

mercaptals, etc., is lacking in the non-reducing sugars, and the in- 
ability of the latter to reduce Fehling's solution, or to react with phe- 
nylhydrazine, dilute alkalies, hydroxylamine, etc., is thus explained. 

The non-reducing sugars give many of the color and spectral re- 
actions of the reducing sugars, sucrose and raffinose, for example, giv- 
ing the a-naphthol reaction with sulphuric acid and Seliwanoff s reaction 
with resorcin and hydrochloric acid. But as previously explained these 
reactions are not given by the original non-reducing sugar, but by the 
reducing sugars derived from this by the hydrolytic action of the acid 
used in making the test. 

A carefully controlled hydrolysis by means of acids or enzymes, 
combined with quantitative measurements of changes in polarization 
or in copper-reducing power, is the most reliable test for the presence of 
non-reducing sugars. Methods involving this principle have been de- 
scribed under the inversion methods for determining sucrose and raffinose, 
and other examples will be given under quantitative chemical methods. 
Individual tests will be described under the heading of each single sugar 
in Part II of this Handbook. 



CHAPTER XIV 

REDUCTION METHODS FOR DETERMINING SUGARS 

THE principal chemical methods for determining sugars are based 
upon the property which all aldehydes and ketones have of reducing 
alkaline solutions of certain metallic salts. The reducing action of 
glucose, lactose and other sugars upon alkaline solutions of copper, 
silver, mercury, bismuth and other metals has already been mentioned. 
In the case of silver and glucose, for example, the reaction when care- 
fully controlled proceeds as follows: 

C 6 H 12 O 6 + 9 Ag2p = 18 Ag + 3 (COOH) 2 + 3 H 2 0. 

Glucose Silver oxide Silver Oxalic acid Water. 

If the weight of reduced silver be determined for this reaction, the 
amount of glucose can easily be estimated. But unfortunately the re- 
ducing action of sugars upon metallic salts does not proceed with the 
quantitative precision of the above equation; the reduction is rarely 
complete and the amount of reduced metal varies with the conditions 
of the experiment. The latter difficulty is obviated, however, in prac- 
tice by controlling the process so that the same weight of reduced 
metal is always obtained for the same weight of sugar. 

Of the various alkaline solutions of metals those of copper are em- 
ployed almost exclusively in sugar analysis. 

COPPER REDUCTION METHODS 

Early Methods. The reducing action of sugars upon different 
salts of copper has been known since the first beginning of chemistry. 
Trommer,* in 1841, first noted the value of alkaline copper-sulphate 
solution as a means of distinguishing grape from cane sugar. Trom- 
mer 's method was improved in 1844 by Barreswil f who made the im- 
portant discovery that addition of potassium tartrate to the alkaline 
copper-sulphate solution greatly increased its stability. Barreswil's 
method was volumetric; the sugar solution was slowly added to the 
boiling copper reagent, which had previously been standardized against 
pure glucose, until the blue color was just discharged. 

* Ann., 39, 360. 

t Journal de Pharmacie [3], 6, 301. 
388 



REDUCTION METHODS FOR DETERMINING SUGARS 389 

Fehling's Method. Pehling,* in 1848, first worked out the details 
of the alkaline copper method, as they now stand, and the copper-sul- 
phate and alkaline-tartrate reagent has since been called by his name. 

The copper solution employed by Fehling consisted of 40.00 gms. 
copper sulphate, CuS0 4 .5 H 2 O, 160 gms. neutral potassium tartrate and 
600-700 gms. sodium hydroxide sol. of 1.12 sp. gr. dissolved in water to 
1154.4 c.c. This is equivalent to 34.65 gms. CuS0 4 .5 H 2 dissolved 
to 1000 c.c., the proportion used by nearly all subsequent workers down 
to the present time. 

Fehling's solution contains 8.822 gms. copper to 1000 c.c. or 0.008822 
gm. to 1 c.c. According to Fehling's experiments 1 c.c. of his solu- 
tion was exactly reduced by 0.005 gm. of anhydrous glucose, or 1 part 
glucose reduced 1.765 parts copper. In terms of the molecular weight 
of glucose the ratio would be 180 X 1.765 = 317.6. Dividing this 
value by 63.6, the atomic weight of copper, the atoms of copper reduced 
by one molecule of glucose is found to be five. The reduction ratio 
1 : 5 was regarded as constant by Fehling and was so employed by all 
chemists until Soxhlet f showed in 1878 that the ratio between sugar 
and amount of copper reduced was not a constant but varied according 
to the excess of copper which is present during the reaction. 

The more modern methods of sugar determination, which employ 
Fehling's solution, may be divided into two general classes. I. Volu- 
metric methods based upon the complete reduction of a measured 
volume of standard solution. II. Methods based upon a gravimetric 
or volumetric determination of the reduced copper. 

VOLUMETRIC METHODS BASED UPON THE COMPLETE REDUCTION OF A 
MEASURED VOLUME OF FEHLING'S SOLUTION 

Soxhlet's Method. Owing to the decomposition which takes 
place in the mixed copper-sulphate and alkaline-tartrate solution upon 
standing, the two solutions employed in the Soxhlet and all other 
modern methods are mixed only just before using. The solutions con- 
sist of the following: Solution A, 34.639 gms. of pure crystallized 
CuSO 4 .5 H 2 are dissolved in water and made up to 500 c.c. Solution 
B, 173 gms. of Rochelle salts are dissolved in water, 100 c.c. of a solu- 
tion of caustic soda, containing 516 gms. NaOH per liter are added, 
and the volume completed to 500 c.c. Previous to analysis mix equal 
volumes of solutions A and B. 

Before using the mixed copper reagent, it shoulc} be standardized 
against glucose, invert sugar, lactose, etc., according to the needs of 
* Ann., 72, 106; 106, 75. t J- prakt. Chem. [2], 21, 227. 



390 SUGAR ANALYSIS 

analysis. Since reducing sugar in sugar-cane, sugar-beet and most 
other food products is most usually expressed as invert sugar, the latter 
is most commonly used for standardization. A standard solution of 
invert sugar has also an advantage in being easily prepared. 

Standard Invert Sugar Solution. Method of the Association of 
Official Agricultural Chemists.* Dissolve 4.75 gms. of pure sucrose in 
75 c.c. of water, add 5 c.c. of 38.8 per cent hydrochloric acid and set 
aside during a period of 24 hours at a temperature not below 20 C. 
Neutralize the acid exactly with dilute sodium hydroxide and make up 
to 1000 c.c.; 100 c.c. of this solution contains 0.500 gm. of invert sugar. 

The amount of standard invert sugar solution necessary to reduce 
100 c.c. of the mixed copper reagent is determined according to the 
details described in the next paragraph. 

Application to Analysis of Sugar Products. Method of the Associa- 
tion of Official Agricultural Chemists.^ Make a preliminary titration 
to determine the approximate percentage of reducing sugar in the ma- 
terial under examination. Prepare a solution which contains approx- 
imately 1 per cent of reducing sugar. Place in a beaker 100 c.c. of the 
mixed copper reagent and approximately the amount of the sugar 
solution for its complete reduction. Boil for two minutes. Filter 
through a folded filter and test a portion of the filtrate for copper by 
use of acetic acid and potassium ferrocyanide. Repeat the test, vary- 
ing the volume of sugar solution, until two successive amounts are 
found which differ by 0.1 c.c., one giving complete reduction and the 
other leaving a small amount of copper in solution. The mean of these 
two readings is taken as the volume of the solution required for the 
complete precipitation of 100 c.c. of the copper reagent. 

Under these conditions 100 c.c. of standard copper reagent require 
0.475 gm. of anhydrous glucose or 0.494 gm. of invert sugar for com- 
plete reduction. Calculate the glucose by the following formula: 

V = the volume of the sugar solution required for the complete 

reduction of 100 c.c. of standard copper reagent. 
W = the weight of the sample in 1 c.c. of the sugar solution. 

rp, 100 X 0.475 

I hen _ = per cent of glucose, 

r X W 

100 X 0.494 . 

or = per cent of invert sugar. 

- v /\ w 

In making the test for unreduced copper a few drops of the filtered 
solution are placed upon a white test plate, acidified with a few drops of 

* Bull. 107 (revised) U. S. Bur. of Chem., p. 42. f Ibid. 



REDUCTION METHODS FOR DETERMINING SUGARS 391 

10 per cent acetic acid and a drop of 2 per cent potassium-ferrocyanide 
solution added. A brown coloration indicates the presence of unre- 
duced copper. 

Volume of Fehling's Solution Reduced by Different Sugars. - 
The ratio between volume of standard Fehling's solution and the amount 
of different sugars, just sufficient to cause complete reduction, was de- 
termined by Soxhlet * to be as follows: 

TABLE LXX 



Volume of Fehling's solution reduced by different sugars. 


Reducing 
power in terms 
of glucose. 


0.5000 
0.5000 
0.5000 
0.5000 
0.5000 


gm. 'glucose reduces 105.2 c.c. Fehling's 
" in vert sugar " 101.2 " 
" fructose " 97.2 " 
" lactose " 74.0 " 
" maltose " 64.2 " 


solution. 


1.000 
0.962 
0.924 
0.703 
0.610 


u 

it 


tc 
It 





The above results calculated to equal volumes of copper reagent 
show that 100 c.c. of mixed standard Fehling's solution are reduced by 
0.4753 gm. of glucose, 0.4941 gm. of invert sugar, 0.5144 gm. of fructose, 
0.6757 gm. of lactose and 0.7788 gm. of maltose. 

Modifications of Soxhlet's Method. Instead of employing 100 c.c. 
of Fehling's solution for the Soxhlet determination, it is more customary 
to use 10 c.c., 20 c.c. or 50 c.c., the quantity thus used being measured 
into a casserole, beaker or flask, and diluted, according to require- 
ments, with a measured volume of water. In case of very dilute sugar 
solutions, as small a quantity as 5 c.c. of Fehling's solution may be used 
to advantage. 

In using any of the numerous modifications of Soxhlet's method, it 
is important that the Fehling solution be standardized under exactly 
the same conditions as in analysis. The same degree of dilution should 
be followed for the mixed copper reagent in all experiments. Soxhlet 
found that 0.5 gm. of glucose reduced 105.2 c.c. of Fehling's solution 
when undiluted and only 101.1 c.c. when diluted with 4 parts of water; 
similar results were also obtained with other sugars. Such differences 
as these might produce a variation of several per cent in the estimation 
of reducing sugars. 

It is also evident that to obtain the most concordant results the 
sugar solutions should always contain about the same percentage of 
reducing sugar. This is accomplished in practice by making a rough 
* J. prakt. Chem. [2] 21, 227. 




392 SUGAR ANALYSIS 

preliminary determination and then making up a fresh sugar solution 
so that the percentage of reducing sugar shall be 0.1 per cent, 0.5 per 
cent or 1.0 per cent, etc., according to the volume of Fehling's solution 
taken and the individual preference of the chemist. In this manner 
approximately the same volume of sugar solution is always used for 
reducing the same volume of copper reagent, and under such con- 
ditions, with a uniform method of boiling, the most accurate results are 
obtained. 

A difference in reducing power is also obtained whether the sugar 
solution be added to the copper reagent in small portions, with suc- 
cessive periods of boiling, or only in one portion with one period of 
boiling. The most accurate results are secured where the test is made 
with the entire volume of sugar solution, necessary for complete reduc- 
tion, with only one period of boiling. 

The following example will give an illustration of the application of 
the method: 

Example. 20 c.c. of Fehling's solution diluted with 80 c.c. of water were 
found to require for reduction exactly 20.2 c.c. of standard invert sugar solu- 
tion or 0.101 gm. 

50 gms. of sugar-cane molasses were diluted to 1000 c.c. Of this solution 
about 8 c.c. were required to discharge the blue color of 20 c.c. Fehling's solu- 
tion. 

80 c.c. of the sugar solution (4 gms. molasses) were then made up to 200 c.c. 
(1 c.c. = 0.02 gm. molasses). Of this solution 19.6 c.c. when boiled with 
20 c.c. Fehling's solution and 80 c.c. of water for 2 minutes showed incomplete 
reduction by the ferrocyanide test and 19.8 c.c. complete reduction. 

Calling 19.7 c.c. the volume of sugar solution necessary to reduce the 

20 c.c. of Fehling's solution, then ^^ ^ = 25.64 per cent invert sugar in 

U.U.Z X 19.7 

the molasses. 

The Ferrocyanide Test for Copper. Several methods are fol- 
lowed for making the ferrocyanide test for unreduced copper. It some- 
times happens that the cuprous oxide is precipitated in a very finely 
divided form, and gives annoyance by running through the filter. 

One method of making the test is to superimpose several small strips 
of filter paper and allow a few drops of the solution to fall upon the 
upper paper. The moistened area upon the second or third underlying 
strip is then treated with a drop of ferrocyanide solution acidified with 
acetic acid. The appearance of a brown spot indicates the presence of 
unreduced copper. 

Another method of removing the portion of solution to be tested is 
by means of a Wiley or Knorr filtering tube, which is prepared as fol- 
lows: 



REDUCTION METHODS FOR DETERMINING SUGARS 393 



Wiley's Filter Tube. The Wiley * filter tube, Fig. 166a, consists of 
a piece of glass tubing, 5 to 7 mm. in diameter and 20 to 25 cm. long, 
one end of which has been softened in a flame and 
then pressed out so as to form a shoulder. A piece 
of fine linen is then stretched tightly over the end and 
tied securely by a thread. In using the tube the 
covered end is dipped into water containing in suspen- 
sion finely divided asbestos, and a film of the latter 
spread over the surface of the filter by suction at the 
upper end. A small portion of the liquid to be tested 
is sucked into the tube and then poured from the 
open end onto the test plate. Knorr's f modification 
of the Wiley tube is of smaller diameter and contains 
a perforated platinum disk in place of the linen (Fig. 
1666). The disk is coated with asbestos and the liquid 
withdrawn for testing as with the Wiley tube. The 
filter tubes should not be reused until after cleaning 
in dilute nitric acid and washing with water. 

Method of Ross. A method due to Ross,| and 
employed quite extensively in Louisiana, is to dip the 
point of a small folded filter, held by means of for- 
ceps, below the surface of the hot solution in the cas- 
serole and withdraw a few drops of the clear liquid 
from the interior of the filter by means of a medicine 
dropper (Fig. 167). The method is simple, and par- 
ticularly useful where there is a large amount of 
routine. 

Conveniences for making the determination by 
Soxhlet's method, such as 2-minute sand glass for 

regulating time of boiling, test plate, dropping bottles pj g 166 Filter 

for ferrocyanide solution and acetic acid, are shown in tubes for deter- 
Fig. 167. mining reducing 

Violette's Method. The volumetric method of su s ars - 
copper reduction, which is used most extensively in France, is that 
of Violette. The proportions of copper sulphate, Rochelle salts and 
alkali employed in the Soxhlet method may be used in the Violette 
determination, or the Violette solution may be taken which consists of 

* Wiley's "Agricultural Analysis" (1897), III, 130. 
t Ibid. 

j Journal of Analytical Chemistry, 4 (1890), p. 427. 
Sidersky's "Manuel" (1909), p. 95. 



394 



SUGAR ANALYSIS 




Fig. 167. Ross's method for determining reducing sugars. 

36.46 gms. CuS0 4 .5 H 2 0, 200 gms. Rochelle salts and 500 gms. sodium 
hydroxide solution of 1.2 sp. gr. made up to 1000 c.c. 

The Violette solution takes a slightly larger amount of copper sul- 
phate than the Soxhlet solution in order that 1 c.c. may correspond to 
the invert sugar derived from 5 mgs. of sucrose or || X 5 = 5.263 mgs. 
of invert sugar. The ratio of invert sugar and copper Sulphate for the 
Soxhlet and Violette solutions is accordingly 5 : 34.64 :: 5.263 : 36.46. 

The Violette solution is preferred by some chemists for convenience 
in determining sucrose by the method of inversion and copper reduc- 
tion. 

The end point of the reduction in Violette's method is determined, 
as in the early process of Barreswil, by the disappearance of blue color 
from the copper solution. The details of the method are as follows: 

Ten cubic centimeters of the mixed copper solution are transferred 
to a large test tube 20 to 22 mm. in diameter and 22 to 24 cm. long; 
5 c.c. of distilled water are added in case the solution is rich in reducing 
sugars and a few small pieces of pumice stone, which have been ignited 
and then washed in acid and water. The copper solution is then 
heated to boiling, the grains of pumice stone giving a smooth ebullition 
and preventing the sudden ejection of liquid from the tube. The sugar 



REDUCTION METHODS FOR DETERMINING SUGARS 395 

solution to be tested, which should have been previously clarified and 
diluted to about 0.5 to 1.0 per cent invert sugar, is then added from a 
burette, a few cubic centimeters at a time, the copper solution being 
boiled for 2 minutes after each addition. As the reduction proceeds 
the blue color of the solution becomes more of a reddish violet, due to 
the diminishing intensity of the blue and the increasing amount of the 
red cuprous oxide. Towards the end of the reduction it is necessary 
to hold the tube against a white wall or paper and observe the color of 
the clear solution, after the red oxide begins to settle. When the final 
drop of sugar solution discharges the last trace of blue color, the read- 
ing of the burette is noted, and the calculation of sugar made as pre- 
viously described. 

A little practice is required in the Violette method in following the 
disappearance of the blue color. The chemist should standardize his 
solution against invert sugar, following the same procedure in deter- 
mining end point as in making an analysis. 

The Violette method is much simpler than the Soxhlet method and 
is for this reason preferred by many chemists. The Soxhlet method, on 
the other hand, owing to the more sensitive method of determining the 
end point of reduction, has a much greater degree of accuracy. 

The Violette method has been modified by Spencer,* so as to in- 
clude the ferrocyanide test for unreduced copper. Some chemists have 
also sought to improve the method by employing larger test tubes and 
using 20 c.c. of the mixed copper solution. The possibilities of modi- 
fication in this direction are of course unlimited and do not require 
special description. 

Pavy's Method. Another volumetric process, using the disap- 
pearance of blue color as end point, is the method of Pavy,f which is 
based upon the fact that when Fehling's solution is reduced in presence 
of ammonia the precipitated cuprous oxide is dissolved as a colorless 
solution, any unreduced copper being indicated by the characteristic 
blue color of the cuprammonium compounds. The disturbing influence 
of the precipitated cuprous oxide upon the color of the solution is thus 
avoided and, in the absence of air, the change from blue to colorless at 
the end point becomes quite sharp. 

Pavy's copper solution is prepared as follows: 34.65 gms. 
CuS0 4 .5H 2 0, 170 gms. Rochelle salts and 170 gms. potassium hy- 
droxide are dissolved in water to 1000 c.c. It is preferable, however, 
as in Soxhlet's method to make up the copper and alkali-tartrate solu- 

* Spencer's "Handbook for Cane Sugar Manufacturers" (1906), p. 131. 
t Pavy's "Physiology of the Carbohydrates" (London, 1894), p. 71. 



396 



SUGAR ANALYSIS 



tions separately to 500 c.c., and to mix equal quantities of the two just 
before using; 120 c.c. of the mixed copper solution are transferred to a 
liter flask, 300 c.c. of ammonia of specific gravity 0.880 are added and 
the volume completed to 1000 c.c. ; 20 c.c. of the ammoniacal Fehling's 
solution are reduced by 0.01 gm. glucose. 

The reduction is carried out in a flask of about 150 c.c. capacity, 
provided with^a two-hole stopper, one opening of which is connected 
with the tip of the burette containing the sugar solution and the other 
with a bent glass tube for the escape of air and steam (Fig. 168). 




Fig. 168. Pavy's method for determining reducing sugars. 

Forty cubic centimeters of the ammoniacal copper solution are placed 
in the flask, and after inserting the stopper the solution is brought to a 
gentle boil. The sugar solution is then added at the rate of 60 to 100 
drops per minute, the discharge being regulated by a Pavy pinch cock 



REDUCTION METHODS FOR DETERMINING SUGARS 397 

(C); the ebullition must be maintained without interruption. When 
the blue color begins to lighten, the sugar solution is added drop by 
drop until the last trace of color is just discharged. The end point is 
made more sensitive by looking through the solution against a white 
plate (P). 

The reduction must be made in complete absence of air, otherwise 
the dissolved cuprous oxide will be reoxidized. A precaution sometimes 
used to prevent the entrance of air, through momentary cooling, is to 
use a bent-glass exit tube, fitted with a rubber valve, dipping into a 
beaker of water. Care must also be taken not to prolong the time of 
reduction, otherwise all the ammonia will be expelled and the cuprous 
oxide not be dissolved. 

In Pavy's method 1 molecule of glucose reduces 6 molecules of 
cupric oxide instead of 5 molecules as by Fehling's solution. These 
proportions vary somewhat, however, according to concentration and 
other conditions of experiment. The solution should, therefore, be 
standardized against glucose or invert sugar following the exact method 
employed in analysis. 

Pavy's method gives good results, when the reduction is carried out 
with complete exclusion of the air. The extra precautions necessary 
for making the determination, and the failure of the method to give 
good results with colored solutions, have prevented the process from 
becoming generally employed. 

Conversion Tables for Volumetric Determination of Sugars. - 
The calculation of reducing sugars by any of the volumetric methods 
is much simplified by the use of appropriate conversion tables. If a 
volume of Fehling's solution be taken, which always corresponds to a 
fixed amount of reducing sugar, as, for example, 0.5 gm. in Table LXX, 
and the sugar solution for titration be made up so as to contain this 
same amount of substance (as 0.5 gm.) in 1 c.c., then the formula for 
determining reducing sugar becomes 

_ 0.5 X 100 _ 100 
: 0.5 XV V 

in which R is the per cent of reducing sugar in the substance and V the 
cubic centimeters of sugar solution necessary for the reduction. 

If the substance be very high or very low in reducing sugar, an even 
fraction or multiple of 0.5 gm. may be taken for the amount of sub- 
stance to be dissolved in 1 c.c. Thus for 0.05 gm. of substance in 1 c.c. 

R = -' and for 1 gm. of substance in 1 c.c. R = * 



398 



SUGAR ANALYSIS 



Under the above conditions of analysis a table giving different 
multiples of the reciprocals of the burette readings will give the cor- 
responding percentages of reducing sugars. The following example 
will illustrate the method for constructing such a table. 

Fehling's solution taken = 0.2 gram of reducing sugar 



Volume of 
sugar solution 
for reduction. 


Reciprocal. 


Weight of substance in 1 c.c. of sugar solution. 


0.40 gm. 


0.20 gm. 


0.10 gm. 


0.04-gm. 


0.02 gm. 


V 


1 
V 


50 

V 


100 

'T 


200 
V 


500 
V 


1000 
V 


c.c. 
20.0 
20.1 
20.2 
20.3 
20.4 

30 io 
40.0 
50.0 


0.05000 
0.04975 
0.04950 
0.04926 
0.04902 


Per cent. 
2.50 

2.49 
2.48 
2.46 
2.45 

i'.G7 
1.25 
1.00 


Per cent. 
5.00 

4.98 
4.95 
4.93 
4.90 

3~33 
2.50 
2.00 


Per cent. 
10.00 
9.95 

9.90 
9.85 
9.80 

6^67 
5.00 
4.00 


Per cent. 

25.00 
24.88 
24.75 
24.63 
24.51 


Per cent. 
50.00 

49.75 
49.50 
49.26 
49.02 

33^33 
25.00 
20.00 


0.03333 
0.02500 
0.02000 


16.67 
12.50 
10.00 



The table can of course be modified in a great variety of ways to 
suit individual requirements. A list of reciprocals for assistance in cal- 
culating such a table is given in the Appendix (Table 25). 

Reischauer and Kruis's Method. In the methods previously 
described a constant volume of Fehling's solution was taken and the 
amount of sugar solution noted necessary to complete the reduction. 
In a process first proposed by Lippmann * and elaborated by Reischauer 
and Kruis f the opposite procedure is followed. A constant volume of 
sugar solution is taken and the amount of Fehling's solution determined 
necessary to oxidize the reducing sugar. 

In the Reischauer-Kruis method the sugar solution is made up so 
as not to contain over 0.58 gm. glucose in 100 c.c. Six numbered test 
tubes holding from 20 to 30 c.c. are taken and 5 c.c. of the sugar 
solution measured into each; 1, 2, 3, 4, 5 and 6 c.c. respectively of 
Fehling's solution are then added to the different tubes, which are 
afterwards shaken and immersed in boiling water for 20 minutes. At 
the end of this time the tubes are examined and the two tubes noted in 
which reduction is just completed and in which the least amount of 
unreduced copper is left. Having noted the limits between which the 
true copper equivalent lies, the volume of Fehling's solution is varied 

* Oester. Ungar. Z. Zuckerind., 7, 256. 
t Oester. Ungar. Z. Zuckerind., 12, 254. 



REDUCTION METHODS FOR DETERMINING SUGARS 399 

within this interval until the exact amount necessary for oxidizing all 
the reducing sugar is found. 

The pipettes employed for this method are graduated in their 
lower part from 1 c.c. to 5 c.c. and in the stem contain an extra 1 c.c. 
graduated into hundredths. With three trials and employment of the 
ferrocyanide test, the volume of Fehling's solution can be determined 
to 0.01 c.c. The following example illustrates the application of the 
method. 



First trial. 


Second trial. 


Third trial. 


