Full text of "Collected Papers Of Wallace Hume Carothers On High Polymeric Substances"

8]<OU 158494 >m

OSMANIA UNIVERSITY LIBRARY

OillNo 7f/-7/ -2? C Accession No. 1.*2-

Authoi
Title

This book shdulcl be retvfrnec/ori or before the date last marked below .

HIGH POLYMERS

A SERIES OF MONOGRAPHS ON THE CHEMISTRY,
PHYSICS AND TECHNOLOGY OF HIGH POLYMERIC

SUBSTANCES

Editorial Board

R. E. BURK, Cleveland, Ohio; H. MARK, New York, N. Y.
G. S. WHITBY, Teddington, Middx.

Volume I

Collected Papers of W. H. Carothers on High Polymeric Substances
Edited by H. MARK and G. S. WHITBY

INTERSCIENCE PUBLISHERS, INC

New York

COLLECTED PAPERS OF

WALLACE HUME CAROTHERS

ON HIGH POLYMERIC SUBSTANCES

Edited by

H. MARK G. STAFFORD WHITBY

Adjunct Professor of Organic Chemistry, Director of Chemical Research, Chemical Research

Brooklyn Polytechnic Institute Laboratory, Tedddington, Middx.

With 35 illustrations

1940

INTERSCIENCE PUBLISHERS, INC.

New York

INTERSCIENCE PUBLIvSHERS, INC.
215 Fourth Avenue, New York, N. Y.

Printed in U. S. A.
MACK PRINTING COMPANY, EASTON, PA.

INTRODUCTION TO "HIGH POLYMERS"

While physics is directing its main effort to exploring the structure of
matter and the nature of energy in the sub-atomic realm, organic chemistry
is becoming increasingly concerned with molecular structures of (a) greater
and greater complexity, and (b) greater and greater size. In both these
latter respects, the study of organic chemistry tends, in its most fruitful and
significant latter-day developments, to become less the wide-ranging study
of "the compounds of carbon" and more the study of the "organic" aspects
of carbon compounds, i. e. t the study of those aspects bearing on the struc-
ture and behavior of living organisms. In both respects, the behavior of
the living cell represents the final problem to the understanding of which
the whole development trends.

Two conditions are necessary for the development of life, viz.:

(a) A structural system, acting as a skeleton, which protects the func-
tionally-active living matter from disturbances and at the same time per-
mits the continuation of diffusion and of chemical reactions in the medium
contained in the system.

(b) Another system of highly reactive and diffusible compounds which
serve as reactants and catalysts iti the processes necessary to maintain life
and direct to a common final result a great number of chemical reactions
and colloidal factors.

In the present generation ever since what was once described as
"Chemical Physiology" and later "Physiological Chemistry" acquired full
status as "Biochemistry" much attention has been paid to the latter
condition, and much illuminating work has been carried out on the chemis-
try of metabolism and on catalysis and promoters of vital processes. At
the same time, arid independently of this, a keen interest has developed iri
the study of the natural skeletal and storage substances of organisms -in
cellulose, keratin, starch, rubber, etc. and as a result a new field of research
has defined itself, namely, the chemistry of high polymers.

During the last twenty years the study of this field has been steadily ex-
tended and intensified. Important forward steps can already be signalized
in the preparation of synthetic high-polymers; in the elucidation of the
mechanism and kinetics of polymerization processes; in the determination
of the chemical structure of natural and synthetic high-polymers; in the
determination and understanding of their colloidal behavior, and in the
development of formulas capable of describing quantitatively the proper-
ties of high-polymers. Generalized structural principles applicable to all
the large molecules in question have become apparent, and, although still

VI INTRODUCTION TO "HIGH POLYMERS"

far from satisfactory in all details, obviously represent a useful frame of
reference within which high-polymers may be classified and described. To
refine and improve the structural picture for every significant high-molecu-
lar material, in the hope of arriving at a fuller and more satisfying explana-
tion of its chemical, colloidal and physical properties, is one of the tasks to
which the further study of high -polymers must be addressed.

New representatives of high-polymers are continually being synthesized,
and thus offering for comparative quantitative investigation an increasingly
large body of experimental material. And, too, new experimental technics,
especially physical technics, are continually being applied to the study of
high-polymers with illuminating results. In short, the whole field is ex-
panding rapidly.

In these circumstances it is felt that a useful purpose should be served
by an attempt, such as this series of volumes represents, to bring together
our present knowledge of the subject in a manner both comprehensive and
critical. Such a survey of the subject should serve to throw into clear relief
the generalizations which it is already possible to make; it should anchor
the new branch of high-polymeric chemistry firmly to the fundamentals of
chemical science, and, generally, it should contribute to fostering and facili-
tating the further study of the subject.

In addition to being of purely scientific interest, high-molecular ma-
terials are of great general and economic importance. Much of our daily
life depends on their use, and a large proportion of the wealth of all civilized
countries is invested in industries which either use natural high-polymers as
raw materials or manufacture and fabricate synthetic polymers. It is cer-
tainly a conservative estimate if one assumes that more than 30% of all
graduate chemists employed in industries are engaged in work connected
with high -polymeric materials. In making such an estimate, one must bear
in mind the fact that high polymeric substances are closely connected with
the various branches of the textile industry dealing with natural and arti-
ficial fibers, the production of natural rubber and its synthetic congeners
and the fabrication of goods therefrom, many branches of the food industry,
the production of molded plastics, of lacquers, of synthetic resins and of
varnishes based on synthetic resins, the industries concerned with starch
and its conversion products, and so on. All who have had experience in
such industries cannot but be aware of the sore need for advancing our
fundamental knowledge of high-polymeric materials, and cannot but have
noted how, speedily, every forward step in such knowledge is reflected in
improvement in the technical processes and in the products of the relevant
industries.

INTRODUCTION TO "HIGH POLYMERS" VII

It may, further, be noted that a very substantial proportion of the stu-
dents obtaining university degrees in chemistry in various parts of the world
are likely to find themselves called upon to apply their knowledge to indus-
tries concerned in some degree or other with high-polymeric materials, and
hence should benefit directly by at least some acquaintance with the funda-
mentals of high-polymeric chemistry.

It should be added that, as notified in the title of the series, it is proposed
to include in the series, not only volumes on High Polymers in the strict
sense, but also certain volumes on "Allied Subjects." This is done because
it is thought undesirable to exclude from the survey certain solid systems,
such as, e. g., those represented by bitumen and clay, which, although not
strictly high-polymeric, nevertheless behave in many respects, especially
Theologically, in a manner closely related to, and in some cases partly de-
pendent on, high-polymers.

It should perhaps be mentioned that, as this series of volumes, although
employing the English language, is international in character, it has been
considered appropriately catholic that contributors to the volumes which
it will comprise should be free to follow in their various contributions either
English or American spellings, where these differ.

R. E. BURK,
H. MARK,
G. S. WHITBY,
May, 1940 Editors.

INTRODUCTION TO "COLLECTED PAPERS OF WALLACE HUME

CAROTHERS"

It is perhaps fair to say that the progress of organic chemistry in our
day is most distinctively characterized by two particularly attractive and
promising lines of research, namely by the investigation of biochemical
catalysts ferments, vitamins, hormones and auxins and by the sys-
tematic development of the chemistry of high polymeric substances.
Both these lines of research are founded solidly on the experimental
methods and theoretical viewpoints of classical organic chemistry. The
former reaches far into biology, physiology and medicine, and touches
some of the central problems of the functioning of living organisms. The
latter cuts deeply into the fundamentals of the structure of matter
especially of those parts of the structure of organisms which are not
obviously involved in their metabolism and, further, is the basis of a
field of chemical industry, already wide, and with great possibilities of
expansion.

In this series on high polymers, of which the present monograph is the
first volume, an attempt will be made to offer a collection of treatises
covering the most significant studies made in the field, in order to show
where the subject now stands and to facilitate the further development
which its study is surely destined to have.

Our present knowledge of high polymers is due to a considerable and
now rapidly growing number of workers, who have contributed essential
material from various points of view. Some have, with success, employed
chiefly the tried and well known methods of organic chemistry. Others
have, with illuminating results, applied the methods of modern physics
and physical chemistry to the study of high polymers. Again, the work
of numerous investigators, whose concern was not primarily or wholly
with the study of high polymers and polymerization, has contributed ex-
perimental facts and theoretical conceptions which form a not unimpor-
tant part of our present body of knowledge of the subject.

No investigator has excelled Wallace Hume Carothers in advancing
our knowledge of high polymeric chemistry and at the same time providing
a basis for the development of technically useful synthetic polymeric
materials. He may rightly be called the outstanding personality in this
new branch of organic chemistry. His brilliant experimental technique,
the abundance of his ideas, and the constructive manner in which he
applied his critical faculty, together with the excellent facilities at his
disposal and the competent collaborators who assisted him, enabled

VIII

INTRODUCTION TO "COLLECTED CAROTHERS' PAPERS" IX

Carothers to carry out in the short space of twelve years more than most
of us accomplish in a lifetime, and to produce an output which is at once
very considerable in volume and compact in character. His publications
are deservedly to be considered as "classical" : they will always remain an
essential part of the foundation on which the high polymeric chemistry
of the future will be erected. Their study should be undertaken by all
who wish to be familiar with the chemistry of high polymers. They are
highly inspiring, and at the same time they point out many of the diffi-
culties which work in this field is bound to meet.

With such thoughts in mind, it was felt that there could be no better
start for this series than to publish, as the first volume in it, a collection
of the papers embodying Carothers' studies of high polymers and closely
related matters. The two editors regard it as a privilege to prepare this
edition, not only because the issue of Carothers' papers in a convenient,
collected form will, it is felt, be an addition to chemical literature which
will be generally appreciated, but also because it will serve as an expres-
sion of the admiration and esteem which they and other workers in the
field of high polymers feel for a great chemist, whose premature death was
a severe loss to chemistry and humanity.

Carothers' work on polymerization and polymers divides itself into
two distinct series of studies devoted to (1) "Studies on Polymerization
and Ring Formation," (2) "Acetylene Polymers and their Derivatives."
The papers in the first of these series are reprinted in Part One of this
volume; those in the second, in Part Two. A few other papers, mostly
concerned with general discussions of the subject of polymerization, are
reprinted in Part Three.

Part One is concerned essentially with condensation polymers produced
by the recurring condensation of long, so-called polyfunctional molecules,
at both ends of which are reactive groups capable of reacting, with the
elimination of water, with the reactive end groups of other molecules of
the same of another type to yield long condensed-polymeric chains, the
products being, typically, polyamides, polyesters and polyanhydrides.
The work, in addition to contributing in the most important way to our
knowledge of condensation polymers, forms the basis of an interesting
industrial development, namely, the manufacture of a synthetic poly-
amide product ("Nylon"), which may be considered, chemically, as repre-
senting a simplified protein-like material, and which can be drawn into

X INTRODUCTION TO "COLLECTED CAROTHERS' PAPERS"

fibers ranging in size from finest textile fibers suitable for making hosiery
and having a tensile strength at least as great as that of natural silk to
coarser fibers suited to serve as "bristles." Not only is the work the basis
of the particular industrial development just mentioned; it also carries
possibilities of other practical developments in the future not yet realized
industrially.

Part Two is concerned with polymers, in the narrower sense of the
term, produced by the self-addition of conjugated systems, especially
conjugated butadienoid systems containing chlorine and other substituents.
In it, also, a good deal of attention is given to the chemical reactions of
vinylacetylene and divinylacetylene, two new polymers of acetylene,
which are the source of the butadienoid systems in question and are the
first readily-available compounds containing the interesting conjugated
enine system, H 2 C=CH C=C . This work is the basis on which there
has been developed the commercial manufacture of a synthetic rubber,
"Neoprene." A survey of the work is given in the Introduction to Part
Two.

A few early papers of Carothers, dealing with subjects other than
polymerization, are omitted from this volume. A bibliography of Caro-
thers' publications relative to polymerization is given (p. 423 et seq.)\
it includes United States patents.

H. MARK

G. STAFFORD WHITE Y
April, 1940

ACKNOWLEDGMENTS

The editors and publishers wish to thank the numerous individuals and
organizations for their assistance in the preparation and publication of this
book and without whose valuable aid, this volume might not have been
possible.

We desire especially to mention the cooperation and encouragement
given the editors by du Pont de Nemours and Company, in which labora-
tories the greater part of Carothers' work was carried out.

We are deeply indebted to :

Dr. E. K. Bolton, Chemical Director of du Pont de Nemours and Com-
pany, in Wilmington, Del.

Dr. A. P. Tanberg, Director of the Experimental Station, and
Dr. A. W. Kenney, of the Experimental Station, Wilmington, Del.

We have also had the opportunity to discuss our plans with several
former collaborators, friends and pupils of W. H. Carothers and wish to
express our gratitude to them. We are particularly grateful to Dr. E. O.
Kraemer and Dr. H. Kienle for their valuable assistance.

It is through the cooperation of those societies and publishers who pub-
lished the original papers, that we were enabled to reprint them.

For this reason, we are very grateful to :

The American Chemical Society, and its Secretary, Mr. Charles W.
Parsons, who permitted the reprinting of all the papers from the
"Journal of the American Chemical Society," and the use of the
illustrations.

The Faraday Society, and its Secretary, Mr. G. S. W. Marlow, who
permitted the reprinting of a paper from the "Transactions of the
Faraday Society."

Mr. H. E. Howe, for permission for the reprinting of a paper from
"Industrial & Engineering Chemistry."

The Williams and Wilkins Company, Baltimore, Mel., for permission
to reprint a paper from the "Chemical Reviews."

Professor Roger Adams and the National Academy of Sciences, Wash-
ington, D. C., for permission to reprint the Biography of W. H,
Carothers by Roger Adams, and for the block for the frontispiece.

The index has been carefully prepared by Dr. Dora Stern, New York.

XII ACKNOWLEDGMENTS

The editors wish finally to express their gratitude for the cooperation
and assistance given them by Interscience Publishers, Inc. and especially,
by Dr. E. S. Proskauer of this firm.

H. MARK
G. S. WHITBY
July, 1940

CONTENTS

PAGE

Introduction to the Series, "High Polymers" V

Introduction to Volume I VIII

Table of Contents VIII

Biography by Roger Adams XV

Part One

Studies on Polymerization and Ring Formation 1

Introduction by H. Mark 3

I. An Introduction to the General Theory of Condensation Polymers. . . 4

II. Polyesters 17

III. Glycol Esters of Carbonic Acid 29

IV. Ethylene Succinates 42

V. Glycol Esters of Oxalic Acid 54

VI. Adipic Anhydride 63

VII. Normal Paraffin Hydrocarbons of High Molecular Weight Prepared by

the Action of Sodium on Decamethylene Bromide 68

VIII. Amides from e-Aminocaproic Acid 78

IX. Polymerization 81

X. The Reversible Polymerization of Six-Membered Cyclic Esters 141

XI. The Use of Molecular Evaporation as a Means for Propagating Chem-
ical Reactions 154

XII. Linear Superpoly esters 156

XIII. Polyamides and Mixed Polyester-Polyamides 165

XIV. A Linear Superpolyanhydride and a Cyclic Dimeric Anhydride from

Sebacic Acid 168

XV. Artificial Fibers from Synthetic Linear Condensation Super polymers. . 179

XVI. A Poly alcohol from Decamethylene Dimagnesium Bromide 190

XVII Friedel-Craf ts Syntheses with the Polyanhydrides of the Dibasic Acids . 192

XVIII. Polyesters from co-Hydroxydecanoic Acid 195

XIX. Many-Membered Cyclic Anhydrides (1) 202

XX. Many-Membered Cyclic Esters 212

XXI. Physical Properties of Macrocyclic Esters and Anhydrides. New

Types of Synthetic Musks 221

XXII. Stereochemistry and Mechanism in the Formation and Stability of

Large Rings 225

XXIII. e-Caprolactone and Its Polymers 235

XXIV. Cyclic and Polymeric Formals 240

XXV. Macrocyclic Esters 248

XXVI. Meta and Para Rings 259

XXVII. Polydecamethylene Oxide 263

XXVIII. Preparation of Macrocyclic Lactones by Depolymerization 265

XIII

XIV

CONTENTvS

Part Two

Acetylene Polymers and Their Derivatives 271

I. Introduction by G. Stafford Whitby 273

II. A New Synthetic Rubber: Chloroprene and Its Polymers 281

III, The Addition of Hydrogen Chloride to Vinylacetylcne 306

IV. The Addition of Hydrogen Bromide to Vinylacetylene, Bromoprene and

Dibromobutene 311

V. The Polymerization of Bromoprene (Third Paper on New Synthetic

Rubbers) 314

VI. Vinylethinylmagnesium Bromide and Some of Its Reactions 321

VII. Sodium Vinylacetylide and Vinylethinylcarbinols 323

VIII. a-Alkyl-/3-Vinylacetylenes 329

IX. l-Alkyl-2-chloro-l,3-butadienes and their Polymers (Fourth Paper on

New Synthetic Rubbers) 331

X. The Chlorination of the Hydrochlorides of Vinylacetylene 335

XI. Dichloro-2,3-butadiene-l,3 and Trichloro-l,2,3-butadiene-l,3 340

XII. The Addition of Thio--cresol to Divinylacetylene 344

XIII. The Action of Chlorine on Dinvinylacetylene 348

XIV. The Dihydrochloride of Divinylacetylene 357

XV. Halogen-4-butadienes-l,2. The Mechanism of 1,4-Addition and of a,y-

Rearrangcment 301

XVI. The Preparation of Orthoprenes by the Action of Grignard Reagents

on Chloro-4-butadiene-l,2 308

XVII. Mercury Derivatives of Vinylacetylene 372

XVIII. l-Halogen-2-vinylacetylenes 375

XIX. The Structure of Divinylacetylene Polymers 378

XX. The Addition of Alcohols to Vinylacetylene, etc 381

XXI. Homologs of Chloroprene and Their Polymers (Second Paper on New

Synthetic Rubbers) 384

XXII. The Synthetic Rubber Problem 391

Part Three

Miscellaneous Papers 399

I. Association Polymerization and the Properties of Adipic Anhydride. . 401

II. Ueber die angeblichen Isomerien bei cyclischen Oxalsaeure-estern 402

III. 6-Caprolactone 407

IV. Polymers and Polyf unctionality 409

Bibliography and Patents 423

Authors Index 433

Subject Index 439

A BIOGRAPHY

By ROGER ADAMS

Wallace Hume Car-others, who died on April 29, 1937, was born in Bur-
lington, Iowa, on April 27, 1896. His contributions to organic chemistry
were recognized as outstanding and, in spite of the relatively short span of
time for his productive accomplishments, he became a leader in his field
with an enviable international reputation.

His paternal forbears were of Scotch origin and settled in Pennsylvania
in prerevolutionary days. His father, Ira Hume Carothers, taught coun-
try school at the age of 19. Later he entered the field of commercial educa-
tion and for forty-five years has been engaged in that type of work as
teacher and vice-president.

His maternal ancestors were of Scotch-Irish stock and were also, for the
most part, farmers and artisans. They were great lovers of music, and
this may account for the intense interest in and appreciation of music which
Carothers possessed. His mother, who was Mary Evalina McMullin of
Burlington, Iowa, exerted a powerful influence and guidance in the earlier
years of his life.

On February 21, 1930, he married Helen Everett Sweetman of Wilming-
ton, Delaware. A daughter, Jane, was born November 27, 1937.

Wallace was the oldest of four children. His education began in the
public schools of Des Moines, Iowa. In 1914 he was graduated from the
North High School. As a growing boy he had zest for work as well as play.
He enjoyed tools and mechanical things and spent much time in experi-
menting. His school work was characterized by thoroughness and his high
school classmates testify that when he was called upon to recite, his answers
revealed careful preparation. It was his habit to leave no task unfinished
or done in a careless manner. To begin a task was to complete it.

He entered the Capital City Commercial College in the fall of 1914 and
was graduated in the accountancy and secretarial curriculum in July, 1915,
taking considerably less time than the average. He entered Tarkio College,
Tarkio, Missouri, in September, 1915, to pursue a scientific course, and
simultaneously accepted a position as assistant in the Commercial Depart-
ment. He continued in this capacity for two years and then was made an
assistant in English, although he had specialized in chemistry from the
time he entered college. During the World War the head of the depart-
ment in chemistry, Dr. Arthur M. Pardee, was called to another institution,
and Tarkio College found it impossible to secure a fully equipped teacher of

XVI A BIOGRAPHY

chemistry. Carothers, who previously had taken all of the chemistry
courses offered, was appointed to take over the instruction. Since he was
rejected as a soldier on account of a slight physical defect, he was free to
serve in this capacity during his junior and senior years. It is interesting
that during his senior year there were four senior chemistry-major students
in his class and every one of them later completed work for the doctorate,
studying in the universities of this country and abroad. Today they bear
testimony to the fact that as undergraduates they owed much to the in-
spiration and leadership of Carothers.

Upon entering college his interest in chemistry and physical sciences was
immediate and lasting, and he rapidly outdistanced his classmates in ac-
complishment. As a student he showed mature judgment and was always
regarded by his fellow students as an exceptional person. Invariably he
was the brightest student in the class regardless of the subject. Financial
necessity required that he earn a large portion of his educational expenses.
He always found time, however, to associate with the other students,
though he showed little interest for the boisterous enthusiasms of the aver-
age underclassman. During his last two years in college he was entrusted
with a number of student offices to which he gave freely of his time and
energy.

Leaving Tarkio College in 1920 with his bachelor of science degree, he
enrolled in the chemistry department of the University of Illinois where he
completed the requirements for the master of arts degree in the summer of
1921. His former instructor at Tarkio College, then head of the chemistry
department at the University of South Dakota, desired a young instructor
to handle courses in analytical and physical chemistry and was fortunate
in securing Carothers for this position during the school year, 1921-1922.
He went to South Dakota only with the intention of securing sufficient
funds to enable him to complete his graduate work, but the careful and
adequate preparation of his courses, as well as his care of the students under
his direction, showed that he could be a very successful teacher of chemistry.
He was still the same quiet, methodical worker and scholar, not forceful as
a lecturer, but careful and systematic in his contact with the students.
He always required adequate preparation of assigned work and was able
to get a large volume in student accomplishment.

Simultaneously with his teaching work he started to develop some inde-
pendent research problems. He was especially interested in the 1916 paper
of Irving Langmuir on valence electrons and desired to investigate some of
the implications it held in organic chemistry. Pursuing this idea he car-
ried out laboratory studies which were reported in his first independent

A BIOGRAPHY XVII

contribution to the Journal of the American Chemical Society, "The
Isosterism of Phenyl Isocyanate and Diazobenzene-Imide." His second
independent paper, published while still a student, was that on 'The Double
Bond." In this he presented the first clear, definite application of the
electronic theory to organic chemistry on a workable basis. He described
the electronic characteristics of the double bond and in essence included in
his discussion everything that has since been written on this particular
subject.

He returned to the University of Illinois in 1922 to complete his studies
for the degree of doctor of philosophy, which he received in 1924. His
major work was in organic chemistry with a thesis under the direction of
Dr. Roger Adams, on the catalytic reduction of aldehydes with platinum-
oxide platinum-black and on the effect of promoters and poisons on this
catalyst in the reduction of various organic compounds. His minors were
physical chemistry and mathematics. He exhibited the same brilliance in
all of his courses and in research which characterized his earlier accom-
plishments. Although specializing in organic chemistry, he was consid-
ered by the physical chemists to have a more comprehensive knowledge of
physical chemistry than any of the students majoring in that field. In
1920-1921 he held an assistantship for one semester in inorganic chemistry
and for one semester in organic chemistry. He was a research assistant
during 1922-1923, and during 1923-1924 held the Carr Fellowship, the
highest award offered at that time by the department of chemistry at
Illinois. During these two years his seminar reports demonstrated his wide
grasp of chemical subjects. The frequency with which his student col-
leagues sought his advice and help was indicative of his outstanding ability.
At graduation he was considered by the staff as one of the most brilliant
students who had ever been awarded the doctor's degree. A vacancy on
the staff of the chemistry department of the University of Illinois made it
possible to appoint him as an instructor in organic chemistry in the fall of
1924. In this capacity he continued with unusual success for two years,
teaching qualitative organic analysis and two organic laboratory courses,
one for premedical students and the other for chemists.

Harvard University, in 1926, was in need of an instructor in organic
chemistry. After carefully surveying the available candidates from the
various universities of the country, Carothers was selected. In this new
position he taught during the first year a course in experimental organic
chemistry and an advanced course in structural chemistry, and during the
second year he gave the lectures and laboratory instruction in elementary
organic chemistry.

XVIII A BIOGRAPHY

President James B. Conant, of Harvard University, was professor of
organic chemistry at the time that Carothers was instructor. He says of
him

"Dr. Carothers' stay at Harvard was all too short. In the brief space of time during
which he was a member of the chemistry department, he greatly impressed both his col-
leagues and the students. He presented elementary organic chemistry to a large class
with distinction. Although he was always loath to speak in public even at scientific
meetings, his diffidence seemed to disappear in the classroom. His lectures were well
ordered, interesting, and enthusiastically received by a body of students only few of
whom planned to make chemistry a career. In his research, Dr. Carothers showed even
at this time that high degree of originality which marked his later work. He was never
content to follow the beaten track or to accept the usual interpretations of organic reac-
tions. His first thinking about polymerization and the structure of substances of high
molecular weight began while he was at Harvard. His resignation from the faculty to
accept an important position in the research laboratory of the du Pont Company, was
Harvard's loss but chemistry's gain. Under the new conditions at Wilmington, he had
facilities for carrying on his research on a scale that would be difficult or impossible to
duplicate in most university laboratories. Those of us in academic life, however, always
cherished the hope that some day he would return to university work. In his death,
academic chemistry, quite as much as industrial chemistry, has suffered a severe loss."

In 1928 the du Pont Company had completed plans to embark on a new
program of fundamental research at their central laboratory, the Experi-
mental Station at Wilmington, Delaware. Carothers was selected to head
the research in organic chemistry. The decision to leave his academic po-
sition was a difficult one. The new place demanded only research and
offered the opportunity of trained research men as assistants. This over-
balanced the freedom of university life and he accepted. From then on
until his death his accomplishments were numerous and significant. He
had the rare quality of recognizing the significant points in each problem
he undertook and unusual ability for presenting his results in a most ex-
plicit and precise way, which led to clarity and understanding. In these
nine years he made several major contributions to the theory of organic
chemistry and discoveries which led to materials of significant commercial
importance. Dr. Elmer K. Bolton, Chemical Director of the du Pont
Company, writes concerning Carothers

"At the time the du Pont Company embarked upon its program of fundamental re-
search in organic chemistry in the Chemical Department, Dr. Carothers was selected to
direct this activity, because he had received the highest recommendations from Harvard
University and the University of Illinois, and was considered to have unusual poten-
tiality for future development. There was placed under his direction a small group of
excellently trained chemists to work on problems of his own selection. The results of
his work, extending over a period of nine years, have been of outstanding scientific in-

A BIOGRAPHY XIX

terest and have been considered of great value to the Company as they have laid the
foundation for several basically new developments of commercial importance.

"In our association with Dr. Carothers, we were always impressed by the breadth and
depth of his knowledge. He not only provided inspiration and guidance to men under
his immediate direction, but gave freely of his knowledge to the chemists of the depart-
ment engaged in applied research. In addition, he was a brilliant experimentalist.
Regarding his personal characteristics, he was modest, unassuming to a fault, most un-
complaining, a tireless worker deeply absorbed in his work, and was greatly respected
by his associates. He suffered, however, from a nervous condition which in his later
years was reflected in poor health and which became progressively worse in spite of the
best medical advice and care, and the untiring efforts of his friends and associates. His
death has been a great loss to chemistry and particularly to the Chemical Department.
In my judgment, he was one of the most brilliant organic chemists ever employed by the
du Pont Pompany."

His reputation spread rapidly; his advice was sought continually, not
only by his colleagues but also by chemists throughout the world. In 1929
he was elected Associate Editor of the Journal of the American Chemical
Society; in 1930 he became an editor of Organic Syntheses. He took an
active part in the meetings of the organic division of the American Chemical
Society. He was invited frequently to speak before various chemical
groups. He addressed the Johns Hopkins summer colloquium in 1935 on
"Polymers and the Theory of Polymerization." That year he also spoke
on the same subject before the Faraday Society in London, when his paper
was considered one of the outstanding presentations on the program. His
achievements were recognized by his election to the National Academy
of Sciences in 1936 the first organic chemist associated with industry to
be elected to that organization. During these years from 1928-1937
several attractive academic positions were offered him but he chose to re-
main to the end with the Company which had given him his opportunity for
accomplishment .

Carothers was deeply emotional, generous and modest. Pie had a lov-
able personality. Although generally silent in a group of people, he was a
brilliant conversationalist when with a single individual, and quickly dis-
played his broad education, his wide fund of information on all problems of
current life, and his critical analysis of politics, labor problems and business,
as well as of music, art, and philosophy. With all his fine physique he had
an extremely sensitive nature and suffered from periods of depression which
grew more pronounced as he grew older, despite the best efforts of his
friends and medical advisers. *

* This biography is an abbreviated reprint of Roger Adams, Wallace Hume Carothers,
1896-1937. National Academy of Sciences of the U, S., Biographical Memoirs. 20.
12th Memoirs. 1939.

PART ONE

STUDIES ON POLYMERIZATION AND RING
FORMATION

PART ONE

STUDIES ON POLYMERIZATION AND RING
FORMATION

INTRODUCTION

This series of articles extends from 1929 to 193(5 and comprises twenty-
eight papers.

Reactions which produce giant molecules can be divided into poly-
condensation and polymerization processes; Carothers has called them
C-polymers and A-polymers, respectively. The articles of Part 1 are
especially devoted to the study of C-polymerization. The following are
the outstanding problems which are open for investigation.

(a) If two bifunctional molecules, e. g., one dibasic acid and one glycol,
or diamide, react, two possibilities occur. The reaction can result (1) in
a chain polymer of lower or higher molecular weight, which still bears
either hydroxyl or carbonyl terminal groups or (2) in a smaller or larger
ring, which does not contain the reactive group. The following questions
arise: Under what conditions does either of these two possibilities take
place and what is the molecular weight of the resulting compound. Papers
I to XIV are chiefly concerned with this problem. A large number of
dibasic acids, glycols, diamides, co-hydroxy and ammo-carbonic acids were
included in the investigation and many formerly unknown compounds were
synthesized and described.

If 5- or 6-membered rings can be formed during a polycondensation of
the type described, the reaction almost invariably leads to ring formation ;
this may be put forward as a general result. There is still a certain possi-
bility also for the production of 7- and 8-membered rings, while larger
rings are not formed during such polycondensation reactions. This rule
is associated with the stereochemical points of view of Sachse and Mohr.

(b) The second problem is concerned with the behavior of the high
polymeric compounds, which can be produced by polycondensation, and
their possible technical utilization. Polyesters and poly amides may be
very readily cast into films or drawn into filaments which display very
interesting technical properties. Papers XV, XVIII and XIX deal in part
with this question. In general, as soon as the degree of polymerization
exceeds a number between 30 and 40, the material begins to exhibit strength
and extensibility. Such products can be cast or spun from solution and
also from the molten state; they can later undergo a stretching process

4 POLYMERIZATION AND RING FORMATION

which increases the strength and pliability of the resulting samples and also
causes the material to assume a lustrous appearance. This investigation
can certainly be regarded as an outstanding example of successful scientific
work, which bears at the same time considerable technical importance.
One of the polyamides, which has been prepared and investigated, was
selected for technical production, and is now coming into great importance
under the trade name of "Nylon."

(c) Another attractive line of investigation was the more thorough
study of the second type of compounds which result from polycondensation
reactions, namely poly-membered rings. Ruzicka has shown that such
compounds exhibit peculiar properties and are of considerable interest from
the point of view of their odor. In papers XX, XXI, XXII, XXIII and
XXVI, a new method for the preparation of such rings is described; their
formation and properties are studied. It turns out that the specific odor
is closely connected with the number of bonds and varies with it in a
characteristic way.

It is of importance to point out that an experimental method has been
developed to deal with such high polymeric chains or multi-membered
rings or mixtures of the two. Pure samples suitable for quantitative
characterization could hardly be obtained by the normal experimental
methods of organic chemistry. Therefore, a new type of molecular still
has been developed, which is briefly described in paper XI and which is
extensively applied in all these investigations.

In the following, each single paper is reprinted without any change;
a short abstract of its contents is given at the beginning in italics; and a
few additional references concerning the recent development and literature
are added. The purpose of this brief introduction is to emphasize the main
general results and to refer briefly to their scientific and technical im-
portance.

I. An Introduction to the General Theory of Condensation

Polymers*

This first paper contains a general introduction to the theory of polymers.
With the aid of a comprehensive table Carothers distinguishes between
two classes of polymers:

A-polymers, produced by recurring addition of monomers. The molecular
formula of the monomer is identical with that of the structural unit. The

*W. H. Carothers; Journ. Am. Chem. Soc., 51, 2548-59 (1929); Contribution No. 10
from the Experimental Station of E. I. du Pont de Nemours and Co.
Received April 13, 1929. Published August 7, 1929.

1. THEORY OF CONDENSATION POLYMERS 5

monomer can be obtained from the polymer by thermal or photochemical de-
composition; the polymer is formed by self -addition,

C-polymers, produced by recurring condensation of monomers. The
molecular formula of the monomer differs from that of the structural unit.
The monomer can be obtained from the polymer by a hydrolytic process;
the polymer is formed by a polyintermolecular condensation.*

Then the conditions for the formation of C-polymers are explored. The
term "functionality is introduced. A functional group in the monomer
is such an arrangement of atoms as might lead to a reaction step.
Such groups are, e. g., OH, NH 2 , COOH, SO 3 H, etc. According as
the monomer contains one, two, three, etc., of them it is called mono
functional, bifunctional, etc.

Multifunctional monomers can react intra- or intermolecularly . In the
first case they lead to rings, in the second to C-polymers.
If during a bifunctional reaction a 5- or 6-ring can be formed intra-
molecuarly, then it almost invariably happens that the reaction leads to
ring formation.

Bifunctional reactions which if intramolecular would lead to larger than
6 -rings generally proceed intermolecularly and give C-polymers. The
formation of 7 -rings, however, may possibly occur, while 8 -rings are much
less probable.

This empirical rule, which has of course also several exceptions is dis-
cussed in the paper from the point of view of the stereochemical considera-
tions of Sachse and Mohr.

Polymerization frequently leads to substances of very high molecular
weights, and the problem of the structure of high polymers is attracting
a great deal of attention, especially because such important materials as
rubber, cellulose, proteins and resins either are high polymers or have
certain properties which are common to high polymers.

The conditions which Berzelius (1) was concerned to recognize by the
term polymer, were the presence of the same atoms in the same proportions
in compounds having different molecular weights. These conditions are
satisfied by the members of a great many thousand pairs of compounds
which are not now regarded as polymers. Thus, of the compounds paracet-
aldehyde, butyric acids and hydroxycaproic acids, only the first would
now be considered a polymer of acetaldehyde, although there is nothing in
the conditions of the Berzelius definition to exclude the others. Hence,

* In modern literature A-polymers are frequently called polymerisation products,
while C-polymers are termed polycondensation products. The shorter terminology of
Carothers would certainly be practical.

6 POLYMERIZATION AND RING FORMATION

whatever the term polymer may mean now, it does not mean precisely
what Berzelius intended, and the conditions which he set up are not
sufficient to define it. In current attempts to define this term (2) it is still
stated that a polymer and its monomer must have the same atoms in the
same proportions. But this condition is not satisfied by the polyoxy-
methylenes (see Table I) which are universally considered to be polymers
of formaldehyde. It seems desirable, therefore, to attempt to formulate
a definition which will be in so far as possible in accordance with both
the current usage and the essential facts.

The structures of a good many polymers, including some of very high
molecular weights, are known either completely or in part and an examina-
tion of their formulas shows some interesting relationships (see Table I).
They are characterized by a recurring structural unit, so that if this is
represented by R , the structure of these polymers may be represented
in part by the general formula R R R R R R R R , etc., or
( R -) n . In this formula n may be small as in paracetaldehyde (n = 3),
or it may be very large as in the polyoxymethylenes. The end valences
may be united as they are in paracetaldehyde to form a ring, or they
may be saturated by univalent groups such as H and OH to form an
open chain of the type H( R ) n OH as they are in a-polyoxymethylene.
It seems probable that cellulose and silk fibroin are of this type, and in
any event it may be observed that no high polymer is certainly known to
be cyclic (3). There are polymers which do not conform to the type
( R ) n but those which do will be called linear whether the chain is open
or closed; and the subsequent discussion is concerned only with these.

The structural units R are bivalent radicals which, in general, are
not capable of independent existence. The presence of a recurring struc-
tural unit is, of course, characteristic of most organic compounds (e. g.,
CH 2 in aliphatic compounds), but in the case of polymers there exists
a molecule, the monomer, corresponding to the structural unit, and from
which the polymer may be formed or to which it may be degraded.

Examination of the formulas of Table I will show that two types of
polymers may be distinguished. In the first type, which includes paracet-
aldehyde, rubber, polystyrene and polyoxymethylenes, the molecular
formula of the structural unit is identical with that of the monomer, i. e. t
the formula of the structural unit, R , is isomeric with that of the
monomer. In the second type, which includes the polyethylene glycols,
cellulose and silk fibroin, the molecular formula of the structural unit
differs from that of the monomer by H 2 O, t . e., the monomer is H R OH.
The transformation of polymers of the first type into their monomers is

J. THEORY OK CONDENSATION POLYMERS 7

brought about simply by heating, and the reverse transformation (poly-
merization) occurs spontaneously or by the action of catalysts. In the
second type, degradation to the monomer occurs by hydrolysis, and if
the reverse process were to take place it would require the elimination of
water among many molecules. This would be polyintermolecular con-
densation (4).

These two classes will be distinguished as (1) addition or A polymers.
The molecular formula of the monomer is identical with that of the struc-
tural unit. The monomer can be obtained from the polymer by ther-
molysis or the polymer can be synthesized from the monomer by self-
addition. (2) Condensation or C polymers: the molecular formula of
the monomer differs from that of the structural unit. The monomer can
be obtained from the polymer by hydrolysis or its equivalent or the polymer
can be synthesized from the monomer by polyintermolecular condensation.
Polymerization then is the chemical union of many similar molecules
either (A) without or (C) with the elimination of simpler molecules (HaO,
HC1, NaCl, NH 3 , etc.) (5).

Assuming that polyintermolecular condensation exists, the above ex-
amples and definitions and their implications provide ample reason for
referring to this process as a type of polymerization. These examples,
of course, do not provide any proof that this process as distinct from
and independent of A polymerization does exist. This proof will appear
incidentally in the following discussion, which is concerned with the
general principles involved in the formation of condensation polymers.

Polyfunctional Compounds. Polyintermolecular condensation re-
quires as starting materials compounds in which at least two functional
groups are present in the same molecule (e. #., hydroxy acids, HORCOOH,
might lead to poly-esters, HORCOORCOORCOORCOORCOORCO ,
etc.; amino acids, to poly-amides, NH 2 RCONHCONHRCONHR-
CONHRCONHRCO , etc.).

Among compounds having more than one functional group, those of the
type x R y may be called bifunctional, R"x 3 , trifunctional, etc. In
these formulas R stands for a bivalent radical (R* for a trivalent radical)
and x and y for functional groups capable of reacting with each other in
a known fashion to form the new functional group z. Thus x R y >
R z , where R z evidently represents a structural unit which
will appear in the product and which may be present 1, 2, 3, ... n times.
Reactions of this type will be called bifunctional regardless of the num-
ber of molecules involved. Reactions of the type x R x + y R'- y
> z R z R' may be called bi-bifunctional. All such reac-

8 POLYMERIZATION AND RING FORMATION

tions may, at least by hypothesis, pass through the stage x R x R' y
which is equivalent to x R y, and for purposes of discussion they
may therefore be classed as simple bifunctional reactions. Reactions
of the type R'x 3 + Ry 2 > product will be called tri-bifunctional. Simi-
larly there may be tetra-bifunctional, tri-trifunctional reactions, etc.,
and all these may be classed together as polyfunctional reactions. The
present discussion is concerned only with bifunctional reactions.

Bifunctional Reactions. These always present two possibilities: they
may be intramolecular or intermolecular. If intramolecular they can
lead only to the simple monomeric ring I. If intermolecular they may
lead either to a polymeric ring II or to a polymeric chain III.

x R y - > R z I

- > ( R z ) II

- ^ X __(__R_ Z __)_R__y HI

n- 1

These obviously represent three possibly competing reactions.

The question now arises, what factors will determine which of these
possible courses a bifunctional reaction will take? It is obvious, for
example, that in general dilution would favor intra- over intermolecular
reactions (6). Temperature and catalysts might favor either one or the
other. It appears, however, that structural and stereochemical factors will
usually be more important than any others. That is, though it may be
possible in some bifunctional reactions to control the choice between intra-
and intermolecular reaction by suitable adjustment of experimental condi-
tions, this choice will, in general, be almost completely determined by the
nature of the reacting molecules.

The effects of these factors may be stated as follows. (1) If the product
of intramolecular reaction would be a ring which, on stereochemical
grounds, is incapable of existence, reaction will be intermolecular. This
is apparently the case with p-N^CcEUCI^CHkCl, which reacts with itself
intermolecularly (7), and not as had previously been supposed (8) with the
formation of the so-called dihydro-^-indole (9). The utility of this very
obvious principle is somewhat diminished by the fact that, in the present
state of stereochemical knowledge, it is sometimes impossible to predict
whether a given ring system will be capable of existence or not.

(2) Bifunctional reactions which can lead to the formation of 5- or 6-
rings almost invariably proceed intramolecularly. This well-established

I. THEORY OF CONDENSATION POLYMERS

*
'8

x-*

V-o

!*:

o-o

-CHi-0-

-CH-CH

-HK
NOO-*HO

-}| WA MA
2 oT or
^i w w

Q S5 ' - - fi

CO T-l L- *5

i o_Sx w \5o >; N- r' ^ K g-

I I S uT g J jjq -

o v w/5 l^i* I

T l^il!

4* o-^^^s^-.

-algal*

r w "^ CQ

* -Sa-:- -2
Aa if 65 ^*

T iitrfsa

*5 J8 5 S - 8

H 2 ^ c

4 a I f

W 01 ^0 ^F ^

10 POLYMERIZATION AND RING FORMATION

fact is responsible for the existence of the majority of the very large
number of 5- and 6-rings which is known.

(3) Bifunctional reactions which, if intramolecular, could lead only to
larger-than-6-rings, generally proceed intermolecularly and lead to poly-
meric products. If this rule were free from any exceptions, it would neces-
sarily follow that the polymeric products would always be of the open-
chain type, III.

Although a great many bifunctional reactions have been studied in the
hope of forming large rings, our information as to the precise nature of the
products of these reactions is very meager. That they do not usually
proceed intrarnolecularly follows from the fact that very few large rings
are known. It is true that such reactions frequently do not proceed at
all under conditions which, in analogous cases, lead to the formation of
5- and 6-rings; but the formation of the latter often proceeds under con-
ditions which do not permit intermolecular reaction even among uni-
functional compounds (compare the formation of 7-lactones in the pres-
ence of a large excess of water, with the esterification of acetic acid).
There are, however, a good many types of reaction which are intermolecular
among simple compounds, which are practically free from side reactions,
and which by suitable adjustment of experimental conditions may be
forced to completion. It should be possible to conduct reactions of these
types even when bifunctional, so as to obtain analytically homogeneous
products whether they proceed intrarnolecularly or not. These simple
principles have been repeatedly applied to bifunctional reactions in this
Laboratory, and they have seldom failed to lead to analytically homogene-
ous products. Moreover, where the formation of 5- or 6-rings was ex-
cluded by the nature of the reacting materials, these reactions have, with-
out exception, led to high polymers. This fact provides a possible explana-
tion for the meagerness of the information which is available concerning
the precise nature of the products of such bifunctional reactions. High
polymers frequently have properties which make their investigation very
difficult. Moreover, if a chemist is expecting a reaction to lead to ma-
terials of simple properties, he is usually inclined to regard the appearance
of a resinous or sirupy product which neither can be crystallized nor dis-
tilled as signifying only that the experiment has failed. When the prod-
ucts of such reactions have been capable of purification and have shown
the expected analytical composition, they have frequently been assumed
to be dimeric (and cyclic) for no other reason than that they were ob-
viously not monomeric. In this connection it should be emphasized that
substances of very high molecular weights may neverthless be micro-
crystalline and very soluble (see the following paper).

I. THEORY OF CONDENSATION POLYMERS 11

Some of the points discussed above are illustrated by the following
examples.

Anhydrides of the acids of the series HOOC(CH 2 ) a: COOH are known
in which x has all values from 1 to 8 inclusive. Of these only succinic and
glutaric anhydrides are monomeric. These are, respectively, 5- and 6-
rings. The other anhydrides which, if monomeric, would be 4-, 7-, 8-,
9-, 10- and 11 -rings, are all highly polymeric (10).

Hydroxy-acids of the series HO(CH 2 ) Z COOH may condense with them-
selves. Only those in which x is 3 or 4 yield monomeric lactones (11).
These lactones are 5- and 6-rings, respectively. This series of acids now
includes those in which x has all values from 8 to 16 inclusive (12). The
lactones corresponding to the acids in which x has the values 12 to 16 are
also known (12b), as well as the 17-membered lactone of the unsaturated
acid, hexadecene-(7)-ol-(16)-acid-(l) (13). All of these lactones are per-
fectly stable substances and can be distilled without decomposition (14);
and yet none of these lactones has been synthesized by a bifunctional re-
action (15). They were synthesized by the oxidation of the corresponding
cyclic ketones with persulfuric acid. By heating the acids, products are
indeed obtained which have the same analytical composition as the lac-
tones, but (12a) they are "polymeric-like." The properties of the known
lactones of this series indicate that lactones cannot be intermediates in the
formation of these polymers. The formation of these must, therefore,
involve C polymerization.

Amino acids of the series NH 2 (CH 2 ), r COOH are known in which x has
the values 3, 4, 5 and 6. When x is 3 and 4, intramolecular anhydride
formation occurs spontaneously at the melting point of the acids, the
products being the 5- and 6-membered rings, pyrrolidone and piperidone
(16). When x is 5 the acid is, on heating, converted to the extent of 20 to
30% of the theoretical into the 7-membered lactam. "The rest of the acid
is converted into a viscous gelatinous mass which could be obtained in a
state of only approximate purity and which, according to the analysis,
is isomeric with the 7-membered lactam; that is, it is a polymeric prod-
uct" (17). Where x is 6 no trace of the corresponding lactam could be ob-
tained by heating the acid, although this lactam (suberone-isoxime) is
a known and stable substance (18). By heating the acid a product was ob-
tained having the same composition as this lactam, but it was an un-
distillable solid, insoluble in most solvents (17).

The compounds Br(CH 2 ) I NH 2 , where x is 4, 5 and 6, react with them-
selves with the formation of secondary amines containing the structural
unit ( CH 2 ) 7 NH . Where x is 4 and 5 the products are the mono-

12 POLYMERIZATION AND RING FORMATION

meric 5- and 6-rings, pyrrolidine and piperidine (19). Where x is 6 the
amount of the monomeric base (7-ring) formed was too small to permit
purification (20). The product was for the most part an oil which solidified
to a waxy solid. It could not be distilled in vacuum without decomposi-
tion, and though no molecular weight determination is recorded, v. Braun
seems inclined to regard it as dimeric. Its properties are obviously more
consistent with a more highly polymeric structure.

An interesting example of a bi-bifunctional reaction is found in the
reaction between glycols and acetaldehyde (or acetylene). This presents

/o\

the possibility of forming cyclic acetals, CH 3 CH R, or polyacetals,

\0/
HO R O CH O R O CH O R O CH O , etc. This re-

action has been studied by Hill and Hibbert (21). Ethylene and tri-
methylene glycols gave in excellent yields the cyclic acetals which are 5-
and (5-rings. Tetramethylene glycol gave in poor yield a volatile com-
pound which was apparently the monomeric cyclic acetal containing a
7-ring. A considerable part of the product was an undistillable sirup. The
products from octamethylene and decamethylene glycols were also un-
distillable sirups. No molecular weight determinations are recorded. Hill
and Hibbert recognized of course that these sirups might have the polymeric
linear structure indicated above, but seemed inclined to regard them as
simple rings of a peculiar kind.

The only reaction which has led to the formation of large carbon rings
from open-chain compounds is the action of heat on certain salts of di-
basic acids (22). In no case are these large rings the chief products of this
reaction, nor can it be established that they are primary products, since
the reaction takes place at high temperature. If polymeric products
were formed they would be decomposed thermally. High polymers
could not appear in the distillate from which the products are isolated,
because high polymers cannot be distilled. These features sharply dis-
tinguish this reaction from other truly bif unctional reactions which might,
but which do not, so far as our information goes, lead to rings of the same
type (23).

Any theory which attempts to explain why 5- and 6-rings are formed
very readily and larger rings with very great difficulty must take into
account the fact that these larger rings are not less stable than the smaller.
A satisfactory theory is already available. The Baeyer strain theory in

I. THEORY OF CONDENSATION POLYMERS 13

its original form is made untenable by the existence of the higher cyclo-
paraffins, and has been replaced by the Sachse-Mohr theory which per-
mits the existence of non-planar and strainless rings and which has be-
sides a great deal of other evidence in its support (24). The features of this
theory essential to the present argument are exceedingly simple. They
involve the assumption of the tretrahedral angle in the carbon valences
and of free rotation about each carbon-carbon single bond in a chain.
No description of stereochemical relations can be as convincing as a
demonstration with suitable models (25). Such models show that there is
a certain inevitability in the formation of 5- and 6-rings and a large degree
of fortuity in the formation of larger rings. This question has been con-
sidered in some detail by Mohr (26). Regarding the possibility of the
formation of a cyclic ketonic ester by the internal condensation of sebacic
ester he says :

The molecule of sebacic ester is a very long chain. The multiplicity of forms which
this molecule can assume is extraordinarily great. It is clear that by the random col-
lisions of the molecules, fewer molecules of sebacic ester will in unit time assume the
form necessary for ring closure than is the case with adipic ester under the same condi-
tions, since this contains in its molecule only four methylene groups, and four less points
of rotation than does sebacic ester. To bring the two ends of the sebacic ester molecule
together in the least forced fashion apparently requires a very small amount of energy
or none at all. Unfortunately, however, we have not the means arbitrarily and with the
least possible expenditure of energy to bring about the desired change in form of the
molecules and to hinder the undesired. In this connection we are left entirely to chance,
that is, to random collision, which will bring about a given form the more rarely the
more forms are possible, i. e., the longer the chain between the two carboxyl groups.

In perhaps two dozen cases it is well established that a bifunctional
reaction proceeds fairly smoothly with the formation of a larger-than-6-
ring. These cases include more 7-membered than larger rings and more
aromatic than aliphatic compounds. The Sachse-Mohr theory would
lead one to expect that the formation of 7-rings would not be very much
more difficult than the formation of 6-rings. The bifunctional reactions
discussed in the examples offered above all involve purely aliphatic com-
pounds, and for the most part the chains are unsubstituted. In most of
these examples the formation of a 7-ring occurs to a certain extent, and
the formation of the corresponding 8-ring not at all. It appears that in
such 7-atom chains the probabilities of intra- and intermolecular reaction
are about equal. The addition of one more atom to the chain diminishes
the probability of intra- with respect to intermolecular reaction to such an
extent that only the latter appears. These relations may be somewhat

14 POLYMERIZATION AND RING FORMATION

modified by the presence of substituents on the chains and even by the
nature of the reacting groups in ways which it is, at present, impossible to
predict. It is evident, however, that if two atoms of such a chain are
adjacent atoms of a benzene ring their position with respect to each other
is fixed, and the chances of intramolecular reaction are greater than in an
analogous simple chain of the same length. (The latter will have one
more axis of rotation than the former.) It is therefore not surprising
that in NH 2 (CH 2 )6C1 intramolecular reaction occurs only to a slight ex-
tent, and in o-NI^CeH^CH^Cl almost quantitatively (27). Other similar
examples might be cited.

There is also some evidence to indicate that even simple substituents
such as methyl groups on a chain may increase the tendency toward the
formation of larger rings (28). Researches in the diphenyl series have es-
tablished that substituent groups suitably placed may completely inhibit
rotation about a nearby single bond, and a similar effect in aliphatic chains
is at least conceivable. Any restriction of the freedom of rotation of the
atoms of a chain would, on the basis of the Sachse-Mohr theory, increase
the chances of ring formation.

The process of polyintermolecular condensation finds no mention in
treatises on polymerization. This may be due to the fact that such a
process is not admitted to exist, or that it is not admitted to be polymeriza-
tion. The examples cited above, and others to be described later, prove
that such a process does exist, and that it may result in the formation
of very large molecules. Whether it is to be regarded as polymerization
or not will depend upon the definition which is adopted for that term.
The definitions offered above include it as a special type of polymerization.
This classification finds more to justify it than the analogies which are
recognized in these definitions.

The process of A polymerization (the only type which appears to have
been generally recognized) results in the formation of large molecules
from small; and it has come about that any process which has this result
is called polymerization. Since, however, A polymerization is by defini-
tion a process of self -addition, chemists have often been misled to the
assumption that condensation leads directly only to small molecules, and
that if large molecules are formed from small as the apparent result of
condensation, this is due to the intervention of some unsaturated molecules,
capable of undergoing A polymerization (29). It is quite certain, however,
that in many such cases (i. e.,m all cases of true C polymerization) no such
intermediate products occur. This is a matter of important practical im-
plications. The reactions involved in the formation of A polymers must,

1. THEORY OF CONDENSATION POLYMERS 15

in the nature of the case, be for the most part reactions which are peculiar
to the process of polymerization. For this reason the mechanism of A
polymerization still remains somewhat obscure. Hence, the mere assump-
tion that unsaturated intermediates intervene in a reaction which leads to
high polymers contributes little to one's understanding of the mechanism
of the process, or the structure of the product. In those cases in which
this assumption is wrong its use leads one to regard as complicated and
mysterious a process which may be simple and obvious. C polymerization
merely involves the use in a multiple fashion of the typical reactions of
common functional groups. Among bifunctional compounds these reac-
tions may proceed in such a way as to guarantee the structure of the struc-
tural unit, R , in the polymer, ( R ) n , formed. It is one of the im-
mediate objects of the researches to be described in subsequent papers to
discover how the physical and chemical properties of high polymers of this
type are related to the nature of the structural unit.

Summary

Linear polymers confrom to the type R R R R R , etc., which
is characterized by a recurring structural unit. The structural unit R
is a bivalent radical. Two types of polymers are recognized. (1) Addition
of A polymers : the polymeric molecule is converted by heat into a monomer
having the same composition as the structural unit, or the polymer is
formed by the mutual addition of a number of such monomers. (2) Con-
densation or C polymers : the polymeric molecule is converted by hydrolysis
or its equivalent to a monomer which differs in composition from the struc-
tural unit by one H 2 O (or HC1, NH 3 , etc.), or the polymeric molecule is
formed from numbers of the monomers by a process of polyintermolecular
condensation. Rubber, polystyrene, polyoxymethylene and paracetalde-
hyde are A polymers. Cellulose, silk fibroin and hexa-ethylene glycol are
probably C polymers.

Substances of the type x R x and x R y are called bifunctional. In
these formulas R represents a bivalent radical and x and y functional
groups capable of reacting with each other in a known fashion to form the
new functional group, z. Reactions of the type x R y > product are
called bifunctional and those of the type x R x + y R y > product
are called bi-bifunctional. Such reactions will lead to compounds contain-
ing the structural units R z and R z R z . Bifunctional reac-

tions will be intramolecular and will lead to the monomeric product, R z

16 POLYMERIZATION AND RING FORMATION

when this can be a 5- or 6-ring. If the monomeric product can only be a
larger-than-6-ring, reaction will usually be intermolecular and the product
a polymer of the type R z R z R z R z R z , etc.

Bibliography and Remarks

(1) Berzeiius, Jahresbericht, 12, 63 (1833).

(2) Hess,"Chetnie der Zellulose," Leipzig, 1928, p. 577; Meerwein,Houben-Weyl,"DieMethoden
der organischen Chemie," Leipzig, 1923, zweite Auflage, dritter Band, p. 1013; Staudinger, Ber., 58,
1074 (1920).

(3) However, Staudinger, in his latest papers, favors the view that polystyrene and rubber are very
large rings, Ber., 62, 241 (1929).

(4) The term condensation is used here to name any reaction which occurs with the formation of a
new bond between atoms not already joined and which proceeds with the elimination of elements (H 2 ,
Na, etc.) or of simple compounds (HsO, CaHtOH, NH 3 , NaBr, etc.) Examples are the Wurtz reaction,
Friedel-Crafts reaction, esterification, etc.

(5) It would not be difficult to suggest examples in which a single polymer might belong to either or
to both of these classes depending upon the method by which it was synthesized. So far as cellulose
is concerned there is little to justify its classification as a C rather than as an A polymer other than the
criteria which are set forth in the above definitions, and these may at present appear somewhat arbi-
trary. Thus even if cellulose were synthesized by the dehydration of glucose, this might occur by the
formation of a glucosan which subsequently polymerized by self-addition. It may be observed, how-
ever, that of the two trimethylglucosans which might conceivably polymerize in this way to form tri-
methylcellulose, one is known and does not polymerize, and the other is probably incapable of existence
on stereochemical grounds. See Freudenberg and Braun, Ann., 460, 288 (1928).

(6) Cf. Ruggli, Ann., 392,92 (1912), where the recognition of this principle made possible the syn-
thesis of large rings containing an acetylenic linkage.

(7) Ferber, Ber., 62, 183 (1929).

(8) v. Braun and Gawrilow, ibid., 45, 1274 (1912).

(9) For some other examples, see Titley, J. Chem. Soc., 2571 (1928).

(10) For malonic anhydride see Staudinger and Ott, Ber., 41, 2214 (1908). For the other
anhydrides see Voerman, Rec. trav. chim., 23, 265 (1904). The cryoscopic data which Voerman
presents on phenol solutions have no bearing on the molecular weights of the anhydrides, since even
succinic anhydride reacts rapidly with phenol to form phenyl succinate; Bischoff and von
Hedenstrtim, Ber., 35, 4076 (1902). Farmer and Kracovski have recently obtained adipic anhydride
in a monomeric form, J. Chem. Soc., 680 (1927).

(11) On this point no data are available for those in which x is 5 and 6 [/. Am. Chem. Soc., 46, 2838
(1924); Ber., 60, 605 (1927); ibid., 33, 864 (1900)]. The analogous acids, CH 3 CHOH(CH 2 ) 2 CH-
(CH 3 )CH 2 COOH [Baeyer and Seuffert, ibid., 32, 3619 (1899)] and C 2 H*CHOH(CH 2 )4COOH [Blaise
and Kohler, Compt. rend., 148, 1772 (1909) ], have been prepared, and by heating are converted at least
partially into the 7-membered lactones. The acid C 2 HjCHOH(CH 2 )5COOH has also been prepared.
On heating, it also loses water, but the 8-membered lactone is not formed. The product is a lactide
like material which cannot be distilled without decomposition (Blaise and Kohler).

(12) (a) Lycan and Adams, J. Am. Chem. Soc., 51,625 (1929) ; (b) Ruzicka and Stoll, #<?/*>. Chim.
Ada., 11, 1159 (1928). Since this paper was written a great deal of further information concerning
these acids has become available in the paper of Chuit and Hausser, ibid., 12, 463 (1929).

(13) Kerschbaum, Ber., 60, 902 (1927).

(14) This indicates that large heterocyclic rings are not less stable than their lower homologs.
Their physical properties are such as would be expected from the known properties of their lower
homologs The mere presence of a large ring does not result in the development of any of those
unusual secondary, residual or supermolecular forces which are sometimes supposed to confer
"polymeric" properties on relatively simple molecules.

(15) Kerschbaum (ref. 13) records various attempts to prepare ambretollid by the methods com-
monly used for the preparation of lactones. By heating the acid he obtained traces of an oil which had
the odor of the lactone. (It is powerfully odorous.) The product was, however, for the most part, a
gelatinous material soluble in alkali. For the usefulness of the method proposed in Gorman Patent
Application 150,677, see Ruzicka and Stoll, ref. 12b.

(16) Gabriel, Ber., 22, 3338 (1889); Schotten, ibid., 21, 2240 (1888).

II. POLYESTERS 17

(17) v. Braun, Ber., 40, 1840 (1907).

(18) B. p. 156 at 8 mm. It is, however, not prepared by a bi functional reaction, but from cy-
cloheptanone oxime by the Beckmann rearrangement. Wallach, Ann., 312, 305; 309, 18 (1899).

(19) v. Braun and Beschke, Ber., 39, 4121 (1906); Blank, ibid., 25, 3044 (1892); v. Braun and
Steindorf.iTm*., 38, 172 (1905).

(20) v. Braun and Steindorf, ibid., 38, 1083 (1905); v. Braun, ibid. 43, 2853 (1910).

(21) Hill and Hibbert, /. Am. Chem. Soc., 45, 3108, 3124 (1923). See also Franke and Gigerl,
Monatsh.,49 t 8 (1928).

(22) Ruzickaand co-workers, Helv. Chim. Acta, 9, 230, 249, 339, 389, 399, 499, 715, 1008 (1926);
10, 695 (1927); 11, 496, 670, 686, 1159 1174 (1928).

(23) For example, the action of metals on a dihalide, Br(CH2)zBr. Thus decamethylene bromide
reacts rapidly and very smoothly with metallic sodium, but no cyclodecane is formed. See Franke and
Kienberger, Monatsh., 33, 1189 (1912). This reaction will be discussed in a later paper

(24) See Hiickel, "Der gegenwartige Stand der Spannungstheorie," Fortschritte Chem. Physik und
physik. Chem., Band 19, Heft 4, 1927.

(25) Ordinary (preferably quite small) wire tetrahedra may be joined by short pieces of rubber tub-
ing in such a way that the arms which are being connected overlap. This provides free rotation about
the union, but prevents bending. This method of union insures that any model which is built up is
practically strainless (. e., the tetrahedral angle is always retained) and at the same time is sufficiently
mobile to illustrate the multiplicity of forms which long chains and large rings can assume.

(26) Mohr, /. prakt. Chem., 98, 348 (1918).

(27) v. Braun and Bartsch, Ber., 32, 1270 (1899).

(28) See "Annual Reports of the Progress of Chemistry," 1927, p. 100; Moyer and Adams, J. Am.
Chem Soc , 51, 630 (1929).

(29) Thus, Drummond, Inst. Rubber Ind., 4, 43 (1928), .states, "Where resins are formed by a pre-
liminary condensation as are the phenol-formaldehyde products, it is logical to assume that in the
preliminary reaction, unsaturated atomic linkings are introduced which provide the necessary arrange-
ment for polymerization subsequently to occur." Scheiber, Chem. Umschau, Fette, Oele, Wachse
Harze, 15, 181 (1928), is even more explicit: "Fur den Aufbau organischer Stoffe kommen bekanntlich
nur zwei Prozesse in Betracht, und zwar Kondensationen und Polymerisationen. Vorgange der ersteren
Art fvihren im allgemeinen zu Molekulverbanden beschr&nkter Grosse . .Bei Polymerisationen hinge-
gen kommt es in manchen Fallen zur Ausbildung extrem grosser Molekiile. . ."

II. Polyesters*

In this paper a bi-bifunctional reaction is studied by condensing di-
basic carboxylic acids with dihydric alcohols. The number of carbon
atoms in the structural units are 7, 8, 9, 10, 11, 12, 14, 15, 16, 18 and 22.
Consequently] only C-polymers and no other products are obtained.
Their molecular (or better average molecular) weights were measured by
the freezing and boiling point methods. They range between 2500 and
5000. With the exception of the liquid ethylene malonate all polyesters
are microcrystalline materials, with a melting point, which is not quite
sharp.

* W. H. Carothers and G. A. Arvin; Journ. Am. Chem. Soc., 51, 2660-70 (1929);
Contribution No. 1 1 from the Experimental Station of E. I. du Pont de Nemours and
Co.

Received April 13, 1929. Published August 7, 1929.

t See the thesis developed on pages 10 and 11.

18 POLYMERIZATION AND RING FORMATION

The following combinations were prepared*

ethylene malonate hexamethylene succinate

ethylene succinate hexamethylene adipate

ethylene adipate hexamethylene sebacate

ethylene sebacate hexamethylene phthalate

ethylene maleale decamethylene succinate

ethylene fumarate decamethylene adipate

ethylene phthalate decamethylene sebacate

trimethylene succinate decamethylene phthalate
trimethykne adipate
trimethylene sebacate
trimethykne phthalate

An example of a bi-bifunctional reaction is found in the reaction between
a dibasic acid and a dihydric alcohol, HOOC R' COOH + HO R"
OH, which, if it is conducted so as to involve both functional groups of
each reactant, must lead to an ester having the structural unit, OC R'
CO O R" O = R . In accordance with the thesis developed in
the previous paper, esters formed in this way will be polymeric unless the
number of atoms in the chain of the structural unit is less than seven.
In this paper esters are described in which the number of atoms in the
chain of the structural unit is 7, 8, 9, 10, 11, 12, 14, 15, 16, 18 and 22
atoms. All these esters are highly polymeric, and, although some of
them have been prepared by various methods, no monomeric form of any
of them has as yet been isolated.

Preparation of the Esters

The following method was used for the preparation of the solid esters whose prop-
erties are listed in the table. The acid together with a 5% excess of the glycol was placed
in a Claisen flask provided with a receiver and condenser, and the flask was heated in a
metal-bath. At about 160 (bath temperature) reaction set in. Water distilled off
freely during the first hour (temp., 175-185) and very slowly if at all during the succeed-
ing two hours at the same temperature. The receiver was now changed, the flask pro-
vided with a very fine capillary and heating continued under a good vacuum (usually
less than 0.2 mm.) for about three hours, the temperature of the bath being raised to
200-250. During this period little or no distillate collected (provided only a 5% excess
of glycol was used). The residue, which was a slightly dark and, at 150, more or less
viscous liquid, was poured from the flask. The amount of this residue corresponded
with the theoretical (based on the acid used), and the amount of water actually collected
approached the calculated more closely the larger the sample used (60-90%). The
esters were purified by crystallization.

* Mostly by direct condensation of the acid with the glycol in different propor-
tions, in one case (ethylene succinate) by combining silver succinate with ethylene
bromide (comp. p. 25).

II. POLYESTERS 19

Ethylene malonate was prepared by heating ethyl malonate and glycol in the same
fashion. There was some decomposition (evolution of gas) when the residue was heated
to 240 in vacuo. This residue was a thick sirup which could not be induced to crys-
tallize. It was dissolved in acetone, filtered and heated for several hours at 175-190
in high vacuum. Nothing corresponding to a monomeric ethylene malonate was found
on redistilling the distillates from this preparation.

The preparation and properties of some esters not included in the table are described
in the Experimental Part of this paper.

Structure of the Esters

The conclusion that these compounds are esters and that they contain
the structural unit R indicated above follows directly from the
method of preparation and is supported by the analytical data and chemi-
cal behavior. Comparison of Cols. 4 and 5 in Table I shows that the
carbon and hydrogen percentages in general agree with those calculated
for R . So also did saponification numbers where these were de-
termined. The products of saponification of the esters are the acids and
glycols from which they were prepared.

That the esters are not monomeric 7-, 8-, 9-, . . ., 22-membered rings is
indicated by their physical properties and by their molecular weights.
In Col. 7 are given the molecular weights calculated from the observed
boiling point elevations or freezing point lowerings, and in Col. 8 the
solvents and methods used. The solvents used in the ebullioscopic de-
terminations include chloroform, ethylene chloride, benzene and acetone.
Those used in the cryoscopic determinations include benzene, glacial
acetic acid and diphenyl ether.

In these determinations it was necessary to use rather large samples
to obtain boiling point or freezing point changes of 0.02 to 0.05. Judged
by their self-consistency, the values obtained with the use of freezing
benzene have about the same relative accuracy as is ordinarily attained
with simple compounds. With other solvents, and especially in the
boiling point determinations, observations were much less consistent.
This was due in part at least to changes of surface tension, associated no
doubt with molecular size, since there was much foaming. The values
obtained by this method are recorded, since for some of the esters they are
the only ones available, and for others they clearly indicate that the order
of magnitude of the apparent molecular weights is quite independent of the
character of the solvent used.

The lowest molecular weight observed for any of the solid esters was
2300 and the highest 5000. The mean of all the determinations is 3200,
and for purposes of calculation the molecular weight of each ester has

POLYMERIZATION AND RING FORMATION

ex
a

53 v*

*?*'! s

ife"

- g g.p

5 I 8 a>

< ii5 $

^JJJSjg JiJ.JzJ ft

<<OOO < < -^'<UO'S
d, a, ex d d a d d d d

ffi O
3 1

1 1

CO Ol to O O

o e* o> oo CN

t-.' oo oo oo o>

CO O C C 00

h- O O CO C3l

0>
00

o
d

^ CO CO cSl 1-1 O

co c3
o oo

i 1 ?

3f
B3'o

H rH O CN 00 CO

b. CD ^ CO O O

M t H 1 i it

II. POLYESTERS 21

been assumed to be approximately 3000. Such a molecular weight cor-
responds with a value of 8 to 20 (depending on the length of the chain
of the structural unit) for n in the general formula ( CO R' CO
O R" O ) tt . It can scarcely be supposed that the molecules present
in a given sample are all identical so far as the values of n are concerned,
but their crystallinity and a certain homogeneity in such physical behavior
as solubility indicate that the varieties of molecular species present in
a given sample probably do not include a very wide range.

The disposition of the valences at the ends of the chains in the above
formula cannot as yet be definitely decided. A cyclic structure is ren-
dered improbable by the same considerations which led us to expect that
these esters would be polymeric and which have been discussed in the
previous paper. Esters prepared by the method indicated above are
definitely not acidic, and since the detection of a car boxy 1 group in a
molecule having a weight of 3000 to 5000 should present no difficulties,
no carboxyl groups can be present at the ends of the chains. Attempts
to detect the presence of hydroxyl groups have not as yet succeeded.
Kthylene succinate crystallizes unchanged from acetic anhydride, and is
not affected by phenyl or naphthyl isocyanate. Nevertheless we are
inclined to assume that hydroxyl groups are present at each end of the
chain due to the presence of one more molecule of glycol than acid : HO
R" O ( CO R' CO O R" O ) M H. The failure of these
groups to react with reagents may be ascribed to the operation of the same
factors which set a limit to the size of the molecules formed.

In general the analytical values agree more closely for such a formula
than for a cyclic formula. In Col. 6 of Table I are given the carbon and
hydrogen percentages calculated for this formula assuming a molecular
weight as close as possible to 3000. The means of the deviations between
the carbon and hydrogen percentages found and those calculated for the
( R ) n formula are 0.48 and 0.01% ; while the deviations calcu-
lated for the HO ( R ) w R' OH formula are -0.31 and -0.11%.

Assuming a molecular weight of 3000 the chains, H [ O (CH 2 ) y
O CO (CH 2 ) CO 1 n O (CH 2 ) y OH, would contain 170-200
atoms and their lengths would lie between 240 and 280 A. (1). Their
lengths are therefore of the same order of magnitude as that assumed by
Meyer (1) for the cellulose chains (100 glucose units per chain = 510 A.).
The molecular cohesions calculated from the data presented by Meyer (2)
lie between 250,000 and 300,000 cai. or 3.3 to 4 times the heat of separa-
tion of a carbon-carbon bond (75,000 cal.) . These molecular cohesions are,
however, only about 10% of those calculated for a cellulose chain of 500 A.

22 POLYMERIZATION AND RING FORMATION

The structure of these esters will be discussed in more detail later in
the light of experimental work which is not yet completed.

Properties of the Esters

All the esters of the type (CO (CHa)* CO O (CH 2 ) y O )
with the exception of ethylene malonate are crystalline solids. The
property of crystallinity is not very highly developed, but it is quite defi-
nitely present (3). This fact is of interest because examples of crystalline
highly polymeric substances are not very numerous. As with all crystal-
line substances of this class, it is not possible to develop large crystals.
Ethylene succinate when it separates from a melt or from concentrated
solutions (in chloroform) crystallizes in doubly refracting spherolites
(microscopic) which grow to what appear to be star-like groups of needles.
From dilute alcohol solutions it separates in discrete microscopic needles

(4).

Crystallized ethylene succinate and decamethylene sebacate are dusty
powders which have a great tendency to become electrified. Some of the
intervening members are quite soft and even somewhat sticky. Ethylene
succinate solidifies from a melt as a hard, brittle, opaque, white mass.
Decamethylene sebacate solidifies to a white, brittle, waxy solid.

The non-crystalline members include ethylene malonate, ethylene
fumarate and ethylene, trimethylene, hexamethylene and decamethylene
phthalates. The phthalates and the fumarate evidently have less sym-
metry than the saturated straight-chain purely aliphatic type. The lack
of crystallinity of the malonate may be associated with the low melting
point of the malonic acid; or it may be due to the fact that considerable
decomposition occurred during the process of esterification.

The melting points of the solid esters are not very sharp and they depend
somewhat upon the rate of heating. This latter effect is noticed especially
with ethylene succinate.

When the crude reaction mixture is dissolved in chloroform and precipitated with
benzene, the resulting dusty white powder melts in a capillary tube at about 102.
If, however, the tube is placed in a bath which has already been heated to 96 the
sample melts at once. The two melting points are called the slow and instantaneous
melting points, respectively, and they are best observed on a bloc Maquenne. Samples
which are dusted on the bloc when it is below 96 do not melt until it reaches 102,
but as soon as the bloc reaches 96 samples melt the instant they touch it. Both these
temperatures may be changed somewhat by repeated crystallization, but a sample whose
melting point has been raised by repeated crystallization may suddenly show the origi-
nal melting point on further crystallization. By long extraction with boiling absolute
alcohol in which it is practically insoluble, ethylene succinate is modified so that the in-

II. POLYEvSTERS 23

stantaneous and slow melting points coincide at about 107-108, and from the extracts
may be isolated a very small amount of lower-melting material. When such a high-
melting sample is again crystallized from a mixture of chloroform and benzene or ether,
the melting point usually drops, and one observes again the instantaneous melting
point of 96 and the slow melting point 102. The melting point of such samples may
also be raised by long heating just below 100, and both melting points then gradually
approach 107. When a 96, 102 sample is quickly melted, then allowed to solidify in
an agate mortar and powdered, the instantaneous melting point is lowered somewhat,
and the slow melting point may be either raised or lowered, depending upon the time
during which the molten material is heated.

Attempts to define the melting point more clearly by heating or cooling curves on
fairly large samples were unsuccessful (5). Apparently the viscosity of the melt is so
great as to make the process of crystallization slow even in the presence of previously
formed crystals and under a considerable temperature gradient .

The erratic behavior of the melting points may be associated with the fact that all
these esters are hygroscopic, but the absorption of water cannot be the sole cause for
this behavior, for samples of ethylene succinate having various melting points were
stored over 20% sulfuric acid for ten days and all of them gained considerably in weight
without changing in physical appearance or melting point.

Anomalies in melting points are not unknown among compounds of definite con-
stitutions and relatively low molecular weights. One may cite, for example, the fact
that sucrose separates from methyl alcohol in crystals melting at 169-170 and from
other solvents in crystals melting at 179-180, and these two forms do not show any
detectable chemical differences (fi). But the case of ethylene succinate appears to be
more complicated and confused than this. Its melting point depends upon its history in
a fashion which can as yet be defined only to the following extent it always rises and
approaches 107 as it is, while still in the solid state, heated or extracted with boiling
absolute alcohol.

Among the other esters these irregularities were less pronounced.
Nevertheless, the melting points are not very characteristic and no attempt
was made to crystallize them to constant melting points. Instead they
were crystallized two to six times depending upon the color and softness
of the sample. The melting points recorded are the highest observed for
complete melting and are usually within 5 of the lowest observed for any
sample of the same material. The melting range was usually less than
2. The melting point of the decamethylene ester of a given acid was
higher than that of the hexamethylene ester. The melting point of the
hexamethylene ester was higher than that of the trimethylene ester,
which was in turn lower than that of the ethylene ester. With one ex-
ception, the adipic ester of each glycol melted lower than either its suc-
cinic or sebacic ester.

All of these esters are, when molten, quite viscous. Those which do not
crystallize on cooling become more or less hard, tough and glassy. The
use as resins of esters formed by the action of polybasic acids on poly-

24 POLYMERIZATION AND RING FORMATION

hydric alcohols has been covered by numerous patents (7). The impres-
sion seems to prevail that all such esters are resins. Of the esters described
in this paper the majority are crystalline. The phthalates which we have
prepared are, however, all resinous. Ethylene phthalate is quite tough
and moderately hard. Neither this ester nor the trimethylene, hexa-
methylene nor decamethylene esters of the same acid have been described
in the scientific literature. As the length of the alcohol chain increases,
these phthalates become progressively softer. Decamethylene phthalate
has the consistency of a moderately thick sirup.

None of these esters is volatile. Each of them has been heated to
200 in vacua without showing any tendency to distil. Ethylene sue-
cinate heated for three hours at 250 under a pressure of 2 mm. is not
changed in properties, and its analytical composition (carbon, hydrogen
and saponification number) is not changed. It shows no tendency to
distil or to evolve any volatile products when heated to 280 at a pressure
of 1 micron. At 350 ethylene succinate undergoes complete thermal
decomposition, the products being ethylene, acetaldehyde, succinic anhy-
dride, carbon, etc. (8).

All of these esters are insoluble, or nearly so in water, alcohol, petroleum
ether and ether (9). The least soluble of the saturated aliphatic esters is
ethylene succinate. It is more or less soluble in hot ethyl acetate, acetic
acid, acetic anhydride, ethyl succinate and acetone, and it crystallizes
from these solvents on cooling. It is insoluble in benzene. The other
saturated aliphatic esters with the exception of ethylene malonate are
readily soluble in cold benzene and at least moderately soluble in acetone,
ethyl acetate and glacial acetic acid. The solutions are noticeably vis-
cous only when fairly concentrated (e. g. t 10% or stronger). This fact
and the rapidity with which solution occurs lead us to believe that the
solutions are true molecular dispersions.

Although none of these esters is soluble in water they are all somewhat
hygroscopic. This property is especially pronounced in those polyesters
which are not crystalline. The drying of these resins was very difficult
and, as the analytical results indicate, was not always successful. Drying
to constant weight was complicated by the tendency to foam when heated
in vacua (10).

Other Methods of Preparation

The only detailed study of polyesters of the type here described which
we have found is reported in a paper by Vorlander (11).

II. POLYEvSTERS 25

Ethylene succinate was first prepared by Lourenco (12) by heating sue-
cinic acid and ethylene glycol to 300 (m. p. about 90). Subsequently
Davidoff (13) prepared Lourengo's ester and identified it with one which he
obtained by heating silver succinate with ethylene bromide. Vorlander
studied this latter compound in detail and showed that its chemical be-
havior was in all respects what would be expected of an ethylene ester
of succinic acid. He also identified it with Lourengo's ester and with one
obtained by the action of succinyl chloride on the disodium derivative of
ethylene glycol.

Vorlander found that his ethylene succinates prepared from silver suc-
cinate and from succinyl chloride had apparent molecular weights ranging
from 265 to 321 in freezing phenol and in freezing acetic acid. He sup-
posed them to be dimeric, and in accordance with this view was able to
prepare the same compound from silver succinate anddi-(/3-chloro-ethyl)-
succinate. This method of synthesis evidently merely establishes that
the compound was probably not monomeric.

The properties which Vorlander records for ethylene succinate agree in
general with those which we have observed. We have found that it is
possible to prepare ethylene succinates having, within certain limits, vari-
ous melting points (and molecular weights) by heating succinic acid and
glycol in various proportions. The highest melting (102) is that de-
scribed above. It was prepared from acid and excess glycol. Vorlander
and other writers on this subject ascribe to ethylene succinate a melting
point of 88-90. We have obtained samples showing this melting point
by using the acid and glycol in equivalent amounts; and, by using excess
acid, have obtained lower-melting samples. These ethylene succinates
differ somewhat in solubility, but they are alike in physical appearance
in their lack of volatility and in their viscous character when molten. All
of these materials are highly polymeric. We have also prepared ethylene
succinate from silver succinate and ethylene bromide according to Vor-
lander's directions. This reaction does not proceed smoothly, and our
product was slightly colored and melted at 75 instead of at 90. Never-
theless, molecular weight determinations in boiling ethylene chloride gave
values ranging from 1400 to 2000. We think, therefore, that Vorlander's
molecular weight determinations must be in error; and this conclusion
is supported by the fact that the properties which he records and which
we have observed are not consistent with so low a molecular weight as 288.

We have also prepared ethylene succinate from ethyl succinate and
glycol. It is also highly polymeric. Thus the methods which have been
used for the preparation of this compound include four separate and dis-

26 POLYMERIZATION AND RING FORMATION

tinct methods commonly used in the preparation of esters. They all lead
to products of the expected composition but of high molecular weights.
There is not the slightest reason for supposing that the monomeric ethylene
succinate should be incapable of existence, or even that it should be un-
stable, but it still remains unknown. These facts find their explanation
in the thesis developed in the previous paper.

Vorlander also prepared ethylene maleate and ethylene fumarate by the
silver salt method. He could not obtain any consistent values for the
molecular weights of these materials. We have prepared the fumarate
from ethyl fumarate and the maleate from maleic anhydride. Our prepa-
rations differ from those of Vorlander in several respects. Thus, our male-
ate was crystalline and our fumarate resinous. The reverse was true for
Vorlander's compounds. Neither our analytical values nor his agree well
with the calculated. These esters appear to be much more complicated
in their behavior than ethylene succinate. They both become completely
insoluble on heating.

Experimental Part

Ethylene Phthalate No. 1. Ethylene glycol 62 g. (I mole) and phthahc anhydride
74 g. (0.5 mole) were heated together for eight hours at 190 under ordinary pressure
and for three hours at 300 under 3 mm. The viscous residue was heated with boiling
water for twenty minutes, dissolved in chloroform, filtered and precipitated with ether
and then dissolved in acetone, filtered and precipitated with water. It was then dried
by heating to 100-1 70. The resulting glassy resin was fairly hard when cold, and be-
came softer on heating. All attempts to induce it to crystallize failed. It was neutral:
2 g. required 0.03 cc. of 0.23 N NaOH for alkalinity toward phenolphthalein. It was
soluble in chloroform, acetone, ethyl acetate and acetic acid; insoluble in petroleum
ether, ether, benzene, alcohol and water.

Anal. Substance dried to constant weight in high vacuum at 70. Calcd. for
CioH s O 4 : C, 62.48; H, 4.19; molecular weight, 192; saponification number, 96. Calcd.
forH[O(CH 2 ) 2 OOCC 6 H 4 CO] 26 0(CH 2 ) 2 OH = (W^Oio*: C, 62.15; H, 4.28; molecular
weight, 4804; saponification number, 97.3. Found: C, 01.86, 61.95; H, 4.29, 4.30;
molecular weight by micro boiling point method in ethylene chloride, 4830, 5070, 4680,
4690; saponification number, 95.7.

Ethylene Phthalate No. 2. This was prepared in the same way as No. 1 , but with
a 20% excess of phthalic anhydride and was purified in a similar fashion. Appearance
and solubility were the same as for No. 1 .

Anal. After drying to constant weight in high vacuum at 70. Found: C, 62.11,
62.07; H, 4.29, 4.33; molecular weight, by method of Menzies and Wright in ethylene
chloride, 2940, 2700, 3020; by micro boiling point method in ethylene chloride, 3030,
2930; saponification number, 96.2.

Ethylene Phthalate No. 3. Prepared by heating diethyl phthalate (0.5 mole) with
ethylene glycol (1 mole) in the same fashion and purified as before. Appearance and
solubility were the same as for No. 1.

II. POLYESTERS 27

Anal, after drying to constant weight in high vacuum at 70. Found: C, 62.35,
62.14; H t 4.40, 4.37; molecular weight, by freezing point lowering in diphenyl ether,
2070, 2030; by method of Menzies and Wright in ethylene chloride, 1990, 2100, 2050,
1770; by micro boiling point method in ethylene chloride, 2100, 1870.

Ethylene Phthalate No. 4. This was prepared by stirring vigorously 19 g. of glycol
(added slowly) with 60 g. of phthalyl chloride and 51 g. of dry pyridine in 125 cc. of
chloroform at 0-5. The reaction mixture was washed thoroughly with dilute acid and
dilute sodium carbonate and water, and decolorized with Darco. After drying the
chloroform solution the ester was precipitated with ether. The yield was 53 g. It
resembled the other ethylene phthalates, but its solutions in chloroform were less viscous
0.016 poise for a 20% solution at 27 as compared with 0.027-0.031 poise for the other
ethylene phthalates. Its apparent molecular weight was also lower.

Anal, after drying to constant weight in high vacuum at 70. Found: C, 61.64,
61.64; H, 4.28, 4.30; molecular weight, by method of Menzies and Wright in ethylene
chloride, 1550, 1610; saponification number, 99.5, 99.6.

Hydrolysis of Ethylene Phthalate. Twenty grams of ethylene phthalate was re-
fluxed for sixty-four hours with 60 g. of 48% hydrobromic acid. After neutralization
with sodium carbonate, the reaction mixture was steam distilled. From the distillate
was isolated 16.7 g. or 86.6% of the calculated amount of ethylene bromide, b. p. 131-
134. The residue on acidification gave 16.3 g. or 93.6% of the calculated amount of
phthalic acid. This on conversion to the anhydride melted at 130-131 .

Trimethylene Phthalate. Seventy-four grams (0.5 mole) of phthalic anhydride
and 38 g. (0.5 mole) of trimethylene glycol were heated at 250 for two hours under
ordinary pressure and under diminished pressure for two hours. The residue was dis-
solved in benzene, treated with Darco, filtered and precipitated with ether. When cold
it was a clear, glassy solid, somewhat softer than ethylene phthalate. It was soluble
in chloroform, benzene, acetone, ethyl acetate and acetic acid; slightly soluble in alco-
hol; insoluble in ether, petroleum ether and water.

Anal. Calcd. for C n HioO 4 : C, 64.08; H, 4.89; molecular weight, 206. Calcd. for
H[OOCC6H 4 COO(CH 2 ) 3 ] 1 60(CH 2 ) 3 OH = C m H m O^: C, 63.63; H, 5.05; molecular
weight, 3168. Found: C, 63.68, 63.51; H, 5.17, 5.12; molecular weight by micro boil-
ing point method in ethylene chloride, 3180, 3030.

Hexamethylene Phthalate. Twenty-nine and six-tenths g. of phthalic anhydride
and 23.6 g. of hexarnethylene glycol were heated to 180-190 under atmospheric pressure
for two and one-half hours and then for one and one-half hours at 250 under 5 mm.
In addition to the water, some phthalic anhydride collected in the receiver. The resid-
ual dark gum was purified by dissolving in benzene, decolorizing with Darco, and pre-
cipitating with ether. It formed a clear, light brown, sticky gum. Solubility: solvents
are the same as for trimethylene phthalate, but hexamethylene phthalate is more soluble.

Anal, after drying to constant weight in high vacuum at 70. Calcd. for Ci4HioO 4 :
C, 67.75; H, 6.50; molecular weight, 248. Calcd. for H[O(CH 2 )6OOCC6H 4 CO]7O-
(CH 2 ) 6 OH = Ci 94 H 126 O 3 o: C, 67.35; H, 6.85; molecular weight, 1855. Found: C,
66.74, 66.84; H, 6.75, 6.85; molecular weight by freezing point lowering in benzene,
1700, 1830.

Decamethylene Phthalate. Seven and four-tenths g. of phthalic anhydride and
9 g. of decamethylene glycol were heated at 190-200 for two hours under atmospheric

28 POLYMERIZATION AND RING FORMATION

pressure, and at 210-220 for one and one-half hours at 5 mm. The residue was dis-
solved in benzene, decolorized with Darco and precipitated by petroleum ether. It
was a clear, light brown, thick, sticky sirup. Solubility: solvents for decamethylene
phthalate are the same as for hexamethylene phthalate, but the former is more soluble.

Anal, after drying to constant weight in high vacuum at 70. Calcd. for CiaHa^:
C, 71.01; H, 7.95; molecular weight, 304. Calcd. for H[O(CH2)ioOOCC 6 H 4 CO]7O-
(CH 2 )i OH = Ci 3 H 19 oO 3 o: C, 70.80; H, 8.39; mol. wt. 2303. Found: C, 70.66, 70.44;
H, 8.21, 8.26; molecular weight by freezing point lowering in benzene, 2250, 2060.

Ethylene Fumarate. Fifty-seven and three-tenths g. of diethyl fumarate
(0.33 mole) and 25 g. of ethylene glycol (0.4 mole) were heated for ten hours in a stream
of nitrogen, the temperature being gradually raised from 190 to 230 and the pressure
being reduced at the end to 4 mm. The residue weighed 35 g. or 75% of the calculated
amount. It was accompanied by some insoluble material from which it was freed by
solution in chloroform, filtration and precipitation with ether. It was washed several
times with dry ether. It formed a transparent, slightly yellow, moderately tough mass.
After drying it became insoluble in the common solvents.

Anal, after drying to constant weight in high vacuum at 70. Calcd. for C HeO4:
C, 50.70; H, 4.22. Found: C, 51.95, 51.89. H, 6.06, 6.12. Molecular weight de-
terminations could not be made because of the lack of solubility after drying

Ethylene Maleate. Thirty-two and five-tenths g. of maleic anhydride (0.33
mole) and 18.6 g. (0.30 mole) of glycol were heated at 195-200 for four hours, and
then for some time at 210-215 under reduced pressure. The residue (40 g.) was
separated from some insoluble material by solution in warm ethylene chloride and fil-
tration. It was precipitated by cold ether. The product separated as an oil but
solidified on standing at 5-10 for two hours. It was a white powder. Most of it
melted between 88 and 95. After drying in vacua it had become insoluble in the com-
mon solvents including ethylene chloride and it did not melt below 250 .

Anal, after drying to constant weight in high vacuum at 70. Calcd. for CeHeO^
C, 50.70; H, 4.22. Found: C, 49.87, 49.70; H, 4.36, 4.28. Molecular weight de-
terminations could not be made because of the lack of solubility after drying.

For his kind assistance in the analytical work we here express our thanks
to Mr. Wendell H. Taylor.

Summary

The following esters have been prepared: ethylene malonate, ethylene
succinate, trimethylene succinate, ethylene adipate, trimethylene adipate,
hexamethylene succinate, hexamethylene adipate, ethylene sebacate, tri-
methylene sebacate, decamethylene succinate, hexamethylene sebacate,
decamethylene adipate, decamethylene sebacate, ethylene maleate, ethyl-
ene fumarate, ethylene phthalate, trimethylene phthalate, hexamethylene
phthalate and decamethylene phthalate. Their molecular weights have
been determined. They are all highly polymeric. Their properties are
described and their structures are discussed.

III. GLYCOL KSTERvS OF CARBONIC ACID 29

Bibliography and Remarks

(1) The values C = 1.5 A. and O 1.1 A. are taken from Meyer, Naturwtsscnschaftcn, 42, 782
(1928).

(2) Ref. 1, p. 21.

(3) So far as they have been examined in this respect, all these esters give sharp x-ray diffraction
patterns.

(4) Incidentally these are the forms in which Hess* trimethylcellulose crystallizes, and the photo-
micrographs (Figs. 78, 80 and 81) which he presents ("Die Chemie der Zellulose," Leipzig, 1928, p
432) for this substance would serve almost perfectly as pictures of ethylene succinate. The crystal-
linity of this trimethylcellulose therefore provides no guarantee either that it contains only a single
molecular species, or that the molecules present are not very large.

(5) Some experiments were kindly made for us by Dr. E. L. vSkau with the special apparatus
which he has devised for the precise determination of the melting points of pure organic compounds.

(6) Pictet and Vogel, Helv. Chim. Ada, 11, 901 (1928).

(7) See for example U. S. Patents Nos. 1,108,332; 1,091,627; 1,091,628; 1,091,732; 1,108,329;
1,108,330; 1,108,331; 1,642,079; 1,678,105; 1,098,776; 1,098,777, 1,424,137, 1,413,144; 1,413,145,
1,667,199; 1,667,200; 1,119,592; 1,141,944; 1,663,183.

(8) Tilitschejew, /. Russ. Phys. Chem. Soc., 57, 143 (1925).

(9) It is interesting to note that Freudenberg and Braun's trimethylcellulose, like ethylene sue
cinate, is insoluble in ether and carbon tetrachloride, but quite soluble in chloroform [Ann., 460, 288
(1928)].

(10) It is interesting to observe that inulin, a polymeric substance of a different type, but probably
of the same degree of molecular complexity as the esters here described is also quite hygroscopic. Cf
Drew and Haworth, J. Chem. Soc., 2690 (1928).

(11) Vorlander, Ann.. 280, 167 (1894). See also Bischoff, Ber.. 27, 2940 (1894); 40, 2779, 2803
(1907). Glycol esters of adipic acid are referred to in German patent 318,222, Chem. Zentr., II, 536
(1920).

(12) Lourenco, Ann. chim. phys., 293 (1863).

(13) Davidoff, Ber., 19, 406 (1886).

III. Glycol Esters of Carbonic Acid*

This paper describes different glycol esters of carbonic acid. The aim is
to check whether condensations which can lead to 5- and 6 -rings remain
intramolecular, while such as can give only larger rings lead to C-
polymers. In fact it was shown that as long as 5- or 6-rings can be formed
(ethylene carbonate and trimethylene carbonate) well crystallized, low
molecular compounds, with molecular weights of about 90 and 110, are
produced; as soon as only larger rings can be formed C-potymers, with
molecular weights up to about 3000, result). ^
Trimethylene carbonate

0=C

/2\
< >CH 2

X) CH 2 '

* W. H. Carothers and F. J. van Natta; Journ. Am. Chem. Soc. 52, 314-26 (1930) ;
Contribution No. 19 from the Experimental Station of E. I. du Pont de Nemours and
Co.

Received July 24, 1929. Published January 8, 1930.

t Compare the thesis developed on pages 10 and 12.

30 POLYMERIZATION AND KING FORMATION

which contains a 6 -ring can be obtained in both forms: as a monomeric di-
ester with a molecular weight of 104 and as a polymer with a polymerization
degree between 38 and 45. This latter is apparently formed by an A -
polymerization of the monomeric trimethylene carbonate.
The preparation and the properties of the following substances are described
and discussed:

trimethylene carbonate decamethylene carbonate

tetramethylene carbonate diethylene carbonate

pentamethylene carbonate p-xylylene carbonate
hexamethylene carbonate

The glycol esters of dibasic acids described in a previous paper (1) are all
highly polymeric. This was expected from the generalization (2) that bi-
functional reactions which, if intramolecular, could lead only to larger-than-
(i rings proceed intermolecularly. It seemed desirable to examine an ho-
mologous series of similar compounds in which the length of the structural
unit might be as short as 5 or 6 atoms, since in these cases reaction should
be intramolecular and the products monomeric, and in the longer chains
of the same series reaction should be intermolecular and the products
polymeric.

Such a series is found in the glycol esters of carbonic acid. The first

yO CH 2
member of this series, ethylene carbonate, O=C<f \ , has long been

X) CH 2

known (3). It is a crystalline solid which boils at 238 arid it is definitely
established to be the monomeric 5-ring by both cryoscopic and vapor
density data (4).

We have now prepared the trimethylene, tetramethylene, hexamethyl-
ene, decamethylene, diethylene and p-xylylene esters of carbonic acid.
The properties of these esters, together with the apparent mean molecular
weights determined by cryoscopic and ebullioscopic methods, are indicated
in Table I. The analytical compositions of all these esters correspond with
the formulas of their structural units O -(CH 2 ) M CO, but where
the length of the chain of this unit is 5 or 6, only one such unit is present
in a molecule of the ester. On the other hand, where the length of the
chain of the structural unit is 7, 8, 9 or 13, eight to twenty-two structural
units are present in each molecule. The method of preparation and the
analytical compositions of the polymeric esters indicate a structure which
may be represented by the general formula O (CH 2 ) M -O CO O
(CH 2 ) n -O-~- CO O(CH 2 ) M -O CO O (CH 2 ) W O CO-, etc. To

III. GLYCOL ESTERS OF CARBONIC ACID

a -S

I " **

. * 2

oop .g

8^-S 2 &
J8"*;3 MO

w 8

da dd

OO O
d
PQ

t? ~

OcM O

o- >.

and
55-

? 2 v

- 2

EX W
00 O
dd d

gj 00 ww uy

CN ^Ht-i -(

oo o

O 001 iO

Q -

> t?5 oo

l d

s
a

o *=

S5
C

cc
pi

^% ^ ^1 00 * S

OW OW OW OW oW

C'O

^-<* oe> pci '"'S! 3?i5 ^2 Sfe 5I'~ < 2Jt:

O>to wO> O OJ l-O> Or* CO c ^ G5^ ^^-i OOC

o o

1 . 1 _ J . [ . r)-o
O

(CHi

9 $

i O C-
O

O O

^O O'

fc 5*

^* w

Namesof
carbonate*

Hthylene

Trimethylen

Tetramethy

Pentamethy

I I

i
-

' O 0C

> >*

1 1

^-X

'Sb

"g So

'E 1

32 POLYMERIZATION AND RING FORMATION

complete this formula, it is necessary to discover whether the valences
at the ends of the chains are saturated mutually with the formation of a
ring, or are saturated by univalent groups to give open chains. The 7-,
8-, 9- and 13-membered rings which might be formed in this reaction are
not found, and there is no reason to suppose that much larger rings would
be formed more readily. For some polyesters similar to those described
here we have direct experimental proof of the open-chain structure. It
seems quite certain, therefore, that these polycarbonates are also open
chains. From the method of preparation it follows that the ends of the
chains must bear hydroxyl groups (from an extra molecule of glycol) or
carbethoxy groups (from an extra molecule of ethyl carbonate). In this
connection it is interesting to observe that when an excess of ethyl car-
bonate was used in the preparation of hexamethylene carbonate, pure
dicarbethoxyhexane, C 2 H 6 O CO O(CH 2 ) 6 O CO OC 2 H 6 , was isolated,
and an oil which, judging by its composition, was composed chiefly of C 2 H&-
O CO 0(CH 2 ) 6 CO 0(CH 2 ) 6 O CO O(CH 2 ) 6 O CO OC 2 H 6 .

Properties of the Polycarbonates. The physical properties of these
polycarbonates again illustrate the fact that neither melting points nor
solubilities offer any general criteria of molecular size.

In this connection trimethylene carbonate is of especial interest since
it has been obtained in both the monomeric form and a form in which
the degree of polymerization is 38-45. The monomeric form is hygro-
scopic and is soluble in water, benzene and alcohol, but only slightly soluble
in ether and ligroin. The polymeric form is not soluble in water and alco-
hol, but is still soluble in benzene. Solubility here diminishes with con-
siderable increase in molecular size, but it does not disappear. Prac-
tically all the polymeric esters described in this and the previous paper
are insoluble in water and alcohol, but soluble in chloroform and somewhat
soluble in acetone. Those polyesters in which the structural units con-
tain polymethylene chains (CH 2 )^, in which x is greater than 3-5, show
great solubility in benzene.

Polyesters of high molecular weight may be either liquid (diethylene car-
bonate, ethylene malonate (1), decamethylene phthalate (1)) or solids.
Most of the solid polyesters are crystalline. These crystalline polyesters
differ from analogously constituted crystalline esters of low molecular
weight in that the crystals are always microscopic. In some cases it is not
possible to decide even by microscopic examination whether these solid
polymers are crystalline, but x-ray examination always shows definite crys-
tallinity. Thus hexamethylene carbonate was an opaque, tough, horny
material with considerable elasticity. Nevertheless, it had a fairly definite

III. GLYCOL ESTERS OF CARBONIC ACID 33

melting point and it gave a quite sharp x-ray diffraction pattern
(Fig. 3).

The following generalizations concerning the influence of the structure
of the structural units of polyesters on their physical properties can now
be offered. As the lengths of the polymethylene chains, (CH2)*, which
separate the ester groups increase, solubility in organic solvents increases
and viscosity of the molten ester diminishes. This would be expected
from the fact that carbonyl groups contribute much more heavily to in-
termolecular association forces than do methylene groups. With one
exception (ethylene malonate (1)) all of the linear polyesters in which the
ester groups are separated only by polymethylene chains, (CH^)*, are
solids. Less symmetrical linear polyesters such as the alkylene phthalates
are resins (transparent glasses or sirups). The solid polyesters when
melted and then cooled yield opaque masses. These may be brittle and
porcelain-like or soft and waxy. The waxy quality increases with the
lengths of the polymethylene chains just as it does in the glycols and the
acids from which the polyesters are derived.

The problem of the more precise expression of the relationship between
the structure and the properties of high polymers is complicated by the
fact that some of the properties of this class of substances which are of
the greatest practical importance and which distinguish them most sharply
from simple compounds cannot be accurately measured and indeed are
not precisely defined. Examples of such properties are toughness and
elasticity. We have found these two properties in only one of the poly-
esters which we have prepared. This polyester is hexamethylene car-
bonate. Like pentamethylene and decamethylene carbonates, this ma-
terial is crystalline; it has a fairly definite melting point (55-60) and it
gives a sharp x-ray diffraction pattern (see Fig. 3). But the higher
and lower members of this series separate from solution as powders, while
hexamethylene carbonate separates in the form of rubbery flakes. It
is not yet certain that this peculiar combination of properties is inherently
associated with the molecular structure of this particular ester (i. e., with
the linear union of a large number of the structural units CO O
(CH 2 )e O ) rather than with some accidental features of its formation
which may have been absent in the preparation of other members of the
series, but the important problem of the relationship between chemical
or physical structure and rubber-like properties is so poorly provided with
material suitable for inductive argument that the discovery of a new syn-
thetic and analytically homogeneous material exhibiting these properties
is of some significance. So far as we are aware there is nothing in current

34 POLYMERIZATION AND RING FORMATION

theories which would lead one to expect that polymeric hexamethylene
carbonate wonld be elastic.

As indicated above, high polymers, like materials of low molecular
weight, may show great variety in certain physical properties they may
be either liquids or high-melting solids; they may show slight or very
great solubility. The following properties, however, are inherently asso-
ciated with the highly polymeric state.

(1) Lack of Volatility. None of the polyesters which we have
prepared can be volatilized as such even at very low pressures and high
temperatures. Volatility diminishes continuously with increase in mo-
lecular size due to the increase in intermolecular forces. It is probably
true in general that compounds whose molecular cohesions lie above 75,000
cal. (the heat of separation of a carbon-carbon bond) cannot be distilled.
The calculated molecular cohesion (about 250,000 cal.) (5) of our poly-
esters lies far above this value.

(2) Viscosity in Liquid State. This would be expected to increase
with the intermolecular forces and so with molecular size. All of the poly-
esters which we have studied are extremely viscous in the molten state.

(3) Micro-crystallinity in the Solid State. This has been referred
to above. In this connection it is interesting to observe that considerable
difficulty was experienced in preparing and keeping crystals of rnonomeric
trimethylene carbonate sufficiently small to yield a powder diagram in
an x-ray diffraction experiment. On the other hand, it has not been pos-
sible to prepare macroscopic crystals of any polymeric ester.

Trimethylene Carbonate. Among all the esters formed from bi-
f unction al reactions which we have studied, trimethylene carbonate and
ethylene oxalate (0) present the peculiar property of exhibiting reversible
transformation between a rnonomeric and a polymeric form. These rnono-
meric esters contain 6-membered rings. The polymerization of trimethyl-
ene carbonate is brought about by heating and is catalyzed by a trace of
potassium carbonate. The polymer is a perfectly colorless, transparent
glass, which on long standing becomes opaque (crystallization). X-ray
diffraction patterns of the rnonomeric and polymeric forms are shown in
Figs. 1 and 2. The polymer has the same apparent analytical composition
as the monomer. Its degree of polymerization as measured by molecular
weight determinations in boiling benzene is 38-45. On heating in vacua,
it distils and the distillate is found to be the rnonomeric form. Examina-
tion of the literature shows that the property of undergoing reversible
polymerization is common to many 6-rings containing ester linkages.
Thus, the lactones of 6-hydroxyvaleric acid and of the hydroxyethyl ether

III. GLYCOL ESTERS OF CARBONIC ACID

of glycolic acid polymerize on heating (7) and so does glycolide (8). The
action of heat on lactic acid leads to polylactyl lactic acids, and these on

further heating are converted to
the 6-ring lactide (9).

The unique position occupied
in this respect by 6-rings contain-
ing an ester linkage is indicated
by the following facts. No record
of the polymerization of a 7-lac-
tone (5-ring) is found. Our own
attempts to polymerize ethylene
carbonate (5-ring) were unsuc-
cessful. The polyesters which we
have prepared and in which the
length of the chain of the] struc-
tural unit is greater than 7 are not
depolymerized on heating. The
monomeric form of none of these

Fig. 1. Monomeric trimethylene car-
bonate (Mo radiation). The intensity of
the numerous outer rings is weak and these
do not appear in the reproductions.

esters is known.

In an attempt to force the de-
polymerization of tetramethylene
carbonate, it was heated above 300 at 0.9 mm. This caused it to de-
compose with the evolution of consider-
able gas. From the distillate was iso-
lated a very small amount (about 1% of
the calculated) of a crystalline solid
having the analytical composition of
tetramethylene carbonate. Molecular
weight determinations indicated that it

is a dimer: COO(CH 2 ) 4 OCOO(CH 2 ) 4 O.
The absence of a monomer was not defi-
nitely established, but it could only have
been present in very small amount.
Thus, the behavior of polymeric tetra-
methylene carbonate is very different
from that of its next lower homolog.
In the self-esters of some of the

higher cu-hydroxy fatty acids, HO(CH 2 )*COOH, both the monomeric lac-
tones and the polyesters are known (10), The polyesters, however, are

Fig.

2. Polymeric trimethylene
carbonate (Cu radiation).

POLYMERIZATION AND RING FORMATION

not depolymerized on heating, and the polyesters, not the lactones, are
formed on heating the acids.

These facts, together with others which will be discussed later, indicate
clearly that polyesters formed by bifunctional reactions in which the
length of the structural unit is greater than 6 atoms are the direct result of
intermolecular condensation (C polymerization) (2). The formation of
polymeric trimethylene carbonate from its monomer is obviously A poly-
merization, and the same is true of the polymerization of lactides and of
5-lactones.

A fairly satisfactory explanation of these peculiarities may be found
in the Mohr theory (11), which has proved adequate to explain all of the
previously observed influences of structural features on the course of bi-
functional reactions (2). In accordance
with the terms of this theory, 5-rings
will be readily formed and very stable
since they are free from strain. The
formation of seven-membered and larger
rings by bifunctional reactions is very
improbable, but such rings once they
are formed will be quite stable, since
they too are practically free from strain.
Six-rings will be readily formed, but
they present the possibility of strains
which are absent or present only to a
diminished degree in larger rings. Thus,
the Mohr theory predicts the possible ex-
istence of two cyclohexanes (cis and trans)

but only one is known. However, where structural features would pre-
clude the possibility of the intercon version of these two forms, as in dec-
alin, both forms are known. It is concluded, therefore, that in simple
6-rings, such as cyclohexane, equilibrium with reversible interconversion
of the two forms exists. At every such interconversion the molecule must
pass through a position of considerable strain. This picture presents the
features necessary to rationalize the peculiar position of 6-membered ester
rings. A polyester in which the recurring structural unit is a chain of
six atoms would be readily converted into the corresponding monomeric
6-ring by a process of ester interchange, as indicated below :

Fig. 3. Hexamethylene carbonate
(Cu radiation).

-etc.

III. GLYCOL ESTERS OF CARBONIC ACID 37

If the chain of the structural unit were longer than 6 atoms, this reaction
would be less probable for the same reason that the higher co-hydroxy
fatty acids do not yield lactones, i. e., the number of points of rotation
in the chain between carbonyl and alcoholic oxygen are so numerous that
the probability of close intramolecular approach" of these groups is very low.
The polymerization of the 6-ring esters seems also to involve ester inter-
change. This process may involve the coalescence of two such rings,
as indicated below

C -- 0-Csr-C

r/0 \

c o-

C-C---O-C
V ^-R ^ x

The resulting 12-ring might then coalesce with another six-ring, and so
011, with the formation of very large rings. It is possible also that traces
of water may intervene at some stage in the process and that the large
molecules produced are open rather than closed chains. In any event
the peculiar mobility of the 6-ring esters may be ascribed to the fact that
these rings are subject to strains, and that the strains may be relieved by
an ester interchange resulting in an enlargement of the molecule.

The precise mechanism of the polymerization of the 6-menibered ester
rings shares to a certain extent the obscurity which is common to most
cases of A polymerization, but further attention is being devoted to this
problem and the study is being extended to the corresponding amides.
In this connection the analogy which exists between lactides and diketo-
piperazines on the one hand, and polyesters and polypeptides on the
other hand, is of especial interest because of its possible relationship to the
structure of proteins.

Dimeric Tetramethylene Carbonate. The dimeric tetramethylene
carbonate referred to above deserves special mention. There can be little
doubt that it is a 14-membered cyclic ester, and as such it represents a
class of compounds of which few members are known. Other dimeric cy-
clic esters have been obtained and will be described in later papers.

Previous Work on the Carbonates of Dihydric Alcohols. Reference
to ethylene carbonate has already been made. Carbonates of the three
dihydroxybenzenes were prepared by Bischoff and v. Hedenstrom (12).

38 POLYMERIZATION AND RING FORMATION

The carbonate of catechol is apparently monomeric since it can be distilled.
The w-phenylene carbonate is a non-volatile, brown, glassy material of
uncertain melting point (197-202) and low solubility, and the corre-
sponding para compound shows similar properties. These two compounds
are evidently polymeric.

Experimental Part

Preparation of Carbonates. The carbonates here described were all
prepared by ester interchange (alcoholysis) between the glycol and ethyl
carbonate. This reaction proceeds very smoothly in the presence of an
alkaline catalyst and may be forced to completion; the process need not
be modified to take into account the solubility of the resulting ester, as
is necessary when phosgene and the glycol are used as starting materials.

Trimethylene Carbonate. A mixture of 60.8 g. (0.8 mole) of trimethylene glycol,
114 g. (16% excess) of ethyl carbonate and a small cube of sodium was warmed in a
Claisen flask until the sodium had completely dissolved. The flask was then immersed
in an oil-bath at 130. During three hours the temperature was allowed to rise to 170.
The evolved alcohol condensed and collected weighed 69.1 g. or 94% of the calculated.

The oily residue was dissolved in benzene, washed twice with 20 cc. portions of
water and dried over calcium chloride. After removal of the solvent, the residue was
subjected to vacuum distillation. A small initial fraction, b. p. 120-135 at 20 mm.,
was unchanged glycol. The pure trimethylene carbonate then distilled fairly constantly ;
b. p. 160-165 at 6 mm.; 135 at 4 mm., 105 at 0.2 mm.; weight 53 g., or 65% of the
calculated. It completely crystallized in the receiver; recrystallization from absolute
ether produced colorless needles, m. p. 47-48. This hygroscopic ester is very soluble in
benzene, alcohol and water and slightly soluble in ether and ligroin. It decomposes on
distillation at ordinary pressure. This ester gave a sharp x-ray powder diagram (Fig. 1)
slightly scattered due to the difficulty of obtaining sufficiently small crystals. For
analysis, see Table I.

Polymerization of Trimethylene Carbonate. In the distillation of trimethylene
carbonate it was frequently observed that the distilling residue was very thick and
viscous while the warm distillate was comparatively thin and mobile. The distillation
of 53 g. of trimethylene carbonate was interrupted while still incomplete. The distillate
was monomeric trimethylene carbonate, m. p. 48. The undistilled residue (15.6 g.)
was very viscous even when hot and it cooled to a colorless sticky resin. Molecular
weight determinations showed it to be polymeric ; molecular weight in boiling benzene,
found: 2390, 2320.

This residue was distilled at 10 mm. It yielded 13.6 g. of monomeric trimethylene
carbonate (observed b. p. 150-190) which completely crystallized on cooling; mol. wt.,
calcd. for monomeric trimethylene carbonate: 102. Found: (in boiling benzene)
115, 118.

The tarry residue (1.7 g.) remaining in the flask was insoluble in the common organic
solvents.

A trace (0.1% of finely powdered anhydrous potassium carbonate was added to a
sample of the pure crystalline monomeric ester and the mixture was heated at 130.

III. GLYCOL ESTERS OF CARBONIC ACID 39

After shaking and heating for ten minutes the mixture suddenly became very viscous
and a slight amount of gas was evolved. After five hours of heating the colorless product
formed a clear glassy mass on cooling; mol. wt., found: (in boiling benzene) 4670,
3880. For analysis see Table I. After standing for a week, this glassy polymer became
opaque and x-ray examination showed it to be crystalline (Fig. 2).

Under similar conditions (temp. 130; time of reaction, four hours) ethylene car-
bonate did not polymerize.

When polymeric trimethylene carbonate was heated to 210 at ordinary pressure,
it was decomposed. Allyl alcohol was identified as one of the products of this reaction
by its boiling point and by the melting point of its phenylurethan. This fact indicated
decomposition according to the scheme

/
O CH 2 CH 2 CH 2 O C > CH 2 =-CH CH 2 OH + CO 2

The allyl alcohol isolated accounted for only about 25% of the ester decomposed.
Higher boiling unsaturated materials present in the products of this reaction were not
identified.

Tetramethylene Carbonate. This was prepared from 9.0 g. of tetramethylcne
glycol, to which a small piece (2 mm. 3 ) of sodium had been added and 11.8 g. of ethyl
carbonate. The mixture was heated to 120 and then the temperature was allowed
gradually to rise to 160 during seven hours. The distillate was poured back several
times to insure complete reaction of the ethyl carbonate. From the distillate was iso-
lated 5.7 g. or 62% of the calculated amount of ethyl alcohol.

The non- volatile reaction product was dissolved in benzene and the solution washed
with water, dilute hydrochloric acid and water. It was dried with calcium chloride and
the solvent was removed by distillation. On heating the residue in vacua there was ob-
tained a very small amount of distillate boiling at 100-110 at 1 mm. This was chiefly
unchanged tetramethylene glycol since it gave a di-p-nitrobeiizoate, m. p. 175. No
further distillation could be effected when the bath was heated to 250 and the pressure
reduced to 0.2 mm.

The residue was purified by dissolving it in chloroform and precipitating it with
absolute alcohol while cold. The oily precipitate soon crystallized to a cream-colored
powder; weight 6.3 g. or 54% of the calculated; m. p. 55-59; very soluble in cold
benzene, chloroform, acetone and acetic acid; insoluble in ether, alcohol and petroleum
ether. It gave a sharp x-ray diffraction pattern. Before analysis it was dried for four
days at 80 in high vacuum. For analytical data see Table I.

Di--nitrobenzoate of Tetramethylene Glycol. Prepared from the glycol and the
acid chloride in pyridine ; crystallized from boiling acetic acid ; m p. 175.

Anal. Calcd. for Ci S H 16 O,N 2 : C, 55.67; H, 4.12. Found: C, 56.19, 56.00; H,
4.31, 4.14.

Thermal Decomposition of Tetramethylene Carbonate. An attempt was made to
bring about the depolymerization of tetramethylene carbonate by heating it to a high
temperature in vacua. The tetramethylene carbonate was prepared from 9.0 g. of glycol
and 11.8 g. of ethyl carbonate in the same manner as described above. It was heated at
0.9 mm. in a small Claisen flask by means of a m.tal bath. Very slight distillation
occurred as the temperature of the bath was raised from 270 to 300. Between 300 and
325 evolution of gas occurred and the pressure increased. From the distillate were is j-

40 POLYMERIZATION AND RING FORMATION

la ted 0.1 g. of crystalline solid and 2.8 g. of oil. The residue remaining in the distilling
flask weighed 3 g. and was apparently identical with the polymeric tetramethylene
carbonate described above; it melted at 55 and had the same physical appearance
and solubility behavior.

The crystalline distillate separated from alcohol in minute prisms melting at 175-
176. Analysis and molecular weight determinations showed it to be dimeric tetra-

r i

methylene carbonate, CO O(CH 2 )4O CO O(CH 2 )4O.

Anal. Calcd. for (C B H 8 8 )2: C, 51.71; H, 6.94; mol. wt., 232. Found: C, 51.77;
51.80; H, 6.95; 7.02; mol. wt., in boiling ethylene chloride, 194, 231; mol. wt., in freez-
ing benzene, 204.

The liquid distillate had a pungent mint-like odor and readily absorbed bromine
in acetic acid solution. It is a mixture and no chemical individuals have as yet been
isolated from it.

Pentamethylene Carbonate. This was prepared from 5.9 g. of ethyl carbonate and
5.2 g. of pentamethylene glycol to which a small amount of sodium had been added.
The mixture was heated for twelve hours at 120-160; 76% of the calculated amount
of alcohol was found in the distillate. The residue after it had been freed from sodium
was heated at 250 under 0.15 mm. The distillate (1.1 g.) boiled at 110-130 and con-
sisted chiefly of unchanged glycol; di-/>-nitrobenzoate, m. p., 104-105. The non-
volatile sirupy residue crystallized on standing to a hard waxy material; weight, 4.1
g. or 63% of the theoretical. It was purified by precipitation with absolute alcohol
from a cold chloroform solution as a granular and slightly colored powder; m. p., 44-
46; very soluble in benzene, chloroform, acetone and acetic acid; insoluble in ether,
alcohol and petroleum ether. Before analysis it was dried to constant weight in high
vacuo. For analysis see Table I.

Di--nitrobenzoate of Pentamethylene Glycol. From the glycol and the acid
chloride in pyridine; crystallized from a mixture of benzene and alcohol; m. p., 104-105.

Anal, Calcd. for C 19 H 18 O 8 N 2 : C, 56.71; H, 4.51. Found: C, 56.95, 57.02; H,
4.59, 4.69.

Hexamethylene Carbonate. This was prepared from 12 g. of ethyl carbonate and
12 g. of hexamethylene glycol to which a small piece of sodium had been added. The
mixture was heated from 130 to 170 during two hours; 86% of the calculated amount
of alcohol was found in the distillate. The residue after the removal of the sodium
was subjected to distillation in high vacuum. A very small amount of distillate was
found to consist of unchanged glycol. The residue, which amounted to 10 g. or 67% of
the calculated, solidified on cooling to a light colored, horny, tough mass. It was soluble
in benzene, acetone and chloroform; insoluble in ether and alcohol; m. p. 55-60. A
sharply defined x-ray diffraction pattern (Fig. 3) showed it to be crystalline. For
analysis see Table I.

Dicarbethoxyhexane. In an experiment similar to the above, ethyl carbonate and
glycol were used in the ratio of 2 moles to 1 (24 g. of ester and 12 g. of glycol). The
calculated amount (9.2 g.) of alcohol was collected. The residue after removal of sodium
was subjected to vacuum distillation. It yielded 5.5 g. of distillate, b. p. 130140 at
0.8 mm. The residue could not be distilled, and neither the distillate nor the residue
could be induced to crystallize. No ethyl carbonate was found in the distillate. The
distillate was dicarbethoxyhexane; rc 2 ,? 1.4310; dig 1-065; d\l 1.056.

III. GLYCOL ESTERS OF CARBONIC ACID 41

Anal. Calcd. for Ci2H 2 2O 6 : C, 54.94; H, 8.41; mol. wt., 262. Found: C, 55.00;
H, 8.41; mol. wt. in boiling benzene, 236, 240, 245.

The oily residue was very soluble in alcohol and ether. Its analytical composition
corresponded with the formula C 2 H 5 O CO O(CH 2 ) 6 O CO O(CH 2 ) 6 O CO O-
(CH 2 ) 6 CO OC 2 H 6 .

Anal. Calcd. for C 2 6H 4 6Oi 2 : C, 56.73; H, 8.36; CO 2 on hydrolysis, 33.1%; mol.
wt., 532. Found: C, 56.77, 57.00; H, 8.41, 8.53; CO 2 , 34.15; mol. wt. in boiling ben-
zene, 573, 549, 528. The CO 2 was determined by saponifying with alcoholic sodium
hydroxide and weighing the sodium carbonate formed.

Carbonate of Diethylene Glycol. This was prepared in 43% yield by the method
used in the preparation of the other carbonates. It was a light colored sirup which
could not be induced to crystallize; insoluble in alcohol and ether, quite soluble in
acetone and benzene, very soluble in chloroform and hot ethyl acetate. On standing it
appeared to decompose to some extent into acetaldehyde and carbon dioxide and for
this reason it could not be dried to constant weight. For analysis see Table I.

Decamethylene Carbonate. This was prepared in 75% yield by the method used
for the preparation of the other carbonates. It was soluble in benzene and chloroform
but insoluble in alcohol. It was purified by precipitation from chloroform with alcohol
in the cold. The oily product gradually solidified to a cream-colored powder; m. p.
55; very soluble in chloroform, slightly soluble in ether, benzene, acetone and acetic
acid; insoluble in alcohol and petroleum ether. This powder gave a sharply defined
x-ray diffraction pattern. Before analysis it was dried to constant weight in high vacuum
at 90. For analysis see Table I.

/>-Xylylene Carbonate. This was prepared from />-xylylene glycol and ethyl car-
bonate as in the preparation of the other carbonates. The residue after completion of
the ester interchange was insoluble in hot benzene, absolute alcohol, chloroform and
carbon tetrachloride ; slightly soluble in acetic acid. It dissolved more or less completely
in ethylene chloride, anisole and dioxane. It was thoroughly triturated with water and
with cold alcohol to remove sodium and excess glycol and ethyl carbonate. The amount
of the white solid remaining corresponded to 82% of the calculated. It was separated
into two fractions by extraction with ethylene chloride.

(a) Soluble Fraction. This was precipitated with ether from ethylene chloride
in the form of white flocks; m. p. 137-138. For analysis see Table I.

(b) Fraction Insoluble in Ethylene Chloride. This was insoluble in the common
organic solvents; m. p. 177-185 (rather indefinite). For analysis see Table I.

Attempted Preparation of Methylene Carbonate. Silver carbonate (81 g.) was
heated with methylene bromide (51 g.) in dry toluene solution. Formaldehyde and
carbon dioxide were formed and free silver was produced. From the reaction mixture
there was isolated only a small amount of oil with a sweet odor probably methylene car-
bonate. This liberated formaldehyde on heating: CH 2 CO 3 -* H 2 CO + CO 2 .

We are indebted to Dr. A. W. Kenney and Mr. Henry Aughey for the
x-ray diffraction pictures, to Mr. W. H. Taylor for many molecular-
weight determinations and to Mr. G. A. Jones for determinations of carbon
and hydrogen.

42 POLYMERIZATION AND RING FORMATION

Summary

The following compounds have been prepared by the action of the ap-
propriate glycols on ethyl carbonate : trimethylene carbonate, tetramethyl-
ene carbonate, pentamethylene carbonate, hexamethylene carbonate, deca-
methylene carbonate, diethylene carbonate and p-xylylene carbonate.
Their properties are described. In accordance with a generalization al-
ready set forth, ethylene carbonate and trimethylene carbonate are mono-
meric and the other carbonates are polymeric. Trimethylene carbonate
undergoes reversible A polymerization. A stable dimeric form of tetra-
methylene carbonate has been prepared by thermal decomposition of its
usual polymeric form. Hexamethylene carbonate is tough and elastic.

Bibliography and Remarks

(1) Carothers and Arvin, Journ. Ant. Chem. Soc., 51, 2560 (1920).

(2) Carothers, ibid., 51, 2548 (1929).

(3) Nemirowski, J. prakt. Chem., (2} 28, 439 (1883); Chem. Zenlr., 23 (1884).

(4) Vorlander, Ann , 280, 186 (1894).

(5) Meyer and Mark, Z. angew. Chem., 41, 943 (1928).

(6) See Bischoff, Ber., 40, 2803 (1907). Our own studies of alkylene oxalates will be described
later.

(7) Hollo, Ber. t 61, 895 (1928).

(8) Bischoff and Walden, Ann., 279, 45 (1894).

(9) Dietzel and Krug, Ber., 58, 1307 (1925).

(10) Ruzicka and Stoll, Helv. Chim. Acta, 11, 1159 (1928); Lycan and Adams, Journ. Am. Chem.
Soc., 51, 625 (1929); Chuit and Hausser, Helv. Chim. Ada, 12, 463 (1929); Lycan and Adams, Journ
Am. Chem. Soc., 51, 3450 (1929).

(11) Mohr, J. prakt. Chem., 98, 348 (1918).

(12) Bischoff and v. HedenstrSm, Ber., 35, 3431 (1902).

IV. Ethylene Succinates*

This contribution contains a more thorough investigation of ethylene suc-
cinate, its formation, structure and bearing on the association theory of
high polymers.

Neutral ethylene succinates are prepared by heating succinic acid with
excess glycol at 180 until the distillation of water stops: they are chains
of about 3000 m. w.

Acidic ethylene succinates are obtained if an excess of succinic acid is ap-
plied. The reactants are heated to 180 until no more water is evolved and
then the temperature is raised to 200-240 for one to five hours to make

* W. H. Carothers and G. L. Borough; Journ. Am. Chem. Soc., 52, 711-21 (1930);
Contribution No, 20 from the Experimental Station of E. I. du Pont de Nemours and
Co.

Received August 7, 1929. Published February 6, 1930.

IV. ETHYLENE SUCCINATES 43

reaction as complete as possible. The products are always acidic.
The polymerization degree varied from 6 to 23. Molecular weights were
determined by the boiling point method and by titration of the end-standing
acid group. Excellent agreement was obtained between the two methods*
and the theoretical values for the above mentioned polymerization degrees.
The mecJianism of the formation of chain polymers is investigated and the
possibility of an association of a number of 8-rings through exceptionally
strong van der Waals forces discussed.
Carothers expresses himself in favor of the main valence chain theory.

Ethylene succinate as a typical example of a condensation polymer (1)
has been submitted to further study with the view of gaining more in-
formation in regard to its structure. The presence of the structural unit,
(1) I, in this ester has been established by the methods used in its synthe-
sis (2, 3), by its chemical behavior and by its analytical composition (2, 3).
O CH 2 CH 2 O OC CH 2 CH 2 CO (I)

Of the various polymeric forms described in this paper, the apparent
molecular weights indicate average values of 6 and 23, respectively, for
the number of structural units contained in each molecule of the lowest
and highest polymers. It is assumed that these units are joined together
in a linear fashion by real primary valences, as in II. Since it is very im-
probable that there are free valences at the ends of the resulting chains,
the problem of the structure of these polymers resolves itself into finding
whether the end valences are mutually saturated with the formation of
very large rings or are saturated by univalent groups of some kind.

~O~-(CH 2 ) 2 O OC(CH 2 ) 2 CO O (CH 2 ) 2 O OC (CH 2 ) 2 CO O

(CH 2 ) 2 O OC (CH 2 ) 2 CO , etc. (II)

Any of the reactions by which this ester is prepared offers the formal

possibility of establishing either rings (closed chains) of 8, 16, 24,

(S)n members or of open chains corresponding to each of these rings. The
observed molecular weights indicate the presence of 50 to 180 atoms in
each chain. This fact and the physical properties of the ethylene succi-
nates indicate the absence of more than traces of rings of less than 32 mem-
bers. Since the same theoretical considerations (1, 4) which predict the
improbability of the formation in bifunctional reactions of rings of 8 to 32
members apply a fortiori to still larger rings, an open chain seemed much
more probable than a cyclic structure, and our efforts were directed to-
ward the detection of the univalent groups which would constitute the ends

* Titration values were 1016, 1344, 1795 and 3417; boiling point values 1070,
1380, 1582 and 3110; theoretical values 982, 1414, 1847 and 3432, respectively.

44 POLYMERIZATION AND RING FORMATION

of the open chains. We have proved the presence of such groups, and the
open-chain structure for the polymeric ethylene succinates is clearly es-
tablished.

Neutral Ethylene Succinate. This ester is prepared by heating suc-
cinic acid with excess glycol at 180 until the distillation of water ceases
and then removing the excess glycol by heating in high vacuum at 200-
250. The ester is purified by crystallization and melts at 102 (5).

Its apparent molecular weight is about 3000.

If this ester is an open chain, its method of formation and its observed
molecular weight would lead one to assign to it the structure III, and it

H [O (CH 2 ) 2 O OC -(CHj)r- CO ] O (CH 2 ) 2 OH (III)
should therefore show the reactions of a dihydric alcohol. The possibility
of applying to it many of the typical reactions for the detection of pri-
mary alcohol groups was excluded for one reason or another. Thus the
presence of ester linkages excluded any reactions which might result to
hydrolysis. The fact that the compound is hygroscopic and insoluble in
ether made it useless to attempt to apply any such methods as the Zere-
witinoff. The compound crystallizes from acetic anhydride and is appar-
ently unchanged even on long boiling in this solvent in the presence of
catalysts, but in so large a molecule acetylation might occur without
any apparent change in physical properties, and the amount of acetic acid
which would be liberated on hydrolysis of even a completely acetylated
product would be so small as to make its estimation difficult. The ester
reacts very slowly and incompletely with phosphorus tribromide.

Six grams of the ester heated for five hours in 30 cc. of boiling chloroform with
6 cc. of phosphorus tribromide after thorough washing and repeated crystallization was
found to contain 0.58% Br.

The ester was also recovered apparently unchanged after treatment
with phenyl and naphthyl isocyanates under various conditions. These
failures appear to be due to a reluctance of the hydroxyl groups to react,
and they are not altogether surprising in view of the diminished reactivity
which is frequently associated with increased molecular size.

Reaction occurred completely in the expected sense with succinic and
with ^-bromobenzoic anhydrides at elevated temperatures. These reac-
tions led to the dibasic acid which agreed in properties and composition
with Formula IV, and to a ^-bromobenzoyl ester which agreed in com-
position with Formula V.

HO OC (CH 2 ) 2 CO [0 (CH 2 ) 2 OC (CH 2 ) 2 CO ] 28 OH (IV)
/>-BrC 6 H 4 CO [ O (CH 2 ) 2 O OC (CH 2 ) 2 CO ] 22 O (CH 2 ) 2 O OC

C 6 H 4 Br-/> (V)

IV. ETHYLENE SUCCINATES 45

Preparation of IV. Twelve grams of neutral ethylene succinate was heated with
3 g. of succinic anhydride at 175-180 for three hours. The reaction mixture was
washed three times with hot water, dried, precipitated from chloroform with ether and
crystallized 6 times from acetone. Its properties identified it as the dibasic acid IV,
which would be formed by the reaction of each hydroxyl group of III with one molecule
of succinic anhydride.

The ester III melted at 102, had an apparent molecular weight of
about 3000, was neutral and did not form a sodium salt. IV prepared
from III melted at 98 and had an apparent molecular weight of 3110
(method of Menzies and Wright in ethylene chloride, observed values
3170, 3040) and a neutral equivalent of 1708. The action of sodium bi-
carbonate on IV led to a sodium salt which melted at 104 and contained
1.23% of sodium. The analytical data are given in Table I.

TABLE I
ACIDIC ETHYLENE SUCCINATES

Sample

M. p.,

Neut.
C. equiv.

Mol. wt. calcd.
from neut.
equiv.

Mol. wt. X =
found struc-
in boiling tural
ethylene units per
chloride molecule

Mol. wt.
calcd.
from
formula

Via

508

1016

1070

982

VIb

672

1344

1380

1414

Vic

898

1795

1582

1847

Anal calcd ,
C

1708 3417

Anal found, %
C H

3110

Sodium sj
Anal
. Calcd.

3432

., sodium
Found

p., C.

48.86 5

.44

48 42 48.

.58

5.57 5.

.29

4.48

4.26

49.22 5

.56

48.84 48

.88

5.74 5,

.62

3.16

3.13

49.38 5

.57

49.34 49

.37

5.86 5,

,81

100

2.43

2.29

49.66 5

.58

49.41 49.

.44

5.70 5.

,69

104

1.32

1.23

Preparation of Di--bromobenzoyl Derivative of III. Two and one-half grains of
III was heated for five hours with 0.75 g. of -bromobenzoic anhydride at 175-185.
The reaction mixture was dissolved in chloroform and precipitated by ether, washed
with hot water, dried and precipitated from acetone by ether. It melted at 93 . Analy-
sis showed it to be the expected di--bromobenzoyl derivative, V, of III.

Anal. Calcd. for V as CmHmO^Bra : C, 49.30; H, 5.20; Br, 4.84. Found: C,
49.47, 49.39; H, 5.59, 5.60; Br, 4.84, 5.01, 4.88, 4.91.

Acidic Ethylene Succinates. If glycol and succinic acid would react
completely in the proportions in which they are brought together, it
would be possible to prepare chains of various lengths by using the glycol
and acid in various ratios. In a number of experiments various excess
amounts of succinic acid were allowed to react with glycol. The reactants
were heated together at 180 until no more water was evolved and the

46 POLYMERIZATION AND RING FORMATION

temperature of the bath was then raised to 200-240 for one to five hours
to make the reaction as complete as possible. Only traces of acid and
glycol appeared in the distillates. The residues were always acidic and,
when the excess of acid was large, always contained some unchanged suc-
cinic acid. Moreover, products of high molecular weight could be iso-
lated from these residues even when the ratio of acid was as high as 2:1.
This showed that under the conditions of these experiments complete con-
trol of the length of the chain produced was not .possible by adjustment of
the ratios of the reactants. On the other hand, this factor had some in-
fluence in controlling the lengths of these chains, and by making use of
such control as it offered, together with fractional crystallization, it was
possible to isolate samples of material representing chains of various
lengths, each of which was quite homogeneous in its physical behavior.
A comparison of the analytical data and molecular weights for these acidic
ethylene succinates clearly indicates that their structure may be repre-
sented by VI, in which x has various average values for different fractions.

HO OC (CH 2 ) 2 CO [ O (CH 2 ) 2 O OC (CH 2 ) 2 CO ]* OH (VI)
Via, x = 6 VIb, x - 9 Vic, x = 12

Preparation of VI, a, b and c. Ethylene glycol, 41 g., and succinic acid, 93.5 g.
(20% excess), were heated at 200-210 for four hours, and the residual water was then
removed as completely as possible by heating in vacua. The residue was dissolved in
chloroform and precipitated with ether. The powder which resulted from the drying
of this precipitate weighed 92.2 g. and melted at 87.5-90. It was extracted several
times with boiling water. The residue (59 g.) was dissolved in chloroform and pre-
cipitated by ether. It melted at 90. This constituted fraction Vic. The hot aqueous
extracts on cooling deposited 11 g. of solid which after drying melted at 73-74 .

In another similar experiment the acid and glycol were heated for two hours at
190-210 under ordinary pressure, and then immediately extracted several times with
boiling water (3 X 300 cc.). The residue amounted to 10.5 g. After solution in
chloroform and precipitation by ether, it melted at 82-83. This constituted fraction
VIb. On cooling, the aqueous extract deposited considerable white solid. This after
precipitation from chloroform by ether melted at 73. This material together with
the fraction from the previous experiment melting at the same temperature constituted
fraction Via. By evaporation of the aqueous extracts from this experiment there was
obtained another fraction melting at 86 . The analytical data for this fraction indicated
that it lay between the 73 and the 82 fractions.

Fraction IV. was prepared from III in the manner already indicated.

The analytical data for these fractions (see Table I) indicate that we
are here dealing with a polymeric series in which the average values for the
degree of polymerization vary from 6 up to 23. These fractions closely
resemble each other in physical properties: they have the same appear-
ance under the microscope, and they show similar solubility relations. All

IV. ETHYLENE SUCCINATEvS 47

the fractions are soluble in cold chloroform and in hot 50% alcohol. The
lowest fraction is quite soluble in hot water and the higher fractions only
very slightly soluble. The melting points rise with increasing degree of
polymerization.

All these fractions form sodium salts on treatment with sodium bicar-
bonate, and these resemble very closely in properties the acids from which
they are derived. Thus they are readily soluble in cold chloroform and in
warm acetone. They are only slightly soluble in cold water. All of them,
however, are readily soluble in warm water, and in this respect the higher
members are sharply differentiated from the corresponding acids. The
magnitude of the change in properties which is produced by the transfor-
mation of the slightly polar carboxyl group into the completely polar
sodium salt will be expected to vary with the size of the chains to which
these groups are attached, but we found it somewhat surprising, never-
theless, that these sodium salts should melt only a few degrees above the
melting points of the corresponding acids. This difference diminishes con-
tinuously with the increase in length of the chain and in the highest poly-
mer, IV, amounts to only 6.

The general plan of the structure of these acid esters is clearly estab-
lished by the formation of the sodium salts. Attempts to prepare other
derivatives from acids met with little success. Treatment with thionyl
chloride and with phosphorus pentachloride under various conditions
furnished products which contained chlorine, but it could not be estab-
lished that these materials were really acid chlorides. Attempts to pre-
pare amides by heating the sodium salts of the acids with ^-toluidine hy-
drochloride led to the formation of N-/>-tolyl succinimide. Heating the
sodium salts with />-bromophenacyl bromide led to the formation of de-
rivatives, but these contained less than the calculated amount of bromine.
Thus the sodium salt of Via led to a derivative, m. p. 73, containing
8.66% of bromine, while the calculated value for a di-(^-bromophenacyl)
ester of Via is 11.61% Br. Similarly, the sodium salt of Vic led to a p-
bromophenacyl derivative containing 4.12 instead of the calculated 7.13%
of bromine.

Mechanism of the Formation of Ethylene Succinate. The formation
of an open-chain polyester of the type exemplified by II might occur in
either one of two ways. (1) The cyclic monomeric ester VII might first
be formed, and this might then undergo A polymerization (self-addition)
with the formation of II. (2) II might be formed directly from the acid
and the glycol (C polymerization) by a series of successive reactions.
That polyesters of this general type may be formed by the first mecha-

48 POLYMERIZATION AND RING FORMATION

nism is clearly indicated by the fact (6) that the monomeric form of tri-

methylene carbonate can be isolated by the distillation of its polymer,

and, by heating the monomer can be changed to the poly-

vCO Ox meric form. This is also true of other polyesters, but, so

CH 2 CHi far as our information extends now, only of those in which

CH CH *ke monomer is a 6-membered ring. The 5-ring esters

\CO O/ ^ no ^ Ply mer i ze ; the polyesters whose monomers would
Vii be larger-than-6-rings are not depolymerized; and, al-

though monomers of this type are known (7) and are
stable, none of them has ever been prepared by a bifunctional reaction.
The general theory underlying these facts has already been discussed (1, 6).

The evidence that ethylerie succinate is formed by Mechanism 2 (C
polymerization) rather than 1 is fairly conclusive. Experimental evidence
in support of Mechanism 2 may be found in the fact that the molecules
of ethylene succinate are open chains with functional groups at the ends.
The supposition that the monomeric ring VII is first formed involves the
tacit assumption that 8-rings are readily formed, but are unstable, whereas
fact and theory alike indicate that such rings are formed only with great
difficulty, but are stable when they have been formed. We have made a
good many attempts to isolate the monomeric 8-ring, VII. All of these
attempts have failed, and all of the observations which we have made are
best interpreted on the assumption that the long chains are built up by a
series of successive reactions and that rings are not formed at any stage of
the process. Some light is thrown at one or two points on the details of the
mechanism of the reaction by the following observations.

When glycol was allowed to react with succinic acid in the proportions
of 1 to 2 moles, the product was found to be a mixture composed chiefly
of unchanged succinic acid and polyester of fairly high molecular weight.
The compound, HO OC (CH 2 ) 2 CO O (CH 2 ) 2 O OC (CH 2 ) 2
CO OH, which might be expected to form under these stoichiometrical
conditions was not found.

Neutral ethylene succinate was prepared by heating the acid with excess
glycol and, after some time, distilling off the excess glycol as completely
as possible at high temperature and under greatly reduced pressure. Under
these conditions the formation of some di-/3-hydroxy ethyl succinate, VIII,
might be expected. This ester was in fact formed, and it distilled

HO (CH 2 ) 2 O OC~(CH 2 ) 2 CO O (CH 2 )2~OH (VIII)

out of the reaction mixture when this was heated in vacua. It was obtained

IV. ETHYLENE SUCCINATES 49

in a state of approximate purity by redistillation in a carefully cleaned all-
glass apparatus under high vacuum.

Di-(/3-hydroxyethyl)-succinate. vSixty-two grams of ethylene glycol (1 mole) and
39.3 g. of succinic acid ( l / 9 mole) in a pyrex Claisen flask fitted with ground-glass stop-
pers and provided with a receiver were heated for six hours by means of an oil-bath at
174-180; 8.8 g. of water, b. p. 100-102, w 2 D 8 1.3323, collected in the receiver. The
residue was heated in vacuo finally to a temperature of 250 at 0.015 mm., until dis-
tillation ceased. The residue was a pale yellow viscous liquid which readily solidified
on cooling. It was dissolved in 100 cc. of warm chloroform and precipitated as a
powder by the addition of 250 cc. of benzene. After drying, this powder weighed 36 g.
(0.25 g. equivalent). Its instantaneous melting point was 98 and its slow melting
point 105.5 . The distillates were redistilled and yielded 9.97 g. or 0.648 mole of water,
33.9 g. or 0.548 rnole of glycol and 14.5 g. or 0.07 mole of di-(/3-hydroxyethyl) succinate.
The total esters (0.7 mole + 0.25 g. equivalent) accounted for 0.32 mole of the 0.333
mole of succinic acid used, and these esters plus the glycol recovered accounted for
94.6% of the glycol used, while the water isolated was 87.5% of the calculated amount.

The di-/3-hydroxyethyl succinate was never isolated in a state of purity. When
distilled in a carefully cleaned all-glass (pyrex) Claisen flask, it boiled at 176-180 at
0.001 mm. as a colorless viscous liquid, leaving only a trace of residue.

Anal. Calcd. for CuHi 4 O 6 : C, 46.58; H, 6.88. Found: C, 45.76, 45.70; H,
6.86, 6.95.

Treatment with ^-nitrobenzoyl chloride in pyridine yielded a di-(/>-nitrobenzoate)
as white needles from alcohol; m. p. 90-91 .

Anal. Calcd. for C 2 2H2oOi2N 2 : C, 52.35; H, 3.97. Found: C, 52.36, 52.36;
H, 4.19, 4.14.

The di-phenylurethan formed by the action of phenylisocyanateondi-(/3-hydroxy-
ethyl) succinate and crystallized from a mixture of benzene and petroleum ether melted
at 113.

Anal. Calcd. for C 2 2H 2 4O 8 N 2 : C, 59.45; H, 5.41. Found: C, 59.40, 59.56; H,
5.53, 5.52.

When heated slowly in vacuo di-/3-hydroxyethyl succinate was com-
pletely converted into neutral (polymeric) ethylene succinate and ethylene
glycol, the former appearing in the distillate and the latter remaining in
the distilling flask. This process evidently involves an ester interchange:
glycol is eliminated between two molecules of the dihydroxy ester with
the formation of a new dihydroxy ester containing two of the structural
units of ethylene succinate. Repetition of this process finally results in
such a molecule as III. It is evident that so long as one of the products
of a possible ester interchange can be eliminated in this way, no merely
stoichiometrical factors can set up a limit to the length of the molecules
which might be produced. Nevertheless a fairly definite limit exists:
polyintermolecular esterification has never led us to molecules of greater
average length than about 200 atoms. No doubt various factors are in-

50 POLYMERIZATION AND RING FORMATION

volved in this point. There can be no question that the reactivity of func-
tional groups diminishes with the size of the molecules which contain them.
It is apparently this factor which accounts for the failure of the ester III to
react with many of the typical reagents for hydroxyl groups. Moreover,
as Staudinger has frequently emphasized (8), the thermal stability of mole-
cules must diminish with increase in their size. These two factors act in
opposition. To force the completion of the reaction between succinic acid
and glycol we have used high temperature (250) to increase reaction
velocity and high vacuum to remove the water as completely as possible.
The ester produced under these conditions has a molecular weight of about
3000. It is interesting to observe that in the thermal polymerization of
styrene the product formed at 250 also has an apparent molecular weight
of about 3000, while the polystyrenes formed at lower temperatures have
much higher molecular weights. And so also the A polymerization of tri-
methylene carbonate described in the previous paper (6), although it also
involves an ester interchange, leads to a polyester with a molecular weight
considerably above 3000, since the reaction consists merely in self-addition
and proceeds rapidly at 100.

Dimeric Ethylene Succinate. Tilitschejew (9) has reported the isola-
tion of a new ethylene succinate from the volatile materials formed by
heating the product of the action of succinic acid on glycol (m. p. 88-89)
to 340-390 under 3-4 mm. pressure. The new product was distinguished
from the usual form not only by its different melting point (129-130), but
by its definite macrocrystallinity. Cryoscopic data on acetic acid solu-
tions of the new ethylene succinate indicated a double formula and, since
Vorlander had already reported (10) that the old form was dimeric, Tilit-
schejew regarded his new ester as an isomer of the old. It has now been
established that the usual form of ethylene succinate is not dimeric, but
much more highly polymeric; the new compound cannot, therefore, be
simply an isomer of the old. Aside from this, however, repetition of Tilit-
schejew's experiments has completely verified the correctness of his claims
as to the nature of the new compound. It crystallizes in thin plates melt-
ing sharply at 130. The analytical data clearly indicate that it is dimeric
ethylene succinate.

Anal. Calcd. for (CeH 8 O 4 ) 2 : C, 50.00; H, 5.50; mol. wt., 288; saponification
equivalent, 72. Found: C, 49.85, 50.00; H, 5.60, 5.57; mol. wt., in boiling ethylene
chloride, 302, 299; in freezing benzene 279; saponification equivalent, 71.97, 71.95,
71.66.

Attempts to partially saponify this ester were unsuccessful a part of
the ester was recovered unchanged and the remainder was degraded to

IV. ETHYLENE SUCCINATES 51

sodium succinate and glycol. Nevertheless, it seems fairly certain that
this compound is the 16-membered ring, IX. No alternative formula seems

(CH 2 ) 2 ~0OC (CH 2 ) 2 CO O (CH 2 ) 2 O OC (CH 2 ) 2 COv

> (IX)
(CH 2 ) 2 CO O (CH 2 ) 2 O OC (CH 2 ) 2 CO O (CH 2 ) 2 (X

plausible. It is very improbable that this compound is present as such
in the polymeric ester from which it is prepared. The dimer is quite
soluble in hot absolute alcohol; the neutral polymer is quite insoluble.
Continuous extraction of the polymer with hot absolute alcohol resulted
in the solution of only a very small amount of material and this had the
properties of the polymer not of the dimer. The dimer must therefore
be formed during the process of thermal decomposition. The trans-
formation of polymeric trimethylene carbonate into mononieric ethylene
carbonate described in the previous paper (6) proceeds smoothly and prac-
tically quantitatively at about 200. The formation of the dimeric
tetramethylene carbonate, described in the same paper, and of dimeric
ethylene succinate occurs only at a much higher temperature. Large
amounts of gaseous and liquid products are formed as well as considerable
carbonaceous residue. The yields of the dimers are quite small, e. .,
3-4% of the theoretical under the best conditions (Tilitschejew, however,
reports 5.5%). In these respects the reaction resembles that used by
Ruzicka for the preparation of the large cyclic ketones (11). As yet we
have not succeeded in isolating any monomeric ethylene succinate from
the products of this reaction.

Ethylene Succinate and the Association Theory of High Polymers.
The discussion in this and the previous papers will, it is hoped, have
amply demonstrated the adequacy of the ordinary structural theory of
organic chemistry to deal with the polyesters, a fairly complicated class
of high polymers.

A possible explanation of the structure of the polyesters which has
not yet been considered is the following. The chemical unit or molecule
of ethylene succinate is the monomeric 8-ring, VII. Because of the great
strains in this structure, or for some other reason, it exhibits exceptionally
strong residual or lattice forces, so that the osmotic unit becomes an
aggregate of a great many of these ultimate chemical units. There is
such a complete lack of any general theoretical justification for this view
that there would be no need to consider it in connection with the poly-
esters were it not that in the field of natural polymers such as cellulose,
rubber, etc., in various slightly differing forms it has been defended at
great length by so many investigators (12).

52 POLYMERIZATION AND RING FORMATION

Among the many facts which are quite incompatible with any association
theory of the structure of polyesters, the following deserve special mention.

(1) All of the known cyclic esters containing larger-than-6-rings are
stable substances with definite properties corresponding with their simple
formulas. They do not show any tendency to associate in solution.

(2) The polymeric esters corresponding to these monomers show no
tendency to dissociate.

(3) In the polymeric ethylene succinates the chemical and osmotic
units have been shown to be identical within the limits of experimental
error, i. e., ebullioscopic data give the same values for the molecular weights
that are given by the determination of hydroxyl or carboxyl.

The association theory is not clearly enough defined to permit any
more crucial tests than these, but the following experiment at any rate
agrees with these in the conclusions to which it leads.

If two different association polymers (A)n and (B)m are mixed in
solution or in the liquid state so that they constitute a single phase, then
since (A)n < > nA and (B)m < > mB, the resulting mixture should be com-
posed at least in part of (AB) P .

Ethylene sebacate (m. p. 78) and ethylene succinate (m. p. 103)
were melted together and thoroughly mixed. After cooling, the mass
was extracted with benzene, which is a solvent for the sebacate but not
for the succinate. The residue melted at 103 (unchanged succinate).
To the benzene solution petroleum ether was added; the precipitated
solid melted at 78-79 (unchanged sebacate). An ester was prepared
by the action of glycol on equivalent amounts of succinic and sebacic
acids. This was quite different in its properties from either the succinate
or sebacate and it was homogeneous in its solubility behavior (m. p.
38-40; mol. wt. 1540). This demonstrates the existence of a mixed
polymer and the absence of any reversible relationship of association be-
tween it and the two corresponding simple polymers.

In the 6-ring esters a reversible relationship exists between the mono-
meric and polymeric forms, and some of the statements made above do
not apply to these polymers. But it is possible, nevertheless, to account
satisfactorily for the behavior of these esters without the assumption of
any special or peculiar kinds of valence (6) .

The view that the ordinary structural theory of organic chemistry
is adequate to deal with high polymers has been now for several years
ably defended by Staudinger and his collaborators (13), and recently
the same view has been applied in a brilliant fashion to such natural poly-
mers as cellulose, rubber, silk fibroin, etc., by Meyer and Mark (14). So

IV. ETHYLENE SUCCINATEvS 53

far as the minor differences (15) in the views of these two groups of investi-
gators are concerned, our own experiments on polyesters incline us to favor
those of Staudinger. That is, we can find no real objection to referring to
primary valence chains as molecules, and among the polyesters these mole-
cules are experimentally identical with the osmotic units of their solutions.

The writers are indebted to Mr. W. H. Taylor for the molecular weight
determinations and to Mr. G. A. Jones for the determinations of carbon
and hydrogen.

Summary

The polymeric ethylene succinate previously described is shown to be
a long chain made up of the recurring unit O (CH 2 )2 O CO
(CH 2 ) 2 CO and bearing hydroxyl groups at its ends. Acidic ethylene
succinates made up of similar chains of various lengths and bearing car-
boxyl groups at their ends have been prepared. Molecular weight deter-
minations of these esters based on ebullioscopic measurements agree with
those based on chemical evidence (estimation of hydroxyl or carboxyl).

A study of the ethylene succinate prepared by Tilitschejew verifies
his claim that it is dimeric, and this ester is undoubtedly a 16-membered
ring.

Bibliography and Remarks

(1) Carothers, Journ. Am. Chem. Soc., 51, 2548 (1929).

(2) Vorlander, Ann., 280, 167 (1894).

(3) Carothers and Arvin, Journ. Am. Chem. Soc., 51, 2560 (1929).

(4) Mohr, J. prakt. Chem., 98, 348 (1918).

(5) Capillary tube melting point. See Carothers and Arvin, Journ. Am. Chem. Soc., 51, 2560
(1929), for a description of the peculiar melting point behavior of this compound.

(6) Carothers and van Natta, J. Am. Chem. Soc., 52, 314 (1930).

(7) Ruzicka and Stoll, Helv. Chim. Acta, 11, 1159 (1928).

(8) See Staudinger, Ber., 59, 3019 (1926).

(9) Tilitschejew, J. Russ. Phys.-Chem. Soc., 57, 143-150 (1925); Chem. Zentr., I, 2667 (1926).

(10) Vorlander, Ann., 280, 167 (1894).

(11) Ruzicka and co-workers, Helv. Chim. Acta, 9, 230, 249, 339, 389, 399, 499, 715, 1008 (1926);
10, 695 (1927); 11, 496, 670, 686, 1159, 1174 (1928).

(12) See for example Bergman, Knehe and Lippmann, Ann., 458, 93 (1927); Hess, Trogus and
Friese, ibid., 466, 80 (1928); Schlubach and Eisner, Ber., 61, 2358 (1928); Pummerer, Nielsen and
Gundel, ibid., 60, 2167 (1927).

(13) Staudinger and co-workers, Helv. Chim. Acta, 5, 785 (1922); 7, 23, 842 (1924); 8, 41, 65, 67
(1925); 9,529(1926); 11,1047,1052(1928); Ber., 53, 1073 (1920); 57,1203(1924); 60,1782(1927);
59,3019(1926); 61,2427(1928); 62,241,263,442(1929); Ann., 447, 97, 110 (1926); 467,73(1928);
Z. physik. Chem., 126, 425 (1927); Kautschuk, 237 (1927); Z. angew. Chem., 42, 37, 67 (1929); Z.
Krist., 70, 193 (1929).

(14) Meyer and Mark, Ber., 61, 593, 1932, 1939 (1928); Meyer, Naturwissenschaften, 16, 790
(1928); Z. angew. Chem., 41, 935 (1928).

(15) Meyer, Naturwissenschaften, 17, 255 (1929).

54 POLYMERIZATION AND RING FORMATION

V. Glycol Esters of Oxalic Acid*

This is a study of the glycol esters of oxalic acid.

Monomeric ethylene oxalate is a macrocrystalline substance with a melting
point of 144; it undergoes polymerization under various conditions and
yields microcrystalline, white, dusty polymers with melting points up to
172. This material is insoluble and no molecular weight determination
was possible. An intermediate polymer showed a melting point at 158-
159; it was called B -polymer, is soluble and gave in boiling acetonitrile a
molecular weight around 2400. The polymers can be depolymerized and
fractionated by extraction with acetonitrile.

Propyleneoxalate was obtained as monomeric crystalline material with a
melting point of 142 and as polymeric resinous substance with an ap-
parent molecular weight of 700.\

Trimethylene oxalate is obtained as dimeric and polymeric material with
molecular weights of 270 and about 2000, respectively.
Ilexamethylene oxalate is a white powder with am. p. of 79 and a molecu-
lar weight around 1150.

Melting points ranging from 110 to 172 have been ascribed to ethylene
oxalate (1). Bischoff (2), by distillation, prepared a form melting
at 142 to 143, which he showed to be monomeric. He observed that the
melting points of this and higher-melting forms change spontaneously 011
standing, and he ascribed this change to reversible polymerization but
without the support of any comparative molecular weight data. In
connection with a study of glycol esters of dibasic acids (3) we have made
some further observations on ethylene oxalate and have also prepared some
other alkylene oxalates.

Preparation of Ethylene Oxalate. The ester was prepared by heating
ethylene glycol and ethyl oxalate in a Claisen flask provided with a re-
ceiver. Alcohol distilled off fairly rapidly when the heating bath was kept
at 180-190. The residue was heated in a vacuum for a time to remove
unchanged reactants. The distillate was found to contain some ethyl
(/3-hydroxyethyl) oxalate.

Ethyl (0-Hydroxyethyl) Oxalate. Color less liquid; b. p. (0.2 mm.), 108-110;
<%> 1.2241; n 2 D 1.4405.

Anal. Calcd. for CeHnA,: C, 44.42; H, 6.22. Found: C, 44.01; H, 6.02.

* W. H. Carothers, J. A. Arvin and G. L. Borough; Journ. Am. Chem. Soc., 52,
3292-3300 (1930); Contribution No. 35 from the Experimental Station of E. I. du Pont
de Nemours and Co.

Received April 23, 1930. Published August 5, 1930.

t Corresponding to a polymerization degree of about 7 to 8.

V. GLYCOL ESTERS OF OXALIC ACID

The residue after crystallization from glacial acetic acid or from ethyl
oxalate was a dusty white powder that usually melted at about 153.
This material was polymeric.

Anal. Calcd.forC 4 H 4 O 4 : C, 41.38; H,3.45; mol. wt., 116; saponification equiva-
lent, 58. Calcd. for CwH^Ow = C 2 H 5 O (COCO O(CH 2 ) 2 O)i 3 H: C, 41.70;
H, 3.73; mol. wt., 1554; saponification equivalent, 59.4. Found: C, 41.66, 4197;
H, 3.76, 4.05; mol. wt. (in boiling
acetonitrile), 1510, 1580, 1610;
saponification equivalent, 59.0, 60.8,
61.0.

When this polymer was
heated in a vacuum, distilla-
tion occurred and from the
distillate Bischoff's 143 ester
was obtained (observed m. p.,
143-144) . Molecular weight
determinations in boiling aceto-
nitrile agreed with those in
freezing acetic acid reported
by Bischoff, indicating that
this material is monomeric.

Anal Calcd. for C 4 H 4 O 4 : C,
41.38; H, 3.45; mol. wt., 116.
Found: C, 41.42, 41.44; H, 3.56,
3.45; mol. wt. (in boiling acetoni-
trile), 118, 120, 123, 126.

Small samples of polymer,
if distilled rapidly, gave better
than 50% yields of the mono-
mer. Under other conditions
the yield was smaller owing
to thermolysis (Id) to ethyl-
ene, carbon dioxide, ethylene

Fig. 1. Crystals of monomeric ethylene
oxalate. X 94. (These crystals were placed
for observation in a microscopic culture plate
and sealed against the access of air and moisture
by means of a cover glass. Mechanical dis-
turbances were avoided, and subsequent ob-
servations, except the last, were made from
exactly the same position as the first.)

carbonate, carbon monoxide

and other gaseous, liquid and tarry products.

Properties of Ethylene Oxalate and its Polymers. Monomeric
ethylene oxalate is definitely macrocrystalline and shows a relatively high
solubility (Table II). It is so readily hydrolyzed that it may be titrated
directly with warm tenth normal alkali. On standing at room tempera-
ture, the sharply defined crystals of the monomer rapidly disintegrate

POLYMERIZATION AND RING FORMATION

(Figs. 1-3) with the appearance of having been violently disrupted, and
finally are transformed into very minute crystals, which, if undisturbed,
may become spontaneously oriented on a glass surface to thread-like aggre-
gates (Fig. 4).

This transformation is due to polymerization. It is accelerated by
moderate heat and catalyzed by acids or alkalies. The polymerization is
accompanied by changes in melting point and solubility. In the first
stages of the polymerization the melting point drops (e. g., to 106-110);
higher-melting polymers are formed by heating the monomer in the pres-
ence or absence of solvents or
catalysts, and the final prod-
uct is an insoluble material
melting at 172-173.

A sample of monomeric
ethylene oxalate was heated
at 135-140 for seven hours,
and then extracted repeatedly
with cold acetone. The ma-
terial that separated on evapo-
ration of the acetone was iden-
tified as unchanged monomer.
The acetone-insoluble mate-
rial was a dusty white powder
which after drying melted at
171-172. It was insoluble in
all common organic solvents,
so no molecular weight deter-
minations could be made.

Fig. 2. The crystals of Fig. 1 after five days
at room temperature. X 94.

Anal. Calcd. for (C 4 H 4 O 4 )x:
C, 41.38; H, 3.45; saponification
equivalent, 58.0. Found: C, 41.48,
41.59; H, 3.71, 3.96; saponification equivalent, 58.45.

By fractional crystallization of polymerized monomer and of the 153
polymer formed directly from glycol and oxalic ester, a great variety of
samples was obtained showing such melting points as 106-108, 122-125,
140-142, 148-150, 153, 155-157, 157-159, 160-163, 163-164, 172-
173, as well as others having much wider melting ranges. Attempts to
segregate homogeneous samples from these fractions met with difficulties
due to the rapidity with which spontaneous transformations occurred even
in the absence of solvents and at room temperature. Portions of three

V. GLYCOL ESTERS OF OXALIC ACID

sharply melting samples were stored in glass-stoppered bottles, and melting
point determinations were made at intervals with the following results.

TABLE I
MELTING POINTS

Nature and m. p.
of sample

144 (monomer)
153 (polymer)
172 (polymer)

* M. p. after standing at room temperature
2 months 4 months 5 l /2 months

105-140 105-135 105-122 (A)
149-155 149-155 149-155 (B)
160-170 160-170 160-170 (C)

To avoid these changes during the process of fractionation, it was
necessary to work rapidly and
to use cold solvents as far as
possible. Solubility data estab-
lished the homogeneity of the
monomer (144) and the ap-
parent homogeneity of the
highest (172) polymer. Inter-
mediate fractions, unlike the
172 polymer, were completely
soluble in warm acetonitrile,
and since the solubility of the
latter was not affected by the
presence of monomer, inter-
mediate polymers must have
been present in these fractions.
This led to the hope that it
might be possible to isolate
polymers sufficiently low for a
study of structure.

The fractions A, B and C of Fig. 3. The crystals of Fig. 1 after seven days
Table I were recrystallized with at room temperature. X 113.

the following results: extrac-
tion of A with cold acetone removed a small amount of monomer. The
residue was treated with warm acetonitrile, which dissolved all but a
small amount of material that melted at 170-172 (high polymer). The
material that separated from the acetonitrile solution, after recrystalli-
zation, melted at 157-159. Similarly from B, fractions were isolated
corresponding with monomer, high (172) polymer, and 157-159 polymer.
C yielded a small amount of material melting at 155-157. The rest
was unchanged 172 polymer. Extractions of various other samples
were made, and in one case a small fraction melting at 106-108 and hav-

POLYMERIZATION AND RING FORMATION

ing an apparent molecular weight of about 900 was isolated. There is still
considerable doubt, however, concerning the homogeneity of this fraction.
The 153 material, which was the usual (recrystallized) product of the
action of ethylene glycol on ethyl oxalate, was not homogeneous in spite of
its fairly sharp melting point, for its apparent solubility changed with
changing ratio of solute to solvent as shown below.

Sample, g.
0.025
.100
.500

Acetonitrile, cc.
15
15
15

Apparent soly. in
g. per 100 g. at 25

0.0821

.2633
.8644

A four-gram sample of the 153 material was then extracted repeatedly
with 25 cc. portions of acetonitrile. With each extraction the apparent
solubility decreased until a constant value was reached.

Extraction
number

1
2
3
4
5
6

7
8
9

Apparent soly. in
g. per 100 g. at 25

1.4321
0.8111

.3240

.2390

.1729

.1315

.0894

.0891

.0890

M. p. of residue, '

152-153
152-155
155-159
155-160
157-160
157-159
157-159
158-159
158-159

The apparent solubility of the residue did not change on increasing the
ratio of solute to solvent five-fold, so this material must be regarded as
essentially homogeneous. Its apparent molecular weight in boiling aceto-
nitrile was about 2380 (observed values, 2070, 2480, 2670, 2370, 2275, 2520).

One may conclude that at least two polymers of ethylene oxalate exist:
a soluble form melting at 158-159 of molecular weight about 2400, and an

TABLE II
SOLUBILITIES OF ALKYLENE OXALATES

Nature of oxalate sample

Monomeric ethylene
Polymeric ethylene
Polymeric ethylene
Monomeric propylene
Monomeric ethylene

heated at 90 for

two weeks

These values have no quantitative significance since the sample was not homo-
geneous.

M. p., C.

144
159

mol. wt.
118-126
2070-2670

Acetonitrile

11.29
0891

Acetone
4.13

Chloroform
0.35

172

<0 01

<0.01

142
148-150

131-147
1620-1640

12.31
0.1823

3.47
0.1630

0.06
0.0390

V. GLYCOL ESTERvS OF OXALIC ACID

insoluble form melting at 172 of unknown but probably much higher mo-
lecular weight. Either of these forms may arise spontaneously from the
other, and both of them may be formed from the monomer. Definite
evidence for the existence of other polymers of ethylene oxalate is lacking.
Chemical Properties of the Polymeric Ethylene Oxalates. Mono-
meric ethylene oxalate is hydrolyzed with extraordinary rapidity. To a
certain extent this property is shared by its polymers. Hence, although
the polymers show acid reactions toward litmus in contact with water,
this fact cannot be used to
argue for the presence of long
primary valence chains bearing
carboxyl groups at the end.
Attempts to prepare sodium
salts from the polymers by
the action of cold sodium bi-
carbonate solution led to the
isolation of sodium oxalate and
unchanged polymer. This ease
of hydrolysis is associated with
great sensitivity toward other
reagents. The attempt to de-
tect hydroxyl groups or car-
boxyl groups by heating poly-
meric ethylene oxalate with
m-bromobenzoic anhydride
and with phenylhydrazine led
to the isolation of ethylene-
fo's-w-bromobenzoate and to
the phenylhydrazide of oxalic
acid. The latter reaction is
what would be expected from
an ester having the structure
O (CH 2 )2 O , etc., whether the chain is open or closed, but the first
reaction requires the elimination of CO CO O residues. These
fragments of the molecules appeared as carbon monoxide and carbon diox-
ide, which were observed to be evolved from the reaction mixture.

Ethylene &-w-Broniobenzoate. Prepared by the action of w-bromobenzoic
anhydride on ethylene glycol or on ethylene oxalate; crystallized from a mixture of
chloroform and alcohol; m. p. 78-79.

Anal. Calcd. for QeH^B^: C, 44.86; H, 2.80; Br, 37.38; mol. wt., 428. Found.

Fig. 4. The crystals of Fig. 1 after two weeks
at room temperature. X 185.

60 POLYMERIZATION AND RING FORMATION

C, 44.85, 45.06; H, 2.92, 2.98; Br, 37.26, 37.03; mol. wt. (in boiling benzene),
430.

Propylene Oxalate. By heating propylene glycol and ethyl oxalate
together and removing unchanged reactants in a vacuum, a colored viscous
resin was obtained. In boiling acetonitrile this showed an apparent mo-
lecular weight of about 700 (observed, 670, 660). No crystalline material
could be isolated from it. When this resin was strongly heated at a pres-
sure of 5 mm., distillation occurred. A crystalline solid that separated
from the liquid distillate melted at 142 after recrystallization from hot
alcohol. This was identified as monomeric propylene oxalate.

Anal. Calcd. for CsHeOi: oxalic acid, 69.24; saponification equivalent, 65;
mol. wt., 130. Found: oxalic acid, 69.82; saponification equivalent, 64.5, 65.1; mol.
wt. (in boiling acetonitrile), 131, 147.

Carbon and hydrogen values (Pregl method) were consistently low, perhaps owing
to the loss of methane.

The monomeric methyl ethylene oxalate showed no tendency to poly-
merize spontaneously at room temperature, but on being heated to 140-
150 for eight hours it was converted to a white powder that melted at 176-
178 and was insoluble in all common organic solvents.

Trirnethylene Oxalate. This ester has been prepared by Tilitcheev (4)
as a solid melting at 82-84 by heating methyl oxalate with trimethylene
glycol first at atmospheric and then under diminished pressure, dissolving
the residue in chloroform, and precipitating it with methyl alcohol. On
distillation at 3-4 mm. pressure it was converted into an "isomeric" form
melting at 186-187.

Since the length of the chain of the structural unit of trimethylene oxalate
is seven atoms, the reaction between oxalic ester and trimethylene glycol
should be intermolecular and the product polymeric, in accordance with
the generalization based on the study of other similar reactions (3, 5).

We prepared trimethylene oxalate by heating ethyl oxalate with tri-
methylene glycol. After three crystallizations from a mixture of chloro-
form and ethyl alcohol it melted at 87-88. Its apparent molecular weight
in boiling acetonitrile was about 2000 (observed value, 2040, 1980).

Anal. Calcd. for C 6 H 6 O4: C, 46.15; H, 4.65. Found: C, 46.03, 46.34; H, 4.90,
4.95.

When this polymeric material was heated at 250 at 3-4 mm., thermoly-
sis and distillation occurred. Gaseous, liquid and carbonaceous products
were formed. The liquid distillate from 52 g. of polymer weighed 23.6 g.,
and on being cooled and treated with alcohol it yielded a small amount of

V. GLYCOL ESTERS OF OXALIC ACID 61

crystalline solid, which after repeated crystallization melted at 186-187.
This was Tilitcheev's "isomeric" trimethylene oxalate.

Anal. Calcd. for (CsHsO^: C, 4.165; H, 4.65; raol. wt., 260. Found: C.
45.75,46.12; H, 4.61,4.69; mol. wt. in boiling acetonitrile, 282, 278; in freezing phenol,
272, 265.

The molecular weight determinations prove this material to be dimeric.

The liquid distillate from which the dimeric trimethylene oxalate was
isolated was redistilled. It boiled from 70 at 20 mm. to 185 at 0.25 mm.
From the higher-boiling fractions a considerable amount of trimethylene
carbonate was isolated. This is a product that Tilitcheev assumed to be
intermediate in the thermolysis of trimethylene oxalate, but he was unable
to isolate it. Saponification of the remaining liquid led to the isolation
(as calcium oxalate) of about 1.5 g. of oxalic acid. Hence it is possible
that some monomeric trimethylene oxalate may have been present in the
mixture.

These results lead to the following conclusions. The action of trimethyl-
ene glycol on ethyl oxalate proceeds intermolecularly and leads to an ester
of the type CO CO O (CH 2 ) 3 O CO CO O (CH 2 ) 3 O
CO CO O (CH 2 ) 3 O , etc. This on being heated to a high tempera-
ture undergoes thermal decomposition and yields a complicated mixture of
products containing a small amount of the 14-membered cyclic ester

CO CO O (CH 2 ) 3 O CO CO O (CH 2 ) 3 O -,

These reactions are analogous to those observed in the formation and
decomposition of ethylene succinate (6) and tetramethylene carbonate (7).
No truly reversible relationship between a monomeric and polymeric form
of trimethylene oxalate exists, as it does in ethylene oxalate and in tri-
methylene carbonate (7). The dimeric trimethylene oxalate shows no
tendency to polymerize spontaneously.

Hexamethylene Oxalate and Decamethylene Oxalate. From con-
siderations which have already been set forth it is to be expected that
these esters, by whatever method they are prepared, will be linear con-
densation polymers. They are readily prepared by heating ethyl oxalate
with the corresponding glycols at first under atmospheric pressure and
finally in high vacuum.

Hexamethylene Oxalate. White powder purified by precipitation from chloro-
form by methyl alcohol, m. p. 66.

Anal. Calcd. for C 8 Hi 2 O 4 : C, 55.79; H, 7.03; mol. wt., 172. Found: C, 55.77,
55.58) H, 7.17, 7.08; mol. wt. (in boiling benzene), 1050, 1160, 1120.

62 POLYMERIZATION AND RING FORMATION

Became thylene Oxalate. White powder purified by precipitation from chloro-
form by methyl alcohol, m. p. 79.

Anal. Calcd. for Ci 2 H 2 4O 4 : C, 62.57; H, 9.63; mol. wt. f 232. Found: C, 62.89,
62.73; H, 8.97, 8.90; mol. wt. (in boiling benzene), 1160, 1190.

Summary

Ethylene oxalate exists in three mutually interconvertible forms: a
monomer (m. p. 144), a soluble polymer (m. p. 159) and an insoluble
polymer (m. p. 172). Ethylene oxalates showing other melting points
are probably mixtures of these three forms, since evidence for the existence
of any other individual forms is lacking. None of these forms are stable at
ordinary temperature. The monomer polymerizes spontaneously, and the
purified polymers are partially depolymerized.

Propylene oxalate exists in at least two mutually interconvertible forms :
a monomer and a polymer. Monomeric propylene oxalate polymerizes
much less rapidly than ethylene oxalate.

Trimethylene oxalate (m. p. 86) prepared from ethyl oxalate and tri-
methylene glycol is a linear condensation polymer. It shows no tendency
to depolymerize spontaneously. At high temperature it undergoes ther-
mal decomposition, and one of the products of this reaction is the dimeric
14-membered heterocycle, m. p. 187. This is stable and shows no tend-
ency to polymerize further.

Hexamethylene oxalate and decamethylene oxalate prepared by the
action of the glycols on ethyl oxalate are linear condensation polymers.

Ethyl (^-hydroxy ethyl) -oxalate and ethylene-fos-ra-bromobenzoate are
described.

Bibliography and Remarks

(1) (a) Bischoff, Ber., 27, 2939 (1894); (b) 40, 2803 (1907); (c) Adams and Weeks, Journ. Am.
Chem. Soc., 38, 2518 (1916); (d) Tilitcheev, Ber., 56, 2218 (1923).

(2) Bischoff, ibid., 40, 2803 (1907).

(3) Carothers and Arvin, Journ. Am. Chem. Soc., 51, 2560 (1929); Carothers and van Natta,
ibid., 62, 314 (1930); Carothers and Borough, ibid., 62, 711 (1930).

(4) Tilitcheev, J. Russ. Phys.-Chem. Soc., 68, 447 (1926); C. A. 21, 3358 (1927); Chem. Zentr.,
II, 440 (1927).

(5) Carothers, Journ. Am. Chem. Soc., 51, 254g (1929).

(6) Carothers and Dorough, Journ. Am. Chem. Soc., 52, 718 (1930).

(7) Carothers and van Natta, ibid., 62, 314 (1930).

VI. ADIPIC ANHYDRIDE 63

VI. Adipic Anhydride*

J. W. Hill describes the preparation of adipic anhydride. In accordance
with the general rule put forward by Carothers, the anhydride of adipic
acid should be capable of existing as a 7 -ring and as a long chain polymer.
The author has succeeded in preparing both forms and also in converting
the monomer (m. w. 128) into the polymer of a molecular weight around
800 (polymerization degree 6-7). The reaction of the monomer with
aniline and of the polymer with aniline and phenol was studied.

The anhydrides of succinic and glutaric acids are known and are mono-
meric rings of five and six atoms, respectively. The monomeric anhydride
of the next member of the series, adipic acid, is a seven-atom ring, and
therefore the direct preparation of this anhydride by the removal of water
from the acid should give a polymolecular product, in accordance with the
generalizations formulated in the previous papers in this series (1). This
has been found to be the case. We have also been successful in preparing
monomeric adipic anhydride for the first time.

Adipic anhydride has been prepared by Voerman (2) and by Farmer and
Kracovski (3), both of whom describe it as a solid. Voerman prepared the
compound by treating adipic acid with acetyl chloride, removing volatile
compounds in vacuo y and finally crystallizing from benzene He gives a
melting point of 98. On the basis of a molecular weight determination in
phenol, with which, however, we have found it to react, he seemed to
regard the compound as a monomeric ring. The abnormally high molecu-
lar weight obtained in boiling acetone, the lack of definite crystallinity
and the relatively low solubility of the compound, however, led him to
consider that it might be polymeric. He does not make his viewpoint
altogether clear, but apparently he regarded the anomalous properties of
the compound as due to some sort of association of simple rings and did not
consider the probability of the existence of a long chain or large ring.

Farmer and Kracovski state that adipic anhydride is definitely uni-
molecular and ascribe to it a melting point of 97 This is certainly not
correct since we have found the monomer to melt at a very much lower
temperature.

* J. W. Hill; Journ. Am. Chem. Soc., 52, 4110-14 (1930); Communication No. 40
from the Experimental Station of E. I. du Pont de Nemours and Co.

Received July 26, 1930. Published October 6, 1930.

This paper does not bear the name of W. H. Carothers, but it is appropriate to
include it in this volume and give a short abstract. This will maintain the full con-
tinuity of this interesting and important series of contributions to high polymeric chem-
istry. Compare also paper No. XVII on page 192.

64 POLYMERIZATION AND RING FORMATION

We have prepared adipic anhydride by both of these methods and have
found that the product before subjection to distillation is polymeric, as
we had anticipated and Voerman suspected. Polymeric adipic anhydride
separates from hot benzene as a microcrystalline powder. Samples pre-
pared in various experiments showed melting points ranging from 70 to 85.
This variation and the difference between the melting points of our prod-
ucts and those of previous investigators may be due to differences in molecu-
lar size. A molecular weight determination in boiling benzene gave a value
about six times the normal value. This value is probably low as the
substance is very sensitive to moisture, and only the ordinary precautions
were taken to exclude it. The compound hydrolyzes in boiling water to
yield the acid and reacts rapidly with aniline and phenol. It is hygro-
scopic and on standing unprotected gradually reverts to the acid. It is not
distillable as such, but when heated in vacua to somewhat above 200, it
breaks down and yields the true monomeric adipic anhydride, which distils
out.

Monomeric adipic anhydride is a liquid freezing at about 20 and boiling
at about 100 at 0.1 mm. It reacts with cold water quickly with the evolu-
tion of much heat to yield the acid. It reacts almost instantly with
aniline to give a very high yield of pure monoanilide. On being heated at
100 for a few hours, it changes to a polymer melting at 80-85. On
standing at room temperature for ten days or more, it gradually solidifies.
It is probable that traces of water bring about the polymerization. When
a sample of liquid anhydride is transferred to a vessel which has not been
very carefully dried, an amorphous skin soon coats the glass.

The reactions of the monomeric and polymeric adipic anhydrides with
aniline establish the structures of these compounds, the former as a seven-
membered ring and the latter as a long chain, which may be a large ring or
an open chain with terminal carboxyl groups. Monomeric anhydride (I),
which has a symmetrical structure, can react with aniline to give but one
product, adipic acid monoanilide (II). This has been verified experi-

^^COOH
(CH,) 4

mentally. Polymeric adipic anhydride, to which we may assign formula
III, on the other hand, is no longer symmetrical as soon as one anhydride
group reacts with aniline. The product formed, therefore, depends on

VI. ADIPIC ANHYDRIDE 65

which side of the oxygen atom the next anhydride group along the chain
breaks. It will be seen from the scheme above that the reaction of poly-
meric adipic anhydride with aniline may lead to three products: adipic

oj o o jo oiro "1

4 C-l-0-C(^

H H H | NHC 6 H fi \ H

1 L J*

o o o o o

(CH 2 ) 4 C
C 6 H 6 NH

CeHNH
monoanilide adipic acid dianilide, etc.

Ill

acid, adipic acid monoanilide and adipic acid dianilide. All of these com-
pounds have been isolated from the reaction of the polymer with aniline.
On the assumption that a random reaction takes place and that the car-
boxyl groups which are probably at the ends of the chain may be neglected,
these products should be formed in the proportion of one mole of acid to
one mole of dianilide to two moles of monoanilide. In a quantitative ex-
periment a 25% yield of dianilide was obtained. The adipic acid and
monoanilide formed were not estimated quantitatively. Similarly the
solid anhydride yields diphenyl adipate on warming with phenol.

The experiments with aniline show definitely that the solid adipic an-
hydride is a true condensation polymer in which the structural units are
held together by primary valence forces. It is presumed that the polymers
are open chains with carboxyl groups at the ends as has been demonstrated
in the case of some of the polymeric ethylene succinates (4) . This view is
further supported by the presence of adipic acid in the distillate when the
polymeric anhydride is cracked to form the monomer.

The writer wishes to thank Dr. W. H. Carothers for hrs interest and
advice, and Mr. S. B. Kuykendall for determinations of carbon and
hydrogen.

Experimental Part

Preparation of Polymeric Adipic Anhydride. One hundred grams of recrystallized
adipic acid and 300 cc. of redistilled acetic anhydride were refluxed together for four
to six hours, and the volatile constituents were removed in a vacuum at 100. The
residue melted at about 70. It was repeatedly recrystallized from dry benzene, from
which it separated as a micro-crystalline powder. Various samples prepared in this
same way melted at temperatures varying from 70 to 85. At temperatures above the
melting point, all these samples were viscous liquids, which solidified on cooling to waxy
masses. In all experiments moisture was carefully excluded. In the most careful
preparation carried out, moisture was excluded by phosphorus pentoxide tubes, the
residue was heated overnight in an exhausted system which contained caustic potash

66 POLYMERIZATION AND RING FORMATION

in an adjacent vessel and then recrystallized four times with one nitration in a closed
system. After each recrystallization the mother liquor was removed by careful de-
cantation. The product melted at 73-75. The polymeric anhydride melted under
boiling water and gradually went into solution. On cooling adipic acid crystallized out.

Anal. Calcd. for (CeHgOa)^ saponification equivalent, 64.0. Found: saponi-
fication equivalent, 63.2, 63.4.

Reaction of Polymeric Adipic Anhydride with Aniline. A sample of polymer pre-
pared directly from adipic acid and acetic anhydride (m. p. 75) was added to a slight
excess of aniline and triturated in a mortar. Reaction took place spontaneously with
the evolution of heat. The mixture was treated with dilute hydrochloric acid and the
solid precipitate was filtered off. The filtrate was evaporated to dryness and the residue
was then taken up in a volume of water not quite sufficient for complete solution, filtered
and washed. The residue after crystallization from water was identified as adipic acid
by a mixed melting point determination. The residue from the hydrochloric acid treat-
ment was boiled with water and filtered. The filtrate deposited needles on cooling, which
after recrystallization from water, were identified as adipic acid monoanilide by a mixed
melting point with an authentic sample. The residue insoluble in boiling water, con-
sisting of adipic acid dianilide, was recrystallized from alcohol; needles, m. p. 240-241 .
(5). This experiment was carried out with various samples of polymeric anhydride:
crude polymer prepared directly from adipic acid and acetic anhydride, recrystallized
polymer, polymer formed by heating the monomer and still-residue from distillation of
the monomer. In every case dianilide was isolated and identified. The experiment with
the heat polymer was carried out quantitatively and gave a 25% yield of dianilide.

Anal. Calcd. for C, 8 H,oO 2 N2: C, 72.96; H, 6.81. Found: C, 73.31, 73.24; H,
6.95, 7.12.

Reaction of Polymeric Adipic Anhydride with Phenol. Polymeric adipic anhydride
was warmed with phenol to fusion and poured into cold water. On agitation a white
solid separated out. This, after two recrystallizations from a mixture of alcohol and
water (1 : 1), separated as lustrous plates of m. p. 105.5-106. The method of prepara-
tion and the ultimate analysis showed this compound to be diphenyl adipate.

Anal. Calcd. for Ci 8 H 18 O 4 : C, 72.45; H, 6.08. Found: C, 72.63, 72.19; H, 6.07,
6.29.

Preparation of Monomeric Adipic Anhydride. One hundred grams of adipic acid
and 300 cc. of acetic anhydride were refluxed together for four hours. The acetic acid
formed in the reaction and the excess acetic anhydride were removed by distillation in a
vacuum. The residue was transferred to a Claisen flask and heated under vacuum.
After the removal of the residual acetic anhydride at 0.1 mm. no further distillation took
place up to 210 bath temperature. At this temperature the pressure rose, and in differ-
ent experiments from 50 to 60 g. distilled between 105 and 125 under pressure from
3 to 8 mm., the bath temperature ranging from 210 to 235. The residue was viscous
when hot and solidified to a dark colored mass on cooling. The distillate consisted of a
light pink liquid containing a small amount of crystalline material. The liquid and solid
phases were separated by centrifuging. The crystalline material was washed with ben-
zene, recrystallized from ethyl acetate, and identified as adipic acid by a mixed melting
point determination. The liquid was redistilled in an all-glass apparatus. The boiling
point was 98 to 100 at 0.1 mm. at a bath temperature of 117 to 120. Only about half
of the material could be recovered by distillation since the residue gradually polymerized.

VI. ADIPIC ANHYDRIDE 67

Samples of the colorless distillate were sealed in small, carefully dried glass tubes for
freezing point determinations. The compound froze at 19 to large translucent plates
andremeltedat22.

Anal. Calcd. for CeH 8 O 8 : C, 56.25; H, 6.30; mol. wt., 128; saponification equiv.,
64.0. Found: C, 55.89, 55.98; H, 6.51, 6.52; mol. wt. (in boiling benzene), 134, 131;
saponification equiv., 63.7, 63.6.

Polymerization of Monomeric Adipic Anhydride. When monomeric adipic an-
hydride was transferred to a vessel which had not been carefully dried, a translucent coat-
ing appeared on the surface of the glass. The coating was amorphous and in places
skin-like. It melted at 60-65. When the liquid was heated in a sealed tube at 100,
it polymerized completely and a maximum melting point of 80-84 was reached in about
seven hours. On being heated at 138, a maximum melting point of 81-85 was reached
in about two hours. In neither experiment did further heating affect the melting point.

Anal. Calcd. for (CeHsOa)^ saponification equiv., 64.0. Found: saponification
equiv., 63.5, 63.3; mol. wt. (in boiling benzene), 860, 710.

Reaction of Monomeric Adipic Anhydride with Aniline. One cubic centimeter
of anhydride was added to about eight cubic centimeters of aniline with stirring. Re-
action was immediate and much heat was evolved. The excess aniline was dissolved in
dilute hydrochloric acid. The solution was diluted to 200 cc. and filtered. The crude
product was dried and weighed; m. p. 152-153; yield 80%. The monoanilide of adipic
acid has previously been prepared by heating C 6 H 6 NHCOCH(COOH)(CH2) 8 COOH (6).

Recrystallization from water did not raise the melting point. The crude com-
pound was completely soluble in hot water (absence of dianilide). A similar experiment
was carried out and worked up in smaller volume; m. p. (crude), 151-1.53; yield, 87%.

Anal. Calcd. for C^HuOsN: C, 65.12; H, 6.80; neutralization equiv., 221.
Found: C, 64.76, 64.86; H, 6.81, 7.01; neutralization equiv., 222, 223.

Summary

The monomeric and polymeric forms of adipic anhydride have been
prepared, the former for the first time, and they have been shown to be
mutually interconvertible. The reactions of the compounds with aniline
have shown the monomer to be a seven-atom ring and the polymer a long
chain or large ring. Diphenyl adipate is described.

Bibliography and Remarks

(1) Carothers, Journ. Am. Chem. Soc., 51, 2548 (1929); Carothers and Arvin, ibid., 51, 2560
(1929); Carothers and van Natta, ibid. t 52, 314 (1930); Carothers and Dorough, ibid. t 52, 711 (1930).

(2) Voerman, Rec. trav. chim. t 23, 265 (1904).

(3) Farmer and Kracovski, J. Chcm. Soc., 680 (1927).

(4) Carothers and Dorough, Journ. Am. Chem. Soc., 52, 711 (1930).

(5) BSdtker \Ber., 39, 2765, 4003 (1906)] gives 240; Balbiano [Gazz. chim. ital., 32, I, 446 (1902)]
gives 240-241; Bouveault and Titry [Bull. soc. chim., [3] 25, 444 (1901)] gives 235.

(6) Dieckmann, Ann., 317, 62 (1901). The melting point given is 152-153.

68 POLYMERIZATION AND RING FORMATION

VII. Normal Paraffin Hydrocarbons of High Molecular
Weight Prepared by the Action of Sodium on Decamethylene-

bromide*

Long chain paraffins can be obtained by treating decamethylenebromide
with metallic sodium in ethereal solution.

After an enumeration and discussion of the different reactions, which are
feasible in this system^ the preparation of the hydrocarbon mixture and its
separation are described. The latter is carried out in a molecular still at a
pressure of about 10~ 5 mm. Hg. Between the melting points of 35 and
106 eighteen fractions were separated. The distillates were grouped
together into eight fractions and redistilled; they crystallized with con-
stant melting points and proved to be hydrocarbons up to C 7 oHi42. A
curve showing the melting points as a function of chain length is given
and thoroughly discussed.

Finally a detailed description of the preparation of decamethylene-
bromide is given.

No rationally synthetic and practical methods for the preparation of
giant individuals of the simpler homologous series are available, and one
of the objects of the experiments here reported was to explore the possi-
bilities of bifunctional Wurtz reactions as a means of access to this realm.

The first step in simple Wurtz reactions is the formation of the corre-
sponding sodium alkyl, RNa (1). This normally couples with the halide,
forming the hydrocarbon R-R, but it may react in other ways and yield
by-products. Sodium ethide reacts with diethyl ether

C 2 H 6 Na + (C 2 H 5 ) 2 O = C 2 H 6 + C 2 H 6 ONa + C 2 H 4 (2)

and one may expect the formation of considerable amounts of the hydro-
carbon RH when the Wurtz reaction is carried out in that solvent.

The experiments presently described are concerned with the action of
sodium on decamethylene bromide in ethereal solution, and the normal
course of this reaction including that due to the participation of the solvent
(but ignoring for the moment the possibility of intramolecular reaction)
may be formulated as follows

Br(CH 2 )ioBr + 2Na - Br(CH 2 )ioNa + NaBr
Br(CH 2 ) 10 Na + (CiH,)jO - H(CH 2 ) 10 Br + CaHiONa
Br(CH : )ioNa + Br(CH 2 )ioBr - Br(CH 2 ) 20 Br + NaBr

* W. H, Carothers, J. W. Hill, G. E. Kirby and R. A. Jacobson; Journ. Am. Chem.
Soc., 52, 5279-88(1930); Contribution No. 46 from the Experimental Station of E. I.
du Pont de Nemours and Co.

Received October 9, 1930. Published December 18, 1930.

VII. NORMAL PARAFFIN HYDROCARBONS, ETC. 69

H(CH 2 )i Br + 2Na = H(CH 2 ) 10 Na + NaBr
H(CH 2 )ioNa + (C 2 H 6 ) 2 O - H(CH 2 ) 10 H + C 2 H 6 ONa + Q^
H(CH 2 ) 10 Na -f H(CH 2 )i Br - H(CH 2 ) 20 H + NaBr
H(CH 2 )ioNa + Br(CH 2 )i Br = H(CH 2 ) 20 Br -f NaBr
H(CH 2 )ioNa + H(CH 2 ) 20 Br - H(CH 2 ) 30 H + NaBr
H(CH 2 )ioNa + Br(CH 2 ) 20 Br = H(CH 2 ) 80 Br + NaBr
H(CH 2 ) 3 oBr + 2Na = H(CH 2 )soNa -f NaBr
H(CH 2 ) 80 Na + Br(CH 2 ) 20 Br = H(CH 2 ) 60 Br + NaBr, etc.

The product will be composed of individuals of the general formula
H[(CH 2 )ioLH and possibly of similar chains terminated at one or both ends
by bromine atoms. The length of these chains will be determined by the
relative rates of reaction of the sodium compounds with the solvent and
with the halides. Solubility effects may also come into play.

The suitability of this reaction as a source of straight chain hydrocarbons
will depend upon the extent to which it is possible to avoid undesirable
side reactions, and some indications on this point may be had from the
known behavior of simple halides. Sodium and w-heptyl bromide at the
boiling temperature give tetradecane, 67%; heptane, 9%; heptene, 3%;
heneicosane, 3%; hydrocarbons of higher molecular weight, some (3).
Sodium and butyl bromide in ether under specified conditions give n-
octane, 68%; butane ?, ?%; butene, ca. 1.5%; dodecarie, ca. 0.15%;
hexadecane, ca. 0.05% (4). In both these typical cases the principal by-
product is probably the hydrocarbon RH. In the reaction of the dihalide,
this type of by-product would be identical with one of the normal reaction
products. The higher hydrocarbons from butyl bromide and from heptyl
bromide are probably not straight chains, and the formation of analogous
products from the dihalide would be definitely pernicious because of the
difficulty of separating isomeric individuals of high molecular weight.
The quoted data on butyl bromide indicate however that in simple Wurtz
reactions in ether not more than traces of these higher by-products need be
formed.

The action of sodium wire on decamethylene bromide in absolute ether
has been studied by Franke and Kienberger (5). The reaction proceeded
smoothly to completion, and the products isolated were: w-decane, 34%;
C 20 hydrocarbon, 30%; a small amount of C 40 hydrocarbon; and a con-
siderable amount of solid hydrocarbon of higher molecular weight. All of
these products were free of halogen. The melting point of the C 2 o hydro-
carbon identifies it as n-eicosane. The formation of large amounts of
w-decane apparently puzzled Franke and Kienberger, and they repeated
their experiments with elaborate precautions to exclude water and alcohol
from the solvent and the halide. The results were the same. n-Decane
and w-eicosane are, however, strictly normal products of the reaction as it

70 POLYMERIZATION AND RING FORMATION

has been formulated above. The fact that Franke and Kienberger could
find neither decene nor cyclodecane in the products is worthy of note, for
any side reaction would almost certainly be accompanied by the formation
of some decene, and any intramolecular reaction by the formation of some
cyclodecane. These data indicate that the course of this reaction is, at
least for the most part, strictly normal and exclusively intermodular. The
major part of the product is composed of relatively short chains, but this
is a matter which may be expected to be susceptible of some control by
changes in the experimental conditions, and, as it appears to be in fact,
especially by increase in the surface of the sodium.

We had already treated decamethylene bromide in ether with finely di-
vided sodium with stirring before Franke and Kienberger's paper had come
to our attention. The product, obtained in good yield, was an ether-in-
soluble solid, and no material boiling as low as w-decane was present. This
product, by extraction and crystallization, was separated into C 2 oH42,
C 3 oH 62 and C^Haz, each fairly pure, representing together about 25% of
the total, and a higher fraction, m. p. 108-112, representing about 75%
of the total. This contained 1.39% organic halogen, from which it was
freed by heating and stirring it with a small amount of molten sodium in
boiling butyl ether. The apparent molecular weight of the resulting hydro-
carbon was about 1000. No individuals could be isolated from it by re-
peated fractional crystallization, and it could not be distilled at 0.1 mm.
pressure. It was finally separated into a series of individuals by distilla-
tion in the molecular still and by crystallization of the distillates.

Preparation of the Hydrocarbon Mixture. In a 500-cc. flask provided with a
reflux condenser and a mercury-sealed mechanical stirrer, 95 g. (1.52 atoms) of sodium
was pulverized under hot xylene. The xylene was removed and replaced by absolute
ether, and then 75 g. (0.25 mole) of decamethylene bromide was added. The reaction
mixture was stirred continuously. It soon developed a deep blue color. It was gently
heated for one hour and then allowed to boil without external heating for two and one-
half hours. Finally it was heated for two and one-half hours more and then allowed to
stand overnight. The excess sodium in the thick mass was decomposed by alcohol, a
large volume of water was added and the mixture was filtered with suction. The only
material found in the ethereal layer of the filtrate was 0.5 cc. of an oil boiling at 240-
330. The soft white residue on the funnel liquefied almost completely on being
stirred with boiling water and solidified to a crystalline mass on cooling. It weighed
30.5 g. (87.1% calculated as CH 2 ). It melted from 85 to 100 and contained bromine
(found, 2.17, 2.27%). This solid was heated and vigorously stirred for five hours
with 7 g. of molten sodium in boiling butyl ether. The excess sodium was decomposed
by alcohol, a large volume of water was added and the mixture was filtered. The solid
residue on the funnel was washed with boiling water and dried. It weighed 21 g.
It was free of halogen and melted from 87 to 105. This is the material which was sub-

VII. NORMAL PARAFFIN HYDROCARBONS, ETC.

mitted to fractionation in the molecular still. After the butyl ether had been removed
from the non-aqueous layer of the distillate, there remained 6 g. of a soft waxy solid
melting at 4&-54 .

Separation of the Hydrocarbon Mixture. Eight grams of the solid
hydrocarbon, m. p. 87-105, described above was heated in a small mo-
lecular still (6) provided with a water-cooled condenser and a trap cooled
with liquid air. The pressure in the system was continually maintained
below 10~ 6 mm. Fractions were collected as follows.

Temp, of bath, C.
130

150-160
150-160
150-160

160

180

190

195

195
200-220

250

300

Wt. of distillate, g.
0.7
.27
.15

.20

.12

.20

.19

.08

.10

.52

.31

.11

.20

.10

.05

.08

.47

.17

Total 4.02
Residue 3.16
Loss 0.82

M. p. of distillate, C.

35
60

60
77
78
79
87
87
90
90.
94.
96
97
98
99
99
103
103

- 55

- 75

- 78

- 79

- 81

- 84

- 89

- 91

- 90.5
5- 92
5- 97

- 98

- 99.5
-100.5
-101
-102
-105
-106.5

TABLE I
INDIVIDUAL HYDROCARBONS

Solvent used

Distn.

Hydro-
carbon

M. p.
found, C.

for crys-
tallization

te o mp.,

Anal.
C

calcd.
H

Anal.
C

found
H

>^2oH42

- 35

Abs. EtOH

.00

85.45

15.03

\JsoHg2

- 66

Abs. EtOH 4- Et 2 O

100

.21

.79

85.28

14.45

C4 H 82

80,

,5- 81

Ethylene chloride

150

.31

.49

85.59

14.74

C6oHio2

91.

,9- 92

Ligroin -f petroleum

ether

200

.37

14.

,63

85.34

14.53

C(M)H 122

98.

5- 99,

Butyl acetate

250

85.

14.

85.66

14.34

C 7 oH 14 2

105

-105,

Butyl acetate

300

85.

14.

85.50

14.58

Soluble
residue 110 -114 Butyl acetate

Not distillable

POLYMERIZATION AND RING FORMATION

100

The loss is due to the impossibility of quantitatively removing the distil-
late from the condenser. In this distillation the average area of the evapo-
rating surface was about 18 sq. cm. About twenty hours was required

for the collection of each of the
above listed fractions. When
the temperature of the heating
bath was raised above 300 with
the view of distilling the residue,
the pressure rose and no distilla-
tion could be effected.

The distillates were grouped
together into eight fractions and
redistilled. The rate of distilla-
tion from these partially puri-
fied materials was very much
higher than from the initial mix-
ture. The distillates were crys-
tallized to constant melting
points. The properties of all
the fractions thus obtained are
indicated in Table I.

Identity and Properties of
the Hydrocarbons. The melt-
ing points 36.7 (7) and 65.6-
66 (8), respectively, have been
assigned to w-eicosane and n-tri-
acontane, and these values are
in good agreement with our
observations. The other four
hydrocarbons listed in Table I
are all new, and the last one

20 40 60 80 100
Melting points of hydrocarbons, C.
Fig. 1. Normal paraffins: , Gascard,
Ann.Chim., [9] 15,332 (1921); + , Hildebrand
and Wachter, Journ. Am. Chem. Soc., 51, 2487
(1929); O, new data. Cyclopolymethylenes :
A, Ruzicka and co-workers, Helv. Chim. Acta,
9, 499 (1926); 11, 496 (1928).

stands six atoms above any
paraffin previously described.
Several [intermediate members
of this same series have been
described, however, and from

the available data it is possible to construct a smooth curve of melt-
ing points covering the range Ci 8 -C 7 o. The usefulness of such a curve
in estimating the most probable values for individual hydrocarbons
has already been emphasized (9), and in the curve presented in Fig. 1 the

VII. NORMAL PARAFFIN HYDROCARBONS, ETC. 73

crosses are Hildebrand and Wachter's estimates of the best values for hy-
drocarbons in the range Ci 9 -C 3 6. We have used a method of plotting simi-
lar to that suggested by Austin (10) and have included the values for the
three previously known hydrocarbons above C 3 6H 7 4. These are C 6 4Hno,
C 6 2Hi26 and C64Hi 30 , all described by Gascard (8). Our data and those of
Gascard fall very close to the same smooth curve which fits the best data
for the lower hydrocarbons, and this fact incidentally affords confirmation
of Hildebrand and Wachter's contention that the melting points accepted
by the "International Critical Tables'* for some of the hydrocarbons in the
range C 2 4 to C 32 are much too high. Hell and Haegele described dimyricyl
in 1889 (11) and ascribed to it the formula CeoHm. This ascription ac-
quired the sanction of recognition by Beilstein (third and fourth editions)
and remained unquestioned until evidence that the myricyl radical contains
31 carbon atoms appeared. In 1921 Gascard (8) concluded that Hell and
Haegele's hydrocarbon was C 6 2Hi 2 4 and that hexacontane had never been
prepared. Our data confirm this conclusion. Hexacontane melts at
98.5 to 99.3, while the observed melting points ascribed to dimyricyl have
ranged from 100.5-102.

Independent proof of the identity of the six hydrocarbons is furnished
by x-ray diffraction patterns which have been obtained by Dr. A. W. Ken-
ney and will be described in a separate publication.

Melting point data are available for only five cyclic polymethylenes
above Cio. These fragmentary data do not fall on any smooth curve (see
Fig. 1), but they are all widely enough separated from the data for the
normal paraffins to provide additional assurance that none of our paraffin
compounds are cyclic.

Under the heading distillation temperature in Table I are listed the mini-
mum bath temperatures required to effect evaporation at a moderate rate
in the molecular still. These temperatures are quite characteristic since
they are, in each case, only a few degrees above the temperatures at which
no distillation occurs. Using the Langmuir formula (6) it is possible to
make a rough calculation of the order of magnitude of the vapor pressures
of these hydrocarbons from the minimum distillation temperatures. Such
calculations indicate a value less than 0.01 mm. for heptacontane at 300
and a value less than 1 mm. for triacontane at 100. Heptacontane has a
molecular weight of 983, and it is probable that no organic compound of
much higher molecular weight than this can ever be distilled under any ex-
perimental conditions however favorable. The residue remaining from the
removal of the heptacontane undoubtedly contained some octacontane,
but none of this distilled out at a bath temperature of 300 and when the

74 POLYMERIZATION AND RING FORMATION

temperature was raised higher than this, decomposition set in. This ex-
perimental result agrees remarkably well with the inferences to be drawn
from the data presented by Meyer and Dunkel (12). From these data one
can calculate that the molecular cohesion of heptacontane will be about
71,000 calories and that of octacontane about 81,000 calories. Since the
heat of separation of the carbon-carbon bond is about 75,000 calories, it
should be possible to distil the first of these compounds, but not the second.
Saturated paraffins have, in general, a lower molecular cohesion for a given
molecular weight than any other types of compounds, and hence the limit
of distillability will be found at a lower molecular weight for other com-
pounds than hydrocarbons. These results are also in agreement with data
of a different kind presented by Burch (13) who, by distilling a Pennsyl-
vania petroleum in a molecular still, obtained as the highest distillable frac-
tion a material having an apparent average molecular weight of 801 and a
residue of apparent molecular weight of 1550.

The physical properties of w-heptacontane are similar to those ascribed
to dohexacontane and tetrahexacontane. It is very slightly soluble in
boiling alcohol, ether or petroleum ether, but crystallizes well in the form
of minute needles from hot butyl acetate or benzene. It dries to a starch-
like powder which has a great tendency to become electrified.

It was hoped that it might be possible to isolate hydrocarbons much
higher than heptacontane, since information concerning simple individuals
of very high molecular weight would be of great importance to the study
of macromolecular materials generally, but there seems little probability
of achieving this result without some change in method or advance in
technique. The residue from which the n-heptacontane was distilled is a
mixture which undoubtedly contains n-octacontane, w-nonacontane, n-
decacontane and still higher hydrocarbons. It is readily soluble in hot
butyl acetate (except for a trace of insoluble and infusible material), and
it separates as a powder melting at 110-114. Its apparent average
molecular weight in boiling benzene is about 1300. (Calcd. for CgoHm,
1263.) By extraction with hot ethylene chloride, it was separated into
fractions melting at 100-107, 106-111 and 110-114. In a sense this
mixture is a polymeric homologous series, but its solubility is much lower
than that of such supposedly analogous series of higher molecular weight as
polystyrene and hydro-rubber. It resembles a probably similar but more
complex mixture of hydrocarbons obtained by Fischer and Tropsch by
the catalytic hydrogenation of carbon monoxide (14).

There is one point bearing on the behavior of highly polymeric materials
which is capable of a rough preliminary test with the series of hydrocar-

VII. NORMAL PARAFFIN HYDROCARBONS, ETC. 75

bons described above. It has been suggested (15) that the apparent de-
crease of molecular weight (decrease of viscosity, etc.), which is observed
when rubber and other very high polymers are gently heated or treated
with certain mild chemical agents or subjected to mechanical stresses, is
real and is due to the fact that the thermal stability of molecules decreases
continuously with increasing size, and in these materials has reached so low
a value that cracking occurs at slightly above room temperature. Samples
of each of the hydrocarbons from C 3 o to CTO were sealed off in small bulbs
under nitrogen and heated side by side in a metal-bath. After five minutes
at 400 all the melting points were unchanged. After five minutes at 410
the melting point of C 70 was lowered from 106-107 to 104-105. The
melt, moreover, was not clear as before but turbid. The turbidity dis-
appeared rather sharply at 1 10 like a liquid crystal. The melting points
of the other hydrocarbons were substantially unchanged, but, except for
C 30 , they all gave hazy melts which cleared up only at temperatures some-
what above their true melting points. After five minutes more at 420 all
the melting points had become lower (that of 70 most) and there was some
coloration. The C 30 was white, the C 4 o was cream-colored and the higher
hydrocarbons in regular order were increasingly darker. When the tubes
were finally opened after five minutes more at each of the temperatures
430, 450 and 470, they were found to contain considerable gas most
above the Cyo and least above the C 30 . Thus, the decrease in thermal
stability with increasing molecular weight which is so marked in passing
from methane (ca. 700) (16) to ethane (ca. 550) (17), and from ethane to
hexadecane (ca. 470) (18) is still detectable in going from triacontane to
heptacontane, but it has already fallen to such a small value that it would
be somewhat unsafe to infer that a paraffin hydrocarbon of molecular
weight 200,000 or even greater might not persist at room temperature.

Mechanism of the Action of Sodium on Decamethylene Bromide. It
has already been pointed out that all the observed products of this reaction
are accounted for by assuming replacement of the bromine atoms of the
halide by sodium, and the subsequent coupling of this with other molecules
of the halide, or its reaction with the solvent. The absence of cyclic hy-
drocarbons is not surprising since, although the higher cycloparaffins are
no less stable than cyclohexane (19), the formation of large rings in bifunc-
tional reactions occurs only under very exceptional conditions (20). In
the formation of esters from co-hydroxy acids, amides from amino acids and
anhydrides from dibasic acids, intramolecular reaction occurs exclusively
only if there is a possibility of forming a 5- or a 6-membered ring. The
behavior of alkylene halides toward metals however is peculiar in that there

76 POLYMERIZATION AND RING FORMATION

seems to be little tendency toward intramolecular reaction even when a 5-
or a 6-membered ring might be formed. Thus the action of magnesium on
alkylene halides leads to considerable coupling, but even from pentamethyl-
ene halides no cyclic hydrocarbon is formed (21). This suggests that the
coupling perhaps occurs largely at the metal surface where the atoms at
the ends of the chains may be fixed by forces of adsorption and their free-
dom of intramolecular approach hindered.

Our main objective in the experiments described above was to prepare
normal paraffins of very high molecular weight, and we made a number of
attempts to adjust the experimental conditions so as to increase the length
of the reaction chains. We treated decamethylene bromide with molten
sodium in boiling butyl ether, with liquid sodium-potassium alloy in ethyl
ether, with molten sodium in boiling octane and with finely divided sodium
in the absence of a solvent. There was no evidence of the formation of
considerable amounts of paraffins higher than Cioo in any of these experi-
ments. We also prepared decamethylene dimagnesium bromide and
treated it with cupric chloride in the expectation that products of the
formula BrMg[(CH 2 )io]*MgBr would be formed. The chief product
(71%), however, after the addition of water, was a volatile liquid appar-
ently chiefly composed of decane and decene. The amount of higher boil-
ing material (20%) did not exceed that which would arise from the coupling
products usually produced in the formation of such reagents as decamethyl-
ene dimagnesium bromide (21).

Preparation of Decamethylene Bromide. The preparation, in good
yields, of a whole series of polymethylene bromides from the glycols via
the esters of the corresponding dibasic acids has been described by Chuit
(22) but without great experimental detail. We first attempted to reduce
ethyl sebacate by Levene and Allen's modification (23) of the original
Bouveault procedure (24). In this modification, which has given good re-
sults in the reduction of esters of monobasic acids, the sodium is first granu-
lated by stirring in hot toluene, and the reaction mixture is vigorously
stirred during the addition of the alcohol-fester mixture. This procedure
gave very poor yields of decamethylene glycol; the yields were increased
by diminishing the speed of stirring and became quite good when the stir-
ring was omitted altogether. The procedure finally adopted was almost
identical with that of Bouveault, and the details are given below.

A condenser having a length of 2 meters and an inside diameter of about 2.5 cm. is
connected with the side neck of a 5-liter two-necked round-bottomed flask, and the
central neck of the flask is provided with a 1 -liter dropping funnel. One hundred and
fifteen grams (5 atoms) of sodium in a single piece or in two or three large pieces is

VII. NORMAL PARAFFIN HYDROCARBONS, ETC. 77

placed in the flask. One hundred and twenty-nine grams (0.5 mole) of ethyl sebacate
is dissolved in one liter of thoroughly dried absolute ethyl alcohol and poured into the
dropping funnel. One hundred and fifty to two hundred cubic centimeters of the al-
cohol-ester mixture is allowed to fall onto the sodium at once. After two or three
minutes the reaction becomes vigorous and the sodium melts. As soon as the sodium
has melted another portion (about 100 cc.) of the alcohol-ester mixture is suddenly
added. This causes the molten sodium to break up into fine particles, and the reaction
becomes much more vigorous. The rest of the alcohol-ester mixture is then added as
rapidly as possible (five to six minutes). As soon as the vigor of the reaction has some-
what subsided, the flask is heated by an oil-bath to 110-115 until most of the sodium
has dissolved. The ethyl alcohol is then removed from the reaction mixture by steam
distillation, and the glycol is removed from the residue by extended (about seventy-two
hours) continuous extraction with ether. The glycol is purified by distillation. The
combined yield from nine such runs was 593 g. or 75% of the theoretical.

Continuous ether extraction as a means of isolating glycols produced by the reduc-
tion of esters has been successfully used by C. S. Marvel (25) ; otherwise the above pro-
cedure is similar to that which has also been used by Miiller for preparing a series of
glycols (26).

Using the method described by Chuit (22), we obtained from decamethylene glycol
only a 70% yield of bromide and this was contaminated with bromohydrin. The
following modification of this method gave a purer product and a better yield.

Dry hydrogen bromide was led into a well-stirred melt of 255 g. (1.47 moles) of
decamethylene glycol in an open flask until the mass was saturated. The temperature
of a metal-bath surrounding the flask was kept at 90-95. After saturation was com-
plete the bath temperature was raised to 135-140 and a slow stream of hydrogen bro-
mide was led in for six hours. The reaction mass was then cooled, diluted with benzene
and decanted from a small amount of water remaining in the flask. The benzene was
then removed by distillation, and the product distilled under diminished pressure;
yield, 376 g. (85.5%); b. p., 168-172 (10 mm.).

Anal. Calcd. for Ci H 20 Br 2 : Br, 53.33. Found: Br, 53.16, 53.16.

We are indebted to Mr. W. L. McEwen for the preparation of deca-
methylene glycol.

Summary

By the action of sodium on decamethylene bromide in ether a mixture
of paraffins of the general formula H [(CH 2 )io]* H has been prepared.
By fractional distillation in the molecular still and by crystallization, the
following individuals have been isolated from this mixture: w-eicosane,
w-triacontane, w-tetracontane, n-pentacontane, n-hexacontane, n-hep-
tacontane. The last four of these are new, and the last one has a higher
molecular weight than any pure paraffin hydrocarbon hitherto described.
The residue remaining from the separation of the hydrocarbons named
amounts to about 25% of the total and is composed of still higher members

78 POLYMERIZATION AND RING FORMATION

of the same series which it has not yet been possible to separate. The
mechanism of the reaction and the properties of the products are discussed.

Bibliography and Remarks

(1) Schorigin, Ber., 40, 3111 (1907); 43, 1931 (1910); Schlubach and Goes, ibid., 55, 2889
(1922); Ziegler and Colonius, Ann., 479, 135 (1930).

(2) Schorigin, Ref. 1; Ber., 41, 2723 (1908).

(3) Bachmann and Clarke, Journ. Am. Chem. Soc., 49, 2089 (1927).

(4) Lewis, Hendricks and Yohe, ibid., 50, 1993 (1928).

(5) Franke and Kienberger, Monatsh., 33, 1189 (1912).

(6) Washbura, Bur. Standards J. Research, 2, 476 (1929).

(7) Krafft, Ber. t 19, 2220 (1886). The Cao hydrocarbon of Franke and Kienberger (Ref. 5)
which melted at 36 was unquestionably n-eicosane; and their Cw hydrocarbon, which melted at 72,
was probably a mixture of n-triacontane and n-tetracontane.

(8) Gascard, Ann. chirn., [9] 15, 332 (1921); Peterson* Ber., 12, 741 (1879).

(9) Hildebrand and Wachter, Journ. Am. Chem. Soc., [9] 51, 2487 (1929).

(10) Austin, ibid., 52, 1049 (1930).

(11) Hell and Haegele, Ber., 22, 502 (1889).

(12) Meyer, Naturw., 16, 781 (1928); Dunkel, Z. physik. Chem., 138 42 (1928).

(13) Burch, Proc. Roy. Soc. (London), 123, 271 (1929).

(14) Fischer and Tropsch, Brennstoff-Chcmie, 8, 165 (1927).