1 c.c. Cu 


all reduced 


4 


15 c.c. Cu 


all reduced 


4 


32 c.c. Cu 


all 


reduced 


2 ' 




it 


(4.30 




S( 


4 


34 


t 







3 




" 


14 


45 


Cu in solution 


J4 


36 


1 


< 


' 


j 4 ' 




11 


4 


60 




11 


14 


38 


' Cu 


in solution 


I 5 ' 


Cu 


in solution 


4 


75 




" 


4 


40 


' 


* 


' 


6 ' 


n 


4 


90 






4 


42 ' 







The quantity of Fehling's solution which exactly oxidizes the reducing 
sugar in the 5 c.c. of solution may, therefore, be placed at 4.37 c.c. 

The amount of glucose corresponding to each 0.01 c.c. between 
1 c.c. and 6 c.c. of Fehling's solution is found from a table calculated 
by Kruis (Appendix, Table 9). 

The Reischauer-Kruis method possesses certain advantages over the 
methods previously described in point of exactness; the error due to 
variation in reducing power with changes in concentration is avoided, 
the amount of reducing sugar in 5 c.c. corresponding to different volumes 
of Fehling's solution being definitely known for the conditions of ex- 
periment. The large amount of labor and time necessary for com- 
pleting a determination has been, however, a serious obstacle against 
the general use of the method. 



METHODS BASED UPON A GRAVIMETRIC OR VOLUMETRIC DETERMINATION 

OF REDUCED COPPER 

In the methods of this class an excess of copper is present in the 
Fehling's solution at the end of reduction. The precipitated cuprous 
oxide after a fixed period of heating is filtered off, and the amount of 
copper determined by any of the numerous gravimetric or volumetric 
processes. The weight of reducing sugar corresponding to a definite 
weight of precipitated copper is then determined by means of formulae 
or tables which have been calculated from results obtained upon known 
amounts of pure sugar under similar conditions of experiment. 



400 



SUGAR ANALYSIS 



Variability in Reducing Power of Monosaccharides. Soxhlet* 
showed that when a solution of glucose acted upon Fehling's solution 
the first portion added reduced most strongly and the succeeding por- 
tions gradually less so. This variability in reducing power is found to 
be different, however, for the monosaccharides, glucose, fructose, invert 
sugar, galactose, etc., than for the disaccharides, lactose and maltose. 

As examples of the variability in reducing power of monosaccharides 
the following results are given. The values, which were calculated 
from Bertrand's sugar tables, represent the milligrams of copper re- 
duced by each succeeding 10-milligram portion of added sugar. 

TABLE LXXI 

Shouting variability in reducing power of monosaccharides 



Number of series. 


Invert sugar. 
Milligrams 
copper. 


Glucose. 
Milligrams 
copper. 


Galactose. 
Milligrams 
copper. 


First 10 m 
Second 10 
Third 10 
Fourth 10 
Fifth 10 
Sixth 10 
Seventh 10 
Eighth 10 
Ninth 10 
Tenth 10 


gS. Of SUj 

< 
< 


;ar red 

< - 
i 


uce 


20.6 

19.8 
18.9 
18.4 
17.7 
17.2 
16.6 
16.1 
15.8 
15.4 


20.4 
19.7 
19.0 

18.4 
17.9 
17.4 
17.0 
16.3 
15.9 
15.8 


19.3 

18.6 
18.3 
17.7 
17.3 
16.9 
16.7 
16.3 
16.3 
16.0 















It is seen that each succeeding 10 mgs. of added glucose undergoes a 
loss in reducing power of about 3 per cent. 

Law of Reducing Action. The reducing action of a monosac- 
charide upon Fehling's solution may be expressed in general terms as 
follows : 

If for the first minute quantity s of a given sugar a definite amount 
c of copper is reduced, then for any multiple n of s the weight of copper 
would be nc, if the same amount of copper in the Fehling's solution were 
always maintained. The latter condition, however, is never realized 
in practice, and with the continuous removal of copper from solution 
the value nc becomes nc (n 1 -{- n 2 + n 3 + n ri)k. 
When working with weighable quantities of sugar, this expression 
should be modified to c + (n l)d (n 2 + n 3 + n ri)k 
in which d is the difference between the weights of copper for the first 
two members of the series s and 2s. The values of d and of the constant 

* J. prakt. Chem. [2], 21, 227; 



REDUCTION METHODS FOR DETERMINING SUGARS 401 

k are easily determined empirically, and knowing these it is possible to 
construct tables for any of the reducing sugars. 

As an example of this method of calculation the following values are taken 
from the experimental work of Allihn : * 

No. of series (ra) . 

1 10 mgs. of glucose reduce 18.0 mgs. copper 

2 20 mgs. of glucose reduce 38.2 mgs. copper 

25 250 mgs. of glucose reduce 463.0 mgs. copper. 

18.0 = c. 

38.2 - 18.0 = 20.2 = d. 

Substituting the above values for c and d in the equation for n = 25, 
18 + (25 - 1) 20.2 - (25 - 2 + 25 - 3 . . . ) k = 463.0 

whence k = 0.14. 

The equation 18 + (n - 1) 20.2 -(n-2 + n-3+- n n) 0.14 will 
give the milligrams of copper reduced by any multiple n of 10 mgs. of glucose 
under the conditions of Allihn's experiments. 

Suppose it is required to find the milligrams of copper reduced by 100 mgs. 
of glucose. 

18 + (10 - 1) 20.2 - (10 - 2 + -10 - 3 . . . ) 0.14 = 194.8 mgs. Cu. 
Allihn obtained by actual experiment 195 mgs. of copper by the reducing 
action of 100 mgs. of glucose. 

Calculation of Reduction Tables. The calculation of tables for 
the copper-reducing power of different sugars is usually made by the 
method of least squares, according to the general formula: 

y = A + Bx + Cx\ 

in which x is the milligrams of copper reduced by y milligrams of sugar 
and A, B and C constants. Having determined by experiment the 
values of x for 10 or more values of y, the calculation of A, B and C is 
made in the same manner as described on page 175. 

As an example of the method of least squares the work of Allihn is again 
quoted. Allihn found that different amounts of glucose under constant con- 
ditions of experiment reduced the following amounts of copper. 



Mgs. of glucose (y) . . . 
Mgs. of copper (x) 


10.0 
18.0 


20.0 

38.2 


25.0 
47.5 


50.0 
99.0 


100.0 
195.0 


125.0 
242.5 


150.0 

287.7 


175.0200.0 
333.0377.7 


225.0 
421.2 


250.0 
463.0 



Substitution of the above values for x and y in the formula y = A + Bx + Cx 2 
gives the general equation 

y = - 2.5647 -f 2.0522 x - 0.0007576 x z , 

by means of which Allihn constructed his table giving the milligrams of glucose 

corresponding to any weight of reduced copper between 10 mgs. and 463 mgs. 

* J. prakt. Chem. [2], 22, 46. 



402 



SUGAR ANALYSIS 



Variability in Reducing Power of Disaccharides. The variability 
in reducing power of maltose and lactose is different from that noted for 
the monosaccharides. According to the amount of free alkali, time of 
boiling and other conditions, succeeding portions of maltose and lactose, 
while usually showing a slight loss, may show either no change at all, or 
even a slight gain in reducing power over preceding portions of the same 
sugar. This peculiarity of maltose and lactose is explained by a slight 
hydrolysis of the sugar into monosaccharides of higher reducing power. 
A slight inversion of this kind takes place with sucrose, which is strictly 
speaking a non-reducing sugar, and it no doubt occurs to a greater or 
less extent with all higher saccharides upon boiling with Fehling's solu- 
tion. 

As an illustration of the reducing power of successive portions of 
maltose, the following results are taken from the tables of Wein and of 
Munson and Walker. 

TABLE LXXII 
Showing variability in reducing power of maltose 



Number of Series. 


Wein. 


Munson and 
Walker. 


First 30 m 
Second 30 
Third 30 
Fourth 30 
Fifth 30 
Sixth 30 
Seventh 30 


gs. of mal 


^ose rec 


uce 


Mgs. Cu. 
35.4 
34.5 
34.0 
33.4 
33.4 
33.8 
33.5 


Mgs. Cu. 

35.9 
33.6 
33.5 
33.8 
33.6 
33.7 
33.6 













It is seen that in both series of experiments there is at first a marked 
decrease and then a slight increase in the reducing power of the suc- 
cessive portions of added sugar. Changes of a similar nature are 
noted in some of the tables for lactose. 

The reducing power of the disaccharides upon Fehling's solution is 
much more subject to change with difference in conditions than the 
monosaccharides. Kjeldahl,* for example, found that increasing the 
amount of alkali in Fehling's solution caused the reducing power of 
maltose and lactose to gain with ten times the rate of increase noted for 
glucose. The same effect is also produced by prolonging the time of 
boiling. This greater sensibility of the disaccharides to disturbing in- 
fluences during reduction involves a greater experimental error in the 
determination when the details of the method are not carefully followed. 
* Neue Z. Riibenzuckerind., 37, 13, 23. 



REDUCTION METHODS FOR DETERMINING SUGARS 403 

Methods and tables for estimating different sugars from the amount 
of copper reduced from Fehling's solution have been devised by Soxhlet; 
Allihn; Wein; Meissl; Herzfeld; Lehmann; Kjeldahl; Pfluger; Ost; 
Honig and Jesser; Brown, Morris and Millar; Bertrand; Defren; 
Munson and Walker; Kendall; and many others. It is impossible to 
describe all these processes and only a few of the more typical methods 
will be selected. The method of Allihn,* which is one of the widest 
known, illustrates well the various principles involved and will be 
described first in somewhat fuller detail. 

Allihn's Method for the Determination of Glucose. The follow- 
ing details of Allihn's method with the description of several processes 
for determining the amount of reduced copper are taken from the 
Methods of Analysis of the Association of Official Agricultural 
Chemists, f 

PREPARATION OF REAGENTS 

Copper-sulphate Solution. Dissolve 34.639 gms. of CuS04.5H 2 in 
water and dilute to 500 c.c. 

Alkaline-tartrate Solution. Dissolve 173 gms. of Rochelle salts and 
125 gms. of potassium hydroxide in water and dilute to 500 c.c. 

DESCRIPTION OF METHOD 

Place 30 c.c. of the copper solution, 30 c.c. of the alkaline-tartrate 
solution and 60 c.c. of water in a beaker and heat to boiling. Add 
25 c.c. of the solution of the material to be examined, which must be so 
prepared as not to contain more than 0.250 gm. of glucose, and boil for 
exactly two minutes keeping the beaker covered. Filter immediately 
through asbestos without diluting, and obtain the weight of copper by 
one of the methods described in the following section. The correspond- 
ing weight of glucose is found from Allihn's table (Appendix, Table 10). 

METHODS FOR DETERMINING REDUCED COPPER 

Reduction of the Cuprous Oxide in Hydrogen.| "Filter the 
cuprous oxide immediately through a weighed filtering tube made of 
hard glass, using suction. Support the asbestos film in the filtering tube 
with a perforated disk or cone of platinum, and wash free from loose 
fibers before weighing; moisten previous to the filtration. Provide the 
tube with a detachable funnel during filtration, so that none of the 
precipitate accumulates near the top, where it could be removed by 

* J. prakt. Chem. [2], 22, 46. 

t Bull. 107 (revised), U. S. Bur. of Chem., pp. 49-53. 

t Ibid. 



404 



SUGAR ANALYSIS 



i ii in 

Fig. 169. Forms of tubes for filtering cuprous oxide. 




Fig. 170. Showing methods of filtering cuprous oxide with filter tube or 
Gooch crucible. 



REDUCTION METHODS FOR DETERMINING SUGARS 405 

the cork used during the reduction of the cuprous oxide. Transfer all the 
precipitate to the filter and thoroughly wash with hot water, following 
the water by alcohol and ether successively. After being dried, con- 
nect the tube with an apparatus for supplying a continuous current of 
dry hydrogen, gently heat until the cuprous oxide is completely re- 
duced to the metallic state, cool in the current of hydrogen and weigh." 
Several forms of tubes for filtering cuprous oxide are shown in 
Fig. 169. Glass wool is sometimes used in place of a platinum disk for 
holding the asbestos, but makes a less resistant support (see Fig. 169 III). 




Fig. 171. Apparatus for reducing cuprous oxide to copper. A, hydrogen generator; 
B and C, gas driers; D, filter tube containing cuprous oxide. 

A convenient method of filtering cuprous oxide by means of suction 
is shown in Fig. 170. A continuous filtration should be maintained 
and all the precipitate should be transferred to the tube before the 
liquid above the asbestos is allowed to run completely through. Too 
rapid or too irregular filtration may cause particles of cuprous oxide to 
run through the asbestos. A fine jet of water will usually bring all the 
cuprous oxide into the filter tube; should any of the precipitate remain 
adhering to the beaker a feather, or a rubber-tipped rod, will assist the 
removal. 

The reduction of the cuprous oxide to copper by means of hydrogen 
is shown in Fig. 171. All air must be expelled from the tube before 



406 



SUGAR ANALYSIS 







Fig. 172. Desiccator for 
filter tubes. 



heating, otherwise there is danger of explosion. The heating should be 
continued until all water is expelled from the tube. A desiccator of the 
form shown in Fig. 172 is convenient for hold- 
ing filter tubes before weighing. 

The asbestos used for loading the filter tubes 
should be of a kind which is not attacked by 
hot Fehling's solution. The following method 
of preparation used by Munson and Walker * 
is recommended. 

Preparation of Asbestos. Prepare the 
asbestos which should be the am phi bole 
variety by first digesting with 1 : 3 hydro- 
chloric acid for two or three days. Wash free 
from acid and digest for a similar period with 
soda solution, after which treat for a few hours 
with hot alkaline copper-tartrate solution of 
the strength employed in sugar determinations. 
Then wash the asbestos free from alkali, finally 
digest with nitric acid for several hours, and 
after washing free from acid shake with water for use. In preparing 
filter tubes or Gooch crucibles load with a film of asbestos one-fourth 
inch thick, wash this thoroughly with water to remove fine particles of 
asbestos; finally wash with alcohol and ether, dry for 30 minutes at 
100 C., cool in a desiccator and weigh. It is best to dissolve the cop- 
per with nitric acid each time after weighing and use the same felts over 
and over again, as they improve with use. 

The method of estimating copper by reduction of the precipitated 
cuprous oxide, although not so exact as the electrolytic method, is 
nevertheless sufficiently accurate for most purposes of analysis. In the 
case of impure sugar products the cuprous oxide is frequently con- 
taminated with mineral or organic matter, and in such cases the 
method gives too high results. 

Determination of Reduced Copper by Electrolysis. Deposition 
from Sulphuric-acid Solution.^ Filter the cuprous oxide in a Gooch 
crucible (as shown in Fig. 1 70) , wash the beaker and precipitate thoroughly 
with hot water without any effort to transfer the precipitate to the filter. 
Wash the asbestos film and the adhering cuprous oxide into the beaker 
by means of hot dilute nitric acid. After the copper is all in solution, 
refilter through a thin film of asbestos and wash thoroughly with hot 

* J. Am. Chem. Soc., 28, 666. 

t Bull. 107 (revised), U. S. Bur. of Chem., pp. 49-53. 



REDUCTION METHODS FOR DETERMINING SUGARS 407 

water. Add 10 c.c. of dilute sulphuric acid, containing 200 c.c. of sul- 
phuric acid (sp. gr. 1.84) per liter, and evaporate the filtrate on the 
steam bath until the copper salt has largely crystallized. Heat care- 
fully on a hot plate or over a piece of asbestos board until the evolution 
of white fumes shows that the excess of nitric acid is removed. Add 
from 8 to 10 drops of nitric acid (sp. gr. 1.42) and rinse into a platinum 
dish of from 100 to 125 c.c. capacity. Precipitate the copper by elec- 
trolysis. Wash thoroughly with water, alcohol and ether successively, 
dry at about 50 C. and weigh. If preferred the electrolysis can be 
conducted in a beaker, the copper being deposited upon a weighed 
platinum cylinder. 

Deposition from Sulphuric- and Nitric-acid Solution.* Filter and 
wash as previously described. Transfer the asbestos film from the 
crucible to the beaker by means of a glass rod and rinse the crucible 
with about 30 c.c. of a boiling mixture of dilute sulphuric and nitric 
acids, containing 65 c.c. of sulphuric acid (sp. gr. 1.84) and 50 c.c. of 
nitric acid (sp. gr. 1.42) per liter. Heat and agitate until solution is 
completed; filter and electrolyze. 

Deposition from Nitric-acid Solution.^ Filter and wash as pre- 
viously described. Transfer the asbestos film and adhering oxide to 
the beaker. Dissolve the oxide still remaining in the crucible by 
means of 2 c.c. of nitric acid (sp. gr. 1.42), adding it with a pipette 
and receiving the solution in the beaker containing the asbestos film. 
Rinse 'the contents of the beaker until the copper is all in solution, 
filter, dilute the filtrate to a volume of 100 c.c. or more and electrolyze. 
When a nitrate solution is electrolyzed, the first washing of the deposit 
should be made with water acidulated with sulphuric acid in order that 
the nitric acid may all be removed before the current is interrupted. 

Leach's Electrolytic Apparatus. A convenient apparatus for 
the electrolytic deposition of copper in sugar analysis is that of Leach 
shown in Fig. 173. A is a hard rubber plate 50 cm. long and 25 cm. 
wide provided with four insulated metal binding posts B, each carry- 
ing at the top a thumb screw by which a coiled-platinum-wire electrode 
may be attached. In front of each post is a copper plate about 4 cm. 
square covered with thin platinum foil P, which is bent around the 
edges of the copper plate and so held in place, the copper plate being 
screwed to the rubber from beneath. On the square platinum-covered 
plate is set the platinum evaporating dish which holds the solution 

* Bull. 107 (revised), U.S. Bur. of Chem., pp. 49-53. 

t Ibid. 

j Leach's "Food Inspection and Analysis" (1911), p. 608. 



408 



SUGAR ANALYSIS 



from which the copper is to be deposited, the inside of the dish forming 
the cathode, while the coiled platinum wire, dipping below the surface 
of the solution, forms the anode. In front of each platinum-covered 
plate is a switch S, and at either end of the hard-rubber plate is a bind- 
ing post R, for connection with the electric current. 




Fig. 173. Leach's electrolytic apparatus for determining reduced copper. 

, Four determinations may be carried on simultaneously in four 
platinum dishes, if desired, the wiring and the switches being so ar- 
ranged that beginning at one end of the plate either the first dish, or 
the first two or the first three, may be thrown in or out of the circuit at 
will without interrupting the current through the remaining dishes. A 
cover with wooden sides and glass top fits closely over the whole ap- 
paratus as a protection from dust, but may easily be lifted off to manipu- 
late the dishes when desired. The sides of the cover are perforated to 
permit the escape of the gas formed during the electrolysis. 

The ordinary street current is used when available, and the strength 
of the current may be varied within wide limits by means of a number 
of 16- or 32-candle-power lamps K, coupled in multiple, and a rheostat 



REDUCTION METHODS FOR DETERMINING SUGARS 409 

L, consisting of a vertical glass tube sealed at the bottom, containing a 
column of dilute acid, the resistance being changed by varying the 
length of the acid column contained between the two platinum ter- 
minals immersed therein, one of which is movable. A gravity battery 
of four cells may be employed if the laboratory is not equipped with 
electric lights. 

In using the apparatus the plating process should go on till all the 
copper is deposited, which requires several hours or over night with a 
current of about 0.25 ampere. Before stopping the process the absence 
of copper in the solution should be proved by removing a few drops 
with a pipette, adding first ammonia, then acetic acid and testing with 
ferrocyanide of potassium. If no brown coloration is produced, all the 
copper has been plated out. Throw the dish out of circuit by means of 
the switch, pour out the acid solution quickly before it has a chance to 
dissolve any of the copper, wash the dish first with water and then with 
alcohol, dry and weigh. 

The copper may be removed from the platinum dish by strong 
nitric acid. 

The electrolytic process for determining reduced copper is the most 
exact of all methods. The determination, however, involves a con- 
siderable expenditure of time and for this reason is but little used in 
sugar laboratories where there is a large amount of routine. 

Electrolytic Method of Peters. Peters * has devised a rapid elec- 
trolytic method for the determination of copper, whereby the metal 
is deposited from an alkaline-tartrate solution, such as is used in 
preparing Fehling's solution. The electrolysis is carried out either 
in platinum dishes placed upon plates of sheet brass to which the 
cathode connection is made, or in glass beakers or large test tubes, 
in which case large cylindrical strips of sheet copper may be used for 
the cathode. The anode consists of a flat or cylindrical spiral of 
platinum wire, which should be placed at a distance of 1 cm. or less 
from the cathode surface. A volume of 10 c.c. copper solution 
(which may be slightly acid or alkaline) is usually taken, to which is 
added an approximately equal volume of a solution containing 35 gms. 
pure Rochelle salts and 25 gms. potassium hydroxide (purified by 
alcohol) in 100 c.c. For copper solutions containing free sulphuric 
or nitric acid, two volumes of the alkaline-tartrate solution may be 
used. From 0.4 to 1.0 c.c. of a saturated aqueous potassium-cyanide 
solution is then added according to the amount of copper present; 
the amount of cyanide solution should be less than sufficient to dis- 
* J. Am. Chem. Soc., 34, 426. 



410 SUGAR ANALYSIS 

charge the blue color. If the copper deposit should be found to be 
too soft or dark colored, more cyanide should be used; an excess of 
the latter, however, greatly lengthens the time for complete deposition 
of the copper. 

In making the determination the direct 110- volt current of a light- 
ing system is used with three 32-candle-power lamps interposed as 
resistance; under these conditions the voltage measures 2.6 and the 
amperage 2.85. During the electrolysis the solution is warmed by a 
small burner placed under the brass plate to one side of the cathode 
vessel; if test tubes are used they are placed upon wire gauze over a 
small flame. The evolution of gas and the circulation of warm 
liquid cause a very rapid deposition of copper, which is usually com- 
plete in less than 30 minutes. The solution should be covered during 
electrolysis to prevent loss by spraying. 

To determine the completion of electrolysis, Peters recommends 
the Endemann-Prochazka * hydrobromic acid test. One volume of 
concentrated sulphuric acid is diluted with 2 to 3 volumes of distilled 
water. About 1 c.c. of the dilute acid is placed in a narrow test tube, 
a few crystals of potassium bromide added and the whole heated to 
boiling. A drop of the solution to be tested is then added; as small 
an amount as 0.007 mg. copper will cause a red color to develop. 

If the deposition of copper is complete, the solution in the cathode 
vessel, without breaking the current, is displaced by a small stream 
of water until the resistance lamps are extinguished; under this pro- 
cedure no copper is lost by solution. The electrode containing the 
deposit of copper is then washed in alcohol and ether, dried and 
weighed. 

On account of the similarity in composition of the electrolyte 
employed by Peters to that of the alkaline-tartrate solution used in 
Allihn's method, the process recommends itself for the determination 
of copper in the original Fehling's solution or in the filtrate from the 
reduced cuprous oxide obtained in the analysis of sugar solutions. 

Several volumetric processes have been devised for determining 
copper in the precipitate of cuprous oxide. Of these the permanganate, 
the iodide and thiocyanate methods will be described. 

Volumetric Permanganate Method, f Filter and wash the cuprous 
oxide as in the previous methods. Transfer the asbestos film to the 
beaker, add about 30 c.c. of hot water and heat the precipitate and 
asbestos thoroughly. Rinse the crucible with 50 c.c. of a hot saturated 

* Chem. News, 42, 8. 

t Bull. 107 (revised), U. S. Bur. of Chem., pp. 49-53. 






REDUCTION METHODS FOR DETERMINING SUGARS 411 

solution of ferric sulphate in 20 per cent sulphuric acid, receiving the 
rinsings in the beaker containing the precipitate. After the cuprous 
oxide is dissolved, wash the solution into a large Erlenmeyer flask and 
immediately titrate with a standard solution of potassium permanga- 
nate. One cubic centimeter of the permanganate solution should equal 
0.010 gm. of copper. In order to standardize the permanganate solu- 
tion, make six or more determinations with the same sugar solution, 
titrating one-half of the precipitations and determining the copper in 
the others by electrolysis. The average weight of copper obtained by 
electrolysis, divided by the average number of cubic centimeters of 
permanganate solution required for the titration, is equal to the weight 
of copper equivalent to 1 c.c. of the standard permanganate solution. 

The reaction between the ferric sulphate and cuprous oxide is ex- 
pressed as follows: 

Fe 2 (SO 4 ) 3 + Cu 2 + H 2 S0 4 = 2 FeS0 4 + 2 CuSO 4 + H 2 0. 

Since 1 atom, or 16 parts, of is required to oxidize the iron reduced 
by 2 atoms, or 127.2 parts, of Cu, and 1 c.c. of n/10 permanganate 
contains 0.0008 gm. of active 0, then 1 c.c. of n/10 permanganate is 
equivalent to 0.00636 gm. Cu. For a solution containing 5 gms. of 
potassium permanganate to the liter, 1 c.c. will be equivalent very 
closely to 0.01 gm. of copper. Owing to slight deviations in practice 
from the above theoretical equation, the copper value of the perman- 
ganate must always be determined by direct experiment. 