(15) Staudinger, Ber., 59, 3037 (1926).

(16) Bone and Wheeler, /. Chem. Soc., 81, 542 (1902).

(17) Williams-Gardner, Fuel Science Practice, 4, 430 (1925).

(18) Gault and Hessel, Ann. chim., [10] 2, 319 (1924).

(19) Ruzicka and co-workers, Helv. Chim. Ada, 9, 499 (1926); 11, 496 (1928).

(20) Carothers, Journ. Am. Chem. Soc., 61, 2548 (1929); and subsequent papers of this series.

(21) Zappi, Bull. soc. chim., [4J 19, 249 (1916); v. Braun and Sobecki, Ber., 44, 1918 (1911).

(22) Chuit, Helv. Chim. Acta, 9, 264 (1926); Chuit and Hauser, ibid., 12, 850 (1929).

(23) Levene and Allen, J. Biol. Chem., 27, 443 (1916); cf. Adams and Marvel, Univ. Illinois
Bull., 20, 50 (1922).

(24) Bouveault and Blanc, Bull. soc. chim , [3] 31, 666 (1904).

(25) Private communication.

(26) Mailer, Monatsh., 49, 27 (1928).

VIII. Amides from e-Aminocaproic Acid*

e-Aminocaproic acid when heated reacts intra- and intermolecularly , and
there is formed 20-30% of the seven-membered lactam ring and 80-70% of
a C-polymer, containing probably at least ten molecules of the amino acid.
No information can be given as to the end group of the chain molecule,
but it is improbable that the molecule is a multi-membered ring.

-Aminocaproic acid on being heated reacts with itself and yields two
products the corresponding 7-membered lactam, 20-30%, and an un-
distillable material of approximately the same composition, 80-70% (1).

* W. H. Carothers and G. J. Berchet; Journ. Am. Chem. Soc., 52, 5289-91 (1930) ;
Contribution No. 42 from the Experimental Station of E. I. du Pont de Nemours and
Co.

Received October 9, 1930. Published December 18, 1930.

VIII. AMIDES FROM e-AMINOCAPROIC ACID /

The following related facts are of interest. 5-Aminobutyric acid and
-aminovaleric acid readily lose water and yield the corresponding 5-
and 6-membered lactams (2). No corresponding polymers are formed or
known. The reaction of f-aminoheptoic acid with itself is apparently
exclusively intermolecular, for it leads to an undistillable product quite
different in its properties from the known 8-membered lactam (3). Similar
relationships are found among the hydroxy acids: 7- and 5-hydroxy acids
yield lactones; e-hydroxy acids yield both lactones and polyesters (4) and
the higher w-hydroxy acids yield only the polyesters (5) , although many of
the higher lactones are known (6). Polyesters have now been studied in
some detail (7), but (except for polypeptides) little attention has been de-
voted to the polyamides. In this paper we record some observations
concerning the polyamide derived from c-aminocaproic acid.

Water is rapidly evolved when e-aminocaproic acid is heated above its
melting point (e. g., to 210-220). The polymer and the lactam which re-
sult are readily separated by allowing the latter to distil under diminished
pressure, or by extracting it with boiling alcohol. In different experi-
ments the yield of lactam was 20-30% and the polymer 80-70%. The
same products were obtained by gradually heating the ethyl ester of the
acid from 160 to 200 during six hours, and the yield of lactam was about
38%.

The polymer is a hard gray wax, insoluble in most organic solvents, but
soluble in hot formamide, from which it separates as a microcrystalline
powder melting at 212-214. Its analytical composition agrees very
closely with that required for the structural unit NH (CH^sCO .
Anal. Calcd. for C 6 H n ON: C, 63.71; H, 9.73. Found: C, 63.35, 63.62; H,
9.93, 10.07. A product from another preparation gave: C, 63.35, 63.45; H, 9.56, 9.43.

It is quantitatively hydrolyzed in six hours by boiling concentrated
hydrochloric acid to -aminocaproic acid (identified by conversion to the
p-toluenesulfonyl derivative, melting point and mixed melting point,
105-106).

By partial hydrolysis, polyaminocaproylaminocaproic acids are formed.
Thus, a sample of the polyamide was heated to boiling for one hour in con-
centrated hydrochloric acid solution, and the solution was diluted with
water to precipitate the unchanged polymer (about 50%). The ma-
terial, remaining in solution, on being treated with alkali and />-toluene-
sulfonyl chloride, gave a solid product which, after repeated crystalliza-
tion, showed a neutral equivalent corresponding with the formula QH?-
S0 2 -[NH(CH 2 ) 5 CO]3-NH(CH 2 ) 6 COOH (calcd., 607; found, 601). The
structure of the polymer may, therefore, be represented by the formula

80 POLYMERIZATION AND RING FORMATION

? NH(CH 2 ) 6 CO [NH(CH 2 ) 6 CO]x~NH(CH 2 ) 6 CO ?

It is a true condensation polymer and is formed directly from the amino
acid by intermolecular reaction. This is proved by the fact that the lac-
tam does not polymerize under the conditions of formation of the poly-
amide either in the presence or absence of catalysts. Moreover, no lactam
can be obtained from the polyamide by heating it to high temperature
in high vacuum. Thus there is no such reversible relationship between
lactam and polyamide as is found (7c, e) to exist between six-membered
cyclic esters and corresponding polyesters. Lactam and polyamide re-
sult from the amino acid by two independent and simultaneous processes.
The lactam incidentally is really the 7-ring r-NH(CH 2 ) 5 CO-| (8). This was

proved by identifying its hydrolytic product with c-ammocaproic acid
through its p-toluenesulfonyl derivative by means of a mixed melting
point.

The polyamide is sufficiently soluble in phenol to permit a determina-
tion of molecular weight by cryoscopy. Values obtained ranged from 800
to 1200, and since the errors of this method are such as to lead to low
results it may be assumed that at least 10 molecules of the amino acid
must have participated in the formation of a single molecule of the poly-
amide.

The question of the nature of the groups at the ends of the polyamide
chains cannot yet be answered. The formation of large rings is very
improbable on theoretical grounds (9), and in the formation of polyesters
ring formation is known not to occur the hydroxyl or carboxyl groups
still persist at the ends of the polyester chain (7b, c, d, 5c). However, no
certain evidence of the presence of terminal amino or carboxyl group in
the molecule of the polyamide could be obtained. The polyamide dissolves
readily in cold concentrated hydrochloric acid, but when the solution is
diluted with water it separates unchanged and free from more than traces
of halogen. It is not soluble in hot or cold aqueous sodium hydroxide nor
does it form any sodium salt. It reacts with molten m-bromobenzoic an-
hydride and, in hot pyridine, with ^-bromobenzene-sulfonyl chloride.
The products contain bromine in amounts which correspond with deriva-
tives of minimum molecular weights about 1100 and 1500, respectively,
but in view of the failure of other reactions these cannot be accepted as
evidence of the presence of free amino groups in the polyamide. At pres-
ent we are inclined to assume that the amino and carboxyl groups which
would normally be present at the ends of the polyamide chains are lost by
pyrolysis. (The evolution of carbon dioxide during the formation of the

IX. POLYMERIZATION 81

polyamide could not be detected, but the vapors evolved had a strong amine
odor.)

Summary

e-Aminocaproic acid (or its ethyl ester) on being heated reacts with
itself both intra- and intermolecularly and yields the 7-ring lactam (ca.
30%) and a polyamide (ca. 70%) . The formation of a molecule of the latter
probably involves at least ten molecules of the amino acid.

Bibliography and Remarks

(1) v. Braun, Ber., 40, 1840 (1907); Gabriel and Maas, ibid., 82, 1266 (1899).

(2) Gabriel, ibid., 22, 3338 (1889); Schotten, ibid., 21, 2240 (1888).

(3) v. Braun, ibid., 40, 1834 (1907); Wallach, Ann., 312, 205; 309, 18 (1899); Manasse, Bfr.,
35, 1367 (1902).

(4) Baeyer and Seuffert, ibid., 32, 3619 (1899); Blaise and Koehler, Compt. rend., 148, 1772
(1909); Marvel and co-workers, Journ. Am Chem. Soc., 46, 2838 (1924).

(5) (a) Blaise and Koehler, Ref. 4; (b) Chuit and Hauser, Helv. Chim. Ada, 12, 4634 (1929);
(c) Lycan and Adams, Journ Am. Chem. Soc., 51, 625, 3450 (1929); (d) Blaise and Marcilly, Bull,
soc. chim., 31, 308 (1904).

(6) Ruzicka and Stoll, Helv. Chim. Ada, 11, 1159 (1928).

(7) (a) Lycan and Adams, Journ. Am. Chem Soc , Ref. 5c; (b) Carothers and Arvin, ibid., 51,
2560 (1929); (c) Carothers and Van Natta, ibid., 52, 314 (1930); (d) Carothers and Dorough, ibid.,
52, 711 (1930); (e) Carothers, Dorough and Arvin, ibid., 52, 3292 (1930); (f) Chuit and Hauser,
Ref. 5b; (g) Blaise and Marcilly, Ref. 5d; (h) Bougault and Bourdier, /. pharm. chim., (6] 29, 561
(1909); [6] 30, 10 (1909).

(8) The shifting of a functional group down a chain in attempts to prepare large rings by bifunc-
tional reactions has been reported by various investigators. See, for example Blaise and Koehler,
Ref. 4.

(9) Carothers, Journ. Am. Chem. Soc., 51, 2556 (1929).

IX. Polymerization*

TABLE OF CONTENTS

/. Definitions 82

1 . Current definitions 82

2. Proposed definitions 83

3. Linear and non-linear polymers 84

4. Types of compounds capable of polymerizing 84

5. Types of polymerization 85

6. Condensation polymerizations and bifunctional reactions 85

II. Condensation polymerization 86

L Polyesters 86

a. The self-esterification of hydroxy acids 86

b. Polyesters from dibasic acids and glycols 90

* Wallace H. Carothers; Chemical Reviews 8, 353-426 (1931); Communication No.
55 from the Experimental Station of the E. I. du Pont de Nemours and Company.
Received March 21, 1931. Published June 1931.

82 POLYMERIZATION AND RING FORMATION

2. Bifunctional Wurtz reactions and Friedel Crafts reactions 93

3. Other bifunclional reactions 95

a. Polyamides 95

b. Polyamines 96

c. Polyacetals 96

d. Poly anhydrides 96

e. Grignard reactions 96

f. Sulfur and selenium compounds 98

g. Miscellaneous 98

4. Stereochemical factors involved in condensation polymerization. . . 100

a. Large rings 104

III. Polymerization involving cyclic compounds 106

1 . Six-membered cyclic esters 106

2. Adipic anhydride no

3. Diketopiperazines and polypeptides Ill

4. Ethylene oxide 113

IV. Addition polymerization of unsaturated compounds 113

1 . Ethylene and other olefines 114

2. Vinyl compounds 115

3. Dienes 119

4. Aldehydes 120

V. Poly functional reactions and non-linear polymers 121

VI. Natural polymers 123

1. The association theory versus the structural theory 123

2. Cellulose 124

3. Rubber 127

VII. The physical properties of high polymers 129

1. Solubility and colloidal properties 130

2. Crystallinity 132

3. Mechanical properties 134

This review is concerned with reactions that result in the combination of simple
molecules to form materials of high molecular weight, and especially with the nature of
such reactions and the structure and properties of the products. For the sake of com-
parison, the structures of certain natural polymers concerning which it is possible to
make definite statements are briefly considered also. The literature on the subject of
high polymers is most profuse, and from the theoretical side it has until very recently
been contradictory and confused. For these reasons the present discussion is of necessity
rather selective and critical. (For other general discussions of certain aspects of this
field the reader may consult references 1 to 7 and 18.) It is necessary also to define
especially the sense in which certain terms are to be used.

I. Definitions

1. Current Definitions. It is generally agreed by organic chemists (9)
that polymerization is chemical combination involving the operation of
primary valence forces, and that the term polymer should not be used

IX. POLYMERIZATION 83

(as it frequently is by physical and inorganic chemists) to name loose or
vaguely defined molecular aggregates. Beyond this, however, there is
not much agreement. The word polymer was introduced by Berzelius
(12) nearly a century ago to recognize the fact that two compounds may
have the same composition but different molecular weights, and he classi-
fied polymerism as a special type of isomerism. The accepted meaning of
these terms has subsequently undergone considerable change, but current
textbook definitions always state that a monomer and its polymer have
the same composition, and it is usually either stated or implied that the
process of polymerization is peculiar to unsaturated compounds and con-
sists in self-addition. But recently discovered facts show that these
conditions really are not satisfied in many instances to which the terms
polymer and polymerization are applied. The formula of a-polyoxy-
methylene, for example, is not (CH 2 O)*, but (CH 2 O)*H 2 O (110).

It is true also that many compounds which are not unsaturated are
capable of reacting with themselves to form products of high molecular
weight, and that such reactions are quite commonly called polymerization.
For example, benzyl chloride in the presence of aluminum chloride:

*C 6 H 5 CH 2 C1 -* (CyHs)* + *HC1

One can assume that the first step in this reaction is the elimination of
HC1 to yield the radical C6H 6 CH = , which, being unsaturated, is capable
of reacting with itself by addition. On the other hand, it may be that this
is simply a polymolecular Friedel Crafts reaction, the first step being the
formation of C6H5CH 2 C6H 4 CH 2 C1, which is capable of reacting with itself
by the same mechanism and finally yielding a very long chain in which the
chlorine content is negligible. If the usual definitions are adopted, this
reaction can properly be called polymerization only if it proceeds by the
first mechanism. Curiously enough, one actually finds in the literature
cases in which the fact that a reaction is called polymerization is used to
prove that the first step must be the formation of an unsaturated inter-
mediate capable of reaction with itself by addition. It is perhaps un-
necessary to say that the questions of the composition of high polymers
and the mechanism of their formation are frequently rather intricate and
cannot be solved in advance by definition.

2. Proposed Definitions. It is more practical and useful (and also
more consistent with actual usage) to define polymerization as any chemical
combination of a number of similar molecules to form a single molecule.
A polymer then will be any compound that can be formed by this process
or degraded by the reverse process: formaldehyde can be regenerated

84 POLYMERIZATION AND RING FORMATION

from polyoxymethylene by the action of heat; ethylene glycol can be ob-
tained from polyethylene glycol by hydrolysis; cellulose can be hydrolyzed
to glucose; and rubber can be formed by the reaction of isoprene with
itself.

3. Linear and Non-linear Polymers. The simplest and perhaps the
most numerous and important high polymers are characterized structurally
by the fact that their molecules are long chains built up from a repeating
radical or unit. This type of structure may be represented by the general
formula,

AAAAAAAAA
..... XV XV XV XV XV XV XV XV XV ' .....

The repeating radical A is called the structural unit. The number
and nature of these units determine the nature of the molecule. To
complete this formula it is necessary to specify the disposition of the
terminal valences; they may conceivably be mutually joined to form a
cyclic structure or saturated by univalent groups. Whether the chain is
open or closed, polymers of this class will be called linear polymers. It
should be added that linear high polymers are usually mixtures containing
chains of different lengths.

If the polymer is derived from a single compound the structural units
of the chain will in general be identical, but the mutual reaction of two or
more compounds that are chemically similar but not identical may lead
to chains made up from two or more different units. Such products will
be called linear mixed polymers.* Many polypep tides and proteins be-
long to this class. Non-linear polymers also exist; they can be formed,
for example, by the cross linking of long chains into two- or three-dimen-
sional structures.

4. Types of Compounds Capable of Polymerizing. The step-by-step
synthesis of long molecular chains containing a repeating unit is illustrated
by Fischer's famous synthesis of polypeptides. Reactions of polymeriza-
tion, however, lead to the formation of polymeric chains in a single opera-
tion. The capacity for self -combination of this kind is found among simple
molecules of three different types: (a) unsaturated compounds; (b) cyclic
compounds; (c) polyfunctional compounds. These are illustrated by the
following examples.

(a) nCH2=O + H 2 O - HO CH 2 O CH 2 O . . . CH 2 O CH 2 O CH 2 O H

Formaldehyde a-Polyoxymethylene

(b) n[CH 2 CH 2 O] + H 2 O -> HO CH 2 CH 2 O CH 2 CH 2 O . . . CH 2 CH,O H

Ethylene oxide Polyethylene glycol

* The term heteropolymer has also been suggested (23).

IX. POLYMERIZATION 85

HO CH 2 CH 2 O CH 2 CH 2 O . . . CH 2 CH 2 O H*
Polyethylene glycol

In each of these examples the product molecule is made up of a repeat-
ing series of identical units. Each unit corresponds with one molecule
of the starting material. The latter, therefore, is called the monomer
(mer = part). Ethylene glycol and ethylene oxide are both monomers
since, although they differ in structure and composition, their molecules
contain one CH 2 CH 2 O unit each. Similarly diethylene glycol,
HO CH 2 CH 2 CH 2 CH 2 H, is a dimer.

5. Types of Polymerization. Molecules can combine to form larger
molecules either by addition or by condensation, f and two corresponding
types of polymerization may be recognized. The formation of polyoxy-
methylene from formaldehyde is addition or A polymerization. The
formation of polyethylene glycol directly from ethylene glycol is condensa-
tion or C polymerization.** This second type of polymerization, although
its existence as a general phenomenon has only recently been recognized
(18), is especially simple both practically and theoretically since it involves
the known behavior of typical functional groups.

6. Condensation Polymerizations and Bifunctional Reactions. A
general class of condensation polymerizations is represented by the equa-
tion

x R y H- x R y + x R y -f -* x R z~R z R . . .

In the formula x R y, R is a bivalent radical and x and y are functional
groups capable of reacting with each other to form the known functional
group z. Thus if x is HO and y is COOH, z will be CO O. The com-
pounds x R y are called bifunctional compounds and their reactions,
bifunctional reactions. Reactions of the type, x R x + y R' y ""*
product, may be included in this class.

* The direct polyintermolecular dehydration of ethylene glycol to polyethylene
glycol not involving the intermediate formation of ethylene oxide is perhaps somewhat
hypothetical. It is used here merely as a formal illustration.

f The term condensation is used here to name any reaction that occurs with the
formation of a new bond between atoms not already joined and proceeds with the
elimination of elements (hydrogen, nitrogen, etc.) or of simple compounds (water, ethyl
alcohol, ammonia, sodium bromide, etc.). Examples are the Wurtz reaction, Friedel-
Crafts reaction, esterification, etc. See Kempf in Houben-Weyl, ref. 6, Volume II, p.
717.

** The term condensation polymerization was at one time applied by Staudinger
(38) to addition polymerizations that involve the migration or displacement of a group
or an atom, but its use in this sense has never become general. In view of its obvious
propriety for the purpose, it was adopted by the writer (18) as a name for polyinter-
molecular condensation.

86 POLYMERIZATION AND RING FORMATION

Bifunctional reactions present the possibility of following various
courses, (a) They may be intramolecular at the first stage. The product
will then be the cyclic monomer

(i) Reaction may be intermolecular at the first stage and intramolecular
at some subsequent stage. The product will then be a cyclic polymer

[( R Z )].

(c) Reaction may be exclusively intermolecular. In this case the product
will be an open chain of the type x R z R z .... R z R y.
These possibilities may be illustrated by the hydroxy acids of the series
H O (CH 2 ) W CO OH. The self-esterification of these might lead
to the simple lactones

[O (CH,) n CO],

or to cyclic polyesters

f[0 (CH 2 ). CO] P ]

or to open chain polyesters

H[O (CH 2 ) n CO] P OH.
All three of these possibilities can be realized.

II. Condensation Polymerization

1. Polyesters. Bifunctional esterifications are especially suitable for
the study of condensation polymerization, because esterification is a re-
versible reaction, and it is entirely free from side reactions under conditions
that are easy to realize.

a. The self-esterification of hydroxy acids. It is well known that
hydroxy acids react with themselves to form cyclic esters when there is the
possibility of forming a 5- or a 6-membered ring: 7- and 5-hydroxy acids
lead to the corresponding lactones, and a-hydroxy acids yield the cyclic
dimers. Thus glycolic acid reacts with itself to form glycolide:*

2HO CH 2 COOH -* [O CH 2 CO O CH 2 CO ] + 2H 2 O

The 7-lactones are stable substances, but the 5-lactones and glycolide
and its analogs on being heated are rapidly transformed into polyesters of

* But it is by no means certain that the glycolide is ever formed through the steps
glycolic acid -* glycolylgly colic acid -* glycolide. Dietzel and Krug (19) have shown
that in the self-esterification of lactic acid the primary steps are: lactic acid -*
lactyllactic acid -* polylactyllactic acids. Lactide when it is formed results from the
depolymerization of these polylactyllactic acids.

IX. POLYMERIZATION 87

high molecular weight. This peculiar behavior of the 6-membered cyclic
esters is considered in more detail in a later paragraph.

No information is available concerning the behavior of the simplest
c-hydroxy acid, hydroxycaproic acid, but the corresponding bromo acid
when treated with sodium ethylate in alcohol solution gives a poor yield
of the seven-membered lactone (37). The chief product is an undistillable
material, undoubtedly polyester. c-Hydroxycaprylic acid, C 2 H 8 CH-
OH (CH 2 ) 4 COOH is partly converted into the corresponding seven-
membered lactone on being heated, but the next member of this series,
C2H5 CHOH (CH 2 )5 COOH, under the same conditions gives only an
undistillable residue (24).

Larger lactones containing rings of fourteen to eighteen atoms are known
(25), and also /3-lactones (4-membered rings) (26, 27). All of these are
fairly stable compounds, but none of them has ever been prepared directly
from the corresponding acid.*

The attempt to prepare self -esters from simple /3-hydroxy acids results
merely in dehydration to the unsaturated acid, but in hydroxypivalic
acid, HO CH 2 C(CH 3 ) 2 COOH, this usual behavior is impossible
because of the absence of any a-hydrogen. The self-esterification of this
acid has been studied by Blaise and Marcilly (28). It occurs at 200C.
or at lower temperatures in the presence of certain catalysts. The product
is a microcrystalline powder insoluble in most organic solvents. It forms
sodium salts insoluble in cold water, and it is readily hydrolyzed to hy-
droxypivalic acid. These facts clearly indicate a structure of the type

HO CH 2 C(CH 3 ) 2 COO[ CH 2 C(CH 3 ) 2 COO ] CH 2 C(CH 3 ) 2 COOH

and on the basis of molecular weight and analytical data Blaise and Mar-
cilly assign to n the value 4.

In regard to the higher o>-hydroxy acids, Chuit and Hausser (29) have
recently synthesized the entire series from HO (CH 2 ) 7 COOH to HO
(CH 2 ) 20 COOH. All these acids readily undergo self-esterification on
being heated. The products are not the corresponding lactones (which
in several cases are known), but solid acidic materials whose properties are

* Kerschbaum (36) records various attempts to prepare the lactone of ambrettolic
acid (hexadecene-7-ol-16-acid-l) from the acid obtained by hydrolyzing the naturally
occurring lactone. By heating the acid he obtained a small amount of oil which, from
its odor, was inferred to contain lactone. The chief product, however, was a non-
volatile material soluble in alkali. Neither Chuit and Hausser (29) nor Lycan and
Adams (30) report any evidence for the presence of lactone in the self -esters prepared
from higher co-hydroxy acids. Unpublished experiments made in this laboratory have
failed to reveal the presence of any lactone among the products of the dehydration of
co-hydroxypentadecanoic acid.

POLYMERIZATION AND RING FORMATION

TJ
8 1

o
a

CL>
T3

d cJ

"rt tn O5
.H > "3
w I-.

1
R

L fS

O "GO O

. V-.

3 o

t/5 r^

t 10

<H
d o

Oj o

I S

o p,

d % d

8 *

y ^

CJ Q^

fi ld

h c2

r? G

8 .S S

E^l d

CJ d

^ c

.S B

^ d

8,-i fe-2 0)

5* 3 l!-

r ^

fc fcjj|> 8

** o| * ""'

lrH

W .S <>_

|||2| *

*a%"*

> i to

W |J

3 1

H <*

VTESTERS FROM

Is O CO

)io O CO

I 2 CO (

> 6 CO (

u w

O HH

ffi

fe o

{

ffi

"rt

1 s

rrt

a<L>
jg

<L>

jj*

*>.

"rt

"85

4-*

4-J

IX. POLYMERIZATION

Physical properties
rocrystalline powd

r^ 77

o
a

;||
*!''

. j/. i-c

i, transparent resii
, transparent resin
, transparent gum
r viscous, transpare
rup

r--,rc.f1i;*m ^^r/4

[ 1

i 1

lisli

rocrystalline powd

r 70

. f . i ^
rocrystalline powd

3.S E

3 ISSr a i

2 a .S {

3 .a

3 it

11 ft 11 2

^O *f t^ b W

C<1 rH r-

*d rt ^,

gS|l> g

g g g g ^

3 8

<ll g w

Ttl CO rH CJ O

5 S

W) t O 0^

00 O5 <N CD b

rH rH

^.s's" 11

? I

I CO

A A I O

996V

1 S

< O

w
w

lula of structural unit

>io CO (CH 2 )

li CO (CH 2 )

| 2 O CO C 6 H 4 -
> 3 CO C 6 H 4 -
-O CO C 6 H 4 C
ho O CO C 6 H 4 -

k r\ r>r\ r*r\

) CO CO

O e?

o W

ffi" ffi" ^ ffi *"

? K

d. y

^0 C.

; u

00 C

I A

4~>

-M 4->
5 J 2 C

) Oj

73 ^

a)
,0

Sltl. -i

\ **

3 2

1 S

l"!ll !

g. Sx-x j

5 ^

flj r*> -f- -M
g r=3 0> OJ rt

? *2

S3 *- c c -*

r=J <U t3 C ;

_ r+

c/j cj

^s 'S

rQ .S g OJ

: S

S H W Q f-

H W

P-I

90 POLYMERIZATION AND RING FORMATION

consistent with the general formula HO R CO O R CO . . . .
O R COOH. Similar products obtained by the self-esterification of
higher hydroxy acids have also been investigated by Bougalt and Bourdier
(31) and by Lycan and Adams (30).

The presence of a terminal carboxyl group in each of these products
proves that the ester formation is exclusively intermolecular. It should
be possible, therefore, by regulating the degree of completeness of the
esterification to obtain molecules of various lengths and, in particular, to
obtain exceedingly long molecules. Such molecules are of especial interest
in connection with the study of high polymers, since the nature of the acid
used absolutely determines the structural unit of the polyester.

b. Polyesters from dibasic acids and glycols. The study of bifunctional
esterifications from this standpoint was first undertaken by the writer
and his collaborators (32 to 35). Dibasic acids and glycols were used as
starting materials since these are more readily accessible than hydroxy
acids, and they also permit more numerous structural variations in the
ester product.

Esters derived from dibasic acids and glycols can be prepared by any of
the typical reactions ordinarily used in the preparation of simple esters,
e. g.:

(a) HO ROH + HOOC R' COOH -* O R O CO R' CO + H 8 O

(b) HO ROH + EtOOC R' COOEt -* O R O CO R' CO -f EtOH

C 6 H 6 OH

(d) HO ROH + C1CO R' COC1 -* O R O CO R' CO + HC1

(e) Br R Br + AgOCO R' COOAg -+ O R O CO R' CO + AgBr

In these equations O R O CO---R' CO represents the struc-
tural unit of the product. The studies have proved that the nature of the
ester is completely determined by the number of atoms in the chain of the
unit: if this number is five, the product is monomeric and cyclic (i. e., it
contains only one unit) ; if the number is six, the product can be obtained
in both monomeric and polymeric forms and these are interconvertible;
if the number is more than six, the product is exclusively polymeric.

The preparation of an ester of this last class may be illustrated by the
action of succinic acid on ethylene glycol.* Succinic acid mixed with a

* The preparation of ethylene succinate by reaction between the acid and glycol
was first carried out by Lourenco (17). Later Davidoff (40) and Vorlander (41) pre-
pared it by various methods, and the latter investigator assigned to it the formula of
a 16-membered cyclic dimer. His molecular weight determinations were apparently in
error. The dimeric ester is now known (42, 34) and its properties are quite different
from those of Vorlander's ester, which closely resembles the polymeric ester prepared
from the acid and the glycol.

IX. POLYMERIZATION 91

slight excess of glycol is heated in a distilling flask. At about 160-175C.
rapid esterification sets in, accompanied by the liberation of water. When
no more water is evolved the temperature is raised to 220-250 C. and the
pressure is reduced below 1 mm., whereupon most of the excess glycol dis-
tils. The ester product is completely non-volatile and remains in the dis-
tilling 1 flask in the form of an exceedingly viscous liquid. When this is
cooled it solidifies to a hard, brittle, opaque, white mass. This can be
dissolved in cold chloroform and precipitated as a granular powder by the
addition of ether or benzene. It can be recrystallized from hot ethyl
acetate or a large volume of hot acetone. The recrystallized ester melts at
about 102C. A product having the same physical properties can be
prepared in the same way by the action of diethyl succinate or diphenyl
succinate on ethylene glycol. The reaction then consists in ester inter-
change and the volatile product liberated is alcohol or phenol.

By the method of direct esterification or by ester interchange the esters
listed in Table I have been prepared. They are without exception highly
polymeric, the molecular weights being on the average in the neighborhood
of 8000. Molecular weights of several of the esters have been determined
in a variety of solvents and by both freezing and boiling point methods,
and within the limits of experimental error the same values have been
found.

These esters show considerable similarity in their physical properties.
They are completely non-volatile and all of them dissolve quite readily in
cold chloroform. Those derived from the higher glycols or acids also dis-
solve in cold benzene, and in the solid state they are less hard and more
wax-like than ethylene succinate. In spite of their high molecular weight
and lack of complete homogeneity all these esters except the phthalates
are crystalline. The melting points vary slightly from one preparation to
another of a given ester and usually cover a range of about 2.

The ethylene succinate described above is neutral, and its composition
and molecular weight indicate the average formula*

HO [(CH 2 ) 2 O CO (CH 2 ) 2 CO O ] 22 (CH 2 ) 2 OH,

* The molecules of this product certainly do not all have the same length, but on
the other hand it appears to be much more nearly homogeneous than any products ever
obtained by addition polymerization (polystyrene, etc.). In its qualitative solubility
behavior it resembles a chemical individual, and fairly elaborate fractional crystalliza-
tions have failed to separate it into portions showing any considerable difference in their
properties. The reason for its relatively great homogeneity lies in the conditions of its
formation, which are such as to force much smaller molecules to react with themselves,
but are not sufficiently drastic to cause the formation of much larger molecules. Under
other conditions, polymeric mixtures having a wide range of molecular weights can bo
obtained.

92 POLYMERIZATION AND RING FORMATION

i. e. y it is an open chain derived from twenty-two molecules of acid and
twenty-three molecules of glycol.

The hydroxyls at the end of this chain do not readily react with the
usual reagents for alcohol groups; their sluggish behavior in this respect
is characteristic of many high molecular weight materials. However, on
being heated to a fairly high temperature with ^-bromobenzoic anhydride
the ester yields a di--bromobenzoyl derivative identified as such by its
bromine content. The ester also reacts with molten succinic anhydride
and yields an acidic ester whose neutral equivalent and observed molecular
weight agree with the formula I d.

(I) HOCO (CH 2 ) 2 CO [ O (CH 2 ) 2 O CO (CH 2 ) 2 CO ] x OH
I a, x = 6; I b, x = 9; I c, x = 12; I d, x 23.

Similar acidic esters of lower molecular weight have also been obtained by
partial esterification of glycol with an excess of succinic acid and fractional
crystallization of the product. The observed molecular weights and
neutral equivalents of the fractions corresponding to various values of x
in formula I are indicated in Table II.

TABLE II
ACIDIC ETHYLENE SUCCINATES

Molecular weight

Sodium salt

Forrfiula
la

Melting
point
C.

Calculated
from
neutraliza-
tion
equivalent

1020
1340
1800
3400

Found by
ebulli-
oscopy

1070

1380
1580
3110

Melting
point
C.

91
97
100
109

Molecular
weight
Calculated
from
sodium
content

1030
1460

2010
3740

I b

1 c

The presence of these terminal groups proves that the esterification of
succinic acid by ethylene glycol is intermolecular at every stage.* The
reaction evidently involves a series of condensations resulting in the pro-
duction of ester molecules of progressively greater length. The first prod-
uct might be

HO(CH 2 ) 2 OCO(CH 2 ) 2 COOH,
and the second

HO(CH 2 ) 2 OCO(CH 2 ) 2 COO(CH 2 ) 2 OCO(CH 2 ) 2 COOH.

* Lycan and Adams (30) also present convincing evidence of the open chain struc-
tures of polyesters derived from w-hydroxydecanoic acid. They isolated fractions
whose molecular weights estimated by titration with alkali ranged from 1000 to nearly
9000. All of these fractions formed potassium salts which were completely soluble in
warm water.

IX. POLYMERIZATION 93

Moreover, since both glycols and dibasic acids have a tendency to esterify
at both ends simultaneously, it is reasonable to suppose that similar chains
will be present, some of which are terminated at both ends by carboxyl
and others by hydroxyl. It is evident that an exceedingly large number of
species may be involved in the ester equilibrium.* But as the amount of
water participating in the equilibrium is diminished by its constant removal,
the smaller molecules are forced to couple with each other until finally
practically none of them remains. A further simplification of the kinds
of molecular species present in the product is effected by using an excess of
glycol, since this makes all the terminal groups alike. For this purpose, it
makes no difference how large an excess of glycol is used , provided the reac-
tion mixture is finally heated for some time in a high vacuum, since under
these conditions the reaction can be propagated by ester interchange.
Thus it is possible to isolate the monomeric ester, foV/3-hydroxyethyl-
succinate,

HO (CH 2 ) 2 O CO (CH- 2 ) 2 CO O (CH 2 ) 2 OH,

and this on being heated to 200C. in a vacuum loses glycol and is converted
into the neutral polyethylene succinate of molecular weight 3000 already
described. In view of this fact it is obvious that the accidental mutila-
tion of the terminal groups (e. g., the loss of OH or CC^) cannot in itself
prohibit the progress of the coupling; the reaction can progress by ester
interchange involving the last ester linkage.

Apparently the ultimate factors that set the attainable limit on the
length of polyester molecules are purely physical. As the molecular weight
becomes greater, the reaction product becomes more viscous and the rate
of diffusion of the volatile product (water or alcohol) to the surface becomes
slower and slower, f Moreover as the reaction progresses, the mobility
of the reacting molecules diminishes and the relative concentration of the
reactive groups becomes smaller. All of these factors affect the rate and
some of them affect the position of equilibrium, but the molecular weight
at which a practical limit is reached will depend upon the temperature and
pressure, the area and thickness of the reacting mass, etc., and by a suit-
able adjustment of these factors it is possible to obtain polyesters having
very much higher molecular weights than those listed in Table I (51).

2. Bifunctional Wurtz Reactions and Friedel Crafts Reactions. The
action of sodium on polymethylene bromides, Br(CH2)*Br, leads in certain

* The kinetics of the reaction between phthalic anhydride and ethylene glycol have
been studied by Kienle and Hovey (43).

t At least in the crystalline state polyesters adsorb water and hold it very tena-
ciously.

94 POLYMERIZATION AND RING FORMATION

cases to the formation of cyclic hydrocarbons. Cyclopropane, cyclopen-
tane, and cyclohexane have been prepared in this way. On the other
hand, the action of sodium on decamethylene bromide does not yield any
cyclic hydrocarbons.

In the presence of absolute ether this reaction proceeds very smoothly
(44, 45). The product, which is for the most part insoluble in ether, con-
sists of a complex mixture formed by the coupling of various numbers of
molecules of the halide with each other, e. g.,

Br(CH 2 ), Br + 2Na + Br(CH 2 )ioBr + 2Na + Br(CH 2 ) 10 Br ->
Br(CH 2 ) 30 Br + 4NaBr.

At the same time, owing to the participation of the ether in the reaction,
most of the terminal bromine atoms are replaced by hydrogen.* This
reduction of the terminal groups can finally be carried to completion by the
action of sodium in boiling butyl ether, and the product then consists of a
mixture having the general formula

H f(CH 8 )io]* H

in which the values of x range from 1 to at least 10. The various members
of this mixture up to and including w-C 7 oHi 42 can be separated and isolated
in a state of purity by fractional evaporation in a molecular still followed
by crystallization. The identity of these fractions is established by their
melting points and their x-ray diffraction patterns. There is no evidence
of the presence of any materials in the mixture not belonging to the series

H [(CH 2 ) 10 ]* H

About 25 per cent, of the total product consists of members standing above
C7oHi42. These are not capable of being distilled, and they cannot be
separated from each other by fractional crystallization. The average
molecular weight of this material indicates that it must contain hydro-
carbons at least as high as CiooH 2 o2.

The following results of unpublished studies made by Dr. R. A. Jacob-
son in this laboratory are more or less closely related to those described
above. The action of metallic sodium on ^-dibromobenzene in absolute
ether leads to a product corresponding in composition to the formula
Br CeH 4 (CeH^e CeH 4 Br. When the reaction is carried out in boiling
toluene the product formed corresponds approximately in composition and

* In its mechanism the Wurtz reaction undoubtedly involves as its first step the
formation of the sodium compound RNa (46). This normally couples with another
molecule of the halide, but if ether is present it may be destroyed by the reaction
RNa + (C ? H 6 ) 2 Q -* H + C 2 H 4 + C 2 H 6 ONa (47).

IX. POLYMERIZATION 95

molecular weight to the formula Br C 6 H 4 (CeH 4 )i 2 CeH 4 Br. Both of
these products are readily soluble in benzene. ^-Xylylene bromide,
BrCH 2 C 6 H 4 CH 2 Br, when treated with sodium in hot toluene, yields a
very insoluble hydrocarbon which does not melt below 350C. Apparently*
a very large number of ^-xylylene units participate in this coupling.

Benzyl chloride in the presence of aluminum chloride or ferric chloride
(see 10) reacts with itself to form resins of high molecular weight. This is
evidently a bifunctional Friedel Crafts reaction involving the progressive
coupling of successively longer chains with the elimination of hydrogen
chloride. Depending upon the conditions of their formation these resins
are fusible and soluble, or infusible and insoluble. Molecular weight values
indicate that the fusible resins are formed from 14 to 25 molecules of benzyl
chloride. The infusible resins are no doubt much more highly polymeric.
Benzyl fluoride in the presence of a trace of acid reacts very vigorously
with itself in a similar manner (11).

3. Other Bifunctional Reactions. The number of possible types of
condensation polymers is practically unlimited. Although very few of
these possibilities have received any considerable study, the following ex-
amples at least indicate that the formation of such polymers is the usual
result of bifunctional reactions when structural features preclude the
formation of a 5- or a 6-membered ring.

a. Polyamides. The acids NH 2 (CH 2 ) 3 COOH and NH 2 (CH 2 ) 4 COOH
are readily dehydrated to the corresponding monomeric lac tarns (48).
There is no record of either of these being caused to polymerize. The
next higher member yields two products (49, 50a). One of these (20-30
per cent.) is the lactam (7-membered ring), a distillable crystalline mate-
rial; the other is a poly amide, an undistillable hard, waxy material insol-
uble in most solvents except concentrated hydrochloric acid, phenol, and
hot formamide. It can be hydrolyzed quantitatively to the amino acid.
Molecular weight determinations indicate the presence of at least ten
structural units in its molecule. The formation of this polymer is due
to direct intermolecular condensation and the formation of the lactam is
an independent reaction, for the latter cannot be polymerized under the
conditions that lead to the production of the former. The acid NH 2 (CH 2 ) 6 -
COOH on being dehydrated yields exclusively a product that is polymeric
(50), although the corresponding monomeric lactam has been prepared by
another method and is a stable substance (164). The acid NH 2 (CH 2 )i -
COOH also yields only poly amide (51).

A number of polyamides have been prepared by Dr. J. E. Kirby in this
laboratory by the action of dibasic acids on aliphatic diamines. These

96 POLYMERIZATION AND RING FORMATION

materials are all much less soluble, and when crystalline have much higher
melting points than the analogous polyesters (51).

b. Polyamines. v. Braun observed (52) that the compound NH 2 -
(CH 2 )eCl, unlike its immediate lower homologs, when it reacts with itself
yields only a very small amount of the volatile cyclic base. The chief
product is an undistillable, waxy solid, but this has never been studied in
detail.

The formation of quaternary ammonium salts from various compounds
of the series (CH 3 )2N(CH 2 ) n Br has been studied (53). The products have
been assumed to be cyclic monomers or dimers, but it still remains to be
demonstrated that none of them are linear polymers.

c. Polyacetals. The reaction between glycols and acetaldehyde (or
acetylene) presents the possibility of forming cyclic acetals,

/\
CH 3 CH R

or polyacetals

HO R O CH O R O CH O R O CH O . . .

I I I

CH 3 CH 3 CH 3

This reaction has been studied by Hill and Hibbert (8). Ethylene and
trimethylene glycols gave in excellent yields the cyclic acetals which are
5- and 6-atom rings. Tetramethylene glycol gave in poor yield a volatile
compound which was apparently the monomeric cyclic acetal containing a
7-atom ring. A considerable part of the product was an undistillable sirup.
The products from octamethylene and decamethylene glycols were also
undistillable sirups. No molecular weight determinations are recorded,
but it may be assumed that the undistillable products are polymeric. An
analogous compound prepared in this laboratory (51) from benzaldehyde
and diethylene glycol had an apparent molecular weight of about 1370.

d. Poly anhydrides. Dibasic acids of the series HOOC(CH 2 )*COOH
are readily converted into the corresponding anhydrides. Malonic an-
hydride, the first member of the series, is polymeric (54). The next two
members, succinic and glutaric anhydrides, are known only as the mono-
meric 5- and 6-membered rings. The anhydrides of all the higher acids
are polymeric.* The monomeric anhydrides are macrocrystalline, readily

* The known examples are anhydrides of adipic, pimelic, suberic, azelaic, sebacic
(55) and hexadecamethylene dicarboxylic (51) acids. These if monomeric would be
respectively, 7-, 8-, 9-, 10-, 11-, and 19-membered rings. Adipic anhydride can be ob-
tained in both a monomeric and a polymeric form (56).

IX. POLYMERIZATION 97

distillable solids; the polymeric anhydrides are non-volatile microcrystal-
line powders or waxes, and they are less soluble than the monomeric an-
hydrides. Both types are very reactive, but qualitatively their chemical
behavior is not the same. The monomers react with aniline to give pure
monoanilide; the polymers give a mixture of acid, monoanilide, and di-
anilide in the ratio 1:2:1. This is precisely in accordance with the cal-
culated behavior of a very long chain having the general structure.

CO R CO O CO R CO O CO R CO O , etc.

e. Grignard reactions. Bifunctional Grignard reagents such as BrMg-
(CH 2 )6MgBr are capable of reacting with bifunctional reactants (dialde-
hydes, dike tones, simple esters, etc.). The products may be rings or long
chains. Several 5- and 6-atom rings have been prepared in this way, and
in poor yield one 7-atom ring (57). In this laboratory it has been found
(51) that decamethylene dimagnesium bromide reacts readily with methyl
formate :

. . . (CH 2 ) 10 MgBr + O=-D OCH 3 + BrMg . . . -*

. . . [ (CH*),- CH(OMgBr) ] x . . .

The final product is a microcrystalline solid readily soluble in various
organic solvents and melting at about 120C. Its chemical and analytical
behavior shows that it is the expected linear poly alcohol containing about
eight of the structural units (CH2) 10 CHOH .

In the formation of bifunctional Grignard reagents some coupling always
occurs: dibromides of the formula Br(CH 2 ) w Br yield considerable amounts
of BrMg(CH 2 ) 2n MgBr and progressively smaller amounts of higher cou-
pling products. It is a curious fact that this reaction does not occur intra-
molecularly even in the case of pen tame thy lene bromide where it would
lead to the formation of a 5-membered ring (154).

The coupling of simple Grignard reagents can be effected by the action
of iodine, and this method has been applied by Grignard and Tcheoufaki
to acetylene dimagnesium bromide (155). Products having the following
formulas were isolated :

HC==C Cs=CI; HC^C C^C -CssC I;

and also a form of carbon, which probably resulted from a continuation of
the initial reaction in the same sense.

The attempt to prepare a Grignard reagent from ^-xylylene bromide,
BrCH 2 C 6 H4CH 2 Br, leads (51) to an insoluble hydrocarbon having the
composition

98 POLYMERIZATION AND RING FORMATION

( CH 2 C 6 H 4 CH 2 ),.

The value of x is probably very large.

/. Sulfur and selenium compounds. The action of sodium sulfide on
alkylene halides of the formula X(CHa) w X leads to considerable yields of
the expected cyclic products (CH2) W >S only when these are 5- or 6-mem-
bered rings (162). Ethylene bromide gives white amorphous insoluble
products (161); when prepared under certain conditions these are capable
of being depolymerized by heat to yield the cyclic dimer, diethylene di-
sulfide. Monomeric ethylene sulfide has been prepared by Delepine (86)
by treating ethylene thiocyanate with sodium sulfide. It polymerizes
spontaneously on standing. The chief product of the action of sodium
sulfide on trimethylene halides is an amorphous polymer, although a small
amount of the cyclic trimethylene sulfide can be obtained under certain
conditions (87). Hexamethylene iodide also yields an amorphous polymer
as the chief product (163, 87). v. Braun suggests that its formula is
probably

I(CH 2 VS-(CH 2 ) fl . . . .S(CH,)J.

but no estimates of its molecular weight have been presented.

The amorphous material formed from ethylene chloride and sodium
sulfide is capable of being molded into a product which is very resistant to
solvents and somewhat resembles rubber in its properties. One may infer
that its molecular weight is very high. In recent patents (159) it is claimed
that similar products can be obtained from alkylene halides generally.

Morgan and Burstall (160) state that the interaction of sodium selenide
and the requisite alkylene dibromide leads readily to the production in good
yields of cyclic seleno hydrocarbons containing 5- or 6-membered rings,
but trimethylene bromide and sodium selenide give little cycloselenopro-
pane. The main product is a sixfold polymer melting at 38-40 C.

g. Miscellaneous. A curious spontaneous progressive coupling is found
in the action of water on methyl orthosilicate. The reaction may be
formulated as follows :

OCH 8 OCH 8

CH 3 O Si OCH 3 -f H 2 O -* CH 3 O Si OH
OCH 3 OCH 3

OCH S OCH 3 OCH 3

2CH 3 OvSi OH -* CH 3 O- Si OSi OCH S + H 2 O

OCH 3 OCH S OCH 8

IX. POLYMERIZATION 99

Partial hydrolysis and coupling of this dimeric product lead to a tetramer,
and so on. By suitably adjusting the initial ratio of water to ester it is
possible to obtain samples representing various degrees of polymerization.
The lower members have been isolated as chemical individuals by fractional
distillation (156). The intermediate members are viscous undistillable
liquids soluble in organic solvents. Solutions of these polymers are used
in the preparation of paints having unusual properties (60). Films pre-
pared from these paints harden by a continuation of the initial progressive
hydrolysis and coupling, which yields as the final product pure silica. This
synthesis incidentally furnishes an elegant proof of the three-dimensional
polymeric structure of silica. The OCH 3 groups are responsible for the
coupling, which results in the formation of long chains from the mono-
meric ester.

OCH 8 OCH 3 OCH 3 OCH 8 OCH 3

^i__ -Si-0~Si-0-Si-0-Si-0~

I 1 ^ ^ f ^

OCH 3 OCH 3 OCH 3 OCH 3 OCH 3

Since each silicon atom in these chains still bears two OCH 3 groups, the
coupling can continue in the other dimensions, yielding finally a 3-dimen-
sional lattice in which each silicon atom is joined (through oxygen) to four
others.

Metallic sodium acts on compounds of the type R2SiCl2 yielding chain
structures composed of SiR2 SiR2 groups (157). These products are
for the most part amorphous and glue-like.

The specially pronounced tendency toward the formation of 5- and 6-
membered rings is frequently manifested in complex coordination com-
pounds (148) and complex formation frequently fails when it would involve
the formation of larger rings. Ethylene and trimethylene diamine yield
crystalline nickelotriene succinimide compounds of the formula [Nien 3 ]-
Sucj and [Nitrs] Suc2, but the compounds from tetra- and penta-methylene
diamines are amorphous and, according to Tschugaeff (114), they must be
represented as chains of unknown length:

Sue Sue

H 2 N(CH 2 )NH 2 . . .Ni. . .H 2 N(CH 2 ) n NH 2 . . .Ni. . .
I I

Sue Sue

Bifunctional couplings probably occur in many reactions of oxidation
and reduction, Busch and Schmidt h^ve shown (22) that the catalytic

100 POLYMERIZATION AND RING FORMATION

reduction of aryl halides under some conditions proceeds to a considerable
extent as follows :

2C 6 H 6 Br + 2H -* C 6 H 6 C 6 H 6 + 2HBr

When the same reaction is applied to -diiodobenzene, terphenyl, CeHb'-
CeH^CeHs, and ,'-diphenyldiphenyl are obtained.

A preliminary stage in the oxidation of aniline to aniline black is the
formation of long chains of the following type (20),

In a similar way the oxidation of phenols may lead to the formation of
polyphenylene ethers (21). Substituted polyphenylene ethers of high
molecular weight are obtained by the condensation polymerization of silver
salts of halogenated phenols (173).

4. Stereochemical Factors Involved in Condensation Polymerization.
Bifunctional reactions always present the possibility of following two
courses they may lead to ring closure, or to progressive coupling but
the very numerous studies of such reactions have been concerned almost
exclusively with ring closure. The possibility of progressive coupling has
frequently been ignored or rejected as unlikely, and highly improbable
cyclic structures have occasionally been assigned to products on the
ground merely that they were the result of bifunctional reactions. In
other cases the actual products have been discarded simply because they
were obviously not the expected cyclic products. This attitude has
certainly been partly responsible for the comparative meagerness of the
literature on condensation polymers. The examples cited in the previous
paragraphs show, however, that bifunctional reactions may proceed in a
strictly normal fashion by progressive coupling and result in the formation
of large molecules, and that, considering any particular homologous series,
this type of reaction is the rule and ring closure is the exception.

This is not an especially astonishing conclusion. Intermolecular reac-
tion is a perfectly general phenomenon; intramolecular reaction is a pecu-
liar and special kind of happening. This is immediately evident from the
consideration that if two groups are to react they must meet, i. e., they
must approach each other very closely. If the groups are not present in
the same molecules, such approach is always possible as long as the mole-
cules are free to move about. But two groups, x and y, that are present at
opposite ends of a single molecule are capable of approaching each other
only if the architecture of the molecule permits, and even if such approach
is permissible it does not necessarily follow that it will occur, or if it occurs

IX. POLYMERIZATION 101

that it will be effective. Meanwhile the x groups of this molecule are
continually colliding with y groups of other similar molecules. From this
standpoint progressive coupling in bifunctional reactions is inherently more
probable, generally speaking, than ring closure, and the latter, if it occurs
to the exclusion of the former, must be peculiarly favored by some special
factors.

The reasons for the great difference in the relative ease with which dif-
ferent types of rings are formed has been the subject of much speculation,
and two factors that have frequently been assumed to be of self-evident
importance are the relative energy content of the rings and their stability
or degree of strain. There is, however, no theoretical justification for this
assumption (58), and the facts show that neither of these factors can pos-
sibly be decisive. Thus, the action of sodium on propylene bromide yields
cyclopropane, not cyclohexane, although the latter has a much lower energy
content per unit and is less highly strained than the former.

On the other hand spatial relations in the reacting compound must of
necessity have a great influence on the possibility of ring formation. The
atoms of the benzene ring, for example, and all the atoms joined directly
to it lie in the same plane (59), and to link the two para positions of the ring
together through a chain of less than four or five accessory atoms would
require a very improbable degree of distortion. Many attempts have
been made to bring about such linkings, but the supposedly successful
examples have proved on reexamination to be fictitious (62). The same
thing is true (61) of alleged examples of ring closure through the p and p r
positions of diphenyl.

The stereochemistry of simple aliphatic chains is a rather complicated
matter. Ruzicka's discovery that large aliphatic rings are no less stable
than small ones is impossible to reconcile with the Baeyer strain theory,
and the latter has been superseded by the Sachse-Mohr theory (66) which
permits the existence of large rings in non-planar and strainless forms and
is supported by a great deal of other evidence besides. The essential
assumption in the Sachse-Mohr theory is simply the usual one that the
four valences of the carbon atom are directed toward the corners of a
tetrahedron and that there is free rotation about each single bond in a
chain. Space models embodying these features can easily be made by
joining small wire tetrahedra with rubber tubing in such a way that the
arms that are being connected overlap. Rings constructed with such mod-
els show graphically that a 5-atom ring is uniplanar and free from strain;
a 6-atom ring is highly strained unless two of the atoms are allowed to lie
in a different plane from the other four; larger rings if they are to exist

102 POLYMERIZATION AND RING FORMATION

must be multiplanar and strainless, and they possess a mobility which
makes it possible for them to assume a great multiplicity of shapes.*

The variety of possible configurations of an open chain is much greater
than in a ring and it increases rapidly with the length of the chain. In
the crystals of fatty acids and paraffins the chains probably are rigidly ex-
tended into a zigzag structure (13); but a model of such a chain is ex-
ceedingly flexible and mobile, and it seems very improbable that in the
liquid condition or in solution the molecules retain their rigidly extended
form. No doubt certain possible configurations are more probable than
others, but there is at present no means of knowing just what these are.f
It is clear, however, that the relative probability of close approach of the
ends of a chain diminishes very rapidly as the length of the chain increases.
Owing to the fewness of their separate points of rotation, 5- and 6-atom
chains can assume relatively few configurations. In fact, if a space model
of such a chain is supported at one of the bonds the entire structure can be
rotated in such a way that the freely moving and unsupported ends collide
at each rotation. This shows that there is a certain inevitability in the
closure of such chains. Longer chains on the other hand can assume an
extraordinary multiplicity of shapes without any close approach of the
ends. The ends of the model can be brought together arbitrarily without
any resistance, but as Mohr has pointed out (66) in the molecule itself one
is dependent upon random collision,** "which will bring about a given form
the more rarely the more forms are possible, i. e., the longer the chain."

One obvious and important implication of this theory is that a cyclic
structure for linear high polymers is very improbable. The formation of
such polymers usually depends upon the absence of any tendency toward
ring closure in the early stages of the coupling, and the probability of ring
closure will diminish as the length of the chain increases. This implica-

* Apparently the shape that they actually tend to assume involves the close ap-
proach of opposite sides of the ring; the ring consists essentially of two parallel chains
joined at the end (67, 144, 145).

f Attempts to decide this question by studies of x-ray diffraction (88), electric mo-
ment (63), ionization constant (65), and numerous other properties (143) have led to
the following rather contradictory conclusions: (a) the chains are straight zigzags;
(b) they are straight zigzags except for the 5-atom chain which is coiled ; (c) they have
a helicoidal configuration which brings the first and fifth atoms very close together and
introduces an anomaly in properties when the length of the chain is 5, 10 or 15 atoms.
None of these configurations provides any mechanism for the close intramolecular ap-
proach of the ends of chains longer than 6 atoms.

** It seems quite possible that this dependence upon random collision is not in-
herently necessary. It may be that some selective control over molecular form is possi-
ble by the use of the orienting effects at surfaces, and such factors may perhaps come in-
to play in the synthesis of large rings in nature. This suggests the possibility of some
very interesting studies of a novel kind in surface chemistry.

IX. POLYMERIZATION 103

tion is also in accord with the facts. It is true that cyclic formulas have
been assigned to various high polymers by Staudinger and by other investi-
gators, but none of these formulas has been established.* On the other
hand, in a variety of instances open chain structures have been experi-
mentally proved.

The presence of substituents and the nature of the terminal groups may
be expected to modify the stereochemical behavior of chains. Thus, if
two atoms of a chain are adjacent atoms of a benzene ring their position
with respect to each other is fixed, and the chances of intramolecular reac-
tion are greater than in an analogous simple chain of the same length.
(The latter will have one more axis of rotation than the former.) It is
not surprising therefore that the majority of the known 7- and 8-membered
rings have at least two of their atoms members of a benzene ring. The
compound NH 2 (CH 2 )6C1 reacts with itself intramolecularly only to a slight
extent, while with <^NH 2 C6H 4 (CH 2 ) 4 C1 ring closure is almost quantitative
(68). Other similar examples might be cited (50, 163).

There is also some evidence to indicate that even simple substituents
such as methyl groups on a chain may increase the tendency toward the
formation of larger rings. Researches in the diphenyl series (69) have
established that substituent groups suitably placed may completely inhibit
rotation about a nearby single bond, and a similar effect in aliphatic chains
is at least conceivable. Almost any restriction of the freedom of rotation
of the atoms of a chain would, on the basis of the Sachse-Mohr theory,
increase the chances of ring formation.

The influence of the nature of the terminal groups is seen in the fact
that co-hydroxy acids on dehydration give both lactone and polyester,
while attempts to prepare 7-atom cyclic esters from dibasic acids and di-
hydric alcohols yield only polyesters. Dilution also may be expected to

* In this connection merely negative evidence, e. g. t the failure to detect terminal
groups, is worthless. Terminal groups may become lost or mutilated, or they may be
present and yet fail to react. High polymers are frequently very sluggish in their
chemical behavior.

The tendency to resort to cyclic formulas does not aid in clarifying the problems of
high polymers. One can assume that the rubber molecule is an enormously large ring,
but rubber is very susceptible to degradation by oxygen, heat, mechanical action, etc.,
and the first step in such degradation must cause the ring to be ruptured. This product
then presents all those problems which the assumption of a cyclic structure was designed
to evade.

It seems evident that all polymerizations, whatever their mechanisms, must be
progressive stepwise reactions. The simultaneous combination of 100 molecules presents
insuperable kinetic difficulties. If this view is correct, ring closure can occur only as the
last step in the reaction chain. But at this stage the ends of the chain must be more
remote from each other than at any earlier stage, and the opportunities for ring closure
must be at a minimum.

104 POLYMERIZATION AND RING FORMATION

favor ring formation as compared with progressive coupling (70), but all
these effects are comparatively small.

a. Large rings. The question arises whether large rings are ever formed
as the normal primary products of bifunctional reactions. Meyer and
Jacobson in their textbook (147) devote over fifteen hundred pages to 5-
and 6-membered heterocyclic compounds and only 17 pages to all larger
rings. Most of the examples of this class of compounds that they accept
as authentic fall into one or more of the following types : the ring contains
only 7 or 8 atoms ; two or more atoms of the ring are members of a benzene
nucleus; the yields are very poor; the compound is formed by ring widen-
ing. There remain a few examples (e. g., the cyclic duplomercaptal de-
rived from acetone and pentamethylene dimercaptan) in which the evi-
dence for a very large cyclic structure is good ; but this evidence is no better
than that on which the 1 6-membered cyclic formula for ethylene succinate
was based (41), and ethylene succinate when reexamined (32, 34) proved to
be a linear polymer of high molecular weight.

Concerning the identity of large rings described by Ruzicka, it is not
possible to entertain any doubt at all, but the methods required for the
preparation of these materials present some illuminating peculiarities.
His lactones were obtained from the corresponding cyclic ketones by oxida-
tion (25),

r-i r\

(CH 2 ),CO -* (CH 2 ),CO

,_! LJ

i. e. y the ring was already established in the starting material. All at-
tempts to prepare the lactones by a bifunctional reaction, e. g., from the
hydroxy acids or from the silver salts of the bromo acids were unsuccessful.
The ketones themselves (71) were obtained by heating the thorium salts
of the dibasic acids to a very high temperature (400-500C.). These
conditions of violence lead to thermal rupture, and, whatever the primary
products may be, the ultimate products must, for the most part, be volatile
materials of fairly low molecular weight. The cyclic ketones are found
mixed with a great variety of other materials among these volatile prod-
ucts. It is at least conceivable that the primary products in this reaction
are linear polyketones and that the cyclic ketones are thermolysis products
of the polymer.

In this connection it is interesting that large heterocyclic rings can also
be obtained by thermolysis of the polyesters. Thus when polyethylene

IX. POLYMERIZATION

105

O 1-1
CO IO

~ g

s ^

c/3

W 04
J U
PQ H
<3 t/3

H W
u

M
_]

<fi bt

d c

r ^

'""'

ffi

s^.

, .

Formul

2 OCO(CH 2 ) 2 CO

4 OCO O(CH 2 ) 4 OC

3 OCOCO O(CH 2 ) 3

9 CO O(CH 2 ) 9 CO1

H 2 ) 8 CO OCO(CH 2

/ *e5'

/ ~V1

<L)

j|j

o
O

"rt

g
S

.y
'o

itramet

jd
1

meric j

106 POLYMERIZATION AND RING FORMATION

succinate is heated to 300C. in a vacuum, it decomposes and yields a
carbonaceous residue and a gaseous and liquid distillate. From the latter
a small amount of the sixteen-membered cyclic dimer can be obtained.
Similar cyclic products have been obtained from tetramethylene carbonate,
from trimethylene oxalate and from the potassium salt of the acetyl de-
rivative of hydroxydecanoic acid. The anhydride of sebacic acid can also
be depolymerized under certain conditions to a cyclic dimer. The proper-
ties of these compounds are listed in Table III. Their identity is estab-
lished by analytical data and repeated molecular weight determinations.
They are distinguished from the corresponding polymers by their definite
macrocrystallinity and relatively great solubility. The attempts to obtain
corresponding monomers which would be 7-, 8-, and 11-membered rings
have been unsuccessful. It seems possible that this failure may be due to
purely practical difficulties, but so far as the data go they are in agreement
with the fact that Ruzicka found the yields of the cyclic ketones of 8 to 12
atoms to be much less than the yields of the larger rings. This fact is not
easy to explain on the basis of the simple Sachse-Mohr theory presented
above. A possible explanation has recently been offered by Stoll and
Stoll (67).

III. Polymerization Involving Cyclic Compounds

1. Six-Membered Cyclic Esters. Among cyclic esters the property of
undergoing reversible polymerization is characteristic of and peculiar to
the 6-membered rings. 7-Butyrolactone cannot be polymerized, and no
corresponding polyester is known. Higher lactones, e. g. t

[O (CH 2 ) 16 CO],

are stable substances which show no tendency to polymerize spontaneously,
although polymers are the only products of the self-esterification of the
corresponding hydroxy acids. On the other hand, 6-valerolactone, a mo-
bile liquid, gradually changes on standing to an opaque solid polymer, and
from this the lactone can be regenerated by heating.

Various isolated examples of the polymerization of the 6-membered
esters have been reported, but only a few of them have been described in
any detail. In this laboratory some study (mostly unpublished) has been
made of the mechanism of this phenomenon.

The esters now known to exhibit this behavior are listed in Table IV.
(In most cases neither the molecular weights nor the melting points of the
polymers can be regarded as very significant, since both are dependent upon
the conditions under which the polyester is formed.)

IX. POLYMERIZATION

107

3 S.

.Ja

tf
W
S

S 2

SS S 5

rH en

s g

EMBERED

o
o

i i
o

ffi

r i

>-v

*3
|

o" 1

* o 1

ffi

[OCH 2 CH(CH 3

[OCH 2 CO OC

ffi

colic acid

a
3

*5c

-s

-4_

! |

)_l

-a
>^

"i,

,a
5

4-1

Lactone

S
^

00*

^0.

' S

> t

Lactide .

108 POLYMERIZATION AND RING FORMATION

Some of these examples present special features worthy of mention.
Monomeric trimethylene carbonate is a very soluble, crystalline solid.
If it is heated with a trace of potassium carbonate to 130C. for a few
minutes, the mobile melt suddenly becomes very viscous and evolves a
small amount of gas. The colorless sirup on being cooled solidifies to a
stiff mass which shows an apparent molecular weight of about 4000.
When heated in a vacuum this mass distils almost quantitatively, and the
distillate consists of pure monomer.

Ethylene oxalate is a solid crystallizing in transparent, flat diamonds,
melting at 143C. These on standing for a few days in a stoppered con-
tainer disintegrate to a powder, which consists of a mixture of polymers
By extraction with cold solvents it can be separated into two definite frac-
tions, one melting at 159C. and having an apparent molecular weight of
about 3000, and one melting at 173C. and having an unknown but prob-
ably much higher molecular weight. These fractions on standing for a
few days lose their identity; they are partly converted into each other and
into monomer. It is interesting to note in connection with the rapidity
of these transformations that monomeric ethylene oxalate is exceedingly
sensitive to hydrolysis.

Drew and Haworth (74) have obtained the lactone of 2,3,4-trimethyl-/-
arabonic acid in crystalline form (m. p. 45C.) and have observed that in
the presence of traces of hydrogen chloride it is converted into a crystalline
polymeric powder. This has a considerably higher melting point, a lower
solubility, and a lower specific rotation than the lactone. Its molecular
weight (about 2000) indicates that it is derived from about 10 molecules of
the latter. At 175C. it distils completely in vacua t and the distillate con-
sists of pure lactone. Drew and Haworth ascribed a linear polyester struc-
ture to this polymer and were inclined to accept its crystallinity as evi-
dence of its absolute homogeneity, but it seems much more probable that it
is a polymeric mixture.

The ease of polymerization of the 6-membered cyclic esters appears to
be related to their susceptibility to hydrolysis. In general, substitution
increases the resistance to hydrolysis and diminishes the tendency to poly-
merize. Thus glycolide polymerizes spontaneously at the ordinary tem-
perature, but lactide only on being heated or exposed to the action of cata-
lysts. Attempts (51) to bring about the polymerization of analogs of
glycolide derived from some of the higher a-hydroxy fatty acids have been
unsuccessful.

These polyesters are formed from the monomers by a process of ester
interchange

IX. POLYMERIZATION 109

? ... O R CO O R CO O R CO R CO ...

I ............. J L ............. J L ............. J L ............. J

and the reverse transformation proceeds by a similar mechanism, as in-
dicated by the arrows. Both transformations are catalyzed by acids and
bases typical interchange catalysts.

Direct evidence for this interchange mechanism is found by polymerizing
5-valerolactone in the presence of various amounts of chloroacetic acid
(51). This acid actually participates in the reaction, and according to the
amount present it regulates the length of the chains produced. The effect
of the acid may be compared with that which water might produce. One
molecule of water with one molecule of the lactone would simply yield the
hydroxy acid. A smaller amount of water would yield some hydroxy acid
and this might react with the lactone to form a dimeric ester :

HO R COOH + O R CO -> HO R CO O R COOH.

If the amount of water were quite small the polyester molecule would
have to be quite large. Apparently chloroacetic acid functions in precisely
the same way, and a comparison of the halogen content and neutral equiva-
lent of the polyesters produced under various conditions not only estab-
lishes the open chain structure of the polymer, but also clearly indicates the
mechanism of the reaction.

It remains to explain why only the 6-membered esters are capable of
undergoing reversible transformation of this kind. The presence of some
strain in such esters is indicated by the great instability of 5-lactones as
compared with y-lactones (78, 79). The Sachse-Mohr theory permits the
existence of 6-atom rings in two isomeric strainless forms; but in simple
rings these two isomeric forms have never been realized, and one is forced
to conclude either that such rings are uniplanar and hence highly strained,
or that the two isomeric forms are in dynamic equilibrium. Practically
this amounts to the same thing, since at each conversion the molecule must
pass through the uniplanar position of strain. These strains can be re-
lieved by an ester interchange resulting in the formation of the polyester.
The easy depolymerization of the resulting polyester is readily explained
by the high degree of probability of the close approach of points 6 atoms
apart in a chain. Polyesters whose structural units are longer than 6
atoms are not readily depolymerized because of the improbability of close
approach of the requisite atoms of the chain. Cyclic esters of 5 atoms or

110 POLYMERIZATION AND RING FORMATION

of more than 6 atoms are not polymerized because their cyclic systems are
free from strain.

In the formation of 6-membered cyclic esters from open chain compounds,
either the monomer or the polymer may first be isolated according to the
experimental conditions, but in either event it is not easy to prove that the
form isolated is the primary product and not a polymerization or depoly-
merization product of the primary product. It seems fairly certain that
simple 5-lactones may be formed directly from the corresponding hydroxy
acids. On the other hand there is evidence (19) that the self-esterifica-
tion of lactic acid yields only polylactyllactic acids and that lactide, when
it is formed, is produced from these by depolymerization. Since lactide
can itself be polymerized to a polylactyllactic acid, the latter furnishes an
example of a polymer that can be formed either by addition or by condensa-
tion polymerization. In most cases, no doubt, either the cyclic ester or
the polymer can be formed first depending upon the conditions.

2. Adipic Anhydride. The behavior of adipic anhydride (56) illus-
trates the fact that the property of undergoing reversible polymerization
of this type is not peculiar to the polyesters. In this case it is the 7-atom
ring that is unstable. Succinic and glutaric anhydrides, the 5- and 6-
atom rings, cannot be induced to polymerize. Sebacic anhydride is poly-
meric and cannot be depolymerized to a monomer. Adipic anhydride,
when it is prepared from the acid, is also polymeric. It is a waxy solid
which separates from solvents as a microcrystalline powder. On being
heated in vacua it is depolymerized to a considerable extent, and distilla-
tion occurs. The distillate is the monomeric adipic anhydride, a liquid
freezing at about 21 C. In the presence of traces of moisture this very
rapidly reverts to the waxy polymeric form.

A direct chemical proof of the actual structural difference between the
two forms of adipic anhydride is possible (56). Both forms react very
rapidly with aniline. The monomer, since it contains in its molecule only
a single anhydride linkage, can yield only the monoanilide and this is in
fact the only product formed.

fCO(CH 2 ) 4 CO O] + C6H 6 NH 2 -* HO CO(CH 2 ) 4 CO NHC 6 H 6 .
But with the polyanhydride the nature of the final monomeric products
will depend upon which side of the anhydride linkage is involved at suc-
cessive steps :

f 1 I ! I
. . . CORCO-7-O CORCO-^O CORCO OCORCO-i-O CORCO O . . .

RNH ! H RNH H H

NHR RNH H RNH

IX POLYMERIZATION 111

If the molecule is infinitely long, i. e. t if it is so long that the terminal groups
can be ignored, considerations of probability indicate that the product will
be 50 per cent, monoanilide, 25 per cent, dianilide and 25 per cent. acid.
Within the experimental error these are the yields actually obtained.
This indicates that the molecule must be made up of at least 10 or 15
structural units. A cyclic polymer would of course give the same result,
but a low molecular weight is excluded by the properties of the material,
and a cyclic structure of any kind is made improbable by considerations
already discussed.

3. Diketopiperazines and Polypeptides. The curious behavior of the
a-hydroxy acids is especially interesting in connection with the structure
and formation of proteins, since these are for the most part derived from
a-amino acids. If glycine were to behave like glycolic acid, the following
transformations could be realized.

NH
-H,0 . / \

NH 8 CH a COOH <, . TT ^ * CO CH, heat

-Hrf)

H,0 \

I I <-

CH, CO

NH
... -NH-CH a -OD~NH-CHr-C-NH-<JHr-^0-NH-CHCX>- ...

It appears that diketopiperazine can be polymerized, though with some
difficulty (83), and polypep tides are readily obtained by the self -condensa-
tion of glycine and its esters.

Balbiano (80) in 1900 showed that when glycine is heated in glycerol
solution it loses water and yields as the principal product a horn-like mass,
together with a small amount of diketopiperazine. The horn-like mate-
rial is practically insoluble in all solvents except hot concentrated hydro-
chloric acid, and by hydrolysis it is converted to glycine. Later this reac-
tion was studied in more detail by Maillard (81), who showed that accord-
ing to the conditions and time of heating more or less of the dimeric an-
hydride or the horn-like polymer can be obtained. He also isolated an
intermediate individual, triglycylglycine,

NH 2 CH-r-CO NH CHr CO NH CH-r- CO NH CH 2 COOH

This polypeptide is soluble in water, and Maillard made the curious ob-
servation that if its aqueous solutions contain diketopiperazine they deposit
on standing an insoluble material. Analytical evidence indicated that this

112 POLYMERIZATION AND RING FORMATION

material is a hexapeptide. Except for the solubility effect in aqueous solu-
tions there is no obvious reason why this coupling of the diketopiperazine
with the polypeptide should stop at the hexapeptide stage. In glycerol
solution one may suppose that it continues progressively and results in the
building up of a very long chain. The final horn -like product would then
be a polypeptide containing a very large number of structural units
perhaps forty or more. Maillard assigns to this horn-like polymer the
structure of a cyclopolyglycylglycine, but no convincing proof of the
cyclic formula is presented, and on general grounds an open chain formula
seems more probable.

The production of glycine anhydrides in glycerol solution probably
involves the formation of glyceryl esters as transitory intermediates, but
diketopiperazines and polypeptides can also be obtained by heating amino
acids in the absence of a solvent. Curtius and Benrath (82) state that at
high temperature the chief product from glycine is a pentapeptide.

The esters of glycine couple with themselves much more readily than the
free acid. Curtius (84) showed that glycine ethyl ester loses alcohol even
at the ordinary temperature. When water is present the chief product is
diketopiperazine, but some polypeptide ester is formed at the same time.
Under anhydrous conditions the latter is the chief product. When glycine
ethyl ester is dissolved in a little dry ether and allowed to stand at the
ordinary temperature, a crystalline precipitate gradually accumulates.
This consists almost exclusively of the ethyl ester of triglycylglycine.
The esters of such polypeptides are of course capable of coupling with
themselves to form still longer molecules. Fischer (85) observed that
when the methyl ester of alanylglycylglycine is heated to 100C. it yields
some hexapeptide ester and some less soluble amorphous material. Curtius
(84) found that when his triglycylglycine ester was heated to 100C. in
vacua it was converted into an insoluble infusible material having the
composition NHCH 2 CO . He assigned to this material the struc-
ture of a cyclic octapeptide, but a much more highly polymeric open chain
structure seems more probable.

In connection with the polypeptides the behavior of the anhydrides of
the N-carboxy amino acids may be mentioned (89). The compound

CH 2 CO

NH CO

dissolves readily in water at 0C. At 15C. carbon dioxide is evolved and
an aqueous solution of glycine results. If, however, the compound is

IX. POLYMERIZATION 113

rubbed with a little water at room temperature or is heated, carbon dioxide
is lost immediately and an insoluble infusible material is the only other
product. This has the composition NHCH^CO and on hydrolysis it
yields glycine. It appears to be similar to the horn-like product obtained
by Balbiano (80). It is quite probably a polypeptide of very high molecu-
lar weight. The anhydrides of other N-carboxy a-amino acids behave
in a similar fashion (90).

4. Ethylene Oxide (14, 15). Ethylene oxide provides another ex-
ample of the polymerization of a cyclic compound not involving double
bond unsaturation. The polymerization is induced by various catalysts
such as alkali metals, tertiary amines, and stannic chloride. Ultraviolet
light or Florida earth is not effective. The product is a solid readily
soluble in water and in most organic solvents except ether. Its properties
vary somewhat according to the method of preparation. By fractional
precipitation it can be separated into fractions ranging in apparent molecu-
lar weight from about 400 to nearly 5000. The lowest member of this
series is liquid and the highest one a solid melting at about 59 C. The
polyethylene oxides are not depolymerized by heat; but above 300C.
they decompose and yield a complicated mixture of products containing
some acetaldehyde and acrolein. Roithner (15) assigns to polyethylene
oxide the following formula,

[O CH 2 -~ CH 2 O CH 2 CHir- O CH 2 CH 2 ]

but the evidence for this structure is largely negative. Unpublished work
in this laboratory has led to the detection of terminal hydroxyl groups,
indicating that the molecule is an open chain.

Ethylene oxide also polymerizes under certain conditions to yield dioxane
(16). Apparently this shows no tendency to polymerize further as the 6-
membered cyclic esters do.

IV. Addition Polymerization of Unsaturated Compounds

This is the only type of polymerization that is recognized by the usual
definitions. Thus Cohen states (91) that "the property of undergoing
polymerization is peculiar to unsaturated compounds, from a natural
tendency to saturate themselves/' So far as the formation of materials
of high molecular weight is concerned, such reactions are much less clear-
cut than bifunctional condensations, for the latter involve only the applica-
tion of the known reactions of typical functional groups, and the general
structural plan of the product may be inferred directly from the structure

114 POLYMERIZATION AND RING FORMATION

of the starting materials. On the other hand, no clue to the intimate de-
tails of the mechanism of self-addition can be found in the reactions of the
compound concerned with any compounds other than itself. Ethylene
oxide and certain cyclic esters and anhydrides already discussed poly-
merize by self -addition, but these reactions in some respects are radically
different from the polymerization of compounds containing multiple link-
ages.

Under sufficiently drastic conditions almost any compound can be con-
verted into a material of high molecular weight. Thus methane when sub-
jected to the action of alpha particles or to the silent electrical discharge is
partly transformed with loss of hydrogen into higher hydrocarbons (165),
and benzene in the electrodeless discharge is converted into insoluble,
amorphous products (166) . The reaction of acetylene with itself illustrates
how complicated the polymerization of an unsaturated compound may be.
Polymerization at elevated temperature in the presence of active charcoal
leads to a complicated mixture of hydrocarbons containing considerable
amounts of benzene and naphthalene (167); the use of copper- or mag-
nesium-containing catalysts yields (168) a completely insoluble amorphous
powder (cuprene) of unknown structure, and similar products are obtained
by the action of ultraviolet light (169), cathode rays (170), or alpha parti-
cles (171) on acetylene. On the other hand, the action of the silent elec-
trical discharge (172) at low temperature leads to the formation of con-
siderable amounts of liquid products which contain highly unsaturated
open chain compounds.

The brief discussion of addition polymerization contained in the follow-
ing paragraphs is confined to a few of the simplest and most thoroughly
studied cases.

1. Ethylene and Other defines. The rather extensive literature on
the polymerization of ethylene has recently been reviewed by Stanley (92) .
Ethylene polymerizes less readily than most of its homologs and derivatives.
The polymerization is accelerated by heat, pressure, ultraviolet light, the
silent electric discharge, and by certain catalysts such as sulfuric acid
(especially in the presence of salts of copper and mercury), zinc chloride,
boron trifluoride, and aluminum fluoride. The products are usually oils
having a wide boiling range. In general, they are not exclusively of the
C n H2 W type; hydrogenation, dehydrogenation, and cyclization may occur
at the same time. These products are commercially valuable, but too com-
plicated to furnish any clue as to the mechanism of their formation.

Mignonac and Saint-Aunay (93), however, have succeeded in isolating
the first products formed in the action of the silent electrical discharge on

IX. POLYMERIZATION 115

ethylene. These products are butene-1 and hexene-1. Pease (94) has
found evidence of the presence of butene-1 in the products of the thermal
polymerization, and from a study of the kinetics has concluded that this is
a chain reaction. In effect at least this reaction involves at the first step
the addition of ethylene, as H + CH=CH 2 , to the double bond of another
molecule of ethylene, and then a similar addition to butylene.

Lebedev and his collaborators (95) have presented interesting data on
the early stages of the polymerization of isobutylene. They used espe-
cially Florida earth as a catalyst and obtained mixtures from which a whole
series of polymers up through the heptamers was isolated. They found
that the trimer was not polymerized under the conditions that lead to
higher polymers from either the monomer or the dimer. Hence the higher
polymers must be built up by the successive addition of monomer or dimer.
As to the precise structure of these polymers very little is known.

2. Vinyl Compounds. Substituted ethylenes of the type CH 2 =
CH R, in which R is a negative group, polymerize much more readily
than does ethylene itself. Examples are:

Styrene (96)

CH 2 =CH Cl(Br) Vinyl chloride (bromide) (97)

CHr=CH O COCH 3 Vinyl acetate (98, 142)

CH^CH COOH Acrylic acid (99)

CH2=CH CHO Acrolein (100)

Indene (101, 105)

The most extensive and important studies of the polymerization of vinyl
compounds are those carried out by Staudinger and his co-workers. The
behavior of styrene may be taken as typical. On standing or on being
heated, this mobile, volatile liquid first becomes more viscous, then changes
to a more or less tough elastic jelly, and finally it may become converted
into an exceedingly hard brittle mass. This change is powerfully catalyzed
by light and by atmospheric oxygen, and it is inhibited by certain anti-
oxidants such as hydroquinone. Other catalysts that are effective in ac-
celerating the change are organic peroxides and certain metallic halides
such as stannic chloride and antimony chloride. If a little stannic chloride
is added to an alcohol solution of styrene, the solution becomes warm and
the polystyrene quickly separates as an amorphous mass.

116 POLYMERIZATION AND RING FORMATION

Polystyrene is not a definite material having a constant set of properties.
By whatever method it is prepared it can be separated by fractional ex-
traction or precipitation into fractions having the same composition but
different properties and molecular weights. The lower members are readily
soluble in ether and the highest members are quite insoluble. The ap-
parent molecular weights range from 1000 up to 25,000 or more (perhaps
as high as 200,000). All these fractions are soluble in benzene, and the
viscosity of the solutions increases progressively with increasing molecular
weight. The average molecular weight of the crude polymer depends
upon the conditions of its formation. Those formed very rapidly, e. g.,
by the action of heat at high temperature or by the action of stannic
chloride at ordinary temperature, have relatively low molecular weights.
Those formed more slowly, e. g., by spontaneous polymerization at room
temperature, have much higher molecular weights.

The polystyrenes having molecular weights above 10,000 show colloidal
behavior; they swell before dissolving, and the viscosity of their solutions
is very high. Nevertheless, there is considerable evidence to show that
these solutions are true molecular dispersions (102), and that the molecule
and the colloidal particle are identical.

Chemically the polystyrols are completely saturated. They do not de-
colorize permanganate or absorb bromine. Under sufficiently drastic
conditions they can be completely hydrogenated (in the benzene nucleus)
without any significant change in molecular weight (103). On being
heated to about 320C. they revert to the monomer, styrene. It is doubt-
ful if this reversion is ever quantitative.

The polymerization of styrene evidently involves the disappearance of
the double bond and the formation of very large molecules. The simplest
and most probable structure of these large molecules is that suggested by
Staudinger.

. . . CH CH 2 CH CH 2 CH CH 2 CH CH 2 . . .

I I I I

C 6 H* C 6 H 6 C 6 H 6 C 6 H 6

The polymerization of vinyl acetate is similar to that of styrene. In
both cases the polymer first formed remains dissolved in the monomer, and
the mixture of monomer and polymers can be obtained in the form of more
or less tough, transparent, elastic masses. The chemical behavior of the
polyvinyl acetate indicates that it is a mixture of molecules of various
lengths built up according to the plan indicated in the formula

. . . CH 2 CH CH 2 CH CH 2 CH CH 2 CH CH 2 CH-. . .

I I I I I

OAc OAc OAe OAc OAc

IX. POLYMERIZATION 117

The number of molecules of monomer involved probably ranges between
forty and one hundred. By fractional precipitation, samples of different
average molecular weights can be obtained, and these naturally differ
somewhat in their physical properties. In general the polyvinyl acetates
are soluble in organic solvents but not in water. As esters they are readily
hydrolyzed. The products are acetic acid and polyvinyl alcohol:

CH 2 CH CH 2 CHCH 2 CH CH 2 CH

OH OH OH OH

The latter as a polyhydroxy compound dissolves in water to form rather
highly viscous solutions, but does not dissolve in organic solvents.

Acrylic acid also polymerizes very readily, yielding a product to which the
following formula may be assigned.

. . . CH 2 CH CH 2 CH CH 2 -CH CH 2 CH . . .

I I I I

COOH COOH COOH COOH

As a highly polybasic acid it forms a sodium salt which dissolves in water
to yield very viscous solutions. The acid itself also dissolves in water
(the most highly polymeric varieties swell very strongly and dissolve with
difficulty), but does not dissolve in the typical organic solvents. On the
other hand, polyacrylic esters prepared by polymerizing acrylic esters are
insoluble in water, but soluble in organic solvents.

The methyl ester of polyacrylic acid on being treated with methyl mag-
nesium iodide yields a product corresponding approximately in com-
position to the expected polytertiary alcohol, and this on reduction yields
a hydrocarbon. Although the composition of this hydrocarbon does not
exactly correspond to that required by the expected structure
. . . CH 2 CH- CH 2 CH. . .
(CH 3 ) 2 HC CH(CH 3 ) 2

its high molecular weight proves that the ester from which it was derived
is also really a material of high molecular weight, and that the units must
be joined by carbon-carbon linkages.

The problem of arriving at a definite mechanism for the polymerization
of vinyl compounds is complicated by the fact that neither the formation
of the polymers nor their chemical behavior furnishes any certain clue to
their structure. The general formula

-CH 2 CH
R

CH 2 CH CH 2 CH CH 2 CH . . .

118 POLYMERIZATION AND RING FORMATION

is perhaps more plausible than any other, since, as Staudinger has pointed
out, it best accounts for the fact that some poly vinyl compounds (e. g.,
polystyrene) are smoothly depolymerized by the action of heat: the re-
curring substituent would weaken the linkages of the chain at the points
indicated by the dotted lines. But depolymerization is never quantita-
tive, and it is at least conceivable that a fraction of the units of each chain
are arranged in the reverse fashion :

. . . CH CH 2 CH 2 CH CH CH 2 CH 2 CH . . .

I II I

R R R R

It has already been mentioned that the first step in the polymerization
of ethylene is the formation of butene-1 and the second step is the formation
of hexene-1. Similar terminal unsaturations have been found by Whitby
and Katz in a still longer series of polyindenes, and this suggests that the
most plausible mechanism for the polymerization of vinyl compounds is
"best represented as proceeding stepwise by the addition regularly of suc-
cessive molecules of monomer to the double bond present at each stage of
the polymerization immediately preceding" (105). This mechanism may
be formulated as follows :

RCH=CH 2 + RCH=CH 2 - RCH 2 CH 2 C(R)=CH 2

RCH 2 CH 2 C(R)=CH 2 + RCH=CH 2 - RCH 2 CH 2 CH(R)CH 2 C(R)=CH 2 , etc.

Staudinger disagrees with this view. He claims that the polyindenes
do not have any terminal unsaturations, and from analogies based on the
behavior of a-methylstyrene (104) he assigns a cyclic formula to poly-
indene and polystyrene. a-Methylstyrene, CH 2 =C(CH 3 )C6H5, poly-
merizes much less readily than styrene, but the reaction may be caused to
occur quite rapidly by certain catalysts, e. g., stannic chloride (104). The
principal product is the saturated dimer, a substituted cyclobutane. A
smaller amount of a saturated trimer is formed at the same time, and in
progressively smaller amounts saturated tetramer, pentamer, decamer,
hexamer, heptamer, and octamer. On the basis of their physical and
chemical properties Staudinger assumes that these higher polymers are
respectively 6-, 8-, 10-, 12-, 14-, and 16-membered rings.

It should be observed however that these products are quite different
from those obtained from styrene under similar conditions ; moreover, the
absence of the a-hydrogen makes it impossible for a-methylstyrene to
polymerize by the mechanism suggested above for vinyl compounds
generally. It is therefore scarcely permissible to conclude that the high
polymers from styrene must be cyclic because the low polymers from a-

IX. POLYMERIZATION 119

methylstyrene are cyclic. Other more general grounds for rejecting cyclic
formulas for linear high ploymers have been presented in previous para-
graphs.

The polymerization of vinyl compounds is enormously susceptible to
catalytic and anticatalytic effects. Heat also accelerates the polymeriza-
tion, and in general the more rapidly the polymer is formed, the lower is
the average molecular weight of the product. Oxygen and peroxides are
catalysts, and antioxidants act as inhibitors. Light, especially the shorter
wave lengths of the visible spectrum, accelerates the polymerization. In
certain cases the presence of oxygen inhibits this photochemical effect.
With the aid of certain specific catalysts, it is possible to convert styrene
and certain other vinyl compounds into dimers, but these are stable sub-
stances that show no tendency to polymerize further, and they differ
structurally from the dimers that have been hypothecated as intermediates
in the formation of the high polymers. Under conditions that result in
the formation of the latter, no polymers of very low molecular weight can
be detected. All these facts indicate that the formation of the high poly-
mer is a chain reaction. The collision of an activated molecule of mono-
mer with another molecule of monomer yields an active dimer capable of
coupling with another molecule of monomer, and the activating energy
persists in the polymeric chain until it has been built up to a considerable
length. Kinetic studies of the polymerization of vinyl acetate (106) and
styrene (107) support the idea of a chain mechanism.

3. Dienes. Butadiene and isoprene are of especial interest in con-
nection with rubber. Nobody knows whether in nature rubber is ac-
tually formed from isoprene or not; but it is true that rubber yields some
isoprene on thermal decomposition, and that isoprene can be polymerized
to a product which more or less resembles natural rubber. Similar prod-
ucts can also be produced from butadiene and from some of its derivatives.
Efforts to produce synthetic rubber have led to hundreds of patents and
various other publications. This subject is reviewed in a book by Schotz
(108). There is space here to mention only two or three points.

Butadiene and isoprene polymerize much less readily than styrene,
vinyl acetate, etc., but apparently they are subject to the same kind of
accelerating influences. Among the catalysts that have been used are
oxygen, peroxides, ozonides, alkali metals, alkali alkyls. When emulsi-
fied, especially in the presence of oxygen, they polymerize more rapidly
than otherwise. It is claimed in many patents that the presence of proteins,
gums, etc., in such emulsions has a favorable effect on the course of the
reaction and results in products more nearly resembling natural rubber,

120 POLYMERIZATION AND RING FORMATION

but it seems probable that the advantages of such additions are largely
imaginary. It is possible to prepare from isoprene and butadiene high
polymers that have very little resemblance to natural rubber, and the
problem of preparing a synthetic rubber of good quality is enormously
complicated and difficult.

The structural unit in the polymeric chain from isoprene appears to be

CH 2 C=CH CH 2
CH 3

In effect, the polymerization involves the union of a large number of these
radicals derived from isoprene by the rearrangement of its bonds. Two
such radicals might unite in three ways: 1,1; 1,4; 4,4. Midgley and Henne
(109) have captured the first step in the reaction by carrying out the poly-
merization in the presence of sodium and alcohol so that the terminal
valences of the dimer are hydrogenated. The structures of the three prod-
ucts, whose formulas are shown below, prove that all three types of com-
bination occur.

CH 3 ~-CH=C CH 2 CH 2 C=CH CH 3

I I

CH 3 CH 3

CH 3 C=CH CH 2 CH 2 C=CH CH 3

CH 3 CH 3

CH 3 C=CH CH 2 CH 2 CH=C CH 3
I I

CH 3 CH 3

This result indicates that the arrangement of the units in synthetic rubber
formed under these conditions is less regular than in natural rubber, but
it by no means proves that the actual mechanism of the reaction consists
in the direct union of radicals corresponding in formula with the structural

unit. The 1,4 addition of H 1 CH=C(CH 3 )CH=CH 2 to isoprene,

for example, would lead to the same result.

4. Aldehydes. The polymerization of formaldehyde under various
conditions leads to polyoxymethylenes. These are microcrystalline pow-
ders, which in general cannot be melted or dissolved in organic solvents
without decomposition. They have been very elaborately studied by
Staudinger and his co-workers (110). These studies have proved that the
molecules of the polyoxymethylenes are long chains of the type

. . . . O CH 2 O CH 2 O CH 2 O CH 2 O CH 2 . . . .

The chains contain from forty to at least one hundred structural units.
The different varieties of polyoxymethylenes are distinguished by the

IX. POLYMERIZATION 121

nature of the terminal groups: in the a-variety the terminal groups are
OH ; in the /3- variety they are OCH 3 , and this variety is more inert chemi-
cally than the a-variety. By the action of acetic anhydride, the poly-
oxymethylenes are simultaneously degraded and acetylated. The product
consists of a mixture of compounds of the series CH 3 CO (O CR^ X O
COCHa. By distillation and crystallization, each individual of this series
from x = 1 to x = 20 has been isolated in a fairly pure state. The melting
points and boiling points of these individuals increase continuously with
increasing molecular weight and their solubilities diminish. The successive
members above x = 20 resemble each other so closely and their solubilities
are so low, that they cannot be separated into pure fractions.

Formaldehyde also polymerizes to yield the cyclic trimer, trioxymethyl-
ene, and analogous trimers are the most common forms of the polymers of
other aldehydes. It seems possible, however, that some of the metalde-
hydes are linear polymers of high molecular weight analogous to the poly-
oxymethylenes. Conant (111) has obtained from butyraldehyde by the
action of very high pressure a solid polymer for which he suggests such a
structure. This polymer is apparently stable only under pressure. Under
the ordinary conditions it rapidly reverts to the monomer.

The thio aldehydes and ketones show more tendency to polymerize than
their oxygen analogs (158). Organic silicon compounds of the types
R 2 Si=O, RSi(=O)OH, etc., are frequently incapable of being isolated in
the monomeric condition, and they invariably polymerize very readily.
It is interesting in this connection to compare carbon dioxide and silicon
dioxide. The latter is probably a three-dimensional polymer.

V. Polyfunctional Reactions and Non-linear Polymers

The polymers discussed thus far, whether formed by condensation or
by self -addition, are of a type that may be symbolized by the formula ....
A A A A A A A . . * . Reactions of condensation are not
limited to bifunctional compounds, however. If one of two reactants con-
tains two functional groups and the other contains more than two, the
product will be not a simple chain but a more complicated structure.
Such reactions may be called bi-trifunctional, bi-tetrafunctional, tri-tri-
functional, etc. (18). Reactions of this class are especially important
technically in connection with the formation of synthetic resins. Two
examples may be considered.

The glyptal resins are formed by the action of phthalic acid on glycerol.
The reaction first leads to the formation of a fairly soft, soluble, thermo-

122 POLYMERIZATION AND RING FORMATION

plastic resin, and this on being heated further yields a hard, insoluble
resin which is completely lacking in thermoplasticity. The only reaction
involved in the process is esterification (112), and the resin can be saponi-
fied completely to yield phthalic acid and glycerol (51). Analysis of the
resin just before the infusible stage is reached shows that the glycerol and
the phthalic acid are far from having reacted completely with each other,
that is, free carboxyl and presumably free hydroxyl groups are still present.
The behavior of a dihydric alcohol such as ethylene glycol in this reaction
presents quite a different picture from glycerol. With the glycol it is
possible to obtain complete esterification, but however far the reaction is
carried the product does not become infusible or insoluble. The reason for
this difference is obvious. The polyester formed from the dibasic acid
and the glycol is linear; as the reaction progresses the molecules grow,
but the growth takes place only in one dimension. Similar chains formed
in the reaction of glycerol with phthalic acid would bear hydroxyl groups

HO -7- -y -y COOH

OH OH OH OH

By reaction of these groups with phthalic acid the chains would be linked
together, and thus a very complicated 3-dimensional molecule would be
built up. After a certain degree of complexity is reached, the possibility
of molecular mobility no longer exists. It is conceivable that this cross-
linking of the chains finally results in a mass that is essentially a single
molecule. In any event it is easy to see how the possibility of further reac-
tion disappears long before all the carboxyl and hydroxyl groups have a
chance to participate. It is easy to see also why the action of any dibasic
acid on glycerol always yields an amorphous resin, whereas the polyesters
from dibasic acids and dihydric alcohols are frequently crystalline.

The formation of Bakelite from phenol and formaldehyde may also be
classified as a tri-bifunctional reaction. The formaldehyde behaves as
though it were HO CH 2 OH, and it reacts no doubt largely at the o-
and ^-hydrogens of the benzene nucleus. With phenol itself the number
of possible products even of quite low molecular weight is so great that
no intermediate polymeric individuals can be isolated. On the other hand
p-cresol has only two readily reactive positions. Thus it is possible as
Koebner has shown (113) to isolate the compounds indicated below
OH OH OH OH OH

CH 8 CH 3

IX. POLYMERIZATION 123

as crystalline individuals, by causing formaldehyde and ^-cresol to react
in the appropriate ratios, and these can be used to build up still longer chains
of the same series. The progressive hydrolysis of silicic esters to silicon
dioxide, already discussed in a previous paragraph, furnishes another ex-
ample of the formation of a 3 -dimensional polymer.

Addition polymerization may also lead to the formation of 3-dimensional
structures. Thus acetylene reacts with itself under the influence of cer-
tain catalysts (not necessarily copper) with the formation of cuprene, an
infusible and insoluble powder, and Staudinger suggests (2) that in this
reaction the first step is the formation of unsaturated chains .... CH=
CH CH=CH CH=CH .... which subsequently combine with one
another. The vulcanization of rubber probably also involves the cross-
linking of the long chains through the agency of the unsaturated linkages
present (39).

VI. Natural Polymers

1. The Association Theory Versus the Structural Theory. The

peculiar and difficult physical and chemical behavior of polymers has
occasionally led to the suggestion that forces of a peculiar kind are involved
in their formation. Thus Thiele in 1899 (132) suggested that perhaps in
such materials as polystyrene the molecules of monomer are bound to-
gether merely by partial valences. Rohm in 1901 (115) concluded that the
transformation of monomeric acrylic esters into the h'ighly polymeric form
is not chemical reaction but a kind of allotropic change. Later Schroeter
(116, 117) suggested that the formation of dimeric ketene and of tetra-
salicylide is due to the manifestation of an excess of peripheral external
force about the monomeric molecules, and that the actual chemical struc-
tures of the monomeric molecules are not changed in the process. These
particular suggestions are not tenable in the light of chemical evidence now
available, but the association theory of polymeric structure reappeared
about 1924 and was widely accepted as an explanation of the peculiarities
of natural high polymers (118). According to one form of this theory,
cellulose, for example, might be an anhydroglucose having the molecular
formula CeHioOs. This molecule, because of the unusual strain of its cyclic
structure or for some other reason, is supposed to exhibit enormously ex-
aggerated forces of association or residual valence, and hence to behave
physically as though it were a material of very high molecular weight.
In the same way proteins might be built up by the mutual association of
various small units, e. g. t diketopiperazines.

In support of this theory various investigators showed that it was pos-

124 POLYMERIZATION AND RING FORMATION

sible by freezing and boiling point methods to obtain small and rapidly
shifting values for the molecular weights of polysaccharides, proteins, and
rubber. Repetition of these determinations by other investigators proved,
however, that the low results were due in most cases to errors in technique.
Other support came from x-ray studies, which indicated that the unit
cell of the crystal lattice of some high polymers is too small to contain a
very large molecule. It was assumed at the time that a unit cell could not
contain less than 1 molecule, but studies of known substances of high
molecular weight proved that this assumption was incorrect. Meanwhile
Staudinger's studies of synthetic materials repeatedly demonstrated that
polymerization may lead to the formation of very long chains built up by
real chemical forces in a regular fashion, and that such synthetic mate-
rials often resemble natural high polymers in many significant physical
and chemical properties. Studies made in this laboratory on high poly-
mers formed by condensation reactions led to the same conclusion. The
idea that natural high polymers involve some principles of molecular
structure peculiar to themselves arid not capable of being simulated by
synthetic materials is too strongly suggestive of the vital hypothesis, which
preceded the dawn of organic chemistry, to be seriously considered.

It should be emphasized in this connection that polymerization is riot
peculiar to unsaturated compounds, and that a very high degree of mobility
in the relation between a monomer and its polymer does not preclude the
intervention of real primary valence forces in the process or the presence of
a definite macromolecular structure in the polymer. This fact is illustrated
especially by adipic anhydride and the 6-membered cyclic esters discussed
in previous sections of this paper. These materials appear to exhibit all
the supposedly diagnostic features of association polymerization; never-
theless their transformation into polymers is a real chemical process, and
the polymers are actually made up of large molecules. No example is
yet known in which a small molecule of known structure simulates a mate-
rial of high molecular weight without undergoing any change in structure.

A return to the simple structural theory of organic chemistry and the
application of modern tools have been responsible, during the past five
years, for very rapid progress in the interpretation of the structure and
properties of some of the simpler, naturally occurring high polymers. It
appears that many naturally occurring macromolecular materials have a
linear polymeric structure. The present status of this subject has been
reviewed by Meyer and Mark (3), and in the following paragraphs only
cellulose and rubber are briefly discussed.

2. Cellulose. Chemical evidence for a linear polymeric structure in

IX. POLYMERIZATION 125

the cellulose molecule has long been available (125). More recently,
Haworth's proof of the structure of cellobiose (120) and certain studies of
the relation of glucose and cellobiose to cellulose (122) furnished the basis
for tentative efforts to determine the nature of the units in the molecular
chain (121). Sponsler and Dore (123) on the basis of x-ray evidence first
put the linear polymeric structure in explicit form. This formula was
further developed by Meyer and Mark (119), and it can now be said that
the structure of cellulose, at least in its essential outlines, is definitely known.
In its simplest chemical form the structural unit of cellulose may be repre-
sented by the bivalent radical derived by the removal of water from glucose
(glucopyranose).

^ -jOHj

CH,OH

Glucose

A large number of these units are united chemically to form a long
chain which constitutes the cellulose molecule.

(For a more detailed picture of the spatial arrangements in this chain,
including the atomic distances, see reference 3.) In the cellulose fiber
(cotton, ramie, etc.) these long molecules are lined up parallel with each
other along the fiber axis.

To complete this formula it is necessary to specify the nature of the
terminal groups and the length of the chain. The most reasonable assump-
tion is that the chains are open and terminated by hydroxyl groups (alco-
holic at one end and semi-acetal at the other), but no very definite experi-

126 POLYMERIZATION AND RING FORMATION

mental evidence on this point is available. Meyer and Mark estimate
that the molecular chains are made up of from sixty to one hundred glucose
units (molecular weight, 10,000 to 16,000), and they assume that the mole-
cules are segregated into compact bundles (crystallites or micelles) from
which the gross structure of the cellulose fiber is built up. Staudinger,
however, has presented evidence (124) that the crystallites do not persist
in dispersions of cellulose and its derivatives, and that the molecules have
a weight much higher than 16,000. Stamm (126) has made direct measure-
ments of particle size in ammoniacal copper dispersions of cellulose by the
Svedberg ultracentrifugal method and obtained a value (on the copper-free
basis) of 40,000 =*= 5 per cent. The molecules in a given sample of cellulose
are probably not all of the same length, but Stamm 's data indicate a much
higher degree of homogeneity than one might expect.

According to this picture cellulose, like other high polymers, is not a
chemical individual in the sense of being composed of identical molecules,
and its structure cannot be completely specified by a single exactly defined
formula. On the other hand, this picture accounts for the physical and
chemical behavior of cellulose just as completely and satisfactorily as the
formulas of simple compounds account for their properties. Cellulose
fibers are very strong because of the parallel arrangement of the long molec-
ular chains along the fiber axis. These chains adhere firmly to one
another because of the cumulative force of association of the numerous
hydroxyl groups. Strong aqueous alkali is able to penetrate this structure,
and the resulting spreading apart of the chains causes lateral swelling.
Other reagents (e. g., nitric acid) can penetrate the structure and esterify
the hydroxyls without completely changing the apparent physical structure
of the fiber. Cellulose is degraded much more rapidly by acids than by
alkalies because the units are joined together by acetal linkages. The
great length of the molecules accounts for the high viscosity of dispersions
of cellulose and its derivatives. The very first stages of degradation
greatly reduce the viscosity (124) because, for example, the hydrolytic
absorption of 1 part of water in 2000 is capable of reducing by half the
average size of the molecules. Undegraded cellulose has little or no re-
ducing power (130) because it contains only one reducing group in an ex-
ceedingly large molecule, but reducing power is manifested and increases
progressively with hydrolysis (124, 127). Complete hydrolysis finally
gives a quantitative yield of glucose (125) ; cellobiose, a triose, a tetrose
(128) and higher polysaccharides are formed as intermediate products.
The hydrolysis agrees in its kinetics with the theoretical requirements for
the chain structure (129). This structure also accounts for the presence

IX, POLYMERIZATION 127

of three esterifiable hydroxyl groups for each Ce unit, and for the fact
that completely methylated cellulose yields 2,3,6-trimethylglucose on
hydrolysis (121).

3. Rubber. It is well known that rubber hydrocarbon has the same
empirical composition as isoprene (CsHs) ; that isoprene can be polymerized
to yield a material resembling natural rubber; and that the behavior of
rubber toward halogens, hydrobromic acid, and hydrogen indicates the
presence of one double bond for each five atoms of carbon. Harries many
years ago (131) showed that the degradation of rubber by ozone yields
chiefly levulinic acid and aldehyde, and this fact indicates that the rubber
molecules must be largely built up by the repetition of the unit

CH 2 C=CH CH 2

in a regular manner as indicated below.

(II) . . . CH 2 CMe-=CHCH 2 ~-CH 2 CMe-=CHCH 2 CH 2 CMe=CHCH 2 . . .

... CMe==CHCH 2 CH 2 CMe=-CH ... __gg omzatlon _ > CHOCH 2 CH 2 COMe
Rubber Levulinic aldehyde

Harries first assumed that the rubber molecule is an 8-membered ring.
Later the discovery of larger fragments in the products of degradation by
ozone led him to suggest a larger cyclic structure. The physical properties
of rubber clearly indicate however that it is macromolecular. Pickles
(133) suggested the more plausible linear polymeric structure indicated
above, and Staudinger (134) has brought forward a large mass of evidence
in favor of this structure. The best evidence available indicates that the
average weight of the rubber molecule is exceedingly large perhaps in the
neighborhood of 70,000 (135). It appears, moreover, that molecules hav-
ing widely different sizes must be present in a given sample of rubber.
Raw rubber is not homogeneous in its behavior toward solvents. When
placed in contact with ether it swells and part of it diffuses into solution
fairly rapidly. The action of fresh ether on the residue is much slower,
and it is possible to carry the process of extraction so far as to obtain ulti-
mately a residue that shows scarcely any tendency to dissolve. On the
basis of such experiments it has been assumed that rubber is made up of
two phases, sol and gel, and that the properties of rubber are due to the
colloidal relationships of these two phases. In support of this idea there
is the fact that colloidal dispersions of the jelly type frequently exhibit
striking elasticity. This is true, for example, of polystyrene when it is
swollen with unpolymerized styrene or with other hydrocarbon solvents.

128 POLYMERIZATION AND RING FORMATION

Whitby (137), however, has shown that sol-rubber (diffused) alone shows
all the characteristic physical properties of raw rubber and differs from the
latter only in degree. Moreover, it appears that sol- and gel-rubber are
not distinct species and that neither of them is homogeneous. Rubber
probably consists of a long continuous series of molecules of differing
lengths. The smallest molecules dissolve quite readily and the largest
ones only with difficulty.

If the rubber molecule is built up uniformly according to the regular
plan indicated in formula II it should yield levulinic acid and aldehyde as
the exclusive products of degradation by ozone. The yields of these prod-
ucts obtained by Harries accounted for only about 70 per cent, of the rub-
ber. Pummerer (136) has recently undertaken a reexamination of the
Harries method, using carefully purified rubber, and has been able to ac-
count for 90 per cent, of the rubber. He finds significant amounts of suc-
cinic acid, acetic acid, and acetone among the products. If the rubber
molecule is an open chain, it might have an extra double bond at one end,
and according to the disposition of this bond any of the above named by-
products might be produced by ozonization. These products might also
result from occasional irregularities in the arrangement of the units along
the chain. The various possibilities have been outlined by Whitby (137).

Although the examination of unstretched rubber by x-rays gives only
an amorphous ring, stretched rubber gives a sharp fiber diffraction pattern
(138). When unstretched rubber is cooled in liquid air and then frac-
tured by impact, it breaks up into irregular fragments; stretched rubber
treated in the same way breaks up into thin fibers along the axis of stretch
( 1 39) . These facts show that stretched rubber is much more highly oriented
than unstretched rubber.

The structural unit of rubber contains a double bond and this brings
about the possibility of stereoisomerism. The units may be in the cis
or the trans form (or both).

/ \

CH 2 H H CH 2

v v

it 4

/ \ / \

CH 2 CH 3 CH 2 CH 2

\ \

cis form trans form

The x-ray data are said to favor the cis orientation (140). A spatial
model of a long chain of such cis units is capable of being coiled up into a

IX. POLYMERIZATION 129

cylindrical spiral, and this spiral can be stretched out into a long chain.
Various writers have suggested that the spiral model is capable of ex-
plaining the reversible stretching of rubber. This model has recently been
discussed in detail by Fikentscher and Mark (140). It is assumed that the
residual valence forces at the double bonds are responsible for holding the
spiral in its compressed form. When the rubber is stretched, work is done
against these forces, and the molecules assume the chain-like form, where
they are much more highly oriented with respect to one another. Vulcani-
zation is assumed to involve the chemical linking of these spiral chains at
occasional points through sulfur atoms. A very small amount of com-
bined sulfur does not interfere with the stretching of the spirals, but it pre-
vents the chains from slipping past one another or being torn apart; con-
sequently vulcanized rubber is not plastic, and it is not dissolved but only
swelled by rubber solvents. As the amount of combined sulfur increases,
the entire structure becomes more rigidly linked together, the plastic prop-
erties are completely suppressed, and the ability to imbibe solvent is lost.
This picture is useful but it can hardly be said to account completely for
the remarkable properties of rubber.

VII. The Physical Properties of High Polymers

Perhaps the most important result of the study of synthetic high poly-
mers has been to establish the fact that such materials are actually made up
of exceedingly large molecules in the sense of the ordinary structural theory
of organic chemistry. This is a point on which considerable scepticism
has prevailed in the past, and the attempt to evade or ignore the idea of
the molecules in dealing with high polymers has led to much speculative
confusion.

It is true that synthetic linear high polymers are invariably mixtures
whose molecules are chains of slightly differing lengths, and it is difficult
to obtain reliable estimates of the average size of these molecules. Never-
theless it must be admitted that a molecule does not lose any of its definite-
ness as an entity either through the fact that it is exceedingly large or
through the fact that it cannot be completely separated from other similar
but slightly different molecules, and the properties of high polymers must
ultimately be conditioned by the kinds of molecules which they contain.

It would be beyond the intended scope of this paper to attempt a de-
tailed discussion of the relation between the molecular structure and the
physical properties of high polymers, but there are two or three points that
deserve some mention.

130 POLYMERIZATION AND RING FORMATION

It is evident that in some respects the physical behavior of a molecule
whose length is 100 times as great as its other dimensions must be pro-
foundly different from that of a small compact molecule. Enormously
long, flexible, and clumsy molecules must be very sluggish in their kinetic
behavior, and it is not surprising that high polymers cannot be distilled or
that they are never obtained in the form of thin mobile liquids.

The cohesive forces which resist the separation of molecules from one
another (as measured, for example, by the heat of vaporization) increase
continuously with increasing molecular weight in a given series, and in high
polymers they reach values greatly in excess of the energy required to
rupture a primary valence linkage in a chain (141). For this reason high
polymers cannot be distilled without decomposition ; indeed it appears that
the upper limit of distillability may lie at as low a molecular weight as 1200
to 1500 (44).

1. Solubility and Colloidal Properties. High polymers are subject to
the same rough qualitative solubility rules that apply to simple com-
pounds: like dissolves like; polar compounds dissolve in polar solvents
and non-polar compounds in non-polar solvents ; solubility in a given series
diminishes with increasing molecular weight. Thus, rubber and poly-
styrenes are soluble in benzene, but not in acetone; polyamides are not
dissolved by the usual organic solvents, but are dissolved by hot form-
amide; polyacrylic acid is soluble in water, while its esters dissolve in
organic solvents but not in water; polystyrenes of low molecular weight
(about 1000) are soluble in ether while polystyrenes of high molecular weight
(about 20,000) are only slightly soluble in ether but are still dissolved by
benzene.

The solubility of high polymers is sometimes surprisingly great com-
pared with that of analogous simple compounds of much lower molecular
weight. The higher normal paraffin hydrocarbons (e. g., heptacontane)
are practically insoluble in any solvents at the ordinary temperature, while
polystyrene and hydrorubber, which are essentially very long paraffin
chains substituted at intervals by phenyl and methyl groups, dissolve
readily in benzene. One reason for this no doubt lies in the fact that most
high polymers are mixtures of molecules of different lengths, and these are
capable to a certain extent of behaving independently in their solubility
behavior. Moreover, the crystal lattices of high polymers are not so well
ordered and rigidly constructed as those of low molecular weight materials
and they may, for this reason, be more susceptible to attack by solvents.

The fact that the cohesive forces operating between large molecules are
exceedingly high does not mean that polymers are incapable of forming

IX. POLYMERIZATION 131

molecular dispersions. Solubility depends upon specific affinities. Solu-
ble linear polymers of relatively low molecular weight (e. g., 1000 to 5000)
dissolve spontaneously and very rapidly in appropriate solvents and yield
solutions which are not highly viscous. The osmotic unit in these solu-
tions is the molecule, not an aggregate of molecules. For various poly-
esters this has been proved (32, 34) by the fact that the same molecular
weight values are obtained in a variety of solvents and by both freezing
point and boiling point methods. Moreover, as the data of Table II show,
direct chemical determinations of molecular weight give the same values
as the osmotic methods. Polyesters having molecular weights consider-
ably above 5000 dissolve in the same solvents as the lower polyesters, but
the process of solution is slower and the solutions are more viscous (51).
The same behavior is observed in other polymeric series. Polystyrenes
having molecular weights of about 1000 dissolve instantly in benzene, and
the viscosity of the solutions is low; polystyrenes having molecular weights
above 10,000 swell before dissolving, and the solutions are highly viscous
(102). These evidences of colloidal behavior are due simply to the fact
that the molecules are exceedingly long. The probable mechanism of
solution of certain polymers is best illustrated by a specific example.

Rubber is made up of enormously long hydrocarbon chains ranging
perhaps from 1000 to 10,000 A. in length. These chains have a high
specific affinity for certain non-polar solvents such as benzene. In a mass
of rubber, adjacent chains are firmly bound to one another by cohesive
forces. The structure is not an entirely regular one and there is no doubt
a considerable amount of purely mechanical entanglement. Benzene, by
virtue of its specific affinity for the chains, is capable of penetrating into
the mass, solvating the chains, and spreading them apart. The structure
thus becomes swollen and more tenuous, and finally individual fragments
are carried away into solution. The fragments may be single molecules
or only incompletely disrupted aggregates, but finally, if sufficient solvent
is present, the latter are broken down and what amounts to an actual mo-
lecular dispersion results. This dispersion has a very high viscosity even
when quite dilute, for the molecules are not only very large, but, owing
to the fact that they are solvated and extended in only one dimension, they
have an effective radius of action quite out of proportion to their size.

The view that lyophilic dispersions of linear high polymers are usually
true molecular dispersions, although it has not yet been generally ac-
cepted by colloid chemists (see 174), has been supported by Staudinger
with a large mass of evidence (146, 102, 124, 135), which in its cumulative
force seems to the writer fairly conclusive. Reference may be made also

132 POLYMERIZATION AND RING FORMATION

to Stamm's determination with Svedberg's ultracentrifuge of the particle
size of cellulose dispersions in copper-ammonia solutions (126). He ob-
tained the value 40,000 =*= 5 per cent. It is quite certain on various
grounds that the average molecular weight of cellulose cannot be less than
about 16,000, so that if Stamm's particle is an aggregate it cannot con-
tain more than two or three molecules. It seems highly arbitrary to
assume that the solvent action of the dispersing agent, which depends upon
a specific affinity for the cellulose molecules, should be capable of carrying
off the molecules only as pairs or triplets and never as single molecules.

It is of course not contended that association never occurs in lyophilic
solutions of high polymers, but merely that association occurs only as the
result of some appropriate peculiarity in the molecular structure of the
polymer, e. g., through the presence of recurring amide or carboxyl groups.

2. Crystallinity. Linear polymers, in spite of their lack of complete
homogeneity and their high molecular weight, are by no means always
amorphous. As indicated in Table I, all the polyesters derived from gly-
cols of the series HO(CH2)*OH and acids of the series carbonic, oxalic,
succinic, etc., separate from solvents in the form of powders which show
quite definite melting points. On the other hand, similar esters derived
from phthalic acid are invariably transparent, amorphous resins. The
ability to crystallize appears to require a high degree of linear symmetry
in the structural unit. The presence of side chains such as methyl or
phenyl groups on the units, and the random mixing of structural units,
which occurs, for example, when polymers are prepared from a single glycol
and two different acids, diminish the tendency toward crystallinity. Thus,
polyesters and poly amides derived from unsubstituted aliphatic com-
pounds are crystalline, and so also are the polyoxymethylenes and poly-
ethylene oxide, while polymers derived from vinyl compounds of the type
XCH=CH 2 are usually amorphous. In these vinyl polymers the X group
diminishes the linear symmetry of the chain ; moreover in the formation of
such polymers occasional inversions of the order of the units probably
occur.

The behavior of ethylene succinate (molecular weight, 3000) on crys-
tallization appears to be typical of many high molecular weight materials.
It separates from a melt or from concentrated solutions in chloroform as
doubly refracting microscopic spherulites which grow to what appear to be
star-like clusters of needles. Further growth leads to frost-like patterns.
The melt finally solidifies to an opaque porcelain-like mass. From dilute
alcohol solutions ethylene succinate separates in very thin, discrete needles,
but these lose their identity as soon as the solvent has evaporated. This

IX. POLYMERIZATION 133

behavior is highly characteristic of very long chains. It is reproduced
in all its details by such diverse materials as triacetylinulin (150), n-
heptacontane (51) and trimethylcellulose. The photomicrographs of
crystalline trimethylcellulose presented by Hess (149) would serve equally
well to represent ethylene succinate. Linear polymers in the form of
microcrystalline powders have a pronounced tendency to become elec-
trified, and they strongly adsorb considerable amounts of water vapor
even when they show scarcely any tendency to dissolve in water.

The crystallization of linear polymers probably involves the parallel
arrangement of the long chains into compact bundles, since this arrange-
ment enables the molecules to exert their maximum cohesive force (151).
Loose parallel swarms of molecules may also exist in melts or solutions of
the polymers. Molecules of identical length might be arranged in bundles
as shown in Fig. l(a), but especially with very long molecules, a less
regular type of structure such as that shown in (b) might be produced.
The arrangement shown in (b) has no sharp boundaries and this defect

FIG. la FIG. Ib

would be exaggerated if the molecules were not all of the same length.
Since the molecules of a specimen of high polymer are very long and have
not all the same length the lattice bundles first formed must more nearly
resemble (b) than (a). In the presence of solvent, such crystals might
persist as discrete particles, but in the absence of a solvent they would tend
to coalesce and lose their identity, owing to the absence of sharp bound-
aries and the incomplete neutralization of the residual forces of the pro-
jecting molecules. Thus, it is never possible to isolate large discrete (un-
solvated) crystals of high polymers. Moreover, though solid masses of
crystalline high polymers may be either hard and brittle or very tough, or
soft and wax-like, they never show any definite planes of cleavage. The
coalescence of the initial crystallites, which occurs as a molten mass of
polyester finally solidifies, must occur in a random and rather disordered
fashion, and it is probable also that the crystallites are cemented together
by molecules that have not succeeded in completely identifying themselves
with any particular crystallite.

The melting points of crystalline linear polymers show certain regulari-
ties. For a given molecular weight the melting points increase with the

134 POLYMERIZATION AND RING FORMATION

cohesive force (polarity) of the structural units. Polyamides have much
higher melting points than analogous polyesters. Polyesters derived
from short chain dibasic acids melt higher than those derived from the
longer chain acids; mixed polyesters melt lower than simple ones. For a
given polyester the melting point usually increases with increasing molecu-
lar weight up to a certain point, and after that it remains unchanged even
though the molecular weight be increased many fold. Melting points are
sometimes rather vague, but more frequently they are surprisingly sharp
even when the molecular weight is so high that the molten polymer shows
no sign of flowing and the only indication of melting is the disappearance
of opacity.

The question of the meaning of the term crystallinity in connection with
high polymers is rather confused. Linear polyesters whose molecular
weights lie below 5000 are definitely crystalline; they have sharp melting
points and the crystals can actually be seen under the microscope. The
evidences of crystallinity in polyesters whose molecular weights lie above
10,000 are somewhat more vague, but even these materials furnish sharp
x-ray powder diffraction patterns. Similar though less sharply defined
patterns are obtained from a transparent sheet of regenerated cellulose.
These patterns indicate that part at least of the molecules of such ma-
terials must be definitely ordered with respect to one another. On the
other hand, certain linear polymers, e. g., polystyrene, can be obtained in
the form of white powders which show no microscopic or x-ray evidence
of crystallinity. In these cases apparently the molecules tend to collect
into discrete aggregates of some kind, but not in a sufficiently orderly
fashion to exhibit any of the usual properties associated with crystals.

3. Mechanical Properties. High polymers are very extensively used
as structural materials in the construction of artifacts. One has only to
mention cellulose, silk, and rubber to indicate the great economic im-
portance of these non-chemical uses of organic materials. These uses
depend upon such properties as mechanical strength, toughness, pliability,
and elasticity. Such properties are found to a useful degree only among
polymers of very high molecular weight. The synthetic materials of this
class that have been most successfully used are 3 -dimensional polymers
such as Bakelite and the glyptals. These materials have considerable
strength, rigidity and toughness, but they are completely amorphous,
and they are greatly inferior to natural fibers in breaking strength and
pliability. The breaking strength of a flax fiber (100 kg. per sq. mm.)
is of the same order as that of a good grade of steel (152). The qualities
necessary for a useful fiber appear to be associated with a very high molecu-

IX. POLYMERIZATION 135

lar weight linear -polymeric structure and a certain degree of crystallinity
or definite order in the arrangement of the molecules. The relation be-
tween molecular structure and arrangement and the physical properties
of fibers has been most clearly recognized and discussed by Meyer and
Mark (3). In a natural cellulose or silk fiber the long molecular chains
are arranged in an ordered fashion parallel with the fiber axis. This state
of affairs is symbolized in Fig. 2. This arrangement provides the maximum

Fig. 2.

possible strength in the direction of the fiber axis since the mutual cohesive
force of the long chains is fully utilized. To rupture the fiber it is necessary
to cause the chains to slip past one another against this cohesive force as
indicated in the dotted line. A transparent sheet of regenerated cellulose
shows (by x-ray patterns) a certain degree of order in the arrangement of
its molecules, but there is no general orientation. This state of affairs is
symbolized in Fig. 3. This more or less random arrangement of ordered
molecular aggregates can be brought into the more highly ordered state
symbolized in Fig. 2 merely by the action of mechanical stress. Thus
the strength of a sheet of Cellophane that initially has approximately the
same strength in all directions can be so changed, merely by careful
stretching, that its strength along the axis of stretch is increased several
fold (153). At the same time its strength along the axis normal to its
stretch is considerably diminished. The strength of a rayon filament can
be increased several fold by the action of stress while it is in the spinning
bath, and a comparison of the x-ray patterns shows a much higher degree
of orientation along the fiber axis in the filament formed under stress. For
a rough mechanical analogy of the mechanism of this process one may pic-
ture a disordered mass of long straws (molecules) coated with a semi-
fluid adhesive (cohesive force). The gradual application of stress to such
a mass would finally bring the straws into parallel alignment where they
would more strongly cohere and resist the further action of stress.

The peculiarities of high polymers are nowhere more strikingly exempli-
fied than in this curious ability to accept permanent orientations through

136 POLYMERIZATION AND RING FORMATION

the action of mechanical stress. The properties of simple organic com-
pounds are, generally speaking, independent of their physical history;
they are completely determined by the nature of the molecules. Very
large molecules, however, are not capable of adjusting themselves instantly
to any changes in physical environment, and the properties of a very high
molecular weight material may vary over a wide range depending upon the
physical treatment it has received.

Bibliography and Remarks

(1) Staudinger, H., Helv. Chim. Ada, 12, 1183 (1929).

(2) Staudinger, H., Z. angew. Chem., 42, 37, 67 (1929).

(3) Meyer, K., and H. Mark, Der Aufbau der hochpolymeren orgamschen Natur staff e. Akade-
mische Verlagsgesellschaft, Leipzig (1930).

(4) Staudinger, H., Ber., 59, 3019 (1926).

(5) Scheiber, J., and K. Sandig, Die kunsllichen Harze. Wissenschaftliche Verlagsgesellschaft,
Stuttgard (1929).

(6) Meerwein, H., Houben's Die Mtthoden der organischen Chemte, 3rd edition, Volume II, p. 593.
Georg Thieme, Leipzig (1925).

(7) Meyer, K., Naturwissenschaften, 16, 781 (1928); Z. angew. Chem., 41, 935 (1928).
(8a) Hill, H. S., and H. Hibbert, Journ. Am. Chem. Soc., 45, 3124 (1923).

(8b) Franke, A., and E. Gigerl, Monalsh., 49, 8 (1928).

(9) Hess, K., Die Chemie der Cellulose, p. 574. Akademische Verlagsgesellschaft, Leipzig (1928).

(10) Ushakov, S. N., and A. V. Kon, C. A., 24, 3796 (1930).

(11) Ingold, C. K., and E. H. Ingold, Journ. Am. Chem. Soc., 1928, 2249.

(12) Berzelius, J. J., Jahresbericht, 12, 63 (1833).

(13) Reference 3, p. 47.

(14) Staudinger, H., and O. Schweitzer, Ber., 62, 2395 (1929).

(15) Roithner, E., Monalsh., 15, 665 (1894).

(16) Faworski, A., Chem. Zentr., 1907, I, 16.

(17) Lourenco, Ann. chim. phys., 67, 293 (1863).

(18) Carothers, W. H., Journ. Am. Chem. Soc., 51, 2548 (1929).

(19) Dietzel, R., and R. Krug, Ber., 68, 1307 (1925)1

(20) Willstatter, R., and C. Cramer, Ber., 43, 2976 (1910).

(21) Goldschmidt, S., E. Schulz and H. Bernard, Ann., 478, 1 (1930).

(22) Busch, M., and W. Schmidt, Ber., 62, 2612 (1929).

(23) Wagner- Jauregg, T., Ber., 63, 3213 (1930).

(24) Blaise, E., and A. Koehler, Compt. rend., 148, 1773 (1909).

(25) Ruzicka, L., and M. Stoll, Helv. Chim. Ada, 11, 1159 (1928).

(26) Johansson, H., Lunds Universitels Arsskrift, II (2), 12, 3 (1916).

(27) Johansson, H., Ber., 48, 1262 (1915).

(28) Blaise, E. E., and L. Marcilly, Bull. soc. chim., [3] 31, 308 (1904).

(29) Chuit, P., and J. Hausser, Helv. Chim. Acta., 12, 463 (1929).

(30) Lycan, W. H., and Roger Adams, Journ. Am. Chem. Soc., 51, 625, 3450 (1929).

(31) Bougalt, J., and L. Bourdier, Compt. rend., 147, 1311 (1908).

(32) Carothers, W. H., and J. A. Arvin, Journ. Am. Chem. Soc., 51, 2560 (1929).

(33) Carothers, W. H., and F. J. van Natta, Journ. Am. Chem. Soc., 52, 314 (1930).

(34) Carothers, W. H., and G. L. Dorougji, Journ. Am. Chem. Soc., 52, 711 (1930).

(35) Carothers, W. H., J. A. Arvin and G. L. Dorough, Journ. Am. Chem. Soc., 52, 3292 (1930).

(36) Kerschbaum, M., Ber., 60, 902 (1927).

(37) Marvel, C. S., and E. R. Birkhimer, Journ. Am. Chem. Soc., 51, 260 (1929).

(38) Staudinger, H., Ber., 53, 1073 (1920).

(39) Meyer, K., and H. Mark, Ber., 61, 1948 (1928).

(40) Davidoff, Ber., 19, 406 (1886).

(41) Vorlander, D., Ann., 280, 167 (1894).

(42) Tilitschejew, M., Journ. Russ. Phys. Chem. Soc., 57, 143 (1925); Chem. Zentr., 1926, I,
2667.

IX. POLYMERIZATION 137

(43) Kienle, R. H., and G. Hovey, Journ. Am. Chem. Soc. t 62, 3636 (1930).

(44) Carothers, W. H., J. W. Hill, J. E. Kirby and R. A. Jacobson, Journ. Am. Chem. Soc., 62,
5279 (1930).

(45) Franke, A., and O. Kienberger, Monatsh., 33, 1189 (1912).

(46) (a) Schlubach, H. H., and E. C. Goes, Ber., 66, 2889 (1922).

(46) (b) Ziegler, K., and H. Colonius, Ann., 136, 474 (1930).

(47) Schorigen, P., Ber., 43, 1931 (1910).

(48) (a) Gabriel, S., Ber., 22, 3338 (1889).

(48) (b) vSchotten, C., Ber., 21, 2240 (1888).

(49) (a) Carothers, W. H., and G. J. Berchet, Journ. Am. Chem. Soc., 62, 5289 (1930).

(49) (b) Gabriel, S., and T. A. Maass, Ber., 32, 1266 (1899).

(50) (a) v. Braun, J., Ber., 40, 1835 (1907).
(50) (b) Manasse, A., Ber., 36, 1367 (1902).

(50) (c) Diehl, I.., and A. Einhorn, Ber., 20, 377 (1887).

(51) Unpublished results.

(52) (a) v. Braun, J., and A. Steindorff, Ber., 38, 3083 (1905).
(52) (b) v. Braun, J., Ber., 43, 2853 (1910).

(52) (c) v. Braun, J., and C. Muller, Ber., 39, 4110 (1906).

(53) (a) Littmann, E. R., and C. S. Marvel, Journ. Am. Chem. Soc., 62, 288 (1930).

(53) (b) Knorr, L., and P. Roth, Ber., 39, 1420 (1906).

(54) Staudinger, H., and E. Ott, Ber., 41, 2214 (1908).

(55) Voerman, G. L., Rec. trav. chim., 23, 265 (1904).

(56) Hill, J. W., Journ. Am. Chem. Soc , 62, 4110 (1930).

(57) Grignard, V., and G. Vignon, Compl. rend., 144, 1358 (1907).

(58) Huckel, W., Forlschritte Chem., Physik. physik. Chem., 19, No. 4, 7 (1927).

(59) Lonsdale, K., Proc. Roy. Soc. (London), 123, 494 (1929).

(60) King, G., J. Oil Colour Chem. Assocn., 13, 28 (1930).

(61) Le Fevre, J. W., Journ. Chem. Soc., 1929, 733.

(62) Ferber, E., Ber., 62, 183 (1929).

(63) Smyth, C. P., and W. vS. Walls, Journ. Am. Chem. Soc., 63, 527 (1931).

(64) Tilitschejew, M., J. Russ. Phys. Chem. Soc., 68, 447 (1926); C. A. 21, 3358 (1927).

(65) (a) Gane. R., and C. K. Ingold, Journ. Chem. Soc., 1928, 1594.

(65) (b) Gane, R., and C. K. Ingold, Journ. Chem. Soc., 1928, 2267.

(66) Mohr, E., J. prakl. Chem., 98, 315 (1918).

(67) Stoll, M., and G. Stoll-Comte, Helv. Chim. Ada, 13, 1185 (1930).

(68) v. Braun, J., and B. Bartsch, Ber., 46, 3376 (1912).

(69) Stanley, W. M., and R. Adams, Journ. Am. Chem. Soc., 62, 1200 (1930).

(70) Ruggli, P., Ann., 392, 92 (1912).

(71) (a) Ruzicka, L., M. Stoll, H. W. Huyser and H. A Boekenoogen, Helv. Chim. Ada, 13, 1153
(1930).

(71) (b) Ruzicka and co-workers, Helv. Chim. Ada, 9, 230, 249, 339, 389, 399, 499, 715, 1008
(1926); 10, 695 (1927); 11, 496, 670, 686, 1174 (1928).

(72) Fichter, F., and A. Beisswenger, Ber., 36, 1200 (1903).

(73) Hollo, E., Ber., 61, 895 (1928).

(74) Drew, H. D. K., and W. N. Haworth, Journ. Chem. Soc. t 1927, 775.

(75) Bischoff, C. A., and P. Walden, Ann., 279, 45 (1893).

(76) Bischoff, C. A., and P. Walden, Ber., 26, 262 (1893).

(77) Bischoff, C. A., and P. Walden, Ann., 279, 71 (1893).

(78) Carter, S. R., W. N. Haworth and R. A. Robinson, Journ. Chem. Soc., 1930, 2125.

(79) Johannson, H., and H. Sibelius, Ber., 62, 745 (1919).

(80) (a) Balbiano, L., Ber., 34, 1501 (1901).

(80) (b) Balbiano, L., and D. Trasciatti, Ber., 33, 2323 (1900).

(81) (a)'Maillard, L. C., Ann. chim., 1, 519 (1914).

(81) (b) Maillard, L. C., Ann. chim., 2, 210 (1914).

(82) Curtius, T., and A. Benrath, Ber., 37, 1279 (1904).

(83) Shibata, K., Ada. Phylochim., 2, 39 (1925); 2, 194 (1926).

(84) (a) Curtius, T., Ber., 37, 1284 (1904).

(84) (b) Curtius, T., Ber., 16, 753 (1883).

(85) Fischer, E., Ber., 39, 2924 (1906).

(86) Delepine, M., Bull. soc. chim., [4] 27, 740 (1920).

138 POLYMERIZATION AND RING FORMATION

(87) Grichkevitch-Trokhimovsky, E., Bull. soc. chim., [4] 24, 541, 543, 546, 548, 549, 551 (1918).

(88) Katz, J. R., Z. angew. Chem., 41, 329 (1928).

(89) (a) Leuchs, H., and W. Geiger, Ber., 41, 1721 (1908).

(89) (b) Leuchs, H., Ber., 39, 857 (1906).

(90) Curtius, T., and W. Sjfjber, Ber., 55, 1543 (1922).

(91) Cohen, J. B., Otganic\:hemistry, Part I, p. 193. Edward Arnold and Company, London
(1928).

(92) Stanley, H. M., Journ. Soc. Chem. Ind., 49, 349T (1930).

(93) Mignonac, G., and R. V. de Saint-Aunay, Bull. soc. chim., 47, 522 (1930).

(94) Pease, R. H., Journ. Am. Chem. Soc., 52, 1158 (1930).

(95) Lebedev, S. V., and G. G. Koblyanskii, Ber., 63, 1432 (1930).

(96) (a) Staudinger, H., M. Brunner, K. Frey, P. Garbsch, R. Signer and S. Wehrli, Ber., 62, 241
(1929).

(96) (b) Staudinger, H., K. Frey, P. Garbsch and S. Wehrli, Ber., 62, 2912 (1929).

(96) (c) Staudinger, H., and M. Brunner, Helv. Chim. Acta, 12, 972 (1929).

(96) (d) Milas, N. A., Proc. Natl. Acad. Sci., 14, 844 (1928).

(97) (a) Staudinger, H., M. Brunner and W. Feist, Helv. Chim. Acta, 13, 805 (1930).

(97) (b) Flumiani, G., Z. Elektrochem., 32, 221 (1926).

(98) (a) Staudinger, H., K. Frey and W. Starck, Ber., 60, 1782 (1927).

(98) (b) Whitby, G. S., J. C. McNally and W. Gallay, Trans. Roy. Soc. Can., 22, 27 (1928).

(98) (c) Hermann, W. O., and W. Haehnel, Ber., 60, 1658 (1927).

(99) Staudinger, H., and E. Urech, Helv. Chim. Acta, 12, 1107 (1929).

(100) Moureu, C., and C. Dufraisse, Ann. chim., 15, 158 (1921).

(101) (a) Staudinger, H., A. A. Ashdown, M. Brunner, H. A. Bruson and S. Wehrli, Helv. Chim.
Acta, 12, 934 (1929).

(101) (b) Staudinger, H., H. Johner and V. Wiedersheim, Helv. Chim. Acta, 12, 958 (1929).

(101) (c) Staudinger, H., H. Johner, G. Schiemann and V. Wiedersheim, Helv. Chim. Acta, 12, 962
(1929).

(102) (a) Staudinger, H., and W. Heuer, Ber., 63, 222 (1930).
(102) (b) Staudinger, H., and K. Frey, Ber., 62, 2909 (1929).
(102) (c) Staudinger, H., and R. Machemer, Ber., 62, 2921 (1929).

(102) (d) Staudinger, H., Kolloid-Z., 61, 71 (1930).

(103) (a) Staudinger, H., E. Geiger and E. Huber, Ber., 62, 263 (1929).

(103) (b) Staudinger, H., and V. Wiedersheim, Ber., 62, 2406 (1929).

(104) Staudinger, H., and F. Breusch, Ber., 62, 442 (1929).

(105) Whitby, G. S., and M. Katz, Journ. A)n. Chem. Soc., 50, 1160 (1928).

(106) Starkweather, H. W., and G. B. Taylor, Journ. Am. Chem. Soc., 52, 4708 (1930).

(107) Taylor, H. S., private communication.

(108) (a) Schotz, S. P., Synthetic Rubber. E. Benn, London (1926).

(108) (b) Pankow, M., Kunststoffe, 20, 219, 248 (1930).

(109) Midgley, T., Jr., and A. L. Henne, Journ. Am. Chem. Soc., 52, 2077 (1930).

(110) (a) Staudinger, H., R. Signer, H. Johner, M. Liithy, W. Kern, D. Russidis and O. Schweit-
zer, Ann., 474, 145 (1929).

(110) (b) Staudinger, H., and M. Liithy, Helv. Chim. Acta, 8, 41, 65, 67 (1925).
(110) (c) Staudinger, H., H. Johner, R. Signer, G. Mie and G. Hengstenberg, Z. physik. Chem.,
126, 425 (1927).

(110) (d) Staudinger, H., and R. Signer, Helv. Chim. Acta, 11, 1047 (1928).
(110) (e) Mie, G., and J. Hengstenberg, Helv. Chim. Acta, 11, 1052 (1928).
(110) (f) Staudinger, H., and R. Signer, Z. Krist., 70, 193 (1929).
(110) (g) Kohlschutter, H. W., Ann., 482, 75 (1930).

(110) (h) Ott, E., Z. physik. Chem., Abt. B, 9, 378 (1930).

(111) (a) Conant, J. B., and C. C. Tongberg, Journ. Am. Chem. Soc., 52, 1659 (1930).

(111) (b) Bridgman, P. W., and J. B. Conant, Proc. Nail. Acad. Sci., 15, 680 (1929).

(112) Kienle, R. H., and A. G. Hovey, Journ. Am. Chem. Soc., 51, 509 (1929).

(113) (a) Koebner, M., Chem.-Ztg., 54, 619 (1930).

(113) (b) Pollak, F., and F. Riesenfeld, Z. angew. Chem., 43, 1129 (1930).

(114) Tschugaeff, L., Ber., 39, 3199 (1906).

(115) Rdhm, Dissertation, Tubingen, 1901. Cited in reference 99.

(116) Schroeter, G., Ber., 49, 2697 (1916).

(117) Schroeter, G., Ber., 52, 2224 (1919).

IX. POLYMERIZATION 139

(118) (a) Abderhalden, E., Naturwissenschaften, 12, 716 (1924).

(118) (b) Bergmann, M., Naturwissenschaften, 13, 1045 (1925).

(118) (c) Pringsheim, H., Naturwissenschaften, 13, 1084 (1925).

(118) (d) Bergmann, M., Ber., 59, 2973 (1926).

(118) (e) Hess, K., Ann., 435, 1 (1924).

(119) (a) Meyer, K. H., and H. Mark, Ber., 61, 593 (1928).

(119) (b) Mark, H., and K. H. Meyer, Z. physik. Chem., Abt. B, 2, 115 (1929).

(119) (c) Mark, H., and G. v. Susich, Z. physik. Chem., Abt. B, 4, 431 (1929).

(120) Haworth, W. N., C. W. Long and J. H. G. Plant, Journ. Chem. Soc., 1927, 2809.

(121) (a) Freudenberg, K., and E. Braun, Ann., 460, 288 (1928).

(121) (b) Freudenberg, K., Ann., 461, 130 (1928).

(122) (a) Irvine, J. C., Chem. Reviews, 1, 41 (1924).

(122) (b) Freudenberg, K., Ber., 54, 767 (1921).

(123) (a) Sponsler, O. L., and W. H. Dore, Colloid Symposium Monograph, 4, 174 (1926). Chemi-
cal Catalog Company, New York.

(123) (b) Sponsler, O. L., Nature, 125, 633 (1930).

(124) (a) Staudinger, H., and O. Schweitzer, Ber., 63, 3122 (1930).
(124) (b) Staudinger, H., and O. Schweitzer, Ber., 63, 2317 (1930).

(124) (c) Staudinger, H., and H. Freudenberger, Ber., 63, 2331 (1930).

(125) Willstatter, R., and L. Zechmeister, Ber., 46, 2401 (1913).

(126) Stamni, A. J., Journ. Am. Chem. Soc. t 52, 3047 (1930).

(127) Bergmann, M., and H. Machemer, Ber., 63, 316 (1930).

(128) Willstatter, R., and L. Zechmeister, Ber., 62, 722 (1929).

(129) (a) Kuhn, W., Ber., 63, 1503 (1930).

(129) (b) Freudenberg, K., W. Kuhn, W. Diirr, F. Bolz and G. Steinbrunn, Ber., 63, 1528 (1930).

(130) Bergmann, M., and H. Machemer, Ber., 63, 2304 (1930).

(131) Harries, C., Ber., 37, 2708 (1904); 38, 3985 (1905).

(132) Thiele, J., Ann., 306, 92 (1899).

(133) Pickles, S. S., Journ. Chem. Soc., 98, 1085 (1910).

(134) (a) Staudinger, H., and J. Fritschi, Hell). Chim. Ada, 5, 785 (1922).
(134) (b) Staudinger, H., Ber., 57, 1203 (1924).

(134) (c) Staudinger, H., and W. Widmer, Helv. Chim. Acta, 7, 842 (1924).

(134) (d) Staudinger, H., Z. angew. Chem., 38, 226 (1925).

(134) (e) Staudinger, H., and E. Geiger, Helv. Chim. Acta, 9, 549 (1926).

(134) (f) Staudinger, H., Kautschuk, 1925, 5.

(134) (g) Staudinger, H., Kautschuk, 1927, 63.

(134) (h) Staudinger, H., M. Asano, H. F. Bondy and R. Signer, Ber., 61, 2575 (1928).

(134) (i) Staudinger, H., H. F. Bondy and E. Geiger, Ann., 468, 1 (1929).

(134) (j) Staudinger, H., and H. F. Bondy, Ann., 468, 1 (1929).

(134) (k) Staudinger, H., Kautschuk, 5, 94, 126 (1929).

(134) (1) Staudinger, H., and H. F. Bondy, Ber., 62, 2411 (1929).

(134) (m) Staudinger, H., and H. F. Bondy, Ber., 63, 724 (1930).

(134) (n) Staudinger, H., and E. O. Leupold, Ber., 63, 730 (1930).

(135) (a) Staudinger, H., and H. F. Bondy, Ber., 63, 734 (1930).
(135) (b) Staudinger, H., Ber., 63, 921 (1930).

(135) (c) Staudinger, H., and H. Joseph, Ber., 63, 2888 (1930).
(135) (d) Staudinger, H., Kautschuk, 6, 153 (1930).

(135) (e) Staudinger, H., and H. F. Bondy, Ber., 63, 2900 (1930).

(136) Pummerer, R., Z. angew. Chem., 43, 757 (1930); Pummerer, R., G. Ebermayer and K.
Gerlack, Ber., 64, 809 (1931).

(137) Whitby, G. S., Trans. Inst. Rubber Ind., 5, 184 (1929); 6, 40 (1930).

(138) (a) Katz, J. R., Chem.-Ztg., 49, 353 (1925).

(138) (b) Meyer, K. H., and H. Mark, Ber., 61, 1939 (1928).

(139) Hock, L., Gummi-Ztg., 39, 1740 (1925).

(140) Fikentscher, H., and H. Mark, Kautschuk, 6, 2 (1930); Rubber Chemistry and Technology, 8,
201 (1930).

(141) Dunkel, M., Z. physik. Chem., Abt. A, 138, 42 (1928).

(142) Whitby, G. S., J. C. McNally and W. Gallay, Trans. Roy. Soc. Can., 22, 27 (1928).

(143) Lee, W. B., and P. J. Van Rysselberge, Journ. Phys. Chem., 33, 1543 (1929).

(144) Katz, J. R., Z. angew. Chem., 42, 828 (1929).

140 POLYMERIZATION AND RING FORMATION

(145) Ruzicka, L., M. Stoll, H. W. Huyser and H. A, Boekenoogen, Helv. Chem. Acta, 13, 1152
(1930).

(146) (a) Staudinger, H., Kolloid-Z., 53, 19 (1930).

(146) (b) Signer, R., Z. physik. Chem., Abt. A, 150, 257 (1930).

(147) Meyer, V., and P. Jacobson, Lehrbuch der organischen Chemie, zweiter Band, drilter Teil.
Walther de Gruyter and Company, Berlin and Leipzig (1920).

(148) Wittig, G., Stereochemie, p. 277. Akademische Verlagsgesellschaft, Leipzig (1930).

(149) Reference 9, p. 432.

(150) Pringsheim, H., and W. G. Hensel, Ber., 63, 1096 (1930).

(151) Reference 3, p. 50.

(152) Reference 3, p. 153.

(153) Mark, H., Melliands' Textilber., 10, 695 (1929).

(154) (a) Zappi, E. V., Bull. soc. chim., 19, 249 (1916).

(154) (b) v. Braun, J., and W. Sobecki, Ber., 44, 1918 (1911).

(155) Grignard, V., and Tcheoufaki, Compt. rend., 188, 357 (1929).

(156) Konrad, E., O. Bachle, and R. Signer, Ann., 474, 276 (1929).

(157) Kipping, F. S., A. G. Murray and J. G. Maltby, Journ. Chem. Soc., 1929, 1180.

(158) Schdnberg, A., Ber., 62, 195 (1929).

(159) Brit, patents, 279,406 and 314,524.

(160) Morgan, G. T., and F. H. Burstall, Journ. Chem. Soc., 1930, 1497.

(161) Mansfield, W., Ber., 19, 696 (1886).

(162) v. Braun, J., and A. Triimpler, Ber., 43, 547 (1910).

(163) v. Braun, J., Ber., 43, 3220 (1910).

(164) Wallach, O., Ann., 312, 205 (1900).

(165) Lind, S. C., and G. Glockler, Journ. Am. Chem. Soc., 51, 2811 (1929).

(166) (a) Harkins, W. D., and D. M. Gans, Journ. Am. Chem. Soc., 52, 5165 (1930).

(166) (b) Austin, J. B., and I. A. Black, Journ. Am. Chem. Soc., 52, 4552 (1930).

(167) (a) Zelinsky, N., Ber., 57, 264 (1924).

(167) (b) Fischer, F., F. Baugert and H. Pichler, Brennstoff-Chem., 10. 279 (1929).

(167) (c) Pease, R. N., Journ. Am. Chem. Soc., 51, 3470 (1929).

(167) (d) Schlapfer, P., and M. Brunner, Helv. Chim Acta, 13, 1125 (1930).

(168) (a) Sabatier, P., and J. B. Senderens, Bull. soc. chim. 25, 683 (1901).
(168) (b) Brit, patent, 303,797.

(168) (c) Schlapfer, P., and O. Stadler, Helv. Chim. Acta, 9, 185 (1926).

(169) (a) Lind, S. C., and R. S. Livingston, Journ. Am. Chem. Soc., 52, 4613 (1930).

(169) (b) Bates, J. R., and H. S. Taylor, Journ. Am. Chem. Soc., 49, 2438 (1927).

(170) (a) McLennon, J. C., M. A. Perrin and J. C. Ireton, Proc. Roy. Soc. (London), 125A, 246
(1929).

(170) (b) Coolidge, W. D., Science. 62, 441 (1925).

(171) Lind, S. C., D. C. Bardwell and J. H. Perry, Journ. Am. Chem. Soc., 48, 1556 (1926).

(172) Mignonac, G., and R. V. de Saint-Aunay, Compt. rend., 188, 959 (1929); Bull. soc. chim., 47,
14 (1930).

(173) Hunter, W. H., and G. H. Woolett, Journ. Am. Chem. Soc., 43, 135 (1921).

(174) (a) Reference 3, p. 173.

(174) (b) Kraemer, E. O., and G. R. Sears, Journ. Rheology, in press.
(174) (c) Loewen, H., Kautschuk, 7, 12 (1931).
(174) (d) Mark, H., Kolloid-Z., 53, 32 (1930).

X. SIX-MEMBERED CYCLIC ESTERS 141

X. The Reversible Polymerization of Six-Membered Cyclic

Esters*

This publication describes the reversible polymerization of six-membered
cyclic esters t which has already been observed in a previous paper of this
series, f

Such behavior has been found in several cases by different authors and
seems to be fairly common with six-membered cyclic esters or 8-lactones;
it does not occur however with e-lactones nor with rings of a higher number
of members.

Different linear polymers of d-valerolactone were prepared up to molecular
weights of 2240, at temperatures between 20 and 150, with and without
catalysts. Their melting points ranged between 35 and 55. Poly-
merization took place in 3 to 29 days. The samples represent crystalline
powders; they are all acidic. Titration gave molecular weights which
closely agreed with the boiling point values.

Depolymerization takes place if the samples are heated and presumably
consists of an ester interchange. The behavior of the six-membered rings
is in accordance with the stereochemical theory of Sachse and Mohr.

Reactions leading to the formation of high polymers have received con-
siderable attention during the past few years both because of their inherent
interest as representatives of a realm that has been relatively little ex-
plored, and because of their bearing on the formulation of naturally
occurring macromolecular materials. In this connection the behavior of
six-membered cyclic esters is of peculiar interest : they combine with them-
selves, in many cases spontaneously at the ordinary conditions, to form
polymers of high molecular weight, and this transformation is reversible.
Since these esters are generally free from unsaturation, their self-combi-
nation must have its origin in some peculiarity of the heterocyclic system.
Six-membered heterocycles are quite common among naturally occurring
materials and are closely associated with natural polymers (sugars and
polysaccharides, diketopiperazines and polypeptides) ; it is therefore rea-
sonable to assume that the behavior of the cyclic esters lies closer to the
natural processes that result in the formation of macromolecular substances

* W. H. Carothers, G. L. Dorough and F. J. van Natta; Journ. Am. Chem. Soc.,
54, 761-72 (1932) ; Contribution No. 72 from the Experimental Station of E. I. du Pont
de Nemours and Co.

Received September 24, 1931. Published February 5, 1932.

t W. H. Carothers and F. J. van Natta; Journ. Am. Chem. Soc., 52, 314 (1930);
pages 29 to 42 of this volume and W. H. Carothers, J. A. Arvin and G. L. Dorough;
Journ. Am. Chem. Soc., 52, 3292 (1930); this volume pages 54 to 62.

142 POLYMERIZATION AND RING FORMATION

than does the polymerization of unsaturated compounds, such as styrene,
which have been more extensively studied.

Several isolated examples of the reversible polymerization of six-mem-
bered cyclic esters have been recorded in the literature but the question of
the mechanism of this phenomenon has received very little attention.
In the present paper, we report some observations bearing on this point.

Previous Work. Fichter and Beisswenger (1) first reported that "5-
valerolactone (I) is characterized by a remarkable and surprising phenome-
non: it polymerizes after a short time; the initially mobile oil becomes
gradually thicker and finally solidifies to a crystalline mass." The polymer
after recrystallization melted at 47-48. Attempts to determine its mo-
lecular weight gave values ranging from a five- to a seven-fold polymer.
Later the same transformation was observed by Hollo (2), who also records
the spontaneous polymerization of the lactone (II) of hydroxyethylglycolic
acid.

Bischoff and Walden (3) in 1893 described the transformation of glycolide
(III) under the influence of heat or a trace of zinc chloride into a polymeric
solid melting at 220. On being distilled in a vacuum it was reconverted
to the monomer, melting at 86-87.

Drew and Haworth (4) have obtained the lactone (IV) of 2,3,4-trimethyl-
/-arabonic acid in crystalline form (m. p. 45) and observed that, in the
presence of traces of hydrogen chloride, it is converted into a crystalline
polymeric powder which has a considerably higher melting point, a lower
solubility and a lower specific rotation than the monomer. Its molecular
weight (about 2000) indicates that it is derived from about ten molecules
of the monomer. At 175 it distils completely in vacua, and the dis-
tillate consists of pure lactone. Drew and Haworth (5) favored a linear
polyester structure for this polymer.

Trimethylene carbonate (V) was first described by Carothers and van
Natta (6). The monomer is a very soluble crystalline solid melting at 47.
If it is heated to 130 for a few minutes with a trace of potassium carbonate,
the mobile melt suddenly becomes very viscous and evolves a small amount
of gas. The colorless, viscous sirup on cooling solidifies to a stiff mass
which shows an apparent molecular weight of about 4000. When this mass
is heated in a vacuum it distils almost quantitatively, and the distillate
consists of pure monomer.

Monomeric ethylene oxalate (VI) was first described by Bischoff and
Walden (7), who obtained it as a crystalline solid melting at 143 by distill-
ing the product of the action of ethylene glycol on monoethyl oxalate.
Later Bischoff (8) obtained less soluble, higher melting forms of ethylene

X. SIX-MEMBERED CYCLIC ESTERS 143

oxalate, and showed that these on distillation were converted into the 143
form. Bischoff also observed that the melting points of the ethylene
oxalates change on standing, and suspected that these changes might be
due to a reversible polymerization, but he reported no comparative mo-
lecular weight data. The relation of the various forms of ethylene oxalate
to one another has lately been studied by Carothers, Arvin and Borough
(9). It was found that the flat, diamond-shaped crystals of the monomer

CO CO CO

R-dS'b R-dH O

o ca o CH-R

YH,

I R = H* II. R - H III. R - H

la, R = n-C,H, Ha, R - C.H, Ilia, R - CH,

CO CO CO

<A> Ao

CH.O-CHCH, CH.CH, R-CH6

\/ \/ \d

CHOCH. CH> CHi

IV V VI. R - H

/ C \

I I /0-(CH,)-<X ^^^9

CH, CH, (fo \ CO ((5H0.4 I

H, CH, \0-(CH,V-0/ ^-O

VII VIII 3t

CO0-- (CH 2 )r-0 CO COO (CH,), O~CO

(CH,),

\CO--O-~(CH 2 ), O (X) CO O (CH 2 )r-0--CO

IX XI

completely disintegrate during the course of a few days, yielding a micro-
crystalline powder which consists of a mixture of polymers. From this
mixture by careful extraction with cold solvents two definite fractions
can be isolated. One of these melts at 159 and has an apparent molecular
weight of about 3000; the other fraction, which is probably much more
highly polymeric, melts at 173 and is too insoluble for molecular weight
determinations. These isolated fractions on standing lose their identity;

144 POLYMERIZATION AND RING FORMATION

they revert spontaneously within a few days to more complicated mix-
tures which usually contain appreciable amounts of monomer. The poly-
meric ethylene oxalates can be converted into the monomer (with con-
siderable loss) by vacuum distillation. Propylene oxalate (Via) (m. p.
142) is converted by the action of heat into an insoluble, microcrystalline
polymer melting at 178 (9).

Generality of the Phenomenon. The examples cited above show that
the capacity to undergo reversible polymerization is quite common among
six-membered cyclic esters. The presence of substituent groups, however,
has a considerable effect and, in general, a depressing effect, on this tend-
ency. The unsubstituted esters I, II and VI polymerize spontaneously
at ordinary temperature in the complete absence of any added catalyst;
substituted esters on the other hand, such as la, Ila, IV and Via, require
the action of heat, or catalysts, or both, and in some cases polymerization
fails altogether.

A number of substituted S-valerolactones (I) have been described in the
literature, but without any record of their polymerization. We have
therefore prepared a-w-propyl-S-valerolactone (la) as described in the
experimental part. Although 5-valerolactone was found to be completely
polymerized to a waxy solid after twenty-nine days at room temperature,
the new a-w-propyl-6-valerolactone was still unchanged after twelve
months, nor did it show any signs of polymerization after being heated to
80 for one month. However, when heated to 80 for one month in the
presence of a trace of potassium carbonate or zinc chloride, it became more
viscous, and its apparent molecular weight rose to 1100-1200.

Hollo (2) records the spontaneous polymerization of the lactone (II) of
hydroxyethylglycolic acid, but makes no mention of the polymerization
of its a-alkyl derivatives (II, R = C 2 H B , w-C 3 H 7 and is0-C 3 H 7 ) which he
prepared at the same time. We have made attempts to polymerize the
lactone (Ila) of hydroxy ethyl- a-hydroxybutyric acid. Sixteen hours of
heating at 140-160 failed to produce any change in the viscosity or the
color of the pure lactone. Under the same conditions in the presence of
a trace of potassium carbonate, only a slight increase in viscosity occurred.

Bischoff and Walden (10) state that lac tide (Ilia) does not polymerize
under the conditions that result in the polymerization of glycolide (III).
We find, however, that at 250-275 the polymerization of lactide is quite
rapid : in two hours a sample was transformed into a resinous mass which
showed an apparent molecular weight of about 3000. A similar effect
can be obtained at much lower temperatures, e. g. t 140-150, if potassium
carbonate is present.

X. SIX-MEMBERED CYCLIC ESTERS 145

We have also examined some of the higher homologs of lac tide. The
cyclic esters derived from a-hydroxycaprylic and from a-hydroxy palmitic
acids were prepared and heated for various periods both in the presence and
absence of catalysts. Evidences of polymerization could be found only at
temperatures sufficiently high to produce considerable decomposition.

Finally, it may be mentioned that although ethylene oxalate polymerizes
spontaneously at the ordinary temperature, propylene oxalate undergoes
a similar transformation only at considerably elevated temperatures (9).

Peculiarity of the Phenomenon. Among cyclic esters the capacity to
undergo reversible polymerization in the manner illustrated by the ex-
amples cited above is peculiar to six-membered rings. The 7-lactones and
other five-membered cyclic esters show no tendency to polymerize, and no
corresponding polymers are known. Thus, we have heated samples of pure
7-butyrolactone both with and without catalysts (zinc chloride, potassium
carbonate) at 80 for twelve months; none of the samples showed any
detectable increase in viscosity. We have also made various attempts to
polymerize ethylene carbonate, but these attempts were all unsuccessful,
although the corresponding six-membered ring, trirnethylene carbonate
(V), is readily polymerized (11).

It appears also that cyclic esters of more than six members show no tend-
ency to polymerize spontaneously. Information on this point is some-
what incomplete. No aliphatic cyclic esters of nine to thirteen atoms
have been described. In fact no practical means for the preparation of
such esters are yet known. In general, the attempt to prepare cyclic esters
of more than six atoms from open chain compounds leads to linear polymers.

Thus the six-atom ring trirnethylene carbonate (V) is readily obtained
by heating diethyl carbonate with trimethylene glycol, and subsequently
distilling the product in vacuo, but when diethyl carbonate is similarly
treated with tetramethylene glycol, the product is a polymeric material
that is not capable of being distilled in vacua, i. e., it cannot be depoly-
merized; only at temperatures above 300, it undergoes complete thermal
decomposition and yields a complicated mixture of products among which
are found very small amounts of the dimeric fourteen-membered ring VIII
(6). Attempts to detect the presence of the monomer VII have been
unsuccessful. Similar behavior is observed with trimethylene oxalate (9)
and ethylene succinate (12) . The preparation of these esters leads to linear
polymeric products of high molecular weight. No smooth depolymeriza-
tion of these polymers can be effected by the action of heat; but at high
temperatures they undergo complete thermal decomposition, yielding a
complicated mixture of gaseous, liquid and tarry products containing

146 POLYMERIZATION AND RING FORMATION

very small amounts of the cyclic dimers which are, respectively, fourteen
and sixteen-membered rings (XI and IX) The corresponding mono-
mers have never been obtained.

The following observations indicate the relatively great stability of the
large cyclic esters. Specimens of dimeric trimethylene oxalate (XI),
dimeric ethylene succinate (IX), and exaltolide (X) have been preserved
in the laboratory for two years without showing any signs of change. The
last two of these compounds have been heated in sealed tubes at 170 for a
considerable period of time. After twelve hours they were quite un-
changed; after forty- three hours they had become slightly colored and their
melting points slightly lowered, but signs of appreciable polymerization
were absent (13).

The peculiar position of the six-membered esters is shown in the follow-
ing outline, in which, for simplicity, the starting material in each case is
represented as an hydroxy acid.

(A) The structural unit is five atoms long

-H 2 O

HO-R COOH o-R CO-i
+H 2 ' >

(B) The structural unit is six atoms long

~H 2 O
HO k COOH

+H 2 1 I -H 2

I \

*CO

HO R -COOH
+H 2 o1 ^-H 2 O
O R CO--O R CO O R CO O R CO *

Under B it should be mentioned that the reversible relation between
the monomer and the polymer complicates the problem of deciding in a
given case which of these is the real primary product of the dehydration
of the hydroxy acid. In any event the conditions of the dehydration
usually favor the polymerization of the monomer, so that the latter is
isolated only by vacuum distillation of the actual reaction product. Diet-

X. SIX-MEMBERED CYCLIC ESTERS 147

zel and Krug (14) have presented evidence to show that the self-esterifica-
tion of lactic acid leads directly only to polylactyl lactic acids and that
lactide, when it is produced, results from the depolymerization of these
polyesters. On the other hand, it is established, at least in certain in-
stances (4), that 5-lactones may be formed as such directly from the
corresponding hydroxy acid. No doubt in most cases, depending upon
the conditions, either the monomer or the polymer may be formed as
the primary product.

Under C it should be mentioned that although attempts to prepare
seven-membered cyclic esters from dibasic acids and dihydric alcohols
lead only to polyesters, e-hydroxy acids appear to be dehydrated with the
simultaneous formation of polymer and monomer (15). There is, how-
ever, no record of the interconversion of these two forms, and the higher
a>-hydroxy acids yield linear polymers exclusively (16).

Mechanism of the Phenomenon. The smooth reversibility of the
polymerization of the six-membered cyclic esters and the absence of double-
bond unsaturation at first suggested that here, if anywhere, might be
found the missing models of that hypothetical phenomenon, association
polymerization, to which for a time the peculiarities of natural polymers
were widely attributed (17). This thought was considerably weakened,
however, when Dr. J. W. Hill presented in the closely analogous case of
adipic anhydride, a direct and decisive proof of the structural difference
between monomer and polymer (18). No similar proof is possible in the
esters under consideration; nevertheless, all the evidence together con-
clusively favors the mechanism suggested by Carothers and van Natta (6),
namely, that the polymers are linear polyesters and that both the poly-
merization and its reversal proceed by a process of ester interchange. This
is indicated in the equation

O R CO + O R CO + O R CO + O R CO -f etc. >

rr_ o R CO^O-~R-~CO--O-~R---- CO O R CO O R CO

The peculiar position occupied by the six-membered cyclic esters may be
explained, as we have already suggested (6), by stereochemical factors:
rings of six atoms are strained (or may pass through positions of strain) ;
rings of five atoms or more than ten atoms are strainless. In the cyclic
esters the strain can be relieved by the process of ester interchange, which
results in polymerization. The resulting polymers (from the six-membered
esters) are easily depolymerized owing to the high probability of the close

148 POLYMERIZATION AND RING FORMATION

approach of atoms six atoms apart in a chain. Polyesters whose de-
polymerization would result in larger monomeric rings are not depoly-
merized; the probability of the close approach of the requisite atoms is
too slight.

The interchange mechanism for the polymerization of the six-membered
esters is supported by the following facts: (1) The polymers have the same
apparent analytical compositions (19), and saponification equivalents
as the corresponding monomers, and they yield the same products on
hydrolysis.

(2) The products of the polymerization of the six-membered cyclic esters
closely resemble in their physical properties the polymers formed by
the action of dibasic acids on glycols, or by the self-esterification of the
higher w-hydroxy acids. It has been definitely established that these
polymers are linear polyesters; the chains are open and are terminated
by hydroxyl and/or carboxyl groups (12, 20).

(3) Both the forward and the reverse transformations are catalyzed by
acids and bases, typical ester-interchange catalysts.

(4) The speed of the polymerization runs parallel with the susceptibility
to hydrolysis. The much greater susceptibility of 5-lactones to hydrolysis
as compared with analogous 7-lactones is well known. The hydrolysis
constant for 5-valerolactone (I) is more than twice that of its a-methyl
derivative (2), and the former polymerizes spontaneously while the latter
does not. Similar relations hold for the lactone of hydroxyethylglycolic
acid (II) and its a-alkyl derivatives (2). The hydrolytic constants for
glycolide in its two stages are, respectively, 0.0179 and 0.119, while the
corresponding values for lactide are 0.00313 and 0.0611 (21). The former
polymerizes much more readily than the latter. Ethylene oxalate, which
polymerizes with extraordinary facility, is so sensitive to hydrolysis that,
like an acid, it can be titrated directly with moderately warm dilute alkali.

Further Observations. Data on some polymers of 6-valerolactone are
presented in Table I.

TABLE I
POLYMERS OF S-VALEROLACTONE

No.

Catalyst

Temp,
during

Time

M. p. of
polymer, C.

Mol. wt. of
polymer
(boiling point in
benzene)

None

Room

29 days

35-40

1060 1060

None

80-85

13 days

52-53

1270 1330

K 2 CO 3

80-85

5 days

53-54

1840 1820

ZnCl 2

80-85

5 days

52-54

2230 1846

None

150

10 hrs.

52-55

2110 2240

Over CHsCOCl

Room

3 days

50-51

1720 1720

X. SIX-MEMBERED CYCLIC ESTERS 149

The melting points are in general somewhat higher than those reported
by Fichter and Beisswenger (1). The polymer is soluble in a variety of
organic solvents and is readily crystallized, e. g., from a mixture of petro-
leum ether and benzene. When allowed to solidify from a melt it forms
an opaque, soft, waxy solid. The crystals are poorly developed; they ap-
pear under the microscope as tiny irregular particles.

Polymeric 5-valerolactone is acidic. In acetone solution it can be
titrated with alkali to a sharp end-point which persists for several minutes.
Sample 5 of Table I thus showed a neutral equivalent of 2153 and 2243
in two determinations. The closeness of this value to the observed mo-
lecular weights indicates the probable presence of one carboxyl group for
each molecule. It is not quite certain that this agreement is not accidental.
A product prepared by heating the monomer at 175 showed an equivalent
weight of 6320 and a molecular weight of about 1500 (observed 1420 and
1660), but the molecular weight values are under suspicion because of severe
foaming. It seems certain however that at least part of the polyester
molecules bear carboxyl groups.

The polymeric 5-valerolactone is not homogeneous, but contains mole-
cules of different sizes. Its sodium salts are not soluble in water. A
sample of recrystallized polymer formed at ordinary temperature (m. p.
53-56) showed by titration in acetone an equivalent weight of 3160.
The acetone solution was evaporated and the resulting dry sodium salt
taken up in hot absolute alcohol, filtered and cooled to crystallize. In this
way fractions of different solubilities and sodium content were isolated:
(1) 0.38 g. containing 0.22% Na; (2) 0.77 g. containing 0.49% Na; (3)
0.14 g. containing 3.92% Na. The equivalent weights inferred from the
sodium contents are: (1) 10,400, (2) 4700, (3) 590. Determination of the
molecular weight of fraction 2 by the boiling point in benzene gave the value
3880.

The polyester derived from the lactone (II) of hydroxyethylglycolic
acid differs in its behavior from polymeric 5-valerolactone. It cannot be
titrated to a sharp end-point with 0.1 N alkali. Attempts to prepare
a sodium salt by treating it with sodium bicarbonate under various condi-
tions have led to mixtures from which two distinct fractions are readily
separated. One of these is unchanged polyester entirely free of sodium;
the other is the pure sodium salt of hydroxyethylglycolic acid. Thus
even so mild an alkali as sodium bicarbonate hydrolyzes the polyester, and
it is impossible to determine whether the polymeric molecules bear car-
boxyl groups. Ethylene oxalate (VI) shows a still more exaggerated sen-
sitivity to alkalies. Trimethylene carbonate (V) appears to be neutral.

150 POLYMERIZATION AND RING FORMATION

In this case any terminal carboxyl groups that might arise during the poly-
merization would be lost spontaneously since they would be linked to
oxygen (R O CO OH). The momentary liberation of a small amount
of gas at the end of the sudden thermal polymerization of trimethylene
carbonate is perhaps due to loss of carbon dioxide from acid carbonic ester
groups formed in this way.

For 5-valerolactone, however, as already indicated, the presence of a
terminal carboxyl group in the polymeric molecule is fairly clearly demon-
strated.

It seems likely that the first step in the polymerization of the cyclic
esters involves the intervention of a trace of the corresponding hydroxy acid
or one of its derivatives. One can imagine that hydroxy valeric acid would
thus react with valerolactone

HO(CH 2 ) 4 COOH + O(CH 2 ) 4 CO > HO(CH 2 ) 4 CO O(CH 2 ) 4 COOH

The dimeric acid would then react in the same manner with valerolactone
to form a trimeric acid, and the reaction would continue in this sense until
all of the lactone was exhausted or the chains became too long for further
reaction. In this mechanism a foreign acid might equally well participate,
and it would be bound in the product as the first unit of the polymeric
chain. Experiments in this direction were made with chloroacetic acid.

One gram of 5-valerolactone was heated with 0.1 g. of chloroacetic acid for fifteen
hours at 150-160. The reaction mixture, which solidified on standing in a refrigerator,
was recrystallized from alcohol and carefully dried. It melted at 42 and contained
3.28% chlorine. It was triturated with an excess of sodium bicarbonate in the presence
of a little water, and the product after drying was extracted with warm alcohol. The
polyester which separated was thrice recrystallized from alcohol. It now melted at
45-46 and contained 2.3% sodium. The molecular weights calculated respectively
from the chlorine and sodium content, assuming a molecule having the formula XII,
are 1008 and 1000.

(XII) C1CH 2 CO [O(CH 2 ) 4 CO ]*OH

In a similar experiment in which one gram of valerolactone was heated with 0.05 g.
of chloroacetic acid the initial product melted at 48 and contained 2.06% chlorine.
The sodium salt melted at 50-51 and contained 1.35% sodium. The molecular
weights calculated from these data are 1720 and 1705.

Chlorine-containing derivatives were also obtained by heating the polymeric
lactone with chloroacetic acid. Two grams of polymer and 0.1 g. of chloroacetic acid
were heated for ten hours at 150-160 . The product was thrice crystallized from alcohol.
It contained 1.75% chlorine, and this corresponds to a molecular weight of 2030. Its
neutral equivalent as indicated by titration with alkali was 2083.

These results are at least consistent with the mechanism suggested above. A
mechanism involving the mutual coalescence of two molecules of monomer to form a

X. SIX-MEMBERED CYCLIC ESTERS 151

twelve-atom ring and the growth of this through the progressive absorption of more
monomeric molecules seems less likely, but some observations lately made in this
laboratory on the polymerization of cyclic anhydrides indicate that this mechanism
is not at all impossible.

Lactone (II) of Hydroxyethylglycolic Acid. Preparation: The method of Hollo (2)
was slightly modified. The crude mixture containing the sodium salt of the hydroxy
acid, sodium chloride and a little glycol was washed with acetone, suspended in abso-
lute alcohol and treated with less than the calculated amount of hydrochloric acid.
The precipitated sodium chloride was removed by filtration and the lactone isolated
from the filtrate by distillation. In this way it was readily obtained free of chlorine.

When allowed to stand it gradually solidified to a pasty mass. After five weeks
the solid was crystallized five times from ethyl acetate. It then melted at 62-64 and
showed a molecular weight in boiling benzene of about 460 (observed, 451, 471). This
probably corresponds approximately with Hollo's polymer melting at 56-63.

A polymer formed by heating the lactone at 150 for five hours after being crystal-
lized four times from ethyl acetate melted at 87-89. It was very soluble in cold
chloroform, hot ethyl acetate and hot alcohol, slightly soluble in ether, cold ethyl
acetate and water. Molecular weight determinations in boiling benzene gave the values
1647 and 1788.

Hollo describes another type of polymer obtained when the sodium salt of the
hydroxy acid is treated with an excess of hydrochloric acid. This melts at 66-68,
boils at 216 to 220, and distils without change at atmospheric pressure.

It seems quite impossible however that a polymer should boil only 5 higher than
the monomer (210-215), and the explanation of Hollo's observation probably lies in
the fact that the hydrochloric acid present in the polymer distils with the monomer
during depolymerization and the polymerization of the distillate is so powerfully cata-
lyzed that it occurs very rapidly. In our experiments a specimen of polymeric lactone
originating from a preparation in which a slight excess of hydrochloric acid was used to
liberate the hydroxy acid was observed to distil at 215-216 and the distillate solidified
immediately on cooling. The solid product after three crystallizations from ethyl
acetate melted at 85 to 87.

This polymer distilled completely, but the distillate was not unchanged polymer.
Molecular weight determinations on the distillate made immediately after the com-
pletion of the distillation gave the values, 144, 140 and 121 (calcd. for monomer, 102).
After two hours the distillate had already begun to crystallize. After twenty-four
hours it had solidified, and the solid after recrystallization thrice from ethyl acetate
showed a molecular weight of about 900 in boiling benzene (observed, 867, 931).

Polymers originating from preparations in which an excess of hydrochloric acid
was carefully avoided behaved in precisely the same way except that the reversion of
the distillate was much slower; it remained fluid for twenty-four hours. The melting
points of the purified polymers obtained from either source approached 89.

tt-w-Propyl-5-valerolactone was prepared through the steps indicated in the follow-
ing equations.

C 6 HBOCH 2 CH 2 CH 2 Br +

C 6 H 6 OCH 2 CH 2 CH 2 (C 3 H 7 )C(COOC 2 H 5 ) 2 --HBr > BrCH 2 CH 2 CH 2 CH(C 3 H 7 )COOH

> BrCH 2 CH 2 CH 2 CH(C 3 H 7 )COONa - > OCH 2 CH 2 CH 2 CH(C 3 H 7 )CO

152 POLYMERIZATION AND RING FORMATION

phenoxypropyl-(w-propyl)-malonic diethyl ester: colorless liquid, b. p. (4 mm.)
195-200; 4 1.0246; n 2 D 5 1.4820.

Anal. Calcd. forCi 9 H 2 8O 6 : C, 67.85; H, 8.33. Found: C, 68.60, 68.04; H, 8.03,
8.09.

5-Bromo-a-tt-propylvaleric acid: b. p. (5 mm.) 148-150; d 1.3851, n 1.4730.

Anal. Calcd. for C 8 Hi 6 O 2 Br: C, 43.05; H, 6.72; neutral equivalent, 223. Found:
C, 44.33; H, 6.84; neutral equivalent, 220, 222.

The high values for carbon are probably due to the presence of some phenol.

a-n-Propyl-S-valerolactone: b. p. (10mm.) 118-120; </J 0.9929; w 2 D 1.4585.

Anal. Calcd. for C 8 H 14 O 2 : C, 67.67; H, 9.85; saponification equivalent, 142.
Found: C, 67.59, 67.50; H, 10.01, 10.18; saponification equivalent, 143, 144.

Preparation of Lac tide. The preparation of lac tide by methods described in the
older literature (22) gave poor yields, but excellent results were obtained by the follow-
ing method based on recent patents (23). Two hundred grams of commercial lactic acid
was heated at ordinary pressure in a Claisen flask by a metal bath at 120 until water
ceased to distil. The temperature was then raised to 140 and the pressure reduced
to 10 mm. During six hours water continued to distil slowly from the flask. The
pressure in the system was reduced to 5 mm. and the temperature raised sufficiently
to cause rapid distillation of the lactide, which solidified in the receiver. The yield
of crude product was 125 g. It was purified by crystallization from ether; m. p. 128.

Summary

Several isolated examples of the reversible polymerization of six-mem-
bered cyclic esters have been briefly reported in the literature. Further
examples are reported in the present paper, together with some experimen-
tal data and speculations on the mechanism of the phenomenon. The
following conclusions are reached.

1. The ability to undergo reversible polymerization is generally char-
acteristic of six-membered cyclic esters.

2. Ester rings of five atoms or more than six atoms do not polymerize
under the action of heat.

3. The tendency of six-membered cyclic esters to polymerize is closely
related to their great susceptibility toward hydrolysis; both tendencies are
diminished by the presence of substituent groups.

4. The polymers formed from six-membered cyclic esters are linear
polyesters and, at least in certain instances, the chains are open and ter-
minated by hydroxyl and carboxyl groups.

5. Both the polymerization and the depolymerization consist essen-
tially in a process of ester interchange.

6. The peculiar position occupied by the six-membered cyclic esters is
readily explained by stereochemical considerations based on the Sachse-
Mohr theory.

X. SIX-MEMBERED CYCLIC ESTERS 153

Bibliography and Remarks

(1) Fichter and Beisswenger, Ber. t 36, 1200 (1903).

(2) Hollo, ibid., 61, 895 (1928).

(3) Bischoff and Walden, ibid., 26, 262 (1893).

(4) Drew and Haworth, J. Chem. Soc. t 775 (1927).

(5) Cf. Haworth, "The Structure of Sugars," Edward Arnold and Co., London, 1929, p. 78.

(6) Carothers and van Natta, Journ. Am. Ghent. Soc., 52, 314 (1930).

(7) Bischoff and Walden, Ber., 27, 2939 (1894).

(8) Bischoff, Ber., 40, 2803 (1907).

(9) Carothers, Arvin and Dorough, Journ. Am. Chem. Soc., 52, 3292 (1930); cf. Bergmann
and Wolff, J. prakt. Chcm., 128, 229 (1930); and Carothers and van Natta, Ber., 64, 1755 (1931).

(10) Bischoff and Walden, Ann., 279, 71 (1894).

(11) 0-Lactones have not been specifically examined from this standpoint, but it appears likely
in view of some observations recorded by Johansson [Lunds Universitets Arsskrift II, [2] 12, 3 (1916))
that simple lactones of this class undergo irreversible polymerization under the action of heat.

(12) Carothers and Dorough, Journ. Am. Chem. Soc., 52, 711 (1930).

(13) We do not intend to suggest that the polymerization of large cyclic esters is impossible. In
fact, since such esters can be hydrolyzed to the corresponding hydroxy acids, it is evident from equation
C that such a polymerization can probably be effected under some conditions. The process would
consist in hydrolysis of the cyclic ester to the hydroxy acid and the dehydration of this to the poly-
ester. Both of these steps might occur in a single operation under appropriate conditions, e. g., at
elevated temperature in the presence of a small amount of water plus mineral acid. But the experi-
ments described above do demonstrate that the large cyclic esters, at least qualitatively, are very
different from the six-membered esters.

(14) Dietzel and Krug, Ber., 58, 1307 (1925).

(15) Marvel and Birkheimer, Journ. Am. Chem. Soc., 51, 260 (1929); Blaise and Koehler, Compt.
rend., 148, 1773 (1909).

(16) No exception to this statement has been observed among the numerous esters we have pre-
pared from dibasic acids and glycols, but we have lately found that in the self-esterification of w hydroxy-
decauoic acid in the absence of solvent, appreciable amounts of the cyclic dimeric ester are formed.
This has already been prepared by a different method and described by Lycan and Adams [Journ.
Am. Chem. Soc., 51, 3450 (1929)].

(17) Abderhalden, Naturwissenschaflen, 12, 716 (1924); Bergmann, ibid , 13, 1045 (1925), Pring
sheim, ibid., 13, 1084 (1925); Hess, Ann., 435, 1 (1924).

(18) Hill, Journ. A. Chem. Soc., 52, 4110 (1930).

(19) But the values for carbon are usually somewhat low and the values for hydrogen somewhat
high. This may signify that the chains are open, or it may mean merely that the polyesters in spite
of the fact that they are insoluble in water are nevertheless somewhat hygroscopic and difficult to dry.

(20) Lycan and Adams, Journ. Am. Chem. Soc , 51, 625, 3450 (1929).

(21) Johansson and Sibelius, Ber., 52, 745 (1919).

(22) Gay-Lussac and Pelouze, Ann., 7, 43 (1833); Bischoff and Walden, Ber., 26, 263 (1893)

(23) Chemische Werke, French Patent 456, 824 (1913); Gruter and Pohl, U. S Patent 1,095,205
(1914).

154- POLYMERIZATION AND RING FORMATION

XI. The Use of Molecular Evaporation as a Means for
Propagating Chemical Reactions*

A new design of molecular still is described, the aim being to remove a volatile
reaction product of low vapor pressure from another product, which is
formed simultaneously and having no volatility. The reaction mixture is
heated in a flat pan, a high vacuum being maintained and a cooled con-
densing plate being disposed in the immediate neighborhood of the evaporat-
ing surface.

Reversible reactions involving the simultaneous formation of a volatile
and non-volatile product are often forced to completion by causing the
volatile product to distil from the reaction mixture as fast as it is formed.
The purpose of the present note is to call attention to the possibility of
extending the application of this principle to instances in which the effective
vapor pressure or escaping tendency of a volatile product or potential
product is very small. In the molecular still (1) distillation or continuous
evaporation can be effected even when the vapor pressure of the distilling
substance is as low perhaps as 10~ 5 mm. The theory of the process has
been discussed by Washburn (Ib), and we need only to mention that suc-
cessful operation requires a highly evacuated system comprising a condenser
placed very close to the evaporating surface. The temperature of the con-
denser must be low enough to reduce the vapor pressure of the distillate
to a negligible value. Under these conditions the mean free path of the
molecules is less than the distance from the condenser to the evaporating
surface; consequently most of the molecules that manage to escape from
the evaporating surface are caught by the condenser with a negligible
probability of return.

Mercury and apparently cane sugar can be evaporated in the molecular
still at room temperature (Ib) and w-heptacontane can be distilled without
decomposition (2). The tendency to distil diminishes with increasing mo-
lecular weight, and it appears that, so far as practically useful rates are
concerned, the upper limit of distillability for paraffin hydrocarbons may
lie at as low a molecular weight as 1100 or 1200. For other types of or-
ganic compounds it will in general lie at still lower molecular weights (2).
Thus all substances of high molecular weight are practically completely
non- volatile. There are many instances of reversible reactions in which

* W. H. Carothers and J. W. Hill; Journ. Am. Ghent. Soc. t 54, 1557-59 (1932);
Contribution No. 74 from the Experimental Station of E. I. du Pont de Nemours and
Co.

Received November 12, 1931. Published April 6, 1932.

XI. MOLECULAR EVAPORATION FOR PROPAGATING REACTIONS 155

only one volatile substance is involved. If this substance has a very low
vapor pressure, or if it is strongly held by other substances involved in the
reaction, or if its equilibrium concentration is very low, then molecular
evaporation may be capable of producing a result not even remotely ac-
cessible with the aid of the usual distillation equipment and methods.

Examples of the use of molecular evaporation as a means of propagating
chemical reactions are provided in other papers of this series (3). The re-
actions involved are reversible bifunctional
condensations. The starting materials are
linear polymers capable, by self-combination,
of yielding a still longer molecule with the
elimination of some volatile product. The
size of the product molecules depends upon
the completeness of the reaction. With the
aid of the molecular still one can obtain prod-
ucts of much higher molecular weight than
with the ordinary distillation equipment.
The same principles can doubtless be applied
to other kinds of reversible reactions.

The simple stills described by Washburn
(Ib), function satisfactorily but they do not
provide ready access to the residue from the
distillation. For this reason we constructed
the instrument shown in Fig. 1 and used it
in carrying out some of the reactions referred
to above. The new instrument has the ad-
vantage that the flat pan containing the dis-
tillation residue can be lifted out when the in-
strument is open.

Fig.

1. Molecular still (Y 7
natural size).

Description of the Instrument

The outer shell (A) of the apparatus is made up of
two domes from pyrex vacuum distilling apparatus

(Corning Glass Company Catalogue, Item No. 56). The ground surfaces are lubri-
cated with a good vacuum grease. (B) is the condenser provided with water
leads. (C) is the glass support for the heater (D) and the distilling pan (E) (4).
The heater and distilling pan are contained in an outer copper pan as shown. The
heater consists of nichrome wire wound spirally on a sheet of mica which is supported
between sheets of thin asbestos board inserted below the distilling pan. The heater leads
(F) are brought up through the support (C) and out as shown. Connection with the
heater is made with spring clamps which make the heater readily demountable. The
lower part of the apparatus is the regular distilling tube provided with the Pyrex

156 POLYMERIZATION AND RING FORMATION

vacuum distilling apparatus (Catalogue No. 56). The rubber stoppers at the top and
bottom of the apparatus are well covered with a suitable wax such as picein. A
thermocouple (not indicated in the drawing) is led in through the lower wax seal and
the junction placed under the pan E. The slight bulges on the condenser and the pan
support where they enter the apparatus are necessary to obviate the danger of these
parts sucking in when the apparatus is evacuated.

Summary

In reversible reactions involving the simultaneous formation of volatile
and non-volatile products, the use of molecular evaporation makes it pos-
sible to realize chemical effects that cannot be achieved with the aid of the
usual distillation equipment. A new design of molecular still is described.

Bibliography and Remarks

(1) (a) Bronsted and Hevesy, Phil. Mag., 43, 31 (1922); (b) Washburn, Bur. Standards J. Re-
search, 2, 476 (1929); (c) Burch, Proc. Roy. Soc, (London), 123, 271 (1929); (d) Hickman, Chcm.
Ind., 48, 365 (1929).

(2) Cat-others, Hill, Kirby and Jacobson, Journ Am. Chem. Soc., 52, 5279 (1930).

(3) Papers XII, XIII, XIV and XVI.

(4) This pan is conveniently made by cutting a pyrex beaker close to the bottom. The material
of which the pan is composed has a considerable effect on the rate of the bifunctioual condensations
described in Papers XII, XIII and XIV. No polyesterification could be effected in a copper pan.
On the other hand, when superpolyesters and polyanhydrides are prepared in a glass pan the products
adhere very firmly to the glass so that the dish is frequently shattered on cooling owing to the force
of contraction.

XII. Linear Superpolyesters*

In previous papers^ polymerization up to molecular weights of about 5000
has been investigated and compared with the tendency to the formation of
low molecular rings. As a general rule it can be stated that if 5- or 6-
rings can be formed by esterification, the intra-molecular type of reaction
prevails; in the case of 6- and 7 -member ed rings both monomolecular rings
and linear polymers of medium molecular weight are produced.
From now on the main interest is concentrated on higher polymers, which
are called superpolymers if the molecular weight exceeds 10,000. At the
same time products with a molecular weight between 800 and 5000 are
termed a-esters, while products above 5000 are called u-esters.
In this paper a series of u-polyesters was prepared with the use of the mo-

* W. H. Carothers and I. W. Hill; Journ. Am. Chem. Soc., 54, 1559-66 (1932);
Communication No. 75 from the Experimental Station of E. I. du Pont de Nemours and
Co.

Received November 12, 1931. Published April 6, 1932.

I See Contribution I to XI of this series on pages 4 to 156 in this volume.

XII. LINEAR SUPERPOLYESTERS 157

lecular still described in a previous paper*. Linear a-esters are mostly
microcrystalline powders (in some cases opaque, brittle waxes) with melt-
ing points around 80. They are mostly quite soluble in chloroform and
ethyl-acetate and on the average represent a polymerization degree around
10.

u-Superpolyesters are resins with considerable hardness and toughness. If
placed in solvents, they first imbibe the latter and swell; later they give solu-
tions of high viscosity. When heated they become transparent around 80,
but do not yet flow at this temperature. Flow only takes place at consider-
ably higher temperatures. In the solid state they give sharp X-ray powder
diagrams, which closely resemble the corresponding patterns of the a-esters.
A few degrees below the melting point, the rings lose their sharpness and at
the melting point a diffuse halo characteristic of a liquid appears. Su-
perpolyesters in the plastic state can be drawn out into very strong, pliable
and as the X-ray investigation shows highly orientated fibers, f
The following a-esters are described in this paper:

u-trimethylene ester of hexadamethylene dicarbonic acid
^-polyethylene succinate
co-polyester from w-hydroxydecanoic acid
(^-polyester from w-hydroxypentadecanoic acid
u-ethylene sebacate

In previous papers we have described polyesters having molecular
weights ranging from 800 to 5000 derived from dibasic acids of the series
HOOC(CH 2 )*COOH and glycols of the series HO(CH 2 )/)H (2). These
esters are microcrystalline solids that dissolve readily in appropriate sol-
vents. Their solutions are not highly viscous and they show no signs of
inherently colloidal behavior. In the present paper we describe poly-
esters of the same series having much higher molecular weights. They
exhibit colloidal behavior and simulate to a remarkable degree some of
the properties of certain naturally occurring high polymers. For the sake
of brevity and convenience we use the designation a-ester or a-form for
the polyesters having molecular weights ranging from 800 to 5000 and the
designation co-ester or co-form for the new superpolyesters.

Nature of the Polyesterification. The mutual esterification of dibasic
acids and dihydric alcohols is a bifunctional reaction (3). As such it pre-
sents the formal possibility of yielding (1) cyclic monomeric esters, (2) cyclic

* See page 154 to page 156 of this volume.

f Compare the next paper on pages 165 to 168 of this volume.

158 POLYMERIZATION AND RING FORMATION

polyesters, and (3) open chain polyesters. Studies already reported have
proved that the first possibility is ordinarily realized only in those com-
paratively rare instances where the nature of the starting material permits
the formation of a ring of five or six atoms. Otherwise reaction generally
is intermolecular and the product is of the third type. The polyesterifica-
tion thus consists in a series of intermolecular couplings resulting in the
formation of progressively longer chains as indicated by way of illustration
in the following equations.

HOOC R COOH 4- HO R' OH > HO R'O CO R COOH

2HO R'O CO R COOH >

HO R' O CO R CO O R' O CO R COOH
(n + 1)HO R' O CO R COOH >

HO R' O(CO R CO O R' O) n CO R COOH

The reaction is reversible and an exceedingly large number of entities are
involved in the ester equilibrium. The equilibrium can be displaced by
the removal of the liberated water, and the more nearly complete the re-
action the longer are the molecules of the polyester product. The reaction
becomes formally complete when all the molecules of the initial reactants
are combined into a single molecule.

A complicating circumstance arises from the fact that the glycol and the
acid need not react in precisely equivalent ratio : at any particular stage
n molecules of acid may be involved with n + 1 molecules of glycol or,
conversely, n molecules of glycol may be involved with n + 1 molecules of
acid. In the former case the product molecules will bear hydroxyl groups
at both ends ; in the latter they will bear carboxyl groups at both ends. In
either event no further progress is possible by esterification. Under favor-
able conditions, however, coupling may continue nevertheless. Two mole-
cules may unite by a process of ester interchange involving their terminal
ester linkages. This mechanism is realized experimentally (2) when di-
(/3-hydroxyethyl)-succinate (I) is heated in vacuo. Glycol is liberated
and the final product of prolonged heating is the same as one obtained by
similarly heating succinic acid with a very small excess of glycol.

(I) HO(CH 2 ) 2 OCO(CH 2 ) 2 COO(CH 2 ) 2 OH + HO(CH 2 ) 2 OCO(CH 2 ) 2 COO(CH 2 ) 2 OH >
HO(CH 2 ) 2 OCO(CH 2 ) 2 COO(CH 2 ) 2 OCO(CH 2 ) 2 COO(CH 2 ) 2 OH + HO(CH 2 ) 2 OH, etc.

The only factor that could theoretically preclude the possibility of pro-
ducing exceedingly long molecules as the result of bifunctional esterifica-
tions is the loss of terminal groups through ring formation. Since in the
polyesters thus far examined the terminal groups are still present, it ap-
pears that such ring formation does not occur (4).

Another factor that might be expected to have the same effect is the

XII. LINEAR SUPERPOLYESTERS 159

mutilation of terminal groups, e. g., through the loss of carboxyl as carbon
dioxide. But our experience indicates that the loss of terminal groups in
this fashion occurs only when patently inappropriate conditions are
adopted. More important than this is the fact demonstrated by the be-
havior of di-(/Miydroxyethyl)-succinate cited above that the coupling can
progress through a process of ester interchange involving the last ester
linkage. The possibility of further coupling therefore does not absolutely
depend upon the integrity of the terminal groups as such.

As a matter of fact, however, as such reactions have been carried out
in the past an apparent limit is reached at a comparatively low molecular
weight. Thus the glycol and the acid are heated together in a distilling
flask provided with a receiver to collect the liberated water, and heating is
finally continued at low pressure and considerably elevated temperature.
The molecular weight of the polyester product increases during the progress
of the heating and finally reaches a value of 3000 to 5000. This value is
not noticeably increased if the period of heating is greatly prolonged.

We ascribe the failure of the reaction to progress further under these con-
ditions to the following factors. The concentration of reactive (terminal)
groups in esters having molecular weights as high as 3000 is rather low;
moreover, such esters are very viscous even at a temperature of 200.
This implies a restricted mobility of the molecules, and a low rate of diffu-
sion of the volatile products to the surface. Beyond this, macromolecular
materials, perhaps because of their exaggerated molecular cohesions (5),
have a very powerful tendency to retain dissolved or adsorbed liquids or
vapors several days' heating in high vacuum are required to remove
benzene completely from macromolecular polystyrenes or water from res-
inous esters. All these factors affect either the position of the equilibrium
or the rate at which the equilibrium is approached, and one may suppose
that if some more drastic and effective means could be found for the re-
moval of water or other volatile reaction products, it would be possible to
force the reaction further.

Methods of Producing Superpolyesters. The molecular still, first
used by Bronsted and Hevesy (6) for the separation of the isotopes of mer-
cury and later by Washburn (7) for the separation of petroleum hydrocar-
bons, provides a means for continuously displacing physical equilibria
involving very minute vapor pressures, and the application of this to the
chemical equilibrium under consideration led to the desired result.

Some of the experiments described below were made in stills of the type
described by Washburn (7) and others in a modified still described in the
preceding paper (8), where the principles involved are also briefly discussed.

160 POLYMERIZATION AND RING FORMATION

For reasons of a theoretical nature that need not be discussed, our first
experiments were made with an ester which has not been described pre-
viously, namely, the trimethylene ester of hexadecamethylene dicarboxylic
acid. This was obtained from the acid and the glycol by the method used
for other a-polyesters (2a).

It was then placed in the molecular still and heated by a bath at 200
for a total time of twelve days. During the first seven days a small amount
of distillate collected on the condenser. No distillate was observed during
the remaining five days. The viscosity of the molten polyester increased
progressively during the heating, and the final product showed an apparent
molecular weight of about 12,000.

The same treatment applied to ethylene succinate, ethylene sebacate
and to polyesters derived from co-hydroxydecanoic acid and from co-hy-
droxypentadecanoic acid, led to similar transformations.

A part of the effect of the molecular still treatment is evidently due to the
greatly prolonged time of heating, and it was found later that a similar,
if somewhat less pronounced, effect could be produced by merely heating
polyesters such as ethylene sebacate in thin layers at ordinary pressure in a
stream of nitrogen for a long period of time, or by bubbling a stream of
nitrogen through the molten polyester.

Physical Properties of Superpolyesters. Linear a polyesters are
microcrystalline powders (in certain cases resins) that dissolve readily in
certain solvents such as cold chloroform and yield solutions which are
quite mobile. The crystalline esters melt fairly sharply and when molten
they are exceedingly viscous. On cooling they solidify to opaque
masses which range from rather hard, brittle, porcelain-like masses to soft,
waxy solids as the length of the polymethylene chain in the acid or glycol
from which they are derived increases.

The superpolyesters in the massive state are harder and much tougher.
They also dissolve in cold chloroform but the process is slow; they first
imbibe solvent and swell, and their solutions are highly viscous. They
separate from solution as more or less coherent powdery or curdy masses.
When heated they melt at approximately the same temperatures as the
a-polyesters from which they are derived. The phenomena associated
with the melting, however, are quite different from those observed in the
a-polyesters. The solid ester at the melting point suddenly becomes trans-
parent, but it does not lose its shape ; a tendency to flow appears only at a
considerably higher temperature. In the solid state the superpolyesters
furnish sharp x-ray powder diffraction patterns that closely resemble similar
patterns obtained from the corresponding a-polyesters. At a temperature

XII. LINEAR SUPERPOLYESTERS

161

a few degrees below the melting point the patterns lose their sharpness.
At the melting point the pattern is a diffuse halo characteristic of a liquid.
The most remarkable property of the superpolyesters is their capacity to
be drawn out into very strong, pliable, highly oriented fibres, and this is
discussed in detail in a subsequent paper (9).

By way of summary, the properties of the a- and w-polyesters from tri-
methylene glycol and hexadecamethylerie dicarboxylic acid are listed in
Table I. Similar relations exist between the a- and the co-forms of other
polyesters. One curious property of the co-esters deserves special mention.
They adhere very firmly to the glass vessel (in the molecular still) in which
they are prepared, and the force of contraction during cooling often causes
the vessel to be completely shattered if the layer of ester is more than two
or three millimeters thick.

TABLE I
PROPERTIES OF a- AND CO-POLYESTERS FROM TRIMETHYLENE GLYCOL AND HEXADECA-

METHYLENE DlCARBOXYLIC AdD

Apparent mol. wt.

At 100 C.

At room temperature

Melting point, C.
Solubility

a-Ester
3000

Viscous liquid

White, opaque, brittle wax;

d\l 1.061
75-76

Very soluble in cold CHC1 3 .
Readily soluble in hot
ethyl acetate ; separates
on cooling as a micro-
crystalline powder

8.6 units

w-Ester

12,0000

Soft, sticky resin

Cream colored, opaque,
horny, elastic; dgo, 1.058

Becomes transparent at 75,
but does not flow

Swells in cold CHC1 3 , then
dissolves. Swells and
slowly dissolves in hot
ethyl acetate; separates
on cooling as a white,
amorphous, curdy pre-
cipitate

166 units

Relative viscosity of 7.3%
weight solution in chloro-
form (chloroform, 216
seconds = 1 unit)

Structure of the w-Polyesters

The simplest and most reasonable assumption concerning the super-
polyesters is that they are formed by a continuation of the same bifunc-
tional esterification that first results in the formation of the a-polyesters,
and that they are exceedingly long polymeric chains bearing hydroxyl or
carboxyl groups at the ends. Their physical, chemical and analytical be-
havior is completely consistent with this assumption, but no actual data

162 POLYMERIZATION AND RING FORMATION

on the presence of terminal groups are yet available. The average length
of the molecules also is still very uncertain. The molecular weights
quoted in Table I were obtained by ebullioscopy and in view of the fact
that the observed elevations did not greatly exceed the experimental error
they can be accepted as signifying only that the molecular weights are
quite large at least as high as 10,000. Comparison of physical properties
indicates that molecular weights of the co-polyesters probably lie in the
neighborhood of 20,000. Further work on the determination of the mo-
lecular weights of the superpolyesters is in progress.

w-Tfimethylene Ester of Hexadecamethylene Dicarboxylic Acid. This ester was
obtained in its a-form by heating the acid together with a 10% excess of the glycol in a
distilling flask at 180-200 for three hours, and then heating the residue to 200 for
six hours at a pressure less than 1 mm. It was thus obtained in the form of a very vis-
cous liquid which on cooling solidified to a hard, waxy mass. It was readily soluble in
cold chloroform and in hot ethyl acetate, from which it separated on cooling as a white
microcrystalline powder. It melted rather sharply at 75-76. Its apparent molecular
weight was about 3300.

Anal. Calcd. for O (CH 2 ) 3 O CO (CH 2 ) 16 CO = C 21 H 38 O 4 : C,
71.13; H, 10.81; saponification equivalent, 177.2; mol. wt. 354.3. Calcd. for
H(~O (CH 2 ) 3 O CO (CH 2 ) 16 CO ) 10 O(CH 2 ) 3 OH = C 2 , 3 H 388 O 42 : C, 70.63;
H, 10.80; saponification equivalent, 181; mol. wt. 3619. Found: C, 70.34, 70.40; H,
10.60, 10.69; saponification equivalent, 175.2; mol. wt. (in boiling benzene), 3600, 3200.

A sample (8.5 g.) of this recrystallized ester was placed in a molecular still of the
type described by Washburn (10). The condenser was cooled with running tap water,
and a trap between the mercury diffusion pump and the still was cooled with liquid air.
The still was heated by means of a metal bath kept at about 200 . The pressure of the
system as measured by a McLeod gage between the trap and the pump was always
less than 10~ 5 mm. Heating was continued for seven days. During the first two days
of heating, 0.4 g. of distillate collected on the condenser. This was removed and the
heating of the residue was continued for five days more. No distillate collected during
this period.

The analytical composition and the saponification equivalent of the residual ester
were the same as those of the starting material. Its molecular weight, however, was
much greater.

Anal. Calcd. for O (CH 2 ) 3 O CO (CH 2 ) 16 CO = C 2 iH 38 O 4 : C,
71.13; H, 10.81; saponification equivalent, 177.2; mol. wt. 354.3. Calcd. for H (O
(CHj)r O CO (CH 2 )i<rCO) 34 O (CH 2 ) 3 OH = C 7 i7H 13 ooO 2 6 8 : C, 70.89; H, 10.83;
saponification equivalent, 178.2; mol. wt., 12,122. Found: C, 70.38, 70.52; H,
10.56; 10.53; saponification equivalent, 178.4; mol. wt. (in boiling benzene) 12,100.

The distillate was not homogeneous. By crytallizing it from alcohol there was
obtained an apparently homogeneous fraction composed of needles melting sharply at
124-125. It depressed the melting point of hexadecamethylene dicarboxylic acid
(which melts at 122-123 ) and was neutral. A single ultimate analysis indicated that
it was possibly a monomeric form of the trimethylene ester of hexadecamethylene di-
carboxylic acid.

XII. LINEAR SUPERPOLYEStERS 163

Anal. Calcd. for O (CH 2 ) 3 O CO (CH 2 )i8 CO; C, 71.13; H, 10.81.
Found: C, 70.06; H, 10.63.

A sample of the a-ester and heated at 200-250 for thirty-two hours with a current
of dry nitrogen bubbling through the molten mass. The viscosity of the molten polymer
increased very greatly and the product exhibited the phenomenon of cold-drawing
described in paper XV.

co-Polyethylene Succinate. A sample of polyethylene succinate having the ap-
proximate formula HO (CH 2 );r-O(CO (CH 2 ) 2 CO O (CH 2 )2 O) 2l H (11) was
heated in the molecular still as described above for the trimethylene ester of hexadeca-
methylene dicarboxylic acid. From the very small amount of distillate no homogeneous
fraction could be isolated. The residue was a tough, horny, somewhat elastic mass
(the original material was hard, brittle, rather porcelain-like). On being heated to 97
is softened to a very viscous and somewhat rubbery mass. Fibers prepared from this
material were short and rather brittle.

Anal. Calcd. for C 6 H 8 O 4 : C, 50.00; H, 5.50. Found: C, 50.30; H, 5.70.

w-Polyester from co-Hydroxydecanoic Acid. co-Hydroxydecanoic acid was prepared
by the method of Lycan and Adams (12). Five grams was heated under a moderate
vacuum for ten hours to bring about incipient polymerization. If this is neglected
and the acid heated directly in the molecular still, it distils practically completely be-
fore polymerization sets in. The material after the preliminary heating was transferred
to the molecular still and heated at 150 for twenty-four hours. Two and four- tenths
of a gram of distillate (m. p. 60-65) collected. This was not further examined. The
residue was a hard brittle wax which was heated for four days more at 150 and then at
200 for two days. The final residue was a light gray, translucent, tough, flexible mass.
On being heated it suddenly became transparent at 65 but did not lose its shape. At
higher temperatures it could easily be drawn into threads which could be cold drawn.

Microscopic Observation at a Magnification of 180 Diameters. A thin hazy film
was obtained by placing a small specimen of the ester on a slide and covering it with a
cover glass. Between crossed nicols the field appeared as an aggregate of very small,
irregular anisotropic particles, each of which showed two extinction positions. The
slide was then warmed to melt the ester and the cover glass moved about 5 mm. The
middle region of the specimen when cold was now transparent and homogeneous; the sur-
rounding portions were unchanged. This effect is apparently due to the action of stress
in moving the slide. The transparent section showed brilliant interference colors be-
tween crossed nicols and extinction parallel and normal to the direction of the stress ap-
plied.

Anal. Calcd.: C, 70.59; H, 10.59. Found: C, 70.24, 70.21; H, 10.74, 10.79;
mol. wt. in boiling ethylene chloride, 24,600.

co-Ester from co-Hydroxypentadecanoic Acid. One gram of "exaltolide" (the lactone
of co-hyroxypentadecanoic acid) was refluxed with 0.5 g. of sodium hydroxide in 12 cc.
of water and 5 cc. of alcohol for five hours. The solution was evaporated to dryness on
a steam-bath, dissolved in 450 cc. of water, acidified with dilute sulfuric acid and ex-
tracted with ether. The ether solution was dried over magnesium sulfate, filtered and
evaporated to dryness; yield, 1.075 g. The product was recrystallized from benzene;
m. p. 83-84. The acid was put in the molecular still and heated overnight at 150. It
distilled completely except for a very small residue of hard wax, The distillate was

164 POLYMERIZATION AND RING FORMATION

heated for six hours at 150 at 1 mm. and again placed in the still and heated for seven
days. A small amount of distillate which melted at 70-75 was not further examined.

The residue consisted of 0.65 g. of a flexible, translucent mass. It became trans-
parent at 95, yielding a stiff mass which held its shape up to 200. It could be drawn
into thin fibers and cold drawn.

Ethylene Sebacate. 101 g. (0.5 mole) of sebacic acid and 32.5 g. (0.525 mole) of
ethylene glycol were heated together at atmospheric pressure in a Claisen flask at 175.

The aqueous distillate was discarded. The residue was then heated at 250 for
five hours at 2 mm. pressure. The light gray-buff product was a hard wax. Threads
drawn from the molten ester, if pulled immediately while still warm, yielded transparent
fibers. The product after recrystallization from ethyl acetate melted at 75 (cop-
per block) and still exhibited the cold drawing phenomenon; molecular weight by
cryoscopic method in benzene, 4800, 6000.

Summary

By prolonged heating in a molecular still or in a stream of inert gas,
the previously described linear polyesters derived from dibasic acids and
glycols or from higher co-hydroxy acids are caused to react with themselves
to produce polyesters of much higher molecular weight. The new super-
polyesters (co-polyesters) are tough, opaque solids which exhibit sharp
x-ray powder diffraction patterns and become transparent at a definite
temperature. They dissolve in chloroform to form highly viscous solu-
tions. The mechanism of the reaction is discussed.

Bibliography and Remarks

(1) The term "superpolymer" is applied to linear polymers having molecular weights above
10,000.

(2) (a) Carothers and Arvin, Journ. Am. Chem. Soc., 51, 2560 (1929); (b) Carothers and van
Natta, ibid., 52, 314 (1930); (c) Carothers and Dorough, ibid., 52, 711 (1930); (d) Carothers, Arvin
and Dorough, ibid. t 52, 3292 (1930).

(3) Carothers, ibid., 51, 2548 (1929).

(4) Valuable indications in this connection are furnished by Lycan and Adams' studies [Journ.
Am. Chem. Soc., 51, 625, 3450 (1929)] of the polyesters derived from w-hydroxydecanoic acid. We
should add that our generalization to the effect that no intramolecular reaction occurs in bifunctional
esterifications involving compounds whose structural units are longer than seven atoms is not altogether
free from exceptions. Dr. van Natta of this laboratory in preparing polyesters of co-hydroxydecanoic
acid has recently observed the formation of appreciable amounts of the dimeric lactone already ob
tained by Lycan and Adams by a different method.

(5) Dunkel, Z. physik. Chem., Abt. A, 138, 42 (1928).

(6) Bronsted and Hevesy, Phil. Mag., 43, 31 (1922).

(7) Washburn, Bur. Standards J. Research, 2, 476 (1929); see also Burch, Proc. Roy. Soc. (Lon
don), 123, 271 (1929); and Hickman, Chem. Ind., 48, 365 (1929).

(8) Paper XI.

(9) Paper XV.

(10) Washburn, Bur. Standards J. Research, 2 480 (1929).

(11) Carothers and Dorough, Journ. Am. Chem. Soc., 52, 711 (1930).

(12) Lycan and Adams, ibid., 51, 625 (1929).

XIII. POLYAMIDES AND MIXED POLYESTER-POLYAMIDES 165

XIII. Polyamides and Mixed Polyester-Polyamides*

Above its melting point t-aminocaproic acid\ splits off water and gives

a) a cyclic lactam, which is a seven-membered ring and

b) a linear polyamide with a molecular weight around 3000,

Linear super poly amides are harder, tougher, less easily fusible and much
less soluble than the corresponding super poly esters owing to the higher
molecular cohesion attributed to the CONH group.
In the molecular still the polyamide obtained from e-amino caproic acid
polymerizes further and exhibits a considerable change in physical
properties.

By heating together e-aminocaproic acid with hexadecamethylene di-
carboxylic acid and trimethylene glycol mixed poly ester -poly amides are
obtained. They lie in their physical properties between the pure poly-
esters and pure polyamides and can be drawn out into strong and fine
transparent fibers .

Above its melting point -aminocaproic acid undergoes dehydration
with the simultaneous formation of two different products (1). One of
these is the cyclic lactam, a seven-membered ring. The other is a poly-
amide having the formula

. . . NH (CH 2 ) 6 CO CH (CH 2 )5 CO NH (CH 2 ) 6 CO . . .

This compound is analogous to the polyesters obtained by the self-esterifica-
tion of higher hydroxy acids or by the action of glycols on dibasic acids.
The binding atom between the units in the case of the esters is O and
in the case of the polyamide NH . In comparison with the esters the
polyamide is harder, tougher, less easily fusible and much less soluble.
The difference lies in the direction required by the much higher molecular
cohesions of amides as compared with esters (2).

The polyamide prepared by dehydration of the amino acid at atmospheric
pressure or under diminished pressure in the usual distillation equipment
has an apparent molecular weight of about 1000 (the actual value probably
lies somewhat higher than this, perhaps in the neighborhood of 3000).
It corresponds, therefore, in molecular weight with the oi-polyesters that
have been described in previous papers (3) . The polyesterification reaction

* W. H. Carothers and J. W, Hill; Journ. Am. Chem. Soc., 54, 1566-69 (1932);
Communication No. 76 from the Experimental Station of E. I. du Pont de Nemours and
Co.

Received November 12, 1931. Published April 6, 1932.

t Compare paper VIII of this series on pages 78 to 81 in this volume.

166 POLYMERIZATION AND RING FORMATION

can be forced further toward completion by the use of molecular distilla-
tion, and this process leads to the formation of superpolyesters described
in the preceding paper (4) . One of the objects of the experiments described
in the present paper was to examine the possibility of forcing the amide re-
action in a similar manner so as to obtain a polyamide of very high molecu-
lar weight.

The polyamide already described was placed in the molecular still and
heated for forty-eight hours at 200. A very small amount of the crystal-
line cyclic lactam distilled from the reaction mixture and the residue was
considerably changed in its properties. It was harder and tougher than
before and in thin sections was flexible and elastic. It softened at 210
with considerable decomposition. Like the initial amide it was insoluble in
common organic solvents with the exception of hot phenol and hot form-
amide.

Anal. Calcd. for C 6 H U ON: C, 63.71; H, 9.73. Found: C, 63.94, 63.94; H, 9.72,
9.76.

The change in properties indicates a considerable increase in molecular
weight but no actual measurements of molecular weight are available
owing to the lack of any method of sufficient reliability.

The superpolyesters described in the preceding paper are especially
interesting because of their ability to furnish strong, pliable, highly oriented
fibers. This property is discussed in detail in paper XV. As synthetic
silk, however, these materials suffer from the defect of low melting point and
considerable solubility in various organic solvents. On the other hand,
the polyamide described above is too infusible and insoluble to allow a
ready test of its ability to furnish fibers. In the hope of obtaining a com-
promise between the properties of the polyesters and the polyamides we
prepared mixed polyester-polyamides. These compounds were obtained
by heating together trimethylene glycol, hexadecamethylene dicarboxylic
acid and e~aminocaproic acid. The glycol was used in a 5% excess over
the amount equivalent to the dibasic acid and the amino acid was varied
in different experiments in the proportion of 1, 2, 3 and 5 moles per mole
of dibasic acid. The mixtures were separately heated in a Claisen flask
for three hours in a current of dry nitrogen by means of a bath kept at
200-220. Heating was continued for five hours more at a pressure of
1 mm. with the b^th at 250-260. The very viscous residue was then
removed from the flask and transferred to the molecular still, where it
was heated for three days at 200. So far as could be inferred from the
physical appearance of these products the molecular still treatment pro-

XIII. POLYAMIDES AND MIXED POLYESTER-POLY AM I DEvS 167

duced very little effect although in each experiment a small amount of the
cyclic lactam from the amino acid was obtained as a distillate.

The products were opaque or translucent solids, hard, very tough and
in thin sections flexible and elastic. As the proportion of amino acid in-
creased, the polymers increased in brittleness, hardness, transparency
and melting point. The effect of the composition on melting point is indi-
cated below

1 mole of amino acid ca. 73 3 moles of amino acid 125

2 moles of amino acid 100 5 moles of amino acid 145

The melting points were determined on a copper block. In the two
polymers of lowest amino acid content the temperatures at which the
specimens became transparent were taken as the melting points. In the
other two cases where the high degree of transparency when cold made this
determination extremely uncertain the temperatures at which the sample
first began to adhere slightly to the block were taken as the melting points.