Volumetric Iodide Method,* Low's Modification.^ Standardization 
of the Thiosulphate Solution. Prepare a solution of sodium thiosul- 
phate containing 19 gms. of pure crystals to 1000 c.c. Weigh accu- 
rately about 0.2 gm. of pure copper foil and place in a flask of 250 c.c. 
capacity. Dissolve by warming with 5 c.c. of a mixture of equal 
volumes of strong nitric acid and water. Dilute to 50 c.c., boil to expel 
the red fumes, add 5 c.c. strong bromine water and boil until the 
bromine is thoroughly expelled. Remove from the heat and add a 
slight excess of strong ammonium hydroxide (about 7 c.c. of 0.90 sp. gr.). 
Again boil until the excess of ammonia is expelled, as shown by a change 
of color of the liquid, and a partial precipitation. Now add a slight ex- 
cess of strong acetic acid (3 or 4 c.c. of 80 per cent acid) and boil again 
for a minute to redissolve the copper. Cool to room temperature and 
add 10 c.c. of a solution of pure potassium iodide containing 300 gms. 

* For a critical study of the iodide method for determining copper in sugar 
analysis see paper by Peters, J. Am. Chem. Soc., 34, 422. 
f J. Am. Chem. Soc., 24, 1082. 



412 SUGAR ANALYSIS 

of potassium iodide to 1000 c.c. Titrate at once with the thiosulphate 
solution until the brown tinge has become weak, then add sufficient 
starch liquor to produce a marked blue coloration. Continue the ti- 
tration cautiously until the color due to free iodine has entirely van- 
ished. The blue color changes toward the end to a faint lilac. If at 
this point the thiosulphate be added drop by drop and a little time be 
allowed for complete reaction after each addition there is no difficulty 
in determining the end point within a single drop. One cubic centi- 
meter of the thiosulphate solution will be found to correspond to about 
0.005 gm. of copper. 

Determination of Copper. After washing the precipitated cuprous 
oxide, cover the Gooch crucible with a watch glass and dissolve the 
oxide by means of 5 c.c. of warm nitric acid (1 : 1), poured under the 
watch glass with a pipette. Catch the filtrate in a flask of 250 c.c. 
capacity, wash watch glass and crucible free of copper; 50 c.c. of 
water will be sufficient. Boil to expel red fumes, add 5 c.c. of bromine 
water, boil off the bromine and proceed exactly as in standardizing the 
thiosulphate. 

In a later modification of the above method, Low has found it 
possible to dispense with the use of bromine, the nitrous acid being 
expelled from the copper solution by boiling, adding ammonia, heat- 
ing, acidifying with acetic acid and again boiling. 

The reaction between the copper acetate and potassium iodide is 
expressed as follows: 

2 Cu(C 2 H 3 O 2 ) 2 + 4 KI = Cu 2 I 2 + 4 KC 2 H 3 2 + I 2 . 

Since 1 atom, or 63.57 parts, of copper liberates 1 atom, or 126.92 
parts, of iodine and 1 c.c. of n/10 thiosulphate solution (24.8 gms. Na 2 S 2 Oa 
+ 5 H 2 to 1000 c.c.) reacts with 0.01269 gm. I,then 1 c.c. n/W thiosul- 
phate corresponds to 0.00636 gm. Cu. For a solution containing 19.5 
gms. of pure sodium thiosulphate to the liter, 1 c.c. will be equivalent 
very closely to 0.005 gm. of copper. In actual practice the above 
reaction does not proceed with absolutely quantitative precision, the 
results of the determination varying somewhat according to concen- 
tration of acid, excess of reagents, temperature and other conditions. 
It is, therefore, important always to standardize the thiosulphate 
solution against pure copper under the exact conditions which are 
followed in analysis. 

Kendall's Modification of the Iodide Method. The removal of the 
nitrous acid, formed in dissolving the copper, is the chief difficulty in 
the iodide method. Kendall * has modified the method by removing 

* J. Am. Chem. Soc., 33, 1947. 



REDUCTION METHODS FOR DETERMINING SUGARS 413 

the nitrous acid with hypochlorite, the free chlorine, which is evolved, 
being afterwards removed with phenol. 

The cuprous oxide, after filtering and washing upon a Gooch cru- 
cible, is dissolved in 10 to 15 c.c. of 30 per cent nitric acid into a 
300 c.c. Erlenmeyer flask. The volume of solution and washings 
should be between 50 and 60 c.c. with an acidity of 4 to 5 c.c. con- 
centrated nitric acid; 5 c.c. of sodium hypochlorite solution are then 
added of such strength that the. iodine liberated by 5 c.c. is equivalent 
to 30 c.c. of n/10 thiosulphate. The solution is allowed to stand 
2 minutes, when 10 c.c. of a 5 per cent colorless phenol solution are 
quickly added. The chlorine gas above the liquid is removed by 
blowing into the flask and the sides are washed down with a jet of 
water. The solution is then made slightly alkaline with sodium 
hydroxide and acidified with acetic acid; 10 c.c. of 30 per cent potas- 
sium iodide solution are then added and the free iodine titrated with 
standard sodium thiosulphate, as under Low's modification, using sol- 
uble starch as indicator. The thiosulphate is previously standardized 
against pure copper under the same conditions as those of the method. 

In working with known weights of copper between 20 and 340 
nigs., Kendall found the error of his method to exceed in no case 0.3 mg. 

Peters' s Modifications of the Iodide Method. Peters* has found that 
boiling the nitric-acid solution of copper in the presence of talcum 
powder will remove completely all lower oxides of nitrogen and leave 
the solution, after cooling and diluting, in suitable condition for titra- 
tion. The copper, or its compound, is dissolved in an Erlenmeyer 
flask in the least possible volume of concentrated nitric acid, to which 
one-half its volume of water has been added; 5 to 10 c.c. of dilute 
acid are sufficient for 0.5 gm., or less, of copper. After solution 15 to 
25 c.c. of distilled water and a little pure powdered talcum are added, 
and the mixture boiled vigorously for 5 to 10 minutes. After cooling 
to room temperature distilled water is added and 10 c.c. of a saturated 
potassium-iodide solution, the dilution being so regulated that the 
final volume of liquid at the end of the thiosulphate titration is about 
120 c.c. 

Peters has also employed the iodide method in the determination 
of copper in the alkaline-tartrate solutions, or filtrates, occurring in 
sugar analysis. In the modification employed, 20 c.c. of Allihn's 
alkaline-tartrate solution, 20 c.c. of Fehling's copper-sulphate solution 
and 20 c.c. of water (as in a blank determination), or of the aqueous 
reducing-sugar solution, were taken, making the total volume for 
* J. Am. Chem. Soc., 34, 422. 




414 SUGAR ANALYSIS 

reduction always 60 c.c. After the reduction the cuprous oxide is 
filtered, washed and the nitrate, which has a volume of 70 to 75 c.c., 
acidified with 4 to 5 c.c. of concentrated sulphuric acid. After cool- 
ing to about 20 C., 10 c.c. of saturated potassium iodide are added 
and the solution titrated with standard thiosulphate in the usual 
way. 

The end point of the titration in the iodide method is best deter- 
mined according to Peters by noting the point at which a drop of the 
thiosulphate solution ceases to produce a perceptible white area upon 
the quiet surface of the titration liquid. As in the case of all other 
modifications of the iodide method, the thiosulphate solution must 
be standardized against pure copper under the exact conditions of 
the analysis. 

Potassium iodide is an expensive reagent and where many deter- 
minations of copper are made by this method, the waste titration 
liquids and cuprous iodide precipitates should be saved for recovery 
of the iodine. 

Volumetric Thiocyanate Method (Volhard-Pfluger).* The fol- 
lowing solutions are required: (a) n/W silver-nitrate solution, (6) n/10 
ammonium-thiocyanate solution, (c) a cold saturated solution of sulphur 
dioxide (S0 2 ) in water, (d) nitric acid of sp. gr. 1.2, free from nitrous 
acid, (e) a saturated solution of ferric alum, (/) normal sulphuric-acid 
solution. 

The filter tube, or Gooch crucible, containing the cuprous oxide is 
weighed and the approximate amount of copper determined. The cuprous 
oxide is then dissolved from the asbestos with nitric acid, the solution 
treated with a slight excess of normal sulphuric acid solution (/) necessary 
to convert all the copper into copper sulphate and evaporated to dryness. 
The copper sulphate is then dissolved in water and washed into a 300-c.c. 
graduated flask. Sodium carbonate solution is added to the point of 
turbidity and then 50 c.c. of the sulphurous acid reagent (c). The 
solution is boiled for 1 minute and then n/10 thiocyanate (6) added 
until there is an excess of about 5 c.c. above the calculated amount 
necessary for precipitating the copper as cuprous thiocyanate Cu 2 (SCN) 2 . 
The solution is then cooled, made up to 300 c.c., shaken and filtered 
through dry filter paper. Should the first runnings appear turbid, they 
are returned to the filter; 100 c.c. of the clear filtrate are diluted with 
100 c.c. of water, 50 c.c. of nitric acid (d) and 10 c.c. of ferric-alum solution 
(e) are added, and the solution titrated with n/10 silver nitrate (a) until 
the red color is discharged. The addition of silver solution is continued 

* Pfliiger's Archiv, 69, 423. 



REDUCTION METHODS FOR DETERMINING SUGARS 415 

to the next even number of c.c., and then the solution titrated back 
with n/10 thiocyanate until the white liquid just begins to turn pink. 

Let A be the cubic centimeters of n/10 thiocyanate added to the 
300 c.c. of solution, B the cubic centimeters of n/10 silver nitrate added 
to the 100 c.c. of nitrate, and C the cubic centimeters of n/10 thio- 
cyanate to titrate back excess of B. 

Since 1 c.c. n/10 thiocyanate = 6.357 mgs. copper then the total milli- 
grams of copper (Cu) are found by the formula Cu = 6.357 (A +3 C 3 B) . 
The thiocyanate solution should be standardized against pure copper 
under the conditions of analysis, as in the permanganate and iodide 
methods. 

Volumetric Cyanide Method. Of other volumetric processes which 
are used for determining reduced copper may be mentioned the well- 
known cyanide method. The unreduced copper in the filtrate from the 
cuprous oxide is titrated with standard potassium cyanide solution until 
the blue color disappears. The difference between the copper in the 
volume of Fehling's solution taken, and that found in the filtrate after 
reduction, is the amount of copper reduced by the sugar. 

Determination of Copper by Weighing as Cupric Oxide. In this 
method the cuprous oxide, after collecting upon a Gooch crucible, is 
heated to redness for about 15 minutes, when it is converted to black 
cupric oxide. To insure complete oxidation care must be taken that 
the oxide is not exposed to the reducing action of the illuminating gas 
during ignition. For this reason the operation is best carried out in a 
muffle. 

If porcelain Gooch crucibles are used they should have open bottoms 
with loose perforated disks for supporting the asbestos (CaldwelPs 
crucible, Fig. 174). The one-piece porcelain Gooch 
crucible is liable to crack at high temperatures of 
ignition. 

Finely-divided cupric oxide is hygroscopic and, 
after cooling in a desiccator, should be weighed as 
quickly as possible. The weight of cupric oxide Fig. 174. Gooch cru- 
multiplied by the factor 0.7989 gives the weight of cible with detach- 
metallic copper. Several sugar tables, as KjeldahPs 
and Defren's, express results in terms of cupric oxide, thus avoiding 
the labor of calculation, when this method of determining copper is 
used. 

The method of estimating copper from the weight of cupric oxide 
is one of the most accurate of the indirect methods. With impure pro- 
ducts, however, the precipitate of cuprous oxide frequently carries 




416 



SUGAR ANALYSIS 



down mineral matter and this contamination will impair somewhat the 
accuracy of the method (see Table LXXIII). 

Determination of Copper by Direct Weighing of the Cuprous 

Oxide. In this method the precipitated cuprous oxide is collected 
in a filter tube or Gooch crucible in the usual way. Wash thoroughly 
with hot water, then with 10 c.c. of alcohol and finally with 10 c.c. of 
ether. Dry the precipitate 30 minutes in a water oven at tlie tem- 
perature of boiling water; cool and weigh. The weight of cuprous 
oxide multiplied by 0.8882 gives the weight of metallic copper. The 
sugar tables of Munson and Walker express results in terms of cuprous 
oxide, and the use of these tables will save much labor of calculation 
when this method of determining copper is used. 

Contamination of Cuprous Oxide. Direct weighing of the cuprous 
oxide is the simplest of the gravimetric methods for estimating reduced 
copper in sugar analysis. The process, however, is less accurate than 
the other methods previously described. The method gives good re- 
sults with sugar solutions of high purity, but with impure products the 
cuprous oxide is contaminated with mineral and organic impurities, 
which may affect considerably the accuracy of the determination. 

The extent of the error in estimating copper from the weight of 
cuprous oxide is shown by the following comparative analyses made by 
Sherwood and Wiley * upon a variety of sugar-containing products. 

TABLE LXXIII 
Comparison of Methods for Determining Reduced Copper 







Reduced Copper 




Material. 


From weight 
of cuprous oxide 


From weight 
of cupric oxide. 


Volumetric 
iodide 
method (Low). 


Molasses residuum 


Gram. 

3753 


Gram. 

0.3594 


Gram. 

0.3494 


, 


0.3905 


0.3634 


0.3470 


u 


2517 


0.2348 


0.2242 


n 


3287 


0.3130 


0.3034 


n 


3291 


0.3134 


0.3029 


(i 


. 2768 


0.2698 


0.2688 


(t 


0.2709 


0.2620 


0.2612 


Pure dextrose 


4619 




0.4617 


(i 


2449 




0.2444 


(i u 


1251 




0.1257 


Beer. . . 


0755 




0.0753 


n 


0746 




0.0748 


Molasses .... 


4628 




0.4520 


Corn juice 


3360 




0.3134 


Malt extract.. 


3322 




0.3048 


(i 


3160 




0.2933 


i< 


2093 




0.1934 











* Bun. 105, U. S. Bur. of Chem., p. 120. 



REDUCTION METHODS FOR DETERMINING SUGARS 417 

The results upon the molasses residuum show a contamination of 
the cuprous oxide with organic matter as shown by the differences in 
copper as calculated from the suboxide and oxide, and with mineral 
matter as shown by the differences in copper as calculated from the 
oxide and by the volumetric method. 

With solutions of pure sugar and such liquids as beer, where the 
organic matter consisted largely of carbohydrates, the calculation of 
copper from the weight of cuprous oxide gave accurate results. In the 
case of the malt extracts, which contained added peptones, the precipi- 
tated cuprous oxide seemed to carry down a considerable amount of 
albuminoid matter from solution; in the case of the molasses the precipi- 
tated copper seemed to be in partial combination with certain nitro- 
genous bases such as xanthine. 

Similar comparisons upon methods of determining copper in the 
analysis of cane-sugar products are given in Table LXXX. 

The chemist is usually able to form an opinion of the purity of the 
cuprous oxide from its physical appearance. If the precipitate is yel- 
low or greenish-red in color, or has a flaky appearance, there is evidence 
of contamination, in which case the reduced copper must be determined 
by one of the direct methods. 



CAUSES AFFECTING THE ACCURACY OF ESTIMATING SUGARS FROM A 
DETERMINATION OF REDUCED COPPER 

In addition to the errors in determining reduced copper, there are a 
number of other causes which affect the accuracy of the analytical 
methods belonging to this class. 

Purity of Reagents. A frequent cause of inaccuracy in deter- 
mining sugars by the methods of copper reduction is the presence of 
organic or mineral impurities in the Fehling's solution. The copper 
sulphate, the caustic alkali and especially the Rochelle salts should be 
of the purest quality. The copper sulphate and alkali-tartrate solu- 
tions should be filtered separately through glass wool, or asbestos, and 
the mixed reagent should be perfectly clear and show no trace of cuprous 
oxide after boiling. A blank determination should be made upon each 
fresh lot of solution; the crucibles, or filter tubes, used in the blank 
test should show no increase in weight under the conditions of experi- 
ment followed in analysis. 

Degree of Dilution and Time of Boiling. The effect of varying 
the dilution of Fehling's solution, or the time of boiling, is shown by 
the following comparison of results from Allihn's table with those 



418 



SUGAR ANALYSIS 



obtained by Wein's, and by Koch and Ruhsam's modifications of 
Allihn's method. 





2 minutes' heating. 


30 minutes' heating. 


Reduced cop- 






per. 


Diluted (Allihn). 
Glucose. 


Undiluted (Wein). 
Glucose. 


Diluted (Koch and 
Ruhsam) . Glucose. 


Mgs. 


Mgs. 


Mgs. 


Mgs. 


10 


6.1 


4.5 


4.1 


50 


25.9 


24.6 


21.3 


100 


50.9 


49 9 


46.9 


150 


76.5 


75.5 


72.0 


200 


102.6 


101.7 


96.8 


250 


129.2 


128.3 


122.7 


300 


156.5 


155.6 


149.0 


350 


184 3 


182 3 


176.2 


400 


212.9 


212.0 


205.0 


450 


242.2 


240.6 


235.9 



It is seen that considerably more copper is reduced by using a more 
concentrated Fehling's solution or by heating for a longer time. 

Incomplete reduction of the copper solution has been raised as an 
objection against such methods as Allihn's, which boil for only 2 min- 
utes. If the time of filtration be too prolonged an additional amount 
of copper is sometimes precipitated, thus increasing the results. It is 
important, therefore, with methods which boil for only 2 minutes to 
filter immediately, and as rapidly as permissable, at the end of the time 
limit. 

Atmospheric Pressure and Temperature of Boiling. Variable 
temperature of boiling, due to difference in altitude above sea-level, 
has been suggested by Traphagen and Cobleigh * as a cause of differ- 
ences in determining reducing sugars. Rosenkranz f has recently 
studied the influence of pressure upon the reducing power of invert 
sugar with the following results: 



Pressure. 


Temperature of 
boiling. 


25 c.c. invert sugar solution plus 


25 c.c. water. 
50 c.c. Fehling's 
solution. 


25 c.c. 10% sucrose 
solution. 50 c.c. 
Fehling's solution. 


Millimeters. 
J775 

1600 
J760 
) 925 


Deg. C. 
103-105 

90- 96 
103-104 
109-110 


Mgs. Cu. 

236.5 
232.5 
235.6 
236.1 


Mgs. Cu. 
260.4 

244.9 

277.7 
296.3 



* J. Am. Chem. Soc., 21, 369. f Z. Ver. Deut. Zuckerind., 61 (1911), 426. 



REDUCTION METHODS FOR DETERMINING SUGARS 419 

The results show for pure invert sugar a slight tendency towards 
increase in reducing power with increase in pressure; the error due to 
this cause, however, is slight and may be neglected for ordinary at- 
mospheric conditions. When sucrose is present the increase in pressure 
causes a marked increase in the amount of reduced copper, owing to the 
much greater degree of inversion. 

Surface Area of Solution. The diameter of the vessel in which 
the Fehling's solution is heated has been found to influence the amount 
of reduced copper. With wide beakers, which expose a larger area of 
solution to the air, more cuprous oxide is lost by oxidation (through 
being redissolved in the alkaline-tartrate solution) than in narrower 
beakers. Kjeldahl has eliminated the error due to oxidation by mak- 
ing the reduction in an atmosphere of hydrogen or of oxygen-free illumi- 
nating gas. 

Under the same set of conditions the oxidation error is a constant 
one and the discrepancies due to this cause are eliminated by making 
the reduction always in beakers of the same size. A 350-400-c.c. 
lipped beaker of Jena, or Non-sol glass, 7-8 cm. in diameter is about 
the proper size. 

SPECIAL COPPER-REDUCTION METHODS 

Modifications of Allihn's Method. Allihn's method gives the 
most accurate results upon sugar solutions containing 0.4 to 1.0 per 
cent glucose, i.e., 0.10 to 0.25 gm. glucose in the 25 c.c. of solution. 
When less than 50 mgs. of glucose are present the method is apt to 
show wide discrepancies in the hands of different chemists. Several 
modifications of Allihn's method, involving a longer period of heating, 
have been devised for the purpose of increasing the accuracy of the 
determination with dilute sugar solutions. 

Pfliiger's Method. Pfliiger,* who uses the same reagents and 
volumes of solutions as in Allihn's method, has modified the determina- 
tion by heating the mixed sugar and Fehling's solutions (145 c.c. in all) 
in a boiling- water bath for exactly 30 minutes and then diluting with 
130 c.c. of cold water before filtering. The cuprous oxide is filtered upon 
asbestos and, after washing and drying, the weight of precipitate deter- 
mined. Owing to the frequent occurrence of impurities in the cuprous 
oxide, especially when working with fluids or extracts of animal origin, 
Pfliiger advises to make also a direct determination of the copper by 
means of the thiocyanate method. 

* Pfliiger's Archiv, 69, 399. 



420 SUGAR ANALYSIS 

Pfliiger's table giving the weights of glucose corresponding to dif- 
ferent weights of cuprous oxide and copper, is found in the Appendix 
(Table 11). 

Koch and Ruhsam's * Method. In this modification the same 
reagents and volumes of solutions are used as in Allihn's and Pfliiger's 
methods. The mixed sugar and Fehling's solutions (145 c.c. in all) 
are first brought to a boil and then set in a boiling-water bath for ex- 
actly 30 minutes. The solution without diluting is then filtered through 
asbestos in a Gooch crucible and the reduced copper determined by any 
of the usual methods. 

The glucose table for Koch and Ruhsam's method is given in the 
Appendix (Table 12). 

Koch and Ruhsam's modification was designed for determining 
glucose in tannin extracts, etc., and is the official method of the Ameri- 
can Leather Chemists and other similar associations. 

The modifications of Allihn's method, using 30-minute heating, are 
considerably more accurate than the original process upon dilute glu- 
cose solutions and should be employed for determining small amounts 
of sugar in urine, tannin extracts and other animal and vegetable sub- 
stances of low glucose content. When, however, the 25 c.c. of sugar 
solution contain over 0.10 gm. of glucose, Allihn's original method of 
2-minute boiling may be followed with perfect safety, and with a con- 
siderable economy of time. The fact that more copper is reduced upon 
longer heating does not affejt the accuracy of the method, since the 
tables were standardized under exactly similar conditions. 

Application of Allihn's Method to the Determination of Other Re- 
ducing Sugars. Allihn's method has been employed for determining 
other reducing sugars besides glucose. Honig and Jesser f have used 
the method for determining fructose and have constructed a table giv- 
ing the copper-reducing power of fructose for different weights of sugar. 
In Table LXXIV the fructose values of Honig and Jesser, and the 
corresponding glucose values of Allihn, are given for several weights of 
reduced copper. The ratio of fructose to glucose, for the same weight 
of copper, is also given. 

For equal weights of sugar the amount of copper reduced by fructose 
is about 92 per cent of that reduced by glucose. Soxhlet found by his 
volumetric method (p. 391) that for equal weights of sugar the reducing 
power of fructose was 92.4 per cent that of glucose. 

* J. Soc. Chem. Ind., 13, 1227. 
t Monatshefte, 9, 562. 



REDUCTION METHODS FOR DETERMINING SUGARS 421 



TABLE LXXIV 
Showing Comparative Reducing Power of Fructose and Glucose 



Reduced copper. 


Fructose (Honig 
and Jesser). 


Glucose (Allihn). 


Ratio , gluc08e . 
fructose 


Mgs. 








32.7 


20 


17.4 


0.870 


70.2 


40 


35.9 


0.898 


107.1 


60 


54.6 


0.910 


143.2 


80 


73.0 


0.912 


178.9 


100 


91.5 


0.915 


213.9 


120 


110.0 


0.917 


248.3 


140 


128.3 


0.916 


282.2 


160 


146.7 


0.917 


315.3 


180 


165.0 


0.917 


347.9 


200 


183.1 


0.916 


379.9 


220 


201.3 


0.915 


411.3 


240 


219.5 


0.915 


Average ratio (excluding first 2 of the series) 0.915 



Reducing Ratios of Sugars. It is seen from Table LXXIV that 
if the values are eliminated for weights of sugar under 50 mgs., for 
which, as previously stated, Allihn's method gives uncertain results, 
the ratio of fructose to glucose for the same weight of reduced copper is 
a constant quantity 0.915. Other monosaccharides show a similar 
constancy of ratio. The following ratios are given by Browne * for a 
number of other sugars, the copper-reducing power in all cases being 
determined by Allihn's method: 

Glucose 



Arabinose 
Glucose 
Xylose 
Glucose 
Invert Sugar 
Glucose 
Galactose 



= 1.032. 



= 0.983. 



= 0.958. 



= 0.898. 



Relative Copper-reducing Power. Instead of using the ratios 
of the weights of sugars for the same amount of reduced copper, the 
ratios of the weights of copper reduced by the same amount of the two 
sugars are frequently used. O'Sullivan f expressed the relative copper- 
reducing power of a sugar by the symbol K and adopted as his standard 
(K = 100) the cupric oxide reduced by a given weight of glucose under 
the conditions of his method. O'Sullivan found, for example, that 1 gm. 






* J. Am. Chem. Soc., 28, 439. 
t J. Chem. Soc. (1879), 72, 275. 



422 SUGAR ANALYSIS 

of glucose reduced 2.205 gms. CuO and 1 gm. of maltose 1.345 gms. 
CuO. The relative copper, or cupric oxide, reducing power of maltose 



would then be K = X 100 = 61. 



In the examination of starch-conversion products the copper-re- 
ducing power of maltose, expressed by the symbol R, is sometimes 
taken as the standard. Taking the previous values of 'Sullivan the 



2 

R of glucose would be = X 100 = 164. 



In place of the constant K, Brown, Morris and Millar * have sub- 
stituted the value K, which is : K. According to this system the rela- 



tive copper-reducing power of maltose (using O'Sullivan's results) is 0.61 
K. The values for K, when determined for the same absolute weights of 
the two sugars, are practically identical with the reducing ratios as cal- 
culated in the previous section. 