The solubility of the mixed polyester-polyamides also diminished with
increasing amino acid content. The first member of the series was swelled
by hot ethyl acetate and finally completely dissolved. It separated from
the cold solution in the form of soft powder. The second was very slightly
soluble in hot ethyl acetate, and the higher members were practically in-
soluble. Since the polyamide derived from e-aminocaproic acid is not
dissolved by hot ethyl acetate, the mixed polyester-polyamides are evi-
dently not merely physical mixtures. From the nature of the reaction
also it seems likely that all the components participate in forming each
molecule of the product. The different units that may be present in the
product molecules are

NH R CO O R' O CO R* CO

which may be joined in any order that leads to ester or amide linkages.

The mixed polyester-polyamides when molten can readily be drawn out
into filaments by touching a specimen with a rod and drawing the rod away.
The successful production of continuous filaments by this method requires
a degree of plasticity in the molten mass that appears only at a temperature
somewhat above the melting point. Filaments obtained in this way are
brittle and opaque. When slightly warmed (e. g., to 40 to 50) they can
be drawn out by the action of stress into transparent fibers which are ex-
ceedingly pliable. Purely qualitative observations indicate that these
fibers are considerably stronger than similar fibers derived from the co-
polyesters.

168 POLYMERIZATION AND RING FORMATION

Summary

The polyamide derived from -aminocaproic acid when heated in the
molecular still undergoes a considerable change in its physical properties,
indicating an increase in molecular weight. By heating together e-amino-
caproic acid with hexadecamethylene dicarboxylic acid and trimethylene
glycol, mixed polyester-polyamides are obtained. These materials in
their physical properties lie between the polyesters and the polyamides
and like the superpolyesters described in the preceding paper they can be
drawn out into strong, pliable, transparent fibers.

Bibliography and Remarks

(1) Carothers and Berchet, Journ. Am. Chem. Soc , 52, 5289 (1930); Gabriel and Maass, Ber., 32,
1266 (1899); Braun, ibid , 40, 1835 (1907)

(2) Dunkel, Z. physik. Chem., Abl. A, 138, 42 (1928).

(3) Carothers and Arvin, Journ. Am. Chem. Soc., 51, 2560 (1929), Carothers and van Natta, tbtd ,
62, 314 (1930).

(4) Paper XII.

XIV. A Linear Superpolyanhydride and a Cyclic Dimeric
Anhydride from Sebacic Acid*

The different products, obtained during the anhydrization of sebacic acid
are described in this paper.

a-sebacic anhydride is formed by the action of acetic anhydride on sebacic
acid and has a melting point of 79-80. In the molten state it is very viscous,
on cooling it crystallizes in needles; it has a molecular weight around 5200.
p-sebacic anhydride, formed by treating the a-anhydride in the molecular
still; is a definitely macrocrystalline solid melting sharply at 68. It dissolves
readily in most of the common solvents and was identified as a twenty-two-
membered cyclic dimer.

y-sebacic anhydride is obtained by heating the ^-product to its melting
Point, cooling down and heating again. Its melting point is 82, its molecular
weight lies in the neighborhood of 600, which shows that it is formed from
the ^-product by a process of polymerization.

u-sebacic anhydride is a super polymer, with a rather high molecular
weight;^ it is formed by heating a-sebacic anhydride in the molecular still.
It is a hard and tough resin, which becomes transparent at 83 but remains

* J. W. Hill and W. H. Carothers; Journ. Am. Chem. Soc., 54, 1569-79 (1932);
Communication No. 77 from Experimental Station of E. I. du Pont de Nemours and
Co.

t No definite figure can be given owing to the unsolubility of this material.

XIV. ANHYDRIZATION OF SEBACIC ACID 169

still hard. At about 130 it gets soft and can be drawn into continuous,
strong and pliable filaments.
The reactions of the different anhydrides with aniline are described.

In a previous paper (1) it has been shown that adipic anhydride as or-
dinarily prepared is a linear polymer of the type formula represented by I
and that under the action of heat in vacuo it is broken down to the cyclic
monomer, a seven-membered ring.
(I) ... O CO R CO O CO R CO O CO R CO. . .

Observations have now been extended to sebacic anhydride, and it is
shown that the anhydride (a-anhydride) prepared by the action of acetic
anhydride or acetyl chloride on sebacic acid is also polymeric. No smooth
depolymerization of this polymer can be effected under ordinary conditions,
but in the molecular still (2) at elevated temperatures two processes occur
simultaneously. The a-anhydride is transformed into a polyanhydride of
much higher molecular weight (w-anhydride) and at the same time de-
polymerization occurs with the formation of a crystalline product ({$-
anhydride), which is shown to be not the eleven -membered cyclic monomer,
but the twenty-two -membered cyclic dimer. At its melting point the
dimer reverts to a higher polymer (7-anhydride). The anhydrides of
different origins are arbitrarily assigned the prefixes a, 0, 7 and co to desig-
nate them for the purposes of discussion.

Sebacic a- Anhydride. Voerman (3) obtained this as a microcrystalline
solid melting at 74.5 by the action of acetyl chloride on sebacic acid. He
seemed unable to decide whether to regard it as monomeric or polymeric.
Compared with succinic and glutaric anhydrides (known to be monomeric)
it showed a diminished solubility and ability to crystallize, and molecular
weight determinations in boiling benzene and boiling acetone gave ab-
normally high values. The doubt concerning its polymeric character
arose from the fact that freezing point determinations in phenol gave values
agreeing with those calculated for the monomer, but Voerman overlooked
the fact that sebacic anhydride reacts very rapidly with phenol to give
phenyl esters of sebacic acid.

We prepared the a-anhydride by heating sebacic acid with excess acetic
anhydride, distilling off the volatile material, and precipitating a benzene
solution of the residue with petroleum ether. The melting points varied
somewhat from one preparation to another and a typical specimen melted
at 79-80. The molten anhydride is exceedingly viscous; on cooling it
crystallizes: minute doubly refracting spherulites first separate and grow
to what appear to be star-like clusters of needles. This behavior is highly

170 POLYMERIZATION AND RING FORMATION

characteristic of crystalline linear high polymers and is observed with such
diverse materials as polyesters (4), trimethylcellulose (5) and triacetylinu-
lin (6).

At a temperature of 200 and under a pressure of 0.1 mm. the a-anhy-
dride shows no tendency to distil, and this behavior is also consistent with
a highly polymeric structure.

We tentatively assign to sebacic a-anhydride the formula II.

(II) CH 3 CO [ O CO (CH 2 ) 8 CO ]* O-CO CH 3

This represents it as a linear polymer and the chains are terminated by
acetyl groups derived from the reagent used in bringing about the anhydride
formation. It reacts with water to form sebacic acid and acetic acid. The
latter may conceivably represent adsorbed acetic acid or acetic anhydride
not removed in the purification process, but we are inclined to the view that
it arises from acetyl groups that actually form a part of the polyanhydride
molecule. This assumption together with the estimation by distillation
and titration of the amount of acetic acid formed on hydrolysis furnished a
value of 5260 for the molecular weight of a typical specimen of the polymer.

A specimen of the recrystallized a-anhydride which had been stored for three days
in a vacuum desiccator over phosphorus pentoxide, potassium hydroxide and paraffin
was refluxed with 3 g. of sodium hydroxide in 40 cc. of water, diluted, acidified with
sulfuric acid and distilled with the occasional addition of water. A total of 415 cc. of
distillate was collected and titrated with 0.1 N sodium hydroxide. Sodium acetate
was identified in the resulting neutralized solution by microscopic observation of the
highly characteristic crystals of sodium uranyl acetate formed upon adding uranyl
nitrate. We are indebted to Mr. W. D. M. Bryant for this identification. Since it was
found that sebacic acid itself tends to steam distil very slowly, a correction was made by
repeating the above determination using 6 g. of pure sebacic acid.

Anal. Subs. 5.51 g.: 0.0999 N NaOH, 26.35 cc. Blank: 0.0999 N NaOH,
5.35 cc. Found: (assuming two acetyl groups per molecule) mol. wt., 5260.

A different though not strictly independent estimate of molecular weight
is found in the quantitative analysis of the behavior of the polymer toward
aniline, which is described in the next section, and this method gives a
value of 5500. Attempted molecular weight determinations in boiling
benzene gave much lower values than this (as low as 650), but their lack of
constancy and reproducibility deprives them of any significance (7). On
the other hand, the a-anhydride in its physical properties closely resembles
polyesters derived from sebacic acid and known to have molecular weights
in the neighborhood of 5000.

The analytical data presented below are consistent with the suggested
structure for the a-anhydride.

XIV. ANHYDRIZATION OF SEBACIC ACID 171

c Calcd.,% H c Found, % ^

On 65.18 8.75 64.19 9.07

CH 8 COO(CioHi6O 3 )3COCH3 62.35 8.31 64.22 9.01

CHgCOCKCioHuCWaCOCHs 64 . 00 8 . 57

CH 3 COO(CioHi6O 3 )3oCOCH3 64.84 8.71

Behavior of Sebacic Anhydrides toward Aniline. Theory requires a
difference, both qualitative and quantitative, between cyclic monomeric
anhydrides and linear polymeric anhydrides in their behavior toward un-
symmetrical reagents such as aniline. This requirement was first pointed
out by one of us and experimentally verified in the case of adipic anhy-
drides (1, 8). The analysis has now been extended and further observa-
tions made with sebacic anhydrides.

The monomeric anhydride (II) can react with aniline to form only a
single product, namely, the monoanilide of sebacic acid.

(II) CO (CH 2 ) 8 CO O j + CH 6 NH 2 > C.H 6 NH CO (CH 2 ) 8 COOH

In the polymeric form, however, as soon as one anhydride linkage reacts
with aniline, reaction at the next linkage will give different products de-
pending upon which side of the oxygen is involved with the reagent. In
III we symbolize aniline (RNH H) as b-a.

(III) . . CO (CH 2 ) 8 CO O CO (CH 2 ) 8 CO O CO (CH 2 ) 8 CO O . .

b-a.

b-a >

a-b

b-a.
a-b >

a-b

Considering the two linkages of an interior unit of the chain, if reaction
occurs in the direction b a at the first one it may occur in the directions
b a, and a b at the second; and if the direction is a b at the first it
may be either b a or a b at the second. If we assume that reaction is
random, i. e., that the direction of reaction at the second linkage is not
affected by the direction that the reaction has already taken at the first
linkage, then the probability of each of these four possibilities is the same,
and the terminal groups attached to the reaction product derived from
the unit under consideration are (1) a b, (2) a a, (3) b b and (4) b a. The
first and fourth of these represent monoanilide, and the second and third
represent, respectively, dibasic acid and dianilide. Thus each interior unit
of the anhydride chain will yield

172 POLYMERIZATION AND RING FORMATION

l /4 molecule of dibasic acid
1 /4 molecule of dianilide, and
*/2 molecule of monoanilide

and this total product will contain one atom of acid hydrogen.

If the poly anhydride is cyclic, all the units are interior units, and the
ratio of products indicated above will be maintained so long as the an-
hydride is polymeric (i. e., not monomeric). But if the chains are open
this ratio will be modified.

We consider the case IV in which the chains are open and terminated by
carboxyl groups.

(IV) HOOC (CH 2 ) 8 CO O [ CO (CH,) 8 ~-CO O L_ 2 CO (CH 2 ) S

COOH

The total number of units in the chain is x. Of these, the x-2 interior
units will furnish acid, monoanilide and dianilide in the ratios already
indicated, while the terminal units will each furnish one-half molecule of
acid and one-half molecule of monoanilide. Thus the total products are

(x 2)/4 -f 1 molecule of dibasic acid

(x 2)/2 4- 1 molecule of monoanilide, and

(.v 2)/4 molecules of dianilide

and the number of atoms of acidic hydrogen in the product is

[(* - 2)/4 4- 1]2 + (x - 2)/2 4- 1 = * 4- 1

If M is the gram molecular weight of the polymer, then, since the molecu-
lar weight of the structural unit is 184

184* 4- 18 = M (1)

and this weight after reacting with aniline will furnish x 4~ 1 mole of hy-
drogen ion. Thus the weight of anhydride required to furnish one mole
of hydrogen ion is

M/(x 4- 1) (2)

Let g represent the weight of a sample of polyanhydride of the formula
IV which is treated with aniline and cc represent the volume in cubic
centimeters of normal alkali required to neutralize the product. Then
from (2)

M/(x 4- 1) = 1000 g(/cc) (3)

and from (1)

x = (M - 18)/184 (4)

XIV. ANHYDRIZATION OF SEBACIC ACID 173

Substituting and solving for M, we have

M * 166,000 g/(184 cc - 1000 g) (5)

If the terminal groups are acetyl as in V

(V) CH 8 CO O [CO (CH 2 )8 CO O ] x COCH,

M = 184 + 102, and this weight after reacting with aniline will furnish

#/4 moles of dibasic acid
#/4 moles of dianilide
x/2 moles of monoanilidc
1 mole of acetic acid, and
1 mole of acetamide

The moles of available hydrogen ion in this product are

2*/4 + x/2 + 1 = x + I
In this case equation (5) becomes

M = 82,000 g/(184 cc - 1000 g) (6)

The manner in which the calculated available acid from the aniline reac-
tion product varies with the molecular weights of different sebacic an-
hydrides is indicated in Table I.

From the data of Table I it is evident that even if one allows an error
of 2% in the titration it should be possible to distinguish an open chain
terminated by carboxyl and having a molecular weight of 7500 or less from a
cyclic anhydride. If the terminal groups of the chain are acetyl, a similar
distinction is possible only if the molecular weight lies below about 3500.

TABLE I
ACID FURNISHED BY ANILINE REACTION PRODUCT OF SEBACIC ANHYDRIDES

Millimoles of available H + in reaction

Mol. wt. product from one gram of anhydride

Structural (for cyclic Cyclic Chain terminated Chain terminated

tinits anhydride) anhydride by carboxyl by acetyl

1 184 5.44 9.9 6.95

2 368 5.44 7.75 6.39

3 552 5.44 7.02 6.12

4 736 5.44 6.62 5.97
6 1,104 5.44 6.24 5.82

10 1,840 5.44 5.92 5.66

15 2,760 5.44 5.76 5.59

20 3,680 5.44 5.68 5.55

40 7,360 5.44 5.56 5.49

60 11,040 5.44 5.52 5.48

100 18,400 5.44 5.49 5.46

200 36,800 5.44 5.46 5.45

174 POLYMERIZATION AND RING FORMATION

The data from which the molecular weight of the a-polyanhydride
described in the preceding section were calculated are as follows.

Weighed samples of the reprecipitated anhydride, which had been stored for several
days in an evacuated desiccator over solid potassium hydroxide, were added to 5 to 8 cc.
of aniline and triturated with a stirring rod until a smooth cream was formed. The
mixture was then diluted with about 30 cc. of 95% alcohol and titrated with 0.1 TV sodium
hydroxide using phenolphthalein as the indicator.

Anal. Subs. 0.5869, 0.4814, 0.5632: 0.0999 N NaOH, 32.40, 26.55, 31.12 cc.
Calcd. by equation (6): M = 5532, 5891, 5248.

Sebacic co- Anhydride. Attempts to depolymerize sebacic a-anhydride
by heating it under greatly diminished pressure in a Claisen flask resulted
merely in thermal decomposition with the formation of tarry products.
But in the molecular still (9), at 200 change occurred simultaneously in
two directions: a crystalline distillate gradually accumulated on the con-
denser, and the residue became more viscous. The rate of distillation was
about one gram in twenty-four hours (from 8 g. of a-anhydride, and an
evaporating surface of about 40 sq. cm.). The final residue when cold was
very much harder and tougher than the a- anhydride. It was opaque and
when heated to about 83 it suddenly became transparent but remained
very stiff. At 130 it was soft enough to be drawn into continuous fila-
ments. These filaments exhibited the phenomenon of cold drawing
described in a subsequent paper (10). The resulting fibers were lustrous
and exceedingly strong and pliable, but after standing for a few days in a
vacuum desiccator they became brittle and fragile. Massive specimens
of the co-anhydride also become brittle under the same conditions and lose
their capacity to be drawn out into filaments. No data concerning the
molecular weight of co-poly anhydride are available, but its toughness and
viscosity indicate a considerably higher molecular weight than for the co-
polyesters described in paper XII.

The brittleness and loss of tenacity that develop when the co-polyanhy-
dride is stored are apparently due to hydrolysis. When a sample of the
anhydride was stored at ordinary temperature in the still in which it was
prepared, it retained its strength and toughness for six days. When the
still was opened and the anhydride transferred to a desiccator containing
anhydrous calcium chloride, brittleness developed within twenty-four hours.

The co-anhydride adheres very firmly to the glass dish in which it is
prepared, and if the specimen is more than 2 or 3 mm. deep the dish is com-
pletely shattered by the force of contraction during cooling. The co-poly-
esters manifest the same behavior.

Anal. Calcd. for (Ci Hi 6 O 3 ) w : saponification equivalent, 92. Found; 92.0, 92.3.

XIV. ANHYDRIZATION OF SEBACIC ACID 175

Sebacic ^-Anhydride. The distillate formed from the a-anhydride
during the treatment in the molecular still is a definitely macrocrystalline
solid melting sharply at 68. It dissolves readily in most of the common
organic solvents with the exception of petroleum ether and ligroin. When
recrystallized from a mixture of petroleum ether and benzene it separates
in the form of fine needles and the melting point remains unchanged. The
analytical behavior of this material identifies it as the twenty-two-mem-
bered cyclic dimer.

Anal. Calcd. for (Ci Hi 6 O 8 ) n : C, 65.18; H, 8.75; mol. wt., 368; saponification
equivalent, 92. Found: C, 65.08, 64.85; H, 8.90, 8.77; mol. wt. in boiling benzene,
386, 393; saponification equivalent 91.9.

Still more decisive support for the cyclic dimeric structure for sebacic
0-anhydride is found in the analysis of its behavior toward aniline when
treated in the manner already described for the a-anhydride.

Anal. Subs. 0.939 (after treatment with aniline): 0.0992 N NaOH, 51.2 cc.

Thus one gram of the anhydride furnished 5.41 millimoles of hydrogen
ion. The calculated value for a cyclic anhydride is 5.44, and, as the
data of Table I show, this agreement demonstrates that the 0-anhydride
cannot be an open chain unless its molecular weight lies above 3500. But
the volatility of the compound indicates that its molecular weight must lie
below 1000 (11).

The neutral solution from the above titration contained both monoanilide and di-
anilide. The alkali used in the titration was neutralized with an equivalent amount of
hydrochloric acid; the alcohol was evaporated; the residue was treated with excess
hydrochloric acid to dissolve the aniline, and filtered. The residue on the filter was
repeatedly extracted with boiling water until no more material dissolved. The com-
bined filtrates on cooling deposited handsome lustrous platelets of the monoanilide,
m. p. 122-123. This compound appears not to have been described before. Re-
crystallization from water did not raise its melting point.

Anal. Calcd. for Ci 6 H 23 O 4 N: C, 69.26; H, 8.37. Found: C, 69.51, 69.58;
H, 8.26, 8.48.

The insoluble dianilide was transferred to a Gooch crucible and weighed: yield,
0.450 g. ; theory requires 0.450 g. After recrystallization from alcohol it melted at
201-202 (12).

Anal. Calcd. for C^HafAN*: C, 74.96; H, 8.01. Found: C, 74.49, 74.67;
H, 8.34, 8.05.

The fact that dianilide was produced in this reaction demonstrates
decisively that the anhydride is not monomeric, and the close agreement
between the dianilide calculated and found indicates again that the
structure is cyclic.

176 POLYMERIZATION AND RING FORMATION

The ^-anhydride is not present as such in more than traces in the a-
anhydride before the molecular still treatment. The /3-anhydride is
exceedingly soluble in cold carbon tetrachloride, whereas rough determina-
tions indicated that the a-compound at 25 dissolves to the extent of
only about 0.04 g. in 100 cc. of carbon tetrachloride.

Sebacic 7-Anhydride. When the 0-anhydride is heated to its melting
point, cooled and then heated again, its melting point is found to have
changed from 68 to 82. This change is due to polymerization. Samples
of the new anhydride were prepared by sealing the /3-anhydride in care-
fully dried glass tubes and heating them in a water-bath. The specimens
melted at 68 to an exceedingly mobile fluid which almost instantly became
so viscous that it failed to flow when the tube was inverted. When cold,
the product was an opaque (or translucent) macrocrystalline wax that
resembled the a-anhydride. It melted at 82, and heating at 90-100 for
five hours did not raise the melting point.

When the 7-anhydride is heated in the molecular still it behaves in the
same manner as the a-anhydride and is converted to the 0- and w-anhy-
drides.

Determinations of the molecular weight of the 7-anhydride in boiling
benzene gave values in the neighborhood of 600, but for reasons already
discussed in connection with the a-anhydrides we think that these values
merely signify that the material is not monomeric or dimeric.

The behavior of the 7-anhydride toward aniline fails to reveal the pres-
ence of terminal carboxyl groups in detectable amounts.

Anal. Subs. 0.2234, 0.2707, 0.1925: (after treatment with aniline) 0.0999 N
NaOH, 12.15, 14.70, 10.50 cc. These values correspond to 5.43, 5.43, 5.45 cc. of N
NaOH per gram of sample.

Thus one gram of the anhydride furnishes 5.44 millimoles of hydrogen
ion, and from Table I it appears that if the polymer is an open chain bearing
carboxyl groups at the ends, its molecular weight must lie above 7500.

Discussion

The a-anhydride probably bears terminal groups in the manner already
indicated. We suppose that it is made up of chains of slightly differing
lengths. Our idea in heating this product in the molecular still was that
under these conditions it would condense or couple with itself by a process
of dehydration or anhydride interchange involving the elimination of ter-
minal groups between adjacent molecules.

A similar mechanism had already been established (13) for the transfor-
mation of the a-polyesters into co-polyesters. The available facts are all

XIV. ANHYDRIZATION OF SEBACIC ACID 177

consistent with this mechanism for the transformation of the a-polyan-
hydride into the co-poly anhydride. The molecules of the latter compound
therefore are very long open chains. Their molecular weights lie perhaps
in the neighborhood of 30,000, certainly at not less than 15,000. The
transformation of this compound into the /3-anhydride evidently involves
a process of anhydride interchange.

*. CO-O-CO R CO

t I t

An analogous transformation occurs by heating polyesters derived from
six-membered cyclic esters (14). But among polyesters smooth depoly-
merization in this fashion can be effected only if the structural unit is six
atoms long. Polyesters whose units are longer than six atoms yield cyclic
degradation products only in small amounts and only when they are heated
to a temperature sufficiently high to produce complete thermal decomposi-
tion (15). The much smoother degradation of the anhydride is consistent
with the much greater mobility and reactivity of anhydrides as compared
with esters.

A peculiarity of this transformation lies in the fact that the product
is not the eleven-membered cyclic monomer but the twenty-two-membered
cyclic dimer. A similar peculiarity exists in the nature of the cyclic
degradation products from the polyesters (15). The significance of this
fact will be discussed in connection with data concerning cyclic anhydrides
from other dibasic acids which will be presented in a future publication.

The formation of the 7-anhydride from the /^-anhydride apparently
involves merely the mutual coalescence of the cyclic molecules of the latter
compound. At any rate this transformation occurs under fairly rigor-
ously anhydrous conditions, and, since no terminal groups could be de-
tected in the resulting 7-anhydride, one is almost impelled to conclude that
the 7-anhydride is initially a very large ring. This conclusion is a little
difficult to reconcile with the fact that in the molecular still the 7-anhy-
dride behaves just like the a-anhydride. According to our hypothesis the
transformation of a-anhydride depends upon the presence in this compound
of terminal groups capable of being eliminated by condensation between
adjacent molecules. There is of course nothing theoretically to preclude
the possibility of the mutual coalescence of very large anhydride rings to
produce still larger rings. Moreover, it is possible that the co-anhydride
derived from the ^-anhydride and that derived from the 7-anhydride are
structurally different: the former may be an open chain and the latter a

178 POLYMERIZATION AND RING FORMATION

large ring. On the other hand, it appears that mere heating does not bring
about the further polymerization of the 7-anhydride. The conditions
of molecular distillation are required for this transformation. It is a
little difficult to see just why the further growth of the 7-anhydride mole-
cules, if they are rings, should occur only under conditions that favor the
simultaneous elimination of smaller cyclic (^-anhydride) residues. Per-
haps the resulting rupture leaves fragments (e. g., bivalent radicals) that
are especially prone to unite with one another.

We think it more likely, however, that traces of moisture inevitably ac-
quired by the 7-polymer during its transfer to the molecular still may suffice
to transform its molecules into open chains terminated by carboxyl groups.
Coupling could then occur by the elimination of water between adjacent
molecules.

One further feature of the behavior of the a-anhydride in the molecular
still remains to be mentioned. It has already been indicated that degrada-
tion to the 0-anhydride and growth to the co-anhydride occur simultane-
ously. The latter process however reaches an apparent limit after a short
time, e. #., twelve hours. But this does not interrupt the formation and
distillation of /3-anhydride, which continues at an only slightly dimin-
ished rate until only a trace of residue remains.

Summary

vSebacic a-anhydride formed by the action of acetyl chloride or acetic
anhydride on the acid is a linear polymer. The chains are probably open
and the molecular weights in the neighborhood of 5000. No smooth de-
polymerization of this anhydride can be effected by ordinary vacuum dis-
tillation, but in the molecular still two different transformations occur:
(1) coupling of the a-anhydride with itself yields w-anhydride, a polymer of
much higher molecular weight; (2) depolymerization of the a- and de-
compounds, which occurs at the same time, yields the definitely crystalline
volatile /3-anhydride melting at 68. The w-anhydride is very tough and
hard and it can be drawn out into exceedingly strong, pliable, lustrous,
highly oriented fibers. The /3-anhydride is a cyclic dimer (twenty-two-
membered ring). When heated above its melting point it instantly poly-
merizes, yielding a 7-anhydride which closely resembles the a-anhydride
in its physical properties. The behavior of these anhydrides toward ani-
line is described and discussed. The prefixes a, 0, 7 and o> used above
have been adapted arbitrarily to designate the anhydrides of different
origins.

XV. ARTIFICIAL FIBERS FROM SUPERPOLYMERS 179

Bibliography and Remarks

(1) J. W. Hill, Journ. Am. Chetn. Soc., 52, 4110 (1930).

(2) Paper XI.

(3) Voerman, Rec. trav. chim., 23, 265 (1904).

(4) Carothers and Arvin, Journ. Am. Chem. Soc. t 51, 2560 (1929).

(5) Hess, "Die Chemie der Cellulose," Akademische Verlagsgesellschaft, Leipzig, 1928, p. 432.

(6) Pringsheim and Hensel, Ber., 63, 1096 (1930).

(7) The low and inconsistent molecular weight values obtained for the polymeric anhydrides in
boiling benzene are probably due to progressive degradation by hydrolysis. Greer has lately shown
[Journ. Am. Chem. Soc., 52, 4191 (1930)] that completely dry benzene seizes with great avidity suffi-
cient moisture to provide 0.000229 g. of water for one gram of benzene. In a molecular weight deter-
mination this would be sufficient to furnish more than one gram of water for each 100 g. of anhydride.
The hydrolytic absorption of water in this proportion (1-100) would reduce the average molecular weight
of the anhydride by 75% if the initial value were 5400 and the polymer were an open chain. If the
initial value were 396 the reduction would be only 14%. If the polymer were cyclic its molecular
weight might actually be slightly increased by partial hydrolysis.

(8) We now find that Etaix [Ann. chim., [7] 9, 356 (1896)] has recorded the fact that the anhy-
drides of adipic, suberic, azelaic and sebacic acids react with ammonia to give the diamides as well
as the amic acids. Einhorn and Diesbach [Ber., 39, 1222 (1906)] have also observed that the anhy-
drides of diethylmalonic acid react with ammonia and with diethylamine to form both diamides and
amic acids. But apparently none of these investigators recognized the bearing of this behavior on
the structure of the anhydrides. Incidentally, it seems certain in view of this behavior that Einhorn
and Diesbach's tetrameric diethylmalonic anhydride is actually a sixteen-membered ring. On the
other hand, their dodecamer is probably a high polymeric mixture rather than a definite individual.

(9) Carothers and Hill, paper XI.

(10) Paper XV.

(11) Carothers, Hill, Kirby and Jacobson, Journ. Am. Chem. Soc., 52, 5284 (1930).

(12) Pellizzari, Gazz. chim. ital., 15, 555 (1885), gives 198; Barnicoat, /. Chem. Soc., 2926 (1927),
gives 200.

(13) Paper XII.

(14) Carothers, Dorough and van Natta, Journ. Am. Chem. Soc., 54, 761 (1932).

(15) Carothers, Chem. Reviews, 8, 381 (1931).

XV. Artificial Fibers from Synthetic Linear Condensation

Superpolymers*

This publication contains a very interesting and complete description of the
mechanical behavior of super polymers, when they are treated mechanically
and especially when drawn into fine filaments .f The experiments refer to
the u-ester of trimethylene glycol and hexadecamethylene dicarboxylic acid.
The filaments can be obtained directly by drawing threads from the molten
state with a rod; they can also be formed by dissolving the material in
chloroform and dry spinning it like cellulose acetate rayon. Diameters
less than Viooo inch can be obtained.

* W. H. Carothers and J. W. Hill; Journ. Am. Chem. Soc., 54, 1579-87 (1932);
Communication No. 78 from the Experimental Station of E. I. du Pont de Nemours and
Co.

Received November 12, 1931. Published April 6, 1932.

t It may be noted here that these investigations led finally to the development of
an all synthetic fiber of very outstanding qualities, which has now been put on the mar-
ket by E. I. du Pont de Nemours and Co. under the name of "Nylon."

180 POLYMERIZATION AND RING FORMATION

// no stress is applied during spinning, the filaments show properties simi-
lar to those of the massive ester from which they are prepared (m. p. about
75). If stress is applied, the fibers get transparent, lustrous, very strong,
pliable and heat resistant.

X-ray investigation shows that in the first case unorientated micellae are
present, while the filaments drawn under stress exhibit a fiber diagram
which shows that they are built up by a main valence chain lattice.
The strength of such fibers ranges from 16 to 24 kg. per sq. mm., which is
comparable with cotton (28 kg. per sq. mm.) and natural silk (35 kg. per
sq.mm.)*

It is concluded from these observations, that in the case of hydroxydecanoic
acid continuous filaments cannot be spun, unless the molecular weight lies
above 7000\; the property of cold drawing appears above 9000 and a techni-
cally useful degree of strength and pliability of the fibers requires a mo-
lecular weight of at least 12,000. This corresponds to an average length of
the chains of about 1000 A or 0.1 JJL and to a polymerization degree of about
70.

The linear condensation superpolyrners described in the three preceding
papers of this series (1) can easily be drawn out into strong, pliable, trans-
parent, permanently oriented (2) fibers. So far as we are aware no strictly
synthetic material has hitherto been obtained in the form of fibers which
thus simulate natural silk.

The superpolyester derived from hexadecamethylene dicarboxylic acid
and trimethylene glycol is for brevity referred to below as the 3-16 co-ester.

Methods of Producing Filaments. Continuous filaments can be ob-
tained from the molten 3-1 (i co-ester by touching it with a stirring rod
and drawing the stirring rod away. More uniform filaments are ob-
tained by dissolving the ester in chloroform and extruding the viscous
solution through an ordinary rayon spinneret into a chamber warmed to
permit the evaporation of the chloroform. The filaments can be picked
up and continuously collected on a motor-driven drum at the bottom of
the chamber. The production of filaments having a diameter of less than
0.001 inch presents no difficulties. It is also possible to extrude the molten
ester through a spinneret that is provided with a heating coil to maintain
the ester in a sufficiently fluid condition. When this method is used it is

* Meanwhile artificial superpolyamide fibers (Nylon) have been spun with tenaci-
ties which considerably exceed the above figures.

t This corresponds to a polymerization degree of about 40 and to a length of the
extended chains of about 400 A or 40/zju.

XV. ARTIFICIAL FIBERS FROM SUPERPOLYMERS

181

easy to apply considerable tension to the filaments as they are collected
on the drum.

Properties of the Filaments. The properties of the filaments produced
by any of these methods de-
pend upon the amount of ten-
sion or stress applied during
the spinning operation. If no
stress is used the filaments
closely resemble the massive
ester from which they are pro-
duced. They melt at 74-75
and, like the initial mass, they
are opaque and devoid of
luster. But if sufficient tension
is applied during the spinning
operation the filaments are
very different from the initial
mass in their physical proper-
ties. Instead of being opaque
they are transparent, and they
have a very high luster. They

also have a very much higher breaking strength than the initial mass, and
they are sufficiently pliable to be tied into hard knots, whereas the opaque

filaments produced without
stress are so fragile that
they can hardly be bent
without breaking. The
opaque polyester and the
transparent filaments pro-
duced from it by spinning
under tension furnish very
different x-ray diffraction
patterns. The pattern for
the opaque ester (Fig. 4) in-
dicates that it is crystalline,
but there is no sign of gene-
ral orientation. The trans-
parent filaments on the

Fig. 2. Same as Fig. 1 under crossed Nicols show- other hand f urnish a fiber
ing parallel extinction of oriented section. pattern (Fig. 5) such as one

Fig. 1. Filament from 3-16 co-ester showing
opaque and transparent sections with boundary.

182

POLYMERIZATION AND RING FORMATION

obtains from certain natural silk and cellulose fibers and rayon filaments
that have been spun under tension. The character of this pattern indi-

Fig. 3. Same as Fig. 2, 45 to cross hairs.

cates a very considerable degree of molecular orientation along the axis of
the filament (3).

Fig. 4. x-Ray diffraction pattern of
unoriented 3-16 w-ester.

Fig. 5. x-Ray diffraction pattern of
oriented 3-16 w-ester.

Cold Drawing. In connection with the formation of fibers the co-poly-
mers exhibit a rather spectacular phenomenon which we call cold drawing.
If stress is gently applied to a cylindrical sample of the opaque, unoriented

XV. ARTIFICIAL FIBERS FROM SUPERPOLYMERS

183

3-16 co-polyester at room temperature or at a slightly elevated tempera-
ture, instead of breaking apart, it separates into two sections joined by
a thinner section of the transparent, oriented fiber. As pulling is con-
tinued this transparent section grows at the expense of the unoriented
sections until the latter are completely exhausted. A remarkable feature
of this phenomenon is the sharpness of the boundary at the junction be-
tween the transparent and the opaque sections of the filament (Fig. 1).
During the drawing operation the shape of this boundary does not change;
it merely advances through the opaque sections until the latter are ex-
hausted. This operation can be carried out very rapidly and smoothly,
and it leads to oriented fibers of uni-
form cross section. The oriented and
unoriented forms of the polyester are
different crystalline states, and in the
cold drawing operation one crystalline
form is instantly transformed into the
other merely by the action of very slight
mechanical stress.

Photomicrographs of filaments of
the 3-16 co-ester are reproduced in
Figs. 1, 2, 3 and 6. The opaque fila-
ment shown in Fig. 1 was obtained
by pulling out a specimen of the molten
ester, and the transparent section
attached to it was produced by cold
drawing this filament. This figure
shows clearly the sharpness of the
boundary between the oriented and
the unoriented sections. Figures 2 and 3 represent the same sample under
polarized light. The oriented fiber is birefringent and shows parallel
extinction. The opaque filaments of Fig. 6 were obtained by spinning a
23 weight per cent, solution in chloroform of the 3-16 co-ester through a
spinneret having 0.0045 inch holes into a chamber four feet deep heated to
about 30. At the bottom of the chamber the filaments were caught and
collected on a motor driven drum. The filaments had a diameter between
0.001 and 0.002 inch. No attempt was made to stretch them during the
spinning operation, and they were opaque, devoid of luster and very fragile.
Nevertheless they furnished a fiber x-ray diffraction pattern practically
identical with that shown in Fig. 5. Thus the x-ray data indicate a high
degree of orientation while the physical properties indicate that scarcely

Fig. 6. Filaments of 3-16 w-ester
spun from chloroform solution showing
transparent fibers produced by cold
drawing (X 15).

184

POLYMERIZATION AND RING FORMATION

any orientation was effected during the spinning operation. When a
bundle of the filaments was warmed to about 35 and pulled it was drawn
out into thin, transparent fibers which were strong and pliable. Rough
determinations of the tenacity of the partially oriented opaque filaments
and the oriented transparent filaments prepared from them by cold draw-
ing indicated that the breaking strength of the transparent filaments was
about six times that of the opaque filaments. This specimen of oriented
fibers was used in some of the physical tests described in the next paragraph.
Tenacity and Elasticity. Rough determinations of the tensile strength
of rather thick fibers of 3-16 co-ester prepared by pulling filaments from the

molten ester and subsequently orient-
ing them by cold drawing gave values
ranging from 16 to 24 kg./sq. mm.
(24,000 to 36,000 Ib /sq. in. These
values compare well with those for
cotton fibers (ca. 28 kg./sq. mm.) and
silk (ca. 35 kg./sq. mm.). Under
gradually increasing stress the fibers
show a more or less definite flow point,
and then, after considerable elongation,
a point of increased resistance to stress.
The fibers are sufficiently tough and
flexible to be tied into hard knots ; thick
fibers are sufficiently stiff to be bristle-
like.

We are indebted to Dr. W. H.
Charch of the du Pont Rayon Company
for a series of physical tests on fibers
of the 3-16 co-ester spun from chloro-
form solution and subsequently ori-
ented by cold drawing. The observed tenacity was about 1.1 g. per
denier, but this value is believed to be too low owing to the fact that no
account was taken of the considerable number of broken filaments in the
specimens tested. The wet tenacity is fully equal to the dry tenacity.
(In the actual tests the fibers gave consistently higher values when
wet than dry, but the differences probably lay within the experimental
error.) A clump of filaments rolled into a small ball and compressed
showed a remarkable springiness resembling wool. In their elastic proper-
ties these fibers are very much superior to any known artificial silk. In
Fig. 7 are presented some slightly idealized curves taken from drawings

Fig. 7. x-Ray diffraction pattern of
oriented ethylene sebacate.

XV. ARTIFICIAL FIBERS FROM SUPERPOLYMERS 185

made by a modified Richards dynamometer. Bundles of filaments of vis-
cose rayon, natural silk and fibers from the 3-16 co-ester were stretched for
one minute and the curves show the extent and rate of recovery from
stretch during the first minute after release from tension. These curves
indicate that the synthetic product is even superior to natural silk, but it is
possible that the silk specimen was somewhat deteriorated.

Fibers from Other Linear Condensation Superpolymers. Fibers
closely resembling those described in the preceding section have been ob-
tained in the same way from a variety of other linear condensation polymers
in the co- or superpolymeric state. The polyesters thus far examined are

Fig. 8. Cold drawn fibers from co-anhydride of deca-
methylene dicarboxylic acid. The lack of definition is
due to specular reflection caused by high luster.

those derived from w-hydroxydecanoic acid, co-hydroxypentadecanoic acid,
ethylene glycol and sebacic acid, trimethylene glycol and adipic acid and
ethylene glycol and succinc acid. All of these materials clearly exhibit
the phenomenon of cold drawing. Wide differences in the nature of the
structural unit naturally have a considerable effect on the capacity to
form fibers and on some of the properties of the fibers. Thus ethylene
succinate in its a-form has a higher melting point and is more brittle than
most other a-polyesters derived from dibasic acids and glycols (4), and
fibers derived from co-ethylene succinate are somewhat short and brittle.
On the other hand, a-polyesters derived from adipic acid have rather low
melting points, and w-polyesters derived from the same acid show a tend-
ency to collapse at slightly elevated temperature.

Polyamides such as that derived from e-aminocaproic acid can also be got

186 POLYMERIZATION AND RING FORMATION

in the superpolymeric state (5). Owing to the very high polarity of the
amide group and the consequent high molecular cohesion, such polyamides
are exceedingly hard and insoluble in most common solvents. They show
signs of melting only at temperatures sufficiently high to produce considera-
ble decomposition. These refractory qualities make it very difficult to draw
satisfactory filaments from pure polyamides. On the other hand, mixed
polyester-polyamides of the type already described (5) yield filaments that
are easily cold drawn and, when oriented, have very high tenacity and pli-
ability.

Polyanhydrides derived from dibasic acids of the series HOOCXCH^)*-
COOH are especially easy to obtain in the superpolymeric state. The
technique involved and the nature of the reaction is discussed in detail for
sebacic acid (x = 8) in paper XIV. The behavior described there is char-
acteristic (so far as the formation of superpolyanhydride is concerned)
for those acids in which x is greater than 4. (Actual observations thus far
include those in which x is 5, 6, 7, 8, 9, 10, 11, 12 and 16.) The fibers ob-
tained from the anhydrides have an especially high degree of strength,
pliability and luster (cf. Fig. 8) but they gradually disintegrate on stand-
ing owing to hydrolytic degradation.

Discussion

Our studies of polymerization were first initiated at a time when a great
deal of scepticism prevailed concerning the possibility of applying the
usually accepted ideas of structural organic chemistry to such naturally
occurring materials as cellulose ; and its primary object was to synthesize
giant molecules of known structure by strictly rational methods. The
use of synthetic models as a means of approach to this problem had already
been undertaken by Staudinger, but the products studied by him (polyoxy-
methylene, polystyrene, polyacrylic acid, etc.) although unquestionably
simpler than naturally occurring polymers, were produced by reactions of
unknown mechanism, and their behavior, except in the case of polyoxy-
methylene, was not sufficiently simple to furnish an unequivocal demon-
stration of their structure. On the other hand, the development of the
principles of condensation polymerization described in preceding papers
of this series has led to strictly rational methods for the synthesis of linear
polymers, and the structures of the superpolymers III to VI follow directly
from the methods used in their preparation. Meanwhile the weight of
authoritative opinion has shifted; further evidence has been accumulated
(6) and cellulose has been assigned a definite and generally accepted struc-

XV. ARTIFICIAL FIBERS FROM SUPERPOLYMERS 187

ture (7). Like the synthetic products III to VI it falls in the class already
defined as linear superpolymers.

Cellulose (polyacetal)

II ... NH R CO NH R' CO NH R CO NH R'-~ CO. . .

Silk (polyamide)

III ... 0-- R -CO O R CO O R CO O R CO O R CO . .

Polyester (from hydroxy acid)

IV ... ~O R-O CO R'COORO CO R' CO-

Polyester (from dibasic acid and glycol)

V ... O - R CO NH R'~ CO NH R' CO O R CO . .

Mixed polyester-polyamide

VI ... - O CO- R CO0 COR CO O CO R CO . . .

Polyanhydride

Addition polymers of very high molecular weight have been synthesized
in the past, but the capacity to yield permanently oriented fibers of any
considerable strength has not been observed hitherto. Why is this, and
what conditions of molecular structure are requisite for the production of
a useful fiber? We regret that the limitations of space prohibit a detailed
discussion of these and accessory questions, and permit only the baldest
statement of our conclusions.

We picture a perfectly oriented fiber as consisting essentially of a single
crystal in which the long molecules are in ordered array parallel with the
fiber axis (Fig. 10). (In actual fibers a considerable number of the mole-
cules fail to identify themselves completely with this perfectly ordered
structure.) The high strength in the direction of the fiber axis and the
pliability are accounted for by the high cohesive forces of the long molecules
and by the absence of any crystal boundaries along the fiber axis (7a).
Fiber strength should depend upon molecular length, and recent work by
Dr. van Natta in this laboratory indicates that it is not possible to spin
continuous filaments from the polyester of hydroxydecanoic acid until
its molecular weight reaches about 7000. The property of cold drawing
does not appear until the molecular weight reaches about 9000. From
these and other facts we conclude that a useful degree of strength and pli-
ability in a fiber requires a molecular weight of at least 12,000 and a mo-
lecular length not less than 1000 A. (The limits for polyamides may
perhaps lie at somewhat lower values.)

Besides being composed of very long molecules, a compound must be
capable of crystallizing if it is to form oriented fibers, and orientation is

188

POLYMERIZATION AND RING FORMATION

probably necessary for great strength and pliability. Linear condensation
polymers are quite generally crystalline unless bulky substituents are
present to destroy the linear symmetry of the chains; addition polymers,
especially those produced from vinyl compounds, are more rarely crystal-
line. Possible reasons for this have already been discussed in part (8).

FOR ONE MINUTE

OH '. MINUTE f

/PCI EASE KCOVCffV

* stl ! Ml^l

Fig. 9. Elastic recovery of various fibers.

Three-dimensional polymers are obviously unsuited for fiber orientation,
and synthetic materials of this class are besides invariably amorphous.
Glyptal resins belong to this class (9). It has been proposed (10) to use
glyptal resins for the production of artificial silk, but our attempts to carry
out the proposed process led to exceedingly fragile (though lustrous)
threads which showed no signs of orientation when examined by x-rays.

Although one will not ex-

. / J ! I TT\"f"*. I 'r I J"TV? \ !.* * P e t OI "i en ted fibers to arise by

1 * T .' ; ; j [ ;*''.. i ' * ' * T"T~ any process of spontaneous

' ! j j 1 j I j V V i !' ! J J r I TT\ crystallization under the ordi-

4 " ' r ' * f * " " " * T * -- * * nary conditions, the phenom-

Fig. 10. Lattice from molecules of unequal enon of cold drawing is per-

length. haps only accidentally asso-

ciated with the capacity to

yield oriented fibers; it apparently requires a certain degree of softness
and suppleness in the molecules ; and its mechanism is doubtless analogous
to that involved in the mechanical orientation of cellulose preparations (11).
(Perhaps however, in the unoriented polyesters the molecules are in spiral
form and become extended during orientation.)

It has been suggested repeatedly by Staudinger that the great sensitivity
of cellulose and rubber to degradation by heat and by certain reagents is

XV. ARTIFICIAL FIBEKvS FROM SUPERPOLYMERvS 189

due to the fact that the upper limit of thermal stability of the very long
molecules lies close to room temperature. It is of interest in this connection
that the synthetic linear superpolyesters, in spite of their very high mo-
lecular weights, are formed at 200 to 250, and that they show no signs
of being degraded by repeated exposure to elevated temperature.

Acknowledgment

We are greatly indebted to Dr. W. H. Charch of the du Pont Rayon
Company and to Dr. E. O. Kraemer of the Experimental Station for valu-
able criticism and advice, and we have also to thank Dr. A. W. Kenney of
the Experimental Station for the X-ray data.

Summary

The linear condensation co -poly esters, polyanhydrides and mixed poly-
ester-polyamides described in preceding papers are easily drawn out into
very strong, pliable, highly oriented fibers which closely simulate natural
silk and cellulose fibers. These materials also resemble cellulose and silk
in the essential details of their molecular structure. The significance of
these analogies is discussed.

Bibliography and Remarks

(1) Papers XII, XIII and XIV.

(2) The orientation implied in this use of the term is general orientation, with reference to the
fiber axis, of the molecules or of some ordered units involving the molecules.

(3) According to some preliminary calculations made by Dr. A. W. Kenney of this laboratory
the identity periods in the oriented fibers derived from ethylene sebacate and from the 3-16 ester lie
within about 10% of the calculated lengths for one structural unit of the chain.

(4) Carothers and Arvin, Journ. Am. Chem. Soc., 51, 2560 (1929).

(5) Paper XIII.

(6) Haworth, Long and Plant, J. Chcm Soc., 129, 2809 (1927) ; Freudenberg, Ber., 54, 767 (1921);
Ann., 461, 130 (1928); Freudenberg and Brauri, ibid., 460, 288 (1929); Freudenberg, Kuhn, Durr
Bolz and Steinbrunn, Ber., 63, 1528 (1930).

(7) Sponsler and Dore, Colloid Symposium Monograph, 4, 174 (1926) Chemical Catalog Com-
pany, New York; Sponsler, Nature, 125, 633 (1930); Meyer and Mark, Ber., 61, 593 (1928); Mark
and Meyer, Z. physik. Chem , [B] 2, 115 (1929), Mark and Susich, ibid., 4, 431 (1929)

(7a) Cf. Meyer and Mark, "Der Aufbau der hochpolyrneren organischen Naturstoffe," Akade-
mische Verlagsgesellschaft, Leipzig, Germany, 1930.

(8) Carothers, Chem. Reviews, 8, 415 (1931).

(9) Ref. 8, page 402.

(10) British Patents 303,867 and 305,468.

(11) Ref. 8, p. 418.

190 POLYMERIZATION AND RING FORMATION

XVI. A Polyalcohol from Decamethylene Dimagnesium

Bromide*

This is a short description of a polyalcohol made from decamethylene di-
magnesium bromide. As this Grignard compound is a bifunctional re-
actant it can build up a linear polymer. By reaction with methyl formate
it yields a crystalline polydecamethylene carbinol. If this is heated in
the molecular still a colorless, insoluble, tough, pliable mass of a very con-
siderable molecular weight is obtained.

The action of bifunctional reactants on bifunctional Grignard reagents
(e. g., on BrMg(CH 2 ) A; MgBr) presents the possibility of producing cyclic
compounds of various types. Cyclopentanone (1), cyclohexanone and
methylcyclohexanol (2) have been prepared thus, and also 5- and 6-atom
rings containing phosphorus, arsenic (3), lead (4) and silicon (5). Concern-
ing larger rings it is recorded (2) that diacetyl reacts with pentamethylene
dimagnesium bromide but yields very little of the product corresponding
in composition with the expected 7-ring glycol. The same reagent is
said (6), to react with dialdehydes and diketones to give ring compounds,
though "not easily"; but none of these compounds are described. The
rarity and obscurity of such allusions to reactions which might yield large
rings is certainly not due to any difficulty of inducing reaction between
MgX and carbonyl or other groups. On the other hand, bifunctional
reactions are not obliged to choose between ring formation and frustration
or abortion; they may be, and in fact when the possibility of forming a
ring of 5 or 6 atoms is absent, they generally are intermolecular, and they
thus lead to linear polymeric chains (7). The diversity of the Grignard
reactions should permit the preparation of linear polymers of many differ-
ent types. Complications due to the usual side reactions of the MgX
group are multiplied when two of these groups are present in the same
molecule. Nevertheless, as the following example shows, it is possible by
such reactions to prepare linear polymers which are homogeneous in their
analytical and chemical behavior.

Methyl formate (0.2 mole) was added dropwise to an ethereal solution (250 cc.)
of decamethylene dimagnesium bromide prepared from 0.2 mole of the bromide. Each
drop of the formate immediately yielded a heavy, voluminous precipitate. After the
addition of all the formate, the mixture was refluxed for one and one-half hours. A

* W. H. Carothers and G E. Kirby; Journ. Am. Chem. Soc., 54, 1588-90 (1932);
Communication No. 48 from the Experimental Station of E. I. du Pont de Nemours and
Co.

Received November 12, 1931. Published April 6, 1932.

XVI. POLY ALCOHOL, ETC. 191

color test (8) then showed that C MgX was absent from the ethereal layer but present
in the insoluble precipitate. The mixture was treated with water and dilute hydrochloric
acid, and the undissolved solid was separated by filtration. The yield was 88% of the
theoretical. The product was completely freed of magnesium by triturating it with
warm dilute hydrochloric acid. It dissolved readily in hot alcohol, acetic acid or ethyl-
ene chloride, but not in acetone, ether or benzene. It melted at about 115-120.
More sharply melting samples (e. g,, 120-121) were obtained by repeated crystalliza-
tion from alcohol, and higher and lower melting samples were obtained in other experi-
ments. The melting points all varied with the rate of heating.

The analytical composition of this solid corresponds with the expected poly second-
ary alcohol, and its apparent molecular weight indicates an average formula
[(CH 2 )io CHOH] 6 . The equation for the reaction may be represented as

. . . (CH 2 )ioMgBr + 0=C OCH 3 + BrMg . . . - >

. . . [ (CH 2 ) 10 CH(OMgBr) ] . . .

Anal. Calcd. for CiiH 22 O: C, 77.65; H, 12.94. Found: C, 77.35, 77.40; H,
13.04,12.79. Mol. wt., calcd. for (C n H 22 O) 6 : 850. Found (in boiling ethylene chloride) :
835, 879.

It is undoubtedly mixed with traces of shorter and longer chains and with analogous
derivatives of the coupling products BrMg(CH 2 ) 2 oMgBr, BrMg(CH 2 ) 3 oMgBr, etc.,
which are always formed in small amounts during the preparation of the reagent. The
chains are probably all open, but there may be considerable variety in the nature of the
terminal groups. However, its lack of homogeneity is probably no greater than that
of a purified sample of poly vinyl alcohol (9), of which it is a homolog.

The polyaclohol is readily converted into such acyl derivatives as bromobenzoyl,
phthalyl, etc., but all of these products were obtained only in the form of oils or tars.
The composition of an oily acetate prepared by refluxing the alcohol with acetic anhy-
dride containing some sodium acetate confirms the empirical structure assigned to the
alcohol.

Anal. Calcd. for (CH 2 ) 10 CH (OCOCH 3 ) = C 13 H 24 O 2 : C, 73.59; H,
11.32; saponificatioti equivalent, 212. Found: C, 74.09, 73.61; H, 11.27, 11.33:
saponification equivalent, 207, 223, 227.

Experiments in other directions remain incomplete.

The behavior of the polyalcohol in the molecular still is noteworthy. A sample
melting at 110-113 was heated for a few hours at about 150 under a pressure less than
10~ 6 mm. A trace of crystalline solid melting at about 80 collected on the condenser
which was within 1 cm. of the evaporating surface. The residue was a colorless, semi-
transparent, very tough pliable mass. It was insoluble in the common solvents. On
being heated, it gradually became completely transparent between 60 and 100. At
215-220 it had become somewhat soft and sticky without losing its shape. Above
this temperature it first became slightly yellow and finally, at about 250, still darker
with considerable decomposition. Its analytical composition was nearly the same
as that of the alcohol from which it was derived. The change in properties indicates
a considerable increase in molecular weight, and the absence of solubility and complete
fusibility is consistent with the development of a three-dimensional polymeric molecular
structure. This might readily occur, with very little change in composition, through

192 POLYMERIZATION AND RING FORMATION

cross-linking of the polyalcohol chains by occasional ether formation. When the
polyalcohol is heated in the same way at ordinary pressure, it remains quite unchanged
in its properties ; so this example provides an even more striking illustration than those
previously recorded (10) of the powerfulness of the molecular still as a tool for displacing
chemical equilibria.

The description of these fragmentary experiments with bifunctional
Grignard reagents is presented now because it has become necessary to
suspend our work in this field.

Summary

A crystalline polydecamethylene carbinol has been prepared by the
action of decamethylene dimagnesium bromide on methyl formate. Its
properties are described. It is converted into a colorless, insoluble, tough,
pliable mass when it is heated in a molecular still.

Bibliography and Remarks

(1) v. Braun and Sobecki, Ber., 44, 1918 (1911).

(2) Grignard and Vignon, Compt rend., 144, 1358 (1907).

(3) Gruttner and Wiernik, Ber., 48, 1473 (1915).

(4) Gruttner and Krause, ibid., 49, 2CG6 (1916).

(5) Bygden, ibid., 48, 1236 (1915).

(6) v. Braun, Chem. Zentr., II, 1993 (1909)

(7) See Carothers, Journ. Am. Chem. Soc., 51, 2548 (1929), and subsequent papers of this series.

(8) Gilman and Schulze, ibid., 47, 2002 (1925).

(9) Herrmann and Haehnel, Ber., 60, 1658 (1927); vStaudinger, Frey and Starck, ibid., 60, 1782
(1927).

(10) Papers XI, XII, XIII and XIV.

XVII. Friedel-Crafts Syntheses with the Polyanhydrides of

the Dibasic Acids*

This short note deals with the Friedel-Crafts reaction of polyanhydrides
of dibasic acids. Such anhydrides react with benzene in the presence of
anhydrous aluminum chloride to yield dibenzoyl-alkane, ^-benzoyl fatty
acid and dibasic acid. The reaction has been actually carried out with
benzene and polymeric adipic and sebacic anhydride; it provides a con-
venient method of synthesis for u-benzoyl fatty acids and is presumably
quite general.

* J. W. Hill; Journ. Am. Chem. Soc., 54, 4105-06 (1932); Communication No. 100
from the Experimental Station of E. I. du Pont de Nemours and Co. See also foot-
note (1) in paper VI on page 63.

Received June 15, 1932. Published October 5, 1932.

XVII. SYNTHEvSRS WITH POLYANHYDRIDE OF DIBASIC ACIDS 193

It has been shown that the polymeric anhydrides of aliphatic dibasic
acids react smoothly and normally as acid anhydrides with aniline (1),
with phenol (la), with ammonia (2a) and with diethylamine (2b). In
each reaction three products are formed. The course of the reactions is
shown by the following scheme, which does not consider the end groups of
the polyanhydride

( OC(CHJ,,CO O), + RH ^*R OC(CH 2 ) M CO R +

4 (a)

*R OC(CH 2 ) n COOH + - A HOOC(CH 2 ), t COOH
2 (b) 4 (c)

where R is C 6 H<>NH , C 6 H 6 O , NH 2 or (C 2 H 6 ) 2 N

It has been demonstrated on theoretical grounds that the products of types
(a), (b) and (c) should be formed in the molecular ratio of 1:2:1. This
has been verified quantitatively in the aniline reaction (1).

The dibasic acid polyanhydrides have now been found to react typically
with benzene in the presence of anhydrous aluminum chloride to yield
dibenzoyl alkane, w-benzoyl fatty acid and dibasic acid.

( OC(CH,) B CO O), + CeHo -f (A1CU) >

~ C 6 H 6 CO(CH 2 )nCOC 6 H 6 -f o C 6 H 5 CO(CH 2 ) n COOH + - HOOC(CH 2 ) n COOH
4 4

This reaction has been carried out with polymeric adipic and sebacic an-
hydrides and benzene. There is no reason to doubt that it is perfectly
general and applicable to any polymeric dibasic anhydride and any suitable
aromatic hydrocarbon.

This reaction is particularly interesting in providing a convenient method
of synthesis for the w-benzoyl fatty acids which have heretofore been pre-
pared by more elaborate syntheses or in very low and uncertain yields as
by-products in the Friedel-C rafts reaction of dibasic acid chlorides with
aromatic hydrocarbons (3).

Experimental Part

Friedel-Crafts Reaction between Adipic Polyanhydride and Benzene. One hun-
dred and forty-six grams (one mole) of adipic acid was refluxed with 400 cc. of acetic
anhydride for six hours. The excess acetic anhydride and the acetic acid formed in the
reaction were removed by distillation in vacua up to 120 bath temperature. The result-
ing polyanhydride (la) was dissolved in 400 cc. of warm, dry benzene. Three hundred
grams of anhydrous aluminum chloride was suspended in 1500 cc. of dry benzene con-
tained in a 3-necked 3-liter flask fitted with a reflux condenser and a mechanical
stirrer.

The anhydride solution was added with stirring over a period of one hour and the
mixture allowed to stand overnight. The product was decomposed in ice and 250 cc.

194 POLYMERIZATION AND RING FORMATION

of concentrated hydrochloric acid added. The benzene layer was separated and ex-
tracted with dilute aqueous sodium hydroxide. The water phase deposited a small
amount of crystals on standing which were identified after recrystallization from water
as adipic acid by a mixed melting point determination.

The alkaline solution was acidified and the crystalline precipitate of cu-benzoyl-
valeric acid filtered off: weight dry, 78 g.; yield, 75%. The product was recrystallized
from a benzene-petroleum ether mixture, m. p. 70-71 (4).

Anal. Calcd. for Ci 2 Hi 4 O 3 : C, 69.9; H, 6.8. Found: C, 70.1; H, 6.7.

The extracted benzene was distilled to a small volume and chilled; 56. 5 g. of dibenzoyl-
butane was obtained, yield 85%. Recrystallized from alcohol, it was slightly pink and
melted at 105-106 (5). It was identified by means of a mixed melting point determina-
tion with an authentic specimen.

Anal. Calcd. for Ci 8 H 18 O 2 : C, 81.2; H, 6.8. Found: C, 80.6; H, 7.0.

Friedel-Crafts Reaction between Sebacic Polyanhydride and Benzene. One
hundred and one grams (0.5 mole) of sebacic acid was refluxed with 300 cc. of acetic
anhydride for five hours. The excess anhydride and the acetic acid formed were dis-
tilled off in vacua up to 120 . The residual polyanhydride was dissolved in 250 cc. of
warm dry benzene and added to 150 g. of anhydrous aluminum chloride suspended in
750 cc. of dry benzene as in the previous preparation. The products were worked up
in the same way. The crude precipitated co-benzoylnonanoic acid was extracted with
hot water to remove sebacic acid. Fifteen grams of sebacic acid separated from the
washings and was identified by a mixed melting point determination with a known
sample. The co-benzoylnonanoic acid weighed 50 g., a yield of 78%. Recrystallized
from dilute alcohol, it melted at 77-78 (6).

Anal. Calcd. for C 16 H 22 O 3 : C, 73.3; H, 8.4. Found: C, 73.3; H, 8.4.

The dibenzoyloctane was recovered from the benzene solution and weighed 35 g.,
a yield of 86%. Recrystallized from alcohol, it melted at 92-93 (7).

Anal. Calcd. for C 2 2H 26 O 2 : C, 82.0; H, 8.1. Found: C, 81.7; H, 8.1.

Summary

The polymeric anhydrides of adipic and sebacic acids react with benzene
in the presence of anhydrous aluminum chloride to yield mixtures of the
appropriate dibenzoyl alkane, co-benzoyl fatty acid, and dibasic acid. The
reaction is doubtless general.

Bibliography and Remarks

(1) (a) Hill, Journ. Am. Chem. Soc., 52, 4110 (1930); (b) Hill and Carothers, ibid., 54, 1569 (1932).

(2) (a) Etaix, Ann. Chim., [7] 9, 356 (1896); (b) Einhorn and Diesbach, Ber., 39, 1222 (1906).

(3) Auger, Ann. chim., [6J 22, 360 (1891); Etaix, ibid.. [7] 9, 391 (1896).

(4) Grateau, Compt. rend., 191, 947 (1930), gives 71.

(5) 6taix, Ann. chim., [7] 9, 372 (1896), gives 102-103.

(6) Auger, ibid. t [6] 22, 364 (1891), gives 78-79.

(7) Auger, ibid., [G] 22 363 (1891), gives 88-89.

XVIII. POLYESTERvS FROM co-HYDROXYDECANOIC ACID 195

XVIII. Polyesters from w-Hydroxydecanoic Acid*

A series of polyesters from u-hydroxydecanoic acid is described in this
paper. It represents without doubt the best established set of synthetic
high polymers concerning identity, uniformity and structure.
The lower members were prepared by heating u-hydroxydecanoic acid at
150 in a normal distilling flask; the higher ones were obtained in the
molecular still previously described.^

The molecular weights range from 780 to 25,000 and were determined by
titration of the terminal acid group; the highest member was also
measured in the Svedberg Ultracentrifuge. The agreement is excellent
(25,660 and 25,200).

The polyesters below 10,000 dissolve rapidly and completely in cold chloro-
form or benzene and in various hot organic solvents. The highest mem-
bers show greatly diminished solubility and swell before going slowly into
solution. The melting points increase with the molecular weight until
about 1000 (being then around 70), which corresponds to a polymerization
degree of jive and to a length of about 100 A for the extended chain. Be-
tween 1000 and 25,000 there is no further increase of melting points. All
esters are microcrystalline; the lower members being waxy and brittle, the
higher ones hard, horny and very tough.

The samples above 10,000 (polymerization degree 50, length of the extended
cJiain around 1000 A) can readily be spun into fibers. The initial fila-
ments are opaque and fragile; by stretching they become transparent,
pliable and very strong. The x-ray diagram shows the presence of fiber
orientation. Up to a molecular weight of 5700 (polymerization degree 28)
no fibers can be obtained. The ester with m. w. 7330 furnished filaments
but they could not be extended. The ester 9330 gave stretchable fibers
but they were very weak. Above 10,000 strong extendable filaments are
obtained.

These esters have also been used to test the viscosity equation of Stand-
inger.** It was found that this surprisingly simple relation between
viscosity and molecular weight holds fairly well up to about 15,000. A hove
this point the viscosities rise more rapidly than the equation requires,
and the law loses its strict validity.

* W. H. Carothers and F. J. van Natta; Journ. Am. Chem. Soc., 55, 4714-19
(1933); Communication No. 129 from the Experimental Station of E. I. du Pont de
Nemours and Company.

Received August 8, 1933. Published November 7, 1933.

t Compare paper XI, pages 154 to 156.

** E. O. Kraemer and F. J. van Natta; /. Phys. Chem., 36, 3175 (1932).

196 POLYMERIZATION AND RING FORMATION

The obvious importance of simple synthetic models as an aid in studying
macromolecular materials has been emphasized repeatedly by Staudinger
(1), who has used polystyrene, polyoxymethylene, polyacrylic acid, etc.,
for this purpose. Our own researches on condensation polymers were
started with the idea that the fact of a proposed model's being synthetic is
of little value unless the method of synthesis is rational, i. e., unless it is
sufficiently clear-cut to leave no doubt concerning the structure of the
product. Polystyrene may, for example, serve as a simplified model of
rubber, but it has the disadvantage that the method used in its synthesis (a
spontaneous polymerization of unknown mechanism) furnishes no certain
clue to its structure. The independent demonstration of its structure pre-
sents the same difficulties as does rubber; in fact today the formula of
rubber can be written with more assurance than that of polystyrene.

In the first paper of this series (2) it was pointed out that bifunctional
condensations frequently proceed by known mechanisms. In the second
paper (3) it was shown for polyesters derived jointly from dibasic acids
and glycols that the reaction consists exclusively in esterification while the
average size of the product molecule increases progressively with the com-
pleteness of the reaction. Rational, deliberate control over the average
molecular weight is thus made possible, and, as was showu later (4), by
using more drastic conditions molecules of average weight greater than
20,000 can be obtained. Meanwhile a study of the self-esterification of
co-hydroxydecanoic acid was presented by Lycan and Adams (5), who con-
cluded that the products must be formulated like the polyesters referred to
above.

We have now greatly extended the range of polyesters derived from
co-hydroxydecanoic acid and have obtained the series of fractions of differ-
ent average molecular weights shown in Table I. The structure of these
polyesters HO[ (CH 2 ) 9 CO O ] M (CH 2 ) 9 CO OH follows from the
method used in their synthesis (see Experimental Part) and the fact that
they can be hydrolyzed quantitatively to co-hydroxydecanoic acid. The
only important uncertainty is the distribution, in a given sample, of the
molecular weight about the observed average (i. e., the degree of homo-
geneity). However, since these esters are all crystalline solids and since
the mer or unit of the chain is quite long, the homogeneity is probably
better than in the case of such materials as polystyrene which can be
purified only by fractional precipitation or extraction. Moreover, a
determination of the molecular weight of the highest member of the series
by the Svedberg ultracentrifugal method indicates that the weight of most
of the molecules lies fairly close to the observed average. It is indeed our

XVIII. POLYESTERS FROM w-HYDROXYDECANOIC ACID

197

opinion that this series of polyesters is more definitely and certainly es-
tablished in identity and structure than any similar series of macromo-
lecular compounds yet described.

TABLE I
POLYESTERS FROM CO-HYDROXYDECANOIC ACID

Mol. M. p., v 85

wt. C. di n D Spinnability Tensile strength

780 66-67 1.0957 1.4494 Absent
1720 72-74 1.0935 1.4506 Absent
3190 74-75 1.0877 1.4517 Absent
4170 74-76 1.0814 1.4517 Absent

5670 73-75 1.0751 1.4518 Very short fibers. No cold drawing
7330 74-75 1.0715 1.4517 Long fibers, but cold drawing absent Very weak
9330 75-76 1 .0668 1 .4518 Long fibers which cold draw Very weak

16900 77-78 1 . 0627 Easily spins and cold draws 13.1 kg./mm. 2

20700 77-78 1 . 0632 Spins with difficulty but easily cold draws 12.3 kg./mm. 2

25200 75-80 1.0621 1.4515 Spins above 210 and cold draws 7.0 kg./mm. 2

TABLE II
MOLECULAR WEIGHTS OF POLYESTERS

Observed tnol. wts.
(by titration)

A ^erage

Calcul ted
Observed mol. wts. length of
(by other methods) molecule in A.

776

784

780

930 (in boiling benzene)

1707

1722

1720

1620 (in boiling benzene)

123

3188

3201

3190

Not measured

188

4114

4226

4170

Not measured

313

5626

5717

5670

Not measured

440

7327

7329

7330

Not measured

570

9331

9335

9330

Not measured

730

16890

17110 16790

16900

Not measured

1320

20380

20930

20700

Not measured

1610

24240

25760 25660

25200

26700 (ultracentrifuge)

1970

Molecular Weights. Molecular weights were estimated by titration
with standard alcoholic potash of the polyesters dissolved in a chloroform
alcohol mixture. Phenolphthalein was used as an indicator. This method
applied to pure lauric acid gave values agreeing sharply with the theoretical
(200.4, 199.6 and 199.2 as against 200.2). Observed values for the poly-
esters are shown in Table II. It is interesting to note that no difficulty
was encountered in titrating the highest member of the series where the
acid hydrogen is only one part in 25,000.

The average equivalent weight measured by titration will of course be
identical with the average molecular weight only if each molecule bears
a terminal carboxyl in the manner required by the indicated structure.

198 POLYMERIZATION AND RING FORMATION

The accidental loss of terminal carboxyls under the conditions of prepara-
tion used seems rather improbable, but indication of the absence of such
loss is provided by independent estimations of molecular weight. The
boiling point method furnished values for the first two members agreeing,
within the probable experimental error, with the much more sharply re-
producible values determined by titration. For compounds above mo-
lecular weight 3000, in our experience, boiling-point or freezing-point
methods are not self-consistent within 10%, and above 10,000 they are
practically worthless. We were, however, very fortunate in being able to
obtain a value for the highest polymer by the Svedberg ultracentrifugal
method (6). The agreement is all that could be desired.

Physical Properties. The polyesters below 10,000 dissolve rapidly
and completely in cold chloroform or benzene and in hot acetone, ethyl
acetate and acetic acid. They are practically insoluble in hot alcohol,
ligroin or water. The highest members show diminished solubility in
benzene and in hot acetone and ethyl acetate. They dissolve copiously in
chloroform, but solution occurs only slowly and is preceded by some
swelling. At ca. 110 the first member of the series is a highly viscous
liquid while the highest member is a transparent resin that is too stiff to
flow and sufficiently elastic to offer resistance to permanent deformation.

Lycan and Adams have pointed out (5) that the melting point of self-
esters of hydroxydecanoic acid increases with increasing molecular weight
up to 1000, but changes little between 1000 and 9000. As the data of
Table I show, there is also no further increase between 9000 and 25,000.
The polyesters separate from solution in the form of white powders which
give sharp x-ray powder diffraction patterns. At least in the lower mem-
bers of the series crystallinity can also be demonstrated by microscopic
observation. A very dilute solution of the 3190 ester in butyl acetate,
when examined at a magnification of 400, showed the separation of tiny
flat glittering plates. The molten esters in thin layers crystallize very
rapidly, but with the lower members one can observe the crystallization to
start with the emergence and growth of innumerable doubly refracting
centers, apparently spherulites. At a magnification of 900 the growth ap-
pears to involve the intermeshing of radiating clusters of needles.

The masses resulting from crystallization of the molten polyesters are
opaque solids. The lower members are waxy and brittle ; when fractured
they show no planes of cleavage. The highest members are harder and so
horny and tough that they can scarcely be fractured.

In Table I are listed the densities of the solid esters at 25. The values
diminish as the molecular weight rises, quite rapidly at first and then more

XVIII. POLYESTERS FROM o>-HYDROXYDECANOIC ACID 199

slowly until at a molecular weight of 16,900 they become almost constant.
This at first sight is very surprising. Polymerization usually involves an
increase in density. There is, for example, an increase of 23% in passing
from chloroprene to polychloroprene (7) The interpretation of the
present case can scarcely be attempted in the absence of data on the
molten esters, which are not yet available ; but it is easy to imagine that
the longer the molecules are the more difficulty they will have in lining
themselves up perfectly in the crystal lattice. The refractive indices of
the molten esters at 85 are presented in Table I. There is a slight in-
crease from 780 to 1720, a smaller increase from 1720 to 3190, and beyond
that no further change.

Fiber Formation. In a previous paper (8) it was shown that linear
polyesters of molecular weight above 10,000, when melted or dissolved,
can be spun into fibers. The initial filaments are opaque and fragile, but
when stress is applied to them they are readily elongated several fold and
then remain permanently extended. The stretching is accompanied by a
loss of opacity, and an enormous increase in tensile strength and pliability.
Examination by x-rays shows the presence of fiber orientation.

These fibers are, we believe, the first examples of a synthetic material
being obtained in the form of fibers having any considerable degree of
strength, orientation and pliability. The analogies in chemical structure
with cellulose and silk are especially interesting, and the esters of the present
paper provide some data (Table I) on the relation between molecular
weight and fiber-forming ability. Until a molecular weight of 5670 is
reached, the viscosity and coherence of the molten esters are so low that
they do not yield continuous filaments. The ester of molecular weight
7330 furnished continuous filaments which, however, could not be stretched
and oriented. The 9330 ester could be stretched, but the oriented fibers
were too weak to permit the determination of tensile strength. Oriented
fibers from the 16,900 ester had a tensile strength of 13.1 kg./sq. mm. or
1.36 g. per denier, which is in the same range as good regenerated cellulose
fibers. The value for the 20,700 ester (12.3 kg./sq. mm.) differs from that
for the 16,900 ester by an amount that probably lies within the experi-
mental error. The tensile strength of the 15,200 ester is, however, defi-
nitely much lower. It is exceedingly improbable that the fiber strength
of completely oriented molecules rises to a maximum and then falls off as
the length of the molecule increases. On the other hand, the degree of
orientation that can be produced by cold-drawing probably depends upon
other factors besides molecular weight. An unusually high temperature
(210) was required to soften the 25,200 ester sufficiently to permit spin-

200 POLYMERIZATION AND RING FORMATION

ning, and its relatively low tensile strength may therefore reasonably be
ascribed to factors associated with orientation. It may be observed that
the ability to form strong, highly oriented fibers does not appear until the
molecular length reaches some value lying between 700 and 1300 A.

Viscosity and Other Properties. The viscosities of dilute solutions
of these polymers in tetrachloroethane have been carefully measured.
The results already reported and discussed (9) show that Staudinger's
simple empirical equations relating viscosity to molecular weight (1) are
satisfied fairly well until the molecular weight reaches 16,900; beyond this
point the viscosities rise more rapidly than the equations require. Staud-
inger has assumed that the relations between viscosity and molecular
weight established for compounds below 10,000 will hold for compounds of
higher molecular weight. The data referred to above provide the first
experimental test of this assumption, and for the polyesters under con-
sideration it is shown to be not strictly valid. Molecular weights that
have been assigned to cellulose, rubber, high polystyrenes, etc., on the
basis of viscosity measurements must therefore be considered as subject to
considerable uncertainty.

Study of the physical properties of polyesters derived from hydroxy-
decanoic acid is being continued, and further results will be reported in
future papers.

Experimental Part

Preparation of Polyesters. The lowest members of the series were obtained by
heating co-hydroxy decanoic acid at 150 in an ordinary distilling flask provided with a
receiver. Diminished pressure and higher temperatures were used for higher members
of the series; the highest members (above 10,000) were obtained by heating the hydroxy-
decanoic acid or its lower polyesters in a molecular still consisting essentially of a suction
flask into which a water-cooled test-tube was inserted to act as a condenser. The
products were crystallized several times to increase their homogeneity. Typical pro-
cedures are indicated below.

Polyester of Molecular Weight 4170. Twenty grams of the acid was heated at
150-175 under atmospheric pressure for one and one-half hours, then at 200 at 1 mm.
pressure for eight hours. The resulting waxy product was crystallized by dissolving in
a small amount of boiling chloroform and adding several volumes of hot acetone. The
product had an apparent equivalent weight of 3906. A second crystallization from hot
acetone by slow cooling raised the equivalent weight to 4170. The ester was obtained
as a chalky powder, soluble in chloroform and benzene and in hot acetone, ethyl acetate
and acetic acid. Hot alcohol, ligroin or water did not dissolve it appreciably.

Anal. Calcd. for HO [(CH 2 ) 9 CO O] 24 H: C, 70.22; H, 10.66; mol. wt.
4101; saponification equivalent, 170.8. Found: C, 70.35, 70.70; H, 10.69, 10.88;
mol. wt. by titration, 4114, 4226; saponification equivalent, 170.7.

Polyester of Molecular Weight 9330. Thirty-six grams of the acid was heated

XVIII. POLYESTERS FROM co-HYDROXYDECANOIC ACID 201

for seven hours at 120-175 at 1 mm. pressure, then for twenty-five hours at 225 at 1
mm. It was crystallized several times from hot acetone by slow cooling and was ob-
tained as a pure white powder. It did not differ appreciably from polymer 4170 in solu-
bility.

Anal. Calcd. for HO[(CH 2 ) 9 CO O] 66 H: C, 70.39; H, 1066; mol. wt.,
9376; saponification equivalent, 170.5. Found: C, 70.56; H, 10.87; mol. wt. by titra-
tion, 9335, 9331; saponification equivalent, 169.6.

Polyester of Molecular Weight 20,700. Fifteen grams of hydroxydecanoic acid
was heated at 230 in a molecular still for thirty hours. During the first few hours a
considerable amount of distillate (dimer and unchanged acid) collected on the condenser.
This was removed and the residue in the flask was stirred frequently. At the end of the
reaction the mass was plastic when hot. It was purified by dissolving in a small amount
of hot chloroform, then adding several volumes of acetone and allowing the product to
crystallize by cooling. This polymer was less soluble than lower members in benzene
or in hot acetone or ethyl acetate. Solution was slow and was preceded by some
swelling.

Anal. Calcd. for HO- [(CH 2 ) 9 CO O ]| 21 H: C, 70.47; H, 10.66; mol. wt..
20,605; saponification equivalent, 170.3. Found: C, 70.72, 70.29; H, 10.62, 10.47;
mol. wt. by titration, 20,380, 20,930; saponification equivalent, 170.8.

Saponification of the polyesters by alcoholic alkali always resulted in complete
solution, indicating the absence of any appreciable amount of non-ester ingredient, and
the following experiment showed that the product of saponification was the initial hy-
droxydecanoic acid: 0.3097 g. of the 13,600 polyester was saponified with alcoholic
sodium hydroxide; the alcohol was removed and the solution was acidified and extracted
with ether. Evaporation of the ether yielded 0.34 g. (99.3% of the theoretical amount)
of hydroxydecanoic acid.

During the preparation of the polyesters a small amount of the cyclic dimeric lac-

I n

tone, a 22-membercd ring O (CH 2 ) 9 CO O (CH 2 ) 9 CO, was always formed. It
was obtained in the form of needles from dilute alcohol; m. p. 95-96.

Anal. Calcd. for C 2 oH 3 6O 4 : C, 70.58; H, 10.67. Found: C, 70.07, 70.04; H,
10.53, 10.55.

This lactone has already been prepared by Lycan and Adams (5) indirectly from the
potassium salt of hydroxydecanoic acid and acetic anhydride followed by dry distilla-
tion at 400-500 .

Summary

A series of polyesters prepared from w-hydroxydecanoic acid and ranging
in molecular weight from 780 to 25,200 is described. Strong, oriented
fibers are obtained only from members having molecular weights above
9330. The influence of molecular weight on some other physical proper-
ties is discussed.

Bibliography and Remarks

(1) Staudinger, "Die hochmolekularen organischen Verbindungen," Julius Springer, Berlin, 1932.

(2) Carothers, Journ. Am. Chem. Soc., 51, 2548 (1929).

(3) Carothers and Arvin, ibid , 51, 25GO (1929).

202 POLYMERIZATION AND RING FORMATION

(4) Carothers and Hill, ibid., 54, 1559 (1932).

(5) Lycan and Adams, ibid., 51, 625, 3450 (1929).

(6) Kraemer and Lansing, Journ. Am. Chem. Soc., 55, 4319 (1933).

(7) Carothers, Williams, Collins and Kirby, Journ. Am. Chem. Soc., 53, 4203 (1931).

(8) Carothers and Hill, ibid., 54, 1579 (1932).

(9) Kraemer and van Natta, J. Phys. Chem., 36, 3175 (1932).

XIX. Many-Membered Cyclic Anhydrides (1)*

This paper contains first some definitions:

Linear polymers are those the molecules of which are built up from a

recurrring bivalent structural unit.

Superpolymers are linear polymers having molecular weights above 10,000.

Unit length is the number of atoms in the chain of the structural unit.

Macrocyclic compounds are rings of more than seven atoms.

Monomers are compounds containing only one structural unit.

Dimers contain two structural units.

Bifunctional compounds are molecules bearing two groups capable of mutual

reaction.

Intramolecular reaction leads to ring closure,

Intermolecular reaction to chain formation.

a-Anhydrides] of a series of dicarboxylic acids were prepared by the action
of acetic anhydride or acetylchloride on

pimelic acid dodecandioic acid

suberic acid brassylic acid

azelaic acid tetradecandioic acid

undecanedioic acid octadecandioic acid

They are all linear polymers with molecular weights between 3000 and
5000. When heated in the molecular still**, they disproportionate, yield-
ing

(a) <jo-anhydrides (residue). (These arc super polymers with very high
molecular weight) and

(b) ^-anhydrides (distillate), which are cyclic monomers or dimers.

The latter are converted by heat or time into y-anhydrides, which are
very similar to the a-modification.

* J. W. Hill and W. H. Carothers; Journ. Am. Chem. Soc., 55, 5023-31 (1933);
Communication No. 131 from the Experimental Station of E. I. du Pont de Nemours
and Co.

Received September 2, 1933. Published December 14, 1933.

t Compare paper XIV on page 168.

** See paper XI on page 154.

XIX. MANY-MEMBERED CYCLIC ANHYDRIDES 203

Terminology. To clarify the text of subsequent papers the following
definitions are presented. Some of the terms here defined were first intro-
duced in previous papers of this series (2), but they have not yet gained
very wide currency, while in some cases they have been used in a sense
different from that suggested.

Linear Polymers, e. g.

. . . CH 2 CH 2 0CH 2 CH 2 O CH 2 CH 2 O CH 2 CH 2 O CH 2 CH 2 O CH 2 CHoO . . .

are those whose molecules are built up from a recurring bivalent radical
or structural unit, e. g., CH 2 CH 2 O . Linear polymers are not neces-
sarily open chains; they may be rings. Super polymers are linear polymers
having molecular weights above 10,000. The unit length is the number of
atoms in the chain of the structural unit ; in polyethylene glycol (or oxide)
the unit length is three. Macrocyclic compounds are rings of more than
seven atoms. Monomers are compounds containing only one structural
unit; ethylene oxide arid ethylene glycol are both monomers, since each
has one CH^CHaO group. Dimers contain two units: dioxane and
diethylene glycol are both dimers. Bifunctional compounds are chains
bearing two groups capable of mutual reaction with the formation of a
new bond. Bifunctional compounds always formally present two possi-
bilities of self -reaction : intramolecular leading to ring closure, and inter-
molecular leading to chain formation.

Cyclic Anhydrides/ In previous papers, studies of adipic (3) and
sebacic (4) anhydrides have been reported. The present paper describes
the extension of these studies to the anhydrides of pimelic, suberic, azelaic,
undecanedioic, dodecanedioic, brassylic, tetradecanedioic and octade-
canedioic acids. The results closely resemble those already reported.

The anhydrides obtained by the action of acetic anhydride or acetyl
chloride on the acids are linear polymers of the type formula O CO
R CO O CO R CO O CO RCO . . . These products are
called a-anhydrides to distinguish them from other forms (/3, 7, cu) that
originate in other ways. They are very reactive, microcrystalline powders
which are only slightly soluble in organic solvents and have molecular
weights in the neighborhood of 3000 to 5000. They react with aniline to
yield a mixture of the dianilide, the monoanilide and the dibasic acid in
the ratio 1:2:1 (3, 4). This reaction serves to distinguish any sort of
polymeric dibasic acid anhydride from the cyclic monomer, which yields
exclusively with monoanilide.

When the a-anhydrides are heated in a high vacuum with a condenser

204 POLYMERIZATION AND RING FORMATION

placed close to the evaporating surface (molecular still) (5), they undergo
the transformations summarized in the following scheme :

a- Anhydride 0-Anydride (distillate)

linear polymer > cyclic monomer or

moLwt.ca.5000 _ dimer

w- Anhydride (residue) -y-Anhydride

superpolymer < linear polymer (similar

mol. wt. very high to a-anhydride)

The a-anhydride yields /3-anhydride and co-anhydride simultaneously.
As the 0-anhydride collects on the condenser, the residue increases pro-
gressively in viscosity and molecular weight. The resulting co-anhydride
when cold is a very tough, opaque, solid; it becomes transparent (melts)
at a definite temperature without flowing; and at still higher temperatures
it can be drawn into pliable, highly oriented fibers (6). The co-anhydride
depolymerizes just as the a-anhydride does, and if heating in the molecular
still is continued long enough complete conversion to ^-anhydride finally
occurs. The /3-anhydrides on being heated or allowed to stand readily
revert to a polymeric form, the 7-anhydride. In their physical properties
the 7-anhydrides are generally practically indistinguishable from the a-
anhydrides. However, the latter structurally are probably open chains
terminated by acetyl groups while the former at least when freshly formed
under anhydrous conditions are perhaps giant rings (4) .

Adipic anhydride differs from the higher homologs in that the a-an-
hydride depolymerizes relatively easily even in an ordinary distillation
apparatus. The other anhydrides differ from each other only in the nature
of the volatile jft-anhydride which they yield on depolymerization. The
facts in this connection are shown in Table II. It will be observed that
those anhydrides whose unit lengths are 7, 8, 10, 12, 14, 15 and 19 yield
monomeric 0- anhydrides and that those whose unit lengths are 9, 11 and
13 yield only dimeric /3-anhydrides. These relations are represented
graphically in Table I.

The three dimeric anhydrides (suberic, sebacic and dodecanedioic) are
sharply crystalline solids sufficiently stable that their molecular weights
can be determined cryoscopically in benzene. When allowed to react with
aniline they furnish the dibasic acid, its monoanilide and its dianilide in
the ratio 1:2:1. When heated above their melting points the dimeric
anhydrides polymerize almost instantly to 7-anhydrides.

The monomeric anhydrides are not sufficiently stable to permit cryo-

XIX. MANY-MEMBERED CYCLIC ANHYDRIDES 205

scopic molecular weight determinations, but their identity as monomers
is established by the fact that they react with aniline to yield pure mono-
anilide. The first and last members of the series are low-melting crystalline
solids; the others are all liquids.

The monomers show characteristic differences in stability. Adipic
0-anhydride, the seven-membered ring, polymerized completely in about
seven hours at 100 under rigorously anhydrous conditions and fairly
rapidly at room temperature in the presence of traces of moisture. The
8-, 10- and 12-membered monomers polymerize so rapidly even at very

TABLE I
NATURE AND V STABILITY OF 0- AN HYDRIDES

Dimers

stable up to Monomers

melting point Unit length Unstable Extremely unstable

(adipic)

7 - --- > 7

(pimelic)

(suberic)

(azelaic)

10 ----- - -------------- ---------- > 10

(sebacic)

(undecancdioic)

10 ._ ...... _ _______ y 19

LA -- ~ ' " " ~ iA

(dodecanedioic)

(brassylic)

14 - ----- - - -- > 14

(let radecanedioic)

15 ---------- > 15

(octadedanedioic)
19 ---- - - -- - ---- > 19

low temperature that a special technique was required to demonstrate
their temporary existence. When distilled in the molecular still onto a
condenser cooled by tap water, they condensed as transparent fluid drops,
which in the course of an hour or less became opaque, and in several hours
set to hard waxes. Distillate collected under these conditions invariably
contained polymer, since its reaction with aniline always yielded dianilide.
In the method devised, the condenser was cooled by means of liquid air.
At this temperature the monomers condensed as solids. When, after
several hours, a sufficient quantity of distillate had collected, a small

206

POLYMERIZATION AND RING FORMATION

reservoir of aniline (Fig. 1) was connected with the vacuum system by
opening a stopcock, and a layer of crystalline aniline was deposited by
evaporation on top of the anhydride distillate. The condenser was then
allowed to warm up gradually, and as soon as the aniline started to melt,
was quickly removed from the still and the lower part, bearing the distil-
late, was submerged in aniline. When the reaction was completed, the
mixture was examined for dianilide and monoanilide.

When this method was ap-
plied to pimelic anhydride (8-
atom unit) it yielded pure
monoanilide. The next two
members, azelaic and unde-
canedioic anhydrides (10- and
12-atom units) gave mixtures
of acid, monoanilide and di-
anilide. However, the pres-
ence of some monomer in the
distillate was established by
the fact that the ratio of mono-
anilide was considerally higher
than that required by the
Fig. 1 . theory for dimer or other poly-

mer. Moreover, when the con-
denser was operated at a higher temperature, these anhydrides first collected
as transparent fluid drops with no evidence of the presence of crystalline
material (dimer). It seems very probable therefore that the distillates

Acid

Adipic

Pimelic

Suberic

Azelaic

Sebacic

Undecanedioic

Dodecanedioic

Brassylic

Tetradecanedioic

Octadecanedioic

TABLE II

CYCLIC ANHYDRIDES

Product

Structural unit

of depolymerization

of anhydride

and size of ring

OC(CH 2 ) 4 CO--O

Monomer 7 1

OC(CH 2 ) 6 O>- O

Monomer 8 1

OC(CH 2 ) 6 CO O

Dimer 18

OC(CH 2 ) 7 CO O

Monomer 10 ]

OC(CH 2 ) 8 CO

Dimer 22

OC(CH 2 ) 9 CO

Monomer 12 ]

OC(CH 2 )i CO O

Dimer 26 J

OC(CH 2 ) n CO

Monomer 14 1

OC(CH 2 ) 12 CO O

Monomer 15 1

OC(CH 2 ) 16 CO

Monomer 19 1

Stability M. p., C.

Unstable 20

Extremely unstable Liq.

Stable up to m. p. 56-57

Extremely unstable Liq.

Stable up to m. p. 68

Extremely unstable Liq.

Stable up to m. p. 76-78

Unstable Liq.

Unstable 36-37

XIX. MANY-MEMBERED CYCLIC ANHYDRIDES 207

were initially pure monomers and that the presence of dianilide in the
aniline reaction product was due to subsequent polymerization.

The next three members of the monomer series were much more stable
than those of 10- and 12-atoms. Brassylic anhydride (14-atom unit)
yielded pure monoanilide when the condenser was cooled with solid carbon
dioxide and acetone. The liquid distillates from the 15- and 19-membered
anhydrides (tetradecanedioic and octadecanedioic) when condensed at
the temperature of tap water yielded pure monoanilide if allowed to react
with aniline within a few hours of distillation. These two monomeric
anhydrides, like that of adipic acid, could also be preserved for several
days at room temperature without complete polymerization.

Experimental Part

Preparation of a-Anhydrides. The a-anhydrides were prepared by the following
method. A small quantity of the acid (5 to 20 g.) was refluxed with three parts by
weight of acetic anhydride for four to six hours. In the case of the less soluble octa-
decanedioic acid six parts of acetic anhydride was used. The volatile material was dis-
tilled off under the vacuum of a water aspirator. The crude anhydrides were dissolved
in hot dry benzene and precipitated, after nitration of the solutions, with petroleum
ether. They were then preserved in a vacuum over potassium hydroxide, phosphorus
pentoxide and paraffin. The a-anhydrides separate from solution as white micro-
crystalline powders. In the molten state they are very viscous liquids which crystallize
on cooling in the form of small radiating clusters of microscopic needles and finally
solidify to hard waxes.

TABLE III
ANALYTICAL DATA FOR a-PoLYANHYDRiDES

a-Anhydride
Pimelic
Suberic
Azelaic
Undecanedioic
Dodecanedioic

Empirical
formula

C 7 H 10 3
C 8 H 12 3
C 9 H 14 3
C n H 18 3
C 12 H 20 3

M. p.,
C

53-55
65-66
53-53.5
69-70
86-87

Calcd.
C H

59.11 7.11
61.50 7.75
63.49 8.29
66.60 9.18
67.86 9.52

Found
C

61.67

63.04
65.90
66.88

61
63
65
67

.04
.09
.71
.01

7,
7,
9
9

,68
.86
.31
.63

7.42
7.91
9.29
9.84

Brassylic Ci 3 H 22 O 3 76-78 68.97 9.81 68.80 68.98 9.96 10.05

Tetradecanedioic C 14 H 24 O 3 89-91 69.94 10.08 69.17 69.32 9.94 9.79
Octadecanedioic Ci 8 H 32 O 3 94-95 72.90 10.90 72.69 72.02 10.69 10.49

Reaction of the Anhydrides with Aniline. A small sample of the anhydride (ca.
2 g.) was added to 5 to 10 cc. of aniline and triturated with a stirring rod until a smooth
cream was formed. The mixture was then treated with 10% hydrochloric acid to dis-
solve the excess aniline, cooled and filtered. The mixtures derived from the anhydrides
below undecanedioic were separated by treatment with boiling water, in which the
dianilides are completely insoluble. The monoanilides separated from the filtrate on
cooling. Sufficient water was used to keep the dibasic acids in solution. The separa-
tion of the mixtures from the anhydrides above sebacic was accomplished by means of

208

POLYMERIZATION AND RING FORMATION

dilute aqueous sodium hydroxide, which dissolved the monoanilide and the dibasic acids
as sodium salts and left the dianilide. The monoanilide and the dibasic acid were pre-
cipitated from the filtered solution by acidification, and separated by means of boiling
water. In the cases of tetradecanedioic and octadecanedioic anhydrides, the alkaline
separation was carried out hot, as the sodium salts of these monoanilides are difficultly
soluble in the cold. Up to and including undecanedioic a-anhydride, the dianilide was
estimated quantitatively and in each case was obtained in 25% yield, as required by
theory. A clean separation was difficult with the higher homologs as the alkaline solu-
tions were increasingly soapy. The properties of the anilides are given in Table IV.

TABLE IV

MONOANILIDES

Empirical

Calcd.

-Analyses,

YO *

Found

M. p., C.

Cryst. from

formula

Pimelic

108-109

H 2 O

CuHivOsN

66.35

.27

66.23

7.55

Suberic

128-129

Benzene

CnHi 9 O 3 N

67.42

.70

67.51

7.86

Azelaic

107-108

Dil. EtOH

CuH 2 iO 3 N

68.39

.05

68.05

68,

.32

7.89

8.02

Undecanedioic

112.5-113

50% EtOH

Ci7H 2 *O 3 N

70.06

.65

70.28

70,

.10

8.94

9.03

Dodecanedioic

123

50% EtOH

CisHavOsN

70.77

.91

70.85

.13

8.96

8.72

Brassylic

118.5-119.5

50% EtOH

CuHaOsN

71.41

.18

71.71

71.

.46

9.60

9.37

Tetradecanedioic

124-125

60% EtOH

C 2 oH 31 3 N

72.01

.39

71.97

.27

9.42

9.10

Octadecanedioic

128-129

60% EtOH

CwHasOsN

73.98

.10

74.15

74.

,09

10.02

9.92

Pimelic

Suberic

Azelaic

Undecanedioic

Dodecanedioic

Brassylic

Tetradecanedioic

Octadecanedioic

DIANIUDES

155-156 MeOH-H 2 O CiH 22 OtN 2 73.50

186-187 MeOH CsoH^Na 74.03

186-187 Xylene C2iH 26 O 2 N 2 74.51

160-161 EtOH C 23 H 30 O 2 N 2 75.35

170-171 EtOH C 2 4H 32 O 2 N 2 75.73

160-161 EtOH Q*H 3 4O 2 N 2 76.08

169.5-170 EtOH C 2 6H 3 eO 2 N 2 76.41

162-163 EtOH C 3 oH 4 4O 2 N 2 77.52

7.16

7.46

7.75

8.27 75.44 75.32

8.49 76.28 76.09

8.70 75.86 76.16

8.89 76.41 ...

9.56 77.38 77.64

73.16 ...
73.87 73.95
74.03 ...
75.44
76.28

7.11
7.52
7.94
8.28
8.47
8.84
8.91 ..
9.76 9.78

7.29

8.44
8.46
8.74

Depolymerization. The molecular still (Fig. 1) used in these experiments was a
cylindrical vessel, 18 cm. high and 5 cm. in diameter, with a rather flat curved bottom
and an outlet to a high vacuum system. Into this vessel was fitted, by means of a large
ground joint at the top, another cylinder 3 cm. in diameter, the rounded bottom of which
was 3 cm. from the bottom of the larger vessel. This inner vessel constituted the con-
denser. A sample of a-anhydride was introduced into the outer vessel, the condenser
was placed in position and the apparatus was evacuated by means of a mercury diffusion
pump backed by a Hyvac oil pump. A trap cooled by solid carbon dioxide in acetone
was placed in the system. The still was heated by means of a metal bath.

Pimelic Anhydride. The a-anhydride was heated at 150. The 0-anhydride
collected on the water-cooled condenser as a dew which solidified to a pasty gel in about
one hour and to a hard wax overnight (7-anhydride). The saponification equivalent
of the wax was 70.7 [(C 7 HioO 3 ) n requires 71.0]. The experiment was repeated with
liquid air in the condenser and the heating bath at 125. The heating was continued
for six hours, after which the still was isolated from the vacuum system by closing a stop-
cock and put in communication with a small evacuated bulb containing aniline by open-
ing a stopcock (see Fig. 1). Aniline was distilled into the still until a layer of crystalline
aniline completely covered the distillate on the condenser. To accomplish this, it was
necessary to let the liquid air evaporate almost completely from the condenser. The

XIX. MANY-MEMBERED CYCLIC ANHYDRIDES 209

liquid air was finally allowed to evaporate completely. As soon as the condenser had
warmed up to a point where the aniline started to melt, the vacuum was relieved and the
condenser removed and immersed in aniline to a level slightly above the coating of
distillate. When the mixture had come to room temperature, it was treated with dilute
hydrochloric acid. Complete solution indicated the absence of any dianilide in the
reaction product. The solution was made alkaline and extracted three times with ether.
It was then concentrated and acidified to Congo Red. The precipitate was filtered off
and treated carefully with cold water to remove sodium chloride. The residue of mono-
anilide was recrystallized from water; m. p. 107-108. It did not depress the melting
point of an authentic sample of melting point 108-109. The absence of any dianilide
in the aniline reaction product demonstrates that the 0-anhydride was monomeric.

Suberic Anhydride. A sample of a-anhydride was heated at 160 in the molecular
still. The )3-anhydride collected as a crystalline solid 011 the water-cooled condenser
at the rate of about 0.1 g. a day. It melted at 55-57. This compound proved to be
the cyclic dimer of suberic anhydride.

Anal. Calcd. for Ci 6 H 2 4O 6 : C, 61.50; H, 7.75; mol. wt., 312.2; saponification
equiv., 78.0. Found: C, 61.57, 61.67; H, 7.72, 7.89; mol. wt. (cryoscopic in benzene),
343, 355, 346; saponification equiv., 77.5.

It reacted with aniline to form the monoanilide (m. p. 125-127) and the dianilide
(m. p. 186-187) of suberic acid, which were identified by mixed melting points. When
heated above its melting point it rapidly polymerized to the 7-anhydride, a waxy solid
melting at 65-68; mol. wt. observed in boiling benzene, 703, 718.

Azelaic Anhydride. The a-anhydride was heated in the molecular still at 150.
The /3-anhydride collected on the water-cooled condenser as a liquid and soon solidified
to a gel and then in about twelve hours to a wax (7-anhydride) . The observed saponifica-
tion equivalent of the wax was 84.8, and 84.5 [(CsHnOa).,. requires 85]. Two runs were
made in which the condenser was cooled with liquid air and the cold distillate was
treated with aniline as described under pimelic anhydride, but products free of dianilide
were not obtained. The amounts of the two anilides were determined quantitatively,
taking advantage of the insolubility of the dianilide in hot water and the insolubility of
the monoanilide in cold.

(1) Four hours: Dianilide, 0.18 g.; monoanilide, 0.35 g. A polymeric anhydride
yields 2 moles of monoanilide for one ot dianilide ; a monomeric anhydride yields only
monoanilide. The dianilide found indicates an amount of polyanhydride that could
produce only 0.28 g. of monoanilide. The excess, 0.07 g., must arise from monomeric
anhydride, and this amount corresponds with 11% of monomer in the total anhydride
sample.

(2) Six hours: Dianilide, 0.22 g., equivalent to 0.34 g. of monoanilide (from poly-
anhydride). Monoanilide found, 0.49 g. ; excess due to monomeric anhydride, 0.15 g.,
equivalent to 18% of monomer in the total anhydride sample.

Since there was no evidence of any dimer (which should be a crystalline solid) in the
distillate, the failure to demonstrate that the distillate was initially pure monomer was
probably due to the fact that the latter polymerized to a considerable extent before the
examination could be completed.

The distillate if removed when still liquid or pasty, had a spicy, aromatic odor,
This disappeared when the distillate solidified completely.

1,11-Undecanedioic Anhydride. The a-anhydride was heated in the molecular

210 POLYMERIZATION AND RING FORMATION

still with a water-cooled condenser. The /3-anhydride condensed first as a dew which
soon became pasty and finally after a few hours set to a hard wax, m. p. 85-88. Before
complete solidification it had a spicy odor. The saponification equivalent of the wax
was 98.8 [(CnHi 8 O 3 ) n requires 99.1]. The experiment was repeated using liquid air in
the condenser and the aniline technique described under pimelic anhydride. As the
reaction product was found to contain dianilide, the amounts of dianilide and mono-
anilide formed were estimated quantitatively. The product from the aniline reaction
was treated with a slight excess of warm dilute aqueous sodium hydroxide and the in-
soluble dianilide was filtered off and weighed (0.079 g.). The filtrate was acidified,
cooled and filtered (weight of crystalline precipitate 0.281 g.). This material was shown
by a determination of the neutralization equivalent to be pure monoanilide (calcd. 291;
found, 289). The dianilide found (0.079 g.) indicates an amount of polymeric anhydride
capable of yielding only 0.125 g. of monoanilide. The excess of the latter (0.156 g.)
must have come from monomeric anhydride, and calculation indicates 38% of mono-
mer ic anhydride in the mixture.

1,12-Dodecanedioic Anhydride. The a-anhydride was heated hi the molecular
still at 110. A beautifully crystalline distillate of melting point 76-78 collected on
the condenser at the rate of about 0. 1 g. in two days. This /3-anhydride was identified
as the cyclic dimeric anhydride of dodecanedioic acid.

Anal. Calcd. for C 2 4H 4 oOe: C, 67.86; H, 9.52; mol. wt., 424.4; saponification
equiv., 106.1. Found: C, 67.15, 67.54; H, 9.46, 9.60; mol. wt. (cryoscopic in ben-
zene), 496, 450; saponification equiv., 105.5.

It polymerized at the melting point and then melted again at 85-87. It reacted
with aniline in the usual way to yield monoanilide of m. p. 123 and dianilide of m. p.
169-170.

Brassylic Anhydride. The a-anhydride was heated in the molecular still for four
hours at 150 with the condenser cooled with solid carbon dioxide and acetone. The dis-
tillate of ^-anhydride was shown to be pure monomer by its reaction with aniline. The
product was completely soluble in dilute aqueous sodium hydroxide. The monoanilide
separated on acidification of the alkaline solution and was recrystallized from 50%
alcohol. It melted at 118.5-119.5 and showed no depression when mixed with an
authentic sample.

When the depolymerization of the anhydride was carried out using a water-cooled
condenser, the /3-anhydride collected as a liquid which changed in the course of a day to
a pasty mass and in the course of two days to a hard wax (saponification equivalent
found 112.7, calculated 113.1). The liquid was fragrant and aromatic and when al-
lowed to react with aniline shortly after distillation yielded only a small amount of
dianilide. The odor disappeared during the change to the wax.

1,14-Tetradecanedioic Anhydride. The a-anhydride was heated four and one-half
hours at 145-150 in the molecular still using solid carbon dioxide in acetone as the
refrigerant in the condenser. The distillate of /3-anhydride was shown to be pure mono-
mer by its reaction with aniline. The product was completely soluble in a large volume
of warm, dilute sodium hydroxide and consequently consisted of monoanilide free of
dianilide. The monoanilide separated on acidification of the alkaline solution and was
recrystallized from 60% alcohol. It melted at 124-125 and did not depress the melting
point of an authentic sample.

When the depolymerization of the anhydride was carried out using a water-cooled

XIX. MANY-MEMBERED CYCLIC ANHYDRIDES 211

condenser, the distillate, like that from brassylic anhydride, collected as a liquid which
changed in the course of a day to a pasty mass and in the course of two to a hard wax,
m. p. 88-91. The distillate, before it changed completely to the hard wax, possessed
a strong odor like that of musk. This odor was completely lost on saponification
(saponification equivalent found, 120.4, 119.1; calcd., 120.1) and on heating to 100
which brought about the change to the 'y-anhydride instantly.

1,18-Octadecanedioic Anhydride. The behavior of this a-anhydride was identical
in every respect with that of tetradecanedioic anhydride. The fresh liquid distillate like-
wise possessed the odor of musk but was rather fainter. It polymerized instantly on
being heated to 100, to a wax of m. p. 98-100 (7-anhydride), losing its odor in the
change. About 1 cc. of /3-anhydride was collected using a still adapted for the collection
of liquids. It crystallized on chilling or seeding and melted at 36-37 (saponification
equivalent found, 146.5; calcd. for CigHa^Oa, 148.1). After two weeks it had changed
to the wax of m. p. 98-100.

Summary

Data are presented on the anhydrides of dibasic acids COOH(CH 2 ) M -
COOH where n is 4, 5, 6, 7, 8, 9, 10, 11, 12 and 16. The anhydrides are
all linear polymers and when heated in a molecular still they are depoly-
merized yielding volatiles products (0-anhydrides). The latter are either
cyclic monomers or dimers depending upon the unit length, n + 3. The
compounds thus obtained are rings of 7, 8, IS (dimeric) 10, 22 (dimeric),
12, 26 (dimeric), 14, 15 and 19 atoms. The dimers are crystalline solids
which polymerize instantly when heated above their melting points.
The monomers are liquids or low melting solids which polymerize at lower
temperatures than the dimers. The monomers of 8, 10 and 12 atoms are
exceedingly unstable and polymerize rapidly even below room temperature.

Bibliography and Remarks

(1) An abstract of papers XIX, XX, XXI and XXII was presented at the Washington meeting?
of the American Chemical Society, March 28, 1933.

(2) Carothers, Journ. Am. Chem. Soc. t 51, 2548 (1929); Carothcrs, Chem. Rev., 8, 353 (1931);
Carothers and Hill, Journ. Am. Chem. Soc., 54 1559 (1932).

(3) Hill, ibid., 62, 4110 (1930).

(4) Hill and Carothers, ibid., 54, 1569 (1932).