Thus from Defren's table for glucose and maltose 44.4 mgs. of glucose 
reduce 100 mgs. CuO and 44.4 mgs. of maltose reduce 61.1 mgs. CuO then 

-J- = 0.611, K for maltose. 
lUU 

Using again Defren's table 44.4 mgs. glucose and 72.8 mgs. maltose reduce 

44 4 
respectively 100 mgs. CuO, then ^ = 0.610, the reducing ratio of maltose to 

7 .G 

glucose. 

If K, however, is calculated from the weights of sugars as determined 
by the solution factor 3.86, as is sometimes done, then the true reducing 
ratio is not found unless a correction be applied as indicated on page 32. 

The disaccharides, lactose and maltose, do not show usually the 
same constancy in reducing ratios for different weights of copper as the 
monosaccharides. This is due to the partial hydrolysis of the disac- 
charides as previously explained; the reducing ratio is usually higher 
the greater the amount of disaccharide. The copper-reducing ratios of 
lactose and maltose are approximately as follows for Allihn's method: 

LactoJhydrate = - 66 to - 71 ' or approximately 0.7. 
jyF~TT- ~ = 0-56 to 0.62, or approximately 0.6. 
* J. Chem. Soc. (1897), 96, 






REDUCTION METHODS FOR DETERMINING SUGARS 423 

If the copper-reducing power of a sugar is determined (as by Allihn's 
method), the corresponding glucose value of Allihn's table divided by 
the reducing ratio of the sugar to glucose will give the weight of sugar 
in the 25 c.c. of solution. 

Example. 25 c.c. of a fructose solution gave by Allihn's method 265.3 
mgs. of copper. 

The amount of glucose corresponding to 265.3 mgs. of copper, according 
to Allihn's table, is 137.45 mgs. 137.45 -f- 0.915 (the reducing ratio of fructose 
to glucose) = 150.2 mgs. of fructose. The amount of fructose corresponding to 
265.3 mgs. of copper according to Honig and Jesser is 150 mgs. 

The reducing ratios of the different sugars, have an important bearing 
upon the analysis of sugar mixtures, as described in Chapter XVI. 

Special copper-reduction methods and tables, similar to those of 
Allihn, have been established for other reducing sugars. It is im- 
possible to describe all of these in detail and only the following examples 
are given for invert sugar, maltose and lactose. The methods and 
tables are taken from Wein's "Zuckertabellen." 

Meissl's * Method for Determining Invert Sugar. The Soxhlet 
formula for Fehling's solution is used; 25 c.c. of the copper-sulphate 
solution and 25 c.c. of the alkaline-tart rate solution are mixed with the 
sugar solution, which should not contain over 0.245 gm. of invert 
sugar. Enough water is added to make the whole up to 100 c.c., the 
liquid is heated to boiling and kept at ebullition for exactly 2 minutes. 
The cuprous oxide is then filtered on asbestos and the reduced copper 
determined by any of the usual methods. The amounts of invert 
sugar corresponding to different weights of reduced copper are given in 
the Appendix in Table 13, which was calculated by Wein from Meissl's 
reduction factors. 

Wein'sf Method for Determining Maltose. The Soxhlet formula 
for Fehling's solution is used; 25 c.c. of the copper-sulphate solution 
and 25 c.c. of the alkaline-tartrate solution are mixed and heated to 
boiling; 25 c.c. of the sugar solution, which should not contain over 
0.25 gm. of maltose, are then added and the liquid boiled for exactly 
4 minutes. The cuprous oxide is filtered on asbestos and the reduced 
copper determined by any of the usual methods. The amounts of 
maltose corresponding to different weights of reduced copper are given 
in the Appendix in Table 14. 

According to Brown, Morris and Millar, J whose results have been 

* Z. Ver. Deut. Zuckerind., 29, 1050. 

f Wein's "Tabellen." % J. Chem. Soc., Trans., 71, 96. 




424 SUGAR ANALYSIS 

confirmed by Ling and Baker,* the table of Wein gives results which 
are about 5 per cent too low. 

Soxhlet'sj Method for Determining Lactose. The Soxhlet formula 
for Fehling's solution is used; 25 c.c. of the copper-sulphate solution 
and 25 c.c. of the alkaline-tartrate solution are mixed with 20 to 
100 c.c. (according to concentration) of the milk-sugar solution, which 
should not contain over 0.300 gms. of lactose hydrate. If less than 
100 c.c. of milk-sugar solution is taken sufficient water is added to make 
the whole up to 150 c.c. The liquid is then heated to boiling and kept 
at ebullition for exactly 6 minutes. The cuprous oxide is filtered on 
asbestos and the reduced copper determined by any of the usual 
methods. The amounts of lactose hydrate corresponding to different 
weights of reduced copper are given in the Appendix in Table 15, cal- 
culated by Wein from Soxhlet's reduction factors. 

UNIFIED COPPER-REDUCTION METHODS FOR SEVERAL SUGARS 

The confusing multiplicity of copper-reducing tables is due to the 
fact that different investigators have confined their work to one single 
sugar for one individual set of conditions. A number of chemists, how- 
ever, have worked with the purpose of establishing one uniform method 
for all reducing sugars. Examples of such unified methods are those of 
Kjeldahl and Woy, Defren, Munson and Walker, and Bertrand. 

Unified Method of KjeldahlJ and Woy. In Kjeldahl's method, 
as modified by Woy, the Fehling's solution is prepared for each analysis 
with a freshly weighed portion of Rochelle salts. The following solu- 
tions are used: 

(A) 69.278 gms. of pure CuS0 4 .5 H 2 are dissolved to 1000 c.c. 

(B) 130 gms. of pure sodium hydroxide (the amount must be 

established by titration) are dissolved to 1000 c.c. 
According to the richness of the sugar solution, 15 c.c., 30 c.c. or 50 c.c. 
of mixed reagent are made up in an Erlenmeyer flask holding about 
150 c.c. 
For 15 c.c. of reagent take 7.5 c.c. of A, 7.5 c.c. of B and 2.6 gms. 

Rochelle salts. 
For 30 c.c. of reagent take 15.0 c.c. of A, 15.0 c.c. of B and 5.2 gms. 

Rochelle salts. 
For 50 c.c. of reagent take 25.0 c.c. of A, 25.0 c.c. of B and 8.65 gms. 

Rochelle salts. 
The sugar solution is then added, the total volume of liquid in the 

* J. Chem. Soc., Trans., 71, 509. I Neue Z. Riibenzuckerind., 37, 29. 
t J. prakt. Chem., 21, 266. Chem. Centralblatt. 97 [2], 986. 



REDUCTION METHODS FOR DETERMINING SUGARS 425 

flask being always brought to 100 c.c. The flask is then plunged in a 
boiling-water bath and heated for exactly 20 minutes, while leading 
through the liquid a stream of hydrogen, or of illuminating gas which 
has been freed of oxygen by passing through a gas washer containing 
pyrogallic acid and sodium hydroxide solution. The reoxidation of the 
cuprous oxide by the air is in this way prevented. At the end of the 20 
minutes the cuprous oxide is filtered on asbestos, washed, ignited and 
weighed as cupric oxide. The amounts of glucose, fructose, invert sugar, 
lactose hydrate or maltose corresponding to different weights of cupric 
oxide or copper are given in the Appendix in Table 16, which was cal- 
culated by Woy for the 15-c.c., 30-c.c. and 50-c.c. volumes of reagent. 

The Kjeldahl-Woy method is one of great exactness, being carried 
out under rigidly defined conditions. The rather complicated details 
in preparing the copper reagent and in conducting the reduction have 
prevented the process from coming into extensive use. 

Unified Method of Brown, Morris and Millar.* In this method, 
which is adapted from a previous process by O'Sullivan, the Fehling's 
solution is prepared by dissolving 34.6 gms. crystallized copper sulphate, 
173 gms. Rochelle salts and 65 gms. anhydrous sodium hydroxide to 
1000 c.c.; 50 c.c. of the reagent are placed in a beaker of about 250 c.c. 
capacity and of 7.5 cm. diameter. The beaker is set in a boiling-water 
bath, and when the solution has acquired the same temperature, the 
measured volume of sugar solution is added and the whole made up 
to 100 c.c. with boiling distilled water. The beaker is covered with a 
clock glass, returned to the bath and heated exactly 12 minutes. The 
cuprous oxide is filtered in a tube and weighed as metallic copper or 
cupric oxide. 

The table of Brown, Morris and Millar (Appendix, Table 17) gives 
the weight of copper and cupric oxide which correspond to the same 
weight of glucose, fructose and invert sugar, the order of arrangement 
being the reverse of that in most tables. 

Unified Method of Defren.f In Defren's method, which is 
adapted from O'Sullivan, Soxhlet's formula for Fehling's solution is 
used; 15 c.c. of the copper-sulphate solution and 15 c.c. of the alkaline- 
tartrate solution are diluted with 50 c.c. of water in a 300-c.c. Erlen- 
meyer flask. The latter is then immersed for 5 minutes in a boiling- 
water bath, when 25 c.c. of the sugar solution are quickly run in from a 
burette. The flask is replaced in the bath and heated for exactly 15 
minutes. The cuprous oxide is then filtered on asbestos, washed, 

* J. Chem. Soc., Trans., 71, 281. 
t J. Am. Chem. Soc., 18, 751. 



426 SUGAR ANALYSIS 

ignited and weighed as cupric oxide. The amounts of glucose, maltose 
or lactose corresponding to different weights of cupric oxide are given 
in the Appendix in Table 18. 

Unified Method of Munson and Walker.* Transfer 25 c.c. each 
of the copper and alkaline-tartrate solutions (Soxhlet's formula) to a 
400-c.c. Jena or Non-sol beaker and add 50 c.c. of reducing sugar- solu- 
tion, or, if a smaller volume of sugar solution be used add water to 
make the final volume 100 c.c. Heat the beaker upon an asbestos gauze 
over a Bunsen burner; so regulate the flame that boiling begins in 4 min- 
utes, and continue the boiling for exactly 2 minutes. Keep the beaker 
covered with a watch glass throughout the entire time of heating. 
Without diluting filter the cuprous oxide at once on an asbestos felt in 
a porcelain Gooch crucible, using suction. Wash the cuprous oxide 
thoroughly with water at a temperature of about 60 C., then with 
10 c.c. of alcohol and finally with 10 c.c. of ether. Dry for 30 minutes 
in a water oven at 100 C., cool in a desiccator and weigh as cuprous 
oxide. The amounts of glucose, invert sugar, lactose or maltose cor- 
responding to different weights of cuprous oxide or copper are given in 
the Appendix in Table 19. 

Unified Method of Bertrand.f The following formula is used in 
preparing the copper reagents: 

(A) 40 gms. of pure CuS0 4 .5 H0 2 are dissolved to 1000 c.c., 

(B) 200 gms. of Rochelle salts and 150 gms. of sodium hydroxide in 

sticks are dissolved to 1000 c.c.: 

20 c.c. of the sugar solution, which should not contain over 0.100 gm. 
of reducing sugars, are transferred to a 150-c.c. Erlenmeyer flask, and 
20 c.c. each of solutions A and B added. The liquid is then heated to 
boiling and kept at gentle ebullition for exactly 3 minutes. The solu- 
tion is then filtered through asbestos, the precipitate of cuprous oxide 
washed with distilled water and the reduced copper determined by 
the volumetric permanganate method. 

The table of Bertrand (Appendix, Table 20) gives the different 
weights of reduced copper which correspond to the same weight of 
invert sugar, glucose, galactose, maltose and lactose, the order of 
arrangement being the same as in the table of Brown, Morris and 
Millar. 

METHODS FOR DETERMINING REDUCING SUGARS IN PRESENCE OF SUCROSE 

Reference has been made to the slight hydrolytic action of hot 
Fehling's solution upon the higher saccharides. While this action in 

* J. Am. Chem. Soc., 28, 663; 29, 541; 34, 202. f Bull, soc. chim., 35, 1285. 



REDUCTION METHODS FOR DETERMINING SUGARS 427 



case of sucrose is slight it is, nevertheless, sufficiently pronounced to 
cause a considerable error in the determination of reducing sugars 
when much sucrose is present. 

Conditions Affecting the Reducing Action of Sucrose upon Fehl- 
ing's Solution. The reducing action of sucrose upon Fehling's 
solution is proportional first, to the concentration of the sucrose and, 
second, to the amount of copper left unreduced. If enough reducing 
sugars are present to precipitate nearly all the copper from the Fehl- 
ing's solution the inversion of the sucrose will be very slight. This is 
shown in Table LXXV, which gives a series of experiments by Browne.* 
Constant quantities of sucrose, and varying amounts of glucose, were 
taken, and a determination of the latter made by Allihn's method. 

TABLE LXXV 

Showing Influence of Sucrose Upon the Reducing Action of Glucose 



A, 

Sucrose taken 
in 25 c.c. 


B. 

Glucose taken 
in 25 c.c. 


c. 

Glucose found 
in 25 c.c. 


D. 

Error 

(C-B). 


E. 
Calculated correc- 
tion, 
/ mgs. sucrose \ 


F. 

Corrected glu- 
cose, 

(C-E). 


\rngs. glucose + 40/ 


Mgs. 


Mgs. 


Mgs. 


Mgs. 


Mgs. 


Mgs. 


250 


50 


52.3 


2.3 


2.7 


49.6 


250 


100 


102.8 


2.8 


1.8 


101.0 


250 


150 


151.8 


1.8 


1.3 


150.5 


250 


200 


199.0 


-1.0 


1.0 


198.0 


500 


100 


104.5 


4.5 


3.5 


101.0 


500 


150 


153.2 


3.2 


2.6 


150.6 


500 


200 


203.2 


3.2 


2.1 


201.1 


500 


250 


251.3 


1.3 


1.7 


249.6 


1000 


50 


60.3 


10.3 


10.0 


50.3 


1000 


100 


108.2 


8.2 


6.8 


101.4 


1000 


200 


205.3 


5.3 


4.1 


201.2 


1000 


250 


252.0 


2.0 


3.4 


248.6 


2000 


50 


66.6 


16.6 


18.8 


47.8 


2000 


100 


113.7 


13.7 


13.0 


100.7 


2000 


200 


207.5 


7.5 


8.1 


199.4 


2000 


250 


255.5 


5.5 


6.8 


248.7 



The error in the glucose determination, when sucrose is present, is 
seen to be considerable; it is even more pronounced in such reduction 
methods as those of Kjeldahl or Pfliiger, which employ a long period of 
heating. 

It is seen from Table LXXV that the error in the glucose determina- 
tion is directly proportional to the amount of sucrose, and inversely 
proportional to the amount of glucose. Browne has proposed the use 

r milligrams sucrose , 

of an empirical formula, .. . Ar .> as a means of correct- 

milligrams glucose r\- 40 

* J. Am. Chem. Soc., 28, 451. 



428 SUGAR ANALYSIS 

ing for the reducing action of sucrose, when using Allihn's method. 
Table LXXV gives a comparison of the actual errors and of the results 
corrected by means of such a formula. 

In the volumetric methods of Soxhlet, Violette, etc., where the in- 
vert sugar solution is added to the point of complete reduction, no 
excess of copper is left in solution, and the error due to the presence of 
sucrose is practically negligible. 

A number of special copper-reduction methods have been designed 
for determining invert sugar in sugar-house products. The methods 
are classified according to the excess of sucrose over invert sugar in the 
material to be analyzed. 

Herzf eld's* Method for Determining Invert Sugar in Raw Sugars 
Containing Less than 1.5 per cent Invert Sugar. This method is 
designed for the analysis of the higher grades of raw sugar. The sugar 
solution, which should contain 20 gms. of material in 100 c.c. and be 
free from suspended or soluble impurities, is conveniently prepared as 
follows : 

Dissolve 44 gms. of sugar in about 100 c.c. of water in a 200-c.c. 
graduated flask. A little normal lead-acetate solution, just sufficient 
for clarification, is then added and the volume completed to 200 c.c. 
The solution is shaken, filtered and 100 c.c. of the filtrate (22 gms. 
sugar) measured into a 100-110 c.c. flask. Sufficient carbonate, or 
sulphate of sodium is then added to precipitate the excess of lead and 
the volume made up to 110 c.c. The solution is shaken, filtered and 
50 c.c. of the filtrate (10 gms. of sugar) used for the determination. 

Heat 25 c.c. each of the copper-sulphate and alkaline-tartrate solu- 
tions (Soxhlet 's formula) to boiling; the 50 c.c. of clarified sugar solution 
are then added and the whole boiled for exactly 2 minutes. The cuprous 
oxide is filtered on asbestos, washed and the reduced copper determined 
by any of the usual methods. The amounts of invert sugar corre- 
sponding to different weights of copper are given in the Appendix, in 
Table 21. 

In case the percentage of invert sugar in the raw sugar exceeds 
1.5 per cent, Herzf eld's method is no longer applicable. 

Meissl and Wein'st Method for Determining Invert Sugar in 
Mixtures of 90 to 99 per cent Sucrose with 10 to i per cent Invert 
Sugar. This method is designed for the analysis of low-grade raw 
sugars, or of other sugar-house products which do not contain a large 

* Z. Ver. Deut. Zuckerind. (1885), 985. 
t Wein's "Tabellen." 



REDUCTION METHODS FOR DETERMINING SUGARS 429 



excess of invert sugar. The sugar solution is prepared as in the previous 
method, the final filtrate being diluted if necessary so as not to contain 
more than 0.2 to 0.245 gms. of invert sugar in 50 c.c. 

Mix 25 c.c. each of the copper-sulphate and alkaline- tartrate solu- 
tions (Soxhlet's formula) with the 50 c.c. of clarified sugar solution; the 
liquid is then heated to boiling and kept at gentle ebullition for exactly 
2 minutes. The cuprous oxide is then filtered on asbestos, washed and 
the reduced copper determined by any of the usual methods. 

For determining the weights of invert sugar corresponding to differ- 
ent weights of reduced copper, for percentages of sucrose between 90 
and 99, the following condensed table has been calculated by Wein. 
Intermediary values can be easily calculated by interpolating. 

TABLE LXXVI 

For Determining Invert Sugar in Presence of Sucrose. (Meissl and Wein.) 



In mixtures of sucrose 
(S) and invert sugar (/) 
in parts per hundred. 


Milligrams of invert sugar. 


245 


225 


200 


175 


150 


125 


100 


75 


50 


Correspond to Milligrams of Copper. 


99 S + 17 






417.3 
393.7 
385.7 
381.7 
379.3 
376.6 
374.6 
373.1 
372.0 
371.1 


370.8 
357.7 
350.6 
339.1 
337.0 
334.7 
332.3 
330.4 
328.8 
327.8 


323.6 
304.7 
298.4 
295.3 
293.4 
290.1 
287.8 
286.3 
285.1 
284.0 


277.5 
259.7 
253.8 
250.8 
249.0 
245.4 
242.9 
241.0 
239.4 
238.2 


230.0 
213.7 
207.9 
205.0 
203.3 
199.8 
197.3 
195.4 
193.9 
192.7 


182.0 
166.0 
158.3 
155.4 
153.6 
151.0 
149.2 
147.9 
146.8 
146.0 


131.5 
113.8 
107.9 
105.7 
103.2 
101.5 
100.2 
99.3 
98.6 
98.0 


98 S+ 27. .. 






97 S+ 37. . . 






96 S + 47 
95 S + 57 
94 S + 67 
93 S+ 77 


439 '.7 

438.5 
437.6 
437.0 
436.5 
436.1 


420.1 
416.5 
413.9 
411.9 
410.3 
409.2 


52 iSf + 87 


91 S+ 9 7. 


90S+107 



The employment of the above table is best understood from an ex- 
ample : 

A sugar, which indicated 96.2 per cent sucrose by Clerget's method, was 
made up so that 50 c.c. of the clarified and deleaded solution contained 
10 gms. of sample. The amount of reduced copper obtained by MeissPs method 
was 324 mgs. Required the percentage of invert sugar. 

The invert sugar corresponding to 324 mgs. copper according to Meissl's 
table for invert sugar alone is 178 mgs. or 1.78 per cent (uncorrected) . The 
percentage composition, in a mixture of 96.2 parts sucrose with 1.78 parts in- 
vert sugar is approximately 98 per cent sucrose and 2 per cent invert sugar. 
Opposite the mixture 98 S + 2 7 of the table it is seen that 

357.7 mgs. of copper = 175 mgs. invert sugar, 
and 304.7 mgs. of copper = 150 mgs. invert sugar, 



430 SUGAR ANALYSIS 

then for the intermediary 324.0 mgs. of copper 

324.0 - 304.7 Q = 15Q Q mgg 



357.7 304.7 

of invert sugar or 1.59 per cent. 

Meissl and Wein's method is not applicable to products which con- 
tain more than 10 parts invert sugar in 100 parts of mixed sugars. For 
this reason the method has largely given place to the more general 
process of Meissl and Hiller. 

Meissl and Killer's * Method for Determining Invert Sugar in 
Mixtures Containing less than 99 per cent Sucrose and more than i 
per cent Invert Sugar. This method is designed for the analysis of 
all sugar-house products except the highest grades of raw sugars. The 
method is based upon the principle of taking such a quantity of material 
for analysis that the invert sugar will reduce nearly all the copper, 
thus reducing the error due to presence of sucrose to a minimum. 

The sugar solution is prepared as in the two previous methods so 
that 100 c.c., after clarification and deleading, contain 20 gms. of 
sample. Prepare a series of solutions in large test tubes by adding 
1, 2, 3, 4 and 5 c.c. of this solution to each tube successively. Add 
5 c.c. of the mixed copper reagent (Soxhlet's formula) to each, heat 
to boiling 2 minutes, and filter. Note the volume of sugar solution 
which gives the filtrate lightest in tint, but still distinctly blue. Place 
20 times this volume of the sugar solution in a 100-c.c. flask, dilute to 
the mark and mix well. Use 50 c.c. of the solution for the determina- 
tion, which is conducted as in the method of Meissl and Wein. The 
invert sugar is then calculated -by means of the following formulae. 

Let Cu = the weight of copper obtained; 
P = the polarization of the sample; 
W = the weight of sample in the 50 c.c. of solution used for 

determination; 

F = the factor obtained from the table for conversion of cop- 
per to invert sugar; 

-rt- = approximate weight of invert sugar = A ; 

100 
AX-yy = approximate per cent of invert sugar = y, 

100 P 

p , = o, approximate per cent of sucrose in mixture of sugars; 

100 S = /, approximate per cent of invert sugar; 
C\iF 
~w~ = per cent of invert sugar. 

* Z. Ver. Deut. Zuckerind. (1889), 735. 



REDUCTION METHODS .FOR DETERMINING SUGARS 431 



The factor F for calculating copper to invert sugar is then found 
from the following table: 

TABLE LXXVII 

Meissl and Hitter 7 s Factors for Calculating Copper to Invert Sugar for Different Ratio* 

of Sucrose to Invert Sugar 



Ratio of 
sucrose to in- 
vert sugar 
=5:7. 


Approximate weight of invert sugar = A. 


200 
Mgs. 


175 
Mgs. 


150 
Mgs. 


125 
Mgs. 


100 
Mgs. 


75 

Mgs. 


50 
Mgs. 


0:100 


56.4 


55.4 


54.5 


53.8 


53.2 


53.0 


53.0 


10:90 


56.3 


55.3 


54.4 


53.8 


53.2 


52.9 


52.9 


20:80 


56.2 


55.2 


54.3 


53.7 


53.2 


52.7 


52.7 


30:70 


56.1 


55.1 


54.2 


53.7 


53.2 


52.6 


52.6 


40: 60 


55.9 


55.0 


54.1 


53.6 


53.1 


52.5 


52.4 


50:50 


55.7 


54.9 


54.0 


53.5 


53.1 


52.3 


52.2 


60 : 40 


55.6 


54.7 


53.8 


53.2 


52.8 


52.1 


51.9 


70:30 


55.5 


54.5 


53.5 


52.9 


52.5 


51.9 


51.6 


80:20 


55.4 


54.3 


53.3 


52.7 


52.2 


51.7 


51.3 


90: 10 


54.6 


53.6 


53.1 


52.6 


52.1 


51.6 


51.2 


91 :9 


54.1 


53.6 


52.6 


52.1 


51.6 


51.2 


50.7 


92:8 


53.6 


53.1 


52.1 


51.6 


51.2 


50.7 


50.3 


93:7 


53.6 


53.1 


52.1 


51.2 


50.7 


50.3 


49.8 


94:6 


53.1 


52.6 


51.6 


50.7 


50.3 


49.8 


48.9 


95:5 


52.6 


52.1 


51.2 


50.3 


49.4 


48.9 


48.5 


96:4 


52.1 


51.2 


50.7 


49.8 


48.9 


47.7 


46.9 


97:3 


50.7 


50.3 


49.8 


48.9 


47.7 


46.2 


45.1 


98:2 


49.9 


48.9 


48.5 


47.3 


45.8 


43.3 


40.0 


99: 1 


47.7 


47.3 


46.5 


45.1 


43.3 


41.2 


38.1 



The use of Meissl and Hiller's formulae and table for calculating 
invert sugar is best understood from an example. 

The polarization of a sugar was 86.4; 50 c.c. of a solution containing 3.256 
gms. of sample, reduced by Meissl and Hiller's method, 0.290 gms. of copper. 
Required the per cent of invert sugar. 
Cu = 0.290 
2 



= 0.145 = A. 



_0145 
- 0.145 



y- 



100 P 



8640 



-- 95.1 = S. 
I = 4.9. 



P + y 86.4 + 4.45 
100 - S = 100 - 95.1 = 

S : I = 95.1 : 4.9. 

By consulting the table it is seen that the vertical column headed 150 is 
nearest to A, 145, and the horizontal column having the ratio 95 : 5 is nearest 
to the ratio of S to /, 95.1 : 4.9. At the intersection of these columns is found 

C\iF 0.290X51.2 



the factor 51.2 which enters into the final calculation 
per cent of invert sugar. 



W 



3.256 



= 4.56 



432 SUGAR ANALYSIS 

Munson and Walker's * Method for Determining Invert Sugar in 
Presence of Sucrose. Munson and Walker have included in their 
unified method for reducing sugars determinations of invert sugar in 
presence of variable amounts of sucrose. Their table (Appendix, Table 
19) gives the weight of invert sugar for different weights of cuprous oxide 
or copper, when the total weight of invert sugar and sucrose in the 
solution taken is 0.4 gm. and 2.0 gms. The 0.4 gm. amount is used 
preferably for sugar products containing between 1 and 9 parts of 
sucrose to 1 part of invert sugar and the 2.0 gms. amount for sugar 
products containing over 9 parts of sucrose to 1 part of invert sugar. 
This range is sufficient to include all the products of the sugar factory. 

The method requires a preliminary investigation of the material in 
order to determine the approximate percentages of sucrose and invert 
sugar for use in making up the solution. 

MISCELLANEOUS COPPER-REDUCTION METHODS 

The large amount of free alkali in Fehling's copper solution has 
proved its most objectionable feature, owing to the influence which it 
has in rendering sucrose and other substances slightly copper reducing. 
Attempts have accordingly been made to devise a copper reagent for 
sugar analysis which would contain no caustic alkali. While none of 
the solutions thus designed has shown the same all around suitability 
as that of Fehling, a few of them have found a certain usefulness in 
special cases. 

Barfoed's f Copper-acetate Method. Barfoed's copper-acetate 
solution (p. 336), which is not reduced by the disaccharides, maltose and 
lactose, has appealed to chemists as a convenient means of determining 
glucose, fructose and other monosaccharides in presence of the higher re- 
ducing sugars. But notwithstanding its value for qualitative purposes, 
attempts to use Barfoed's reagent for the quantitative determination of 
glucose and other monosaccharides have always given unsatisfactory 
results. 

Soldaini's J Copper-bicarbonate Method. Soldaini's copper-bi- 
carbonate solution (p. 337) has also appealed to chemists as a means 
of avoiding certain errors resulting from tiie use of Fehling's solution. 
Soldaini's method, however, has usually given unreliable results, when 
used for quantitative purposes, the principal objections being the de- 
position of copper hydroxide and the precipitation of lime and other 
mineral impurities with the reduced copper. 

* J. Am. Chem. Soc., 28, 663. 

t Z. analyt. Chem., 12, 27. $ Ber., 9, 1126. 



REDUCTION METHODS FOR DETERMINING SUGARS 433 



Ost's * Copper-bicarbonate Method. Ost has modified Soldaini's 
reagent in order to eliminate its objectionable features. In his latest 
improvement of the method the copper reagent is prepared as follows: 
250 gms. of chemically-pure potassium carbonate and 100 gms. of chemi- 
cally-pure potassium bicarbonate are dissolved in water, and a solution 
containing 17.5 gms. of chemically-pure crystallized copper sulphate 
slowly added. The volume is then made up to 1000 c.c. and the solu- 
tion filtered through asbestos, the first runnings of the filtrate being 
rejected. 

In making the determination 100 c.c. of the copper reagent are 
treated with 50 c.c. of the sugar solution and the liquid boiled for 10 
minutes. The precipitate is then filtered upon asbestos and the re- 
duced copper determined by any of the usual methods. 

Ost has unified his method for a number of reducing sugars; a few 
of the values for different weights of reduced copper are given in Table 
LXXVIII. 

TABLE LXXVIII 
Showing Reducing Power of Different Sugars upon Ost's Copper Solution 



Reduced copper. 


Glucose. 


Fructose. 


Invert sugar. 


Maltose. 


Mgs. 


Mgs. 


Mgs. 


Mgs. 


Mgs. 


100 


30.7 


29.0 


30.0 


57.9 


150 


45.4 


42.7 


44.4 


85.4 


200 


60.7 


57.0 


59 


112.9 


250 


76.5 


71.6 


74 3 


141.1 


300 


93.0 


87.5 


90.9 


170.3 


350 


112.8 


106.4 


109.8 


201.5 


400 


134.9. 


128.2 


131.0 


235.6 



The method has not been found to give good results with lactose. 
Glucose, by Ost's process, reduces about 60 per cent more copper than 
by Allihn's method. 

For determining small amounts of reducing sugars Ost recommends 
the use of his ^-normal copper solution which contains 250 gms. chemi- 
cally-pure potassium carbonate, 100 gms. chemically-pure potassium 
bicarbonate and 3.6 gms. chemically-pure crystallized copper sulphate 
to the liter. In using this solution, which is very sensitive towards 
small amounts of reducing sugars, the time of boiling is reduced to 5 
minutes. 

Ost's method has given good results in the analysis of pure sugar 
solutions, but has proved less reliable in the examination of low-grade 
products owing to the precipitation of lime and other mineral impurities. 

* Chem. Ztg., 19, 1784, 1829. 



434 



SUGAR ANALYSIS 



This difficulty, according to Ost, may be obviated by precipitating the 
lime with ammonium oxalate during the clarification. The method 
upon the whole has not offered sufficient advantages over Fehling's 
solution to come into general use. 

Bang's Copper-bicarbonate Method. Bang * has recently em- 
ployed the copper-bicarbonate method for the volumetric determina- 
tion of very small amounts of glucose. In this method the excess of 
copper, which remains in solution after reduction, is titrated with a 
standard hydroxylamine-sulphate solution in presence of potassium 
thiocyanate. 

TABLE LXXIX 



Hydroxyl- 
amine. 


Glucose. 


Hydroxyl- 
amine. 


Glucose. 


Hydroxyl- 
amine. 


Glucose. 


Hydroxyl- 
amine. 


Glucose. 


c.c. 


Mgs. 


c.c. 


Mgs. 


c.c. 


Mgs. 


c.c. 


Mgs. 


43.85 


5 


29.60 


19 


17.75 


33 


7.65 


47 


42.75 


6 


28.65 


20 


16.95 


34 


7.05 


48 


41.65 


7 


27.75 


21 


16.15 


35 


6.50 


49 


40.60 


8 


26.85 


22 


15.35 


36 


5.90 


50 


39.50 


9 


26.00 


23 


14.60 


37 


5.35 


51 


38.40 


10 


25.10 


24 


13.80 


38 


4.75 


52 


37.40 


11 


24.20 


25 


13.05 


39 


4.20 


53 


36.40 


12 


23.40 


26 


12.30 


40 


3.60 


54 


35.40 


13 


22.60 


27 


11.50 


41 


3.05 


55 


34.40 


14 


21.75 


28 


10.90 


42 


2.60 


56 


33.40 


15 


21.00 


29 


10.20 


43 


2.15 


57 


32.45 


16 


20.15 


30 


9.50 


44 


1.65 


58 


31.50 


17 


19.35 


31 


8.80 


45 


1.20 


59 


30.55 


18 


18.55 


32 


8.20 


46 


0.75 


60 



The unreduced copper and hydroxylamine react as follows: 
4 CuO + 2 NH 2 OH = 2 Cu 2 + N 2 + 3 H 2 O. 

The Cu20, which is thus formed, is immediately precipitated as white 
cuprous thiocyanate Cu 2 (SCN) 2 . The hydroxylamine solution is added 
until the blue color, due to the excess of unreduced copper, just dis- 
appears. The following solutions are employed: 

(A) 250 gms. of pure potassium carbonate, 50 gms. of pure potas- 
sium bicarbonate and 200 gms. of potassium thiocyanate are dissolved 
by warming in'about 600 c.c. of water. The liquid is cooled and a cold 
solution of 12.5 gms. crystallized copper sulphate in about 75 c.c. of 
water slowly added. The solution is made up to 1000 c.c. and, after 
standing 24 hours, filtered. 

(B) 6.55 gms. of pure hydroxylamine sulphate and 200 gms. of 
potassium thiocyanate are dissolved to 2000 c.c. 

One cubic centimeter of B should correspond to exactly 1 c.c. of A. 
* Biochem. Zeitschr., 2, 271. 






REDUCTION METHODS FOR DETERMINING SUGARS 435 

In making the determination 10 c.c. of the sugar solution, which 
should not contain over 60 mgs. of glucose, are measured into a 200-c.c. 
flask and 50 c.c. of solution A added. The liquid is heated to boiling 
and kept at ebullition for exactly 3 minutes. The solution in then 
cooled and solution B added from a burette until the blue color just dis- 
appears. Table LXXIX gives the milligrams of glucose correspond- 
ing to the cubic centimeters of hydroxylamine solution used. 

Kendall's Alkaline-salicylate Method. Kendall* has recently 
devised a method for determining reducing sugars, in which salicylic 
acid and potassium bicarbonate are used in place of the ordinary 
alkaline-tartrate mixture of Fehling's -solution. The advantages 
claimed are that the alkaline-salicylate mixture has no copper-re- 
ducing power of its own and that much larger amounts of copper are 
reduced by a given weight of sugar when the carbonates of the alka- 
lies are used in place of the hydroxides. 

The sugar solution is measured into a 200-c.c. Erlenmeyer flask, 
and the volume made up to 100 c.c. with distilled water. There are 
then added in succession 5 gins, salicylic acid, 15 c.c. copper-sulphate 
solution, containing 133.33 gms. CuS04.5 H 2 O per liter, and 25 c.c. 
potassium-carbonate solution, containing 600 gms. K 2 C03 per liter. 
The flask is shaken until the salicylic acid has completely dissolved, 
and then placed in a boiling- water bath for exactly 20 minutes; the 
reduced cuprous oxide is then filtered upon asbestos, washed with hot 
water, and the copper determined by Kendall's modified iodide 
method (p. 412). From the milligrams of copper thus found the 
corresponding weights of glucose, invert sugar, lactose hydrate and 
maltose hydrate are determined from a specially calculated table. 

VOLUMETRIC-REDUCTION METHODS BY MEANS OF MERCURY SOLUTIONS 
Of other metallic salt solutions besides copper only those of mercury 
have been used to any great extent for determining reducing sugars. 

Knapp'sf Alkaline Mercuric-cyanide Method. The solution used 
in Knapp's method is prepared by dissolving 10 gms. of pure mercuric 
cyanide and 100 c.c. of sodium-hydroxide solution of 1.145 sp. gr. to 
1000 c.c. The solution contains 7.9363 gms. of metallic mercury per liter. 
In making the determination a measured volume of the reagent, 
previously standardized against a known weight of the pure sugar, is 
heated to boiling and the sugar solution added from a burette until 
a drop of the filtered solution shows upon acidifying with acetic acid 
no coloration with ammonium-sulphide solution. The calculation of 



J, Am, Chem. Soc., 34, 317. t Z. analyt. Chem., 9, 395. 



436 SUGAR ANALYSIS 

sugar is made in the same manner as described under Soxhlet's volu- 
metric method with Fehling's solution. 

The end reaction in Knapp's method has been found uncertain 
and the process at present is but little used. 

Sachsse's * Alkaline Mercuric-iodide Method. The solution of 
Sachsse is prepared as follows: 18 gms. of pure dry mercuric iodide 
(prepared by precipitating mercuric-chloride solution with potassium 
iodide, and washing and drying at 100 C.) are dissolved in a solution 
containing 25 gms. of pure potassium iodide; a solution containing 80 
gms. of potassium hydroxide is then added and the volume completed 
to 1000 c.c. The solution contains 7.9323 gms. of metallic mercury 
per liter. 

An alkaline stannous-chloride solution, prepared by treating a 
solution of stannous chloride with an excess of potassium hydroxide, is 
used for determining the end point. 

In making the determination a measured volume of reagent is 
heated to boiling, and the sugar solution added until a drop of the 
filtered solution shows no coloration with the alkaline tin solution. 
The comparative reducing power of several sugars upon Sachsse's solu- 
tion is given in Table LXXXII, page 474. 

ESTIMATION OF HIGHER SACCHARIDES BY DETERMINING THE COPPER- 
REDUCING POWER AFTER HYDROLYSIS 

The methods previously described in this chapter for determining 
reducing sugars are equally applicable to the analysis of the higher non- 
reducing saccharides provided the latter first undergo a quantitative 
hydrolysis into sugars of known reducing power. 

The best examples of such applications of the method are the de- 
terminations of sucrose, starch, dextrin and glycogen by means of 
Fehling's solution. 

DETERMINATION OF SUCROSE BY MEANS OF FEHLING'S SOLUTION 

Sucrose upon treatment with invertase or acids is hydrolyzed quan- 
titatively, 95 parts of sucrose yielding 100 parts of invert sugar. 
If the copper-reducing power of an inverted-sucrose solution be deter- 
mined, the equivalent of invert sugar multiplied by the factor 0.95 will 
give the amount of sucrose present. 

In making the determination care must be taken that the amount 
of sugar after inversion does not exceed the limit of the tables, which 
for 50 c.c. of mixed Fehling's solution is about 240 mgs. of invert sugar, 

* Z. Ver. Deut. Zuckerind., 26, 872. 



REDUCTION METHODS FOR DETERMINING SUGARS 437 

or the equivalent of about 225 mgs. of sucrose. The chemist should 
check the method with pure sucrose, in which case the following pro- 
cedure may be followed. 

Dissolve 1.9 gms. of pure sucrose in about 75 c.c. of water in a 
500-c.c. graduated flask and invert the solution according to the method 
of Herzfeld, or any of the processes described in Chapter X. After 
cooling, the solution is nearly neutralized with sodium hydroxide 
(carefully avoiding any excess) and the volume completed to 500 c.c.; 
50 c.c. of this solution (containing 200 mgs. invert sugar = 190 mgs. 
sucrose) are then treated according to any of the copper-reduction 
methods for invert sugar and the weight of reduced copper determined. 
The milligrams of invert sugar, corresponding to this weight of cop- 
per, multiplied by the factor 0.95 gives the milligrams of sucrose. 

In applying the method to the determination of sucrose in sugar- 
house products, and other substances, which contain invert sugar, 
the difference between the invert-sugar equivalents before and after 
inversion is multiplied by 0.95. The same methods for determining 
invert sugar should be employed in both cases. The method of cal- 
culation is best illustrated by an example: 

Four grams of apple must were made up to 100 c.c. (solution A). Four gms. 
of the same must were inverted, nearly neutralized and made up to 100 c.c. 
(solution B). 

50 c.c. of sol. B gave by MeissFs method 407 mgs. Cu = 230 mgs. invert sugar 
50 c.c. of sol. A gave by Meissl's method 235 mgs. Cu = 126 mgs. invert sugar 

Difference = 172 mgs. Cu 104 mgs. invert sugar. 

104 mgs. invert sugar X 0.95 = 98.8 mgs. or 4.94 per cent sucrose. 

The mistake is sometimes made of taking the difference between 
the weights of reduced copper before and after inversion and calculat- 
ing the invert sugar and sucrose from this. The extent of this error, 
which is due to the variation in the copper-reducing power for different 
parts of the table (as shown in Table LXXI), may be seen from the 
previous example, where a difference of 172 mgs. of copper was found. 
172 mgs. of copper according to Meissl's table correspond to 90.8 mgs. 
of invert sugar. 90.8 X 0.95 = 86.2 mgs. or 4.31 per cent of sucrose, 
a result considerably less than that obtained by the other method. 

In calculating sucrose by any of the chemical methods, the reducing 
sugars before inversion must always be expressed as invert sugar, al- 
though it may actually exist as glucose, lactose, maltose, etc., or a 
mixture of several of these. This, of course, applies only to the sucrose 
calculation and not to that of the reducing sugars. 



438 SUGAR ANALYSIS 

Example. 5 gms. of a sirup containing sucrose and maltose were made up 
to 500 c.c. (solution A). 5 gms. of the same sirup were dissolved, inverted, 
nearly neutralized and made up to 500 c.c. (solution B). 



Maltose - 



Mgs. Mgs. Mga. 

50 c.c. of sol. B gave by Munson and Walker's method 390 = 215.0 

50 c.c. of sol. A gave by Munson and Walker's method 199 = 103.7 = 175.5 

Difference 191 111.3. 

111.3 X 0.95 = 105.7 mgs. = 21.14 per cent sucrose in sirup. 
175.5 mgs. = 35.10 per cent maltose in sirup. 

Calculating the sucrose from the difference in copper, as is sometimes 
wrongly done, would give the following: 191 mgs. Cu = 99.3 mgs. invert 
sugar (by Munson and Walker's table), 99. 3X 0.95 = 94.3 mgs. = 18.86 per 
cent sucrose in sirup. 

The unified methods and tables are most convenient for converting 
the equivalents of any reducing sugar into that of invert sugar. The 
same result, however, may be accomplished by means of the copper- 
reducing ratios given on page 421. 

Example. 10 gms. of a sirup containing sucrose and fructose were made 
up to 500 c.c. (solution A). 10 gms. of the same sirup were dissolved, inverted, 
nearly neutralized and made up to 500 c.c. (solution B). 
25 c.c. of sol. B gave by Allihn's method 414 mgs. Cu = 221 mgs. glucose 
25 c.c. of sol. A gave by Allihn's method 195 mgs. Cu = 100 mgs. glucose 

Difference =121 mgs. glucose. 

The reducing ratio of invert sugar to glucose is 0.958 for Allihn's method. 
121 -r- 0.958 = 126.3 mgs. invert sugar. 126.3 X 0.95 = 120 mgs. = 24.00 per 
cent sucrose in sirup. 

The reducing ratio of fructose to glucose is 0.915 for Allihn's method. 
100 -^ 0.915 = 109.3 mgs. = 21.86 per cent fructose in sirup. 

Owing to the slight variation in the reducing ratios of some of the 
sugars, as maltose and lactose, it is more accurate to determine the 
equivalents by one of the unified methods. 

DETERMINATION OF STARCH BY MEANS OF FEHLING's SOLUTION 

Starch upon heating with dilute hydrochloric acid is hydrolyzed al- 
most quantitatively according to the equation (C 6 Hi 05) n -f- nH 2 = 
nC 6 Hi 2 6 , in which 90 parts of starch yield 100 parts of glucose. The 
conversion of starch into glucose may be accomplished either by direct 
acid hydrolysis, as in Sachsse's method, or by first converting the starch 
into soluble products, as with diastase, and then hydrolyzing the filtered 
solution with acid. 



REDUCTION METHODS FOR DETERMINING SUGARS 439 

Method of Sachsse, as modified by the Association of Official 
Agricultural Chemists.* Stir a convenient quantity of the sample 
(representing from 2.5 to 3 gms. of the dry material) in a beaker with 
50 c.c. of cold water for an hour. Transfer to a filter and wash with 
250 c.c. of cold water. Heat the insoluble residue for two and a half 
hours with 200 c.c. of water and 20 c.c. of hydrochloric acid (sp. gr. 
1.125) in a flask provided with a reflux condenser. Cool, and nearly 
neutralize with sodium hydroxide; complete the volume to 250 c.c., 
filter and determine the glucose in an aliquot of the filtrate by any of 
the usual methods of copper reduction. The weight of glucose multi- 
plied by 0.90 gives the weight of starch. 

Owing to the fact that a perfect theoretical yield of glucose is never 
obtained from starch by acid hydrolysis, Ost f recommends the use 
of the factor 0.925 for converting glucose into starch by Sachsse's 
method. 

Sachsse's method is one of the simplest processes for estimating 
starch, but has the objection of converting pentosans and other hemi- 
celluloses into reducing sugars. The method for this reason gives too 
high results in the analysis of starchy substances which contain much 
cellular tissue. In order to eliminate this error the starch must be 
dissolved from cellular substances before hydrolyzing with acid; solu- 
tion of starch may be effected by heating under pressure or by the 
action of diastase. 

Method of Determining Starch by Solution under Pressure.! 
Three grams of the finely-ground sample are extracted with cold water, 
as in the previous method in order to remove sugars, dextrin, gums, etc. 
If much oil or fat is present the material should first be extracted with 
ether. The residue is then heated in a covered flask or metal beaker, 
of about 200-c.c. capacity, with 100 c.c. of water in an autoclave, a 
form of which designed by Soxhlet is shown in Fig. 175. The heating 
is continued for 3 to 4 hours at 3 atmospheres pressure. If an autoclave 
is not available, Lintner pressure bottles (Fig. 176) may be used; the 
bottles are immersed in a glycerol bath and heated for. 8 hours at 
108 to 109 C. 

When the digestion is finished the pressure is first allowed to sub- 
side, when the autoclave, or pressure flask, is opened and the solution 
filtered through asbestos. The insoluble residue is well washed with 
hot water, and should show no blue reaction with iodine when ex- 

* Bull. 107 (revised), U. S. Bur. of Chem., p. 53. 

f Chem. Ztg., 19, 1501. 

t Konig's "Untersuchung" (1898), p. 221. 



440 



SUGAR ANALYSIS 



amined under the microscope. The filtrate is made up to 200 c.c. and 
then heated with 20 c.c. of hydrochloric acid, of 1.125 sp. gr., for 3 




Fig. 175. Soxhlet's autoclave. 




Fig. 176. Lintner's pressure bottle. 



hours in a boiling-water bath, the flask, which holds the solution, being 
connected with a reflux condenser. The solution, after cooling, is near- 
ly neutralized with sodium hydroxide and made up to 500 c.c. The 
copper-reducing power of the solution is then determined; the glu- 
cose equivalent of the copper multiplied by 0.9 gives the corresponding 
equivalent of starch. 

Method of Determining Starch by Solution with Diastase. - 
Marcker* found that the best method of dissolving starch from hemi- 
celluloses was by means of diastase. The method of Marcker, as modi- 
fied by the Association of Official Agricultural Chemists, is as follows: 

Preparation of Malt Extract. Digest 10 gms. of fresh, finely- 
ground malt 2 or 3 hours at ordinary temperature wi-th 200 c.c. of 
* "Handbuch der Spiritusfabrikation" (1886), 94. 



REDUCTION METHODS FOR DETERMINING SUGARS 441 

water and filter. Determine y the amount of glucose in a given quantity 
of the filtrate after boiling with acid, etc., as in the starch determina- 
tion, and make the proper correction in the subsequent determination. 

Determination. Extract a convenient quantity of the substance 
(ground to an impalpable powder and representing from 4 to 5 gms. of 
the dry material) on a hardened filter with 5 successive portions of 
10 c.c. of ether; wash with 150 c.c. of 10 per cent alcohol and then 
with a little strong alcohol. Place the residue in a beaker with 50 c.c. 
of water, immerse the beaker in a boiling-water bath and stir constantly 
for 15 minutes or until all the starch is gelatinized; cool to 55 C., add 
20 c.c. of malt extract and maintain at this temperature for an hour. 
Heat again to boiling for a few minutes, cool to 55 C., add 20 c.c. of 
malt extract and maintain at this temperature for 1 hour or until a 
microscopic examination of the residue with iodine shows no starch. 
Cool and make up directly to 250 c.c.; filter. Place 200 c.c. of the 
filtrate in a flask with 20 c.c. of hydrochloric acid (sp. gr. 1.125); con- 
nect with a reflux condenser and heat in a boiling-water bath for two 
and one-half hours. Cool, nearly neutralize with sodium hydroxide 
and make up to 500 c.c. Mix the solution well, pour through a dry 
filter and determine the glucose in an aliquot of the filtrate by any of 
the usual methods of copper reduction. The weight of glucose multi- 
plied by 0.90 gives the weight of starch. 

Wein * has calculated a table for the above methods which gives 
the milligrams of starch or dextrin corresponding to milligrams of re- 
duced copper as obtained by Allihn's method. The table was con- 
structed by simply multiplying the milligrams of glucose in Allihn's 
table by the factor 0.9. 

Of the various processes for determining starch the diastase method 
secures the most perfect solution of starch with the least solution of 
accompanying hemicelluloses. In cases, however, where much cellular 
matter is present the hot water and malt solution may dissolve a small 
amount of pentosans, which, by being afterwards hydrolyzed into re- 
ducing pentose sugars, introduce a slight error in the determination. 

A more serious error than the above consists in the incomplete 
hydrolysis of starch into glucose. Experiments by W. A. Noyes,f 
and his coworkers, testing the action of 2.5 per cent hydrochloric acid 
upon the malt conversion of starch, show a hydrolysis, into glucose 
which is about 97 per cent of the theoretical. A diminished yield of 
glucose necessitates the use of a conversion factor somewhat greater 
than 0.9. 

* Wein's "Tabellen." t J- Am. Chem. Soc., 26, 266. 



442 SUGAR ANALYSIS 

Modification of Noyes* for Determining Starch by the Diastase 
Method. In the modification recommended by Noyes the filtrate 
from the malt digestion is treated with one-tenth its volume of hydro- 
chloric acid of sp. gr. 1.125. " After heating for 1 hour in a flask 
immersed in a boiling-water bath, making allowance for the time re- 
quired for the solution to attain the temperature of the bath, the solu- 
tion is cooled, enough sodium hydroxide is added to neutralize 90 per 
cent of the hydrochloric acid used, the solution made up to a definite 
volume, filtered on a dry filter, if necessary, and the reducing power de- 
termined by Fehling's solution; 100 parts of glucose found in this 
manner represent 93 parts of starch in the original material." 

Noyes emphasizes the importance of each chemist determining for 
himself with pure glucose the ratio between glucose and copper for the 
particular solutions and method which he uses. 