(5) Cf. Carothers and Hill, Journ. Am. Chem. Soc., 64, 1557 (1932). The Washburn type of still
was used in this work.

(6) Carothers and Hill, ibid., 54, 1579 (1932).

212 POLYMERIZATION AND RING FORMATION

XX. Many-Membered Cyclic Esters*

Only one general method of producing cyclic esters is described in the
literature and this gives very low yields. In the present study a new and
very effective way for the preparation of cyclic esters has been found.
If linear polyesters are heated in vacuo and in the presence of a catalyst
just below their point of thermal destruction, the monomer or the dimer
cyclic ester is formed in good yield. Sodium or other ester interchange
catalysts are most effective. No fewer than thirty new cyclic esters are
described; in some cases (ethylene carbonate, decamethylene malonate, etc.)
they were exclusively monomeric, in others (pentamethylene carbonate) they
were only dimeric, in others again (tetramethylene carbonate, etc.) they
were mixtures between monomers and dimers. The ratio in which the two
forms are obtained is determined in part by the experimental conditions
and in part by the nature of the ester, especially by its unit length.

The only generally applicable method known for the synthesis of macro-
cyclic esters consists in oxidation of cyclic ketones with Caro's acid. This

""** (( r H>)n ? ==

I II (n = 12, 13, 14, 15, 16)

method, discovered by Baeyer and Villiger (1), has been applied by Ruzicka
and Stoll (2) to the synthesis of the lactones, II.

The reaction is far from clean-cut, while the requisite ketones (I) are
obtained in yields not exceeding 5% (3). The necessity for adopting this
devious and extravagant method for synthesizing macrocyclic lactones is
due to the fact, repeatedly illustrated in previous papers (4) of this series,
that bifunctional esterifications involving unit lengths greater than seven
yield almost exclusively linear polyesters instead of the desired cyclic
esters. When linear polyesters of unit length greater than seven are heated,
e. g., to 275, there is generally no evidence of depolymerization or distilla-
tion (5), but at temperatures somewhat above 300 destructive general de-
composition occurs, and from the complicated mixture of distillable prod-
ucts in four cases (6) true depolymerization products have been isolated.

* J. W. Hill and W. A. Carothers; Journ. Am. Chem. Soc., 55, 5031-39 (1933);
Communication No. 132 from the Experimental Station of E. I. du Pont de Nemours
and Co.

Received September 2, 1933. Published December 14, 1933,

XX. MANY-MEMBERED CYCLIC ESTERS 213

These cyclic esters however are not monomers but dimers, and the yields
are very poor (5% or less).

When linear polyesters of unit length greater than seven are heated
(e. g., to 200-250) under the conditions of molecular distillation (7), the
long chains couple to form still longer chains (8). The products (super-
polyesters) have molecular weights above 10,000 and are capable of being
drawn out into tough, pliable, highly oriented fibers (9).

It was recognized in advance that the conditions of the molecular still
at high temperature would be likely to favor either this type of coupling
or a depolymerization by ester interchange (10) :

Evidences of such depolymerization, at least in traces, were in fact fre-
quently observed in the formation of superpolyesters, and after methods
had been developed for the smooth depolymerization of poly anhydrides,
attention was again turned to the polyesters. The result was the develop-
ment of the method presently described by means of which it became pos-
sible to prepare with good yields and without unreasonable difficulty a
whole series of macrocyclic esters, both monomeric and dimeric. Briefly,
the method consists in heating linear polyesters in vacuo at a temperature
just below the point of thermal destruction under conditions that permit
any volatile product to be removed by distillation as fast as it is formed.

The success of the new method, practically, depends upon numerous fac-
tors not all of which have yet been completely isolated and defined; the
two most important however are the identity of the polyester and the use
of catalysts. On the first point it may be said that esters of carbonic acid
and of oxalic acid depolymerize more readily and smoothly than any others
yet examined. On the second point, ester interchange catalysts (e. g.,
sodium added as metal in the preparation of the initial polyester) are al-
most indispensable.

After some experimentation it was discovered that esters of carbonic acid
and of oxalic acid can be depolymerized merely by heating them (with cata-
lyst) under diminished pressure in an ordinary distilling flask. Most of
the experiments now reported were however made in a simplified molecular
still (described below) operated at pressures ranging from 0.1 to 2 or 3 mm.

The depolymerization of the polyester in favorable cases progresses
smoothly and fairly rapidly, the cyclic ester distilling to the condenser
and being collected as it is formed. The residual ester at the same time
progressively increases in viscosity (formation of superpolymer) ; if this

214 POLYMERIZATION AND RING FORMATION

increase in viscosity does not proceed too far, depolymerization still con-
tinues and is ultimately almost complete and quantitative. In many
cases, however, the residue, before complete depolymerization, becomes con-
verted to a gel; even at 250 it is completely immobile and resembles a
piece of porous, vulcanized, gum rubber. Depolymerization then practi-
cally ceases. This residue incidentally is extraordinarily resistant to attack
by solvents and chemical agents generally. It evidently results from a
polymerization which has progressed beyond the linear superpolymer
stage. It seems likely that some quantitatively minor side-reaction
(dehydrogenation at a venture) occurs permitting accessory polymeriza-
tion to form a three-dimensional macromolecule.

The product of the depolymerization is generally a mixture of cyclic
monomer and cyclic dimer although in many cases one of these forms pre-
dominates to the practical exclusion of the other. The ratio of the two
forms is in some cases quite sensitive to the experimental conditions
(temperature, pressure, shape of apparatus) but the data on this point
recorded in Table I were all obtained under closely similar conditions.

Preparation of Polyesters. Polymeric carbonates derived from the
glycols HO(CH 2 ) M OH where n is 3, 4, 6 and 10 and from diethylene glycol
have already been described (11). The new polymeric carbonates pre-
pared as intermediates in the present work include those derived from the
glycols HO(CH 2 ) M OH where n is 5, 7, 8, 9, 11, 12, 13, 14 and 18, and from
diethylene and triethylene glycols. The fact that these compounds are
lacking in crucial physical properties (e. g. t sharply characteristic melting
points) makes it unnecessary to describe them individually. They are all
insoluble in water and soluble in certain organic solvents such as chloro-
form. They separate from solvents in the form of soft powders (micro-
crystalline) . The molten polyesters are very viscous liquids which crystal-
lize on cooling in the form of small radiating clusters of microscopic needles
or spherulites and finally solidify to more or less hard, tough waxes. The
polymeric carbonate derived from triethylene glycol, however, failed to
solidify and was obtained only as a thick sirup.

The general method of preparation used for the new polycarbonates was the same
as that already described (11), but, on account of the boiling points involved, dibutyl
carbonate was found to be a more convenient source of the acid radical than diethyl car-
bonate. To the mixture of glycol and alkyl carbonate a small amount of sodium was
added; the mixture was heated at 170-220 until the distillation of alcohol ceased, and
removal of the alcohol was completed by continuing the heating for two hours or more
in vacuo. The residue was usually depolymerized without any purification. The
alkylene oxalates and malonates were prepared from the glycols and the diethyl esters of

a'J

888

^ ^

I H IQ O f^*

liii iiil ill!

l>C^IOOv> OOOOOCI O5 OO tO >O

- -i i-i O -OOO ^tHi-4*-*

o i

8
888

8^88
Q "^ 8 S

i CO O '
'

88
88

a Ss

to a <

Sa--
i!il

!!!

c* <M

t~ ^ ^

7 crT d-^ d-

S 3 33 S3 3

I I 1
8888

* o-s-s h g
s S o s S "
8 ? ? ? I 4 S

s lf

5
^
?

216 POLYMERIZATION AND RING FORMATION

the acids in the same manner as the carbonates. The other esters were prepared di-
rectly from the acids and the glycols.

Depolymerization. The apparatus used in the depolymerizations for which data are
listed in Table I consisted simply of a 250-cc. Pyrex suction filter flask. Through a
rubber stopper in the neck of the flask was inserted a test-tube about 1.5 cm. in di-
ameter. Its bottom was about 3 cm. distant from the bottom of the flask. The outside
of this test-tube served as the condenser and receiver; the inside was cooled with a
stream of tap water (in some cases a mixture of solid carbon dioxide and acetone) ; the
distillate collected on the outside. At intervals the test-tube was removed and the solid
or pasty distillate was scraped from it. When the distillate was sufficiently fluid to flow,
a small glass thimble was hung on the end of the test-tube to catch the drip.

The side-tube of the flask was connected to a vacuum line. This line was connected
to an oil pump and the pressure in the line was about O.I to 2 mm. No attempt was
made to secure an exceedingly high vacuum. The flask was immersed to a depth of
about 2 cm. in a metal bath which was kept at 210-240 by means of an electric heater.

Nature of the Depolymerization Products. For the sake of complete-
ness Table I includes the physical constants of ethylene and trimethylene
carbonates. The former, a 5-atom ring, is known only as the monomer;
the latter, which has a unit length of 6 atoms, can be obtained either as
monomer or polymer, but the polymer depolymerizes so rapidly and
smoothly on distillation that the use of a molecular still is quite unneces-
sary (11).

The rest of the compounds listed in Table I are obtained by the ordi-
nary methods of preparation only as linear polyesters ; and the polyesters
were all depolymerized in the apparatus and under the conditions de-
scribed above.

The behavior of tetramethylene carbonate (7-atom unit) was exceptional :
the distillate was found to contain a considerable amount of tetrahydro-
furan, which obviously might arise by the loss of carbon dioxide from the
structural unit. The chief product, however, was the dimer, a 14-mem-
bered ring which has already been described (11). No detectable amount
of monomeric ester was formed. The products from the next five mem-
bers of the series (unit lengths 8 to 12 atoms) also were chiefly dimeric.
Pentamethylene carbonate (S-atom unit) gave no detectable amount of
monomer. The presence of some monomer in the distillate from hexa-
methylene carbonate was inferred from the odor, but the attempt to
obtain isolable amounts of monomer by carrying out the depolymerization
in ordinary distillation equipment gave only indefinite unsaturated prod-
ucts. The odors of the distillates from heptamethylene, octamethylene
and nonamethylene carbonates also indicated the presence of monomer,
and in the last two of these cases monomer was actually isolated and id en-

XX. MANY-MEMBERED CYCLIC ESTERS 217

tified when the depolymerization was carried out in ordinary distillation
equipment.

The next member of the series, decamethylene carbonate, showed a sharp
change in the nature of the depolymerization products: considerable
amounts of the monomer were formed even in the molecular still. Undeca-
methylene carbonate (14-atom unit) behaved similarly. Beyond this
point the ratio was reversed: dodecamethylene carbonate (15-atom unit)
gave only a small amount of dimer, and with higher members of the
carbonate series the isolated products were exclusively monomeric.

The oxalates (decamethylene and undecamethylene) showed a greater
tendency to yield monomers than did the carbonates of the same unit
length. Decamethylene malonate also gave almost exclusively the
monomer.

The rest of the compounds listed in Table I, compared with the car-
bonates, exhibited a peculiar reluctance toward the formation of monomers.
Even ethylene tetradecanedioate yielded a 36-membered dimer and no
appreciable amount of the 18-membered monomer. Similarly co-hydroxy-
pentadecanoic acid yielded the 32-membered dimer instead of the 18-
membered monomer which has already been obtained by oxidation of the
corresponding cyclic ketone (12). It should be added that the depoly-
merization of all the esters referred to in this paragraph is especially slow
and difficult. Depolymerization of carbonates, oxalates, and even malo-
nates, proceeds much more rapidly and completely.

Effect of Conditions on the Ratio of Monomer to Dimer. The data on
the carbonates listed in Table I indicate that polyesters of this series on
depolymerization yield dimers almost exclusively when the unit length is
7 to 12 inclusive, both monomer and dimer when the unit length is 13
and 14, and monomers almost exclusively when the unit length is more than
14. It will be shown later that this characteristic change in the nature of
the products under a given set of conditions is probably due to stereo-
chemical factors inherently associated with the structure of the cyclic
esters. It is also true, however, that the ratio of monomer to dimer can
be controlled within certain limits by the experimental conditions. Thus
octamethylene and nonamethylene carbonates in the modified molecular
still yield almost exclusively the cyclic dimers, but when the depolymeriza-
tion is conducted in an ordinary distilling flask the product is almost en-
tirely the monomer. The distilling flask (unlike the molecular still)
permits refluxing, so that while most of the monomer escapes most of the
dimer is returned to the residue, where it is either cracked to monomer or
converted to higher polymer. Limitations of this method of experimental

218 POLYMERIZATION AND RING FORMATION

control are indicated by the fact that when it was applied to hexamethylene
carbonate no smooth depolymerization occurred and only indefinite un-
saturated products were obtained. Esters of dibasic acids above malonic
also depolymerize very slowly if at all in an ordinary distilling flask.

The depolymerization of polymeric decamethylene carbonate in an
ordinary distilling flask gave a much higher yield of monomer than dimer
and the examination of a large sample of crude product showed the pres-
ence of a by-product, decen-9-ol-l, CH 2 =CH(CH 2 )7CH 2 OH. The fol-
lowing data are typical.

The depolymerization was conducted in ordinary distilling flasks in batches of vari-
ous sizes up to 100 g. and at (bath) temperatures up to 290. The product was a mix-
ture of colorless liquid and crystalline material. The crystals (a mixture of dimeric
carbonate and decamethylene glycol) were removed by filtration, and the filtrate (2425
g.) was distilled, yielding 1276 g. of pure monomer, 400 g. of crystalline residue (chiefly
dimer) and 742 g. of low-boiling material. The latter by redistillation was separated
into four fractions of which the two largest were: A, b. p. 76-81 (2 mm.), 81%, and
B, b. p. 87-89 (2 mm.), 9%. The latter was practically pure monomer. Analytical
data for B indicated that it was a mixture of decamethylene carbonate (monomer)
and decenol (apparently a constant-boiling mixture). The ester was destroyed by
saponification with alcoholic sodium hydroxide, and the recovered alcohol was purified by
distillation.

Decen-9-ol-l. B. p. 85-86 (2 mm.); w 2 D 1.4480; 4 0.8446; M D calcd., 49.53;
M D found, 49.45.

Anal. Calcd. for C 9 H 18 O: C, 76.9; H, 12.8. Found: C, 76.51; H, 13.09.

Its structure was established by the fact that it was oxidized to azelaic acid by neu-
tral aqueous permanganate (in one experiment a small amount of suberic acid was found)
and hydrogenated to decanol-1 which was identified by comparison of the phenyl-
urethan with a known specimen (crys. from 80% alcohol, m. p. 61-62).

Decen-9-ol-l yields a phenylurethan of m. p. 49-50 (crys. from 80% alcohol).

Polymerization of Macrocylic Esters. Macrocyclic esters, unlike 6-
membered cyclic esters (11), show no tendency to polymerize sponta-
neously. They do, however, polymerize at elevated temperatures especially
in the presence of catalysts for ester-interchange, and the following ob-
servations are typical.

Small samples of monomeric decamethylene, dodecamethylene, tridecamethylene,
and tetradecamethylene carbonates were heated for twenty hours at 200 together with
parallel samples each containing a trace of potassium carbonate. The uncatalyzed
samples of tridecamethylene and tetradecamethylene carbonates showed a considerable
increase in viscosity; the decamethylene and dodecamethylene carbonates a much larger
increase. The catalyzed samples of decamethylene and dodecamethylene carbonates
were completely changed to solid, waxy polymers; while the catalyzed samples of the
other two esters were still semi-solid.

Samples of dimeric hexamethylene and decamethylene carbonates, both with and

XX. MANY-MEMBERED CYCLIC ESTERS 219

CO O O CO 00 00 rl r-f -l ^1

co -< c< >o eo r- co id 00 o>

EC ^ ^ ^ (

O M MCOrHCO^CO(N^W.

?iigsl^a

I K Ti S
J- O 3 2 .

Ib^lltfflMISaTl

1^ ! I giBB SiT 8t|

iiiiiiiui!!iiiipiiinii

S S 8 a

W4>W4WyOTr'Y&)'T*T

'HSSgSSjJS-aSSs^

f|fi-si i l'lllllllll|5|l3^;:;-:;i2

a!il!aa|!!!l!j3ssIS8g8SSSSS

IllllUlliillirislHimmil

tHWOO^^QQPPQfiHh'OH HQ^QPWWWWWWWHcX

220 POLYMERIZATION AND RING FORMATION

without traces of potassium carbonate, were heated for eight hours at 200. The
catalyzed samples changed to very viscous liquids which solidified to hard waxes; the
uncatalyzed samples were unchanged.

Summary

By heating linear polyesters with catalysts under certain conditions it
is possible in many cases to bring about a smooth depolymerization to the
corresponding monomeric and/or dimeric esters. This method makes it
possible for the first time to prepare macrocyclic esters in good yields.
Thirty new macrocyclic esters are described, mostly esters of dibasic acids.
The cyclic carbonates and oxalates are obtained most easily. The ratio
in which the two forms, monomer and dimer, are obtained is determined
in part by the experimental conditions and in part by the nature of the
ester especially by its unit length. Monomers of 7 to 1 2 atoms are espe-
cially difficult to obtain.

Bibliography and Remarks

(1) Baeyer and Villiger, Ber., 32, 3625 (1899).

(2) Ruzicka and Stoll, Ilelv. Chim. Acta, 11, 1159 (1928).

(3) Since this paper was prepared for publication Ziegler, Eberle and Ohlinger [Ann., 504, 94
(1933)] have described an ingenious method based on Ruggli's dilution principle for the preparation
of large cyclic ketones in good yields This principle has been referred to in previous papers of the
present series [e g , Journ. Am Chem. Sac., 51, 2551 (1929)] and we hope to describe its application
to the synthesis of cyclic esters in future papers.

(4) Carothers, Journ Am Chem. Soc , 51, 2548 (1929); Carothers and Arvin, ibid, 52, 711;
Carothers, Arvin and Dorough, ibid , 3292; Carothers, Chem. Reviews, 8, 353 (1931). Cf. Chuit and
Hausser, Helv. Chim Acta, 12, 463 (1929), Lycan and Adams, Journ. Am. Chem Soc., 51, 625, 3450
(1929).

(5) Compounds of molecular weight higher than 1000 cannot be distilled even under the highest
vacua, Carothers, Hill, Kirby and Jacobson, Journ Am. Chem. Soc., 62, 5279 (1931).

(6) The cases referred to are ethylene succinate, Tilitschejew, J. Russ Phys -Chem. Soc , 57, 143
(1925), Carothers and Dorough, Journ. Am Chem. Soc , 52, 711 (1930); tetramethylene carbonate,
Carothers and van Natta, ibid , 52, 314 (1930), trimethylene oxalate, Tilitschejew, /. Russ. Phys -
Chem. Soc , 58, 447 (1926); Carothers, Arvin and Dorough, Journ. Am. Chem. Soc , 52, 3292 (1930);
self-ester of hydroxydecanoic acid, Lycan and Adams, ibtd , 51, 625, 3450 (1929).

(7) Carothers and Hill, ibid , 54, 1557 (1932).

(8) Ibid., p. 1559.

(9) Ibid., p. 1579.

(10) Cf. the discussion of 6-membered cyclic esters and their polymers in Carothers, Dorough and
van Natta, ibid., 54, 761 (1932).

(11) Carothers and van Natta, Journ. Am. Chem Soc , 52, 314 (1930).

(12) Ruzicka and Stoll, Helv. Chim. Acta, 11, 1159 (1928).

XXI. PHYSICAL PROPERTIES OF MACROCYCLIC ESTERS, ETC. 221

XXI. Physical Properties of Macrocyclic Esters and
Anhydrides. New Types of Synthetic Musks*

L. Ruzicka has shown that some large ring lactones have very valuable
properties as perfume ingredients; this paper investigates the odor as a
function of the molecular structure of multi-membered rings.
Rings of 5 and 6 atoms smell like bitter almonds and menthol; from 9 to
12 they exhibit the odor of camphor; 14 and 15 give musk, 17 and 18 civet.
The molecular refractions and melting points of the carbonate and oxalate
series are determined.

Odors. The monomeric cyclic anhydrides and cyclic esters described
in the two preceding papers have highly characteristic odors. In particular
some of the higher members have odors closely resembling musk. This
observation is especially interesting because of the bearing of odor on the
general problem of macrocyclic compounds.

The essential principles of musk and civet are the macrocyclic ketones
I and II. Ruzicka's demonstration (1) of this fact was followed by the
discovery (2) that the lactones III and IV are odorous principles of angelica
oil and musk-seed oil. These materials are highly valued as perfume in-
gredients. In spite of the great difficulties involved in their synthesis, a
synthetic ketone ("Exaltone," cyclopentadecanone) and a lactone "Exalto-
lide") have been placed on the market as substitutes for the natural musks.

CH 3 CH (CH 2 ) 7V i O CH (CH 2 )g O

I || >C=0 (CH,)u ' " '

CH (CH 2 )7 X

1 - ic"

CH CH 2 CH (CHi)?' ' CO CH (CH 2 ) 5 CO

(CH 2 )i 2 CO

II III IV

The observations on the anhydrides and carbonates are assembled in
Table I, together with Ruzicka's data on the ketones and lactones. It will
be noted that in each of the four series, with the possible exception of the
ketones, odors are rather vague and indefinite until the number of atoms
in the ring reaches 9. A camphoraceous or minty note then appears, and
at about 13 atoms a woody or cedar-like quality. Beyond this point there
is a nuance of musk which however in the anhydride series does not be-

* J. W. Hill and W. H. Carothers; Journ. Am. Chem. Soc., 55, 5039-43 (1933);
Communication No. 133 from the Experimental Station of E. I. du Pont de Nemours
and Co.

Received September 2, 1933. Published December 19, 1933.

222 POLYMERIZATION AND RING FORMATION

come definite and pronounced until the 15-atom compound is reached.
At 15 or 16 atoms the musk quality approaches a maximum of fullness and
homogeneity. Tetradecamethylene carbonate (17 atoms) shows a remark-
able resemblance to "Exaltolide" (16-atom lactone). Beyond 18 or 19
atoms the odors in each series practically disappear.

The structural feature common to the four series is the presence of one
or more C=O groups as a member of a ring. It is rather extraordinary
that the manner in which the C=O group is linked in the ring should make
so little difference in the quality of the odor. On the other hand it appears
that any modification of the C=O group itself completely changes the odor.
Ruzicka, Schinz and Seidel report (3) that the alcohol corresponding to
civetone has a faint and only slightly characteristic odor.

The fact that the only requirement (there are no doubt other limits not
yet known) for musk-like odor is the presence of a C=O group in a ring of
approximately 15 atoms is further illustrated by the following compounds
not included in Table I (other physical properties of these compounds are
listed in Table I of Paper XX)

L( CH 2 CH 2 O ) 4 CO O-" tetraethylene carbonate, 14 atoms, fresh, faint, musk-like
L(CH 2 )io O CO CO OJ decamethylene oxalate, 14 atoms, fresh, musk-like

L -(CH 2 )n O CO CO O^ undecamethylene oxalate, 15 atoms, musk-like

i 1 decamethylene malonate, 15 atoms, faint, musk-

L (CH 2 )io O CO CH 2 CO O- 1 like

H 2 ) 8 CO O CH 2 CH 2 O-" ethylene sebacate, 14 atoms, musk-like

L-CO(CH 2 ) 9 CO O CH 2 CH 2 OJ ethylene undecanedioate, 15 atoms, musk-like

The first of these contains six oxygen atoms in the ring of which two are
directly connected to the carbonyl group; the second and third have two
adjacent carbonyls to which annular oxygens are attached, in the fourth a
methylene has been interposed between the two carbonyls, and in the
fifth and sixth eight and nine methylenes are inserted between the car-
bonyls. It would be of great interest to have information on similar com-
pounds in which the carbonyls are replaced by methylenes; unfortunately
macrocyclic ethers are as yet entirely unknown.

From the practical standpoint the cyclic anhydrides are of no interest
as odorous materials since they polymerize spontaneously on standing and
the odors then disappear. (This fact incidentally demonstrates that the
odors of the cyclic anhydrides are not due to any traces of the corresponding
cyclic ketones which might conceivably be present as impurities.) On the

XXI. PHYSICAL PROPERTIES OF MACROCYCLIC ESTERS, ETC. 223

a 'a, 'a
c/2 en en

O 3

SJS

224 POLYMERIZATION AND RING FORMATION

other hand the cyclic esters, with the exception of the oxalates, are quite
resistant to hydrolysis or polymerization and from the practical standpoint
they have the advantage over the previously described ketones and lac-
tones that they are formed in relatively high yields.

Molecular Refractions. Ruzicka and Stoll have shown that the
molecular refractions of macrocyclic lactones (4) and hydrocarbons (5) are
lower by about 0.6 unit than the calculated values. The macrocyclic
carbonates exhibit the same peculiar negative exaltation and to about the
same degree. Typical data are assembled in Table II.

E M D
-0.28

+ .15

- .41

- .43

- .71

- .54

- .67

- .63

- .83
+ .01
+ .42
+ .40
+ .22
4- .29

The last four compounds of the table appear to have a positive exalta-
tion, but the amounts of these materials available were quite small and
their purity was somewhat uncertain. The compounds for which negative
values are listed were for the most part purified by crystallization. No
significance can be attached to the fluctuations in these negative values
from one member of the series to the next since the molecular refractions
were not all taken at the same temperature.

Melting Points. The melting points of the known macrocyclic com-
pounds show little regularity in the nature of the change produced by in-
creasing molecular weight but in the range of 8 to 14 atoms there is generally
an oscillation from one member to the next (5). The cyclic carbonates
show a similar oscillation, the melting points and ring size for some of the
monomeric polymethylene carbonates being

11 12 13 14 15 16 17

23 35 11 41 12 25 22

TABLE II
MOLECULAR REFRACTIONS OF CYCLIC ESTERS

Atoms

Compound

in ring

A/D calcd.

MD found

Ethylene carbonate

17.15

16.87

Trimethylene carbonate

21.77

21.92

Octamethylene carbonate

44.86

44.45

Nonamethylene carbonate

49.48

49.05

Decamethylene carbonate

54.20

53.49

Undecamethylene carbonate

58.71

58.17

Dodecamethylene carbonate

63.33

62.66

Tridecamethylene carbonate

67.95

67.32

Tetradecamethylene carbonate

72.57

71.74

Octadecamethylene carbonate

91.04

91.05

Tetraethylene glycol carbonate

49.79

50.21

Decamethylene oxalate

58.72

59.12

Undecamethylene oxalate

63.34

63.56

Decamethylene malonate

63.34

63.63

XXII. FORMATION AND STABILITY OF LARGE RINGS 225

Summary

Macrocyclic esters and anhydrides have odors closely resembling those
of the ketones and lactones of the same ring size. The rings in the neigh-
borhood of fifteen atoms have musk-like odors. The molecular refractions
show a negative exaltation. Melting points in the carbonate series oscil-
late from one member to the next.

Bibliography and Remarks

(1) Ruzicka, Helv. Chtm. Ada, 9, 230, 716, 1008 (192G).

(2) Kerschbaum, Ber. t 60, 902 (1927); Ruzicka and vStoll, Heh, Chim. Acta, 11, 1159 (1928).

(3) Ruzicka, Schinz and vSeidel, Helv. Chim. Acta, 10, G95 (1927).

(4) Ruzicka and Stoll, Helv. Chim. Acta, 11, 1159 (1928).

(5) Ruzicka, Stoll, Huyser and Boekenoogen, ibid., 13, 1152 (1930).

XXII. Stereochemistry and Mechanism in the Formation and
Stability of Large Rings*

The first systematic investigation of large rings was carried out by Ruzicka.
His method gives low yields and does not allow a reliable insight into the
mechanism of the reaction. The new data] on polyesters and anhydrides
make it possible to discuss more thoroughly the formation of large rings.
This is done in the present paper.

First the thermalysis of thorium octadecanedioate was investigated as car-
ried out by Ruzicka. During this reaction first a linear polymer a
polyketone is formed; if this is heated in the molecular still, it cracks
and gives appreciable amounts of the monocyclic ketone. As was
found** in the case of the polyesters and polyanhydrides the formation
of the macrocyclic product proceeds by way of the linear polymer. The
cracking of the chains may be monomolecular or birnolecular.
To explain the easiness of ring closure, the density of large ring com-
pounds and the stability of the rings, a thorough discussion of the sterical
conditions and interferences is made. If a radius of 0.77 A. is assumed
for the carbon atom and a covering sphere with a radius of 1.1 A. for the
hydrogen atoms attached to the carbons, models representing to an appre-
ciable degree the actual conditions are obtained. There is a minimum of

* W. H. Carothers and J. W. Hill; Journ. Am. Chem. Soc., 55, 5043-52 (1933);
Communication No. 134 from the Experimental Station of E. I. du Pont de Nemours
and Co.

Received September 2, 1933. Published December 19, 1933.

t Compare paper number XII on page 156.

** Compare paper number XIII on pages 165 to 168.

226 POLYMERIZATION AND RING FORMATION

probability for the formation of rings between 10 and 17 carbon atoms
owing to the fact that in such molecules a considerable strain is effected
through the compression of the hydrogen domains inside the ring; if the
rings get larger this stress decreases and the ease of formation increases.

A great deal of discussion has been devoted to stereochemistry and
mechanism in the formation of rings, but on account of the limited range of
experimental facts available, fruitful discussion has been largely restricted
to rings containing fewer than 8 atoms. The only important series
exemplifying closure of long chains hitherto has been the macrocyclic
ketones which Ruzicka obtained by heating salts of the higher dibasic
acids. As a basis for theoretical inferences this general reaction suffers
from two disadvantages. The macrocyclic ketones are formed in such
small yields (0.1 to 5%) that they can hardly be regarded as major reac-
tion products, and the nature of the reaction (thermal destruction at high
temperatures) is such that its mechanism is inherently obscure. Data on
the polyesters and anhydrides now provide the possibility for a much
clearer insight into the mechanism of the formation of large rings.

Following the demonstration of the stable existence of large carbon
rings, Baeyer's theory of negative strain has been generally abandoned and
replaced by the idea of Sachse and Mohr (1) that large rings may exist in
strainless and non-planar forms. A carbon chain constructed from con-
ventional wire tetrahedra in such a way as to allow free rotation about
each single bond clearly shows that the chain can assume a great multi-
plicity of shapes. If the number of atoms is 5 or 6, the model is readily
rotated so that the ends collide once in each complete rotation. If the
number is more than 6 the ends can be brought together arbitrarily
without any difficulty to produce a non-planar ring, which is quite flexible
and can be bent into a variety of shapes without any appreciable strain.
Considering the chain model as a representation of the bifunctional mole-
cule, it is evident why rings of 5 and 6 atoms are readily formed and stable.
Larger rings should be equally stable, but it is obvious that their formation
may present some difficulties. The chain can assume a great multiplicity
of shapes; the particular configurations requisite for ring closure are rela-
tively few. If then the molecule is placed under conditions where mutual
reaction of its terminal groups can occur, the probability of iw/ennolecular
reaction is very much greater than that of iw/ramolecular reaction. Hence,
as has been demonstrated in previous papers of this series, reactions that
might conceivably yield large rings from open chains almost invariably
yield linear polymers instead. It is necessary again to emphasize this point

XXII. FORMATION AND STABILITY OF LARGE RINGS 227

since even some of the most recent writings on the subject of forming large
rings still convey the impression that attempts to close long chains do not
result in any clear-cut reaction at all. In fact, in most cases, reaction
occurs perfectly smoothly without any difficulty but the reaction is inter-
molecular not intramolecular.

What means are available for controlling this situation? Obviously
nothing much can be expected from any modification of the nature of the
terminal groups. Even if these groups are of such a nature (e. g., NH 2
and COOH) as to exercise a strong attraction for each other, this pre-
existing attraction may itself be intermolecular rather than intramo-
lecular (2).

There is in fact no method known for controlling experimentally the
shape that molecules assume. The possibility of such control may perhaps
exist in the use of surface forces, but nothing is known about this matter.
There are, however, two possible methods of controlling the result of the
reaction. (1) High dilution will increase the relative probability of intra-
molecular reaction (3). (2) If a series of mutually dependent and quan-
titatively reversible reactions is involved, constant removal of any traces
of cyclic product will cause a displacement of the equilibrium with the
ultimate conversion of the entire sample into the cyclic product. This is
the principle involved in the synthesis of the cyclic esters and anhydrides
described above.

In speculations devoted to Ruzicka's ketone synthesis it has been sug-
gested (2, 4) that the peculiarity of the reaction which makes possible this
very exceptional closure of large rings lies in the ability of the metal ion
(e. g., thorium) to bring the ends of the chain into a position of close intra-
molecular approach. But no reason is offered to explain why the metal
ion should have such a peculiar effect; and the force of this suggestion is
moreover considerably weakened by the claim (5) that macrocyclic ketones
are obtained by thermal destruction not only from salts of dibasic acids
but also from the acids themselves and their anhydrides. It appears
to have been taken for granted in these speculations that reaction is intra-
molecular and leads directly to the cyclic ketone. This is a point that
appears to be open to test and the following experiment was accordingly
made.

Thermolysis of Thorium Octadecanedioate in Dixylylethane. Twenty-five grams
of thorium octadecanedioate, prepared by the method of Ruzicka, and 150 cc. of un-
syrnmetrical dixylylethane were placed in a 300-cc. flask fitted with a stirrer and ther-
mometer. The mixture was heated by means of a metal bath to 325 . In the course
of ten minutes the suspended solid became gummy and attached itself to the stirrer.

228 POLYMKR^ATION AND RING FORMATION

At the end of fifty minutes the mixture had become homogeneous but very viscous. At
the end of two hours, heating was stopped. On cooling, the melt solidified to a slightly
elastic gel. Some solid material (apparently unchanged salt) around the top of the
flask was discarded. The gel was continuously extracted with ether for eighteen hours,
whereupon it disintegrated to a light gray powder. It was further extracted for four
hours with benzene, dried, heated with concentrated hydrochloric acid for three hours,
separated, ground in a mortar, extracted for three hours with hot hydrochloric acid,
then twice with hot alcohol and once with ether. The dried residue (10 g.) was an almost
pure white powder. It was practically free of ash and melted at 126-128; soluble in
hot toluene, butyl alcohol and acetylene tetrachloride.

Anal. Found: C, 78.22, 78.89; H, 12.29, 12.41.

When triturated with aqueous sodium carbonate and thoroughly washed and dried
it was converted to a salt (Na found, 1.16, 1.36%). The analytical data are consistent
with the formula HOOC[(CH 2 )i 6 CO] 7 OH, the calculated values for C and H being 78.9
and 12.4% while a monosodium salt would contain 1.25% Na.

A sample of this material was placed in a molecular still and heated by a metal bath
at 300-305. At the end of one day a film of white solid distillate had collected on the
condenser. It had a very pronounced fragrant musky odor.

Our interpretation of this experiment is that the thorium salt of the acid,
as in other bifunctional reactions involving long chains, breaks down with
the formation of a linear polymer

. . . RCO RCO RCO RCO RCO . . .

The resulting polyketone comprises the solid product remaining from the
extraction. When the polyketone is heated in the molecular still, it cracks
and appreciable amounts of the monomeric cyclic ketone (cyclohepta-
decanone) are produced; the characteristic musk-like odor therefore first
makes its appearance at this point in the experiment.

The facts recited above place the macrocyclic ketones, esters (including
lactones) and anhydrides on a common basis for discussion so far as the
stereochemistry and mechanism of formation are concerned.

The a-polyanhydrides for example may be represented by the formula
. . .CORCOO CORCOO CORCOO CORCOO . . . . They are ap-
parently open chains, and contain no more than traces of cyclic anhydrides
of low molecular weight. The possibility of converting the polymeric
a-anhydrides into ^-anhydrides of low molecular weight depends in the
first place on the very great reactivity of carboxylic anhydrides and the
fact that they are capable of reacting with themselves. A mixed anhy-
dride such as that derived from acetic and butyric acids will, at least at
elevated temperatures, pass into an equilibrium mixture containing a con-
siderable proportion of the two simple anhydrides (6). The process is
analogous to ester interchange and, like that, doubtless proceeds through

XXII. FORMATION AND STABILITY OF LARGE RINGS 229

the formation of an addition product involving the ether oxygen of one
molecule and the carbonyl carbon of another (A) . A linear polyanhydride

// /& AAAAAAlAAA

'/ '/ . . f\ /V f\ f\. "/V 'f\ | f\~ f\ f\ . . .

R C O C R' A l !

A ! '!

/It/ ... A- A A A A A-I-A-- A- -A- - . .

R^_C O CR' i

(A) (B)

is not a chemical individual. It is a mixture of long chains of slightly dif-
fering lengths each of which bears a series of anhydride linkages. It pre-
sents a very complicated series of possibilities of reacting with itself. Ad-
jacent chains can react with themselves to produce simultaneously longer
and shorter chains (B). Reaction in a similar manner at two points will
yield large rings and these may mutually coalesce by the same mechanism
to produce still larger rings. Intramolecular reaction may result in the
formation of cyclic monomers and dimers, etc.

The potentialities of the situation are sufficiently complex to suggest
that at equilibrium the number of entities involved will be limited only by
the magnitude of the sample. No doubt a condition of genuine equilibrium
is impossible to achieve; nevertheless, experiment shows that at elevated
temperatures quite a considerable series of transformations occurs, and
it leads chiefly to the formation of very large molecules. At the same
time appreciable amounts of smaller cyclic fragments (/3-anhydride) are
produced. The actual concentration of these /3-anhydrides present in the
reaction mixture at any time must be small, since the /3-anhydrides poly-
merize almost instantly at the temperature involved; and their rate of
formation cannot be exceedingly great because even at elevated tempera-
tures where the speed and amplitude of molecular vibrations are greatly
increased, the relative probability of the configurations necessary for the
formation of such cyclic compounds must be rather low. However, in
the molecular still a mechanism is provided for removing and isolating the
^-anhydrides as fast as they are produced. The equilibrium is thus con-
tinuously displaced and the entire specimen is finally converted into the
^-anhydride.

A precisely similar mechanism is unquestionably involved in the forma-
tion of macrocyclic esters from polyesters and it is significant in this con-

230 POLYMERIZATION AND RING FORMATION

nection that smooth transformation requires the presence of ester-inter-
change catalysts.

The ketones obviously present a more difficult case. No mechanism cor-
responding to ester interchange exists for the smooth rupture of the link-
ages joining the structural units in a chain of the type . . . CO R
CO R CO R . . . ; nor would one expect a macrocyclic ketone to
polymerize with the formation of the corresponding polyketone. Since the
reactions involved are not strictly reversible and there are many side reac-
tions, the formation of cyclic ketones offers scarcely any possibilities for
rational and deliberate control. The yields are therefore generally only a
small fraction of those obtained with the cyclic carbonates, oxalates and
anhydrides.

Steric Interferences. The outline presented above is incomplete since
the possible influence of atoms attached to the carbon chain has been ig-
nored. Such atoms are capable of acting as obstacles to ring closure, and
they may also introduce strains into rings. For the series under considera-
tion hydrogen is the most numerous and important peripheral atom. The
iiitermiclear C-H distance is known from spectroscopic data to be 1.08
A., and since the (aliphatic) carbon radius is 0.77 A., the distance from the
center of the hydrogen to the surface of the carbon is 0.31 A. This indi-
cates an atomic diameter of 0.62 A. for hydrogen attached to carbon.
But hydrogens not mutually joined will be expected to exercise a mutual
repulsion preventing close approach, and data on the densities of hydro-
carbons, collision areas, etc., show that the average distance between the
centers of hydrogens belonging to separate molecules is always greater
than 0.62 A. The combined hydrogen atom must therefore be assigned,
in addition to its internal radius of 0.31 A., a larger external radius, which
defines the closest average approach of other atoms (7). This external
radius will vary with the compound and the conditions; it will decrease
with increasing temperatures and will be larger in a crystal than a gas.
The most elaborate experiments and speculations on the external radius of
hydrogen are those of Mack (8), who in different cases uses values ranging
from 0.49 to 1.26 A. Stoll and Stoll-Comte (9) had, however, already
made specific application of the external domain of hydrogen atoms in
connection with macrocyclic hydrocarbons. They pointed out that the
cyclic paraffins exhibit two anomalies. (1) When density is plotted against
number of CH2 groups a maximum appears in the range of 10 to 17 atoms,
and (2) it is within this range that the yields of the cyclic ketones fall to an
exceedingly low minimum. The explanation which they offer is very
briefly this. Consideration of the densities of cyclic and open-chain hydro-

XXII. FORMATION AND STABILITY OF LARGE RINGvS

231

carbons indicates a domain for the CH 2 group that can be accounted for by
representing the hydrogen as spheres of diameter 2.3 A. (compared with
1.54 A. for carbon). Rings of 5, 6 or 7 members therefore consist of an
approximately flat cyclic chain of carbon atoms with a shell of much larger
hydrogen atoms around the periphery. In rings of 8 to 15 atoms the
geometrical limitations imposed by the valence angles of the carbon atoms
force some of the hydrogens toward the center of the ring. The space
available is not sufficient to receive them; the "domain" of the hydrogens
(or of the CH2 group) is therefore reduced or compressed. The resulting
strain explains the minimum in the
yield of the cyclic ketones, and the com-
pression explains the maximum in the
density curve of the hydrocarbons.
The turning in of hydrogens toward
the center of the ring is illustrated for
cyclononane by Fig. 1 in which the di-
ameter of the hydrogens (1.18 A.) is
much smaller than that assumed by
Stoll and Stoll-Comte (2.3 A.). This
configuration is quite rigid and it is
evident that if the hydrogens were twice
as large, the space available would not
be adequate to receive them. As the
size of the ring increases the number
of hydrogens forced toward the center
increases; above 15 atoms half of them
are forced toward the inside of the
ring, but there is sufficient space to
accommodate them and no strains are
developed.

This idea of steric interferences due to "external" radii is capable of
many interesting applications to the cyclic esters and anhydrides. It is
unfortunately for this purpose impossible to assign any exact value to the
external radius of hydrogen. Results obtained in the diphenyl problem
(10) suggest that the value implied in a spherical hydrogen of 2.3 A. is
probably much too large; however, for immediate qualitative purposes the
exact value adopted is not particularly important. The following discus-
sion is based on observation of models in which the hydrogen is represented
as a sphere of radius 2.2 A.

When a zig-zag hydrocarbon chain is constructed with such large hydro-

Fig. 1. Cyclononane (carbon, d =
1.54 A.; hydrogen, 1.1). This model
illustrates the cramped nature of the
ring and the turning of hydrogen
toward the center.

232

POLYMERIZATION AND RING FORMATION

1.35
o' 1.30
1.25

rt 1.20

8 1.15

2 1.10

x> x x

g 1.05

XX * N *^

Q 0.95

^"^^^^

0.90

gens, it is immediately evident that the mobility of the chain is greatly
reduced. The hydrogens interfere with many rotational movements that
would present no difficulty if the hydrogens were absent. The interfer-
ences toward ring closure become exaggerated as the length of the chain
is increased beyond 6 or 7 atoms. For a cycloparaffin of 9 atoms the inter-
ferences are so serious that the model is not constructible.

If an oxygen atom is inserted in the chain, the flexibility of the model is
considerably increased, since this insertion is practically equivalent to the
removal of two of the interfering hydrogens (11). Actually also oxygen
probably presents less resistance to the deflection of its valence angles,
although this is not illustrated by the models. In the anhydrides and esters

under consideration each struc-
tural unit contains in its chain
at least one oxygen and at least
one carbonyl carbon (which also
bears no hydrogens). The re-
sult is a great increase in the
constructibility of the models.
Construction of models of the
cyclic hydrocarbons in the range
of 9 to 15 atoms requires either
a compression of the hydrogen
spheres or a considerable deflec-
tion of the angles. In the car-
bonates and anhydrides the en-
tire series of monomers from the
5-atom ring up can be con-
structed from spherical wooden
atoms without more than very

slight deflection of the angles. From this one may infer that no maxi-
mum should appear in the curve relating density to ring-size of the car-
bonates. The experimental curve is shown in Fig. 2. Unfortunately data
are lacking in the range of 7 to 10 atoms, but the nature of the curve
makes it very improbable that any maximum exists.

Although it appears from the models that no compression of the hydro-
gen domains is involved in any of the cyclic ester or anhydride molecules,
the manipulation of the models shows that there is a great difference in
ease of ring closure depending upon the length of the chain. In the range of
8 to 14 atoms the interferences are so numerous that many trials and errors
must occur before the particular configuration that permits ring closure is

6 8 10 12 14 16 18 20

Number of atoms in ring.
Fig. 2. Densities of monomeric cyclic
polymethylene carbonates at 20: O, values
determined at 20; El, values obtained by ex-
trapolation via molecular refractions from de-
terminations of density made at 25; A, values
obtained by extrapolation via molecular refrac-
tions from determinations made at 50.

XXII. FORMATION AND STABILITY OF LARGE RINGS 233

found; and the ring when closed is then very rigid. In the carbonate
series with the 15-membered ring the model acquires considerable flexi-
bility.

The greater ease in constructibility of rings of 15 atoms or more ex-
plains why in the preparation of carbonates in the range of 7 to 12 atoms
the products are almost exclusively the dimeric forms. At equilibrium
the concentration of dimer will be large compared with the concentration
of monomer. But as the ring size increases not only does the ease of
monomer formation approach (and ultimately exceed) that of the dimer,
but the latter becomes relatively more and more difficult to remove by
evaporation. The ring-size at which the two forms begin to appear in
equal amounts will depend partly upon the experimental conditions; but
it will also be controlled to a certain extent by the nature of the ring. It is
interesting in this connection to compare the three 14-membered rings
undecamethylene carbonate, tetraethylene glycol carbonate and deca-
methylene oxalate. The first has two annular oxygens and one carbonyl,
the second two oxygens and two carbonyls and the third five oxygens and
one carbonyl. The first compound under the conditions used gave a mix-
ture of monomer and dimer, the last two gave exclusively monomer.

The alkylene esters of dibasic acids above malonic yield chiefly dimers
even when the unit length is as great as 18. The models do not furnish any
very clear explanation of this fact although it appears that some advantage
in ease of ring formation may be expected if the oxygens and carbonyls
are all adjacent (as in oxalic esters).

The cyclic anhydrides present the same reluctance toward the formation
of rings of intermediate size as do the esters, but they show some peculiari-
ties. The products appear to be invariably either monomer or dimer, not
mixtures of the two. The range of dimer formation is from 9 to 13 atoms
(unit length). Very peculiar is the fact that the two even membered
compounds (10 and 12 atoms) in this range appear to yield exceedingly
unstable monomers rather than dimers. The result is the alternating effect

shown graphically in Table I (p. 20o ). The manipulation of models

in this range furnished the impression that even-membered rings were more
readily constructed than odd, but the geometrical peculiarity responsible
for this effect was not identified. All of the cyclic anhydrides are exceed-
ingly unstable at the high temperatures used in their formation, and the
nature of the /3-anhydrides produced might therefore be expected to be very
sensitive to very slight differences in the ease of constructibility.

Ring Stability. Ruzicka showed (12) that his macrocyclic paraffins
were not destroyed by the action of phosphorus and hydriodic acid at 250

234 POLYMERIZATION AND RING FORMATION

and his cyclic ketones resisted the action of fuming hydrochloric acid at
180-190. This was an indication that large rings are no less stable than
those of 5 or 6 atoms and a refutation of the idea that large rings are
strained. It appears, however, in view of the results obtained with the
cyclic esters and anhydrides, that this conclusion requires some revision.
The higher cyclic anhydrides all polymerize very readily. In this respect
the large rings are very unstable as compared with those of 5 to 6 atoms,
which do not polymerize at all. The instability rises to a maximum with
the rings of 10 to 12 atoms, which polymerize rapidly even at temperatures
below 0. At elevated temperatures (e. g., 200) the macrocyclic esters
also polymerize. It is therefore very difficult to accept the conclusion that
large rings are entirely free from strain. Our interpretation of the facts is
as follows.

Rings of 3 or 4 atoms have very large strains owing to the necessarily
large deflection of the annular valences. Rings of 5 atoms are practically
free of strain. Most rings of 6 atoms are strained. It is true that cyclo-
hexane can be represented as existing in two strainless forms (cis and trans)
but since it has not been possible to isolate two cyclohexanes, it seems
likely that the two forms are in dynamic equilibrium. On each conversion
of one form into the other the ring must pass through a nearly planar posi-
tion of strain. The existence of such strain is indicated by the fact that all
simple 6-membered cyclic esters polymerize very readily. (Glutaric an-
hydride does not polymerize but the oxygen valences in this case may per-
mit sufficient deflection to avoid strain.) Larger rings are all strained.
In the range of 7 to 14 atoms where the models are very rigid the strain
may be pictured as due to the mutual repulsion of non-linked peripheral
atoms that are crowded against one another. In larger, more flexible
rings the vibrations due to thermal agitation constantly present the possi-
bility of introducing momentary strains. This effect is similar to that
pictured for the cyclohexane ring but is probably less pronounced. At
any rate, 6-membered cyclic esters polymerize more readily than macro -
cyclic esters.

The ease of polymerization of the macrocyclic anhydrides is unquestion-
ably associated with the extraordinary reactivity of anhydrides generally.
A facile mechanism for the rupture of the ring is provided by the nature
of the anhydride linkage and hence very slight strains or distortions, which
may arise merely through small interferences among substituent atoms,
need not be tolerated. The macrocyclic esters polymerize by a similar
mechanism (i. e., by ester interchange) but less readily because they are
less reactive. The apparent high stability of the macrocyclic hydrocar-

XXIII. 6-CAPROLACTONE AND ITS POLYMERS 235

bons and ketones is due to the fact that they present no point of easy
chemical attack; if a sufficiently delicate chemical probe were available
they would probably prove to be somewhat less stable than those of 5
atoms.

Summary

It is shown that the formation of macrocyclic ketones from salts of the
dibasic acids probably involves first a linear polyketone which is subse-
quently cracked or decomposed. The ketones thus follow a course al-
ready established for esters and anhydrides. The characteristic analogies
and differences in the three series can be explained by taking into account
the nature of the reactions involved and the steric effects of peripheral
atoms. Rings of more than 5 atoms cannot be regarded as entirely strain-
less. The probable nature of the strains in large rings is indicated.

Bibliography and Remarks

(1) Mohr, /. prakt. Chem., 98, 348 (1918).

(2) The idea that such attractions may favor intramolecular reaction has been suggested by Mills,
"Proceedings, Fourth International Solvay Conference," 1931, p. 20.

(3) Cf. Ruggli, Ann., 392, 92 (1912).

(4) Ruzicka, Stoll and Schinz, Helv. Chint. Acta, 11, 670 (1928).

(5) Ruzicka, Brugger, Seidel and Schinz, ibid., 11, 496 (1928).

(6) Cf. Autenrieth, Ber., 34, 168(1901).

(7) Cf. the excellent review by Sidgwick in "Annual Reports of the Progress of Chemistry for 1932."

(8) Melaven and Mack, Journ. Am. Chem. Soc., 54, 888 (1932); Sperry and Mack, ibid. t 904;
Mack, ibid., 2141.

(9) Stoll and Stoll-Comte, Helv. Chim. Acta, 13, 1185 (1930).

(10) Cf. Adams and Yuan, Chem. Reviews, 12, 261 (1933).

(11) This effect is probably considerably exaggerated in the models since the oxygens are repre-
sented only by their internal diameters.

(12) Ruzicka, Brugger, Pfeiffer, Schinz and Stoll, Helv. Chim. Acta, 9, 499 (1926).

XXIII.* -Caprolactone and Its Polymers

Bifunctional esterifications yield either cyclic monomers or linear poly-
esters.^ Choice between these possibilities is controlled by

(a) the unit length of the reactant,

(b) the nature of the unit,

* F. J. van Natta, J. W. Hill and W. H. Carothers; Journ. Am. Chem. Soc., 56,
455-7 (1934) ; Communication No. 118 from the Experimental Station of E. I. du Pont
de Nemours and Co.

Received October 20, 1933.

f Compare paper No. XII on page 156.

236 POLYMERIZATION AND RING FORMATION

In the present paper these conditions are studied during the self esteri-
fication of i-hydroxycaproic acid. Under certain experimental condi-
tions the principal primary product is the monomeric lactone; a trace of
dimeric ester is formed, but little if any of the high polymers.
Under the action of heat (about 150) the lactone polymerizes to a linear
polymer with a molecular weight around 4000.

Bifunctional esterifications (1) generally yield either cyclic monomers
or linear polyesters. Choice between these possibilities is controlled by
(a) the unit length of the reactant, (b) the nature of the unit and (c) the
experimental conditions (especially dilution). Factor (a) is generally by
far the most important and its effect has been illustrated in previous
papers. Unit lengths of 7 and 8 however constitute transition cases and
factor (b) may here become the controlling one. It is well known that
substitution (e. g., by methyl) favors ring closure but otherwise the effect
of variations in the nature of the unit are not easy to foresee. The sim-
plest possible structural situation for self-esterification is found in the o>-
hydroxy acids, HO(CH 2 ) W _ 2 COOH, and information concerning them is
available for all values of n from 3 to 22 (2) excepting 7 and 8. These
acids have never been isolated or their self-esterification studied (3).

The experiments presently reported were prompted by the considerations outlined
above and by the fact that several grams of -hydroxycaproic acid had become avail-
able as a by-product in the preparation of hexamethylene glycol by the reduction of
diethyl adipate. The acid could neither be distilled nor crystallized, and it was therefore
impossible to isolate it as such in a state of purity. It was however found possible,
as described in B (below), to isolate an oil composed essentially of the acid (80%) and
its lactone (20%). This mixture was dehydrated by heating it in a distilling flask at
150 to 210. The residue was completely volatile, and, when purified by redistillation,
it was found to consist of the lactone of -hydroxycaproic acid, a colorless liquid having
a pleasant, spicy odor. It crystallized when strongly cooled and melted at about 5.
The yield of purified lactone was about 63% of the theoretical. The only other product
of the reaction consisted of a very small amount (1%) of a volatile crystalline solid melt-
ing at 111 to 113. This was identified as the dimeric, cyclic self-ester of e-hydroxy-
caproic acid, a 14-membered ring.

The monomeric lactone showed no appreciable tendency to polymerize spontane-
ously when allowed to stand at the ordinary laboratory conditions; but when heated at
150 in a sealed tube it gradually became more viscous and, after twelve hours, when
cooled to room temperature, it solidified to an opaque mass. When a trace of potassium
carbonate was added to the lactone the same result was obtained in five hours at 150.
The product crystallized from alcohol as a white powder melting at 53 to 55. It was
very soluble in ethyl acetate, acetone and benzene but only slightly soluble in alcohol
or ether. The analytical composition and chemical behavior of this material demon-

XXIII. e-CAPROLACTONE AND IT V S POLYMERS 237

strated that it was a linear polyester, and molecular weight determinations in freezing
benzene gave values about 4000.

Anal. Calcd. for HO [ (CH 2 )6 CO O ] 35 H: C, 62.83; H, 8.84; mol.
wt., 4008. Found: C, 62.15; H, 8.86; mol. wt., 3660, 4300.

All three of the above esters (lactone, dimer and polyester) were actually deriva-
tives of e-hydroxycaproic acid, and no shift of the hydroxyl oxygen was involved in their
formation (4), since they all yielded the same hydrazide when treated with hydrazine
hydrate. This hydrazide was also obtained from the ethyl ester described in C.

To test its susceptibility to depolymerization the polyester was placed in a molecular
still and heated by a bath at 250 and at a pressure of 1 to 2 mm. for ninety hours (5).
During this time only a very small amount of distillate was collected a viscous, dark-
colored oil containing a few minute crystals of the cyclic dimer. The polymeric resi-
due was darker in color but otherwise appeared to be unchanged.

Conclusions. Under the conditions described, the principal primary
product of the self-esterification of -hydroxycaproic acid is the monomeric
lactone; a trace of the dimeric ester is formed at the same time but little
if any of the higher polyester. Like 6-membered cyclic esters, the lactone
of -hydroxycaproic acid is polymerized by the action of heat. (Appar-
ently it polymerizes only slightly less readily than 5-valerolactone.) The
polyester thus obtained can be depolymerized only with great difficulty
under the action of heat. In this respect it differs from the polyesters
that result from 6-membered cyclic esters, and resembles polyesters de-
rived from the higher co-hydroxy acids.

The following comparisons with other compounds having 7-atom units
are also of interest. Tetramethylene carbonate and trimethylene oxalate
are obtained only in the form of linear polyesters. These are depolymer-
ized with great difficulty and the only products identified are the cyclic
dimers. The lactone of 3,7-dimethyl-6-hydroxyoctanoic acid is reported
(6) to exist in two forms, a liquid and a solid. The authors suggest that
these are stereoisomers, but no molecular weight determinations are re-
corded, and it seems more likely that they represent a monomeric and a
dimeric form. Another e-lactone results from tetrahydrocarvone by oxi-
dation with Caro's acid (7). It is hydrolyzed to the acid, HO CH-
(CH 3 ) (CH 2 ) 2 -~CH(C 3 H7) CH 2 COOH, from which the lactone is
regenerated by the action of heat. A simple lactone is also obtained from
e-hydroxyoctanoic acid by direct distillation (5). When -aminocaproic
acid is heated it is partly converted to the monomeric lactam (about 30%)
and partly to polyaniide (about 70%). These are not directly intercon-
vertible.

Preparation of -Hydroxycaproic Acid. The acid was isolated from the product
obtained by reducing a large amount of ethyl adipate. After the hexamethylene glycol

238 POLYMERIZATION AND RING FORMATION

had been removed by ether extraction of the completely saponified reaction mixture, the
latter was acidified with sulfuric acid and continuously extracted with ether. From the
concentrated ethereal extract, a part of the adipic acid was removed by filtration. At-
tempts to remove the e-hydroxycaproic acid from the filtrate by distillation were un-
successful. The mixture was therefore acetylated (8), and the acetate of the hydroxy
acid was isolated by distillation (b. p. 134 to 145 at 2 mm.). This was saponified,
acidified with dilute sulfuric acid, extracted with ether and concentrated in vacua.
The acetic acid was removed by evaporation in vacua at room temperature in an all
glass apparatus, the condenser bulb being cooled with carbon dioxide snow and acetone.
After four days the bulbs were opened. The distillate consisted of dilute acetic acid.
The residue was a pale yellow, viscous liquid having a slightly rancid odor. It solidified
to a glass when cooled with solid carbon dioxide. Titration indicated that it consisted
of a mixture of e-hydroxycaproic acid and its self -ester containing about 80% of the
free acid.

Anal. Calcd. for hydroxycaproic acid: neutral equivalent or saponification equiva-
lent, 132. Found: neutral equivalent, 164.1, 164.3; saponification equivalent, 130.5.

-Caprolactone. The lactone prepared as described in A was very soluble in alcohol,
benzene, ether, ethyl acetate and water, but insoluble in petroleum ether; b. p. 98 to
99 at 2 mm.; 4 4 1.0698; w 2 D 4 1.4608; M R calcd., 29.44; MR found, 29.15.

Anal. Calcd. for C 6 Hi O 2 : C, 63.11 ; H, 8.84; mol. wt., 114; saponification equiva-
lent, 114. Found: C, 62.81, 63.39; H, 8.68, 9.07; mol. wt. (in freezing benzene), 120,
122; saponification equivalent, 113.4, 113.3.

Dimeric e-Caprolactone. Obtained as described in A: granular crystals; soluble
in most organic solvents, but insoluble in petroleum ether or water; m. p. 112 to 113.

Anal. Calcd. for Ci 2 H2oO 4 : C, 63.11; H, 8.84; mol. wt., 228. Found: C, 62.98,
63.39; H, 8.84, 9.09; mol. wt. (in freezing benzene), 243, 226.

Hydrazide of c-Hydroxycaproic Acid. The monotneric and dim eric lac tones as
well as the polyester described in A, and the ethyl ester described in C (below) were
separately warmed with hydrazine hydrate for several hours on a steam-bath. In
each case the mixture solidified on being cooled; and after crystallization from ethyl
acetate or alcohol, the hydrazide was obtained in the form of white crystals melting at
114 to 115. It was very soluble in water.

Anal. Calcd. for C 6 HuO 2 N 2 : N, 19.18. Found: N, 18.68, 18.74.

Attempts to Prepare -Caprolactone from -Bromocaproic Acid. Prior to the ex-
periments described above, attempts had been made to prepare -caprolactone by the
action of sodium thylate on -bromocaproic acid. By this method Marvel and Birk-
himer (9) obtained a small amount of material "which seemed to be slightly impure
c-caprolactone," but its identity was not established.

e-Bromocaproic acid with an equivalent amount of sodium ethylate in absolute
alcohol was refluxed for eight hours. The mixture was made very slightly acid with a
few drops of hydrochloric acid, then filtered, concentrated and distilled in vacuo.
After a small preliminary fraction, a halogen-free product distilling at 104 to 106 at
4 mm. was obtained in 37% yield.

This,, presumably, is similar to the product described by Marvel and Birkhimer,

XXIII. e-CAPROLACTONE AND ITS POLYMERS 239

but it proved to be the ethyl ester of c-hydroxycaproic acid. Unlike e-caprolactone it
was insoluble in water; and it did not crystallize on being strongly cooled but merely
became very viscous. It was soluble in most organic solvents; d? 5 0.9944; w 2 D 5 1.4381;
MR calcd., 42.32; MR found, 42.24.

Anal. Calcd. for C 8 Hi 6 O 3 : C, 60.0; H, 10.0; mol. wt., 160; saponification equiva-
lent, 160. Found: C, 59.98; H, 10.08; mol. wt. (in boiling acetonitrile) 151, 154;
saponification equivalent, 158.7, 158.8.

When treated with hydrazine hydrate it yielded the hydrazide already described
in B (above).

The residue from the distillation of the ethyl ester was distilled further and yielded
a small fraction boiling at 110-200 at 4 mm. From this halogen-free distillate granular
crystals of the caprolactone dimer described in B separated (identified by mixed melt-
ing point).

The residue still remaining (41%) solidified to a paste on standing, and was ob-
tained as a light colored powder melting at 51 to 53 after several crystallizations from
alcohol. It was similar to the polyester described in A, but it had a somewhat lower
apparent molecular weight.

Anal. Calcd. for HO [ (CH,), CO O ] ir - H: C, 62.7; H, 8.8; mol. wt.,
1842. Found: C, 63.56; H, 9.08; mol. wt. (in freezing benzene) 1980, 1660.

When treated with hydrazine hydrate it yielded the hydrazide already described.

From the above described reaction of sodium ethylate on bromocaproic acid no
rnonomeric lactone was isolated. It is, however, not permissible to infer that no capro-
lactone was formed, since hydrochloric acid was present, and this, doubtless, would
strongly catalyze the polymerization of any caprolactone that might have been formed.

Summary

-Caprolactone has been prepared for the first time. It is the principal
product of the self-esterification of the corresponding acid. A small
amount of the cyclic dimeric ester (14-membered ring) is formed at the
same time. Under the action of heat c-caprolactone is converted to a poly-
ester of high molecular weight. The process is not easily reversible. This
behavior is compared with that already observed for other cyclic and poly-
esters.

Bibliography and Remarks

(1) For terminology.

(2) Chuit and Hausser, Helv. Chim. Ada, 12, 463 (1929); Bougault and Bourdier, Compt. rend ,
147, 1311 (1908); Lycan and Adams, Journ. Am. Chem. Soc., 51 625, 3450 (1929).

(3) Some of their derivatives are meagerly described in the following references: Baeyer and
Villiger, Ber., 33, 863 (1900); Helferich and Malkomes, ibid., 55, 704 (1922); Marvel and Birkhimer,
Journ. Am. Chem. Soc., 51, 260 (1929).

(4) Cf. Blaise and Koehler, Compt. rend., 148, 1772 (1909).

(5) Under these conditions polyesters having 6-atom units depolymerize very rapidly and smoothly
Those having longer units depolymerize very slowly and incompletely if at all, although in many such
cases a smooth depolymerization can be effected in the presence of an ester-interchange catalyst.

(6) Baeyer, Ber. t 32, 3619 (1899); Baeyer and Villiger, ibid., 3628.

(7) Baeyer, ibid., 29, 27, 30 (1896); Baeyer and Vill^er, ibid., 32, 3629 (1899).

(8) Cf. Chuit and Hausser, Helv. Chim. Acta, 12, 463 (1929).

(9) Marvel and Birkhimer, Journ. Am. Chem. Soc., 51, 260 (1929).

240 POLYMERIZATION AND RING FORMATION

XXIV. Cyclic and Polymeric Formals*

Polymerization and ring formation follows the general scheme, f

Reaclants

heat and action of
a dehydrating medium

a-polymer
linear polymers of molecular wt. between 1000 and 5000

heating in the molecular still

(distillate)

^-product; cyclic monomer <

or dimer heat

heat

(residue)

co-product; linear polymer ofmol.
wt. up to 20,000.

y-product; similar or identical
with a-product.

It is to be expected that a similar scheme holds for the acetals derived
from glycols. In the present paper the action of glycols onformals is de-
scribed.

A new class of linear polymers and large rings is obtained.
Heating a glycol with dibutylformal and an acidic catalyst at about 150
yields an a-polyacetal. The molecular weight is around 2200. When
a-polyacetals are heated at 230 to 250 in the molecular still, b- and
(^-products are formed. The cyclic monomers (^-products) are liquids
with characteristic odors fitting completely in the scheme developed by

* J. W. Hill and W. H. Carothers; Journ. Am. Chem. Soc. t 57, 925-8 (1935); Com-
munication No. 152, from the Experimental Station of E. I. du Pont de Nemours and
Co.

Received March 21, 1935.

t Compare paper No. I on page 4 and No. XII on page 156.

XXtV. CYCLIC AND POLYMERIC FORMALS 241

Ruzicka and discussed in a previous paper.* The ^-polymers are
hard and tough microcrystalline masses, which can be drawn out into
strong, pliable, silk-like filaments : X-ray diagrams indicate high orien-
tation. Depolymerization was studied. The relative rate of hydrolysis
of the ethylene acetals derived from formaldehyde, acetaldehyde and ace-
tone is approximately 1:4000:44,000.

Macrocyclic esters and anhydrides can be obtained by depolymerizing
the corresponding linear polymers (1). The various relations involved
are shown in the following diagram.

Reagents > a
(linear
mol. wt

-polymer
polymer of
. 1000-5000)
heating
in vacuum

/3-product distillate (cy-
clic monomer or dimer) <

i .

co-polymer residue (lin-
ear polymer of mol. wt.
10,000-20,000)

heat

7-polytner (similar to or identical
with a-polymer)

Heating the a-polymer causes (a) formation of traces of cyclic monomer
and dimer derived from one and two structural units of the molecular
chain and (b) coupling of the a-polymer chains to form the still longer
chains of co-polymer. Linear polyesters and polyanhydrides result from
reversible bifunctional condensations. In the ideal case, it would be pos-
sible to establish an equilibrium among all the species of a-, ft- and co-forms
and constant removal of the /3-forms by distillation should result in a com-
plete transformation of the sample into cyclic monomer and/or dimer.
This ideal is closely approached in some cases, the chief critical factors in-
volved being inherent mobility of the link, catalysis, temperature and speed
with which possible volatile products are withdrawn.

The fact that acetal interchanges are smoothly reversible reactions sug-
gests that it should be possible to realize all these transformations among
the acetals derived from glycols. For formals this is indeed true, and the
compounds now reported constitute a new family of linear polymers and
large rings. Here also, as with the esters and anhydrides, a single member
may be obtained in the form of a macrocrystalline solid, an odorous (e. g.,

* Compare paper No. XX on page 212 and No. XXI on page 221.

242 POLYMERIZATION AND RING FORMATION

musk-like) liquid, or a tough microcrystal line mass capable of being drawn
into strong, pliable, highly oriented, silk-like filaments.

Preparation and Behavior of Formals. Heating of a glycol with dibutyl
formal and an acidic catalyst at about 150 results in acetal interchange
with the distillation of butanol

HO(CH 2 ) n OH + BuOCH 2 OBu > . . . CH 2 O(CH 2 )nO . . . + 2BuOH

Trimethylene or tetramethylene glycol thus yields a mobile liquid easily
distilled in vacuo, and the major primary product is, therefore, presumably
the cyclic monomer. When the glycol is pentamethylene, or a higher one
the residue remaining from the removal of the alcohol is a viscous liquid.
To ensure complete removal of volatile products, it is heated to 200-220
at low pressure. The remaining non-volatile product is then the a-poly-
acetal. That derived from decamethylene glycol, for example, is a waxy
solid which crystallizes as a powder from hot ethyl acetate and then melts
at 56.5-57. Its observed apparent molecular weight was 2190. Similar
a-polyformals were obtained from pentamethylene, hexamethylene, tetra-
decamethylene and octadecamethylene glycols as well as from triethylene
glycol. The last gave a sirupy product which showed 110 tendency to
crystallize; the others were all solids.

When the a-polyformals were heated to 230-250 at low pressure in a
still provided with a condenser placed close to the evaporating surface (3),
conversion to the ft and w forms occurred. Compared with the polyesters
derived from carbonic acid, the rate of distillation was quite slow; in this
respect, the polyformals resembled the previously described (2) polyesters
derived from the higher dibasic acids such as sebacic; the tendency toward
co-polymer was greater than toward the ft forms, and as the viscosity of the
residue increased, the rate of distillation became less. The ultimate yield
of distillate was, therefore, relatively small, and, as in the case of those
polyesters that manifested a similar behavior, the distillates were largely
the cyclic dimers which were without exception definitely crystalline solids
of sharp melting points. The presence of cyclic monomer in the distillate
could be inferred from the characteristic odor, and the monomer from
pentamethylene formal was obtained in sufficient quantity to permit iso-
lation and purification. The a-polyformal from triethylene glycol differed
from the other formals of similar unit length. It was a viscous liquid and
when heated in vacuum at 200-250 it depolymerized very rapidly. The
product apparently consisted essentially of monomer, which was isolated
in a state of purity and in considerable quantity by fractional crystalliza-
tion of the distillate.

XXIV. CYCLIC AND POLYMERIC FORMALS 243

Polymerization of Cyclic Formals. The possible polymerization of
cyclic acetals has occupied an important place in speculations concerning
cellulose, starch, etc. (3), but no clear example of a polymerization originat-
ing in an acetal linkage per se has been adduced. Among cyclic esters,
rings of more than 5 atoms generally polymerize on being heated with
catalysts and those of 6 atoms occupy a peculiar position because this
transformation occurs with especial facility and is easily reversible (4).
It is of interest in this connection that trimethylene formal (6-ring) could
not be induced to polymerize. The monomeric tetramethylene, penta-
methylene and triethylene glycol formals, however, quickly became more
viscous when heated (e. g., at 150) in the presence of a trace of sulfonic
acid. Dimeric decamethylene formal (26-ring) when treated in this man-
ner was converted to a microcrystalline solid having approximately the
same melting point and molecular weight (2500) as the a-polymer from
which the dimer was originally derived. Results obtained with monomers
are indicated in the table.

VISCOSITY CHANGE (IN ARBITRARY UNITS) OF MONOMERIC FORMALS HEATED WITH A
TRACE OF CAMPHOR SULFONIC ACID

Successive Time of

Formal Ring size heating intervals flow, sec.

Trimethylene 6 Initial 4.5

+2hrs. at 100 4.5

-fl.Shrs. at 150 4.5

Tetramethylene 7 Initial 4.5

+2hrs. at 100 4.5

+Q.5hr. at 150 35
+ 1.0 hr. at 150

Pentamethylene 8 Initial 4.5

+2hrs. at 100 13

4-0. 5 hr. at 150 16

Odors. The remarkable parallelism between the odors of macrocyclic
ketones, lactones, carbonates, malonates, oxalates, sebacates, etc., of simi-
lar ring size previously reported (5) suggested that for musk-like odor the
only requirement is a C=O group in a ring of suitable size. Meanwhile,
Ruzicka (6) has extended his own researches and observed that the cyclic
imine | (CH 2 )ie NH | ^as a musk-like odor, thus demonstrating that

the presence of the carbonyl group is not necessary. The odors of the cyclic
formals indicated in Table I now point to the same conclusion. Only the
compounds marked* were actually isolated, but there is no doubt that the
characteristic odor of the crude distillate was in each case due to cyclic
monomer. The odors of the formals I were so strikingly similar to those

244 POLYMERIZATION AND RING FORMATION

of the corresponding carbonates II as to be almost indistinguishable from
them.

(ino-iH, W<Uc

(I) (ID

TABLE I
ODORS OF MONOMERIC FORMALS

Glycol from which Ring

formal is derived size Odor

*Tetramethylenc 7 Sweet and penetrating

*Pentamethylene 8 Minty, camphoraceous

Hexamethylene 9 Minty

Nonamethylene 12 Earthy, camphoraceous

Decamethylene 13 Cedar, camphoraceous

Tetradecamethylene 1 7 Musk-like

Octadecamethylene 21 No odor

*Triethylene glycol 11 Faint, flowery

Fibers from co-Polyformals. w-Polyesters and polyanhydrides can be
drawn out into continuous filaments which readily accept permanently a
high degree of orientation along the fiber axis and have very good strength
and pliability (7). These materials are in fact the only truly synthetic
fibers for which any measured strengths have been reported although it
appears that continuous oriented filaments can also be drawn from poly-
oxymethylene and polyethylene oxide (8) when their molecular weights
are sufficiently high. In all these classes of compounds as in cellulose it-
self, very long molecules are
made up of units joined through
C-O bonds. In the co-polyfor-
mals as in cellulose, this linkage
is actually an acetal linkage.
It is not surprising then that
the co-polyformals can also be
drawn into oriented filaments.
X-ray diagrams for unoriented
and oriented w-polydecamethyl-
ene formal are shown in Figs. 1
and 2.

Fig. 1. x-Ray diffraction pattern of w-deca- Ease of Depolymerization.

methylcnc formal. The most obvious factor in-

XXIV. CYCLIC AND POLYMERIC FORMALS

245

fluencing ease of depolymerization is the reactivity characteristic of the type
of linkage involved. Polyanhydrides are depolymerized much more readily
than polyesters. The relative ease of hydrolysis of the ethylene acetals de-
rived from formaldehyde, acetaldehyde and acetone is approximately
1:4000:44,000 (9). That the alkylene formals depolymerize with dif-
ficulty is therefore not surprising. One might expect that it would be
very much easier to form macrocyclic acetals from other aldehydes and
ketones; unfortunately, various other complications arise in these cases.

Rings of 9 to 12 atoms are
especially difficult to form and
in some cases exceptionally un-
stable. This may be attributed
to repulsions arising from meth-
ylene hydrogens whose ex-
ternal radii are forced by the
shape of the molecule into a
space too small to receive them
(10). As we have already
pointed out (11), oxygens in a
chain may relieve strains of this
type since they carry no hydro-
gens and their valences are prob-
ably more flexible than those of

carbon (12). Further presumed illustrations of the oxygen effect: poly-
formals (III) are definitely more difficult to depolymerize than are poly-
carbonates (IV), and trie thy lene glycol formal (V) is very much more
easily depolymerized than alkylene formals of similar unit length.

. . (O R O CH 2 )x (O R O CO)* . . .

(HI) (IV)

. . . (OCH,CH 2 OCH a CH20CH,CH,OCH 2 ) x --. . .
(V)

Experimental Part

Tetramethylene Formal. Difficultly separable mixtures were formed when the inter-
change method was applied to this compound. The most convenient method was to dis-
til the glycol (30 g.) with trioxymethylene (10 g.) and a trace of camphor sulfonic acid
(bath temperature, 210). Fractionation of the distillate gave some tetrahydrofuran,
and then a liquid boiling at 112-117 which was washed with strong caustic and redis-
tilled at low pressure. It had dj 1.0022; w 2 D 1.4310; M D found, 26.34; M D calcd.,
26.38.

Decamethylene Formal. Dibutyl formal (0.2 mole) with a 5% excess of decamethyl-
ene glycol and 0.1 g. of ferric chloride gave fairly rapid reaction at 165 (bath). The

Fig. 2. x-Ray diffraction pattern of cold-drawn
o)-decamethylene formal.

246

POLYMERIZATION AND RING FORMATION

temperature was raised to 200 during two hours and heating continued for one and one-
half hours in a good vacuum. The distillate was 98% of the theoretical calculated as
butyl alcohol; the yield of residual a-polymer was 103%. When cold, it was a light
brown rather hard wax. When dissolved in hot ethyl acetate (150 cc. for 17.5 g.), it
separated in the form of a microcrystalline powder; soluble in chloroform, benzene, car-
bon tetrachloride and xylene; insoluble in alcohol, ether, petroleum, hydrocarbons and
acetone.

0-Product. Eight grams of the crude a-polymer was heated in a 250-cc. suction
flask provided with a test-tube through which water could be circulated to act as an in-
ternal condenser (13) ; temperature, 230-250 (bath) ; pressure, about 1 mm. After forty-
eight hours 2 g. of distillate had collected. It was a pasty mixture of liquid and crystals
having a pleasant camphoraceous odor. The solid portion after crystallization from
alcohol was odorless and melted at 93-94 . It was the cyclic dimer.

w-Polymer. The residue from the above was a hard, very tough, opaque, leather-like
mass. It melted (became transparent) at 58-63 , but at this temperature it was too
stiff to flow and showed considerable resistance to deformation. At slightly higher
temperatures, it could be drawn out into thin strips or filaments which could be stretched
and cold-drawn (14). The product then showed fiber orientation (Fig. 2) and also ex-
hibited parallel extinction between crossed Nicols. The cold-drawn material was ex-
ceedingly strong, tough and pliable.

TABLE II
DATA ON FORMALS

Name of formal
Tetramethylene,

monomer
Decamethylene, a-

polymer

Decamethylene, dimer
Pentamethylene, dimer
Pentamethylene,

monomer

Hexamethylene, dtraer
Nonamethylene, dimer
Tetradecamethylene,

a-polymer
Tetradecamethylene,

dimer
Triethylene glycol,

monomer

In boiling benzene;

7-Polymer. The cyclic dimer (0.5 g.) with a trace of camphor sulfonic acid heated
at 150 soon became very viscous and the characteristic odor of the cyclic monomer ap-
peared. After an hour, the melt when cooled set to a hard wax, easily electrified when
powdered. Purified out of benzene, it separated as a microcrystalline powder, m. p. 58-
59; molecular weight observed in freezing benzene, 2580.

Other Formals. The other a-polyformals were for the most part brittle waxy solids,
and the observed melting points were: pentamethylene 38-39, hexamethylene 38,
nonamethylene 54-55, tetradecamethylene 68-69, octadecamethylene 71-72 and

Calcd.

Anaiyt
Mol.

i cm ua

t jj ,
Found
Mol. wt. in
freezing

M. p., C.

Formula

wt.

benzene

, p. 112-117

C 6 HioO 2

102

59 6

103

56-57

(CuHOj) x

11.

(186) x

68.9

2190

93-94

(CuHwCfeh

11.

372

70 8

, 1

368

3.5-56

(CeHi20 2 ) 2

232

61.8

262

. p. 40-44

CeHisOs

(11 mm.)

71-72

(C7Hl 2 2 )2

260

64.0

257

68-69

(CioH 2 o0 2 )i

11.

344

69 8

11.

334 a

68-69

(CisHaoOs)*

(242)*

73 . 1

2480

103.5-104

(CwHaoOz)*

484

74 4

503

18-20

C:HuO4

162

51.6

161

XXIV. CYCLIC AND POLYMERIC FORMALS 247

triethylene, sirup. The 0-polymers were generally pasty solids having the odors indi-
cated in Table I and, when purified by crystallization from alcohol, yielded the odorless
crystalline dimers whose melting points are shown in Table II. Pentamethylene formal
depolymerized more readily than its higher homologs, and the monomer was isolated
from the 0-product as a colorless liquid of camphoraceous odor; b. p. 40-44 at 11 mm.
Triethylene glycol formal depolymerized more readily than any of the others and gave a
70% yield of 0-product which apparently consisted largely of monomer melting at 18-
20.

Summary

A new class of linear polymers represented by the general formula
. . . [CH 2 O R O] x . . . is obtained by the action of alkyl formals on
the higher glycols (above tetramethylene). These a-polyformals can in
part be depolymerized to the j3-forms (cyclic monomer and dimer) and in
part converted to the higher co-polyformals. The latter can be drawn out
into strong, pliable, highly oriented fibers. The /3-forms constitute a new
type of large rings; the monomers have odors scarcely distinguishable
from the corresponding carbonates j O R O CO j and in particular

tetradecamethylene formal, the 17-membered ring, has a musk-like odor.
The dimers are odorless crystalline solids. Trimethylene formal (6-mem-
bered ring) does not polymerize, and no polymeric form is known, but
tetramethylene formal (7-membered ring) and the higher ones polymerize
rapidly (e. #., at 150) when catalyzed by acid.

Bibliography and Remarks

(1) Hill and Carothers, ibid., 64, 1569 (1932); 55, 5023, 5031 (1933).
.(2) Ibid., 55, 5035 (1933).

(3) See, for example, Hibbert and Timm, Journ. Am. Chem. Soc., 45, 2433 (1923); Hill and
Hibbert, ibid., 3108 3124; Helferich and Sparmberg, Ber., 64, 104 (1931); Bergmann and Miekeley,
ibid., 62, 2297 (1929).

(4) Carothers and van Natta, Journ. Am. Chem. Soc., 52, 314 (1930); Hill and Carothers, ibid.,
55, 5037 (1933).

(5) Ibid., 55, 5039 (1933).

(6) Ruzicka, Salomon and Meyer, Helv. Chim. Ada, 17, 882 (1934).

(7) Journ Am. Chem. Soc., 64, 1579 (1932).

(8) Staudinger, "Die hochmolekularen organischen Verbindungen," Julius Springer, Berlin, 1932,
p. 262; Sauter, Z. physik. Chem., 21B, 161 (1933).

(9) Leutner, Monatsh., 60, 317 (1932).

(10) Stoll and Stoll-Comte, Helv. Chim. Acta, 13, 1185 (1930); Hill and Carothers, Journ. Am.
Chem. Soc., 55, 5023, 5031 (1933).

(11) Journ. Am. Chem. Soc., 55, 5050 (1933).

(12) Cf. Sidgwick, "Annual Reports of the Progress of Chemistry for 1932," p. 73.

(13) Journ. Am. Chem. Soc. t 55, 5035 (1935).

(14) Ibid. t 54, 1580 (1932).

248 POLYMERIZATION AND RING FORMATION

XXV. Macrocyclic Esters*

Cyclic esters can be prepared^ with alkaline catalysts from linear poly-
esters derived from carbonic, oxalic or malonic acid, but not from higher
members of the homologous series.

It has been found that by depolymerization at exactly 270 at a pressure
below 1 mm. Hg in the presence of hydrous metal chlorides, such as SnCl 2 --
2H 2 O, KnCl 2 -4H 2 O, FeCl 2 -4H 2 O, or of magnesium powder good yields
of macrocyclic esters are obtainei.

Thirty -six new esters of this type partly monomer ic, partly dimeric
derived from succinic acid and from other, higher dibasic acids, have been
prepared and are described.

The molecular refractions of large rings are generally considerably less than
the calculated values, the depression, ranging from 0.4 to 0.7, being most
marked in the region between 8 and 15 carbon atoms. This is attributed
to the strain and interference of the peripheral atoms described in a pre-
vious paper** and connected with the relatively large sphere of interference
attributed to the hydrogen atoms of the diain.

The depolymerization of linear polyesters derived from carbonic oxalic,
or malonic acid to the corresponding cyclic monomers and dimers proceeds
smoothly and rapidly in the presence of alkaline catalysts (1), but the same
conditions give very poor results when applied to polyesters derived from
succinic acid or from higher acids of the series. We have, however, now
found a number of catalysts that act effectively in these cases. The result
is a considerable extension of the useful range of the interchange method
in the preparation of macrocyclic esters. Since n different glycols and an
equal number of dibasic acids may give 2n 2 different cyclic monomers and
dimers, the number of possible compounds is very large. In this paper,
we describe thirty-six new macrocyclic esters. Their properties are listed
in Table I and are discussed in a later paragraph. Details of procedure are
presented in the Experimental Part, but points of essential importance are
as follows.

The depolymerizations are carried out in the glass apparatus depicted
in Fig. 1 (2). The outer flask contains a liquid whose refluxing vapors
heat the inner chamber where the depolymerization occurs. This makes it

* E. W. Spanagel and W. H. Carothers; Journ. Am. Chem. Soc. t 57, 929-34 (1935) ;
Communication No. 153 from the Experimental Station of E. I. du Pont de Nemours
and Co.

Received March 21, 1935.

t Compare paper No XX on page 212.

** Compare paper No. XXII on page 22.~>.

XXV. MACROCYCLIC ESTERS

249

possible to control the temperature very exactly and ensures an adequate
heat input. The optimum temperature appears to lie in the neighborhood
of 270; the pressure should be about 1 mm. or lower. Materials that
act as effective catalysts with esters from higher acids are SnCl2.2H 2 O,
MnCl 2 .4H 2 O, FeCl 2 .4H 2 O, MgCl 2 .6H 2 O, CoCl 2 .6H 2 O, MnCO 3 , MgO, Mg-
CO 3 and Mg powder.

By way of specific example, it may be mentioned that 20 g. of polymeric
hexamethylene sebacate with 0.2 g. of MgCl 2 .(>H 2 O, in two hours gave 15
g. of distillate from which 11.4 g. of pure crystalline monomer was isolated.
In the preparation of other esters listed in Table I, the crude yields gen-
erally ranged from 40 to 85%, while in favorable cases the yields of pure
monomers approached 70%. Presum-
ably both monomers and dimers are
always present in the crude depoly-
merizate. In fact, for unit lengths of
8, 9, 10 and 11, the dimers tend to
predominate whereas for unit lengths
above 13 very little dimer, if any, is
ordinarily isolated. The ratio of dimer
to monomer can be controlled within
certain limits since if refluxing in the
depolymerization vessel is permitted
relatively little dimer can escape.

Physical Properties. The molecu-
lar refractions of large rings are
generally considerably less than the
calculated values (3), the depressions
commonly ranging from about 0.4 to
0.7. This peculiarity is, however,
for the most part restricted to rings of 8 to 15 atoms, and some of the largest
rings (e. g., 30-34 members) even show considerable exaltations (3). In
Table I molecular refractions for 26 new macrocyclic esters are listed. The
average depression is 0.32. The values fluctuate rather widely, and this
may be due in part to experimental error since the measurements had to be
made at a relatively high temperature. It must, however, be regarded
as a significant fact that, in general, the largest depressions are found for
the smaller rings (9 to 14 atoms) where the strains or interferences of the
peripheral atoms are greatest, while depressions very close to zero are found
only in three cases. These are listed below together with one of the cyclic
carbonates (4) which shows a similar peculiarity.

Fig. 1. Apparatus for depolymerizing.

260

POLYMERIZATION AND RING FORMATION

Ring size
21

20
16

Formula

+0.01
.00

, 0(CH 2 )i 8 OCO i

, Q(CH a ) l4 OCO(CH 8 )aCO |

r (OCH 2 CH2),OCO(CH 8 ) 4 CO [ - .09

4- .03

The first two of these are larger rings than any other esters for which values
are available. The other two have in the ring extra oxygens which are
capable of relieving steric interferences (5).

Densities of the new esters for
which values are available are
plotted in Fig. 2. The curve
shows no maximum, and the den-
sities of isomers lie quite close
together.

The odors of the macrocyclic
esters are especially interesting,
but characterizations are not
sufficiently exact to justify a de-
tailed discussion. It may be
pointed out, however, that Table I
includes 3 new rings of 14 atoms,
3 of 15, 4 of 16, 7 of 17, 2 of 18,
and one each of 19 and 20 atoms.
Those of 18, 19 and 20 atoms are
practically odorless; the odors
most prominent in the others are
woody or cedary, earthy, cam-
phoraceous and musk-like. The
musk-like odor is very pronounced
in all of the 17-atom rings.

Data on melting points are rather incomplete, but here, as in the series
previously examined, there is little to indicate any regular relation between
ring size and melting point. Values for succinates are shown in Fig. 3.
As the ring size increases, the melting point first falls to a minimum, then
it rises to a maximum, falls again and becomes almost constant. The
jsomeric 17-membered rings have the general formula i OCO(CH 2 ) n COO-

(CH 2 )m | w j iere n = 17 ( m 4. 4). The melting points for various
values of n are

1.18

1.16

1.14

1.12

1.08

a 1.06
Q
1.04

1.02
1.00
0.98

20 22

12 14 16 18
Atoms in ring.

Fig. 2. Densities (60/4) of cyclic esters:
A, succinates; O, other cyclic esters.

XXV. MACROCYCLIC ESTERS

251

M.p.,

3
14

4
26

7
59

8
37

11
-8

Ease of Ring Formation. As measured by yields, it has been found for
every series thus far examined that rings of 9 to 12 atoms are more difficult
to form than rings of 15 to 20 atoms. Stoll and Stoll-Comte (6) have sug-
gested that this is due to a type of steric hindrance caused by mutual re-

10 12 14 16 18 20 22 24

Atoms in ring.
Fig. 3. Melting points of monomeric alkylene succinates.

pulsions of peripheral hydrogens which are crowded against one another
in rings of intermediate sizes. The position of the minimum in the yield
vs. ring-size curve is not the same in different series. For cyclic ketones it
lies at 10 to 11 atoms; for esters derived from dibasic acids, it apparently
lies at 7 to 8 atoms. At any rate, monomers of this size have not yet been
obtained among the dibasic esters, while, for example, the 7-atom lactone
is obtained as a major primary product from c-hydroxycaproic acid (7).

Another point of interest is that in the intermediate range odd-mem-
bered rings are relatively more difficult to produce than the adjacent even
ones. An alternating effect of this kind was first observed in the dibasic
acid anhydride series (8). Although rings of 8. 10 and 12 atoms were
formed, those of 9, 11 and 13 atoms were not (the corresponding dimers
were formed instead). More recently, an alternation in yield of cyclic
ketones in passing from one member to the next has also been observed by
Ziegler and Aurnhammer (9) in syntheses by the dilution method. The
type of alternation observed in the present study is illustrated by Fig. 4
where the yield of dimer is plotted against unit length for glycol esters of
succinic acid. The material not dimer is mostly monomer, and so by in-
verting the ordinate scale, a rough plot of monomer yield is obtained. The

252

POLYMERIZATION AND RING FORMATION

alternating effect is more clearly brought out by the inset of Fig. 4 which
shows the rate of change of dimer yield with increasing unit length. One
is tempted to compare this behavior with the well-known alternation in
physical properties of certain series of open chain compounds, but the
responsible factors are probably quite different in the two cases. Ex-
amination of models (with large hydrogens) shows that the interferences to
ring closure in chains of 9 to about 13 atoms appear to be greater for the

100

h iU .U 1-t lo Irt 2U

Atoms in cyclic monomer.

Fig. 4. Yield vs. ring size in the preparation of alkyl-
ene succinates.

odd members (10), arid it is probable that these interferences are still opera-
tive after the chains are closed so that the odd-membered rings are less
stable than the even ones.

Since the rates at which different rings are formed are sure to be dif-
ferently affected by changes in conditions, the inherent relative ease of ring
formation is a concept to which no quantitative significance can be at-
tached. Nevertheless, the effect of structure and particularly of chain
length is so pronounced that even crude data permit a qualitative compari-

XXV. MACROCYCLIC ESTERS

253

son of the inherent relative ease of forming rings of different sizes. A
comparison of this kind for glycol esters of dibasic acids is shown in Fig. o,
and it includes a similar plot for the stability of cyclic anhydrides. Ease
of ring formation and stability will, in general, run parallel, but 6-membered
cyclic esters furnish a notable exception to this rule. They are both more
easily formed and less stable than larger ester rings.

Methods of Ring Formation. For the synthesis of large rings, there
are now available two methods based upon rational principles : the dilution
method first explicitly formulated and applied by Ruggli in 1912 (11) and
the interchange method described above.

Ruggli's dilution method was cited in early papers of this series (12)
where the general theory of bifunctional reactions was developed, but it

Pirfec

Low

Zero

16 18

8 10 12 14
Atoms in ring.
Fig. 5. Ease of formation and stability vs. ring si/A'.

received no further published attention until very recently. Ziegler and
his co-workers (13) have now applied it in a very ingenious fashion to the
synthesis of cyclic ketones in good yields, and further successful applica-
tions have been made by various investigators to the preparation of macro -
cyclic amides (14), imines (15) and lactones (1(5).

In a recent discussion of ring formation (17), Ziegler suggests that the
interchange method will be severely restricted in the range of its applica-
bility. On the contrary, as we have shown in this and the preceding paper,
it presents extensive possibilities, and it may be appropriate to point out
some of the peculiar advantages of this method.

It is true, of course, that the interchange method is inherently restricted
to reversible bifunctional condensations, but almost all condensations are

264

POLYMERIZATION AND RING FORMATION

o
CO CO
O t<

><5 O
CO C

CO >O -i CO ^ ** t- '

I I I I I I J I I

i 00 O3 O

I t 1 I I I I I I I I I + I I I I

) oo oo <

I CO r- !

' - 5
s o c

OO^OO

to o co co o

l^ CO 10 O 00

"H CO Q CO i i

o o o o o

"S fi

<N W - O *-<

DOiOOO
o w S S

J CM t-"

01 M (N (M

fOCOQOO

cc a o o o> o o>
04 ^ececioio cc

S) ; I

"2 S fi

Cyclic ester

3
% Si d

U*{,u. l <U4,rtrt-rJ.2 -^

s 1 1 1 3 * 3 1 -S I i i .& 5

WhwWOQQhHOQHfcw

Tetramethylene azelate

Hexamethylene azelate

Ethylene sebacate

Trimethylene sebacate

J3
8
V

Pentamethylene sebacate

Diethylene glycol sebacate
Hexamethylene sebacate

Ethylene decamethylene dicai

Ethylene brassy late
Decamethylene octadecanedic

a
.2

XXV. MACROCYCLIC ESTERS 255

reversible to a certain extent. We have already presented some evidence
(18) indicating that even polyke tones can be depolymerized and that this
is what is involved in Ruzicka's thorium salt method for the synthesis of
large carbon rings. Practically, the interchange method will be restricted
to those bifunctional condensations that are easily reversible, but, as the
present paper shows, much can be accomplished in that direction by the
choice of suitable catalysts, and the formation of esters and anhydrides by
no means represents the limit of the possibilities.

One other point that deserves mention is this: there are two general
types of bifunctional reactions symbolized in the two formulas

x R y > products, and

x R x + y R' y > products

In the first type (simple bifunctional), both the mutually reactive groups
are present on the same molecule; in the second (bifunctional) (12) the
mutually reactive groups are initially present on separate molecules. An
example of the first type is an hydroxy acid, of the second type, a dibasic
acid plus a glycol. The second type is much the more numerous, and it is
one to which the theoretical advantages of the dilution method are not
applicable with full efficiency since the initial step toward ring formation
necessarily involves an intermolecular reaction. Practical advantages of
the interchange method in connection with such problems as the prepara-
tion of lactones and Brings will be illustrated in future papers.

Experimental Part

The apparatus used in the present work is shown in Fig. 1 . The flask B contains a
liquid whose refluxing vapors heat the chamber A. This method makes it possible to
control the temperature very exactly and ensures an adequate heat input. The de-
phlegmator C may be cooled with steam if reflux is desired to restrict the distillation of
dimer, or it may be omitted entirely. For amounts of polymer in the neighborhood of
50 g., the chamber A should be of about 800 cc. capacity since considerable foaming
occurs at the start of the depolymerization.

Preliminary experiments were made with decamethylene carbonate using a sodium
catalyst. The importance of low pressure was soon demonstrated, and in further work
a pump with high capacity was used and the pressure kept at 1 mm. or below. The
rate increased, but the purity of the distillate dropped off somewhat as the temperature
was raised above 250. A temperature of 270 appeared to be about the optimum, and
it was conveniently supplied by boiling 0-chlorodiphenyl. These observations were
applied to the preparation of monomeric tetradecamethylene carbonate on a one hundred
gram scale using an alkaline catalyst; a 93% yield of monomer having a purity of 97%
was obtained in less than two hours.

These conditions, however, gave no satisfactory results with polyesters derived from
higher dibasic acids, and attention was therefore turned to a further exploration of

POLYMERIZATION AND RING FORMATION

catalysts. The polyesters used were prepared from the appropriate dibasic acid and a
slight excess of glycol by heating at 200 (bath) finally at a pressure of 2-3 mm. Prod-
ucts obtained in this manner were gray waxy masses of varying degrees of hardness.
One of the first catalysts tried was stannous chloride, and the following results are
typical.

TABLE II
ANALYTICAL DATA FOR NEW CYCLIC ESTERS

Mol wt.

f Calculated

C, % H, % Mol. , Found

Compound

Trimethylene succinate, Monomer
Trimethylene succinate, Dimer
Tetramethylene succinate, Monomer
Tetramethylene succinate, Dimer
Pentamethylene succinate, Monomer
Pentamethylene succinate, Dimer
Hexamethylene succinate, Monomer
Hexamethylene succinate, Dimer
Heptamethylene succinate, Monomer
Heptamethylene succinate, Dimer
Octamethylene succinate, Monomer
Octamethylene succinate, Dimer
Nonamethylene succinate, Monomer
Decamethylene succinate, Monomer
Dodecamethylene succinate, Monomer
Tridecamethylene succinate, Monomer
Tetradecamethylene succinate, Monomer
Octadecamethylene succinate, Monomer
Decamethylene glutarate, Monomer
Hexamethylene adipate, Monomer
Triethylene glycol adipate, Monomer
Nonamethylene adipate, Monomer
Heptamethylene suberate, Monomer
Ethylene azelate, Monomer
Ethylene azelate, Dimer
Tetramethylene azelate, Monomer
Hexamethylene azelate, Monomer
Ethylene sebacate, Monomer
Trimethylene sebacate, Monomer
Tetramethylene sebacate, Monomer
Pentamethylene sebacate, Monomer
Diethylene glycol sebacate, Monomer
Hexamethylene sebacate, Monomer
Ethylene decamethylene dicarboxylate, Monomer
Ethylene brassylate, Monomer
Decamethylene octadecanedioate, Monomer

Fifty grams of polymeric hexamethylene succinate with 0.5 g, of vSnCl 2 .2H 2 O was
heated at 270 at 1 mm. or less. The polymer became soft and bubbled, and after
about ten minutes distillation started. The rates observed were : after fifteen minutes,
12 drops/min.; after thirty minutes, 14 drops/min. At this point the residue became a
stiff porous gel occupying about half the volume of the flask, but distillation continued :
after forty-five minutes, 10 drops/min. ; after sixty minutes, 8 drops/min. ; after seventy-
five minutes, 3 drops/min. The ultimate residue was a light porous mass of tough, in-

\_ttiu uiaicu *

% H, % Mol.
wt.

-Found
% H, %

freezing
benzene

53 .

158

53.

162

159

53 .

316

53 .

270

55.

172

55.

170

166

55 .

344

56.

332

58.

186

57.

177

182

58.

372

57.

395

376

GO.

200

59.

199

197

00.

400

60.

,41

418

393

01.

214

60.

,01

,07

200

01.08

428

61.

,91

,82

428

03.

228

63.

228

03 .

,77

456

03.

,40

,83

498

04.

,46

,09

242

.04

.08

236

229

05 .

,62

.37

250

.30

.35

250

240

07.

,85

284

67,

.77

.03

283

271

08.

10.

,06

298

.98

.27

290

277

09,

.23

10,

.25

312

.24

.20

314

300

.74

10,

.88

368

.58

.73

365

348

06.

.67

.03

270

.14

.64

271

265

63,

.16

.77

228

.21

.81

214

217

.96

.39

262

.20

.74

283

275

.67

.63

270

.54

.10

266

256

60,

.67

.63

270

.92

.79

272

269

.15

.41

214

.93

.25

216

212

.15

.41

428

.13

.81

466

431

.46

.09

242

.55

.21

244

234

.07

.63

270

.98

.44

260

269

228

232

224

.40

.09

242

.05

.89

248

236

.02

.37

256

.55

.58

268

254

.67

.63

270

.08

.37

264

258

.76

.82

272

.37

257

258

.60

.85

284

.41

.75

260

256

.62

.37

256

.41

.45

256

249

.67

.63

220

.88

.93

242

252

.33

.50

452

.72

.83

465

443

XXV. MACROCYCLIC ESTERS 257

soluble, yellow resin. The distillate (40 g.), consisting of a light yellow liquid and a
white solid, was filtered. The solid on recrystallization from alcohol yielded 8 g. of
dimer, white plates melting at 110. Any catalyst present in the monomer was removed
by dissolving the filtrate in ether and washing with water. The dried ether solution was
concentrated and distilled. The yield of cyclic monomeric hexamethylene succinate
boiling at 108-110 (2 mm.) was 21 g. This general procedure was used in isolating
other cyclic esters except that when the monomer was a solid it was usually purified by
crystallization from alcohol at low temperature. Yields of crude distillate varied from
40 to 85%.

For further study of catalysts, experiments were made with hexamethylene sebacate
since its monomer is a readily purified crystalline solid. The polyester (20 g.) was heated
with 0.2 g. of the possible catalyst in a glass still at 270 (1 mm ) for two hours. The
crude distillate was dissolved in 50 cc. of alcohol, cooled, and the pure monomer isolated
by filtration. Results are shown in Table III.

Pure
monomer, %

45
55
57
57
(55
55
60
40
45
55

8
18

TABLE III

CATALYSTS

FOR DEPOLYMERIZATION

Catalyst

Distillate, %

SnCl 2 .2H 2 O

MnCl 2 .4H 2 O

FeCl 2 .4H 2

MgCl 2 .6H 2

CoCl 2 .6H 2

Co(N0 3 ) 2 .6H,0

Mg (powd.)

MnCO 3

MgO

MgC0 3

PbCl 2

FeCl 3

T1 2 C0 3

SbCl 3

Th(NO 3 ) 4 .12H 2 O

None (control)

The following substances were also tried but gave very small distillates, from which
no monomer could be isolated: NiCl 2 .6H 2 O, FeSO 4 , TiCl 4 , tin dust, Mg 3 (PO 4 ) 2 , CrCl 3 ,
CaCO 3 , CrCl 2 , ZnCl 2 .

Typical yields of cyclic esters obtained at 270 /l mm. using 1 to 3% SnCl 2 .2H 2 O
as catalyst are shown below.

Losses in purification are due in part to polymerization during redistilla-
tion, and these can be largely avoided by first washing out the catalyst.
Considerable improvements can also be effected by suitable precautions in
preparing the polymer.

258

POLYMERIZATION AND RING FORMATION

TABLE IV
YIELDS OF CYCLIC ESTERS

Unit

Ester length

Ethylene succinate 8

Trimethylene succinate 9

Tetramethylene succinate 10

Pentamethylene succinate 1 1

Heptamethylene succinate 13

Octamethylene succinate 14

Nonamethylene succinate 1 5

Octadecamethylene succinate 24

Decamethylene glutarate 17

Nonamethylene adipate 1 7

Heptamethylene suberate 1 7

Ethylene azelate 13

Trimethylene sebacate 1 5

Tetramethylene sebacate 16

Pentamethylene sebacate 17

Hexamethylene sebacate 18

Diethylene glycol sebacate 17

Ethylene brassylate 1 7

Time,
hrs.

Crude
distillate

Distillate
isolated as
monomer dimer

70
68
52
40
51
73
67
72

48
58
79
84
70
60
73
81
58

Summary

The optimum conditions for the depolymerization of linear polyesters
have been more clearly defined and new catalysts have been found which
make it possible to apply this method generally to the rapid preparation of
macrocyclic esters in good yields. Thirty-six new cyclic esters, monomeric
and dimeric, derived from succinic acid and from other higher dibasic
acids are described. Conclusions developed in the study of other macro-
cyclic compounds have been confirmed and extended.

Bibliography and Remarks

(1) Hill and Carothers, ibid , 55, 5031 (1933).

(2) This design was suggested to us by Mr C H Greenewalt.

(3) Ruzicka and Stoll, Helv. Chim. Acta, 11, 1159 (1928), Ruzicka, Stoll, Huyser and Boeke
noogen, ibid , 13, 1152 (1930), Ruzicka and Boekenoogen, ibid., 14, 1319 (1931); Ruzicka, Hurbin,
and Furter, ibid , 17, 78 (1934), Hill and Carothers, Journ. Am. Chem. Soc. f 55, 5042 (1933)

(4) Journ Am. Chem Soc , 55, 5042 (1933)

(5) Hill and Carothers, ibid , 57, 925 (1935); 55, 5049 (1933)

(6) Stoll and Stoll-Cornte, Helv. Chim. Acta, 13, 1185 (1930)

(7) van Natta, Hill and Carothers, Journ. Am. Chem Soc , 56, 455 (1934).

(8) Hill and Carothers, Journ Am Chem. Soc , 55, 5025 (1933)

(9) Ziegler and Aurnhammer, Ann., 513, 43 (1934)

(10) Hill and Carothers, Journ. Am Chem Soc , 55, 5050 (1933).

(11) Ruggli, Ann., 392 92 (1912)

(12) E f> , Journ Am Chtm Soc , 51, 2551 (1929)

XXVI. META AND PARA RINGS 259

(13) Ziegler, Eberle and Ohlinger, Ann, 504, 94 (1933); Ziegler and Aurnhammer, ibid., 513,
43 (1934)

(14) Reid and Lippert, St. Petersburg Meeting of the A. C. S , March, 1934

(15) Salomon, Helv. Chim. Ada, 17, 851 (1934), Ruzicka, Salomon and Meyer, ibid , 17, 882
(1934)

(16) Stoll, Rouve and Stoll-Comte, ibid , 17, 1289 (1934).

(17) Ziegler, Ber., 67, 139 (1934).

(18) Joitrn Am. Chem Soc., 55, :>04, r > (1933).

XXVI. Meta and Para Rings*

This paper deals with rings closed through the m- and p- positions of the
benzene nucleus. Six meta rings were prepared by combining resorcinol
diacetate with

ethylene glycol
trimethylene glycol
tetrametkylene glycol
hexamethylene glycol
nonamethylene glycol and
decamethylene glycol

Five para rings were obtained from hydroquinone diacetate with

ethylene glycol
trimethylene glycol
tetramethylene glycol
hexamethylene glycol and
decamethylene glycol

Again there is formed first a polyester, which gives the required ring by
depolymerization. Mela rings much smaller than 13 atoms and para
rings much smaller tlian 16 atoms are not likely to be obtained. All
samples prepared were pure white sharply melting macrocrystalline
solids. They are easily soluble in most organic solvents.