DETERMINATION OF DEXTRIN BY MEANS OF FEHLING's SOLUTION 

The principle of the method is the same as that described for starch. 
In the process described by Konig f a weighed amount of the dextrin is 
dissolved in cold water, made up to 1000 c.c. and filtered. Three 
portions of 200 c.c. each of the filtrate are heated in a boiling-water bath 
with 20 c.c. of hydrochloric acid of 1.125 sp. gr. for periods of 1, 2 and 
3 hours. The solutions are cooled, nearly neutralized with sodium 
hydroxide and made up to volume so that the solution does not contain 
over 1 per cent glucose. The glucose is then determined by any of the 
usual methods, and the highest results of the three experiments taken as 
the correct value. The weight of glucose multiplied by the factor 0.9 
gives the equivalent of dextrin. 

If sugars are also present, the glucose equivalent of these must be 
subtracted from the glucose equivalent after hydrolysis and the differ- 
ence calculated to dextrin. 

The hydrolysis of dextrin by dilute hydrochloric acid was found by 
W. A. Noyes t and his co-workers to, be a little less than 95 per cent 
complete at the end of 2 hours' heating and the results seemed to indi- 
cate that the theoretical yield of glucose could not be obtained even by 
prolonged heating. The theoretical factor 0.9 for converting glucose 
to dextrin is no doubt considerably too low for the method of acid 
hydrolysis. 

* J. Am. Chem. Soc., 26, 266. 

t Konig's "Untereuchung" (1898), p. 215. 

j J. Am. Chem. Soc., 26, 266. 



REDUCTION METHODS FOR DETERMINING SUGARS 443 

DETERMINATION OF GLYCOGEN BY MEANS OF FEHLING's SOLUTION 

Pfluger's* Glycogen Method. The method is based upon the 
hydrolysis into glucose of the impure glycogen (C 6 Hio0 5 )n, which has 
previously been precipitated from the solution of animal substance. 

One hundred grams of the finely divided tissue are heated with 100 c.c. 
of 60 per cent potassium-hydroxide solution, in a boiling-water bath for 
3 hours, the flask, which contains the solution, being shaken at frequent 
intervals. The cooled solution is made up to 400 c.c. and treated with 
800 c.c. of 96 per cent alcohol. After standing 24 hours the clear solu- 
tion is decanted through a filter, the precipitate of impure glycogen 
stirred with an excess of 60 per cent alcohol and again set aside. The 
settling of the glycogen in the numerous treatments may be hastened 
by adding a few drops of concentrated salt solution. The clear liquid 
is again decanted and the process repeated for a third time. The puri- 
fication is then continued in the same way, twice with 96 per cent alco- 
hol, once with absolute alcohol, three times with ether and once again 
with absolute alcohol. Any material adhering to the filter is then 
removed to the main portion of precipitate, and the raw glycogen dis- 
solved in hot water. The solution is then neutralized with hydro- 
chloric acid of 1.19 sp. gr., and transferred to a 500-c.c. flask; 25 c.c. of 
hydrochloric acid (sp. gr. 1.19) are then added and the liquid heated 
in a boiling-water bath for 3 hours. The solution is then cooled, neu- 
tralized, made up to 500 c.c., filtered and the glucose determined in 
the filtrate by Pfluger's method. The amount of glucose multiplied by 
the factor 0.927 gives the corresponding amount of glycogen. 

EXTRACTION OF SUGARS AND PREPARATION OF SOLUTIONS FOR CHEMICAL 
METHODS OF ANALYSIS 

The methods and precautions previously given for the extraction of 
sugars and preparation of solutions for polarimetric examination hold 
also for the chemical methods of analysis. 

Clarification of Solutions. With products which contain but 
little insoluble matter, such as sugars, molasses, sirups, jellies, honeys, 
etc., the weighed amount of material is dissolved in water, clarified, if 
necessary, with a minimum of neutral lead-acetate solution, made up to 
volume and filtered. The filtrate, after deleading by means of sodium 
carbonate, sodium sulphate, potassium oxalate or other means, as de- 
scribed on page 276, is then ready for analysis. 

With products of high purity, which contain but little mineral 
matter or organic non-sugars, the use of lead acetate may be dispensed 

* Pfluger's Archiv, 114, 242. 



444 



SUGAR ANALYSIS 



with, and a few cubic centimeters of alumina cream be used for clari- 
fication. 

Precipitation of Reducing Sugars by Basic-lead Salts. Lead sub- 
acetate, or other basic salts of lead, which are employed as clarifying 
agents in the polarimetric determination of sucrose, should never be 
used upon solutions in which reducing sugars are to be determined. 
The action of such compounds in causing a precipitation, or occlusion, 
of reducing sugars in the lead precipitate has already been mentioned. 
Bryan * found that basic-lead salts, in presence of magnesium sulphate 
and ammonium tartrate, precipitated in case of glucose from 3 per cent 
to 17 per cent, and in case of fructose from 8 per cent to 35 per cent, of 
the total amount of sugar in solution. Neutral lead acetate under the 
same conditions caused the precipitation of only 0.9 per cent of the total 
glucose and 0.0 per cent of the total fructose.. (See Table XL, p. 216.) 

In a series of independent experiments made by Bryan and Home f 
upon raw cane sugar and cane molasses the following results were 

obtained. 

TABLE LXXX 

Showing Influence of Clarification with Lead Subacetate upon Determination of Reduc- 
ing Sugars 



Clarifying agent and analyst. 


Allihn's method. 


Munson and Walker's Method. 


Weigh- 
ing as 
Cu 2 0. 


Weigh- 
ing as 
CuO. 


Titration 
of Cu by 
Low's 
Method. 


Weigh- 
ing as 
Cu 2 0. 


Weigh- 
ing as 
CuO. 


Titration 
of Cu by 
Low's 
method. 


r 


' No Clarifying Agent 
A. H. Bryan 
W. D. Home 

Average 


Percent 
6.45 

7.08 


Per cent 
6.22 

7.05 


Per cent 

5.88 
7.02 


Per cent 
6.29 

6.43 


Per cent 
5.98 
6.51 


Per cent 

5.83 
6.37 


6.77 


6.63 


6.45 


6.36 


6.25 


6.10 


Lead-subacetate Solution 
A. H. Bryan. . 


6.14 
6.61 


5.67 
6.51 


5.67 
6.51 


5.76 
6.19 


5.51 
6.01 


5.30 
5.99 


W. D. Home 


L. Average . 


6.38 


6.09 


6.09 


5.98 


5.76 


5.65 


' No Clarifying Agent 
A. H. Bryan 
W. D. Home 


19.77 
20.60 


19.37 
20.06 


19.45 
19.97 


19.20 
20.00 


18.34 
19.43 


18.43 
19.44 


Average . . 


20.19 


19.72 


19.71 


19.60 


18.89 


18.94 


Lead-subacetate Solution 
A. H. Bryan 


17.51 
19.45 


16.47 
19.16 


16.29 
19.16 


17.27 
19.00 


16.26 
18.53 


15.97 
18.26 


W. D. Home 


Average 


18.48 


17.82 


17.73 


18.14 


17.39 


17.12 





* Bull. 116, U. S. Bur. of Chem., p. 73. t Bull. 116, U. S. Bur. of Chem., pp. 72, 74. 



1 



REDUCTION METHODS FOR DETERMINING SUGARS 445 

Clarification with lead subacetate caused a loss of about 10 per 
cent of the total reducing sugars present. The variable results, due to 
method of estimating copper, show a contamination of the cuprous 
oxide as explained on page 416. The higher results by Allihn's method 
are due to the greater inverting action of the more strongly alkaline 
Fehling's solution. 



PREPARATION OF SUGAR SOLUTIONS FROM PLANT SUBSTANCES 

If the material to be analyzed contains much insoluble matter, as 
is the case with plant substances containing cellular tissue, the sugars 
must first be extracted by means of water or alcohol. In the case of 
grains, cattle-feeds, etc., the following provisional method is used by 
the Association of Official Agricultural Chemists.* 

Extraction of Sugars with Cold Water. Weigh into a flask or 
bottle, suitable for stirring or shaking, 10 to 20 gms. of the material, 
depending upon the amount of soluble carbohydrates present. Add 
250 c.c. of ice-cold water, less the volume of water present as moisture 
in the material, and stir or shake for 4 hours. If enzymatic action is 
feared the extraction should be made at a low temperature, preferably 
by surrounding the extraction flask with broken ice; or extract at ordi- 
nary temperature with 40 to 50 per cent alcohol. If there is present in 
the material much soluble substance, correction should also be made for 
the increase in volume due to solution. If necessary for clear filtration, 
add from 5 to 10 c.c. of alumina cream, just before filtering. The volume 
of alumina cream to be added must be taken into account in determin- 
ing the amount of water used for the extraction. After the extraction 
filter immediately, pouring back upon the filter the first portions of 
cloudy filtrate until the filtrate is clear. To free from soluble impuri- 
ties add sufficient normal lead-acetate solution to 200 c.c. of the filtrate 
to precipitate all impurities, make up to 250 c.c. and filter. Remove 
the excess of lead by means of anhydrous sodium carbonate or anhy- 
drous sodium sulphate, followed in the latter case by a small amount of 
anhydrous sodium carbonate, care being taken not to use an excess. 
Filter again and use the clear filtrate for the determination of reducing 
sugars. 

The extraction of sugars from plant substances by means of cold 
water is not always trustworthy owing to the action of enzymes upon 
sucrose, starch and other higher saccharides. The employment of hot 

* Bull. 107 (revised), U. S. Bur. of Chem., p. 57. 



446 SUGAR ANALYSIS 

water is also often unreliable on account of the solution of hemicellu- 
loses, starch and gums. 

Extraction of Sugars with Dilute Alcohol. Bryan, Given and 
Straughn * have recently made experiments upon the extraction of 
sugars from grains and similar products, using as solvents 50 per cent 
alcohol and 0.2 per cent sodium-carbonate solution. Both of these 
solvents inhibit the action of enzymes and were found to give con- 
cordant results upon certain classes of products. In many cases, how- 
ever, the sodium-carbonate extraction gave much higher amounts of 
reducing sugar after inversion a result, perhaps, of the solvent 
action of the alkali upon pentosans and other hemicelluloses. Bryan, 
Given and Straughn believe that extraction with 50 per cent alcohol, all 
points considered, is the most reliable method for general sugar work. 
The method outlined by them is as follows: 

Method of Bryan, Given and Straughn. Place 12 gms. of the finely 
ground substance in a 300-c.c. graduated flask, adding, in case the 
material is acid, from 1 to 3 gms. of precipitated calcium carbonate. 
Add 150 c.c. of neutral alcohol of 50 per cent volume strength; mix 
thoroughly and boil on a hot-water bath for 1 hour, placing a small 
funnel in the neck of the flask to condense the vapor. Cool and make 
up to 300 c.c. with neutral 95 per cent alcohol. After mixing and set- 
tling transfer 200 c.c. of the clear solution to a distilling flask and distil 
off the excess of alcohol, which is thus recovered for future use. The 
liquid residue is evaporated to a volume of 20 to 30 c.c. (but not to 
dryness), and then washed with water into a 100-c.c. graduated flask. 
The solution is clarified with the necessary amount of neutral lead- 
acetate solution, and, after standing 15 minutes, made up to 100 c.c. 
Pass through a folded filter, carefully saving all of the filtrate, to which 
add enough anhydrous sodium carbonate to precipitate the excess of 
lead; allow to stand 15 minutes and then pour through an ashl< 
filter. Over 75 c.c. of filtrate should be obtained; 25 c.c. of the cl< 
filtrate (equivalent to 2 gms. of original material) are diluted with 25 c.c. 
of water and used for the determination of reducing sugars; 50 c.c. oi 
the same filtrate are transferred to a 100-c.c. flask, inverted with 5 c.c. 
of concentrated hydrochloric acid, neutralized and made up to 100 c.c. 
Filter, if necessary, and take 50 c.c. (equivalent to 2 gms. of original 
material) for the determination of reducing sugars after inversion. 
The percentages of invert sugar and sucrose are calculated in the 
usual way and the results multiplied by the factor 0.97 to correct for 
the volume of insoluble matter. 

* Circular 71, U. S. Bur. of Chem. 



REDUCTION METHODS FOR DETERMINING SUGARS 447 



PREPARATION OF SUGAR SOLUTIONS FROM ANIMAL SUBSTANCES 

Clarification. Liquids of animal origin, such as blood, serum, 
urine, milk, secretions, extracts, etc., frequently contain large amounts 
of albuminoids and other nitrogenous substances which interfere with 
the determination of reducing sugars by the methods of copper re- 
duction. The clarifying agent which is most used for such liquids is 
mercuric nitrate. 

Mercuric-nitrate Solution. Treat 220 gms. of yellow oxide of mer- 
cury with 300 to 400 c.c. of water; then add nitric acid in small portions, 
with warming and stirring, until the precipitate is dissolved. Dilute to 
1000 c.c. and filter. 

The liquid to be clarified is treated with mercuric-nitrate solution 
until no more precipitate forms; the solution is then nearly neutralized 
with sodium-hydroxide solution of 1.3 sp. gr., made up to volume and 
filtered. A measured portion of the slightly acid filtrate is then freed 
from excess of mercury by precipitating with hydrogen sulphide; the 
solution is filtered, the hydrogen sulphide removed by a current of air 
and the reducing sugars determined by any of the usual methods. 

Clarification of Milk. For the clarification of milk, the use of 
copper sulphate and potassium hydroxide will be found more con- 
venient. The following is the official method of the Association of 
Agricultural Chemists.* 

Dilute 25 c.c. of the milk with 400 c.c. of water and add 10 cc. of a 
solution of copper sulphate of the strength given for Soxhlet's modi- 
fication of Fehling's solution. Add about 7.5 c.c. of a solution' of 
potassium hydroxide of such strength that one volume of it is just 
sufficient to completely precipitate the copper as hydroxide from one 
volume of the solution of copper sulphate. Instead of a solution of 
potassium hydroxide of this strength, 8.8 c.c. of a half-normal solution 
of sodium hydroxide may be used. After the addition of the alkali 
solution the mixture must still have an acid reaction and contain 
copper in solution. Fill the flask to the 500-c.c. mark, mix and filter 
through a dry filter. Determine the lactose by any of the usual 
methods. 

In determining reducing sugars in substances of animal origin, the 
precipitate of cuprous oxide is often badly contaminated with mineral 
and organic impurities, so that the reduced copper should be deter- 
mined directly and not by weighing as suboxide or oxide. 
* Bull. 107 (revised), U. S. Bur. of Chem., p. 119. 



448 SUGAR ANALYSIS 



CONCENTRATION OF SUGAR SOLUTIONS 

In working with very dilute solutions, such as contain only a few 
hundredths of a per cent of sugar, it is often necessary to concentrate 
the liquid to one-half, one-fifth or one-tenth the original volume be- 
fore a satisfactory determination of the copper-reducing power can be 
made. It is exceedingly important in evaporating such solutions that 
the liquid be kept exactly neutral, otherwise changes may result in the 
composition of the sugars. Traces of free acid may become sufficiently 
concentrated towards the end of evaporation to hydrolyze higher saccha- 
rides, and traces of free alkali may modify or destroy reducing sugars. 

The evaporation of solutions containing reducing sugars must be 
conducted in vessels which do not give up soluble alkali; the concen- 
tration of sugar solutions in glass vessels, unless of perfect resistant 
non-soluble quality, is for this reason to be avoided. The author has 
found flasks and basins of tinned copper to be very suitable for con- 
centrating sugar solutions, there being no change in reducing power 
after diluting and evaporating to the original volume. 

If the solution to be concentrated is slightly acid an excess of finely 
powdered calcium carbonate (alkali free) will prevent the hydrolysis of 
higher saccharides. If the solution, is alkaline, dilute acetic acid is first 
added to faint acidity, and then an excess of calcium carbonate. When 
the evaporation is completed, the residue of insoluble matter is removed 
by filtration. 



CHAPTER XV 

SPECIAL QUANTITATIVE METHODS 

THE determination of sugars by means of their reducing power upon 
Fehling's solution, Sachsse's solution or other metallic salt combina- 
tions is a general method, and has no value for the selective determi- 
nation of particular groups of reducing sugars. For such purposes 
more special processes of analysis must be adopted. The present 
chapter will describe a number of the best known of such special quan- 
titative methods. 

DETERMINATION OF PENTOSES AND PENTOSANS 
Theory of Method. The methods for determining pentoses and 
pentosans are due to the researches of Tollens,* and his school; they all 
depend upon the conversion of the pentose sugars into furfural by dis- 
tilling with hydrochloric acid, according to the principles described on 
p. 374. The amount of furfural, which distills over, is determined and 
calculated to pentoses. The yield of furfural does not correspond per- 
fectly to the equation, 

C 5 H 10 5 C 5 H 4 2 + 3H 2 0, 

100 parts pentose 64 parts furfural 

being for arabinose about 75 per cent and for xylose about 90 per cent 
of the theoretical. Yet by making the distillation under carefully con- 
trolled conditions, it is possible, by means of formulae or tables which 
have been established for different weights of pure pentoses, to make 
a determination with a very close degree of approximation. 

Different reagents have been used for precipitating the furfural in 
the determination of pentoses. Tollens and Stone first attempted to 
determine furfural by precipitating with ammonia as furfuramide. 
An important advance was then made by Tollens, in company with 
Giinther, de Chalmot, Flint and Mann, in using phenylhydrazine for 
precipitating the furfural. The use of phenylhydrazine was attended, 
however, with certain inconveniences and was finally abandoned upon 
the discovery by Councler f of the precipitating action of phloroglucin. 

* For a review of the subject see papers by Tollens with bibliography in Abder- 
halden's "Arbeitsmethoden," 1909, II, 130, and in Papier-Zeitung, 1907, Nos. 56, 60 
and 61 (Reprint). 

t Chem. Ztg., 17, 1743; 18, 966. 

449 



450 SUGAR ANALYSIS 

The phloroglucin method, as first developed by Tollens and Kriiger,* 
was further improved by Tollens and Rimbach, and finally established 
in its present form by Tollens and Krober.f 

Description of the Method. The necessary apparatus for making 
the determination is shown in Fig. 177. From 2 to 5 gms. of substance, 
according to the richness of the material in pentoses or pentosans, are 
placed in a 300-c.c. distillation flask with 100 c.c. of hydrochloric acid 




Fig. 177. Apparatus for determining pentoses and pentosans by distillation with 

hydrochloric acid. 

of 1.06 sp. gr. The flask is closed with a two-hole rubber stopper, 
one opening of which is fitted to the connecting tube* of a condenser 
and the other to a small separatory funnel. The latter is preferably 
of cylindrical form with graduation marks at 30 c.c. and 60 c.c. The 
flask is then placed in a bath of Rose's alloy (1 pgh*t lead, 1 part tin 
and 2 parts bismuth, melting near 100 C.), which, after Beating just 
beyond the point of fusion, is brought up slightly above th6 level of 
the bottom of the flask. The distillate is received in a graduated 
cylinder; when 30 c.c. of liquid have passed over, which should re- 
quire from 10 to 11 minutes, 30 c.c. more of the hydrochloric acid of 
1.06 sp. gr. are added from the separatory funnel. The process is con- 
tinued in this way until fr drop of the distillate shows no pink colora- 

* Z. Ver. Deut. Zuckerind., 46, 21, 195. 

t Jour. f. Landwirtsch. (1900), 355, (1901), 7. 



SPECIAL QUANTITATIVE METHODS 451 

tion with aniline-acetate paper (see p. 375). From 9 to 12 portions of 
30 c.c. usually require to be distilled over, depending upon the amount 
of furfural. The distillation is then suspended and the furfural de- 
termined by precipitation with phloroglucin. 

Preparation of Phloroglucin.* Dissolve a small quantity of phlo- 
roglucin in a few drops of acetic anhydride, heat almost to boiling and 
add a. few drops of concentrated sulphuric acid. A violet color indi- 
cates the presence of diresorcin. A phloroglucin which gives more than 
a faint coloration may be purified by the following method: 

Heat in a beaker about 300 c.c. of hydrochloric acid (sp. gr., 1.06) 
and 11 gms. of phloroglucin, added in small quantities at a time, stirring 
constantly until it has almost entirely dissolved. Some impurities may 
resist solution, but it is unnecessary to dissolve them. Pour the hot 
solution into a sufficient quantity of the same hydrochloric acid (cold) 
to make the volume 1500 c.c. Allow it to stand at least over night 
better several days to allow the diresorcin to crystallize out, and 
filter immediately before using. The solution may turn yellow, but 
this does not interfere with its usefulness. In using it, add the volume 
containing the required amount to the distillate. 

Precipitation of Phloroglucide. The distillate obtained by the 
method previously described is treated in a 500-c.c. lipped beaker with 
a measured volume of phloroglucin solution, so that the amount of 
phloroglucin is about double that of the furfural expected. The solu- 
tion first turns yellow, then green and finally becomes almost black 
when the amorphous dark-green precipitate of furfural phloroglucide, 
CnH 8 O4, begins to deposit. The liquid is then made up to 400 c.c. 
with the 12 per cent hydrochloric acid (1.06 sp. gr.) and allowed to 
stand over night. The solution, after testing with aniline-acetate 
paper to make sure that all furfural has been precipitated, is filtered 
through a weighed Gooch crucible; the precipitate of phloroglucide is 
brought carefully upon the asbestos and washed with 150 c.c. of water 
in such a way that the water is not entirely removed from the crucible 
until the y^ry last. The crucible is then placed upon a support, so 
that the bottom is free to the air, and dried for 4 hours in a boiling- 
water bath; i| is then placed in a weighing bottle, cooled in a desiccator 
and weighed. The increase in weight is the amount of furfural phloro- 
glucide which is calculated to furfural, pentose or pentosan according to 
the table of Krober (Appendix, Table 22). 

The weights of pentose in Krober's table are the averages of the 
corresponding weights of xylose and arabinose. The weights of pen- 
* Bull. 107 (revised), U. S. Bur. of Chem., p. 54. 



452 SUGAR ANALYSIS 

tosan are obtained by multiplying the corresponding weights of pen- 
tose by the factor 0.88, which represents the ratio of ?iC 5 Hi O 5 to 
(C 5 H 8 O4) n or ijj$. The table of Krober has a range for weights of 
phloroglucide between 0.030 and 0.300 gms. For weights of phloro- 
glucide outside of these limits Krober gives the formulae: 

For weight of phloroglucide "a" under 0.03 gm. 

Furfural = (a + 0.0052) X 0.5170 gm. 

Pentoses = (a + 0.0052) X 1.0170 gm. 

Pentosans = (a + 0.0052) X 0.8949 gm. 

For weight of phloroglucide " a " over 0.300 gm. 

Furfural = (a + 0.0052) X 0.5180 gm. 

Pentoses = (a + 0.0052) X 1.0026 gm. 

Pentosans = (a + 0.0052) X 0.8824 gm. 

The factor 0.0052 represents the weight (5.2 mgs.) of phloroglucide, 
which remains dissolved in the 400 c.c. of acid solution. 

For weights of phloroglucide which exceed 0.5 gm. it may be found 
necessary to dry for a longer period than 4 hours in order to attain 
constancy in weight. It is always better in making the determination 
to regulate the weight of material so that the amount of phloroglucide 
falls within the range of the table. 

Precautions and Limitations. In making the determination of 
pentosans by the method of acid distillation, several precautions should 
be noted. It is important first that the heat be applied to the flask in 
such a way that charring of solids upon the surface of the glass above 
the liquid be avoided. Such charring is very apt to occur when the 
flask is heated over the open flame or upon wire gauze; the use of the 
metal bath for heating is for this reason to be preferred. It is also im- 
portant that the distillate be perfectly clear, and free from suspended 
impurities, before adding the solution of phloroglucin. With sub- 
stances which contain much oil or wax, fatty decomposition products 
are sometimes carried over into the distillate; in determining pentoses 
in the urine of herbivorous animals, benzoic acid (a decomposition prod- 
uct of hippuric acid) is distilled over in considerable amount. In all 
such cases the distillate must be filtered from suspended matter before 
precipitating the furfural with phloroglucin. 

Two important limitations of the distillation method for determin- 
ing pentoses should be mentioned. 1. Furfural is formed from other 
substances than pentoses (the so-called furfuroids). 2. Other sub- 
stances, which form a precipitate with phloroglucin, are distilled over 
besides furfural (the so-called furaloids). 



SPECIAL QUANTITATIVE METHODS 453 

" Furfuroids." The formation of furfural from glucuronic acid 
and oxy cellulose has already been considered (p. 375). The presence of 
glucuronic acid in urine, or of oxycellulose in plant substances, will in- 
troduce, therefore, a certain error in the determination of pentoses in 
such materials. Cross and Bevan * for this reason propose that the 
names furfurose, furfurosan or furfuroid be used to designate the fur- 
fural-yielding complex of plants. The researches of Tollens show, 
however, that the pentosans are by far the most important of the fur- 
fural-yielding groups; the term pentosans, though not a perfectly correct 
expression, seems destined to remain until more accurate methods are 
devised for determining the different furfural-yielding groups. 

The distillates obtained by boiling cellulose, starch, sucrose, fructose, 
glucose and other hexose carbohydrates with hydrochloric acid give 
with phloroglucin a small yield of phloroglucide corresponding to 0.5 to 
1.0 per cent pentosans. Whether the reacting substance in such dis- 
tillates is furfural, oxymethylfurfural or mixtures of these has not 
been definitely determined. A slight error is, nevertheless, introduced 
into the pentose, or pentosan, determination by the phloroglucin 
method and the chemist should always bear this fact in mind when only 
small amounts of phloroglucide are obtained. 