The closure of rings through the m- and /^-positions of the benzene nu-
cleus is one of the conventional problems of organic chemistry that long
resisted solution. As soon as a rational theory of ring closure had been
developed (1), it became apparent that success might easily be achieved
by the use of either the dilution or the interchange principle. Meanwhile,

* E. W. Spanagel and W. H. Car-others; Journ. Am. Chem. Soc., 57, 935-6 (1935);
Communication No. 154 from the Experimental Station of E. I. du Pont 4e Nemours
and Co.

Received March 21, 1935,

260 POLYMERIZATION AND RING FORMATION

Ruzicka, Buijs and Stoll (2) applied their thorium salt method to the acids
p-C6H 4 [(CH2) B COOH]2 and m-C 6 H 4 [(CH2) 6 COOH] 2 and from the latter
obtained a small amount of the 16-membered cyclic ketone, while Ziegler
and Luttringhaus (3) have applied the dilution method to the nitriles m-
and />-C8H4[O(CH 2 )eCN]2. The cyclic iminonitriles of 18 and 10 atoms
were obtained in good yields.

In the experiments presently reported, the acids m- and -C 6 H4(OCH 2 -
COOH) 2 were estcrified with glycols of the series HO(CH 2 ), 7 OH and the
resulting polyesters depolymerized. In this manner, we have obtained
w-rings (I) of 13, 14, 15, 17, 20 and 21 atoms, in yields ranging from 10 to
35%. In the ^-series (TI) none of the 14- or 15-membered ring was ob-
tained, but those of 1G, IS and 22 atoms were isolated in yields of 12 to
18%.

O CH Z CO O

(CH,),

-O -CH 2 CO -

(I) (ID

Data concerning yields and properties arc shown in Table I.

TABLE I
PROPERTIES AND VSUMMARV OF PREPARATION OF META AND PARA

Poly- Crude Pure M. p of Ring size

ester, Heating, yield, yield, monomer, of

Polyester diacetate K hrs. % % C monomer

Ethylene resorcinol 23 2.5 30 21 100 13

Trimethylene resorcinol 29 3.5 41 24 134 14

Tetramethylene resorcinol 22 5 23 16 112 15

Hexamethylene resorcinol 22 3 45 35 115 17

Nonamethyleiie resorcinol 20 5 45 35 86 20

Decamethylene resorcinol 20 5 55 35 86 21

Ethylene hydroquinone 48 14 9

Trimethylene hydroquinone 35 65 10

Tetramethylene hydroquinone 28 4 42 12 140 10

Hexamethylene hydroquinone 50 7 30 12 124 18

Decamethylene hydroquinone 50 7 38 18 58 22

These results demonstrate clearly the formation and stable existence of
a 13-membered w-ring, and they suggest that in the ^-series a ring of 16
atoms is the smallest possible, but caution must be used in extending or
generalizing these conclusions. On the one hand, it has already been

XXVI. META AND PARA RINGS 261

shown (4) that the replacement of CH 2 by O - in a chain consider-
ably increases the ease of ring formation; thus, in the carbocyclic series
the limits may lie at values different from those obtained for rings contain-
ing two ester groups and two ether oxygens. On the other hand, the fail-
ure to obtain a ring by no means demonstrates that its formation is im-
possible or that it would be very unstable if formed. Nevertheless, in
view of the observed facts and the implications of space models, it seems
that m-rings much smaller than 13 atoms and ^-rings much smaller than 16
atoms are not likely to be obtained.

Experimental

Preparation and Properties of Polymers. Polymeric resorcinol diacetatcs were
prepared by heating equivalent amounts of the acid and glycol at 19O 200 for about
three hours and finally heating the residue to 210 in vacua one or two hours.
The polymers were light brown viscous resins. Nonaniethylciie resorcinol diacetate,
however, crystallized on long standing; the melting point of the crude polymer was
35-40.

Owing to the high melting point of the acid, polymeric hydroquinonc diacetatcs were
prepared by heating equivalent amounts of the glycol and the ester (ethyl hydroquinone
diacetate) with a small crystal of stannous chloride to 190 for about three hours. Final
traces of alcohol were removed by heating the mixture in a vacuum to 210. Ethylene
and trimethylene hydroquinone diacetates are low-melting brown glassy resins. The
remaining polymers were solids which had the following melting points:

Tetramethylene hydroquinone diacetate 45-50

Hcxamcthylene hydroquinone diacetate 50-55

Decamethylene hydroquinone diacetate 00-65

Depolymerization. The polyesters were depolymerized at 270 (1
mm.) in an 800-cc. vapor-heated still or clepolymerizer of the type described
in the preceding paper (5), using 1 to 2% of SnCl 2 .2H 2 O as catalyst.
Rates, yields and melting points are given in Table I. The monomers,
purified by crystallization from alcohol, are all pure white sharply macro-
crystalline solids. They are very soluble in ethyl acetate, benzene, acetone
and ether; moderately soluble in petroleum ether and carbon tetrachloride.
In the latter solvent, the higher members are more soluble than the lower
ones. The compounds are all neutral (i. e., they do not decompose po-
tassium carbonate) , and they react only slowly with a carbon tetrachloride
solution of bromine. Two of the monomers, hexamethyletie hydroquinone
diacetate and hexamethylene resorcinol diacetate, were saponified; the
corresponding dibasic acids were obtained in good yields and identified by
mixed melting points.

2 POLYMERIZATION AND RING FORMATION

Summary
Polyesters prepared from the acids m- and ^-CeH^OCI^COOH^ by

causing them to react with glycols HO(CH 2 ) n OH have been depolymerized.
Cyclic monomers were thus obtained in the w-series, having rings of 13,
14, 15, 17, 20 and 21 atoms. In the ^-series none of the 14- or 15-mem-
bered rings could be isolated, but those of 16, 18 and 22 atoms were readily
formed.

TABLE II
ANALYTICAL DATA FOR META AND PARA RINGS

C H Mol. wt. C H Mol wt rt

Ethylene resorcinol diacctate Decamethylene resorcinol diacetate

Calcd. for Ci,H 12 O 6 57 .14 4 70 252 Calcd. for C 20 H, 8 O fi 65 . 92 7 . 69 364

Found 57.37 4 76 252 Found 66.02 7.75 364

Trimethylene resorcinol diacetate Tctramethylene hydroquiuone diacctale

Calcd. for C,,H u Oc 58.64 5.96 266 Calcd. for C 14 H 16 O 6 60.00 5.71 280

Found 58.69 5.38 260 ^.M 586 268

Tetrarricthylene resorcinol diacetate

r* 1 A f r* ri r\ I'M MM - ~ i OOM Hexamethylcnc hydroquinone diacetate
Calcd. for Ci4Hi 6 O 6 ()() 00 o./l 280 J J H

Found 6004 5.65 290 Calcd. for C,H 20 O 6 6233 6.49 308

Found 62.40 6.84 307

Hexamethylcnc resorcinol diacetate

Calcd. for CieH^Oe 62 33 6 49 308 Decamethylene hydroquinone diacetate

Found 61.96 629 302 Calcd. for C 20 H 28 6 65.92 7.69 364

Nonamethylene resorcinol diacetate Found ^ 7 ' 84 356

Calcd. for Ci 9 H 26 O 6 65 .14 7 . 43 350
Found 65 02 7 31 360

a Determinations made by freezing point in benzene.

Bibliography and Remarks

(1) Journ Am. Chem Soc , 51, 2548 (1929); 54, 1569 (1932); 55, 5043 (1933)

(2) Ruzicka, Buijs and Stoll, Helv. Chtm. Ada, 15, 1220 (1932).

(3) Ziegler and Luttringhaus, Ann., 511, 1 (1934).

(4) Journ Am. Chem. Soc., 55, 5050 (1933).

(5) Ibid , 57, 929 (1935, this volume, page 248).

XXVII. POLYDECAMETHYLENE OXIDE 2(53

XXVII. Polydecamethylene Oxide*

^During the course of the investigations which belong to this series it was
observed] that the depolymerization of polydecamethylene carbonate
gave a rubbery residue. This material was hydrolyzed and gave deca-
methylene glycol, decene-9-ol-l and polydecamethylene oxide. The
latter has a molecular weight of about 1200.

In a previous paper of this series (1) it was shown that polyesters, when
heated under appropriate catalytic conditions in an evacuated vessel
designed for the quick and irreversible removal of volatile material, undergo
the series of transformations represented in the scheme

cyclic monomer

and/or dimer

(volatile)

a-polyester / A

molecular weight ca. \ |

2000-5000 \ |

w-polyester (residual)

molecular weight above

10,000

In the case of the polymethylene carbonates, the end-products under
favorable circumstances were almost all obtained as volatile depolymerizate
and very little residue remained. However, in certain instances in which
it is now believed that the pressure rose too high on account of inadequate
pumping, the residue became rubbery early in the experiments and distilla-
tion soon practically ceased. During the course of a large number of
experiments on the depolymerization of polymeric decamethylene carbon-
ate, a considerable amount of these residues accumulated, and it was
thought worth-while to attempt the recovery of decamethylene glycol from
them by hydrolysis. Comparatively little glycol was recovered in this
way. The major product isolated from the hydrolysis mixture was a
wax-like material which examination showed to be polydecamethylene
oxide (2)

. . . (CH 2 ) 10 O (CHa)ia O (CH 2 )io O (CH 2 ) 10 O. . .

The hydrolysis of the rubbery residue was carried out by means of
alcoholic potassium hydroxide. The resulting solution was filtered from
a small amount of insoluble matter. Most of the alcohol was distilled and

* J. W. Hill; Journ. Am. Chem. Soc., 57, 1131-2 (1935); Communication No. 156
from the Experimental Station of E. I. du Pont dc Nemours and Co.
Received April 26, 1935.
t Compare paper No. Ill on page 29.

264 POLYMERIZATION AND RING FORMATION

replaced by hot water. The insoluble upper layer was allowed to solidify
and separated. It was then heated under vacuum in a Claisen flask until
the distillation of decamethylene glycol ceased. The residue was re-
peatedly extracted with boiling water until no more glycol could be re-
moved in this way. It was then recrystallized once from acetone, once
from alcohol, dried and analyzed.

Anal. Calcd.for(CH 2 )i O: C, 76.92; H, 12.82; mol. wt., 150. Found: C, 7(5.18,
76.31; H, 12.44, 12.37; mol. wl. (cryoscopically in benzene), 1200, 1200.

It was soluble in benzene, carbon tetrachloride, chloroform and con-
centrated sulfuric acid, insoluble in cold alcohol, ethyl acetate and acetone,
and insoluble, hot or cold, in petroleum fractions. From recrystallizing
solvents, it separated as a microcrystalline powder of melting point 58-
60. From a melt it was obtained as a soft wax.

The probable course of the reaction leading to the polyether may be
represented by the following scheme in which it is supposed that the poly-
ester undergoes scission and gives carbon dioxide and decamethylene
oxide radicals. These may rearrange to decene-9-ol-l or combine to give
the polymeric ether.

-C0 2

( O (CH 2 ) 1(r - OCO ) H >

a-poly decamet hy lene
carbonate

w[ (CH 2 ), O] > ( (CH 2 ) 10 0-) M

CH 2 CH (CH 2 )iOH polydecamethylene oxide
decene-9-ol-l

Support for this mechanism is the fact that in the depolymerization ex-
periments yielding considerable polyether residues, the distillates were
always contaminated with decene-9-ol-l (3).

The constitution of the material was established by its analyses and
molecular weight and by its transformation by the following reactions to
decamethylene glycol.

HI KG Ac

( (CHOur-O ) > I(CH 2 ) 10 I >

HoO
AcO(CH 2 )i OAc > HO(CH 2 ) ]0 OII

The intermediate compounds were not purified.

Twenty grams of recrystallized polymer, 40 cc. of glacial acetic acid, and
75 cc. of constant boiling hydriodic acid were refluxed together for six
hours. The resulting dark red lower layer was separated and refluxed
briefly with a suspension of calcium carbonate which removed the color.

XXVIII. MACROCYCLIC LACTONES BY DEPOLYMERIZATION 265

The heavy oil was recovered by means of an ether extraction and then re-
fluxed with 40 g. of powdered freshly fused potassium acetate, 15 cc. of
glacial acetic acid and 3 cc. of acetic anhydride for four hours. This mix-
ture was drowned in water and the product recovered by ether extraction.
This was then hydrolyzed by refluxing for ten hours with a solution of 15
g. of sodium hydroxide in 150 cc. of 50% alcohol. The alcohol was re-
moved by steam distillation which was continued for several hours to re-
move a very small amount of oil, which smelled like decene-9-ol-l. The
oil on top of the residual liquid solidified and from it was isolated, by two
crystallizations from benzene, 5 g. of pure decamethylene glycol, m. p.
71-72. It showed no depression in melting point when mixed with a
known sample of decamethylene glycol.

Summary

The depolymerization of polydecamethylene carbonate, under certain
experimental conditions, pursues an abnormal course and yields a distillate
contaminated with decene-9-ol-l together with a large amount of residue.
This residue has been examined and found to consist largely of a poly-
meric substance which has been characterized as polydecamethylene oxide.
A mechanism for the side reaction has been suggested.

Bibliography and Remarks

(1) Journ. Am. Client Sac., 55, 5031 (1933)

(2) It is of interest to mention that Franke \Monatsh , 53-54, .177 (1929), and earlier papers]
obtained, by the dehydration of decamethylene glycol with sulfuric acid, not a decamethylene oxide,
but tt-amyl pentamethylene oxide.

(3) Journ. Am. Chem. Soc., 55, 5037 (1933)

XXVIII. Preparation of Macrocyclic Lactones by
D epolymerization *

The depolymerization method^ is used in this publication to prepare
macrocydic lactones. A yield of 70% of pure monomer could be obtained
from the polyester of hydroxy tetradecanoic acid. Depolymerization
was carried out with the aid of catalysts, mostly hydrous magnesium
chloride at 270 in vacua. The lactones of the following acids were pre-
pared:

* E. W. Spanagel and W. H. Carothcrs; Journ. Am. Chem. Soc., 58, 654-6 (1936);
Communication No. 16(5 from the Experimental Station of E. I. du Pont de Nemours
and Co.

Received January 18, 1936.

t Compare e. g. paper No. XXV on page 248.

266 POLYMERIZATION AND RING FORMATION

14-hydroxy-12-oxatetradecanoic (mono and dimer)
1 6-hydroxy-l 2-oxahexadecanoic (monomer) .

Lactones of the higher co-hydroxy fatty acids were first successfully syn-
thesized by oxidizing the corresponding cyclic ketones (1). Preparation
by the conventional methods is impossible because the acids tend to react
intermolecularly yielding linear polyesters. Once this fact was clearly
recognized it became evident (2) (p. 8 . . .) that direct lactonization might
be favored by application of the dilution principle first utilized by Ruggli
(3) ; and Stoll and Rouve (4) have showed experimentally that this is in-
deed the case.

A third method for the preparation of large ester rings consists in de-
polymerizing the corresponding linear polyester. This was first realized
with esters of carbonic and oxalic acids (5). Later it was shown that, by
the proper control of temperature and selection of catalysts, excellent re-
sults could also be obtained with glycol esters of other dibasic acids (0).

We have now applied this method to the preparation of several lactones.
The possibilities are indicated by the fact that a 70% yield of pure mono-
mer was obtained from the polyester of hydroxytetradecanoic acid. Stoll
and Rouv (4) in a similar case report yields by the dilution method con-
siderably higher than this, but their scale of operation was small (10 g.),
the volume large (10 liters) and the time long (six days). Our depoly-
merization required only three and one-half hours for 32 g. of polyester.
It appears therefore that this method has some advantages over the use of
high dilution.

One of the objects of the present work was to obtain some lactones which
had not been reported previously. Meanwhile Stoll and Rouve have pub-
lished a much more extended study (4) of lactonization by the dilution
method covering in part the same ground. The overlapping results of the
two investigations are in good agreement, but in several cases we are able
to report slightly higher melting points than were found by Stoll and
Rouve\

As new compounds we report

r O(CH 2 ) 2 0(CH 2 ) 10 CO- 1
j-O(CH 2 )4O(CH 2 )ioCO- 1 and
r O(CH 2 ) 2 O(CH 2 ) 1 oCOO(CH 2 ) 2 0(CH 2 ) 10 CO^

The first two are rings of 15 and 17 atoms and their odors are definitely
musk-like. These compounds are also of interest because the required hy-

XXVIII. MACROCYCLIC LACTONES BY DEPOLYMERIZATION 267

droxy acids are relatively easily accessible from undecylenic acid through
the addition of hydrogen bromide followed by reaction with the sodium
derivative of the glycol.

The greater ease of forming rings of large size Compared with inter-
mediate size is again illustrated in the present study by the fact that hy-
droxytetradecanoic acid gave monomer and dimer in the ratio of 15/1
while for hydroxyclecanoic acid the corresponding ratio was 0.17/1. Fur-
ther evidence of specificity in the catalysis of ester depolymerizations was
also found. Whereas in most of the cases previously examined, stannous
chloride and various magnesium salts were almost equally effective, in the
present work magnesium chloride was much more effective than any other
material tested (cf. Table I).

TABLE I
EFFECT OF CATALYST ON YIELDS

Catalyst

MgCl 2 .6H 2 O

MgO

MnCl 2 .4H 2 O

SnCl 2 .2H 2

CoCl 2 .6H 2

Mg (powd.)

Remarks
No reflux

Steam cooled reflux
No reflux
No reflux
No reflux
No reflux
No reflux

Crude
yield,

%of

distillate
dimer

%of
distillate
monomer

TABLE II
EFFECT OF TEMPERATURE ON YIELDS

Temp., C. Time, hours % crude yield % dimer

270 2 73 26

260 2 30 12

260 5 58 15

250 2 28 10

250 5 52 12

Experimental Part

Polyesters were prepared by heating the hydroxy acids at 180-250 for three or four
hours in a Claisen flask without added catalyst but with diminished pressure during the
last hour.

Depolymerizations were carried out in the vapor heated still already described (6) at
a pressure of 1 mm. or less. In some experiments a dephlegmator cooled with steam was
inserted at the top of the still. It lowered the rate of distillation and increased the ratio
of monomer to dimer.

Catalysts were explored using polyester (15 g.) from hy droxy undecanoic acid with
0.2 to 0.3 g. of the proposed catalyst at 270 during two hours. The distillate was dis-

268

POLYMERIZATION AND RING FORMATION

rt rj

Sfc~
tn rt fl

ail

-s 65

w rt O

'Ss

O CD O CD
O5 CD O I-

o co

iO iO

I I

-f-

'-1^

O 1- CD

o o

t- 10

*!j

O t"-

p ^ >

^ o

00
CD

s l

o KL;

Ipj

->o

10 rN,

6 1

f^i

tt ^| Q

O clo
CD CD

O O Q

't CO -t

~t CO

fc *S.s

""O

. p

W g WJ K

6 tf

o- rt

O h" <c '

^ ^ ^o

CD 0-1

d C/3

*fe

o 9. Q.

O O

S S Ifv

S '

O ^

ffl

hH W W

PQ c/3 Q
< W *o

Q ~ v

W CD CD

CD tD

H P

1^ ^

>4 -2*O

e->

C ?l

<M -1

^ co

M r^ rt

n o o

O CJ

^ CO

U b/Q bC

be be

"<2 rH

co S S

^ ^

< OP

^ T _ H

~^^

-1^

^ ^~^

CL H
0,oU

C o
g

o" -\

rH t^.

CO CD

CT)

. ^

rH C^-l

r^ ~

^i W)

rH rH

' r ~ |

_. in

1*2

LO (M O
O5 CO l^~

HH CD

" t/

^ ^

did

CD CO

cT rt"

. G o

P *~*

s l

O GJ

i ^

tf) 1 o

rP C

s r! w

h-" hr<

o a ' c a

rH <M rt<

I-

2 ^

Polyester from

10-Hydroxydecanoic acic

13-Hydroxytridecanoic a
14-Hydroxytetradecanoi<
HO(CH 2 ) 2 O(CH 2 ) )0 COO1

HO(CH 2 ) 3 O(CH 2 )i COOI
HO(CH 2 )4O(CH 2 ) 10 COOI

At
i
m<
m
Monomeric lactone from ri

10-Hydroxydecanoic

acid 1
1 1-Hydroxyundeca-
noic acid a - c ]
13-Hydroxytrideca-
noic acid ]

14-Hydroxy-12-oxa-
tetradecanoic acid ]

15-Hydroxy-12-oxa-
pentadecanoic acid c \

l6-Hydroxy-12-oxa-
hexadecanoic acid ]

C/. Stoll, Rouve and 5
50 (1929); Carothersand

XXVIII. MACROCYCLIC LACTONES BY DEPOLYMERIZATION 269

solved in 50 cc. of alcohol and the crystalline dimer which separated on cooling was
filtered off and weighed. The accumulated nitrates on distillation were shown to be
largely monomer, and in the last column of Table I the indicated yields are based on the
assumption that all not dimer was monomer.

The effect of temperature is indicated in Table II. In each case 20 g. of polyester
from hydroxyundecanoic acid was heated with 0.5 g. of MgCl 2 .6H 2 O. The ratio of
monomer to dimer rises as the temperature falls, but the rate falls off quite rapidly at the
same time.

The results for the depolymerization of polylactone derived from 10-hydroxy-
decanoic, 13-hydroxytridecanoic, 14-hydroxytetradecanoic acid and the ether acids are
listed in Table III. Properties and analyses are indicated in Tables IV and V.

TABLE V
ANALYTICAL DATA

C H Mol wt "

Lactone of 10-hydroxydecanoic acid

Calcd. for C,oH,,sO> 70.58 10.58 170

Found 70.31 10.13 162

Dimcric lactone of 13-hydroxytridecanoic acid

Calcd. for C 26 H 18 O 4 73 24 11.26 426

Found 73.46 11.30 407,432

Dimeric lactone of 14-hydroxytetradecanoic acid

Calcd. for C 2 sH fi2 O, 74 33 11 .50 452

Found 74.29 11.79 414

Lactone of 14-hydroxy-12-oxatetradecanoic acid

Calcd. for CiaHajOs 68.42 10.52 228

Found (38 57 10.17 232

Dimeric lactone of 14-hydroxy-12-oxatetradecanoic acid
Calcd. for C>6H 4 sO 6 68 . 42 10. 52 456

Found 68.16 10.36 466

Lactone of 15-hydroxy-12-oxapentadecanoic acid

Calcd. for Ci 4 H 2l O 3 69 . 42 10 . 74 242

Found 69.76 11.00 234

Lactone of 16-hydroxy-12-oxahexadecanoic acid

Calcd. for Ci6H 2s O3 70.3 10.93 256

Found 70.31 11.01 258

a Determinations made in freezing benzene.

Preparation of Materials

11 -Hydroxyundecanoic was prepared from undecylenic acid and hydrobromic
acid following Ashton and Smith (7) . Our best yields were 72 % of the theoretical. Con-
version of this to 11 -hydroxyundecanoic acid was carried out as follows: a solution of
60 g. (1.5 rnolcs) of sodium hydroxide and 132 g. (0.5 mole) of the bromo acid in one liter

270 POLYMERIZATION AND RING FORMATION

of water was refluxed for five hours. The filtered solution was cooled to about 5 for
several hours where upon the sodium salt of the acid separated. It was filtered off and
redissolved in hot water. Acidification liberated the hydroxy acid as an oil which
crystallized when cooled. One recrystallization from benzene gave 74 g. (72% of the
theoretical amount) of fine crystals melting at 70.

The Ether Hydroxy Acid HO(CH 2 )2O(CH 2 )i COOH. One mole of sodium (23 g.)
was dissolved in 248 g. of ethylene glycol. The mixture was then heated to 110-115
under reflux and stirred while 133 g. (0.5 mole) of 11-bromoundecanoic acid was added
in four portions during one hour. Stirring and heating were continued for four hours
and most of the glycol was removed by further heating in vacua . The residue was dis-
solved in water, acidified and extracted with ether. The crude acid after crystallization
from petroleum ether was obtained as a white powder melting at 48-50.

Anal. Calcd. for Ci 3 H 26 O 4 : C, 63.41; H, 10.56; neut. equiv., 246. Found: C,
63.27; H, 10.36; neut. equiv., 244.

The ether hydroxy acid HO(CH 2 )3O(CH 2 )ioCOOH was obtained similarly, m.
p. 51.

Anal. Neutral equivalent calcd. for Ci 4 H 2 8O 4 : 260. Found: 263.

The ether-hydroxy acid HO(CH 2 ) 4 O(CH 2 ) 10 OOH, m. p. 53.

Anal. Calcd. for Ci 5 H 30 O 4 : C, 65.69; H, 10.95; neut. equiv., 274. Found: C,
65.39; H, 11.01; neut. equiv., 281.5, 281.3.

Summary

The interchange method for preparing macrocyclic esters is shown to be
applicable to the preparation of many-membered lactones and to have ad-
vantages in speed and simplicity over the high dilution method.

Three new large cyclic esters are described: the monomeric and dimeric
lactones of 14-hydroxy-12-oxatetradecanoic acid and the monomeric lac-
tone of 16-hydroxy-12-oxahexadecanoic acid.

Bibliography and Remarks

(1) Ruzicka and Stoll, Helv Chim. Acta, 11, 1159 (1928).

(2) Carothers, Journ. Am. Chem. Soc., 51, 2548 (1929).

(3) Ruggli, Ann., 392, 92 (1912).

(4) Stoll and Rouv, Helv. Chim. Acta, 17, 1283 (1934).

(5) Hill and Carothers, Journ. Am. Chem. Soc., 55, 5031 (1933).

(6) Spanagel and Carothers, ibid., 57, 929 (1935).

(7) Ashton and Smith, J. Chem. Soc., 1308 (1934).

PART TWO

ACETYLENE POLYMERS
AND THEIR DERIVATIVES

PART TWO

ACETYLENE POLYMERS AND THEIR DERIVATIVES

I. Introduction

This introduction, by drawing to the reader's attention the salient and
most significant features of Carothers' series of papers on Acetylene
Polymers and their Derivatives, will, it is hoped, add to his satisfaction
in perusing the papers themselves, which are reprinted hereafter.

The outstanding element of interest and importance in the work de-
scribed in the series is the discovery of a synthetic rubber, polychloroprene,
which has been manufactured since 1933 and represents one of the most
striking achievements of industrial chemistry in recent years. In addition
however, to providing the basis for this industrial development, the work
also makes a brilliant and important contribution to our knowledge of
(1) the influence of substitution on the polymerization of conjugated
systems and (2) the chemical behaviour of conjugated systems in which
an ethylenic and an acetylenic bond are disposed in a conjugated relation-
ship to one another.

In the first paper of the series Acetylene Polymers and their Derivatives,
by Nieuwland, Calcott, Downing and Carter (1) there is described the
discovery that, by means, as a catalyst, of cuprous chloride (preferably
together with ammonium chloride) acetylene can be polymerized in high
yield to (1) a dimer, vinyl acetylene, CH 2 :CH.C:CH, b. p. 5 and (2) a
trimer, divinylacetylene, CH 2 :CH.C:C.CH:CH 2 , b. p. 83.5. With
these two highly reactive substances thus made readily available, Carothers
and his co-workers proceeded to an extensive study of their chemistry,
the results of which are described in subsequent papers of the same series.

Chloroprene

Of outstanding interest was the discovery that 2-chloro-l,3-butadiene,
CH 2 : CC1 . CH : CH 2 , obtainable by the action of aqueous hydrogen chloride
on vinyl acetylene under appropriate conditions in the presence of cuprous
and ammonium chlorides, polymerizes readily to a rubber-like product
(Paper II, p. 281). By analogy with isoprene, the compound is commonly
referred to as chloroprene. It polymerizes far more rapidly than does
isoprene, thanks to the activating effect of the /3-chlorine atom, and, con-
publications listed in the Bibliography, p. 423, el seq., are referred to by Reference
numbers in large Roman corresponding to the Serial Numbers of the papers.)

274 ACETYLENE POLYMERS AND THEIR DERIVATIVES

trary to what might perhaps have been expected (by, e. g., analogy with
poly vinyl chloride) the product turns out to be highly elastic and indeed
similar in elastic properties to vulcanized natural rubber. By inter-
rupting the polymerization before it is complete, it is possible to isolate
an intermediate polymeric material (the so-called a-polymer) which,
being plastic, can be handled by much the same techniques as raw rubber
and its stocks are handled, and which having thus been fabricated into
rubber goods of various shapes, can be converted into the final elastic
material (the so-called ju-polymer) by the further application of heat,
which completes the polymerization.

Polychloroprene is now being manufactured on a substantial scale, be-
cause, despite the fact that it is more costly than natural rubber, it has
special properties which make it better adapted than the latter to certain
applications. Chief of these properties is its lower imbibition of aliphatic
hydrocarbons, which, presumably, is dependent on its high content of
chlorine. It is, further, more stable than vulcanized rubber. This sta-
bility is exhibited in its resistance to deterioration on aging, by heat, by
ozone, etc., and is probably due largely to the fact that in it the poly-
merization process to which it owes its formation is substantially complete,
whereas vulcanized rubber is susceptible to further polymerization or if
views now in vogue as to the nature of vulcanization are adopted to
further cross-linking.

Under certain special conditions chloroprene yields a polymer of different
properties, distinguished as the w- or granular polymer. It is lacking in
plasticity and imbibes organic liquids hardly at all. It can be produced,
for example, by exposing chloroprene vapour to light from a mercury arc,
and after its formation has been thus initiated, it will without further
irradication continue to grow in the vapour or in liquid chloroprene.
Carothers (2) in some interesting comments calls attention to the curious
and perhaps suggestive analogy of such a process to vital growth.

Influence of Substituents on Polymerization

Carothers' studies afford much valuable information as to the influence
of substitution in the butadienoid system on the readiness with which
polymerization occurs and the degree to which it proceeds. Considering
first halogen substitution: chloroprene, with chlorine in the 2-position,
polymerizes about 700 times as quickly as isoprene and yields a product
of excellent rubber-like properties; the analogous bromo derivative
(Paper V, p. 314) polymerizes somewhat more rapidly, but the product,
although having good extensibility and strength, shows, presumably owing

I. INTRODUCTION 275

to the high percentage by weight of halogen in the molecule, less com-
pletely recovery from extension than does polychloroprene, and the lower
limit of the temperature range within which it is elastic is higher than for
polychloroprene; 2-iodo-l,3-butadiene (Paper XV, p. 361) polymerizes at
a markedly higher rate than either of the preceding, but the product
(which contains 70 per cent of iodine) is rubber-like only under certain
conditions (3).

The introduction of a second chlorine atom to one of the interior carbon
atoms of the butadienoid system yields a compound CH 2 : CC1 . CC1 : CH 2 ,
which polymerizes even more rapidly than chloroprene CH 2 : CC1 . CH : CH 2
(Paper XI, p. 340). The product, either because of the high degree to
which polymerization has proceeded or because of its high chlorine content,
is a hard tough mass lacking in rubber-like properties. (The high speed
of polymerization of CH 2 :CC1. CC. :CH 2 is perhaps somewhat unexpected,
if one were to judge by analogy with the effect of chlorine substitution in
the ethenoid system, for, whereas here the introduction of one atom of
chlorine, which renders the system highly polar, gives a compound (vinyl
chloride) which is highly polymerizable, the introduction of a second atom
of chlorine, symmetrically disposed with respect to the first, gives a com-
pound (sym.-dichloroethylcnc) in which the ability to polymerize has been
lost. It is, however, to be noted that in 2,3-dichloro-l,3-butadiene un-
substituted terminal :CH 2 groups are present and that the presence of
such may well be a condition for chlorine to exert an " activating" effect
on polymerizability.) The substitution of chlorine in the terminal posi-
tions of the butadiene has no such striking effect on polymerization as its
substitution in the interior positions. l-Chloro-l,3-butadiene, C1CH:-
CH.CH:CH 2 , polymerizes not much more quickly than isoprene and the
product has only inferior rubber-like properties (Paper XI, p. 340) (3).
Similarly, l,2,3,-trichloro-l,3-butadiene, C1CH : CC1 . CC1 : CH 2 polymerizes
much more slowly than 2,3-dichloro-l,3- butadiene, CH 2 : CC1 . CC1 : CH 2
(Paper XI, p. 340).

Observations on a number of chlorine compounds prepared from divinyl-
acetylene by steps involving the addition and the removal of the elements
of hydrogen chloride provide further data on the influence of chlorine-
substitution on the polymerization of conjugated systems (Paper XIII,
p. 348). The compounds in question, each containing a conjugated
system, are as follows:

1. CH 2 :C:CC1.CC1:C:CH 2
: 2. CH 2 :C:C:CC1.CH:CH 2

276 ACETYLENE POLYMERS AND THEIR DERIVATIVES

3. Cl CH 2 . CH : C : CC1 . CH 2 : CH 2

4. C1CH 2 .CH:CC1.CC1:C:CH 2

5. Cl CH 2 . CH : CC1 . CC1 : CH . CH 2 C1

Of these, 1 showed a strong tendency to polymerize: it changed to a
hard resin when allowed to stand for 24 hours; 2 and 3 polymerized only
slowly (3 changed to a viscous syrup on standing for three months); 4
and 5, in which one end or both ends of the molecule carries chlorine and
in which a terminal vinyl group is lacking, showed no tendency at all to
polymerize.

The effect of introducing other groups into the 2-chloro-butadienoid
system, especially their introduction into the terminal positions, is to
depress the poly merizability . 3-Methyl-2-chloro- 1 ,3-butadiene, CH 2 : -
CC1.C(CH 3 ):CH 2 , polymerizes somewhat more slowly than chloroprene,
and the polymer is somewhat inferior to polychloroprene in rubber-like
properties: although tough and elastic, it has a rather low extensibility
(Paper XXI, p. 384). The introduction to the last-mentioned butadiene
of a second, terminal methyl group has a more profound effect, 3-4-di-
methylchloroprene, CH 2 :CC1.C(CH 3 ) :CH.CH 3 , polymerizing not much
more rapidly than isoprene and yielding a produce which, although elastic,
recovers only sluggishly from deformation (Paper XXI, p. 384). As con-
trasted with the relatively slight effect of a 3-methyl group on the poly-
merization of chloroprene, a terminal methyl group, as in 1-methylchloro-
prene, CH 3 CH : CC1 . CH : CH 2 , reduces the speed of polymerization greatly
and detracts markedly from the rubber-likeness of the polymeric product
(Paper IX, p. 331). Increase in the magnitude of the terminal alkyl
group magnifies the effect in the same direction, w-heptylchloroprene,
CyHis . CH : CC1 . CH : CH 2 , yielding on polymerization only a sticky solid
with very slight elasticity (Paper IX, p. 331). Such observations are in
accord with previous findings that the introduction of terminal alkyl
groups to butadiene reduces polymerizability and that, the larger the
alkyl group, the greater is the reduction (4).

In papers of the Acetylene Polymers series published subsequent to
Carothers' association with it, the influence of certain other substituents
on the polymerizability of chloroprene is recorded. The introduction of
a dialkylmethylamino group in a terminal position depresses the poly-
merization of the chloroprene system very greatly, the 1- (dialkylmethyl-
amino) -chloroprenes, R 2 N . CH 2 . CH : CC1 . CH : CH 2 , polymerizing very
slowly indeed and yielding no rubber-like product (5). Alkoxy-methyl or
alkoxy- (methyl) methyl groups in the terminal position have some but a

I. INTRODUCTION 277

far smaller effect on the polymerizability of chloroprene, CHj>O.CH 2 .-
CH : CC1 . CH : CH 2 polymerizing at one-sixth the rate of chloroprene and
yielding a product with somewhat rubber-like properties (6).

In view of the profound effect on polymerization of the presence of
chlorine in the 2-position of the butadienoid system, the effect of other
substituents in this position was examined. As compared with isoprene,
which has a methyl group in the 2-position, 2-heptyl-l,3-butadiene was
found to polymerize rather less readily and to yield a weaker and softer
rubber (Paper XVI, p. 368). 2-Phenyl-l,3-butadiene was found to poly-
merize considerably more rapidly than isoprene but to yield under ordinary
conditions only a dimeric product (Paper XVI, p. 368). In continuation
of Carothers' studies, the effect of introducing (a) alkoxy groups, (b) ali-
phatic acid residues with the 2-position has been examined. 2-Ethoxy-
1,3-butadiene, CH 2 :C(OC 2 H5).CH:CH 2 , was found to polymerize at
about one-third the rate of isoprene and to give a product of inferior
rubber-like properties (7) . 2-Acetoxy-l,3-butadiene, CH 2 . C(O . COCH 3 ) . -
CH:CH 2 , and analogous compounds on the other hand were found to
polymerize at a rate not vastly different from that of chloroprene and to
yield rubber-like products (8).

4-Cyano-l,3-butadiene, CH 2 :CH.CH:CH .N;C, despite the fact that
the substituent group is in a terminal position, has been found to polymerize
about twenty times as quickly as isoprene (9). Perhaps the extra conju-
gation which the cyano-group provides is responsible for this result. The
final product of polymerization, although rubber-like, is markedly inferior
to polychloroprene or polyisoprene : it could be stretched 1300 per cent
without breaking, but ten minutes after release still had an extension of
200 per cent.

Chemical Behaviour of Vinylacetylene and Divinylacetylene

Turning attention now to the acetylene polymers, vinylacetylene and
di vinylacetylene, note may be made of the salient features of observations
regarding (a) the polymerization of these substances, (b) their addition
reactions other than self -addition (i. e., polymerization), (c) their other
reactions.

Polymerization. In the presence of cuprous chloride, vinylacetylene
gives a liquid dimer, which is probably octatriene-l,5,7-ine-3, CH 2 :CH.-
CiC.CH:CH.CH:CH 2 (10, 11), and which, on exposure to air, becomes
converted into an insoluble product which explodes violently when
heated (11). When heated in the absence of a catalyst, as, e. ., at 105
for six hours (11) or in the presence of peroxide catalysts (10), vinyl-

278 ACETYLENE POLYMERvS AND THEIR DERIVATIVES

acetylene polymerizes first to viscous oils and finally to hard, resinous
solids. The available evidence (11) suggests that these polymers, which
contain acetylenic hydrogen, are cyclobutane and cyclobutene derivatives,
as follows :

: C.CH . CH.C: CH, CH: C.CH . CH.

II' II

CH 2 CH 2 CH 2 .CH 2

/C : CH\
(I I )

\CH 2 CHA.CiCH,

their formation involving only the ethylenic linkages of vinylacetylene.

When heated in the presence of acids, acid anhydrides or phenols, vinyl-
acetylene, in addition to forming resinous products, yields, unexpectedly,
styrene in amount corresponding to about one-half of the vinylacetylene
used (11). Styrene, C 8 H 8 , in this connection figures as a dimer of vinyl-
acetylene, C 4 H 4 . Its formation can be represented as, analogously to the
dimerization of butadiene, involving first the addition, in a 1,4-sense, of
the acetylenic portion of one molecule of vinylacetylene to the conjugated
system of another, and then a rearrangement of bond. Thus:

C CH C=-C CH CH

^ \ / \ / \

CH CH 2 CH CH 2 CH CH

+ > \ / * \ /

CH^C CH=C CH=-CH

I I I

CH=CH 2 CH CH 2 CH=CH 2

Divinylacetylene, when heated in the absence of air (in the presence of
air it absorbs oxygen and becomes dangerously explosive) polymerizes to
first a viscous liquid and finally a hard, brittle, insoluble resin. The inter-
mediate, oily material, when exposed as a film to air, takes up oxygen and
"dries" to a hard, insoluble product (12) (Paper XIX, p. 378). It
has been referred to as a "synthetic drying oil." From the inter-
mediate, oily polymeric material it was possible to isolate a dimer
of vinylacetylene, which was shown to be divinylethinyl-l,2-cyclobutane,
CH 2 : CH . C : C . CH . CH . C i C . CH : CH 2 (Paper XIX, p. 378), its formation

I I
CH 2 CH 2

depending as in the case of the analogous dimer of vinylacetylene (supra)
on the mutual addition of an ethylenic part of each of two molecules of the
monomer.

The work provides some information regarding the polymerization of
certain substitution products of vinyl- and divinyl-acetylene. l-Halo-2-
vinyl acetylenes, CH 2 :CH.CH:C Hal., on standing for several months,
polymerize to black, highly explosive solids (Paper XVIII, p. 375).

I. INTRODUCTION 279

l-Alkyl-2- vinyl acetylenes, CH 2 : CH . CH ; CR, polymerizes only slowly;
they become viscous syrups after 2-3 months' standing (Paper VIII, p. 329) .
Replacement of the acetylenic hydrogen of vinylacetylene by a secondary
carbinol residue has a favourable effect on the speed of polymerization,
the tertiary carbinols all polymerizing spontaneously with considerable
rapidity (Papers VI, p. 321; VII, p. 323). Thus, e. g. t (vinylethinyl)
methylethyl carbinol, CH 2 :CH.C:C(OH)(C 2 H 5 )CH3, on standing in a
stoppered bottle, had become a thick syrup in one week, then passed
through a tough elastic, rather rubberlike stage, and in six weeks had
become a very hard mass (Paper VII, p. 323).

Addition Reactions. Prior to Carothers' study of the chemistry of
vinyl- and divinyl-acetylenes little was known about the additive be-
haviour of -enine systems or, alternatively, enyne systems in which
contiguous carbon atoms carry respectively double and triple bonds, thus :
-.C:C.C : . C. Can such a system of fonds behave as a conjugated
system? Yes; for it was shown rather conclusively that the first result
of treating vinylacetylene with hydrochloric acid is to form 4-chloro-l,2-
butadiene (P^per III, p. 306).

HC;C:CH:CH 2 + HC1 > H 2 C : C : CH . CH 2 C1

This compound, which can readily be isolated, can be put through a series
of methathetical reactions without any rearrangement of the bonds
(Paper XV, p. 361). Thus, e. g., by means of hot water, preferably in the
presence of a mild alkali, the chlorine is replaced by a hydroxyl group
(Paper XV, p. 361; U. S. Patent 2,073,363, p. 431; 2,136,178, p. 432).
It readily rearranges, however, under many conditions of treatment, espe-
cially in the presence of cuprous chloride, and gives 2-chloro-l,3-butadiene,
CH2 : CC1 . CH : CH 2 , i. e., chloroprene, the allene arrangement changing
to the more stable conjugated one, in which the chlorine atom, attached
to a doubly bonded carbon, has lost its mobility (Paper III, p. 306).
Carothers discusses the mechanism of the a,7-shift involved and concludes
that it is dependent, not on ionization, but on the formation of a complex
(coordination compound) and chelation (Paper XV, p. 361).

In the presence of a little sodium alcoholate, alcohols readily react with
vinylacetylene to yield acetylene ethers of the type, RO . CH 2 . C : C . CH 3
(Paper XX, p. 381). This reaction is considered to involve first 1,4-
addition, and then a rearrangement similar to the rearrangement of mono-
substituted allenes to disubstituted acetylenes which Favorsky found to
be produced by sodium alcoholates. Thus :

CH 2 :CH.C'CH, i. e., .CH 2 . CH:C:CH. > (RO.CH 2 . CH:C:CH 2 ) >

RO.CH 2 .C:C.CH 3

280 ACETYLENE POLYMERS AND THEIR DERIVATIVES

When two atoms of chlorine are added to divinylacetylene, the product
is reasonably concluded to have the structure C1CH 2 . CH : C : CC1 . CH : CH 2 ,
and hence to have been formed by 1,4-addition. When to this product
two more atoms of chlorine are added, the tetrachloride formed has ap-
parently the structure C1CH 2 . CH : CC1 . CC1 : CH . CH 2 . Cl, and to be found
by a further 1,4-addition (Paper XIII, p. 348). This last conclusion
makes it appear that in a system . C:C:C.C:C. a carbon atom with
twinned double bonds can act as one of the ends of a conjugated system
(cf. the remarks on the polymerization of chlorine compounds derived
from divinylacetylene, p. 378).

It is to be remarked that some of the chlorides derived from divinyl-
acetyelene and also certain derivatives of vinylacetylene, despite the fact
that they contain conjugated double bonds, fail to react with maleic
anhydride or naphthoquinone (Paper XIII, p. 348).

Although in most of their addition reactions vinyl- and divinylacetylenes
behave as conjugated systems, two exceptions were encountered. (1) In
reaction with thiophenols, only the ethylenic bond of the -enine system is
involved, and in view of this Carothers made considerable use of thio-p-
cresol in studying the constitution of acetylene polymers and their deriva-
tives. Divinylacetylene readily adds, at room temperature, two molecules
of thio-/?-cresol to give a crystalline product (Paper XII, p. 344), thus:
CH 2 :CH.CiC.CH:CH 2 ->C 7 H 7 S.CH 2 .CH 2 .C:C.CH 2 .CH 2 S.C 7 H 7 . (2)
A reaction dependent on addition to the acetylenic bond of vinylacetylene
is the formation of 1,3-butadienyl esters by treating vinylacetylene with
an aliphatic acid in the presence, as catalyst, of a mercuric salt or boron
fluoride (13). Thus:

CH 2 :CH.C;CH + R.COOH > CH 2 :CH.C(O.COR) :CH 2

Other Reactions. Reactions of vinylacetylenes dependent on the re-
activity of the acetylenic hydrogen (CH 2 : CH . C : . CH) are as follows:

(a) With Grignard reagents, vinylethinyl magnesium halides, e. g.,
CH 2 : CH . C i CMgBr, are formed, and the latter react in a way typical
of organo-magnesium compounds, giving, with acetone, a carbinol,
CH 2 :CH.C:C(OH)(CH 3 ) 2 , and with carbon dioxide, an acid, vinyl-
propiolic acid, CH 2 :CH.C : -C.COOH (Paper VI, p. 321).

(b) With sodium or sodamide, there is formed sodium vinylacetylide,
CH 2 : CH . C : CNa, which is very reactive; giving with aldehydes secondary
carbinols, CH 2 :CH.C;C. (OH)CHR; with ketones, tertiary carbinols,
CH 2 :CH.C:C(OH)CRR / (Paper VII, p. 323), and with alkyl halides,
alkyl-vinylacetylenes, CH 2 :CH.C:CR (Paper VIII, p. 329).

II. A NEW SYNTHETIC RUBBER 281

(c) With potassium mercuric iodide or with mercuric acetate in acetic
acid, vinylacetylene yields di-vinylethinyl mercury (CH 2 :CH.C:C)2Hg
(Paper XVII, p. 372).

(d) Treatment of vinylacetylene with alkali hypohalides replaces
the acetylenic hydrogen by halogen and yields the 1 -halo-vinylacetylenes,
CH 2 :CH.C;C Hal. (Paper XVIII, p. 375).

(e) By treating vinylacetylene with formaldehyde and a secondary
amine, the acetylenic hydrogen is replaced by a secondary-amino-methyl
residue (14,15), thus:

CH 2 :CH.C;CH + CH 2 O + HNR, > CH 2 :CH CiC.CH 2 NR 2 + H,O

Bibliography and Remarks

(1) J. A. Nieuwlancl, U. S. Pat. 1,811,959 (1931), Journ Am. Client. Soc , 53, 4197-4202 (1931)

(2) W. H. Carothers, Trans. Farad. .Sac., 32, 39 (193(5).

(3) W. H. Carothers, Ind. Eng. Chem , 26, 30 (1934).

(4) Whitby and Gallay, Can. Jour Res., 6, 280 (1932); Whitby and Cro/ier, ibid., 6, 203 (1932)

(5) D. D. Coffman, Journ. Am. Chem. Soc., 57, 1978 (1935)

(6) H. B. Dykstra, ibid., 58, 1747 (1936)

(7) H. B. Dykstra, ibid., 57, 2255 (1935).

(8) J. H. Werntz, ibid , 57, 204 (1935)

(9) D. D. Coffman. ibid., 57, 1981 (1935)

(10) Nieuwland, Calcott, Downing and Carter, ibid., 53, 4197 (1931).

(11) H. B. Dykstra, ibid , 56, 1025 (1934)

(12) Nieuwland, Calcott, Downing and Carter, ibid , 53, 4197 (1931)

(13) J. H. Werntz, ibid., 57, 204 (1935).

(14) D. D. Coffman, ibid., 57, 1978 (1935); also II V S Patent 2,136,177
CIS) Journ. Am Chem. Soc., 53, 4197 (1931).

G. vS. W.

II. A New Synthetic Rubber: Chloroprene and Its Polymers*

Study of the reactions of vinylacetylene, a compound which has be-
come available through discoveries described in the preceding paper (1),
has led to the synthesis of a series of new analogs arid homologs of isoprene.
The present paper is concerned with one of the simplest of these, namely,
chloro-2-butadiene-l,3 (I). This compound is especially interesting for
the following reasons. It is easily prepared in quantity in a state of purity;
it differs structurally from isoprene only in having a chlorine atom instead
of a methyl group; like isoprene it reacts with itself to yield a synthetic

* W. H. Carothers, I. Williams, A. M. Collins and J. E. Kirby, Journ. Am. Chem.
Soc. t 53, 4203-25 (1931) ; Contribution from the Experimental Station of the Central
Chemical Department (Communication No. 83) and from the Jackson Laboratory of
the Dyestuffs Department (Communication No. 26), E. I. du Pont de Nemours and Co.

Received October 3, 1931. Published November 5, 1931.

282 ACETYLENE POLYMERS AND THEIR DERIVATIVES

rubber; but the transformation occurs with much greater velocity than
in the case of isoprene; and the product is distinctly superior to natural
rubber in some of its properties.

In order to recognize the analogy in structure and behavior which ex-
ists between isoprene and chloro-2-butadiene-l,3, we call the latter com-
pound chloroprene, and this name also serves to distinguish it from other
chlorobutadienes that will be described in future papers.

Preparation of Chloroprene. Chloroprene is obtained by the addition
of hydrogen chloride to vinylacetylene

CHEEEC CH=CH 2 + HC1 > CH 2 =C CH=CH 2

Other products, which will be described in future papers, are also formed
by the action of hydrogen chloride on vinylacetylene under certain condi-
tions; but under the conditions indicated in the following example, chloro-
prene is practically the only product.

Fifty grams of cold vinylacetylene is placed in a pressure bottle containing a
thoroughly chilled mixture composed of 175 g. of coned, hydrochloric acid (sp. gr. 1.19),
25 g. of cuprous chloride and 10 g. of ammonium chloride. The bottle is placed in a
water-bath the temperature of which is held at approximately 30, where it is shaken
for a period of four hours. The contents of the bottle are placed in a separatory funnel,
the lower aqueous layer is drawn off and the oily layer is washed with water, dried
with calcium chloride, mixed with a small amount of catechol or pyrogallol and dis-
tilled in vacuo through an efficient column provided with a refrigerated dephlegmator
and receiver. Pure chloroprene is thus obtained in yields of about 65% of the theo-
retical based on the vinylacetylene applied. Some vinylacetylene is recovered in the
distillation. The yields can be considerably improved if the chloroprene is separated
from the reaction mixture by steam distillation in vacuo (100-250 mm.).

Physical Properties and Analysis. Chloroprene is a colorless liquid
with a characteristic ethereal odor, somewhat resembling ethyl bromide.
It is miscible with most of the common organic solvents, but only slightly
soluble in water. v Some of its other properties are boiling point (2), 59.4
at 760 mm., 46.9 at 500 mm., 40.5 at 400 mm., 32.8 at 300 mm., 6.4
at 100 mm.; vapor pressure, logio v. p. (mm. Hg) = l545.3/T(abs) +
7.527; calculated molal latent heat of evaporation, 7090 cal.; density,
d% 0.9583; refractive index, nf, 1.4540, n 1.4583, n, 1.4690.

Found M c 25 . 06 MD 25 . 26

Calcd. 24.67 24.66

Difference 0.39 0.60

Viscosity at 25, 0.394 centipoise.

II. A NEW SYNTHETIC RUBBER 283

Anal. Calcd.forC^HfiCl: C, 54.25; H, 5.69; Cl, 40.06; mol. wt., 88.5. Found:
C, 54.37; H, 5.95; Cl, 39.51, 38.81; mol. wt. (in freezing benzene) 89.0, 89.0.

Chemical Properties and Proof of Structure. The structure of chloro-
prene as chloro-2-butadiene-l,3 is established by its analytical composi-
tion and by the following reactions.

It reacts readily with maleic anhydride and yields, after hydrolysis
with water, a crystalline product to which we assign the structure chloro-4-
tetrahydro-l,2,3,G-phthalic acid (II). The chlorine atom of this product,
in accordance with its assumed structure, is very resistant to the action
of concentrated boiling alkali, and oxidation with boiling nitric acid
yields a crystalline compound identical in melting point and composition
with the known acid, butane-a,/3,7,5-tetracarboxylic acid (III).

CCr CH C(X H 2 CCK X CH COOH

I + || )>0 > || I

CHv X CH- CCK CHv /CHCOOH

VHj x CHo/

I II

CH 2 COOH

CH - COOH
CH - COOH

CH 2 COOH
III

Chloroprene further reacts readily with naphthoquinone, and the pri-
mary product, which presumably has the structure chloro-2-tetrahydro-l,-
4,4a,9a-anthraquinone-9,10 (IV), is smoothly oxidized by air in the presence
of alkali to 0-chloroanthraquinone (V). The identity of this is established
by the method of mixed melting point.

/CHfv /COv /

CCK N CH/ XX

II I .

CHv /CHs

CH 3 CC1=CH CH 2 C1
VI

In view of the recent studies of Diels and Alder (3) this result decisively
demonstrates the presence in chloroprene of a pair of conjugated double
bonds, and it fixes unequivocally the position of the chlorine atom.

284 ACETYLENE POLYMERS AND THEIR DERIVATIVES

In chloroform solution chloroprene readily adds approximately two
atoms of bromine before substitution begins. It rapidly decolorizes
alkaline permanganate solution. In the presence of cuprous chloride it
reacts with aqueous hydrochloric acid. The reaction consists in 1,4-
addition and the product is dichloro-l,3-butene-2 (VI), which will be de-
scribed in a future paper. The chlorine atom of chloroprene is very
firmly bound. Only traces of chloride ion appear on boiling with alcoholic
silver nitrate, alcoholic sodium hydroxide, or pyridine.

Reaction of Chloroprene with Maleic Anhydride. Preparation of Chloro-4-
tetrahydro-l,2,3,6-phthalic Acid (II). 21.2 g. of chloroprene was warmed with 19.6 g.
of maleic anhydride. The anhydride dissolved readily and when 50 was reached
sufficient heat was developed from the reaction to maintain this temperature for some
time. After standing overnight the reaction product was boiled with 200 cc. of water to
remove unchanged chloroprene and filtered. On cooling stout rectangular plates sepa-
rated; yield 31.5 g. or 77% of the theoretical; m. p. 171-172 (corr.). Recrystalliza-
tion from water raised the melting point to 173-175.

Anal. Calcd. for C 8 HO 4 C1: C, 46.94; H, 4.43; Cl, 17.34; neutral equivalent,
102.3. Found: C, 47.54; H, 4.78; Cl, 17.24, 17.71; neutral equivalent, 103.5.

The acid was boiled with 25% potassium hydroxide for three hours. No significant
quantity of chloride ion was produced. This indicates that the chlorine is attached to a
double-bonded carbon atom.

Oxidation to Butane-a,/3,7,S-tetracarboxylic Acid (III). A sample (8.2 g.) of the
above-described acid was warmed with 30 g. of 70% nitric acid until a rather violent
reaction took place with the evolution of nitrogen oxides. The unused nitric acid was
removed in vacuo and the partly crystalline residue taken up in a small volume of hot
water. On cooling a thick mass of flat needles with square ends separated After a
second crystallization, the product melted at 192-193 with effervescence.

Anal. Calcd. for CsHioO 8 : C, 41.05; H, 4.27; neutral equivalent, 58.5. Found:
C, 41.32, 41.46; H, 4.66, 4 67; neutral equivalent, 58 8.

This acid has already been described by Auwers and Jacob (4) and also by Farmer
and Warren (5).

Action of tt-Naphthoquinone on Chloroprene. Conversion to 0-Chloroanthra-
quinone. To 10 g. of naphthoquinone dissolved in benzene 12 g. of chloroprene was
added. The mixture was refluxed for three hours and then allowed to stand overnight.
The benzene was removed under reduced pressure, and the residual mass was dissolved
in warm alcohol and cooled. A considerable amount of unchanged naphthoquinone
separated. This was filtered off. Dilution of the mother liquor with water gave a solid
which separated from alcohol in fine needles melting at 76 . It was still contaminated
with naphthoquinone. It was suspended in alcohol containing a little sodium hydrox-
ide, and air was bubbled through the suspension for twenty minutes. The suspended
solid was crystallized three times from amyl alcohol; small needles, m. p. 209.5; mixed
melting point with /3-chloroanthraquinone, 209.5.

Spontaneous Polymerization of Chloroprene. The following example
is typical of the spontaneous polymerization of chloroprene.

II. A NEW SYNTHETIC RUBBER 285

About 40 cc. of chloroprene is placed in a 50-cc. bottle of soda glass,
closed with a cork stopper, and allowed to stand at the laboratory tem-
perature (about 25) in the absence of direct light. After twenty-four
hours the viscosity of the sample has considerably increased; after four
days it has set to a stiff, colorless, transparent jelly, which still contains
a considerable amount of unchanged chloroprene. As the polymerization
proceeds further this jelly contracts in volume and becomes more tough
and dense. After ten days all the chloroprene has polymerized. We
call this product ju-polychloroprene to distinguish it from other chloro-
prene polymers that will be described later in this paper.

Properties of ^-Polychloroprene. The product of the above de-
scribed reaction is a colorless or pale yellow, transparent, resilient, elastic
mass resembling a completely vulcanized soft rubber. Its density at
20 is about 1.23, and its refractive index (n%) is about 1.5512. It has
a tensile strength of about 140 kg./sq. cm. and an elongation at break
of about 800%. It is not plastic; that is, it does not sheet out smoothly
on the rolls of the rubber mill nor break down on continued milling. It
is not thermoplastic. It swells strongly but does not dissolve in carbon
tetrachloride, carbon disulfide, benzene, nitrobenzene, pyridine, aniline,
ethyl acetate and ether. Compared with natural rubber the tendency
of this material to imbibe gasoline and lubricating oil is very slight. When
a stretched sample is immersed in liquid air for a moment and then struck
with a hammer it shatters into fibrous fragments (6).

The properties of the ju-polymer vary somewhat depending upon the
conditions under which it is formed. When the chloroprene has access
to large amounts of air or oxygen during the polymerization the product
is dark in color and harder and stiffer than otherwise. The polymer
formed at elevated temperature is inclined to be soft and it has a distinct
terpene-like odor. This is due to the presence of the volatile ^-polymer,
which will be described in a subsequent paragraph.

x-Ray Diffraction Pattern. Although the x-ray diffraction pattern
of unstretched rubber shows only a single diffuse ring, characteristic of a
liquid or an amorphous solid, stretched rubber shows a point diagram (7).
On the other hand, according to Mark (8), all synthetic, polymerized
products from isoprene or other unsaturated hydrocarbons so far investi-
gated have given, even when stretched, a diffraction pattern analogous
to that of a liquid.

It is therefore a matter of considerable interest that the ju-polychloro-
prene described above (as well as the cured plastic polymer described in
a later paragraph), when stretched about 500%, exhibits a fully developed

286 ACETYLENE POLYMERS AND THEIR DERIVATIVES

fiber diagram showing a number of definite layer lines. One of these
diffraction patterns is reproduced in Fig. 1, together with a diagram of
stretched rubber (Fig. 2). The identity period along the fiber axis is
4.8 A. This length corresponds rather closely with the calculated length
for one chloroprene unit. The agreement is better if one assumes a trans
instead of a cis configuration since the calculated identity period in a cis
polyprene chain is about 2 X 4.1 A., whereas in a trans chain it is about
4.8 A. (9). Incidentally, 4.8 A. is exactly the identity period observed for
0-gutta percha by Hauser and v. Susich (10). Unstretched samples of
poly chloroprene give an amorphous ring (Fig. 3) entirely like natural
rubber. The spacing corresponding to this ring is 4.86 A. We are in-
debted to Dr. A. W. Kenney for these observations.

Chemical Properties of M-Polychloroprene. The /i-product has the
composition required for an addition polymer of chloroprene.

Anal. Calcd. for (CJIsCl),: C, 54.25; H, 5.69; Cl, 40.06. Found: C, 53.74,
54.83; H, 5.70, 5.93; Cl, 40.06, 39.32.

Molecular weight determinations are not possible on account of its
lack of solubility.

The /i-polymer is unsaturated toward bromine but no quantitative
data on this point are yet available. The chlorine atoms are very firmly
bound. Only slight traces of chloride ion are liberated when the compound
is heated for six hours in boiling alcoholic potash or boiling pyridine. This
fact suggests that the chlorine atoms of the polymer are still attached to
carbon atoms bearing double bonds.

The oxidation of the jit-polymer with hot nitric acid leads to the isolation
of succinic acid.

No attempts have yet been made to degrade the /^-polymer with ozone
completely, but when stretched it is much more resistant than natural
rubber to the deteriorating effect of ozone-containing air.

It is well known that purified rubber hydrocarbon is very susceptible
to autoxidation. A similar and perhaps more exaggerated sensitivity is
characteristic of synthetic rubbers derived from diene hydrocarbons.
In this respect ju-polychloroprene appears to be considerably more resistant.
Nevertheless it does not long remain completely unaltered when freely
exposed to air and light. It gradually becomes darker in color and finally
after two or three weeks is dark brown. At the same time, it becomes
harder, especially on the surface. These changes are accompanied by
the liberation of traces of hydrogen chloride. The autoxidation can be
suppressed by treating the polymer with small amounts of antioxidants.

II. A NEW SYNTHETIC RUBBER 287

Chemical Structure of /n-Polychloroprene. The molecules of natural
rubber are long chains built up from the unit (VII) derived (formally)
from isoprene (VIII). This structure follows from the fact that the ozoni-
CH 2 C=CH CH 2 CH 2 =C CH=CH 2

CH 8 CH S

VII VIII

zation of rubber (IX) leads to levulinic acid and levulinic aldehyde (X)
as the principal products (11).

. . .CH 2 C=CH CH 2 CH 2 C=CH CH 2 . . . >

I I

CH 3 CH 8 OCH CH 2 CH 2 COCH 3

IX X

In their behavior toward ozone the so-called normal synthetic rubbers from
isoprene very closely resemble the natural products (12), and their mole-
cules must therefore for the most part be built up on the same general plan.
Analogy suggests that the molecules of the chloroprene polymer are
similarly built up from the units (XI) derived from chloroprene (XII).

CHj C-=CH~-CH 2 CH 2 = C CH CH 2

Cl Cl

XI XII

The resulting chains would have the formula XIII.

O
. . .CH 2 C=:CHCH 2 CH2C=-CHCH2CH 2 C=CHCH 2 . . . > HOOC(CH 2 ) 2 COOH

Cl Cl Cl

XIII XIV

This formula readily accounts for the fact that oxidation of the ju-poly-
chloroprene yields succinic acid (XIV) . It also explains why the chlorine
atom is very resistant to the action of alkalies : as in vinyl chloride, the
chlorine is attached to a carbon atom bearing a double bond. This situa-
tion is changed however by autoxidation. This must lead to some such

'. CC1 CH

grouping as \ o / m which the chlorine atom would be^xceedingly

mobile.

Some evidence for this formula is also found in the physical properties
of the /x-polychloroprene. The molecular refractivity calculated for this
formula (22.95) agrees exactly with the experimental value (22.95). The
x-ray diffraction pattern indicates an identity period of 4.8 A., which
corresponds quite well with that calculated for one chloroprene unit.

288 ACETYLENE POLYMERS AND THEIR DERIVATIVES

It appears to be generally true that the presence of a chlorine atom at
a double bond decreases the tendency of the double bond to react with
ozone, and in this connection it is significant that /i-polychloroprene is
much more resistant than natural rubber to the deteriorating action of
ozone.

In the formulas IX and XIII the isoprene and chloroprene units have
been represented as being united regularly in 1,4-1,4- . . . order. The
units however are not symmetrical arid in joining of two units one or both
of them might be inverted. This would lead to the arrangements 1,4-
4,1- . . . and 4,1-1,4- .... It has been demonstrated that such inver-
sions (13) do occur in the polymerization of isoprene in the presence of
sodium and alcohol. On the other hand isoprene rubber formed by ther-
mal polymerization, since its behavior toward ozone is normal (14), must
be free from any considerable proportion of such inversions in its molecules.

The oxidation of /x-polychloroprene to succinic acid gives no indication
as to whether the arrangement of the units is normal as represented in
XIII or inverted as in XV and XVI, since the latter would also yield suc-
cinic acid.

. . CH*CC1=CII CHjr CH 2 CH= CCi CH 2 . . . XV

. . CH 2 --CH=CC1 CHr CH 2 - CC1=CH CH 2 . . . XVI

On the other hand the fact that ju-polychloroprene like natural rubber
yields a sharp x-ray diffraction pattern whereas this property is absent
from other synthetic rubbers perhaps indicates that the polychloroprene
is freer from irregularities in the structure of its molecules than other
synthetic rubbers.

The spontaneous polymerization of isoprene requires several years for
its completion ; with chloroprene the transformation is complete in a few
days. The great difference in speed may be ascribed, in part at least,
to the activating influence of the chlorine atom. The methyl group is too
feebly polar to exert any such effect. A similar difference exists between
vinyl chloride and propylene in their tendency to polymerize. The
chlorine atdm here functions not only to powerfully activate the double
bond, it also exerts a greater effect than methyl on the direction of addi-
tion reactions at the double bond : it more effectively controls the polarity
of the molecule. A similar directive effect in chloroprene is demonstrated
by its behavior toward hydrogen chloride. Thus the relation between
chloroprene and isoprene may be symbolized by the formulas XVII and
XVIII.

II. A NEW vSYNTHETIC RUBBER 289

2 = C-CH=CH 2 * ~ CH 2 =CCH=CH 2

CH 3
xvn xvin

It seems highly probable therefore on theoretical grounds not only that
the polymerization of chloroprene will proceed much more rapidly than
the polymerization of isoprene, but that there will be much less chance
of inversions of the units in the polychloroprene chains.

Formula XIII adequately represents the chemical behavior of /x-poly-
chloroprene, but is not sufficiently complete to account for the remarkable
physical behavior of this material. The difficulties in this connection
are precisely the same as those presented by natural rubber. The linear
polymeric structure partially represented in formula IX furnishes a suffi-
cient basis for describing the chemical behavior of rubber. It is known
further that the molecular weight of rubber is exceedingly high perhaps
in the neighborhood of 70,000 (15); but no entirely adequate explanation
of the elastic properties of rubber in terms of this structure has been
offered. It seems scarcely necessary to review the numerous speculations
that have been devoted to this subject (16). We merely point out that
the units in a substituted polyprene chain (e. g., rubber or polychloroprene)
present the possibility of geometrical isomerism and that they may be
arranged in cis-cis-cis- . . . order or trans-trans-trans- . . . order; or both
arrangements may be present in a single chain. The molecules may be
coiled into spirals rather than rigidly extended. On these questions we
have no data concerning polychloroprene (see however the above paragraph
entitled x-Ray Diffraction Pattern). Moreover, we have no direct infor-
mation concerning the molecular weight of polychloroprene. It seems
certain, however, that the chains must be very long.

We observe finally that ju-polychloroprene resembles vulcanized rather
than unvulcanized rubber. It is not plastic; it does not become plastic
when heated; and it does not dissolve but merely swells in rubber solvents
such as benzene and chloroform. This behavior is much more consistent
with a three-dimensional polymeric structure than with a simple linear
structure. We assume therefore that in /i-poly chloroprene chains of the
type already described are chemically linked together at occasional points.
The resulting structure as we conceive it may be symbolized by formula
XIX in which A stands for the structural unit. The cross-linking may
occur through the mediacy of oxygen atoms, but it seems more likely that
the double bonds of parallel chains mutually saturate one another directly.
Since the chains are exceedingly long they need to be linked together
only at occasional points to produce a non-plastic structure.

290 ACETYLENE POLYMERS AND THEIR DERIVATIVES

XIX . . . A A A A -A A A A . . .

A A A A A A A A

. . XX XX XX XV XX XX XX XX . . .

I I

. . . xx A xx x\~ xx xx A A . . .

/3-Polychloroprene. When chloroprene is polymerized at elevated
temperature, e. g, t at 60 especially in the absence of air, the product is
somewhat less dense and softer than that formed at ordinary temperature,
and it has a pronounced terpene-like odor. This odor is due to the presence
of volatile chloroprene polymers (/3-polymer) . This product can be pre-
pared in quantity by storing or heating chloroprene in the presence of sub-
stances such as pyrogallol or trinitroberizene that inhibit the transfor-
mation of chloroprene into the ^-polymer. By distillation the ^-polymer
can be separated into two fractions, one boiling at 92 to 97 at 27 mm.
and the other boiling chiefly at 114 to 118 at 27 mm. The odor of the
fractions is very similar fragrant and terpene-like. The constitution
of these products has not been determined, but there is little doubt that
they are cyclic dimers of chloroprene. It is already known that analogous
products are formed from isoprene (17), butadiene (18) and dimethyl-
butadiene (19). The /3-polychloroprenes are stable substances, which show
no tendency to polymerize further. They have therefore no direct bearing
on the formation of the rubber-like polymers, and they are relatively un-
important, especially since they appear in significant amounts only under
rather exceptional conditions.

Influences Affecting the Polymerization of Chloroprene. A few typical
data are assembled in Table I.

(a) Catalysts. Oxygen is an exceedingly powerful catalyst for the
transformation of chloroprene into the /-i-polymer. Samples of chloro-
prene distilled in high vacuum and sealed off in glass tubes without ex-
posure to the air show an appreciable increase in viscosity only after a
period of one or two months, and the transformation to /^-polymer is still
incomplete after twelve months. From some recent studies of the behavior
of isoprene (20) it appears that the active catalyst in such transformations
is a volatile peroxide. Since it is practically impossible to prepare a sam-
ple of chloroprene without exposing it to air at some time during the course
of its preparation, it seems probable that even the products distilled in
high vacuum are not altogether free from catalyst.

The amount of oxygen necessary to produce an optimum catalytic
effect is quite small. At ordinary temperatures the presence of a volume
of air equal to about 10% of the volume of the chloroprene sample causes
the transformation to /^-polymer to be completed in about eight to ten

II. A NEW SYNTHETIC RUBBER

291

TABLE I

ESTIMATED TIME REQUIRED FOR 90% OF A SAMPLE OF CHLOROPRENE TO POLYMERIZE
UNDER VARIOUS CONDITIONS (ABSENCE OF DIRECT LIGHT)

Time,
days Character of product

8 Colorless, strong, tough

400 Soft, low density, strong odor of dinier

2 Strong, tough, slight odor of dimer

10 Strong, tough, odor of dimer

< 1 Semi-fluid, black, much dimer

<2 Transparent, rather hard, very tough

0.7 Transparent, rather hard, very tough

2 Rather soft, odor of dimer

1 Product variable

1 Strong, tough, colorless
vStill fluid after G5 days
vStill fluid after 13 months; strong
odor of dimer

Temp.

, Pressure

Other added

No.

Air

atm.

substances

Present

None

Absent

None

Present

None

Absent

None

100

Absent

None

Present

4500

None

Present

6000

None

Absent

6100

None

Absent

1 . 5% Benzoyl

peroxide

Present

0.5% Benzoyl

peroxide

Absent

0.1% Catechol

Present

1.6% Thiodi-

phenylaminc

days. If the volume of air is much smaller than this the time required
for the transformation is somewhat greater, but much larger ratios of air
or oxygen do not greatly increase the velocity. Large quantities of oxygen
do however affect the character of the final product. In general they
lead to a product which, instead of being colorless or only slightly yellow,
is dark brown and considerably harder and stiffer than the usual products.

Peroxides such as benzoyl peroxide also function as catalysts. Their
use, however, presents no practical advantages, and is in fact somewhat
hazardous. In samples containing benzoyl peroxide there is frequently
a considerable induction period, and then reaction suddenly starts at
some point in the sample and spreads very rapidly through the mass.
The heat of reaction may be sufficient to char the sample.

The rate of formation of /3-polychloroprene is not appreciably acceler-
ated by oxygen or peroxides.

(b) Temperature, Pressure and Light. Catalytic and anticatalytic
effects in the polymerization of chloroprene are very powerful and difficult
to regulate exactly ; for this reason it is impossible to obtain precise quanti-
tative correlations concerning the effect of temperature, pressure and
light on the reaction velocity. It appears, however, that in the presence
of air the transformation of chloroprene into the ^-polymer occurs about
four times as fast at 62 as at 25. Thus the temperature coefficient of
the catalyzed reaction is abnormally low. The relative rate of the for-

292 ACETYLENE POLYMERS AND THEIR DERIVATIVES

mation of 0-polymer, which is negligible at ordinary temperature, is more
strongly affected by rise in temperature. Polymers produced at tempera-
tures above 50 contain appreciable proportions of /3-polymer. These
proportions are still further increased if the reaction is carried out in the
absence of air since this greatly reduces the speed with which ju-polymer
is formed without affecting the formation of /3-polymer. In the absence
of air the effect of increased temperature on the rate of transformation
of chloroprene into /z-polymer appears to be much greater than in the
presence of air.

When chloroprene is polymerized at temperatures above 80 (in the
absence of solvents) considerable decomposition occurs with liberation
of appreciable amounts of hydrogen chloride, and the product is dark in
color and tarry in consistency.

At a pressure of 6000 atmospheres the polymerization of chloroprene
occurs about ten times as rapidly as at ordinary pressure.

Light has a considerable accelerating effect on the transformation of
chloroprene into the //-polymer. The rate of formation of /3-polymer
is not affected. The active wave lengths He in the blue, violet and near
ultraviolet.

(c) Inhibitors. -Substances that generally function as antioxidants
act as powerful inhibitors for the transformation of chloroprene into the
ju-polymer. Under ordinary conditions a sample of chloroprene will set
in four days to a stiff jelly containing about 40% polymer but the pres-
ence of 0.1% of catechol will permit the sample to remain fluid for several
months. This fact confirms indications already mentioned that the spon-
taneous transformation of chloroprene into ju-polymer is normally de-
pendent upon the presence of traces of autoxidation products of the chloro-
prene. The formation of /3-polymer is not thus dependent upon oxida-
tion. Samples of chloroprene containing inhibitors, after several months
at the ordinary temperature, are found to contain several per cent, of
jS-polymer, and frequently they are quite free of ju-polymer. A sample of
chloroprene containing 0.2% of pyrogallol yielded 49% of crude 0-polymer
after being heated for forty days at 62.

The following types of compounds generally function as inhibitors:
phenols, quinones, amines, mercaptans, thiophenols, aromatic nitro com-
pounds, halogens. Some compounds in each of these classes function as
powerful inhibitors, others have a feebler effect. It is somewhat sur-
prising to find aromatic nitro compounds in this list. Trinitrobenzene
is among the most powerful of the inhibitors.

II. A NEW vSYNTHETIC RUBBER 293

In the presence of relatively feeble inhibitors or small amounts of the
more powerful inhibitors the polymerization of chloroprene can be effected
at a somewhat diminished rate. The non-volatile polymer formed under
these conditions, however, differs very considerably in its properties from
that formed in the absence of inhibitors.

(d) Solvents. The polymerization of chloroprene can be effected in
the presence of solvents. If the solvent is one such as benzene, toluene,
ethylene chloride, or carbon disulfide that powerfully swells M-poly chloro-
prene, the resulting polymer remains dissolved. The polymerization of
solutions containing as much as 50% by volume of chloroprene leads to
the formation of stiff jellies. Even as little as 10% of chloroprene leads
to highly viscous solutions. The solutions are generally colorless and
transparent. The polymers contained in these solutions are somewhat
different in their properties from the /i-polychloroprenes produced under
ordinary conditions. They are softer, and unless the solution is very
old they can usually be redissolved in benzene.

The presence of the solvent considerably diminishes the rate of the
polymerization, and dilute solutions polymerize more slowly than con-
centrated ones. Solvents also frequently exert a specific effect. Chloro-
prene dissolved in benzene polymerizes very much more rapidly than
chloroprene dissolved in ether or pyridine.

Chloroprene may also be polymerized in the presence of non- volatile
solvents, inert fillers and foreign materials of various kinds.

Granular Polymer (co-Polymer). The polymerization of chloroprene
occasionally leads to a coherent mass of glistening, hard, rubbery granules
or globules (co-polymer). This material is non-plastic and it shows scarcely
any tendency to imbibe solvents. The conditions favoring its formation
are not very clearly understood, since it occasionally appears under the
most diverse conditions. It seems certain, however, that its formation
is autocatalytic. When a speck of this polymer appears in a sample of
chloroprene during the early stages of its polymerization the granular
growth continues to spread through the whole sample. Because of its
cell-like structure it occupies more volume than the same amount of M~
polymer, and if the growth begins to spread laterally through a sample
it may burst the walls of a heavy Pyrex container even when the total
volume of the container is much greater than the volume of the product.
The presence of metallic sodium especially favors the formation of the
granular polymer. It frequently appears under other conditions that
result in very slow polymerization.

According to experiments made by Dr. H. W. Starkweather the forma-

294 ACETYLENE POLYMERS AND THEIR DERIVATIVES

tion of the co-polymer is initiated (or accelerated) by light of 3130 A. wave
length. The following observation is especially interesting.

Chloroprene containing pyrogallol to inhibit polymerization was placed
in the bottom of a long Pyrex tube. The chloroprene was cooled to 80,
and the tube was evacuated and sealed off. The lower half of the tube
was covered with black friction tape to exclude light and the upper part
was exposed to light from a mercury arc. During the exposure the lower
part of the tube was kept in a bath at 10 ; the upper half was at 00-65.
After twelve hours there was a white deposit at the top of the tube. This
deposit gradually increased during two and one-half days. Exposure
to the light was then discontinued. The solid deposit, however, continued
to form at the top of the tube as a white, crinkly mass, until the liquid
in the bottom of the tube was completely exhausted.

It is interesting to observe that products similar to this granular poly-
mer have been obtained from other dienes. Kondakow (21) observed that
dimethylbutadiene in a closed flask in diffused daylight is gradually
transformed into a white, insoluble mass, and Harries (22) obtained a
similar product by the action of ultraviolet light on isoprene.

The great resistance of the co-polychloroprene as compared with the
co-polymer to the swelling action of solvents indicates a considerably
higher degree of cross-linking of the chains in the co-compound. If such
a cross-linking should occur in a sufficiently regular fashion, it would lead
to a three-dimensional primary valence lattice, a type of structure which
is illustrated by the diamond, but is not known among synthetic organic
compounds. The conditions under which the co-polymer is formed are
such as might be especially favorable to the development of a regular
three-dimensional structure. The formation of co-polymer is catalyzed
by an co-polymer surface ; the process is one of heterogeneous autocatalysis.
It seems most probable that the function of the co-polymer surface in this
connection is not (or at least not wholly) to activate adsorbed molecules
of the monomer, but rather to orient the adsorbed molecules into a con-
figuration favorable for mutual union. The incidence of activating energy
could then bring about the combination of a very large number of mole-
cules in a single act.

As a matter of fact the granular polymer at first sight gives the im-
pression of being definitely macrocrystalline, but on closer observation
the crystals turn out to be globules. x-Ray examination gives only an
amorphous pattern.

It is interesting to note that the granular polymer never appears in

II. A NEW SYNTHETIC RUBBER 295

samples of a-poly chloroprene (described later) that contain phenyl-/3-
naphthylamine.

Progressive Changes during the Spontaneous Polymerization of
Chloroprene. The following table illustrates in more detail the changes
in properties and composition that occur when a sample of chloroprene
is allowed to stand under ordinary conditions in the presence of a little air.

TABLE II
CHANGES DURING THE POLYMERIZATION OF CHLOROPRENE

Time, Viscosity in

days Polymer, % Density centipoises

0.952 0.4

1 4 .. 6.0

2 14 0.98 550.0

4 45 1.06 vSt iff jelly

10 99-h 1.23 Non-plastic

The polymer formed during the early stages of the reaction can be
isolated by precipitation with alcohol, or by distilling off the unchanged
chloroprene in vacuo. This material is very different in its properties
from the final product, the ^-polychloroprene already described. It is
soft, plastic, and completely soluble in benzene. We call this plastic
polymer a-poly chloroprene. When allowed to stand at the ordinary
temperature it slowly reacts with itself and in the course of a day or two
is transformed into a product apparently identical with the ju-polymer.

Mechanism of the Formation of a- and ju-Polymers. The isolation
of the a-polymer demonstrates that the transformation of chloroprene
into the /x-polymer is a step-wise reaction. Some of the facts concerning
the two polymers and their relation to each other are best correlated by
a brief discussion of mechanism.

The transformation of chloroprene into the a-polymer is evidently a
chain reaction. It is enormously susceptible to catalytic and anticatalytic
effects; it is accelerated by light; and although a large number of mole-
cules is combined to form a single larger molecule, the apparent order of
the reaction is low. The reaction probably first involves the coupling
of an activated molecule of chloroprene with another chloroprene mole-
cule. The activating energy persists in the polymeric chain until it has
been built up to a considerable length. The molecules of a-polymer
thus formed doubtless have the linear structure already suggested in
formula XIII. The formation of a molecule of a-polymer involves a
series of separate acts, but these follow one another in very rapid succession
Under ordinary conditions the a-polymer present when 4% of a sample

296 ACETYLENE POLYMERS AND THEIR DERIVATIVES

of chloroprene has polymerized is indistinguishable from the polymer
present when 20% of the chloroprene has polymerized.

The transformation of the a-polymer into the ju-polymer consists in the
cross-linking of the long chains into a three-dimensional structure of the
type represented in formula XIX. In a sample of chloroprene under-
going spontaneous polymerization this process becomes noticeable when
the concentration of polymer has reached about 25%. It is marked by
an abrupt change in properties. The viscosity increases very rapidly
and the sample soon sets to a stiff jelly. If the polymer is isolated just
before this point is reached, it is found to be soft and plastic. Polymer
isolated just after this abrupt change is still soft, but the manner in which
it resists permanent deformation indicates the presence of a considerable
proportion of the ju-polymer.

The reactions, (1) chloroprene > a-polymer, and (2) a-polymer >*
^-polymer, are not merely two stages of a single process, but are different
reactions. The polymerization of chloroprene is rather strongly inhibited
by primary aromatic amines, such as aniline, the naphthylamines, and
benzidine; but these same compounds when mixed with isolated a-poly-
mer accelerate its conversion into /z-polymer. The temperature coefficients
of the two reactions are different. For reaction (2) the ratio of the velocity
constants for a temperature increase of 10 is about two; for reaction (1)
the ratio is considerably less than two.

Other Polymers; Balata-like Polymer. The a-, M- and co-types arc
not chemical individuals but rather qualitatively different species of poly-
meric mixtures. The properties of each type may vary over a considerable
range and in practice no doubt one generally has to do not with a pure
species but with a mixture in which one of the species may preponderate.
A consideration of the formulas assigned to the a- and /z-polymers will
suggest some of the complications that might arise. The molecules of
the a-polymer are no doubt chains of very great length, but in the poly-
merization of chloroprene under certain conditions, e. g., at elevated
temperature, the process of cross-linking may set in before the chains
have attained the usual length of a-polymer chains. One will then have
a product very different in its properties from that produced by the vul-
canization of a-polymer. Stereochemical factors (e. g., cis-trans isomerism)
may also produce great variations in the character of the products.

The a-y 0-, /A- and co-polymers by no means exhaust the different types
of polychloroprenes. Anything that influences the velocity of the poly-
merization has some effect on the properties of the product, and the modifi-
cations produced by inhibitors and catalysts are especially marked. The

II. A NEW SYNTHETIC RUBBER 297

phenomena in this connection are very complex. It would be useless to
attempt to recognize as distinct species all the distinguishably different
polymeric products derived from chloroprene.

There is, however, one type of product that appears to be qualitatively
different from the a-, /3-, ju- and w-polymers. This material rather closely
resembles balata in its properties. It is obtained more or less contami-
nated with the other types of polymers under various conditions, but es-
pecially by the polymerization of chloroprene in the presence of inhibitors
such as iodine or the tetraalkyl thiuramdisulfides.

A typical specimen of this balata-like material when cold is a hard,
amorphous, non-brittle mass. When warmed to 60 it becomes soft
and plastic. At higher temperatures it is quite sticky. When heated
under vulcanizing conditions the plasticity is partly lost, but the trans-
formation to the elastic condition is very incomplete. We make no at-
tempt to suggest a structure for this material.

Conditions for the Isolation of a-Polymer. Owing to the effect of
changing concentrations, the rate of formation of a-polymer progressively
decreases during the polymerization of chloroprene, and the rate of the
conversion of a-polymer into ^-polymer progressively increases. For
this reason if pure a-polymer is to be obtained, the reaction must be in-
terrupted before all of the chloroprene has polymerized.

Under the most favorable conditions the concentration of the a-polymer
in the polymerizing mixture can be built up to 30 or 40% before any ap-
preciable transformation to ju-polymer occurs. The reaction is best con-
ducted in glass vessels under strong illumination from a Mazda lamp or
a mercury arc in glass. The most effective wave lengths lie in the long
ultraviolet, but the use of a quartz container with mercury arc radiation
is not advisable on account of the danger of forming granular polymer.
The temperature should be kept in the neighborhood of 35. Under these
conditions about 30% of the chloroprene is polymerized in sixteen to
twenty-four hours. The product is a thick, colorless, transparent sirup.
If this sirup is poured into a large volume of alcohol, the a-polymer sepa-
rates as a colorless mass and the unchanged chloroprene remains dis-
solved in the alcohol. The a-polymer can also be separated by allowing
the unchanged chloroprene to distil out of the mixture under diminished
pressure. The mixture is preferably stirred during the distillation.

Properties of the a-Polymer. In density and refractive index the
a-polymer lies very close to the /^-polymer. The a-polymer resembles
milled smoked sheets in its physical properties and mechanical behavior.
It is plastic and it dissolves completely in benzene to form highly viscous

298 ACETYLENE POLYMERS AND THEIR DERIVATIVES

solutions. It can be calendered into thin sheets or extruded with the
usual rubber machinery

At 30 the a-polymer loses its plastic properties and becomes completely
changed to the elastic form (ju-polymer) in about forty-eight hours. At
130 the transformation is complete in about five minutes. This process
corresponds to the vulcanization of natural rubber, but sulfur is not
needed and when present it takes no part in the process. The speed of
this transformation can be greatly modified by the addition of various
substances, some of them materials that are used in the vulcanization
of natural rubber.

Zinc oxide brings about the vulcanization of a-polymer in eight to ten
hours at 30. Zinc chloride, zinc butyrate and ferric chloride arc even
more active catalysts. The most effective organic catalysts are primary
aromatic amines such as aniline, the naphthyl amines and benzidine.
Diphenylguanidine, which is a relatively active vulcanization accelerator
for natural rubber, is a mild accelerator for a-polymer. On the other
hand mercaptobenzothiazole and tetraalkylthiuram sulfides, which are
active natural rubber accelerators, have no accelerating action on the
a-polymer. Basic inorganic materials such as lime and magnesium oxide,
which may accelerate natural rubber through their action with sulfur,
have a slight retarding influence on the vulcanization of a-polymer.
Strong acids and acidic materials that retard the vulcanization of natural
rubber have no influence on the curing of plastic polymer.

Secondary aromatic amines such as phenyl-0-naphthylamine power-
fully inhibit the vulcanization of a-polymer at ordinary temperature.
This fact is of considerable practical importance since it brings about the
possibility of storing the plastic polymer over long periods of time. The
phenyl-jft-naphthylamine also acts as an antioxidatit arid confers age-
resisting properties on the final product. The inhibiting effect of the
phenyl-j3-naphthylamine on the curing of the a-polymer largely disappears
above 100.

Behavior of a-Polymer in Compounding and Properties of the Cured
Rubber. The compounding ingredients that can be used with a-poly-
mer are similar to those used with natural rubber, but there are a number
of important additions. Materials such as ground leather and cork which
strongly retard the vulcanization of natural rubber act as inert ingredients
in a-polymer. Carbon black and zinc oxide act as reinforcing agents as
they do in natural rubber and impart good abrasion resistance. In con-
trast to their action in rubber, whiting and clay are perfectly wet by a-
polymer, and they produce compounds having good tear resistance. Cot-

IT. A NEW vSYNTHETIC RUBBER 299

ton and other vegetable fibers are also much more perfectly wet by a-
polyrner than by rubber. Most plasticizing and softening materials
such as mineral oil, stearic acid and pine tar are insoluble in a-polymer
and have little true softening action. Milling also produces little soften-
ing other than a temporary thermal effect. Mineral rubber and similar
asphaltic materials act as diluents with little effect on the physical proper-
ties of the vulcanized material.

Natural rubber can be successfully milled into a-polymer, although
there is little affinity between the two. Sheets of the two rubbers before
vulcanization may be firmly pressed together and easily separated. A
reasonably firm union can be obtained between natural and chloroprene
rubber when they are vulcanized together under sufficient pressure. Ben-
zene solutions of rubber and a-polymer are not compatible; when thor-
oughly mixed they quickly separate into two layers.

Table III gives the composition of compounded stocks prepared from
a-poly chloroprene and from smoked sheets. These were used for a series
of parallel tests to compare the behavior of the two rubbers. All the
processing and testing were carried out with the usual rubber laboratory
equipment.

TABLE III

COMPOSITION OF COMPOUNDED STOCKS FROM CK-POLYCHLOROPRENE AND FROM SMOKED

SHEETS

Amount in grams

Compound from Compound from

Materials a-polychloroprene smoked sheets

aj-Polychloroprene 100

Smoked sheet rubber . 100

Zinc oxide 10 10

Sulfur ... 3

Stearic acid 2 2

Diphenylguanidine ... 1
Benzidine (accelerator) 0.5

Phenyl-/3-naphthylamine 1 1

The effect of different times and temperatures of vulcanization on the
physical properties of the two compounds is shown in Table IV. The
chloroprene rubber reaches a maximum tensile strength after only five
minutes at 140 ; but the strength is not adversely affected if the curing
time is extended to sixteen hours. In contrast to this, the natural rubber
compound vulcanizes more slowly and softens rapidly under the action
of prolonged vulcanization. The maximum tensile strength for the chloro-
prene compound is slightly lower than that for the natural rubber com-

300

ACETYLENE POLYMERvS AND THEIR DERIVATIVES

pound. In contrast to the smoked sheet compound the chloroprene
compound shows no reversion and the load supported at 500% elongation
continues to increase slowly throughout the curing range. When the
polychloroprene compound is heated longer than sixteen hours at 140
a material resembling hard rubber is formed.

TABLE IV
EFFECT OF VULCANIZATION ON PHYSICAL PROPERTIES OF COMPOUNDED STOCKS

Vulcaniza-
tion
temp.,

110
110
120
120
140
140

Minutes
vulcaniza-
tion
time

20
40
20
40
;5
10

Load at 500% elongation,
kg. per sq cm. hr
Chloroprene vStnoked Chi
rubber sheet r

1(3
28.1

28.2
30.2
29 9
33.4

Pensile strength at
eak, kg. per sq cm.
oroprene Smoked
ubber sheet

!(>(> 9

177.5
109 3
71.2
70.5 . .
70 5

% Elongation
at break
Chloroprene Smoked

920

800
800
820

825
780

140

38.0

770

8(30

140

1(3

700

830

140

175

()

750

820

140

(>()

1 1(>7

210

750

700

140

4(3

2 1(57

192

720

080

140

120

42.0 181

205

740

720

140

480

1 05

700

800

140

9(>0

(>3

9 151

(540

820

The chloroprene compound is very resistant to the action of ozone.
The two compounds were stretched about fifteen per cent, and exposed
to ozone-containing air. The natural rubber compound was ruptured
in three minutes; the chloroprene compound was not detectably affected
during an exposure of three hours.

Table V shows the results of artificial aging tests on the two compounds
of Table III. The chloroprene compound was cured for fifteen minutes
and the natural rubber compound for sixty minutes at 140. The tests
were carried out at 70 in oxygen at twenty atmospheres. Under these
conditions twenty-four hours is generally considered to approximate one
year of natural aging for rubber. These data indicate that the chloro-
prene rubber is considerably more resistant to oxidation than natural
rubber. Natural aging tests have not been carried out for a length of
time sufficient to confirm this conclusion, but samples of the chloroprene
rubber compound that have been kept for one year show no deterioration.

Chloroprene rubber is much more resistant than the natural product
to the action of solvents and many chemicals. After seventy- two hours
the chloroprene compound had increased 7% in weight by immersion in

II. A NEW SYNTHETIC RUBBER 301

TABLE V
AGING PROPERTIES OF COMPOUNDS FROM SMOKED SHEETS AND FROM

or-POLYCHLOROPRENK

Tensile strength,

kg. per sq cm. % Elongation at break

Days in oxygen Chloroprene Smoked Chloroprene Smoked

bomb at 70 rubber sheets rubber sheets

172.2 232.0 890 720

1 212.6 195.0 845 660

2 181.0 163.4 820 660

3 193.3 159.9 805 670
8 209.1 116.1 720 595

14 165.2 ,54.5 690 510

light machine oil and 25% by immersion in kerosene, and it had retained
more than half of its original tensile strength in each case. The tensile
strength of the natural rubber was destroyed under these conditions.
In contrast to natural rubber, Chloroprene rubber is not attacked by hy-
drogen chloride, hydrogen fluoride, sulfur chloride, ozone and many other
chemicals. The high chlorine content of the chloroprene rubber also
renders it very resistant to combustion. Measurements of diffusion of
both hydrogen and helium through a poly chloroprene membrane show it to
be only 40% as permeable as natural rubber. The absence of water-
soluble materials in chloroprene rubber makes it very resistant to pene-
tration by water.

Synthetic Latex. Chloroprene is readily emulsified by shaking or
stirring it with water containing an emulsifying agent such as sodium
oleate. The resulting emulsion polymerizes very rapidly and completely.
The polymer remains suspended or emulsified and constitutes an artificial
latex. When the water is allowed to evaporate from a layer of this latex,
a thin coherent, strong, elastic film remains. This film in its physical
and mechanical properties very closely resembles the ^t-polychloroprene
already described: it is strong, extensible, elastic, resilient, non-plastic
and not thermoplastic, and it is swelled but not dissolved by benzene.

The following example illustrates the preparation of a synthetic latex.
Four hundred grams of chloroprene is slowly added with vigorous stirring
to 400 g. of water containing 8 g. of sodium oleate in a wide-mouthed
bottle. A smooth emulsion results. After a time (usually about thirty
minutes) the temperature of the mixture begins to rise and it may quickly
reach the boiling point of the chloroprene unless cooling is applied. After
standing for two to eight hours at room temperature the polymerization
is complete. The mixture is then practically odorless.

More uniform products are obtained if the temperature is carefully

302 ACETYLENE POLYMERS AND THEIR DERIVATIVES

controlled during the emulsification and polymerization. At a tempera-
ture of 10 the process is complete in about twenty-four hours always
in less than forty-eight hours. The reaction is always characterized by
an induction period, which at 10 usually lasts forty to sixty minutes. It
is evident that the speed of polymerization of chloroprene is much greater
(apparently at least 20-fold) in emulsion than otherwise. The particular
nature of the interface appears to be of great importance in determining
the rate of polymerization. The rate is much more rapid with sodium
oleate than with egg albumen although both of these emulsifying agents
produce very small particles.

A small amount of free acid is developed during the polymerization
of the emulsions, and this gradually brings about coagulation during stor-
age. However, if a little ammonia (e. g., 5 g. of NH 3 per liter) is added
to the latex after completion of the polymerization this tendency is avoided.
Latex stabilized in this way can be stored indefinitely without change.
In addition to the ammonia it is ordinarily desirable to add an antioxidant
such as phenyl-|3-naphthylainine since this greatly prolongs the life of
articles prepared with the latex.

The particle size of latex prepared according to the above example is
very small and remarkably uniform. Figure 4 gives the results of some
measurements made by Dr. J. B. Nichols with the ultracentrifuge. When
sodium oleate is used as the emulsifying agent the mean radius of the
particles is about 0.063/x and more than (>0% of the particles lie between
0.05 and O.OTju- It appears that the ultracentrifugal method has not
been applied to natural latex, but according to Hauser (23) the latex from
mature Hevea trees contains particles ranging in diameter from 0.5 to
3/x as well as a considerable proportion of smaller particles. The particle
size of the synthetic latex can be controlled to a certain extent by suitably
modifying the nature and the amount of the emulsifying agent. The
use of lithium oleate in the ratio of two grams to one hundred grams of
chloroprene gives particles having a mean radius of about 0.087/x.

In the preparation of the synthetic latex the ratio of chloroprene to
water can be varied over a wide range. As long as the concentration lies
below 55% by weight of poly chloroprene, the latices are very fluid. Above
this concentration there is a sharp increase in viscosity, and a 60% latex
is quite thick.

The synthetic latex is rapidly coagulated by acids, alcohol, acetone and
many salts. The polychloroprene separates as a coherent mass, which is
at first quite soft and plastic. However, as soon as the water is squeezed
out, this plasticity is lost. The mass then has the properties already

II. A NEW vSYNTHETIC RUBBER

303

indicated for /z-polychloroprene it is analogous to a soft, vulcanized
natural rubber.

Applications of the Latex. The chloroprene latex can be directly
applied to many uses after the manner of vulcanized natural latex. In
this connection it has some special advantages owing to its peculiar prop-
erties. Thus on account of its small particle size it penetrates porous

articles such as leather and .__________________,____

wood in a manner that can
hardly be approached with
natural latex.

Shaped articles are readily
prepared by dipping forms of
glass, metal or porcelain into
the latex and coagulating the
resulting film or allowing it to
dry. By repeated dipping,
articles of any desired thick-
ness can be built up.

The latex can be mixed with
dyes, fillers and modifying and
protective agents of various
kinds to adapt it to specific
uses.

Polymerization of Chloro-
prene in Porous Materials.
Attempts to impregnate such
porous materials as leather,
wood and tile with natural
rubber are unsuccessful,
whether the rubber is used in
the form of latex or dissolved
in a solvent such as benzene;
but a very intimate impreg-
nation of these materials with

chloroprene rubber can be accomplished by soaking them in chloroprene
and then allowing the rubber to be formed in place. Chamois or kid leather
saturated with its own weight of chloroprene and sealed to prevent evapo-
ration until polymerization is complete becomes translucent and assumes
a rubber-like flexibility without extensibility. Impregnated spruce be-
* Fraction of total weight (ck) corresponding to a certain radius (r) .

100 200 300

Radius in millimicrons.

Fig. 4. Weight distribution curve for chloro-
prene lattices. The units are such that the
area under the curves for any radius interval
equals the fraction of the total weight of parti-
cles whose radii lie in that interval. Curve A,
latex made with sodium oleate. Curve B, latex
made with lithium oleate.

304 ACETYLENE POLYMERS AND THEIR DERIVATIVES

comes water resistant but its appearance is unchanged. This process
may be used with any porous or bibulous material that does not con-
tain inhibitors for the spontaneous polymerization of chloroprene.

Conclusion. If space permitted it would be possible to review the
literature on synthetic rubber from diene hydrocarbons and to show that
almost every recorded peculiarity and complication in this field finds
some analogy in the behavior of chloroprene. Chloroprene is therefore
capable of serving adequately as a representative diene in studying as
a scientific problem the synthesis of rubber-like materials, and for this
purpose it has the great advantage of its very high speed of polymeriza-
tion as compared with dienes previously available.

On the economic side the greatly diminished costs of producing natural
rubber have obviated the need for an artificial material having the same
properties. There remains, however, the need for a synthetic rubber that
is free from some of the inherent defects of the natural product. The
differences between polychloroprene and natural rubber are sufficient to
suggest considerable potentialities for the new synthetic product.

Summary

Chloro-2-butadiene-l,3 (chloroprene) is described arid its structure
established through reactions leading to its conversion into butane- a,
/3,7,5-tetracarboxylic acid, and into /3-chloroanthraquinone.

Within ten days under ordinary conditions in a closed vessel containing
a little air, chloroprene spontaneously changes into a transparent, resilient,
strong, non-plastic, elastic mass resembling vulcanized rubber. This
product is called ^-poly chloroprene. By interrupting the polymerization
before it has proceeded to completion one obtains a soft plastic product
(a-polymer) that resembles unvulcanized rubber. Under the action of
heat the a-polymer rapidly changes to the ^-polymer. Other polymers
of chloroprene described are volatile (0-) polymer, granular (co-) polymer,
and balata-like polymer. The structures of the polymers are discussed
as well as the effect of conditions on the formation of each type.

Unlike any previously described synthetic rubbers, ^-polychloroprene
resembles natural rubber in the fact that when it is stretched its x-ray
diffraction pattern shows a point diagram.

The transformation of chloroprene into /i-polychloroprene occurs very
rapidly in aqueous emulsion. The resulting product constitutes a syn-
thetic (vulcanized) latex. It has a much smaller particle size than natural
latex and it penetrates porous materials more readily.

II. A NEW vSYNTHETIC RUBBER 305

Chloroprene can also be polymerized in the pores of porous or bibulous
materials. The materials thus become intimately impregnated with syn-
thetic rubber.

Compared with natural rubber the new synthetic rubber is more dense,
more resistant to absorption or penetration by water, less strongly swelled
by petroleum hydrocarbons and less permeable to many gases. It is much
more resistant to attack by oxygen, ozone, hydrogen chloride, hydrogen
fluoride and many other chemicals.

Bibliography and Remarks

(1) Nieuwland, Calcott, Downing and Carter, Journ. Am. Chem Sac , 53, 4197 (1931).

(2) The.se data are based on vapor pressure measurements made with an isoteniscope by Dr.
H. W Stark weather

(3) Diels and Alder, Ber , 62, 2337 (1929).

(4) Auwers and Jacob, tbid., 27, 1114 (1894)

(5) Farmer and Warren, /. Chem. Soc , 897 (1929).

(6) The same behavior has already been observed in rubber by Hock, Gummi-Ztg., 39, 1740
(192. r >)

(7) Katz, Chem -Zlg , 49, 353 (1925); Meyer and Mark, Ber , 61, 1939 (1928).

(8) "Die Rontgentechnik in der Materialprufung," Egbert and vSchiebold, Akademische Verlags-
gesellvschaft, Leipzig, 1930* p. 142

(9) Meyer and Mark, "Der Aufbau der hochpolymeren organischen NaturstofFe," Akademische
Verlagsgesellschaft, Leipzig, 1930

(10) Hauser and v Susich, Kantschuk, 7, 145 (1931).

(11) Harries, Ber., 37, 2708 (1904); 38, 3985 (1905); Pummerer, Kbermayer and Gerlach, ibid ,
64, 809 (1931).

(12) Harries, " Untersuchungen," Julius v Springer, 1910, p 222

(13) Midgley and Henne, Journ. Am. Chem Soc , 52, 2077 (1930).

(14) Cf. Pummerer and Koch, in Memmler's "Handbuch der Kautschukwissenschaft," S Hirzel,
Leipzig, 1930, p. 270.

(15) Staudinger and Bondy, Ann , 488, 127 (1931)

(16) vSome of the most suggestive are Staudinger, Kautschuk, 5, 911, 1261 (1929), Meyer and
Mark, "Der Aufbau der hochpolymeren organischen Naturstoffe," Akademische Verlagsgesellschaft,
Leipzig, 1930; Fikentscher and Mark, Kautschuk, 6, 2 (1930).

(17) Pummerer and Koch, loc. cil , p 284

(18) Hofmann and Tank, Z aneew Chem , 25, 1405 (1912).

(19) Van Romburgh and Van Romburgh, Proc Roy. Acad Amsterdam, 34, 224 (1931)

(20) Couant and Peterson, private communication

(21) Kondakow, / prakl. Chem., [2J 64, 109 (1901).

(22) Pummerer and Koch, loc at , p 268.

(23) Hauser, "Latex," Theodor vSteinkopff, Leipzig, 1927, p. 56.

306 ACETYLENE POLYMERS AND THEIR DERI V ATI VES

III. The Addition of Hydrogen Chloride to Vinylacetylene*

Vinylacetylene (I) constitutes the simplest possible example of a con-
jugated enine system. Recorded information concerning the addition
reactions of such systems is exceedingly meager. It has already been re-
ported (1) that chloroprene (chloro-2-butadiene-l,3, III) can be obtained by
the addition of (aqueous) hydrogen chloride to Vinylacetylene, and the
present paper is concerned with a further description of the mechanism
of this reaction and the nature of the products to which it leads.

It is shown that the initial step consists in 1,4 addition, and the primary
product thus formed is chloro-4-butadiene-l,2 (II) (b. p. <SS), a new com-
pound of rather unusual structure and curious properties. Under certain
conditions this chloro-4-butadiene-l,2 can be isolated as the major reaction
product, but it readily undergoes an isomerization involving migration of
the chlorine atom and a shift of a double bondf. Chloroprene (b. p. 59.4)
is formed thus, and the transformation occurs with such facility in the
presence of hydrogen chloride that chloroprene always constitutes a part
of the reaction product. Certain salts reinforce the catalytic effect of
hydrogen chloride on this transformation, and when cuprous chloride is
present no chloro-4-butadiene-l,2 is found in the reaction product. When
sufficient amounts of hydrogen chloride are present the reaction proceeds
further with the formation of dichloro-2,4-butene-2 (IV).

HC1

CHC CH=CH 2 > CH,=C==CH CH,,C1 >

I II

HC1
CH 2 =C CH=-CH 2 > CH 3 C=CH CH 2 C1

I I

Cl Cl

III IV

Influence of Conditions on the Reaction. The reaction between
Vinylacetylene and hydrogen chloride is conveniently carried out in the
following manner. Fifty grams of Vinylacetylene and 175 cc. of concen-
trated hydrochloric acid (about 2.2 moles of hydrochloric acid) are placed in
a pressure bottle, and the bottle is shaken to promote contact between the
aqueous and the hydrocarbon layers. After the completion of the reaction
the oily layer is separated, dried, mixed with a small amount of an atiti-

* W. H. Carothers, G. J. Berchet and A. M. Collins, Journ. Am. Chem. Soc., 54,
4066-70 (1932); Contribution No. 96 from the Experimental Station of E. I. du Pont
de Nemours and Co.

Received June 9, 1932. Published October 5, 1932.

f For a further study of the chemical behavior of chloro-4-butadiene-l,2 and for
a discussion of its rearrangement to chloroprene, see this volume, page 335.

III. VINYLACETYLENE WITH HYDROGEN CHLORIDE 307

oxidant such as catechol or pyrogallol, and distilled in vacuo through an
efficient column. Data from a very large number of experiments of this
type are available. Owing to the fact that in most cases neither the tem-
perature nor the speed of shaking was precisely controlled these data
cannot be used as a basis for a quantitative description of reaction veloci-
ties; nevertheless, they give a clear idea of relative rates.