"Furaloids." - The distillation of other products, which give pre- 
cipitates with phloroglucin, besides furfural has also been long recog- 
nized. Methylfurfural, which is obtained by the distillation of methyl- 
pentoses with hydrochloric acid, forms for example a red precipitate 
with phloroglucin, which, unless removed by solution in alcohol, as 
afterwards described, will give too high a weight of furfural phloro- 
glucide. In the same way oxymethylfurfural (see p. 620) which is 
formed in slight amounts by the action of hydrochloric acid upon 
fructose, sucrose and other hexose carbohydrates, forms a precipitate 
with phloroglucin. 

Frapsf has estimated that the amount of foreign products ("fur- 
aloid ") in the hydrochloric-acid distillate of different plant substances 
may vary from 7 to 23 per cent of the crude furfural. The " furaloid " 
is decomposed according to Fraps by redistilling the acid distillates; 
the pure furfural thus obtained is precipitated with phloroglucin, the 
weight of phloroglucide corresponding to the amount of furfural-yield- 
ing bodies (pentosans or furfuroids); the difference between the 
weights of phloroglucide for distillate and redistilled distillate corre- 
sponds to the amount of furaloid-yielding bodies, the exact nature of 

* Gross and Bevan's "Cellulose" (1895), p. 99. 
t Am. Chem. Jour., 25, 501. 



454 SUGAR ANALYSIS 

which Fraps did not determine. Furaloid does not seem to be formed 
from the pure pentose sugars. 

Precipitation of Furfural by Means of Barbituric Acid. Jager 
and linger * have suggested barbituric acid for precipitating furfural 
in presence of foreign distillation products. Cellulose, starch, sucrose 
and other hexose carbohydrates give hydrochloric-acid distillates 
which, though reacting with phloroglucin, form no precipitate with 
barbituric acid. Jager and Unger claim that the reagent offers, there- 
fore, a more accurate means of estimating pentosans. 

In making the precipitation the hydrochloric-acid distillate is treated 
with a solution of pure barbituric acid in hydrochloric acid of 1.06 sp. gr., 
using 8 parts of barbituric acid to 1 part of estimated furfural. The 
solution is stirred and after standing 24 hours the yellow granular precip- 
itate filtered into a Gooch crucible, washed with water and dried for 4 
hours at 105 C. The weight of precipitate is increased by 0.0049 gm. 
for the amount of substance dissolved in the 400 c.c. of acid solu- 
tion. 

The reaction between furfural and barbituric acid proceeds as fol- 
lows: 

yCO-NH v /CO-NHv 

C 4 H 3 - CHO+H 2 C ' ; CO =C 4 H 3 - CH C ( ; C0+H 2 0. 

X CO-NH X X CO-NH X 

Furfural (96) Barbituric acid (128) Condensation product (206) . 

One hundred parts of condensation product thus correspond to 46.6 parts 
of furfural. 

The barbituric-acid method for determining pentosans offers several 
good features, but the process has not been tried sufficiently as yet by 
chemists to form a conclusion as to its reliability. 

Jolles's Method of Determinating Pentoses. Jolles f has recently 
proposed a method for determining pentoses which differs in several 
particulars from that of Tollens. The substance to be distilled is 
placed in a 1500 c.c. flask with 200 c.c. of 12 per cent hydrochloric acid; 
the flask is heated, while a current of steam is passed through the 
liquid, the distillation being regulated so that the volume of solution 
does not fall at any time below 100 c.c. By distilling the furfural with 
steam the formation of humus substances is said to be prevented and a 
quantitative yield of furfural obtained. The process is continued until 
1 c.c. of the distillate shows no coloration with Bial's orcin reagent 
(p. 382); 100 c.c. of the distillate (usually between 2 and 3 liters) 

* Ber., 35, 4440; 36, 1222. 

t Sitzungsber. Wiener Akad., 114 (II b), 1191 (1905). 



SPECIAL QUANTITATIVE METHODS 455 

are neutralized with sodium hydroxide, and then made faintly acid 
to methyl orange with a few drops of half-normal hydrochloric acid. 
A measured volume of T Vnormal sodium-bisulphite solution is then 
added, and the solution allowed to stand 2 hours. The amount of bisul- 
phite, remaining after the reaction with the furfural, is then titrated 
back with T Vnormal iodine solution, using starch solution as indicator. 
The difference between the volumes of bisulphite and iodine solutions 
gives the amount of bisulphite which entered into combination with 
the furfural. The reaction between the two is expressed by the equa- 
tion : 

/OH 
C4H 3 O CHO-{-NaHSO 3 = C4H 3 O CH 

x SO 3 Na 

The titration of an aliquot, which is less than 5 per cent of the total 
distillate, involves a very great multiplication of any experimental 
errors. Jolles's process has not as yet demonstrated its superiority over 
the much shorter and simpler method of Tollens. 

The method of Tollens for determining pentoses gives good results 
with pure arabinose or xylose but, as has been shown, yields only 
rough approximations in the case of the various furfuroids. Even in 
the case of pure pentosans the calculation of furfural to a mixture of 
araban or xylan in equal amounts, when perhaps the pentosan itself 
may consist almost entirely of one substance, may involve an error of 
several per cent in the calculation. In certain plant exudations, as 
cherry gum, the pentosans consist almost entirely of araban; in the 
hemicelluloses of certain woods, as the beech, almost entirely of xylan; 
in the encrusting substances of most cellular tissues of variable mix- 
tures of araban and xylan. Until accurate methods are available for 
the estimation of xylan and araban, and for the determination of oxy- 
cellulose and other furfuroids, the calculation of furfural to a mixture 
of xylan and araban in equal amounts can be regarded only as a con- 
ventional approximation. 

Applications of Pentosan Method. The determination of pen- 
tosans, notwithstanding certain limitations of the method, has found 
numerous applications in the assay of plant gums, in the analysis of 
feeding materials, in the examination of forestry products and in other 
ways. A single example of such application is given in the analysis of 
paper stock. Krober,* for example, gives the following determinations 
of pentosans in different raw materials used in paper manufacture. 

* Jour. f. Landwirtsch. (1901), 1. 



456 



SUGAR ANALYSIS 
TABLE LXXXI 



Material. 


Pentosans 
calculated to 
ash-free dry 
substance. 


Mechanical wood pulp 


Per cent. 
12 24 


Mechanical wood pulp 


11 93 


Cotton 


1 03 


Linen 


2 20 


Bleached straw. . ... 


26 76 


Bleached raw cellulose (soda process) .... 


6 41 


Bleached raw cellulose (sulphite process) 


7.09 



An application of the above results to a special problem, which may 
confront the paper chemist, is taken from the work of Tollens.* 

A sample of newspaper is known to be made up of mechanical wood pulp 
and sulphite cellulose; it is desired to know the percentages of each which 
were used. The sample of paper upon analysis showed 10 per cent pentosans 
calculated to ash-free dry substance. Calling the percentage of pentosans in 
the ash-free dry substance of mechanical wood pulp 12 per cent and of sulphite 
cellulose 7 per cent, then 

X 100 = 60 per cent mechanical wood pulp. 



12-7 
12-10 
12-7 



X 100 = 40 per cent sulphite cellulose. 



For other applications of the method the chemist is referred to the 
original paper by Tollens. 

DETERMINATION OF METHYLPENTOSES AND METHYLPENTOSANS 
The conversion of methylpentoses into methylfurfural by distilla- 
tion with hydrochloric acid was described on p. 377. The method for 
determining methylpentoses, or methylpentosans, is based upon de- 
termining the amount of methylfurfural which is thus produced. The 
details of the method, which were first worked out by Tollens and 
Ellett,f and further elaborated by Tollens and Mayer, { are practically 
the same as described for the determination of the pentoses. The 
same apparatus (Fig. 177) is used and the substance is distilled with 
12 per cent hydrochloric acid (1.06 sp. gr.) until a drop of the distillate 
gives no yellow coloration with aniline-acetate paper. The methyl- 
furfural is then precipitated with phloroglucin and the solution allowed 
to remain over night, when the red precipitate of methylfurfural 

* Reprint Papier-Zeitung (1907), p. 17. 

t Ber., 38, 492. 

t Z. Ver. Deut. Zuckerind. (1907), 620; Ber., 40, 2441. 



SPECIAL QUANTITATIVE METHODS 457 

phloroglucide is filtered, washed, dried and weighed in exactly the 
same manner as described for furfural phloroglucide. 

The weight of methylfurfural phloroglucide is then calculated 
either to rhamnose by the table of Ellett and Tollens or to fucose by 
the table of Mayer and Tollens. The rhamnose, CHaCaHgC^ H 2 O. 
is calculated to rhamnosan (CH3C 5 H 7 04)n by multiplying by the 
factor 111 = 0.80; and the fucose, CHsCsHgOs, to fucosan by the 
factor ||| = 0.89. The combined table giving the weights of rham- 
nose, rhamnosan, fucose, fucosan, and methylpentosan (mixture of 
equal parts rhamnosan and fucosan) corresponding to different weights 
of methylfurfural phloroglucid is given in the Appendix (Table 23) . 

Instead of the tables the following formulae may be used in which 

Ph is the weight in grams of methylfurfural phloroglucide. 

Fucose = 2.66 Ph - 12.25 Ph 2 + 0.0005. 

Rhamnose = 1.65 Ph - 1.84 Ph 2 + 0.0100. 

Methylpentosan = 1.85 Ph - 6.25 Ph 2 + 0.0040. 

Fucose decomposes slower than rhamnose with hydrochloric acid, 
so that the distillation must be continued longer. More decomposition 
products of methylfurfural are consequently formed in distilling fucose 
with a corresponding less yield of phloroglucide. 

Methylfurfural, according to Fromherz,* may also be estimated by 
precipitation with barbituric acid in the same manner as described for 
furfural. The reaction takes place according to the equation: 

/CO-NH X 

CH 3 C 4 H 2 O - CHO +H 2 C CO 

X CO-NH X 

/CO-NH, 

= CH 3 C 4 H 2 CH - C ; , CO +H 2 0. 

X CO-NH X 

Methylfurfural (110) Barbituric acid (128) Condensation product (220) 

Two parts of condensation product thus correspond to exactly one 
part of methylfurfural. The yellow crystalline precipitate is filtered 
in a Gooch crucible, washed with water and then dried for 5 hours in a 
steam bath. The precipitate is then weighed, and after correcting for 
its slight solubility in the 12 per cent hydrochloric acid (2.29 mgs. in 
100 c.c.), calculated to methylfurfural by dividing by 2. 

According to Jolles f methylfurfural may also be determined by his 
method of steam distillation and titration with bisulphite and iodine 
solutions. The reaction between bisulphite and metyhlfurfural is 
similar to that described for bisulphite and furfural, and the details of 
the two methods are exactly alike. 

* Z. physiol. Chem., 60, 241. 

t Ann., 361, 41. 



458 SUGAR ANALYSIS 

DETERMINATION OF PENTOSES AND METHYLPENTOSES IN MIXTURE 

Method of Tollens and Ellett. The method of determining 
pentoses and methylpentoses in mixture was first worked out by 
Tollens and Ellett,* and is based upon the solubility of methylfurfural 
phloroglucide, and the insolubility of furfural phloroglucide in warm 95 
per cent alcohol. 

In making the determination the material is distilled with 12 per 
cent hydrochloric acid, the distillate precipitated with phloroglucin, 
and the mixed phloroglucides of furfural and methylfurfural filtered in a 
Gooch crucible, dried and weighed according to the usual process. 

The crucible containing the mixed phloroglucides is then placed in a 
smaller beaker with 95 per cent alcohol which is heated nearly to boil- 
ing. The brown-colored solution is then sucked off through the cru- 
cible by means of a filter pump, and the extraction with hot 95 per 
cent alcohol repeated twice more in the same way. The crucible con- 
taining the insoluble furfural phloroglucide is then dried for 2 hours in a 
hot-water bath and reweighed in a weighing bottle. The residual 
weight of furfural phloroglucide is then calculated to pentoses or pen- 
tosans and the loss in weight, due to methylfurfural phloroglucide, 
calculated to methylpentoses, or methylpentosans, by means of the 
respective tables or formulae. 

Trials of this method of separation upon known mixtures of pentoses 
with methylpentoses were made by Ellett and Tollens, and by Mayer 
and Tollens with very close agreements. 

Modification by Hay wood of the Tollens-Ellett Method. Haywood,f 
who has recently tested the method of Tollens and Ellett, believes that 
a correction should be made for the slight solubility of the furfural 
phloroglucide in 95 per cent alcohol. Experiments made by Hay wood 
upon the phloroglucide obtained from pure arabinose showed that for 
varying weights of substance, and extracting 3 to 5 times with alcohol, 
a very uniform weight of about 0.0037 gm. was always dissolved. Hay- 
wood believes the substance thus dissolved to be occluded phloroglucin 
and not phloroglucide. The following slight modification of the Tollens- 
Ellett method is proposed by Hay wood: 

Place the Gooch crucible containing the mixed phloroglucides in a 
100-c.c. beaker and pour into the crucible 30 c.c. of 95 per cent alcohol 
heated to 60 G. Place the beaker for 10 minutes in a water bath 
heated to 60 C. Remove the beaker and crucible and suck from the 

* Z. Ver. Deut. Zuckerind. (1905), 19. 
t Bull. 105, U. S. Bur. of Chem., p. 112. 



SPECIAL QUANTITATIVE METHODS 459 

latter all alcohol remaining tnerein with a suction pump. Repeat this 
alternate extraction and sucking dry of the precipitate 3 to 5 times, 
according to the color of the nitrate obtained. After the final ex- 
traction place the Gooch crucible in a water oven and dry four hours, 
making the final weighing in a closely stoppered glass weighing bottle. 

The difference in weight between the furfural phloroglucide plus 
methylfurfural phloroglucide first obtained and the furfural phloro- 
glucide remaining after extraction with alcohol, minus 0.0037, repre- 
sents the amount of methylfurfural phloroglucide present, from which 
the methylpentose or methylpentosan is calculated by the tables or 
formulae. 

To obtain the weight of pentosans, subtract the corrected weight 
of methylphloroglucide from the weight of the mixture and calculate 
according to Krober's tables or formulae. 

DETERMINATION OF GALACTOSE OR GALACTAN 

Tollens * and his co-workers have developed a method for estimat- 
ing galactose, and its higher condensation product galactan (C 6 Hi 05) n , 
which is based upon a determination of the mucic acid formed by oxi- 
dation of the substance with nitric acid. The oxidation of galactose to 
mucic acid according to theory proceeds as follows: 

C 6 Hi 2 6 + 2HNO 3 = C 6 H 10 O 8 + 2H 2 + 2NO. 

Galactosg^lSO) Mucic acid (210) 

100 part^of gatactose thus equal 116.66 parts of mucic acid. In actual 
experiment only about 75 per cent of the weight of galactose is obtained 
as mucic acid. This yield, however, is fairly constant for the given 
conditions of analysis, so that the weight of mucic. acid multiplied by 
1J gives the weight of galactose. 

The method of Tollens as employed by the Association of Official 
Agricultural Chemists f is as follows: 

Extract a convenient quantity of the substance, representing from 
2.5 to 3 grams of the dry material, on a hardened filter with 5 suc- 
cessive portions of 10 c.c. of ether; place the extracted residue in a 
beaker about 5.5 cm. in diameter and 7 cm. deep, together with 60 c.c. 
of nitric acid of 1.15 sp. gr., and evaporate the solution to exactly one- 
third its volume in a water bath at a temperature of 94 to 96 C. 
After standing 24 hours, add 10 c.c. of water to the precipitate, and 
allow it to stand another 24 hours. The mucic acid has in the mean- 
time crystallized but it is mixed with considerable material only par- 

* Ann., 227, 223; 232, 187. 

f Bull. 107 (revised), U. S. Bur. of Chem., p. 55. 



460 SUGAR ANALYSIS 

tially oxidized by the nitric acid. Filter the solution, therefore, through 
filter paper, wash with 30 c.c. of water to remove as much of the nitric 
acid as possible, and replace the filter and contents in the beaker. 
Add 30 c.c. of ammonium-carbonate solution, consisting of 1 part 
ammonium carbonate, 19 parts of water and 1 part strong ammonium 
hydroxide, and heat the mixture on a water bath, at 80 C., for 15 min- 
utes, with constant stirring. The ammonium carbonate takes up the 
mucic acid, forming the soluble mucate of ammonia. Then wash the 
filter paper and contents several times with hot water by decant ati on, 
passing the washings through a filter paper, to which finally transfer 
the material and thoroughly wash. Evaporate the filtrate to dry ness 
over a water bath, avoiding unnecessary heating which causes decom- 
position; add 5 c.c. of nitric acid of 1.15 sp. gr., thoroughly stir the 
mixture and allow to stand for 30 minutes. The nitric acid decomposes 
the ammonium mucate, precipitating the mucic acid; collect this on a 
tared filter or Gooch crucible, wash with from 10 to 15 c.c. of water, 
then with 60 c.c. of alcohol and a number of times with ether; dry at 
the temperature of boiling water for 3 hours, and weigh. Multiply 
mucic acid by 1.33, which gives galactose and multiply this product by 
0.9 which gives galactan. 

The method of Tollens has been used considerably by Schulze and 
Steiger * for determining galactan groups in different plants of the 
Leguminosse and also by Bauer f for estimating galactose jf$ lactose 
in the urine. * ^ 

The presence of large amounts of foreign organic matter hinders 
the precipitation of mucic acid, and in case of only small amounts of 
the latter may prevent its separation entirely. The tendency of the 
method is, therefore, to give too low rather than too high results. 

FERMENTATION METHODS FOR DETERMINING SUGARS 

A method for estimating sugars has been described (p. 299) which is 
based upon the change in polarization which the solution undergoes 
after fermenting with yeast. 

The fermentation methods for determining sugars are more usually 
carried out by weighing or measuring the carbon dioxide which is 
evolved. The theoretical yield of carbon dioxide from glucose, accord- 
ing to the equation C 6 Hi 2 O 6 = 2 C 2 H 5 OH + 2 CO 2 , is 48.88 per cent. 
In actual experiments only about 45 per cent of CO 2 is obtained, this 
figure varying, however, by several per cent according to the variety 

* Landw. Vers. Stat., 36, 11; 36, 438, 465. 
t Z. physiol. Chem., 61, 159. 



SPECIAL QUANTITATIVE METHODS 



461 



of yeast, influence of non-sugars and other conditions. The weight of 
carbon dioxide obtained during a normal fermentation multiplied by 
the factor 2.2 will give the approximate amount of fermentable hexose 
sugars present. The fermentation method is employed almost en- 
tirely for determining small percentages of sugar,- and has found its 
widest application in the determination of glucose in urine. 

Direct Method by Weighing Carbon Dioxide. The most accurate 
method for determining the yield of carbon dioxide upon fermentation 




Fig. 178. Apparatus for determining sugars from weight of carbon dioxide given 

off by fermentation. 

is shown in Fig. 178. A known amount of the solution is sterilized in a 
small flask, then cooled and inoculated with a pure culture of yeast. 
The flask is then connected by means of a condenser with a train of 
absorption tubes, or bulbs. Bulb I (Fig. 178) contains a few cubic 
centimeters of water, the U-tubes II and III contain calcium chloride 
for removing all moisture from the current of gas, the Liebig potash 
bulb IV, which has been previously weighed, serves to absorb the 
carbon dioxide, and the safety tube V, containing calcium chloride and 
soda lime, prevents back absorption of water, or carbon dioxide, from 
the outside air. The fermentation is allowed to proceed either at room 
temperature, or, if desired, at 30 C., in which case the flask is immersed 
in a water bath carefully maintained at this temperature. At the end of 
1 to 2 days, when no more gas passes through the bulb I, the tube V is 
connected with the aspirator bottle B, the pinchcock at p, which is 
previously closed, opened and a slow current of air, freed from carbon 
dioxide by passing through potassium hydroxide solution, led through 



462 



SUGAR ANALYSIS 



the apparatus. At the end of an hour the liquid in the flask is heated 
nearly to boiling, while a current of cold water circulates through the 
condenser; in this manner the last traces of dissolved carbon dioxide 
are expelled from the liquid. The aspiration is continued for another 
hour, when the potash bulb IV is disconnected and reweighed. The 
increase in weight gives the amount of carbonic acid. 

The more usual process, in the fermentation method of estimating 
sugars, is to estimate the carbon dioxide by measuring the volume of 
gas; 1 c.c. of evolved carbon dioxide (at C. and 760-mm. atmospheric 
pressure) corresponds to 1.96 mgs. carbon dioxide or about 4 mgs. of 
glucose. For determining sugars by this method special forms of appa- 
ratus known as fermentation saccharometers have been devised, of which 
the two forms devised by Einhorn and by Lohnstein are selected as 
examples. 

Einhorn's Fermentation Saccharometer.* This apparatus, which 
is designed for the estimation of small amounts of glucose in diabetic 

urine, is shown in Fig. 179. One 
gram of commercial pressed "yeast is 
shaken thoroughly in the graduated 
test tube with 10 c.c. of the urine. 
The mixture is then poured into the 
bulb of the saccharometer, the ap- 
paratus being inclined so that the 
graduated tube is completely filled. 
The saccharometer is then set aside 
for 20 to 24 hours at ordinary 
temperature. If the urine contains 
sugar, fermentation will usually be- 
gin in about 30 minutes. When the 
fermentation is finished the volume 
of gas is measured in the graduated 
tube, the divisions of which indicate 
cubic centimeters of gas and also the 
approximate fractions of per cent 
glucose. If the urine contains more 
than 1 per cent glucose it must first 
be diluted with water, the reading of the saccharometer being then 
multiplied by the degree of dilution. For diabetic urines of straw 
color and a specific gravity of 1.018 to 1.022 it is recommended to dilute 
twice; of 1.022 to 1.028 sp. gr. 5 times, and 1.028 to 1.038 sp. gr. 10 times. 
* Circular of information. 




Fig. 179. Einhorn's fermentation 
saccharometer. 



SPECIAL QUANTITATIVE METHODS 



463 



It is always desirable in making the test to make a duplicate de- 
termination upon a normal urine. The latter should show at most 
only a small bubble of gas at the top of the tube; should a larger 
amount of carbon dioxide be obtained with normal sugar-free urine, the 
yeast is probably impure and the determination should be repeated. 
If the suspected urine shows no more gas than the control experiment 
the absence of glucose is indicated. 

Lohnstein's * Fermentation Saccharometer. In Lohnstein's sac- 
charometer (Fig. 180) the liquid is fermented over mercury in a closed 
bulb; the carbon dioxide, which is evolved, forces 
the mercury into an upright tube, the amount of 
displacement indicating the per cent of glucose 
present. 

In making a determination the detachable scale 
S is hung in position over the open end of the tube 
T, and a quantity of mercury poured into the bulb 
B until its level in the tube is just opposite the 
zero mark of the scale. The standard weight of 
mercury, necessary for the adjustment, accompanies 
each instrument. 

A small piece of pressed yeast is rubbed with 2 to 
3 times its volume of ordinary water to a thin paste ; 
0.5 c.c. of the urine, or other liquid to be tested, is 
then measured with a special pipette into the bulb; 
the pipette is rinsed into the bulb with a little ordi- 
nary water and 2 to 4 drops of the yeast water 
added. The glass stopper, which should be evenly 
greased, is then inserted, and turned so that the 
small opening on its inner surface comes directly 

opposite a similar opening in the stem of the bulb. Fi g- 180 -- Lohnstein ' s 
. , ,. ,, fermentation sac- 

Any pressure of air, due to inserting the stopper, is c h a rometer 

thus released. The stopper is again slightly turned, 
so as to seal the contents of the bulb hermetically, and then securely 
fastened by the weight W. The apparatus is then set aside until fer- 
mentation is finished, which is indicated by the stationary position of the 
mercury column. The length of time necessary for completing the test 
will depend upon the temperature but does not ordinarily exceed 1 day at 
20 C.; if an incubator is available the time may be shortened con- 
siderably by fermenting at 35 C. When fermentation is finished the 
scale division opposite the top of the mercury column indicates the 
* Miinchener med. Wochenschr. (1899)-, No. 50; also circular of information. 




464 SUGAR ANALYSIS 

percentage of sugar; for percentages of sugar below 2.0 the scale may 
be read to 0.01 per cent and for percentages between 2.0 and 10.0 
to 0.05 per cent. The scale is calibrated upon one side for 20 C. and 
upon the other for 35 C.; if the readings be made at intermediary 
temperatures the percentage of sugar is calculated by interpolating. 
Thus: 

The reading of the mercury column at 25 C. was 4.0 on the 20 C. 
scale and 3.6 on the 35 C. scale. The corrected percentage of sugar is 

then 3.6 + 4 '!? ~ ^ 6 (35 - 25) = 3.87 per cent. 

oO ZO 

Instead of finding the weight or volume of carbon dioxide the per- 
centage of fermentable sugar may also be calculated from the amount 
of alcohol which is found by the action of yeast, or from the difference 
in specific gravity of the solution before and after fermentation. A 
valuable check upon the accuracy of the results obtained by the fer- 
mentation methods is to determine the loss in reducing sugars by 
means of Fehling's solution. 

COLORIMETRIC METHODS FOR DETERMINING SUGARS 

A number of colorimetric methods have been devised for determin- 
ing small amounts of different sugars in solution. The first process of 
this kind was due to Dubrunfaut who determined small percentages of 
glucose by comparing the color, which was produced by heating the 
solution with alkalies, with the colors of solutions containing known 
amounts of pure glucose, which had been similarly treated. 

In addition to the alkalies many of the special reagents, used in 
making color and spectral reactions, such as a-naphthol, resorcin, etc., 
have been employed for the colorimetric estimation of sugars. The 
principal requirement in the use of such reagents for quantitative pur- 
poses is that the color produced must be perfectly soluble and of a 
fair degree of stability. The insoluble, or evanescent, colors, which 
are produced in many of the reactions for sugars, are valueless for 
colorimetry. 

For making accurate comparisons of intensity of color, a special 
apparatus, called a colorimeter, must be used. The colorimeter of 
Duboscq is one of the best known and is selected for description. 

Duboscq's* Colorimeter. The colorimeter of Duboscq, as mod- 
ified by Pellin, is shown in Fig. 181. The apparatus consists of an 
upright case, the front and sides of which are in one piece B, and 
hinged to the back. At the bottom of the case is a shelf S, containing 

* Circular of information. 



SPECIAL QUANTITATIVE METHODS 



465 



two circular openings, above which rest the two cylinders C and C'. 
The latter are very carefully constructed, being closed at the bottom 
by disks of glass whose upper and lower surfaces are perfectly plane 
parallel. Two immersion rods of solid glass, T and T' the ends of 





M 



Fig. 181. 



Fig. 182. 



Duboscq's colorimeter. 



which are also plane parallel are attached to movable slides in the 
back of the case and can be raised or lowered within the cylinders. 
The height of the lower surface of each rod above the bottom of its 
cylinder is indicated upon a scale, which by means of a vernier can be 
read to 0.1 mm. The colorimeter is illuminated by light from the re- 
flector M, which from its opposite surfaces gives either bright or diffused 
light according to the requirements of sensibility. The light, as shown 
in Fig. 182, passes upward through each cylinder and immersion rod to 



466 SUGAR ANALYSIS 

the prisms P and P', from which it is reflected upwards into the tele- 
scope A. The field, when the telescope is focused, consists of a circle 
F, divided into equal parts, exactly resembling the double field of a 
polariscope. Daylight is to be preferred for illuminating the colorim- 
eter although artificial white, or monochromatic, light may be used 
according to requirement. In preparing the instrument for use, the 
mirror must be adjusted so that both halves of the field appear of ex- 
actly equal intensity. 

The sugar solution which is to be tested is placed in one cylinder 
and the standard solution, containing a known percentage of the same 
sugar, in the other, both solutions having been previously treated under 
similar conditions with alkali or other color-producing reagent. The 
door of the case is then closed and the rod immersed in the solution to 
be tested to some convenient scale division, as 100 mm., 50 mm., etc., 
at which point the color of its half of the field should be of suitable in- 
tensity for comparison. The other rod is then immersed in the cylinder 
of standard solution, and lowered or raised until the two halves of the 
field are of equal intensity. The heights of the immersion rods above 
the bottoms of the cylinders will then be inversely proportional to the 
depth of color and hence to the amount of sugar in solution. The cal- 
culation is made as follows: 

If A = the elevation of rod in standard solution, 
B = the elevation of rod in solution to be tested, 
P = the per cent of sugar in standard solution, 
X = the per cent of sugar in solution to be tested, 
AXP 



then X = 



B 



Example. 50 gms. of a glucose solution of unknown strength were made 
up to 500 c.c. with water, adding 5 c.c. of dilute NaOH solution (solution I). 

One gram of pure glucose was dissolved in water and the solution made up 
to 500 c.c. adding also 5 c.c. of the same NaOH solution (solution II). 

Both solutions were heated in a hot-water bath for the same length of time 
and after cooling compared in a Duboscq colorimeter. 

When the immersion rod in solution I was set at 100 mm., the immersion 
rod in solution II gave equal intensity to the field at 160.2 mm. 



1 

Then VL = 1-60 gms. of glucose in the 500 c.c. of solution I, or 3.2 
per cent in the original sample. 

Johnson* has recommended heating with alkaline picric-acid solu- 
tion for the colorimetric determination of glucose. Picric acid is reduced 
* Mon. sclent., Ill, 13, 939. 



SPECIAL QUANTITATIVE METHODS 467 

by glucose and other sugars in alkaline solution to picramic acid, 
the deep red color of which is sharply developed by less than 0.01 per 
cent of sugar. As stable color standards Johnson recommends solu- 
tions of ferric acetate, or of ferric chloride and acetic acid, which have 
been prepared so as to match the color produced by a known weight of 
sugar under the conditions of the method. 

Many of the color reactions of sugars are affected by the presence 
of organic or mineral impurities; the usefulness of colorimetric methods 
in estimating sugars is for this reason largely curtailed. 

Ehrlich's Colorimetric Method for Estimating Caramel. Ehrlich* 
has devised a colorimetric method for estimating caramel, in which the 
standard of comparison is saccharan. This dark-colored caramel sub- 
stance is produced by heating sucrose in a flask immersed in oil to 
about 200 C. under vacuum. The residue, after extracting with 
boiling methyl alcohol, is dissolved in water, filtered and evaporated. 
The saccharan, Ci 2 Hi 8 9 , is obtained as a dark-brown residue (about 
20 per cent of the weight of sucrose) which is easily pulverized to an 
amorphous powder. One part of saccharan in 10,000 of water colors 
the solution a deep brown, which is intensified by the addition of 
alkalies. Saccharan is not precipitated by lead sub-acetate solution, 
so if the latter is used for precipitating other coloring substances from 
solutions of sugars, molasses, etc., the percentage of saccharan in the 
neutralized filtrates may be estimated by comparison in a colorimeter 
with a solution containing a known weight of saccharan. The amount 
of saccharan multiplied by 5 indicates the approximate amount of 
sucrose destroyed by superheating during manufacture. 

Stammer's Colorimeter. Colorimeters are employed in technical 
sugar analysis for grading sirups, for estimating the decolorizing power 
of bone black or other clarifying agent, and for many other purposes 
in which degree of color, and not determination of color-producing sub- 
stance, is desired. For determinations of this kind colored plates, or 
disks, of glass are usually employed as a standard of comparison, the 
results being expressed in units of an arbitrary color scale. 

A colorimeter which is used extensively in the sugar industry is 
that of Stammer f (Fig. 183). The general principle of this apparatus 
is the same as that of Duboscq. The liquid to be tested is placed in 
the cylinder a, which is closed by a glass plate at the bottom. The 
measuring tube c, also closed at the bottom by a glass plate, fits 

* Z. Ver. Deut. Zuckerind., 59, 746. Proceedings, Seventh International Con- 
gress of Applied Chem., Sect., V, p. 92. 

t Stammer's " Zuckerf abrikation " (1887), p. 747. 



468 



SUGAR ANALYSIS 



loosely into a and can be' raised or lowered to any desired level. The 
comparison tube b, which is open at the bottom, is joined to c, the 

two being moved in conjunction by a 
slide in the back of the instrument. 
The colorimeter is illuminated by a re- 
flector at the bottom, the light passing 
upward through b and c into the prisms 
in d which produce the same double- 
field effect as in the Duboscq apparatus. 
In operating the colorimeter the 
standard plate of colored glass is placed 
upon tube 6, which together with tube c 
is then raised or lowered until the in- 
tensity of shade for solution and color 
plate is the same in both halves of 
the field. A millimeter scale upon the 
back of the instrument marks the eleva- 
tion of the measuring tube above the 
bottom of the cylinder, thus indicating 
the thickness of the column of liquid. 

Stammer gives a solution which 
matches the standard plate for a scale 
reading of 1 mm., a color value of 100. 
The color value of any liquid is found 
by dividing 100 by the reading of the 
scale in millimeters. 

In measuring the color of sugars, 
molasses, etc., a weighed amount of 
substance is dissolved in water, made 
up to a definite volume and, if the solution is not clear, filtered. 
The color value of the solution is then calculated either to the 
original amount of substance, or to a polarization of 100, according 
to requirement. 

Example. 20 gms. of a sugar, polarizing 92.4, were dissolved to 100 c.c. 
and filtered. The solution gave a reading of 15 mm. upon Stammer's colorim- 
eter. Then W = 6.666 the color value of the solution. The color value 
calculated to 100 parts sugar would be 20 : 6.666 :: 100 : x = 33.33. The 
latter calculated to 100 polarization would give 92.4 : 33.33 :: 100 : x = 36.07. 

For determining the decolorization produced by bone black the 
color value of the solution is taken before and after filtration. If the 




Fig. 183. Stammer's colorimeter. 



SPECIAL QUANTITATIVE METHODS 469 

original solution is too dark for reading in the colorimeter, it is diluted 
with water, in which case the filtered solution is also diluted to the same 
density. 

Example. An unfiltered sirup diluted to 10 degrees Brix gave a reading 
of 8 mm., or l | a = 12.5 color units, using a Stammer colorimeter. The liquid, 
after filtering through bone black, and diluting to 10 degrees Brix gave a read- 
ing of 40 mm., or W = 2.5 color units. The amount of color removed by the 

10 K 9 ^ 

bone black is then - X 100 = 80 per cent. 

12.5 

A table of reciprocals (Appendix, Table 25) will be found convenient 
for converting the scale measurements of Stammer's colorimeter into 
color units. 

DETERMINATION OF SUGARS BY WEIGHING AS HYDRAZONES AND 

OSAZONES 

The varying solubility of the different hydrazones and osazones of 
sugars in presence of impurities, or of other similar derivatives, has 
prevented the general employment for quantitative purposes of this 
means of separating sugars. In certain cases, however, where the 
hydrazone, or osazone, is characterized by great insolubility a fairly 
accurate determination of several of the sugars has been found possible. 

Determination of Arabinose as Diphenylhydrazone. According 
to Neuberg * arabinose is precipitated quantitatively by treating the 
sirupy solution of sugar with a slight excess of diphenylhydrazine. 
Sufficient alcohol is added to form a perfectly clear solution, and the 
mixture heated to boiling for 30 minutes in a water bath in a flask con- 
nected with a reflux condenser. The solution is cooled, allowed to 
stand for several hours and the white crystalline hydrazone filtered 
into a weighed Gooch crucible. After washing with a few cubic centi- 
meters of cold alcohol, the crucible is dried in a water oven and weighed. 

The weight of arabinose diphenylhydrazone, CsHioC^N N(-C 6 H 5 ) 2 , 
is calculated to arabinose, C 5 Hi 5 , by multiplying by M = 0.4747. 
This method of analysis has been used by Neuberg for estimating 
arabinose in the urine and by Maurenbrecher and Tollens f for de- 
termining arabinose in cacao. 

Determination of Mannose as Phenylhydrazone. The property 
of mannose in forming with phenylhydrazine a very insoluble hydra- 
zone, discovered by Fischer and Hirschberger,t has been used for the 
quantitative estimation of mannose. The precipitation, according to 

* Ber., 36, 2243. f Ber., 39, 3578. | Ber., 21, 1805. 



470 SUGAR ANALYSIS 

Bourquelot and Herissey,* is best accomplished by treating a 3 to 6 
per cent solution of the sugar with an excess of phenylhydrazine acetate 
at a temperature not above 10 C. After standing 24 hours, the white 
crystalline hydrazone is filtered upon a weighed Gooch crucible, washed 
with a little cold water, dried in a water oven and weighed. The solu- 
bility of the hydrazone is 0.04 gm. in 100 c.c. of solution, and the 
weight of precipitate should be corrected accordingly. 

The weight of mannose phenylhydrazone, CeH^C^^HCeHs, is cal- . 
culated to mannose, C 6 Hi 2 O 6 , by multiplying by |f = f , or 0.6666. 
The method is well adapted for determining mannose in presence of 
other sugars and has been employed by Pellet f for estimating small 
amounts of mannose in sugar-cane molasses. 

Determination of Fructose as Methylphenylosazone. According 
to Neuberg { fructose may be determined with a fair approximation 
by precipitating as its methylphenylosazone, CeHioO^^CHsCeH^. 
About 10 c.c. of the concentrated sugar solution are treated with a 
slight excess of methylphenylhydrazine, and sufficient alcohol added 
to give a clear solution. If other sugars than fructose are present the 
solution is slightly warmed and allowed to stand 24 hours for the sepa- 
ration of any insoluble hydrazones of mannose, galactose, etc. After 
removing any precipitate by suction, the filtrate is treated with 4 c.c. 
of 50 per cent acetic acid, heated 5 to 10 minutes upon the water bath, 
and then set aside in the cold for 24 hours. The reddish-yellow crys- 
tals of the osazone are filtered in a weighed Gooch crucible and cal- 
culated to fructose, C 6 Hi 2 6 , by multiplying by J|J = 0.4663. The 
method is only approximate as 10 per cent or more of the osazone 
remains in solution. By using a very cold freezing mixture the sepa- 
ration has been made almost quantitatively. 

SIEBEN'S METHOD FOR ESTIMATING FRUCTOSE 

Sieben in 1884 proposed a method for determining fructose which 
is based upon the destruction of this sugar when heated with dilute 
hydrochloric acid. The method was designed for estimating fructose 
in honey, sirups and other products which contain glucose. The latter 
sugar, like other aldoses, is much less susceptible to the destructive 
action of acids, so that the difference in the reducing power of a solu- 

* Compt. rend., 129, 339. 

t Bull, assoc. chim. sucr. dist., 16, 1181; 18, 758. 

t Ber., 35, 960. 

Z. Ver. Deut. Zuckerind. (1884), 837, 865. 



SPECIAL QUANTITATIVE METHODS 471 

tion before and after treatment by Sieben's process is taken as the 
equivalent of the fructose present. 

In making the determination 100 c.c. of the solution, which should 
contain about 2.5 gms. of total reducing sugars, are heated in a 
250-c.c. graduated flask with 60 c.c of 6-normal hydrochloric acid 
(36.47 X 6 = 218.8 gms. HC1 per liter) for 3 hours in a boiling-water 
bath. A funnel is placed in the neck of the flask to prevent evapo- 
ration. The solution is then cooled and neutralized with 6-normal 
sodium hydroxide (40 X 6 = 240 gms. NaOH per liter), of which from 
56 to 58 c.c. are usualty required. The contents of the flask are then 
made up to 250 c.c., filtered and the reducing sugars determined in 25 c.c. 
of the filtrate by Allihn's method. The reducing sugar thus found is 
calculated as glucose, and the difference in reducing sugar before and 
after the acid treatment estimated as fructose. 

According to Sieben only about 1.5 per cent of the total glucose is 
destroyed under the conditions of his method. Herzf eld * found, how- 
ever, that the destruction of glucose may exceed 7 per cent. Wiech- 
mannf also showed that the complete destruction of the fructose is not 
always assured so that " the results obtained by this method must be 
received with some caution." Dammullert found that the destructive 
power of the acid depended largely upon the ratio of glucose to fructose; 
with mixtures of glucose and fructose in equal proportions only 1.28 
per cent of glucose was destroyed, with pure glucose on the other hand 
the loss exceeded 28 per cent. Attempts to modify and improve the 
process so as to overcome these objections have not been wholly suc- 
cessful. 

* Z. Ver. Deut. Zuckerind., 35, 967. 
t Wiechmann's "Sugar Analysis" (1898), p. 54. 
4 Z. Ver. Deut. Zuckerind., 38, 751. 



CHAPTER XVI 

COMBINED METHODS AND THE ANALYSIS OF SUGAR MIXTURES 

IN previous chapters upon polariscopic and chemical methods 
several instances were given of the application of certain processes to 
the analysis of sugar mixtures. In the present chapter the problem of 
determining several sugars in presence of one another will be taken up 
in somewhat fuller detail. 

If the sum of the specific rotations, copper-reducing powers or 
other properties of the different sugars in a mixture can be expressed 
by a sufficient number of equations, the problem of determining the 
percentage of each sugar in the mixture may be solved by simple alge- 
braic analysis. By thus combining the results of several distinct 
methods it is possible by indirect means to make an analysis of many 
sugar mixtures with a fair degree of accuracy. The combinations of 
methods, which have been proposed for this purpose, are almost number- 
less and only a few examples will be chosen to illustrate the general 
principle. The methods will be grouped for convenience under 
(1) Combined polariscopic methods; (2) Combined reduction methods; 
(3) Combined polariscopic and reduction methods. 

COMBINED POLARISCOPIC METHODS 

If two sugars, A and B, exhibit a known variation in specific rota- 
tion under different conditions of polarization, then the percentages, x 
and y, of the two sugars may be determined by means of the following 
equations: 

ax + by = lOOWz), (1) 

a'x+b'y=lW[a] D ', (2) 

in which [oi\D and [U]D are the specific rotations of the mixture A + J5, 
a and a' the known specific rotations of sugar A and b and b f the known 
specific rotations of sugar B, under the respective conditions of (1) and (2) . 
By determining [O\D and [O\D , the percentages x and y are readily cal- 
culated. 

As an example of this method of analysis the determination of 
glucose and fructose by polarization at 20 C. and 87 C., under the 

472 



COMBINED METHODS AND ANALYSIS OF SUGAR MIXTURES 473 

conditions previously described (p. 296), is given. If the []*> and [a]% 
of glucose are +52.5 and of fructose 92.5 and 52.5 respectively, then 
the []g and [a]g of a mixture containing x per cent glucose and y per 

cent fructose are 

52.5 x - 92.5 y = 100[a] 

52.5 x - 52.5 y = 100[a]g 

By determining the [a]" and []^ of the mixture the percentages of 
glucose and fructose are readily calculated. 

Any other temperature, at which the [a]^ of each of the sugars is 
known, may of course be taken instead of 20 C. and 87 C. The re- 
sults as thus calculated are of course only approximate and require to 
be corrected for the influence of concentration. 

In addition to varying the temperature, changes of condition may 
be accomplished by making one polarization in neutral and the other 
in acid solution; or one polarization in water, and the other in some 
other solvent; or one polarization in the absence and the other in the 
presence of borax or other substance; in all of which changes of con- 
dition a definite known alteration in the polarizing power of one or 
both sugars must be produced. Obviously the greater the degree of 
this change in polarizing power, the less will be the influence of ex- 
perimental errors. 

COMBINED REDUCTION METHODS 

If two sugars, A and B, exhibit a known variation in reducing 
power under different conditions of analysis, then the percentages x 
and y of the two sugars may be determined by means of the general 
equations : 

ax + by = 100/2, (1) 

a'x+b'y = lQOR', (2) 

in which R and R f are the reducing powers of the mixture A + B, a and 
a' the known reducing powers of sugar A, and b and b f the known re- 
ducing powers of sugar B, under the respective conditions of (1) and (2). 
By determining R and R', the percentages x and y are readily calculated. 

A good example of the application of the above formulae is given by 
Soxhlet's * well-known method for determining two sugars in mixture. 

A comparison of the reducing powers of different sugars upon Feh- 
ling's copper solution (Soxhlet's formula) and Sachsse's mercury solu- 
tion was made by Soxhlet with the following results: 

* J. prakt. Chem. (1880), 21, 300; Konig's " Untersuchung " (1898), 217. 



474 



SUGAR ANALYSIS 



TABLE LXXXII 
Showing Relative Reducing Power of Fehling's and Sachsse's Solutions 



Sugar. 


1 gm. sugar in 1 per cent solu- 
tion reduces 


Milligrams of sugar in 1 per cent 
solution reduce 


Fehling's solu- 
tion. 


Sachsse's solu- 
tion. 


100 c.c. 

Fehling's solu- 
tion. 


100 c.c. 
Sachsse's solu- 
tion. 


Glucose 


c.c. 

210.4 
194.4 
202.4 
196.0 
148.0 
202.4 
128.4 


c.c. 

302.5 
449.5 
376.0 
226.0 
214.5 
257.7 
197.6 


Mgs. 
475.3 
514.4 
494.1 
510.2 
675.7 
494.1 
778.8 


Mgs. 

330.5 
222.5 
266.0 
442.0 
466.0 
388.0 
506.0 


Fructose 


Invert sugar 


Galactose 


Milk sugar ... 


Milk sugar hydrolyzed 
Maltose 





The results show that the various sugars differ very decidedly in 
their relative reducing powers upon the two reagents, glucose, for ex- 
ample, reducing more Fehling's but less Sachsse's solution than fructose. 

The combined influences of two sugars, A and B, in their reducing 
powers upon Fehling's and Sachsse's solutions may be expressed as 
follows : 

Let x = gms. of reducing sugar A in 100 c.c. of the 1 per cent sugar 

solution. 
Let y = gms. of reducing sugar B in 100 c.c. of the 1 per cent sugar 

solution. 
Let a = c.c. of Fehling's solution reduced by 1 gm. of sugar A in 100 

c.c. of solution. 
Let b = c.c. of Fehling's solution reduced by 1 gm. of sugar B in 100 

c.c. of solution. 
Let o! = c.c. of Sachsse's solution reduced by 1 gm. of sugar A in 

100 c.c. of solution. 
Let b' = c.c. of Sachsse's solution reduced by 1 gm. of sugar B in 

100 c.c. of solution. 

Let F = c.c. of Fehling's solution reduced by 100 c.c. of sugar solution. 
Let S = c.c. of Sachsse's solution reduced by 100 c.c. of sugar solution. 
Then ax + by = F, 

and a 'x + b'y = S. 

For a mixture of x per cent glucose and y per cent fructose, and 
taking Soxhlet's values in Table LXXXII for a, 6, o! and b', the equa- 
tions would be 

210.4 x + 194.4 y = F 

302.5 x + 449.5 y = S. 



COMBINED METHODS AND ANALYSIS OF SUGAR MIXTURES 475 

By determining the values F and S of the mixture of sugars, the 
percentages x and y are readily calculated. 

In using the above, or other combined reduction methods, the con- 
stants a, bj a' and &' should be determined empirically by the chemist 
for the particular sugars with which he is working. 

As another example of combined reduction methods may be men- 
tioned Kjeldahl's* process of determining the reducing power of the 
mixture of two sugars in both dilute and more concentrated solution, 
using respectively 15 c.c. and 50 c.c. of mixed Fehling's solution ac- 
cording to the details of his reduction method (p. 424). The relative 
differences in the copper-reducing powers under the two conditions of 
analysis are not sufficiently pronounced, however, to afford a reliable 
basis of calculation and the method has been generally condemned. 

The use of combined polariscopic, or of combined reduction, methods 
alone for analyzing sugar mixtures has largely given place to the more 
accurate procedure of combining these two distinct physical and chem- 
ical methods in one. 

COMBINED POLARISCOPIC AND REDUCTION METHODS 
1. ANALYSIS OF MIXTURES CONTAINING TWO SUGARS 
The calculation of the percentages of two sugars in mixture by com- 
bining the results of polarization and copper reduction was first at- 
tempted by Neubauerf in 1877, and the principle of his indirect method 
has been that of most subsequent modifications. In the earlier methods 
of this class the total reducing power of the mixture was determined as 
glucose, fructose or invert sugar, the percentage thus obtained being 
taken as the total amount, or sum, of the sugars present. In the case 
of two sugars, A and B, the percentages x and y of each were expressed 
by the formula 

x + y = R 

in which R was the percentage of total reducing sugar determined as 
glucose, fructose or invert sugar. The results calculated by such a 
formula have, however, only an approximate value, as the difference 
in copper-reducing power of the two sugars A and B has not been taken 
into account. 

The error last mentioned has been largely obviated in the later 
methods of this class through the use of reduction factors (p. 421) by 
means of which the copper-reducing power of a sugar can be con- 
verted into the equivalent of any other reducing sugar which is selected 
as a standard of comparison. For the latter purpose glucose is usually 
* Z. analyt. Chem., 35, 345-347. f Ber., 10, 827. 



476 



SUGAR ANALYSIS 



selected, this being the most common of the reducing sugars and the 
one most easily obtained in a pure condition. 

It was shown upon p. 421 that the different monosaccharides bear 
a constant ratio to glucose for the same weight of reduced copper. 
This ratio was given for several sugars and was found by Allihn's method 
to be 0.915 for fructose, 0.958 for invert sugar, 0.898 for galactose, 
0.983 for xylose and 1.032 for arabinose. 

For a solution containing a mixture of monosaccharides, the sum of 
the glucose equivalents of the individual sugars should equal the total 
reducing sugars estimated as glucose. This is shown in the following 
experiments by Browne, * who mixed known weights of different sugars 
and compared the calculated glucose equivalents with the amount of 
glucose corresponding to the reduced copper obtained by Allihn's 

method. 

TABLE LXXXIII 
Showing Glucose Equivalents of Mixed Reducing Sugars 



Sugars. 


Grams sugar in 25 c.c. 


Total 
weight 
of 
sugars. 


Glucose equiv- 
alent. 


Error. 


1. 


2. 


3. 


Calcu- 
lated. 


Found. 


Glucos6 fructose 


0.0967 
0.0484 
0.0461 
0.0231 
0.0740 
0.1786 
0.0893 
0.0265 
0.0681 
0.0155 
0.1853 
0.0927 
0.2162 
0.1081 
0.1513 
0.0757 
0.0495 
0.0248 
0.1371 
0.0646 


0.0904 
0.0452 
0.1408 
0.0704 
0.0198 
0.0585 
0.0293 
0.0960 
0.0175 
0.1070 
0.0569 
0.0285 
0.0429 
0.0215 
0.0433 
0.0217 
0.1535 
0.0768 
0.0226 
0.0822 




Gram. 

0.1871 
0.0936 
0.1869 
0.0935 
0.0938 
0.2371 
0.1186 
0.1225 
0.0856 
0.1225 
0.2422 
0.1212 
0.2