In a typical experiment of the kind described above, about 43% of the
vinylacetylene was utilized in seven hours, and analysis of the reaction
product yielded chloro-4-butadiene-l,2 and chloroprene in the ratio 2.2:1.

In general this ratio is diminished by increase in the temperature, con-
centration of hydrogen chloride, or time of contact; and the proportion of
the chloro-4-butadiene-l,2 in the product is less the more completely the
vinylacetylene is utilized. This fact demonstrates that part at least of
the chloroprene is formed by the isomerization of the chloro-4-butadiene-l,2
during the course of the reaction, and it seems reasonable to conclude that
all of it is formed by this demonstrated mechanism.

The reaction is considerably accelerated by the presence of certain salts.
Thus when 23 g. of calcium chloride is present in the reaction mixture the
time required to obtain 40% conversion is decreased by about one-half,
but the ratio of the two isomeric chlorobutadienes present in the reaction
product at a given percentage conversion remains practically unaffected.
Cuprous chloride is a much more powerful catalyst. When 23 g. of this
salt is present in the reaction mixture, about 90% of the vinylacetylene
reacts in four hours at 20. In this case the product consists for the most
part of chloroprene, and no chloro-4-butadiene-l,2 is present. It is, how-
ever, not necessary to assume that the cuprous chloride directs the* reaction
so that addition occurs at the acetylenic linkage. Separate experiments
show that cuprous chloride reinforces the catalytic effect of hydrogen
chloride on the isomerization of chloro-4-butadiene-l,2 and this effect seems
adequate to explain the absence of chloro-4-butadiene-l,2 in the reaction
product.

Chloroprene reacts further with hydrogen chloride to produce dichloro-
2,4-butene-2. This reaction is also catalyzed by cuprous chloride, and the
relative velocities are such that any conditions of concentration, tempera-
ture, and time of contact that suffice to convert all of a given sample of
vinylacetylene result in the formation of a certain amount of the dichloro-
butene. The following experiment is illustrative. Fifty grams of vinyl-
acetylene, 173 cc. of concentrated hydrochloric acid, 23 g. of cuprous chlo-
ride, and 10 g. of ammonium chloride were placed in each of forty bottles.
The bottles were gently shaken in a bath at 20 for four hours and then

308 ACETYLENE POLYMERS AND THEIR DERIVATIVES

allowed to stand for twelve hours at 0. The contents of all the bottles
were combined and the mixture was steam distilled in vacua into a receiver
placed at the bottom of a 2-meter jacketed, carborundum-packed column
provided with a dephlegmator and a second receiver each cooled to 80.
The pressure was kept at about 150 mm. until all of the unchanged vinyl-
acetylene had collected in the second receiver, and the pressure was then
gradually reduced to 10 mm. until all of the chloroprene had collected in the
second receiver. The dichlorobutene remained in the first receiver. The
yields were: unchanged vinylacetylene, 115 g.; chloroprene, 2802 g.; di-
chlorobutene, 117 g. In mole percentages these values are 5.7, 84.2
and 2.5, respectively. The deficit amounting to 7.6 mole per cent, was
mostly comprised in intermediate fractions. When this deficit is dis-
tributed proportionately among the three major fractions, the percentage
yields become: unreacted vinylacetylene, 6.1%; chloroprene, 91.2%;
dichlorobutene, 2.7%. The corresponding figures for the calculated yields
based upon unrecovered vinylacetylene are chloroprene, 97%, and di-
chlorobutene, 3%. Further experiments indicate that the conditions of
this experiment lie very close to the optimum for the conversion of vinyl-
acetylene to chloroprene in a batch process : a higher ratio of chloroprene to
dichlorobutene can be obtained only by utilizing a smaller proportion of the
applied vinylacetylene, and a more complete utilization of the vinyl-
acetylene results in a larger proportion of dichlorobutene.

The Properties of Chloro-4-butadiene-l,2 and the Proof of Its Struc-
ture. Chloro-4-butadiene-l,2 is a colorless liquid boiling at 87.7 to
88.1. It is only slightly soluble in water but miscible with most of the
common organic solvents. It has a peculiar, sharp odor. Some other
properties are : n 1.4775; df, 0.9891; M tt calcd., 24.61; A/ R found, 25.30.
Anal. Calcd. for C 4 H 5 C1: C, 54.23; H, 5.64; Cl, 40.11. Found: C, 55.04, 55.11 ;
H, 5.70, 5.90; Cl, 39.75, 40 03.

Its chlorine atom is exceedingly reactive. When mixed with alcoholic
silver nitrate it rapidly yields a copious precipitate of silver chloride.
This fact in itself indicates that the compound is not a 1,3-diene, since in
such a structure the chlorine atom would, of necessity, be attached to a
doubly bonded carbon. Furthermore, the compound does not react with
maleic anhydride or with naphthoquinone. It is not a true acetylenic
compound either since it does not yield any derivative with ammoniacal
cuprous chloride. When treated with ozone it yields formaldehyde and
(after oxidation with potassium permanganate) chloroacetic acid. The
compound has also been directly oxidized with potassium permanganate.
The only product obtained was chloroacetic acid. Acetic acid and oxalic

III. VINYLACETYLENE WITH HYDROGEN CHLORIDE 309

acid were absent. This behavior demonstrates the presence of the groups
CH 2 = and =CH CH 2 C1, and the compound must therefore have the
structure chloro-4-butadiene-l,2. This structure is further confirmed by
the fact that under the action of cold concentrated sulfuric acid, chloro-4-
butadiene-1,2 is readily converted into chloro-4-butanone-2 (2).

The transformation of chloro-4-butadiene-l,2 into chloroprene exempli-
fies a type of reaction that is common to many substituted allyl halides.
Such halides (e. g. t CH 3 CHXCH=CH 2 or CH 3 CH=CHCH 2 X) arise
quite generally by the addition of halogens or hydrogen halides to 1,3-
diencs, and in these cases the possibility of isomerization frequently makes
it difficult or impossible to determine whether the primary product is the
result of 1 ,2 or 1,4 addition. Chloro-4-butadiene-l, 2, however, differs from
other allyl halides: the a, 7 shift brings the adjacent double bonds into the
more stable conjugated configuration and the chlorine atom becomes at-
tached to a doubly bonded carbon where its mobility is lost. A reversal of
the isomerization is therefore impossible, and chloro-4-butadiene-l,2 can-
not be other than a primary product of the addition of hydrogen chloride to
vinylacetylene.

The transformation of chloro-4~butadiene-l,2 into chloroprene has been
observed under a variety of conditions : by the action of powdered potas-
sium hydroxide, by the action of hot quinoline (140-150), by the action of
heat (290) in the presence of silica gel, and by the action of hot dilute
hydrochloric acid (3) . However, the isomerization occurred most smoothly
and rapidly in the presence of hydrochloric acid containing cuprous chlo-
ride. Fifty grams of chloro-4-butadiene-l ,2 was refluxed for three and one-
half hours with 20 g. of cuprous chloride in 100 cc. of 18% hydrochloric
acid. The oily layer was decanted, dried and distilled. The entire speci-
men except for a small amount of undistillable residue came over between
o9 and ()3 and the distillate was pure chloroprene. When the chloro-4-
butadiene-1,2 was similarly treated with aqueous cuprous chloride alone it
was recovered unchanged.

Chloro-4-butadiene-l, 2 unlike its isomer chloroprene shows no tendency
to polymerize. It can be distilled unchanged at ordinary pressure, and
specimens stored under the ordinary laboratory conditions for many months
remain unaltered. It undergoes no change even when submitted to a
pressure of 6000 atmospheres for forty- five hours at 50 (4).

Experimental Part

Oxidation of Chloro-4-butadiene-l, 2. To a mixture of 30 g. of chloro-4-butadiene-
1,2 with 250 cc. of water containing a little sodium carbonate was added in small por-

310 ACETYLENE POLYMERS AND THEIR DERIVATIVES

tions with constant stirring 214 g. of potassium permanganate. The mixture was fil-
tered, the filtrate acidified and continuously extracted with ether for several hours.
The ether solution on distillation gave a liquid boiling at 185 which solidified on cooling.
This was chloroacetic acid, identified by its melting point, neutralization equivalent
(found, 96; calcd., 94.5) and transformation into chloroacetamide, m. p. 119. No
acetic or oxalic acid was found.

Ozonization of Chloro-4-butadiene-l,2. A solution of 20 g. of chloro-4-buta-
diene-1,2 in 20 cc. of chloroform was treated with ozone for twelve hours at 0. The
solvent and the unchanged material were evaporated in vacua and the remaining ozonide.
was decomposed with water. Formaldehyde was detected in the aqueous solution by its
strong odor and the formation of methylene-di-/3-naphthol; white needles melting at
204 (corn).

Chloroacetaldehyde was not detected directly. The aqueous solution was treated
gradually with 50 g. of potassium permanganate, the excess permanganate destroyed
with sulfur dioxide, and the filtrate extracted with ether. Distillation of the ether so-
lution left a residue which crystallized on cooling. This product was chloroacetic acid,
identified by its melting point and neutralization equivalent.

Hydration of Chloro-4-butadiene-l,2.- Into 250 cc. of concentrated sulfuric acid
88.5 g. of chloro-4-butadienc-l,2 was added dropwise with stirring, the temperature be-
ing maintained at 5 to 3. The dark reaction product was poured onto cracked ice,
partly neutralized with sodium carbonate and extracted with ether. The ethereal
solution was washed, dried and distilled, yielding 58 g. of crude chloro-4-butanone-2,
b. p. 110-123. On redistillation it boiled at 120 to 122 at 760 mm.

Anal. Calcd. for C 4 H 7 OC1: C, 45.07; H, 6.57; Cl, 33.33. Found: C, 45.42,
45.42; H, 6.81, 6.47; Cl, 32.26, 32.42.

When treated with pheiiylhydrazine it gave a derivative having the correct melting
point (77) and analysis for phenylmethylpyrazoline (5).

Anal. Calcd. for CioHi 2 N 2 : C, 75.00; H, 7.50; N, 17.50. Found: C, 74.32;
H, 7.60; N, 17.10.

Dichloro-2,4-butene-2. This material is obtained as a by-product in the prepara-
tion of chloroprene and it is readily prepared in quantity by shaking vinylacetylene with
an excess (4 moles) of hydrochloric acid containing cuprous chloride. It is a colorless
liquid having a characteristic odor; other properties are: boiling point 127-129 at
756 mm., 61-63 at 70 mm., 53 to 54 at 50 mm., d* 1.1591, 2 D 1.47239, w 2 c ? 1.46988,
w 2 F 1.48187, A/R calcd. 29.94, M R found 30.27.

Anal. Calcd. for C 4 H 6 C1>: C, 38.40; H, 4.80; Cl, 56.80. Found: C, 38.53;
H, 4.87; Cl, 56.90.

The proof of the structure of this compound will be presented in a future paper deal-
ing with its reactions.

Acknowledgment. The writers are indebted to Mr. O. R. Kreimeier
for assistance in the experiments on the addition of hydrogen chloride to
vinylacetylene.

Summary

The results of experiments on the action of aqueous hydrogen chloride
on vinylacetylene are described. The initial step consists in 1,4 addition

IV. VINYLACETYLENE WITH HYDROGEN BROMIDE 311

and the primary product is chloro-4-butadiene-l,2. This readily under-
goes isomerization, yielding chloroprene, which always constitutes a part of
the reaction product. When cuprous chloride is present in the reaction
mixture the isomerization proceeds more rapidly and no chloro-4-buta-
diene-1,2 is found in the reaction product. When sufficient hydrogen
chloride is present the reaction proceeds further, yielding dichloro-2,4-
butene-2.

Bibliography and Remarks

(1) Carothers, Williams, Collins and Kirby, Journ. Am. Chem. Soc., 53, 4203 (1931); this volume,
page 281.

(2) Cf. Gustavson and Demjanoff, J. prakt. Chem., [2] 38, 201 (1888); Bouis, Ann. chim., [10] 9,
402 (1928).

(3) We are indebted to Dr. D. D. Coffman for some of these observations.

(4) We are indebted to Dr. H. W. Starkweather for this observation.

(5) Maire, Bull. soc. chim., [4] 3, 272 (1908).

IV. The Addition of Hydrogen Bromide to Vinylacetylene,
Bromoprene and Dibromobutene*

Observations on the combination of vinylacetylene with hydrogen chlo-
ride (1) have been extended to the analogous case of hydrogen bromide.

The two reactants are closely similar in their behavior, but hydrogen
bromide appears to act somewhat more slowly than hydrogen chloride.
Concentrated aqueous hydrobromic acid containing cuprous bromide,
when shaken with vinylacetylene at the ordinary temperature, yields the
two products, bromo-2-butadiene-l,3 (bromoprene, II) and dibromo-2,4-
butcne-2 (III). It seems likely, in view of the results already described
for the hydrogen chloride reaction (Ib) that the primary product of reaction
between hydrogen bromide and vinylacetylene is bromo-4-butadiene-l,2
(I), which then rearranges to yield bromoprene, but no decisive direct
evidence for the formation of this primary product is yet available. Its
presence among the reaction products has not been established even when
no cuprous bromide (catalyst) was used. The structure of the dihydro-
bromide has not yet been directly established either, but formula III,
in view of the results with hydrogen chloride, is not open to serious doubt.

* W. H. Carothers, A. M. Collins and J. E. Kirby, Journ. Am. Chem. Soc., 55,
786-8 (1933); Contribution No. 106 from the Experimental Station of the E. I. du
Pont de Nemours and Co.

Received August 11, 1932. Published February 9, 1933.

312

ACETYLENE POLYMERS AND THEIR DERIVATIVES

CH^C CH=CH 2

HBr

2 =C CH CH 2 Br

HBr

Br
I II

CH 8 CBr= CH CH 2 Br
III

The third step in this series of reactions has a relatively high velocity,
and appreciable amounts of the dihydrobromide (III) are present in the
reaction mixture even at an early stage when the major part of the vinyl-
acetylene applied remains unchanged. In this respect also the behavior
of hydrogen bromide appreciably differs from that of hydrogen chloride.
Bromoprene is an oil having a faint greenish-yellow color and an odor
rather closely resembling that of butyl bromide. It boils at 42 to 43 at
165 mm. The proof of its structure follows the course already indicated
for chloroprene. It reacts with maleic anhydride to yield, after hydrolysis,
a product whose composition agrees with that required for bromo-6-
cyclohexene-5-dicarboxylic-2,3 acid (IV). Bromoprene furthermore reacts
readily with naphthoquinone. The primary product is a crystalline solid,
which probably has the formula bromo-2-tetrahydro-l,4,4a,9a-anthra-
quinone-9,10 (V). It is rapidly oxidized by air in alkaline solution to 0-
bromoanthraquinone (VI).

CH,

BrC

CH,

CH 2 CO CH
BrC/WVH

CH 2
H 2 O BrC/^CH COOH

IxCH COOH
CH,
IV

O
VI

Experimental Part

Preparation of Bromoprene. Fifty grams of cold vinylacetylcne was placed in a
pressure bottle containing 225 cc. of a thoroughly chilled solution prepared by dissolving
75 g. of moist, freshly prepared cuprous bromide in 200 g. of concentrated hydrobromic
acid (sp. gr. 1.55). The bottle was allowed to warm up to room temperature and was
then mechanically shaken for seven hours. After standing overnight and shaking for

IV. VINYLACETYLENE WITH HYDROGEN BROMIDE 313

another two hours the unused vinylacetylene was allowed to evaporate. The residual
oil was separated, washed with water, dried over calcium chloride and distilled from
hydroquinone under reduced pressure. Dry, oxygen-free nitrogen was led through the
capillary tube during the distillation. The yield of bromoprcnc boiling at 42-43 at
165 mm. was 30 g. (24%); df , 1.397; w 2 D 1.4988; Af R calcd , 27.50; MR found, 27 94.

Anal. Calcd. for C 4 H 5 Br: C, 30.11; H, 3.78; Br, 60.15. Found: C, 36.32, 36.47;
H, 4.13, 4.21; Br, 60.91, 61.43.

Higher yields were obtained by the use of more concentrated hydrobromic acid.
For example, by using 2 moles of hydrobromic acid (sp. gr. 1 .66) for each mole of vinyl-
acetylene a yield of 44.8% of bromoprenc was obtained.

Dibromo-2,4-butene-2. After the distillation of the bromopretie a considerable
amount of higher-boiling residue remained. From this residue there was obtained 12 g.
(6%) of a strongly lachrymatory, light yellow oil corresponding in analysis to a dibromo-
butene. It had the following physical constants: b. p. (760 mm ) 168 to 169 with loss
of HBr; b. p. (23 mm.) 73; c/J 1.8768, n 1.5485; M R calcd., 35.74; MH found,
36.25.

Anal Calcd for C 4 H 6 Br,: C, 2243, H, 2.80; Br, 7477. Found: C, 2298,
23.06; H, 3.05, 2.95; Br, 74 94, 75 34

Reaction of Bromoprene with Maleic Anhydride. Preparation of Bromo-6-cyclo-
hexene-5-dicarboxylic-2,3 Acid. To 5 g. (0037 mole) of bromoprenc was added 3 g
(0.030 mole) of maleic anhydride. After the mixture had stood at room temperature
for about one hour a spontaneous reaction set in and after three or four hours the mass
solidified. The reaction product was then dissolved in benzene and shaken for twenty
minutes with an excess of 10% sodium hydroxide. The aqueous layer was separated
and acidified. The crude acid which separated was purified by recrystallization from
water. It separated from water in the form of a mixture of thick plates and blunt needles
melting sharply at 186.5 to 187.

Anal Calcd. for C 8 H 9 O 4 Br: C, 38.57; H, 364, neutral equivalent, 124.5.
Found- C, 3837, 38.65; H, 3.75, 3.65; neutral equivalent, 1254

Action of a-Naphthoquinone on Bromoprene. Conversion to 0-Bromoanthraqui-
none. To a benzene solution of 15.4 g. (0 12 mole) of brornoprene was added 10 g
(0.06) mole of a-naphthoquinone. After standing at room temperature for two days
the solution was gently refluxed for one hour The benzene was then distilled off under
reduced pressure. The residual sticky, dark red solid was pressed on a porous tile and
then extracted with warm alcohol The alcohol was decanted from an insoluble tar and,
on cooling, 4.9 g. (49%) of unchanged naphthoquinonc crystallized out. When the al-
cohol mother liquor was diluted with water, 2.6 g of a nearly black solid separated.
After recrystallization from alcohol it was obtained in the form of small, soft, nearly
white crystals. When slowly heated on a copper block it began to turn blue at about
115 and slowly deepened in color as the temperature was raised. At 138 it melted
sharply; as the temperature was raised further it volatilized completely.

The analyses corresponded with the values calculated for bromo-2-tetrahydro-
1 ,4,4a,8a-anthraquinone-9, 1 0.

Anal. Calcd. for CnHuOaBr: C, 57.74, 11,3.99. Found: C, 57.06, 57.65, H,
3.89, 4.04.

About 0.1 g. of the substance was dissolved in alcohol and a few drops of 10% so-

314 ACETYLENE POLYMERS AND THEIR DERIVATIVES

dium hydroxide solution added. The solution was dark red in color. As air was bub-
bled through the solution, the red color gave way to green, which in turn disappeared
leaving a yellow solid. After recrystallization from amyl alcohol the oxidation product
melted at 205-207. /3-Bromoanthraquinone melts at 207 (2).

Summary

Vinylacetylene reacts with aqueous hydrobromic acid to form bromo-2-
butadiene-1,3 (bromoprene) and dibromo-2,4-butene-2. Bromoprene re-
acts with maleic anhydride yielding bromo-6-cyclohexene-o-dicarboxylic-
2,3 acid, and with naphthoquinone yielding a bromotetrahydroanthra-
quinone which is readily oxidized to /3-bromoanthaquinone.

Bibliography and Remarks

(1) (a) Carothers, Williams, Collins and Kirby, Journ. Am. Chem. Soc., 53, 4203 (1931); this
volume, page 281; (b) Carothers, Berchet and Collins, Journ. Am Chem. Soc., 54, 4066 (1932); this
volume, page 306

(2) Heller, Ber., 45, 672 (1912).

V. The Polymerization of Bromoprene (Third Paper on New
Synthetic Rubbers*)!

Chloroprene (I, X = Cl) polymerizes spontaneously to yield a rubber-like
product (1), and the speed of this transformation is roughly 700 times
greater than the analogous transformation of isoprene. The present paper
deals with the behavior of bromoprene (2) (I, X = Br), and the results may
be summarized in the statement that it shows no significant qualitative
differences from chloroprene, although its speed of polymerization under
most conditions appears to be somewhat greater.

CH 2 C CH==CH 2 CH=C H

I I

X X

I II

In connection with the behavior of these materials the following analo-
gies are of interest. Vinyl chloride and vinyl bromide (II, X = Cl and
Br) polymerize spontaneously yielding products of very high molecular
weight (3), but this behavior is not shown by propylene, which polymerizes
only in the presence of special catalysts or under drastic conditions and

* For the Second Paper in this sub-series, see this volume, page 384 reference to
the reprinted paper by Carothers and Coffman, Journ. Am. Chem. Soc., 54,4071(1932)

f W. H. Carothers, J. E. Kirby and A. M. Collins, ibid., 55, 789-95(1933); Con-
tribution No. 107 from the Experimental Station of E. I du Pont de Nemours and Co.

Received August 11, 1932. Published February 9, 1933.

V. POLYMERIZATION OF BROMOPRENE 315

then yields products having only moderately high molecular weights. The
haloprenes (I) bear the same structural relationship to isoprene that the
vinyl halides bear to propylene. Thus the very powerful activating effect
of a single halogen atom on ethylene is also manifested in butadiene
(when the halogen atom is on the j3-carbon). The rate of polymerization
is of great importance in studying the behavior of dienes since a high rate
not only makes it possible to obtain experimental results in a reasonable
length of time, but it permits one to extend the observations over a wide
range of conditions. It becomes possible then to recognize the different
types of reactions involved in the spontaneous polymerization and to obtain
data on the way in which these different types of reactions are affected by
changes in the conditions. Data of this type on chloroprene have already
been presented, and it will be useful to review them briefly (with some
extensions) and to make comparisons with those now available for bromo-
prene and for other dienes.

The polymerization of chloroprene leads to the four well defined and
qualitatively distinct types of polymers shown in the chart.

Volatile ((3) polymer -< - monoprene - > plastic (a) polymer

/ /

granular (co) polymer non-plastic GU) polymer

The influence of various conditions on each of these reactions is indicated
in Table I.

TABLE I
INFLUENCE OF CONDITIONS ON THE POLYMERIZATION OF CHLOROPRENE

Monoprene to Monoprene to a-Polyprene to Monoprene to

Condition 0-polymer a-polyprene /i-polyprene w-polyprene

Temperature + + + + + Autocatalytic, ini-

Pressure + -f- + + ? tiated by strong

Light ++ 0? ultraviolet light

Oxygen + + + + + ? and by metal

Antioxidants --- a surfaces

Certain substances commonly classified as antioxidants (e. g., phenyl-0-naphthyl-
amine) act as inhibitors; others (e. g. t benzidine) act as accelerators: + accelerates,
no effect, inhibits.

Of these different types of polymers the most important are the a- and
the /x-polyprenes since the former corresponds to unvulcanized rubber
and the latter corresponds to vulcanized rubber.

The ju-polyprene is the final product of the spontaneous polymerization of
chloroprene; the a-polyprene is an intermediate step in the formation of
the ^-product. Isolation of the a-polymer free of At-polymer is possible

316

ACETYLENE POLYMERS AND THEIR DERIVATIVES

only when the reaction is conducted under certain conditions : the tempera-
ture must not be too high, since elevated temperature accelerates the
transformation of a-polymer to /z-polymer more than it accelerates the
formation of a-polymer from monomer; the reaction must not be too slow,
since most inhibiting influences have a greater decelerating effect on the
formation of the a- than on the /i-product; moreover, very slow reaction
frequently leads to the formation of the co-polymer. Light and pressure

both appear to have a greater ac-
celerating effect on the formation of
the a-polyprene than on the trans-
formation of this into the ju-poly-
prene.

The /3-polymer, a terpene-like
material, is an undesirable by-prod-
uct in the rubber synthesis. In the
chloroprene polymerization the pro-
duction of 0-polymer becomes ap-
preciable only at temperatures higher
than are necessary to bring about
a i rapid formation of a-polyprene.
Butadiene, isoprene and dimethyl-
butadiene polymerize very slowly;
to get the transformation to proceed
at a reasonable rate, elevated tem-
peratures are generally used, and this
results in the formation of relatively
large amounts of /3-polymer.

The co-polychloroprene is also a
useless product; it is made up of
discrete rubber-like particles (ir-
regular globules) which are non-plastic and not even swelled by rubber
solvents. The formation of this polymer is autocatalytic. When a speck
of the co-polymer appears (or is introduced) in a specimen of incompletely
polymerized chloroprene, the entire specimen is soon more or less com-
pletely converted into the co-polymer. The formation of nuclei of the
co-polyprene is favored by strong ultraviolet light and by metal surfaces
(e. g. y sodium, potassium, mercury, iron, copper and aluminum). The op-
portunities for the formation of such nuclei are also increased by a long
reaction time under any particular set of conditions. Similar polymers, fre-
quently described as cauliflower-like masses, have been obtained from

Fig. 1. x-Ray diffraction pattern
of /z-polybromoprene (from latex)
stretched 500%.

V. POLYMERIZATION OF BROMOPRENE 317

isoprcne, butadiene and dime thy Ibutadiene, and one suspects that they
may be the forms in which the polymers of these dienes are most frequently
obtained. The very slow rate of the polymerization of these dienes would
be especially favorable to the formation of co-polymer.

The term rubber-like is vaguely used to cover a multitude of the most
diverse properties, and the literature of synthetic rubber is exceedingly
obvscure. Most of the agencies that are available to hasten the very slow
polymerization of isoprene and butadiene are such as have been found
in the case of chloroprene to affect more strongly the conversion of the
a- into the jii-polymer than the formation of the a-polymer. One may
expect therefore that the isolation of a true a-polyprene (4) from the iso-
prene and butadiene products will present especial difficulties. So far as we
are aware no clear disclosure has ever been made of an a-polyprene from iso-
prene or from butadiene. However, when isoprene is subjected to a pres-
sure of 12,000 atmospheres until 30% of the isoprene has polymerized,
the polymer is at least 90% soluble in ether (5). If the reaction is allowed
to proceed further until 80% of the isoprene has polymerized, the product is
completely insoluble (5, 0) . Thus the formation of a completely vulcanized
rubber-like product without the aid of sulfur is by no means peculiar to
chloroprene. In fact the differences between the behavior of chloroprene
and the behavior of other dienes appear to be differences of degree rather
than differences of kind.

The Spontaneous Polymerization of Bromoprene. A sample of bromo-
prene when allowed to stand at the ordinary laboratory conditions in a
stoppered flask containing a small amount of air becomes noticeably more
viscous after twelve to fifteen hours. As the reaction proceeds, the vis-
cosity increases; after about five days the sample sets to a stiff, elastic jelly
containing a considerable amount of unchanged bromoprene. Usually
after eight to ten days all of the bromoprene has reacted, but the time
required varies considerably in different experiments. The product has a
density of about 1.74, and this is 24% greater than the density of the
bromoprene. (The increase in density in the formation of /x-polychloro-
prene is about 28%.)

This product, ju-polybromoprene, corresponds in its properties with the
//-polychloroprene already described. It is tough, resilient and elastic
but harder than the analogous product from chloroprene. On standing it
gradually undergoes further change: it becomes still harder and less ex-
tensible and the resemblance to soft vulcanized rubber pretty largely dis-
appears although it still remains very tough and retains considerable
elasticity and resiliency. These changes in the nature of the product are

318 ACETYLENE POLYMERS AND THEIR DERIVATIVES

due, at least in part, to a progressive action of air and they can be retarded
by the application of antioxidants to the surface of the sample. ju-Poly-
bromoprene is similar to ju-polychloroprene in its behavior toward solvents.
It is greatly swelled by chloroform, carbon tetrachloride and aromatic
hydrocarbons but remains practically unchanged after prolonged immer-
sion in alcohol, ether, or aliphatic hydrocarbons such as gasoline.

a-Polybromoprene. The preparation of an a-polybromoprene (plastic
polymer) presents no difficulties. A sample of bromoprene was exposed to-
light from a Cooper-Hewitt lamp at 25. After twenty-four hours 50%
of the material had been converted into polymer. (This rate is about
40% greater than that usually obtained with chloroprene under the same
conditions.) The product was an exceedingly viscous, yellow sirup.
When it was mixed with a large volume of alcohol the a-polybromoprene
was precipitated as a soft, plastic mass. This product showed no tendency
to resist permanent deformation and sheeted out very readily on cold rolls.
Two per cent, of phenyl-jS-naphthylamine was worked into the plastic
mass to prevent spontaneous conversion into an elastic polymer (7). The
sample was compounded with about 5% of its weight of zinc oxide and
heated in a mold at 120 to 125 for twenty minutes. The product was
non-plastic, strong, resilient and extensible (500 to 700%). However,
compared with a similar product from chloroprene it was somewhat lacking
in snap, and its permanent set was rather high. It also showed a greater
tendency to "freeze/' After about two hours at ordinary temperatures it
became very stiff, but the original pliability was restored when it was
heated to 80 for a few minutes.

The Polymerization of Bromoprene in Aqueous Emulsion. Like
chloroprene, bromoprene is readily dispersed in water and the resulting
smooth emulsion polymerizes with great rapidity to form a stable latex.
The preparation and polymerization of such an emulsion is illustrated in
the following example.

Twenty-five cubic centimeters of 2% aqueous sodium oleate in a wide-mouthed
bottle was surrounded by a bath of ice and water. Two drops of triethanolamine was
added to the solution, and then, with vigorous stirring, 25 g. of bromoprene. A smooth,
milk-like emulsion resulted. After the mixture had stood in the ice-bath for five hours
an aliquot portion was removed and poured into a large volume of alcohol. The weight
of the precipitate thus obtained indicated that 78% of the bromoprene had polymerized.
Under the same conditions chloroprene is only about 20-30% polymerized. The bromo-
prene emulsion described above was transferred to a refrigerator. After seventeen hours,
more than 95% of the bromoprene had polymerized. Five cubic centimeters of 3%
ammonium hydroxide was added to stabilize the emulsion and a small amount of an

V. POLYMERIZATION OF BROMOPRENE 319

aqueous dispersion of phenyl-/3-naphthylamine (2% on the rubber content) added to
function as an antioxidant. The resulting latex was very stable.

As in the case of polychloroprene, the dispersed particles in the synthetic
latex derived from bromoprene correspond more closely to the p.- than to
the a-polymer. Nevertheless, when the latex is coagulated by the addi-
tion of acids, the particles coalesce and cohere very firmly. The coagulum
is at first soft and plastic, but it quickly becomes tough, elastic and non-
plastic. Homogeneous, coherent films are obtained by allowing the water
to evaporate from a thin layer of the fluid latex on a plate of porous por-
celain. The films are readily stripped from the plate and the removal
of the water can be completed by drying them for a few hours in an oven
at 80. Such films are exceedingly tough and more resistant to tearing
than analogous films prepared from chloroprene latices. A typical speci-
men had a breaking strength of 160 kg./sq. cm. and an elongation at break
of 740%. (The elongation of similar polychloroprene films is usually
about 800%.) The films exhibited the high permanent set and the tend-
ency to freeze or stiffen already referred to in connection with the vul-
canized a-poly bromoprene. Like the latter, they were non-plastic and
did not dissolve but merely swelled in chloroform and benzene.

/3-Polybromoprene.^The conversion of bromoprene into a volatile
liquid polymer (0-polybromopreiie) occurs under conditions similar to
those already described for chloroprene (8) . A sample of bromoprene con-
taining about 5% of thiodiphenylamine, a substance which powerfully
inhibits the conversion of the haloprenes to rubber-like polymers, was
heated in a sealed tube at 80 for five days. When the resulting black
oil was poured into alcohol a small amount of a black tar separated. From
the alcohol there was obtained by distillation a small amount (about 15%)
of a yellow oil boiling at 104 to 110 at 11 mm. It had a fragrant, ter-
pene-like odor very similar to that of /3-polychloroprene and was a mild
lachrymator. It showed no tendency to polymerize further.

co-Polybromoprene. The formation of co-polybromoprene, like the
formation of co-polychloroprene, occurs under conditions which result in
very slow polymerization. For example, samples of bromoprene con-
taining 0.2% of phenyl-/3-naphthylamine or 0.5% of tetramethylthiuram
disulfide and 0.3% of sulfur slowly became more viscous and, after two to
six weeks, a white deposit having a crystalline appearance began to form.
After formation of this substance had started, conversion of the whole
mass was complete in a few days. The product turned dark brown in
color on standing in air. co-Polybromoprene is soft, opaque and non-

320 ACETYLENE POLYMERS AND THEIR DERIVATIVES

coherent, while a>-polychloroprene is a mass of glistening, hard, rubbery
granules. Like co-polychloroprene, it is not swelled by benzene.

The Structure of Polybromoprene. Evidence has already been pre-
sented for concluding that a-polychloropreiie is precisely analogous to
natural rubber in its chemical structure. The analogies between poly-
chloroprene and polybromoprene are sufficiently close to justify the as-
sumption of a similar structure (III) for the latter compound. The x-ray
evidence is especially interesting. The fact that ju-polychloroprene when
stretched furnishes a fiber diffraction pattern has already been disclosed (1).
Polybromoprene shows a similar behavior, but it furnishes an even sharper
pattern (Fig. 1) (9). So far as we are aware, it has not been possible to ob-
tain fiber diagrams from any other synthetic rubbers, and this fact per-
haps justifies the conclusion that polychloroprene and polybromoprene
are more regular in their molecular structure than any other known syn-
thetic rubbers.

(Til) . . . CH 2 CH=C CH 2 - -CH-r CH=C CH., CHa CH=C CH,

I I I

Br Br Br

Summary

The polymerization of bromoprene is closely analogous to that of chloro-
prene, but somewhat more rapid. Spontaneous polymerization yields as
the final product ju-polybromoprene, which resembles vulcanized rubber
but is more dense than rubber or /x-polychloroprene. A plastic (a) poly-
bromoprene is readily isolated from partially polymerized bromoprene,
and it is converted to the //-product by the action of heat. At elevated
temperatures in the presence of inhibitors a volatile liquid (|8) polymer is
formed. w-Polybromoprene, a granular, insoluble, rubber-like mass, is
produced under conditions that lead to very slow polymerization.

Bibliography and Remarks

(1) Carothers, Williams, Collins and Kirby, Journ. Am. Chem. Soc., 53, 4203 (1931; this volume,
page 281).

(2) Carothers, Collins and Kirby, Journ Am Chem Soc., 65, 786 (1933); this volume, page 311.

(3) Staudinger, Brunner and Feisst, Helv Chim. Acta., 13, 805 (1930).

(4) With the aid of swelling agents, softener and lubricants it is possible to confer a certain amount
of plasticity on /^-polychloroprene It is also possible to obtain from chloroprene plastic polymers
that, on being heated, losa their plastic properties very incompletely or not at all. The material that
we refer to as a-polychloroprene is an inherently plastic, polymerizable polymer; its plastic properties
are completely lost and its elastic properties become fully developed when it stands or is heated.

(5) Conant and Tongberg, Journ. Am. Chem. Soc., 52, 1667 (1930)

(6) According to observations made by Dr. H. W. Starkweather in this laboratory, isoprene
polymers prepared in this manner are also completely non-plastic

(7) Ref 1, page 4219, this volume, page 298.

(8) Ref. 1, page 4211; this volume, page 290.

(9) The x-ray data will be discussed in more detail in a future paper by Dr A. W. Kenney [Such
a paper has not apparently been published hitherto. Kus.]

VI. REACTIONS OF VINYLMAGNESIUM BROMIDE 321

VI. Vinylethinylmagnesium Bromide and Some of Its

Reactions*

As might be expected from its structure, vinylacetylene (I) reacts rapidly
with ethylmagnesium bromide. The reaction proceeds smoothly and
apparently involves only the acetyleriic hydrogen; the behavior of the
product indicates that it is Vinylethinylmagnesium bromide (II). It re-
acts in the typical manner with a variety of reagents.

Acetone yields vinylethinyldimethylcarbinol (III), a colorless liquid
whose structure is established by its hydrogenation to n-butyldimethyl-
carbinol. On standing, it polymerizes to a colorless, transparent resin.

The action of carbon dioxide on Vinylethinylmagnesium bromide ap-
parently gives vinylpropiolic acid (IV), but it was not found possible to
isolate this substance in a state of high purity. Above 110 it polymerizes
explosively, and even at lower temperatures it is rapidly converted into a
tough, insoluble, rather elastic mass.

The products (V and VI) obtained from a-naphthyl isocyanate and from
triphenylchloromethane are stable crystalline solids.

CH2=CH C=iC MgBr CH, CTI C=C COH(CH 3 ) 2
II III

CHa= CH-C-EC COOH CH 2 CH C=C CO NH CioH l7
IV V

CH2=CH C--C C(C 6 H 5 ) S
VI

Experimental Part

Preparation of Vinylethinylmagnesium Bromide (II). The reaction vessel was
provided with a refrigerated return condenser (ice salt) and dropping funnel. The
vinylacetylene (10 to 20% excess) dissolved in ether was added in portions with con-
tinuous stirring to the ethylmagiiesium bromide. The reaction proceeded smoothly
with sufficient evolution of heat to keep the ether refluxing gently. The reaction prod-
uct remained dissolved in the ether.

Vinylethmyldimethylcarbmol (III). Forty grams of purified acetone was added
slowly to a 15% excess of Vinylethinylmagnesium bromide. The product of the reaction
distilled without residue at 5059 (15 mm.). On redistillation, 40 g. of a colorless
liquid boiling at 5961 (17 mm.) was collected. The yield in pure vinylethinyldi-
methylcarbinol was 53%. B. p. 67 (24 mm.); 2 T ? 1.4778; <# 0.8872; MR calcd.,
33.32; found, 35.07. Exaltation, 1.75.

Anal. Calcd. for C 7 H 10 O: C, 76.36i ; H, 9.09; mol. wt., 110. Found: C, 76.18;
H, 8.88; mol. wt. (cryoscopic in benzene), 119.

* W. H. Carothers and G. J. Berchet, Journ. Am. Chern. Soc. t 55, 1094-6(1933) ; Con-
tribution No. 110 from the Experimental Station of E. I. du Pont de Nemours and Co.
Received August II, 1932. Published March 7, 1933.

322 ACETYLENE POLYMERS AND THEIR DERIVATIVES

The carbinol became increasingly viscous on standing. After two weeks it was a
hard, tough, transparent mass insoluble in the common organic solvents. This trans-
formation was accompanied by the absorption of oxygen; the analytical values for car-
bon and hydrogen became progressively lower. The values for the completely poly-
merized product were C, 70.7, and H, 8.7. The polymerization was greatly retarded by
the presence of hydroquinone.

Hydrogenation of Vmylethmyldimethylcarbinol. Twenty grams of the carbinol
was dissolved in 75 cc. of alcohol. Four-tenths of a gram of platinum oxide was added
and the mixture was shaken in a reduction apparatus. It absorbed 0.506 mole of hy-
drogen in forty-five minutes, or about 93% of the theoretical amount, calculated for
three moles of hydrogen absorbed per mole of carbinol. After evaporation of the alco-
hol the residue distilled at 71-72 at 48 mm. It was a colorless liquid with a pleasant
camphor-like odor. Its physical constants agreed closely with those given in the litera-
ture for dimethylbutylcarbinol.

Found Given in the literature

B. p. 139.5-141 (7B1 mm.) B. p. 141-142 (755 mm.)

w 2 D 3 1.4189 n D 1.41592

dll 0.817 d 0.8155

a Henry and Dewaei, Bull. Acad. Roy. Belg., 957 (1908); Chem. Z., I, 1854 (1909).

Anal. Calcd. for C 7 H 16 O: C, 72.41; H, 13.79. Found: C, 72.31, 72.37; H,
13.68, 13.64.

Vinylpropiolic Acid.- One mole of vinylethinylmagnesium bromide was treated with
dry carbon dioxide at until a color test showed the absence of any RMgBr. The
product was then decomposed with water and the aqueous solution was submitted
to continuous extraction with ether for eight hours. The solvent was evaporated in
vacua. The residue weighed 50 g. (calcd., 96 g.). Attempts to distil the product at
this stage by the usual methods always resulted in explosions. A partially successful
distillation was effected at low pressure (about 0.05 mm.) in an all-glass apparatus.
This consisted of a flask sealed to a receiver cooled in liquid air. The flask was heated to
about 60. Evaporation occurred at a moderately rapid rate, and the distillate froze
to a crystalline solid in the receiver. It became liquid below room temperature. On
redistillation in a stream of carbon dioxide most of the volatile product came over be-
tween 64 and 71 at 2 mm. It was a colorless, water-soluble liquid which turned yellow
on standing. Its molecular weight, determined by titration with TV/10 sodium hydrox-
ide, was 102, instead of the calculated value 96. The product was evidently not quite
pure. It reduced permanganate instantly in acetone solution.

A tube containing a sample of the acid was evacuated with a water-pump and sealed
off. The tube was heated for eighty-five minutes at 75, at the end of which time a rub-
ber-like yellow substance had formed. This was insoluble in water, alcohol, ether, ben-
zene and acetic acid at the boiling points of these solvents. It was partly soluble in
hot sodium hydroxide, imparting a yellow color to the solution, from which an amor-
phous solid separated on neutralization.

tt-Naphthylamide of Vinylpropiolic Acid (V). A solution of 9 g. of a-naphthyl
isocyanate in anhydrous ether was added slowly to an excess of vinylethinylmagnesium
bromide. The reaction proceeded smoothly. The mixture was refluxed for thirty
minutes, then poured onto crushed ice. On extraction of the aqueous solution with

VII. SODIUM VINYLACETYLIDE 323

ether, 1.5 g. of dinaphthylurea was left undissolved. On evaporation of the ether solu-
tion, a yellowish solid separated. After two crystallizations from 50% alcohol it was
obtained in the form of small yellowish needles melting at 125 to 126 (copper block);
yield, 9 g. It was readily soluble in ether, benzene, methanol and ethanol. It reduced
permanganate in acetone solution very rapidly. In chloroform solution it absorbed
bromine slowly with the evolution of hydrogen bromide.

Anal. Calcd. for Ci 6 H n ON: C, 81.44; H, 4.97; mol. wt., 221. Found: C,
81.50, 81.27; H, 4.92, 4.85; mol. wt. (in boiling ethylene chloride), 221, 219.

Vinylethinyltriphenylmethane (VI). A slight excess of vinylethinylmagnesium
bromide was treated with a solution of 10 g. of triphenylchloromethane in anhydrous
ether. After completion of the reaction, the mixture was worked up as usual. The
ether solution left on evaporation 7.5 g. of a yellowish crystalline solid. After recrys-
tallization from hot alcohol, it melted at 134-135 (copper block). It reduced perman-
ganate in acetone solution and absorbed bromine, though slowly, in chloroform solution.

Anal. Calcd. for C 23 H 18 : C, 93.87; H, 6.12; mol. wt., 294. Found: C, 94.17,
93.52; H, 6.14, 6.24; mol. wt. (in boiling benzene), 320, 315.

Summary

Vinylacetylene reacts with ethylmagnesium bromide yielding vinyl-
ethinylmagnesium bromide. This behaves in the typical manner toward
acetone, carbon dioxide, a-naphthyl isocyanate and triphenylchlorometh-
ane. The derivatives thus produced are described.

VII. Sodium Vinylacetylide and Vinylethinylcarbinols*

Vinylacetylene, like other true acetylenic compounds, reacts with Grig-
nard reagents to form the corresponding organo magnesium halide, and
this may be used to introduce the vinylethinyl group into compounds of
various types (1). The present paper is concerned with the formation of
sodium vinylacetylide and with its use in a similar manner, especially in the
synthesis of Vinylethinylcarbinols.

Sodium vinylacetylide is readily obtained by the action of Vinylacetylene
on metallic sodium. The metal may be applied directly to the liquid
hydrocarbon, or the latter may be diluted with an inert solvent such as
ether, toluene or liquid ammonia. The reaction occurs with great ra-
pidity if the sodium is dissolved in liquid ammonia. These methods,
however, have the disadvantage that the acetylide produced is frequently
contaminated with appreciable amounts of polymeric material. More

* W. H. Carothers and R. A. Jacobson, Journ. Am. Chew. Sac., 55, 1097-1101
(1933); Contribution No. Ill from the Experimental Station of E. I. du Pont de
Nemours and Co

Received August 11, 1932. Published March 7, 1933.

324 ACETYLENE POLYMERS AND THEIR DERIVATIVES

uniformly satisfactory results are obtained by the action of powdered
sodamide on the hydrocarbon. The latter is preferably dissolved in ether
or liquid ammonia.

The sodium vinylacetylide obtained in this manner is a dusty, white
powder, and vinylacetylene is regenerated in high yield when it is cau-
tiously treated with water in the presence of a diluent. Its stability is
sufficient to permit its storage for two or three days in a stoppered bottle,
and with some care it can be handled in the presence of air. However,
if the air is moist the acetylide sometimes ignites spontaneously. It
attacks many organic reagents with explosive violence, but cooling and
dilution permit sufficient control to obtain smooth reaction in most cases.
Isolation of the sodium compound is not necessary for its application as a
reagent; the intended reactant is preferably added directly to the mixture
resulting from the action of sodamide on vinylacetylene.

In dealing with aldehydes and ketones a still simpler and more satis-
factory procedure consists in adding powdered sodamide in portions to a
mixture of the carbonyl compound with vinylacetylene. In most cases it is
advantageous to have present a diluent such as ether. With simple
aliphatic and alicyclic ketones this method is very satisfactory. The
yields are good and large scale operations are much simpler than with the
vinylethinylmagnesitim halides. However, sodium vinylacetylide has a
more limited range of applicability in the synthesis of carbinols than has
vinylethmylmagnesium bromide. Aliphatic aldehydes are partly resinified
by the sodium compound, and un saturated aldehydes and ketones arc
usually resinned completely. The magnesium compound, however, even
with very sensitive aldehydes and ketones, generally yields the expected
monomeric carbinols (2) .

Vinylethmyldimethylcarbinol has already been described (1) and the
other carbinols listed in Table II resemble it in their properties. They react
readily with bromine and decolorize permanganate; in the presence of
platinum they are smoothly hydrogenated to the corresponding w-butyl-
carbinols. They can be distilled in vacuo without decomposition, but if the
temperature of distillation is too high the tertiary carbinols tend to lose
water with the production of the corresponding substituted divinylacety-
lenes. Vinylethmyldi-w-propylcarbinol, for example, thus leads to the
compound C^HB CH=C (C 3 H 7 )C=C CH=CH 2 . Divinylacetylene
itself was obtained by heating vinylethinylmethylcarbinol with p-toluene-
sulfonic acid.

The tertiary carbinols all polymerize spontaneously on standing (3).
The transformation sets in rather quickly and progresses to the stage of a

VII. vSODIUM VINYLACETYLIDfi 325

thick sirup during the course of a few days. The mixture then sets to a
tough rather elastic mass, but the final product, a hard, transparent glass-
like mass, is obtained only after several weeks or months. The trans-
formation is greatly accelerated by certain catalysts, e. g., benzoyl peroxide,
especially in the presence of light. The final glass-like products adhere
very tenaciously to glass. They are insoluble in the common organic
solvents. The polymerization of the carbinols also proceeds more rapidly
at elevated temperature, e. #., 100, but the final product obtained under
these conditions is fusible and completely soluble in the common organic
solvents. The spontaneous polymerization of the carbinols is strongly
inhibited by the presence of a small amount of hydroquinone. The
secondary carbinols polymerize very much more slowly than the tertiary
carbinols.

Experimental Part

Preparation of Sodium Vinylacetylide. Powdered sodamide (19.5 g, 0.5 mole)
was slowly added to a solution of 75 g. of vinylacetylene in 250 cc. of liquid ammonia.
The mixture was stirred for six hours and the ammonia evaporated in a stream of nitro-
gen, finally at (30 . The residue was a white powder (38 g., ealcd. 37 g.) which showed a
tendency to ignite spontaneously when exposed to the air. It was covered with toluene,
and water was slowly added with constant stirring at 75. The acetylide finally dis-
solved without appreciable residue in the aqueous layer. Vinylaeetylene was distilled
from the mixture and collected in a cold receiver. The yield was 21.3 g. or 82%.

Preparation of the Carbinols. -The general procedure used for the preparation of
the tertiary carbinols is illustrated by the following example. The reaction mixture
consisted of 555 g. (7.7 moles) of methyl ethyl ketone (Eastman Kodak pract.), 551 g.
(10.6 moles) of vinylaeetylene, and 500 cc. of dry ether. The solution was contained in
a 3-necked flask provided with a mercury-sealed stirrer and a coil condenser, which in
turn was connected to a trap. The reaction flask and trap were surrounded with carbon
dioxide snow and the coil condenser was kept cold in the same manner. To the cold
solution was slowly added 300 g. (7.7 moles) of powdered sodamide and stirring was con-
tinued for a total of six hours. The reaction mixture was made acid to litmus by means
of 10% sulfuric acid, the ether layer separated and dried with sodium sulfate, and then
distilled in a vacuum; 677 g. of pure vinylethinylmethylethylcarbinol was obtained;
yield 71%. The carbinol was stabilized with 0.1% hydroquinone in order to prevent
spontaneous polymerization.

The same procedure was applied to the preparation of the secondary carbinols but
the results were less satisfactory. The preparation of vinylethinylmethylcarbinol is
used as an illustration. To a solution of 44 g. (I mole) of freshly distilled acetaldehyde
and 75 g. of vinylaeetylene in 75 cc. of dry ether at -10 was slowly added 39 g. of
powdered sodamide. After a few grams of sodamide had been added, the mixture be-
came so gummy that additional ether was added. Finally, the mixture formed a cake
and stirring was discontinued. After four hours the cake was broken up with a stirring
rod, more ether added, and the suspension stirred at room temperature for three hours.
The reaction mixture was allowed to stand overnight and then decomposed with water

320 ACETYLENE POLYMERS AND THEIR DERIVATIVES

and dilute sulfuric acid, The mixture was extracted several times with ether, and the
latter dried with sodium sulfatc and distilled. A considerable quantity of ether-insoluble
resin remained in the reaction mixture. After removal of the ether and a small amount
of low boiling liquid, 25 g. (26%) of vinylethinylmethylcarbinol was collected. It was
a colorless liquid with an alcoholic odor slightly resembling that of butyl alcohol. It
reacted with dinitrobenzoyl chloride to form the 3,5-dinitrobenzoate of vinylethinyl-
methylcarbinol, white needles from dilute alcohol; m. p. 106-106.2.

Anal. Calcd. for C,3H 10 N 2 O 6 : C, 54.27; H, 3.41. Found: C, 54.34; H, 3.47.

Catalytic Reduction of the Vinylethinylcarbinols. The carbinols derived from
acetone and from methyl ethyl ketotie absorbed four atoms of hydrogen very rapidly
when dissolved in alcohol and shaken with hydrogen in the presence of Adams' platinum
oxide catalyst. The resulting w-butylcarbinols corresponded in their properties with
those already reported in the literature, and the yields were almost quantitative. Hy-
drogenation of the carbinols derived from methyl octyl ketone and from acetophenone
yielded the new saturated carbinols described below.

w-Butylmethyloctylcarbinol. -Colorless liquid, b. p. (3 mm ) 94; n 1.4418;
rf? 0.8318; MR calcd., 68.38; MR found, 68.03.

Anal. Calcd. for C u H 2 oO: C, 78.50; H, 14.02. Found: C, 77.20, 77.30; H,
14.18, 14.25.

tt-Butylmethylphenylcarbinol. Colorless liquid, b. p. (6 mm ) 107 to 109; w 2 ?
1.5118; df 0.9616; MR calcd., 55.52; M R found, 55.52.

Anal. Calcd. for C 12 H, 8 O: C, 80.90; H, 10.11. Found: C, 81.02, 81.16; H,
10.38, 10.24.

n-Propyl-5-octadiene-l,5-ine-3. When the attempt was made to distil vinyl-
ethinyldi-w-propylcarbinol in vacuo in a flask provided with a long column, dehydration
occurred and the non-aqueous distillate was the hydrocarbon n-propyl-5-octadiene-
l,5-ine-3. On redistillation it was obtained as a pale yellow liquid having a characteris-
tic odor; b. p. (6 mm ) 57 to 58; </? 0.8047; w 2 D 1 4949; MR calcd., 50.06; MR found,
53.62. The sample was perhaps not quite pure.

Anal. Calcd. for C U H 16 : C, 89.19; H, 10.81. Found: C, 87.47, 90.27, 87.73;
H, 11.11, 9.94, 10.61.

Preparation of Divinylacetylene by the Dehydration of Vinylethinylmethylcarbi-
nol. To 25 g. of -toluenesulfonic acid in a 500-cc. flask provided with a stirrer, separa-
tory funnel, and condenser was slowly added 20.8 g. of vinylethinylmethylcarbinol.
The flask was warmed on a water-bath to start the reaction, after which it proceeded
vigorously. After two hours, the condenser was replaced by a distilling column and the
contents of the flask distilled in vacuo. Divinylacetylene distilled from the mixture and
collected in a trap surrounded by solid carbon dioxide and acetone. The divinyl-
acetylene was not further purified, but was converted into the hexabromide. The
melting point of this hexabromide was identical with that of the hexabromide from a
known sample of divinylacetylene. The melting points were as follows

Hexabromide from above synthesis m. p. 105-106

Hexabromide from known sample of divinylacetylene 104-106

Mixed m. p. 105-106

VII. SODIUM VINYLACETYLIDK

327

f jj

.- g

rH rH O

Sf 3 fc 2 ?

05 a o T- a>

co 00 O i i O> T- co O co

OOr-lOCOOOrHCOl-<M

rH t^ CO CO CO CO

| "g r^-.rs,^rv,^(M^

IS -3 O>?slCOC<100OO5CD<M

^ ^ ^COCOCO^iOCO^^

c/) rH >O GO O5 O *jft ^ 00 Gft

.2 - '<? r- r- P o r* co ?? co

' J

3 1

1 2 2 S ff R co

oo oo ^ cL

CO ^ O 00

HJ O5 (M rH CO

CO CO O <M
CO CO 'O t^

o 'I

2 ^

' rH O^) GO 00 00 1^-

^ V 05 00 GO 05 00 00

CO rH t^ ^

s ^

3 ^

rH ^

J3 (N -1< -* <M

iO t^ (M -^

ro ^ ^ J^ ^

H-t -rH Cl CO l"^ 1

X 1

* a ^ ^ ^ ^ ^

GO *Q 00 CO 0>

g '

j CO O2 Ci G5 O

O rH 00 O5 CO
rH rH

E
w

d s ^

O *O CO
O O 1C <N

rH rH 00 rH

a I

CX O ^ GO O <M 00
. co t"- co r-*- co O5

d o oo c^i

w 2

W r^*

rH rH

ij ^

a ^

S(M CO (M CO
^ CO -f (M

<M t- ^ O CJ

^ g
*

O t- CO t^. 00

Oi O O> O CO
t- 00 t^ 00 00

H g

K
W O ^ W

W ffi

Q -

^ |-3 o q q q q ^ S, q S:

E
u

i "

I- w

li v

^S ,,,
2 a o Ji

^11"

W o

ji S Sf i * I
w V II o o w

O W u II || O

' ^ -. a a i

3
3

g YcJrSoO

r? I. HI S? V |]| o

III I < \ I - v

fsLirl

i W o S u

rH
_1
<

U
c/)
>>

V w^* V

mula of car
produced

o y

| III

o o
III 1

o o

-^ ffi

o o

. Ill III

"B o o,-^

r- CJ O ^

_^J D w J^ c3 4>

^ a 3 ZH o o

3 | I T^5 ^^ J*^ I"

<<^'QSSoa<

328 ACETYLENE POLYMERS AND THEIR DERIVATIVEvS

Polymerization of the Carbinols. The behavior of vinylethinylmethylethylcarbinol
is typical of the tertiary carbinols. On standing in a stoppered bottle its viscosity in-
creased very rapidly during the first few days, and at the end of a week the product was a
thick, colorless, transparent sirup. After three weeks it had set to a tough, elastic,
rather rubber-like mass which still contained a considerable proportion of unchanged
carbinol. After six weeks it had changed to a very hard, translucent mass.

Fifty grams of vinylethinylmethylethylcarbinol containing 0.5 g. of benzoyl perox-
ide was exposed to a Cooper-Hewitt light (mercury arc in glass). The product obtained
after seventy- two hours was a hard, transparent, pale yellow, glass-like resin. It was in-
soluble in the common organic solvents. When heated it softened somewhat at 125
to 150, but it did not liquefy completely even at much higher temperatures.

Twenty-five grams of the carbinol containing 0.25 g. of benzoyl peroxide was ex-
posed to a 150-watt Mazda light. In four days a product similar to that described in the
preceding example was obtained.

Twenty-five grams of the carbinol containing 0.25 g. of uranyl nitrate was exposed
to a Cooper-Hewitt light. In seventy-two hours, a hard, transparent, amber-colored
resin was obtained. When no catalyst was present under the same conditions, the
transformation to a hard resin required about one week.

A sample of vinylethinylmethylethylcarbinol was heated for several hours at 100
while a slow stream of air was bubbled through it. The product was a brown sirup. At
room temperature it solidified to a brittle resin which dissolved readily in the common
organic solvents.

Summary

Sodium vinylacetylide obtained by the action of sodium or of sodamidc
on vinylacetylene is a very reactive white powder which shows a tendency
to ignite spontaneously in the air. Under properly controlled conditions
it is a convenient reagent for introducing the vinylethinyl group into
reactive organic compounds. Vinylethinylcarbinols are readily obtained
by treating a mixture of vinylacetylene and a ketone with sodamide.
The same method can also be applied to aldehydes but with less favorable
results. Nine new carbinols prepared by this method are described.

Bibliography and Remarks

(1) Carothers and Berchet, Journ. Am Chan. Soc. t 55, 1094 (1933), this volume, page 321.

(2) Unpublished results.

(3) The behavior of the vinylethinylcyclopentauol was exceptional. It had not polymerized after
standing for one year at the ordinary conditions.

VIII. -ALKYL-0-VINYLACETYLENES 329

VIII. a-Alkyl-0-Vinylacetylenes *

Vinylacetylene is a prolific source of new and interesting compounds.
The present paper is concerned with homologs of vinylacetylene, which are
readily accessible through the mediacy of sodium vinylacetylide (1).
Vinylethinylmagnesium bromide reacts with very active alkyl halides such
as triphenylchloromethane (2), but not with simple alkyl halides. Sodium
vinylacetylide on the other hand reacts almost explosively with simple
alkyl halides. The reaction can be moderated, however, by employing
low temperatures, and when the halide is cautiously added to the acetylide
in liquid ammonia it proceeds smoothly and furnishes good yields of the a-
alkyl-/3-vinylacetylenes. Alkyl sulfates or sulfonates can be used with
similar results. The properties of alkyl vinylacetylenes obtained by these
methods are indicated in Table I. The compounds are colorless liquids
with characteristic odors. On standing they slowly polymerize, yielding
viscous, yellow sirups.

TABLE I
PHYSICAL PROPERTIES OF CH 2 =CH C-r^C R

a ^ B. p , C. </4 2 i? M R calcd - M n found Exaltation

CH 3 59 2 at 700 mm. . 7401 1 . 4490 22 . 82 23 . 94 1.12

C 2 H { , 84 . 5-85 . 3 at 758 mm . . 7492 1 . 4522 27 . 44 28 . 82 1 . 38

H-C 4 H 9 02-03 at 01 mm. . 7830 1 . 4592 30 . 08 37 .71 1 . 03

?7-C 7 H,, 74. 5 at 9 mm. .7902 14000 50.53 51.05 1.12

Experimental Part

l-Methyl-2-vinylacetylene. A one-liter, 3-necked flask was fitted with a mercury-
scaled stirrer, a dropping funnel, and an exit tube. The exit tube was connected to a
vertical condenser which in turn led to a gas-washing train consisting of an empty bottle,
a second bottle containing water, and a third containing 10% sulfuric acid. The exit
tube from the latter was connected to a calcium chloride drying tower and this in turn
led to a receiver immersed in a Dewar flask maintained at 78.

To a solution of 104 g. of vinylacetylene in GOO cc. of liquid ammonia was slowly
added 58.5 g. (1 5 moles) of powdered sodamidc. The mixture was stirred for three
hours and then concentrated to approximately 300 cc. by evaporating the ammonia
in a current of nitrogen; 189 g. (1.5 moles) of dimethyl sulfate was added slowly through
the separatory funnel. The reaction was very vigorous and about four hours were re-
quired for the addition. The ammonia was allowed to evaporate and the reaction flask
was finally heated on a water-bath. Part of the l-methyl-2-vinylacetylene collected in
the first bottle and part in the second. The portions were combined, dried over calcium

* R. A. Jacobson and W. H. Carothers, Journ. Am. Chem. Soc., 55, 1022-4(1933);
Contribution No. 112 from the Experimental Station of E. I. du Pont de Nemours and
Co.

Received September 21, 1932. Published April 0, 1933.

330 ACETYLENE POLYMERS AND THEIR DERIVATIVES

chloride, and distilled. Some low-boiling material came over first and then 37.6 g.
(38%) of l-methyl-2-vinylacetylene was collected. It was a colorless, volatile liquid
possessing a powerful hydrocarbon odor somewhat similar to that of vinylacetylene.

Anal. Calcd. for C 6 H 6 : C, 90.91 ; H, 9.09. Found: C, 90.97; H, 8.63. Mol.Wt.:
calcd., 66; found, 66.1, 66.3 (cryoscopic, benzene).

l-Ethyl-2 -vinylacetylene. A solution of 104 g. (1.5 moles) of vinylacetylene in
500 cc. of liquid ammonia was treated with 58.5 g. (1.5 moles) of powdered sodamide as in
the preceding experiment. After three hours, 231 g. of diethyl sulfate was added slowly
through the dropping funnel. The mixture was allowed to stand overnight while the
ammonia evaporated. Water was added to the reaction flask and the upper layer,
weighing 51 g., was separated. After drying with calcium chloride, the liquid was dis-
tilled. A small amount of low-boiling liquid came over first, after which 37 g. of 1-ethyl-
2-vinylacetylene distilled. The product was a colorless liquid with an odor similar to
that of l-methyl-2-vinylacetylene.

Anal. Calcd. for C 6 H 8 : C, 90; H, 10. Found: C, 89.47, 89.73; H, 10.20, 9.44.
Mol. Wt. : calcd., 80; found, 78.4, 79.5 (cryoscopic, benzene).

l-Ethyl-2- vinylacetylene was also prepared by treating sodium vinylacetylidc
with ethyl ^-toluenesulfonate according to the method recently employed by Truchet
(3): 58.5 g. (1.5 moles) of sodamide was slowly added to a solution of 104 g. of vinyl-
acetylene in 200 cc. of butyl ether at 10 and the mixture was stirred for three hours.
A solution of 300 g. (1.5 moles) of ethyl ^-toluenesulfonate in 200 cc. of butyl ether was
then added drop by drop during several hours. The thick mixture was heated in a water-
bath at 80 for three hours and allowed to stand overnight. Water was added but such
a troublesome emulsion formed that the mixture was set aside for twenty-four hours in a
separatory funnel. The ether layer was separated, dried with calcium chloride, and dis-
tilled. A considerable quantity of low-boiling material first distilled, and then a fraction
weighing 45 g. and boiling at 78-88 was collected. Upon redistillation 28.5 g. (23.7%)
of 1 -ethyl-2-vinylacetylene boiling at 8485 was obtained. Of the two methods of pre-
paring this compound, the first was the better.

l-Butyl-2-vinylacetylene. A solution of 100 g. of vinylacetylene in 400 cc. of
liquid ammonia was treated with 39 g. (1 mole) of powdered sodamide. After three
hours, 137 g. (1 mole) of butyl bromide was slowly dropped into the solution during
about four hours. The mixture was allowed to stand overnight, water was added, and
the upper layer separated. The liquid was dried with calcium chloride and distilled.
The product was a colorless liquid with a characteristic hydrocarbon-like odor. The
liquid polymerized during the course of three months to a yellow viscous sirup.

Anal. Calcd. for C S H I2 : C, 88.88; H, 11.12. Found: C, 88.65; H, 10.70.
Mol. Wt. : calcd., 108; found, 103, 104 (cryoscopic, benzene).

l-Heptyl-2-vinylacetylene. A solution of 75 g. of vinylacetylene in 400 cc. of
liquid ammonia was treated with 39 g. (1 mole) of powdered sodamide. After three
hours, 150 g. (0.84 mole) of heptyl bromide was slowly added during four hours. The
mixture was allowed to stand overnight, water was added, and the upper layer (128 g.)
separated. After drying with calcium chloride, the liquid was distilled; 101 g. of 1-
heptyl-2-vinylacetylene was collected. Based on the heptyl bromide used, the yield
was 80% of the theoretical. On standing for two months the liquid polymerized to a
yellow viscous sirup.

IX. 1-ALKYL-2-CHLORO-1.3-BUTADIENES 331

Anal. Calcd. for d,H 18 : C, 88.00; H, 12.00. Found: C, 87.65, 87.51; H,
11.55, 11.81. Mol. Wt. : ealcd., 150; found, 143, 145 (cryoscopic, benzene).

Summary

Sodium vinylacetylide reacts with alkyl halides, sulfates or sulfonates,
yielding a-alkyl-/3-vinylacetylenes. Compounds of the formula CH^-^
CH C^C R are described in which R is methyl, ethyl, n-butyl and
w-heptyl.

Bibliography and Remarks

(1) Carothers and Jacobson, Journ. Am. Chem. Soc. t 55, 1097 (1933), this volume, page 323.

(2) Carothers and Berchet, Journ. Am. Chem. Soc., 55, 1094 (1933), this volume, page 321.

(3) Truchet, Compt. rend., 191, 854 (1930).

IX. l-Alkyl-2-chloro-l,3-butadienes and their Polymers
(Fourth Paper on New Synthetic Rubbers)*

Chloroprene (I) polymerizes very rapidly to form a rubber-like product of
excellent quality (1).

1234
CH 2 -=C CH CH 2 (I)

Replacement of the hydrogen at the 3-position by methyl does not appreci-
ably affect the rate of spontaneous polymerization, but the rubber-like
product is somewhat deficient in extensibility (2). On the other hand, the
introduction of methyl at both the 3- and the 4-positions greatly diminishes
the tendency to polymerize, and the product, although highly extensible,
is lacking in resilience (2).

The present paper is concerned with chloroprenes in which a hydrogen
in the 1 -position has been replaced by alkyl. These compounds are readily
obtained by the action of hydrogen chloride on the corresponding a- sub-
stituted vinylacetylenes. The latter have already been described (3).
Doubtless because of their lesser solubility in water they react more slowly
with aqueous hydrochloric acid than does the parent hydrocarbon, which
reacts practically completely when shaken for five hours at room tempera-
ture with two moles of concentrated hydrochloric acid containing cuprous
chloride (4). Under the same conditions a-methylvinylacetylene is less

* R. A. Jacobson and W. H. Carothers, Journ. Am. Chem. Soc., 55, 1624-7 (1933);
Contribution No. 113 from the Experimental Station of E. I. du Pont de Nemours and Co.
Received September 28, 1932. Published April 6, 1933,

332 ACETYLENE POLYMERS AND THEIR DERIVATIVES

than 40% utilized. Reaction of the higher homologs is still slower, so that
elevated temperature was needed for the butyl compound, and the heptyl
compound required the addition of alcohol to function as a solvent.

The physical properties of the new homologs of chloroprene are indicated
in Table I. They are colorless liquids with characteristic odors. Like
chloroprene, they react with a-naphthoquinone to form addition products,
which are readily oxidized to a-alkyl-/3-chloroanthraquinoiies (II), and
their identity is established by this reaction.

No additions to systems of the type alkyl C^C -CH^- CH 2 have been
recorded hitherto. It is therefore of interest to observe that the result
of adding hydrogen chloride under the conditions described is to place H
at (1) and Cl at (2), but in view of the mechanism already established for
vinylacetylene (4) it seems probable that the first product is alkyl CH --
C -CH CH 2 C1, which then rearranges to the substituted chloroprene. In
any event, current theories, either empirical or electronic, do not appear to
account for this result.

TABLE I
PROPERTIES OF CH,-=CH CC1--CH R

Nature of
R

B. p , C.

Pressure
in mm

"1?

</r

A/K

Calcd.

A/K

Found

Kxalla
tion

H 1

59.4

700

1.4583

0.956,3

24.61

25.27

CH 3

99.5-101.5

759

1 . 4785

9576

29.22

30.32

1.10

C 2 H 6

08.2-09

117

1.4770

9390

33 . 84

35.05

1 21

w-C 4 H 9

04 65

1.4794

. 9360

43 . 04

43 77

o 7;j

w-C 7 H 15

74-76

1 4785

.9141

56 9.3

57 79

Polymerization of the Substituted Chloroprenes. Chloroprene polymerizes prac-
tically completely to an elastic, rubber-like mass in forty-eight to eighty hours when di-
rectly illuminated by a 150-watt Mazda lamp at 30 to 35. The substituted chloro-
prenes listed in Table I all polymerize much more slowly. The methyl compound re-
quires about six or seven weeks, and the higher members of the series require still longer
times. The products, with the possible exception of that derived from the heptyl com-
pound, are definitely rubber-like but much inferior in quality to poly chloroprene. The
product from the methyl compound is the best. Polymerization under direct light from
a 150-watt Mazda lamp during one month at the ordinary temperature gave a soft mass
containing a considerable proportion of unchanged monomer. The polymer precipi-

IX. l-ALKYL-2-CHLORO-l,3-BUTADIENES 333

tated by the addition of alcohol was a soft, plastic mass resembling milled smoked sheets.
It was compounded with 10% of its weight of zinc oxide and then heated at 120 for
twenty minutes. The plastic properties were quite largely suppressed by this treat-
ment, but the product appeared to be incompletely vulcanized. It was strong and
tough and had a high extensibility, but it recovered from stretch rather slowly, and was
deficient in resilience. Vulcanizates obtained from higher members of the series were
still vSofter and more deficient in resilience.

Preparation of l-Alkyl-2-chloro-l,3-butadienes. The general procedure was simi-
lar to that already described for chloroprene (4). One mole of the hydrocarbon was
shaken with approximately 2.2 moles of concentrated aqueous hydrochloric acid con-
taining about 0.25 mole of cuprous chloride and 0.2 mole of ammonium chloride. The
nonaqueous layer was separated, stabilized with hydroquinone and distilled; or, in some
cases, the reaction mixture was distilled with steam after the addition of hydroquinone.
The reaction time varied from five to sixteen hours and the temperature from 23 to

ANALYTICAL DATA FOR THE l-ALKYL-2-CHLORO-l,3-BUTADiENES

Nature of
alkyl C

-Calculatec
Cl

Fo
I

und

Mol wt
(cryoscopic
in benzene)

.
Mol.

wt.

CH 3 58.56

6.83

.61

102.

58.49

.92

34.63

102

103

C 2 H 6 61.82

7.73

30.

.45

116.

61.64

,97

30.28

116

119

tf-C 4 H 9 66.45

8.99

24,

.56

144.

65.73

.76

24.57

152

1.54

>/-C 7 Hs & 70.79 10.19 19.02 186.5 70.39 10.04 19.26 189 193

45. The product in each case appeared to consist entirely of the substituted chloro-
prcnc and the unchanged hydrocarbon. Yields were good, but conversions were not
complete. No appreciable addition could be obtained with the heptyl compound in
aqueous solution, but a fair conversion was obtained when the hydrocarbon was shaken
with 3.4 moles of hydrochloric acid in ethyl alcohol with 0.35 mole of cuprous chloride
and 0.44 mole of ammonium chloride for five hours at 70 to 80.

Condensation of Naphthoquinone with the l-Alkyl-2-chloro-l,3-butadienes.
The substituted chloroprenes were each heated with a-naphthoquinone in the ratio of
about 1 g. to 0.5 g. at 100 for about two hours. Alcohol containing sodium hydroxide
was added, and air was bubbled through the resulting suspension. The solids were
crystallized as indicated below and were thus obtained in the form of yellow crystals
(generally needles).

SUBSTITUTED ANTHRAQUINONES OF FORMULA II

Calcd.

Found

R =

Cryt>t. from

M. p , C.

CH 3

Acetic acid

181

70.19

3.51

70 16

3.50

C 2 H h

Alcohol

151-152

70.99

4.06

70.93

4.18

-C 4 H 9

Methanol

129-130

72.37

5.03

71.83

5.16

tt-C 7 Hi5 a

Alcohol

112.5-113.5

74.02

6.17

74.12

6.40

a In this case the intermediate tetrahydro compound was isolated, probably 1-
heptyl-2-chloro-4,4a,9,9a-tetrahydro-9, 10-anthraquinone ; white crystals from methanol,
m. p. 96-98.

Anal. Calcd. for C 2 iH 2 6O,.Cl: C, 73.16; H, 7.26. Found: C, 72.48; H, 7.44.

334 ACETYLENE POLYMERS AND THEIR DERIVATIVES

Polymerization of the l-Alkyl-2-chloro-l,3-butadienes

l-Methyl-2-chloro-l,3-butadiene.-~ A sample of l-methyl-2-chloro-l,3-butadiene
was exposed at room temperature to an ordinary incandescent light (150-watt Mazda)
for one month. During this period the liquid increased in viscosity slowly at first,
but more rapidly later until a soft, transparent, elastic, rubber-like solid was obtained.
The rubber-like solid was macerated with alcohol to remove monomer and polymers of
low molecular weight. The residual polymer was a tough, rubbery, plastic material.
Ten per cent, by weight of zinc oxide was incorporated by means of steel rolls and the
plastic mass then heated at 120 for twenty minutes. The product was strong, tough
and elastic, but recovery from stretch was rather slow.

A sample of l-methyl-2-chloro-l,3-butadiene was exposed to a 150-watt Mazda
light at 30-35 for six and one-half weeks. The liquid progressively increased in vis-
cosity until finally a pale yellow rubbery solid was obtained. This product was more
completely polymerized than the product of the preceding experiment. The polymer
was highly elastic and resembled cured natural rubber.

l-Ethyl-2-chloro-l,3-butadiene. A sample of l-ethyl-2-chloro-l,3-butadiene was
exposed to the light of a 150-watt Mazda lamp at 30 to 35 for about four weeks. The
product was a pale yellow, transparent, viscous sirup. This was macerated with a large
volume of alcohol to remove unchanged monomer. The soft, sticky, coherent mass that
remained undissolved by the alcohol was mixed with 10% of zinc oxide and heated in a
mold for twenty minutes at 120. The product was a tough, elastic material resembling
the vulcanized product obtained from l-methyl-2-chloro-l,3-butadiene, but it was
softer and less resilient.

l-Butyl-2-chloro-l,3-butadiene. A sample of l-butyl-2-chloro-l,3-butadiene,
when allowed to stand at the ordinary conditions in a stoppered bottle, after nine and
one-half months had changed to a thick, sticky, brown sirup. Precipitation with alco-
hol then gave a 60% yield of rather soft, rubber-like material. The reaction was accel-
erated by light. A sample directly exposed to a 150-watt Mazda lamp at 30 to 35 for
six weeks had changed to a yellow, viscous, sirup. After five weeks more it was consider-
ably thicker. It was washed with alcohol arid the residual sticky solid was mixed with
zinc oxide. It vulcanized very incompletely on being heated.

A sample of l-butyl-2-chloro-l,3-butadiene was subjected to a pressure of 6000
atmospheres at 38. At the end of ninety-six hours it had polymerized to a transparent
soft, sticky solid; 90% of this solid was now insoluble in alcohol. The portion insoluble
in alcohol was mixed with 10% of its weight of zinc oxide, 2% of stearic acid, and 1% of
benzidine and then heated at 120. The physical properties of the product indicated
that it was very incompletely vulcanized. It was elastic but rather weak and somewhat
sticky.

l-Heptyl-2-chloro-l,3-butadiene. A sample of this dierie showed no apparent
change in color or viscosity when allowed to stand at the ordinary conditions for nine
months. Polymerization occurred slowly when a sample was exposed to light from a
Mazda lamp at 30 to 35. After three and one-half weeks the sample had changed to a
colorless, transparent sirup. After six more weeks it had become thick and very vis-
cous. It was finally washed with alcohol and attempts were made to vulcanize the in-
soluble polymer in the presence of zinc oxide. The product was soft and sticky.

A sample of l-hcptyl-2-chloro-l,3-butadiene was subjected to a pressure of 6000
atmospheres at 38. At the end of ninety-six hours it had polymerized to a transparent,

X. CHLORINATION OF VINYLACETYLRNR HYDROCHLORIDE 335

sticky, elastic mass. Only 4% of the material was soluble in alcohol. The alcohol-
insoluble polymer was mixed with 10% of zinc oxide, 2% of stearic acid and 1% of
benzidine and heated at 120. The product was a sticky solid possessing very slight
elasticity.

Acknowledgment. We are indebted to Dr. H. W. Starkweather for
the experiments at high pressure.

Summary

Substituted chloroprenes of the formula CH 2 CH CC1=CH R in
which R is methyl, ethyl, w-butyl and w-heptyl are described. These
compounds all polymerize much more slowly than chloroprene and the
polymers, though rubber-like, are inferior in quality to poly chloroprene.
The methyl compound polymerizes most rapidly and yields the best poly-
mer, but compared with polychloroprene the polymer is lacking in re-
silience.

Substituted anthraquinones derived from the substituted chloroprenes
are described.

Bibliography and Remarks

(1) Carothers, Williams, Collins and Kirby, Journ Am Chem. Soe , 53, 4203 (1931), this volume,
page 281

(2) Carothers and Coffman, Jonrn Am. Chem Soc., 54, 4071 (1932), this volume, page 384.

(3) Jacobson and Carothers, Journ Am Chem Soc., 55, 1622 (1933), this volume, page 329.

(4) Carothers, Berchet and Collins, Journ Am Chem Soc , 54, 4000 (1932); this volume, page 3()<>.

X. The Chlorination of the Hydrochlorides of Vinylacetylene*

The action of hydrogen chloride on vinylacetylene gives rise to the two
monohydrochlorides, chloro-4-butadiene-l,2 (I) and chloroprene (II), and
the dihydrochloride, dichloro-2,4-butene-2 (III) (1).

HC1
CHE=C- CH CH, > CH 2 ==C==CH CH,C1 >

(I)
HC1

CHa=-= CC1 CH=CH 2 > CH 3 CC1=CH CH 2 C1
(II) ' (HI)

All of these compounds react rapidly with chlorine and we now record some
observations on the products to which they lead.

* W. H. Carothers and G. J. Berchet, Journ. Am. Chem. Soc., 55, 1628-31(1933);
Contribution No. 144 from the Experimental Station of E. I. du Pont de Nemours and
Co.

Received September 28, 1932. Published April 0, 1933.

,336 ACETYLENE POLYMERS AND THEIR DERIVATIVES

Chloro-4-butadiene-l,2 in a series of experiments was chlorinated under
various conditions and the combined distillable product was redistilled
through an efficient column. Most of the material segregated into two
fractions, one boiling at 40 to 41 at 10 mm. and the other at 64 to 65 at
10 mm. Each had the composition C 4 H 5 Cl3. The lower boiling com-
pound on oxidation with permanganate yielded a,/3-dichloropropionic acid
as the only recognizable product, and was thus identified as trichloro-1,2,3-
butene-3 (IV). The higher boiling compound when similarly oxidized
yielded only chloroacetic acid. The compound was therefore trichloro-
l,3,4-butene-2 (V). The two compounds evidently arise by the addition
of chlorine at the interior and the terminal members of the pair of con-
tiguous double bonds. Other allenes have been observed to behave in a
similar manner (2).

, CHr-CCl CHC1 CH 2 C1 v

CHo C^CH CHoCl ( (IV) 2 CH 2 C1CC1 2 CHC1 CH 2 C1

N CH,C1 CC1=CH CH 2 C1 ' (VII)

(V)

Further experiments showed that although the two trichlorobutenes were
the principal products when one mole of chlorine was added to chloro-4-
butadiene-1,2, the ratio in which they were formed varied considerably with
the conditions used. At 40 to 50 the 1,3,4-compound predominated,
at 60 to 70 the 1,2,3-compound. Saturation of chloro-4-butadiene-
1,2 with chlorine led to a compound having the composition C^sCU.
In view of its origin it may be assigned the formula pentachloro- 1,2, 3,3,4-
butane (VII).

In all of the chlormations of chloro-4-butadiene-l,2 a considerable frac-
tion (up to about 20% of the total product) consisted of undistillable ma-
terial. In the chlorination of chloroprene (1 mole : 1 mole) the proportion
of undistillable product was still greater (up to about 50%). The distil-
lable product formed either at 40 to 50 or at 60 to 70 was a difficulty
separable mixture, but the principal fraction (25 to 30% of the total
product) was closely similar in its physical properties to the compound al-
ready identified as trichloro-l,3,4-butene-2. This obviously might arise
from chloroprene by a process of 1,4 addition.

The chlorination of dichloro-2,4-butene-2 leads to the formation of
trichloro-l,2,3-butene-3 (IV), tetrachloro-l,2,3,3-butane (VI) and penta-
chloro-l,2,3,3,4-butane (VII). This series of products might be accounted
for as follows

X. CHLORINATION OF VINYLACETYLENE HYDROCHLORIDES 337

+ Cl a -HC1
CH 3 CC1=CH CH 2 C1 > CH 3 CC1 2 -~CHC1--CH 2 C1 >

(VI)

+ C1,

CH 2 CC1 CHC1 CH 2 C1 > C1CH 3 CC1 2 CHC1 CH 2 C1

(IV) (VII)

The isolated tetrachlorobutane however is a stable compound, that is,
it shows no tendency to lose hydrogen chloride spontaneously. More-
over, a very curious feature of the reaction lies in the fact that at elevated
temperature (e. g., 40 to 60) very little hydrogen chloride is evolved
and the predominating product is the tetrachlorobutane; at low tempera-
tures (e. g., 60 to 70), on the other hand, hydrogen chloride is formed
in copious amounts, and the predominating product is the trichlorobutene
or the pentachlorobutane, depending upon the amount of chlorine applied.
It appears therefore that the trichlorobutene does not originate from the
tetrachlorobutane as such, but that both of these compounds arise from a
common prior intermediate. Such an intermediate might be VIII, which
would be formed by the addition of the chlorine molecule at the deficient
carbon in the active form of the double bond (3). Rearrangement of this
intermediate would lead directly to tetrachlorobutane; loss of hydrogen
chloride followed by rearrangement would lead to trichlorobutene. A
sufficient difference in the temperature coefficients of the primary processes
would account for the observed facts.

CH 3 C CH CH 2 C1
C1+

Cl
(VIII)

The above statements in regard to the influence of temperature on the
chlorination of dichloro-2,4-butene-2 are illustrated by the following ex-
periments.

One mole (125 g.) of the dichlorobutene in a flask cooled with a slush of
solid carbon dioxide and acetone was treated with chlorine gas until 1.33
moles of chlorine had been absorbed. The mixture was allowed to warm
up while a slow stream of air was passed through it to remove dissolved
gases. During this operation there was a sudden and very copious evolu-
tion of hydrogen chloride. The loss in weight corresponded with the
evolution of 0,9 mole of hydrogen chloride. The hydrogen chloride

338 ACETYLENE POLYMERS AND THEIR DERIVATIVES

collected and titrated was only 0.73 mole, but some was lost owing to
the suddenness of the evolution. Distillation of the mixture gave

Trichloro-l,2,3-butene-3 . 440 mole

Tetrachloro-1 ,2,3,3-butane . 089 mole

Pentackloro-l,2,3,3,4-butane 0. 198 mole

The rest (0.273 mole) was contained in intermediate fractions and residue;
the latter comprised 10.3% of the total product. If the losses are dis-
tributed proportionately among the three chief fractions, the percentage
yields of these are

Trichlorobutene 60.6% Tetrachlorobutane 12.2% Pentachlorobutane 27.2%
Since in the formation of the pentachloro compound the trichloro com-
pound is a necessary intermediate, 88% of the dichlorobutene was con-
verted to the trichlorobutene, and only 12.2% to the tetrachlorobutane.

One mole (250 g.) of the dichlorobutene was chlorinated at ordinary
temperature, the rate being controlled so that the temperature of the mix-
ture was between 45 and 60. The increase in weight corresponded with
the absorption of 0.01 mole of chlorine, and at the same time 0.36 mole of
hydrogen chloride was liberated. Distillation of the mixture gave

Trichloro-1 ,2,3-butene-3 . 207 mole

Tetrachloro-l,2,3,3-butane . 343 mole

Pentachloro-l,2,3,3,4-butane . 100 mole

Chlorination in this case was incomplete and there was considerable loss in
residue and intermediate fractions, but the data show that the yield of
tetrachlorobutane was greater than the combined yields of the trichloro
and the pentachloro compounds. Further experiments showed that by
careful adjustment of the amount of chlorine applied at low temperatures
the isolated yield of trichloro-l,2,3-butene-3 could be raised to 80%, while
at high temperatures the yield did not exceed 25%.

Oxidation of Trichloro-1, 2, 3-butene-3. Twenty grams of the com-
pound was stirred with 200 cc, of water while 70 g. of potassium perman-
ganate was added in portions. The alkaline solution was filtered, treated
with sulfur dioxide, filtered, acidified and continuously extracted with
ether for nine hours. Evaporation of the ethereal solution gave 8.5 g. of
acidic oil which distilled at 125 at 25 mm. It was identified as a,/3-di-
chloropropionic acid by its melting point (49-50) and its neutralization
equivalent (calcd., 143; found, 141.1).

Oxidation of Trichloro-1, 3, 4-butene-2. Oxidation of this compound
under the same conditions as those described above gave chloroacetic acid

X. CHLORINATION OF VINYLACETYLENE HYDROCHLORIDES 339

TABLE I
CHLORINATION PRODUCTS

Name

B p,

d'i

Calcd.

Found

Trichloro- 1

,2,3-butene-3

40-41

1.3430 1.4944

34.78

34.

Trichloro- 1

,3,

4-butene-2

64 -(

1 3843 1.5175

34.78

34.

Tetrachloro-1

,3,3-butanc

1 4204 1.4958

40.14

40.

55-57

Pentachloro-l,2,3,3,4-butanc

1.5543 1.5157

Anal. - -

45 01

44,

,77

Empirical formula

Calcd.
H

Found
H

C 4 H 6 C1 3

30.09

3.13

.77

29 . 95

3.30

.34

C 4 H fi Cl 3

30.09

3.13

.77

29.09

3 . 01

.33

C 4 H 6 C1 3

.40

71,

.83

C 4 H B C1 5

20.82

2. If)

.00

21.22

2.36

77,

.05

Structural
formula

VII

IV
V
VI
VII

as the only product. It was identified by its melting point (o'3), mixed
melting point and neutralization equivalent (calcd., 94.5; found, 95.1).

Summary

Chlorination of chloro-4-butadiene-l,2 gives mixtures of trichloro- 1,2, 3-
butene-3 and trichloro-l,3,4-butene-2 which react further to produce
pentachloro-l,2,3,3,4-butane. The chlorination of chloroprene (chloro-2-
butadiene-1,3) gives considerable amounts of trichloro- l,3,4-butene-2.
The chlorination of dichloro-2,4-buteiie-2 at 40 to 60 proceeds with little
loss of hydrogen chloride, and the product formed in largest amount is
tetrachloro-l,2,3,3-butane. At low temperatures (-60 to -70) large
amounts of hydrogen chloride are liberated during the chlorination, and the
principal products are trichloro- 1, 2, 3-butene-3 and pentachloro- 1, 2,3,3,4-
butane.

Bibliography and Remarks

(1) Carothers, Berchct and Collins, Journ Am Chcm Soc , 54, 40(iG ( 1932) , this volume, page 300.

(2) Bouis, Ann. clnm , [10] 9, 451 (1928)

(3) Carothers, Journ Am Chem. Soc , 46, 2227 (1924). Evidence that chlorinalions involve
primarily attack by the chlorine molecule has been presented by Sopcr and vSmitli, J. Chem. Soc. t 1582

340 ACETYLENE POLYMERS AND THEIR DERIVATIVES

XL Dichloro-2,3-butadiene-l,3 and
Trichloro-1 ,2 ,3-butadiene-l ,3 *

Chloroprene (I) under ordinary conditions polymerizes to a rubber-like
product about 700 times as rapidly as isoprene (or butadiene) (1). Hence
in the diene polymerization a chlorine atom at the /^-position has a powerful
activating influence. A bromine atom in the same position has an even
greater positive effect (2). On the other hand a-chlorobutadiene poly-
merizes not much more rapidly than isoprene, and the product, though
elastic, has very little strength (3). The corresponding bromine compound
also polymerizes spontaneously but no rubber-like properties have been
ascribed to it (4) . Hence the activation produced by the halogen atom is
very sensitively related to its position, and multiple substitution does not
modify this conclusion, for tetrachloro-l,2,3,4-butadiene-l,3 has been de-
scribed without any indication that it polymerizes at all (5). The effect
of a single ^-halogen atom on diene behavior is in fact unique so far as
recorded facts go. No other type of substitution has yielded a compound
that greatly exceeds butadiene in the speed of its spontaneous polymeriza-
tion and at the same time leads to a rubber-like product. In the methyl
series isoprene polymerizes somewhat more rapidly than butadiene, and
/3,7-dimethylbutadiene perhaps yet faster ((5); and the product from the
latter, though somewhat inferior in snap and extensibility, is still rubber-
like. On the other hand, a-substitution by methyl reduces the tendency
to polymerize and unfavorably affects the quality of the product (7). It
appears that both the terminal methylene groups of butadiene must be
free : if either of them is substituted even by an activating group the speed
of polymerization is diminished and/or the properties of the product are
adversely affected (8).

These conclusions stimulated interest in dichloro-2,3-butadiene-l,3 (II)
which has at once the two unsubstituted terminal methylene groups and two
chlorines properly located to produce a doubly activating effect. To provide
further comparisons trichloro- 1,2, 3 -butadiene- 1,3 (III) was also prepared.

Dibromo-2,3-butadiene-l,3 has already been described, but not very
fully: "In a sealed tube it remains limpid for several hours, then becomes
turbid, and is gradually transformed into a white polymer, but the trans-
formation is complete only at the end of several days." (9). Otherwise
information concerning its behavior is lacking.

* G. J. Berchet and W. H. Carothers, Joum. Am. Chem. Soc., 55, 2004-8 (1933);
Contribution No. 115 from the Experimental Station of E. I. du Pont de Nemours and
Co.

Received October 3, 1932. Published May 6, 1933.

XI. DICHLORO-2,3- AND TRICHLORO-1,2,3-BUTADIENE-1,3 341

Preparation of Dichloro-2,3-butadiene-l,3 and Trichloro-l,2,3-buta-
diene-1,3. Starting materials for the preparation of the two dienes were
found in the series of products obtained by chlorinating the hydrogen chlo-
ride addition products of vinylacetylene (10). The dichlorobutadiene was
obtained from trichloro-l,2,3-butene-3, and the trichlorobutadiene from
pentachloro-l,2,3,3,4-butane. The reactions starting from vinylacetylene
are

HC1 HC1
CH-=C CH=CH 2 > CHjp=CCl CH^= CH, >

(I)
Cl,
CH 3 CC1----CH CH 2 C1 > CHa=CCl CHC1 CH,C1

CH 2 C1 CC1 2 CHC1 CH 2 C1 -

Cl,

-HC1

CH 2 = CC1 CC1=CH 2
(III) (IT)

The elimination of hydrogen chloride from the trichlorobutene proceeds
very rapidly under a variety of conditions. When the trichloro compound
is mixed with a slight excess of approximately N methyl alcoholic potas-
sium hydroxide, the theoretical amount of potassium chloride precipitates
out within fifteen minutes. Dilution with water then precipitates the di-
chlorobutadiene as a heavy oil, which, after being stabilized with hydro-
quinone, is obtained in 86% yield by vacuum distillation. The losses are
almost entirely due to polymerization. There is no evidence of the forma-
tion of any by-products. Dichloro-2,3-butadiene-l,3 is also obtained by
the action of alcoholic potash on butadiene tetrachloride (tetrachloro-
1,2,3,4-butane) but the yields are less favorable.

The dichlorobutadiene could not be induced to react with naphthoqui-
none or with maleic anhydride, and hence no direct and decisive proof of
its structure was possible. The structure dichloro-l,3-butadiene-l,3 is not
excluded by the methods of synthesis used, but this structure seems very
unlikely in view of the rapidity with which the compound polymerizes.

Pentachloro-l,2,3,3,4-butane also reacts rapidly with alcoholic potash.
Besides the trichlorobutadiene the product was found to contain some tetra-
chloro-l,2,2,3-butene-3 whose identity was established by its oxidation to
a,o:,j3-trichloropropionic acid (11). This compound is doubtless an inter-
mediate in the formation of the trichlorobutadiene, and the latter should
therefore have the structure (III).

342 ACETYLENE POLYMERS AND THEIR DERIVATIVES

Polymerization of Dichloro- and Trichlorobutadiene. -Under ordinary
conditions dichlorobutadicne polymerized completely in about twenty-
four hours, at 85 to 90 in about forty minutes. The corresponding times
for chloroprene are about ten days, and about twelve hours (estimated).
Thus the dichloro compound reacts about ten times as rapidly as chloro-
prene and about seven thousand times as rapidly as isoprene. The poly-
merization is inhibited by hydroquinone and accelerated by air or by ben-
zoyl peroxide. The polymer is almost entirely devoid of rubber-like
properties. It is a white, opaque, tough, hard mass, non-plastic and lack-
ing in extensibility. It is somewhat (or partly) soluble in chloroform, but
very slightly soluble in other common organic solvents including dichloro-
butadiene, and under most conditions it separates as an opaque precipi-
tate from the dichlorobutadiene as it is formed. At elevated temperatures
(e. g., 85 to 90), however, the polymer remains dissolved until the poly-
merization is practically complete. The product is at first a soft, colorless,
transparent, sticky, elastic mass, but it becomes opaque, hard, and tough
in a very short time. The absence of rubber-like properties in the polymer
is consistent with other observations on the influence of substitutents : the
spontaneous polymers from CH 2 ^CCl--CMe==CH 2 (12) and from CH2-
CMe CMe=CH 2 are both deficient in extensibility and snap as com-
pared with those from the monosubstituted dienes CH2~CC1 CH=CH2
and CH 2 =CMe~^CH=CH 2 .

The trichlorobutadiene polymerized much more slowly than the dichloro
compound. At the ordinary conditions it changed to a jelly-like mass con-
taining about 50% of unchanged diene in ten to twelve days. After one
month polymerization was apparently complete. The product was a dark
colored, rather soft and friable mass.

Conclusions. The chlorinated 1,3-butadienes now known may be ar-
ranged in the following order so far as the speed of their spontaneous trans-
formation into high polymers is concerned

j8, 7 , 0, 7 > a, 0, 7, 5

and only the second member of the series yields a polymer that is definitely
rubber-like. The generalizations outlined in the introduction thus receive
further confirmation.

Preparation of Dichloro-2,3-butadiene-l,3 (a) from Trichloro-l,2,3-butene-3.
253 g. (15% excess) of potassium hydroxide dissolved in 760 cc. of methanol was placed
in a flask provided with a stirrer and an efficient reflux condenser, and 6,34 g. of trichloro-
l,2,3-butene-3 was added with stirring at such a rate that the temperature of the mixture
remained between 10 and 15, the flask being cooled in a bath of ice and water. After
all the trichlorobutene had been added the mixture was stirred for two hours, and the

XI. DICHLORO-2,3- AND TRICHLORO-1,2,3-BUTADIENE-1,3 343

potassium chloride was filtered off; yield, 293 g. or 99.4% of the theoretical. The filtrate
was poured into a large volume of water and the heavy oil was separated, dried with
calcium chloride and distilled in the presence of hydroquinone. The product boiled
at 39 to 41 at 80 mm. ; yield, 304 g. or 86% of the theoretical. Other properties deter-
mined on a purified specimen are : b. p. 41 to 43 at 85 mm . , 98 at 7(30 mm. ; n 2 ,? 1 .4890 ;
dl 1.1829; MR calcd., 29.47; found, 30.21.

Anal. Calcd. for C 4 H 4 C1 2 : C, 39.02; H, 3.25; Cl, 57.72. Found: C, 38.76;
H, 3.78; Cl, 57.09.

(b) From Tetrachloro-l,2,3,4-butane. -The tetrachlorobutaiie was prepared by
chlorinating butadiene. The sample used was a mixture which contained the solid and
liquid isomers in the ratio 1 :1.12 (13). In a flask provided with a stirrer and a reflux con-
denser was placed a solution of 150 g. of potassium hydroxide in 500 cc. of mcthanol.
To this solution 234 g. of the tetrachlorobutane mixed with 100 cc. of methanol was
slowly added. The temperature of the mixture was kept between 10 and 18. Po-
tassium chloride separated immediately. After the addition was complete, the mixture
was stirred for two hours at 25 and then filtered. The yield of potassium chloride was
156 g. (88%). The filtrate was poured into a large volume of water and the heavy oil
which separated was dried and distilled over hydroquinone. The distillate segre-
gated into two fractions (1) b. p. 39 to 45 at 80 mm., 57 g., and (2) b. p. 45 to 110 at
80 mm., 41 g. The first fraction was /3,7-dichlorobutadiene; on redistillation it boiled
39 to 40 at 80 mm. and it showed the correct refractive index and chlorine content.
The second fraction on redistillation boiled chiefly at 84 to 86 at 27 mm. Its chlorine
content (found, 57.67, 57.75) agreed with that required for a dichlorobutadiene, and it
was perhaps an impure isomeric dichlorobutadiene: d% 1.287; n 2 1.4999.

Preparation of Trichloro-l,2,3-butadiene-l,3 and Tetrachloro-l,2,2,3-butene-3.
One-half mole (115 g.) of pentachloro-l,2,3,3,4-butane was slowly added with stirring
to a solution of one mole of potassium hydroxide in 270 cc. of methanol. The mixture
was stirred at room temperature for about two hours. The precipitated potassium
chloride was filtered off; yield, 61 g. or 82%. The filtrate was poured into a large volume
of water, and the precipitated oil was dried and separated by distillation into two frac-
tions (a) crude trichlorobutadiene boiled at 56 to 65 at 26 mm., 44 g.; and (b) crude
tetrachlorobutene boiling at 66 to 75 at 26 mm., 24.5 g. They were further purified
by distillation.

Trichloro-l,2,3-butadiene-l,3. B. p., 33 to 34 at 7 mm ; n 1.5262; d 1.1060;
MR calcd., 34.34; MR found, 34.39.

Anal. Calcd. for C 4 H 3 C1 3 : C, 30.47; H, 1.90; Cl, 67.61. Found: C, 29.50,
29.43; H, 2.33, 2.19; Cl, 68.42, 68.50.

Tetrachloro-l,2,2,3-butene-3.- B. p. 41 to 42 at 7 mm.; >/, 2 D 1.5133; rfj 1.4602;
Af R calcd. 39.67; MR found, 39.91 .

Anal. Calcd. for C 4 H 4 C1 4 : C, 24.74; H, 2.06; Cl, 73.19. Found: C, 25.49,
25.16; H, 2.23, 2.24; Cl, 72.66, 72.16.

Oxidation. A sample (23 g.) of the tetrachlorobutene was oxidized with excess
aqueous potassium permanganate. After filtration and treatment with sulfur dioxide
the solution was extracted continuously with ether. Evaporation of the ether gave a
liquid residue which distilled at 120 to 125 at 22 mm. It crystallized on cooling and
after being washed with petroleum ether melted at 48 to 50 . Neutral equivalent found,
176.4; calcd. for trichloropropionic acid, 177.5.

344 ACETYLENE POLYMERS AND THEIR DERIVATIVES

Summary

Dichloro-2,3-butadiene-l,3 and trichloro-l,2,3-butadiene-l,3 have been
prepared and their properties are described. The dichloro compound
polymerizes more rapidly than chloroprene, the trichloro compound more
slowly than chloroprene. The polymers are not rubber-like in either case.
The chlorobutadienes now known may be arranged in the following order
so far as their speed of spontaneous polymerization is concerned: 0,y ^>
^> a, 0,7 > a ^> a,#,7,5; and only the second member of the series
(chloroprene) yields a definitely rubber-like polymer.

Bibliography and Remarks

(1) Carothers, Williams, Collins and Kirby, Journ. Am Chcm Soc , 53, 4203 (1931), this volume,
page 281.

(2) Carothers, Collins and Kirby, Journ Am Chem. Soc , 55, 789 (1933), this volume, page 314

(3) Cf. Ind Eng Chcm 26, 30 (193O).

(4) Willstatter and Bruce, Ber., 40, 3979 (1907).

(5) Muller and Huther, ibid., 64, 589 (1931). Cf. also pentachloro and hcxachlorobutadienc-
1,3, Beilstein, 4th ed., Vol. 1, page 250.

(6) Whitby and Crozier, Canadian J. Research, 6, 203 (1932); Whitby and Kat/, tbtd , 6, 398
(1932).

(7) Macallum and Whitby, Trans. Roy. Soc. Can., 22, 39 (1928); Fisher and Chittendcn, Ind
Eng. Chcm , 22, 869 (1930); Whitby and Gallay, Can. J. Res , 6, 280 (1932).

(8) This statement is also supported by observations on the following compounds in which the
speed of polymerization is in the indicated order: CH 2 =CC1 CH=CH 2 , CIl2=CCl CMe=CH 2 ^>>
MeCH=CCl CH=CH 2 . CH 2 =CC1 CMe=CHMe; Jacobson and Carothers, Journ. Am Chem. Soc ,
65, 1624 (1933); this volume, page 331; Carothers and Coffman, Journ Am Chem. Soc., 54, 4071 (1932).

(9) Lespieau and Prevost, Cnmpt rend., 180, 675 (1925).

(10) Carothers and Berchet, Journ. Am. Chem. Soc., 55, 1628 (1933), this volume, page 335

(11) The structure CHC1=CH CC1 2 CH 2 C1 is not absolutely excluded but it seems very uii
likely in view of the ease with which hydrogen chloride is lost from the group CH 2 C1 CHCla .

(12) Carothers and Coffman, Journ Am. Chem. Soc., 54, 4071 (1932); this volume, page 384

(13) Cf. Muskat and Northrup, Journ Am Chem. Soc , 52, 4054 (1930).

XII. The Addition of Thio-p-cresol to Divinylacetylene*

The reaction of acetylene with itself to form an open-chain trimer was
described in the first paper of this series (1). This compound, which has the
molecular formula C 6 H 6 , contains no true acetylenic hydrogen; on hydro-
genation it yields w-hexane; and it differs from the already known di-
methyldiacetylene (I). It was therefore assigned the structure divinyl-
acetylene (V) (2). The formulas II, III and IV, which are equally con-
sistent with the above indicated properties, may perhaps be rejected as
inherently unlikely; but in connection with a detailed study of the acety-

* W. H. Carothers, Journ. Am. Chem. Soc., 55, 2008-12 (1933); Communication
No. 86 from the Experimental Station of E. I. du Pont de Nemours and Co.
Received September 29, 1932. Published May 6, 1933.

XII. ADDITION OF THIO--CRESOL TO DIVINYLACETYLENE 345

lene trimer it seemed desirable to obtain direct and decisive experimental
proof of the structure represented by formula V.

CH 3 C^C C=C CH 3

(I)

CH 3f -=C=C==C=CII CU CH2=-C==C=CH CH=-CH 2

(II) (III)

CH2=O=CH CH=C=CH 2 CH#=CH C^C CH==CH 2

(IV) (V)

Efforts in this direction at first met with difficulties : the compound has a
great tendency to combine with itself in the presence of polar reagents,
and it does not readily yield well-defined derivatives capable of easy identifi-
cation, but a solution of the problem was found in the reaction with thio-p-
cresol.

It has been shown by Posner (3) that vinyl compounds readily add thio-
phenols. In the case of styrene the reaction proceeds in accordance with
the equation

CeHsCH CH 2 + RvSH > C6H 5 CH 2 CH 2 - vS R

Since the direction of addition alleged by Posner is exclusively the reverse
of that required by well-known empirical generalizations and by currently
popular electronic theories, the reaction was lately re-examined by Ash-
worth and Burkhardt with the result that Posner's conclusion was com-
pletely verified (4). Thus for the purpose in view the thiophenols as re-
agents have the advantage that their mode of addition is thoroughly estab-
lished. Moreover, neither heat nor catalysts are required to induce their
reaction with reactive carbon double bonds.

Acetylene trimer (one mole) readily dissolved thio-/?-cresol (two moles).
When the fluid mixture was allowed to stand at the laboratory temperature
during ten clays it gradually set to a magma of thin, transparent, leaf-
like crystals, and the odor of the thiocrcsol almost entirely disappeared.
The reaction was powerfully accelerated by light, and when the mixture
(40 g. of thiocresol and 12.5 g. of acetylene trimer) in a reagent bottle of
soda glass was directly illuminated by a mercury arc, reaction was com-
plete in about five hours. The yield of crystalline product was about 80%
of the theoretical, and some unidentified oily material was formed.

After crystallization from alcohol and from acetic acid the reaction prod-
uct melted at 74.5-75.5. Its composition agreed with that required
for di-(-tolylthio)-l,()-hexine-3 (VI). The verification of this structure
through reactions described below decisively establishes the structure of
the acetylene trimer as divinylacetylene (V). In its reaction with thio-p-

340 ACETYLENE POLYMERS AND THEIR DERIVATIVES

cresol the double bonds of divinylacetylene function independently of the
triple bond. As will be shown in future papers, however, this behavior is
not typical: other reactions indicate conjugation between the ethylenic
and acetylenic linkages.

Di-(-tolylthio)-l,6-hexine-3 readily ad4s two atoms of bromine to
form the dibromide X. The action of potassium permanganate leads to a
series of oxidation products. When the di-(/?-tolylthio)-l,6-hexine-3 is
dissolved in chloroform and shaken with cold dilute sulfuric acid to which
permanganate is added in portions, the acetylenic disulfone, IX, is obtained.
When carbon tetrachloride is used as the solvent in this process oxidation
proceeds further and one obtains the dikctone, VIII. In acetone solution,
alkaline permanganate causes rupture of the carbon chain with the forma-
tion of the already known /3--tolylsulfonepropionic acid (VII). The
acetylenic disulfone, IX, readily adds bromine to form the dibromide XI,
and sulfuric acid converts it into the ketone XII.

(X) C 7 H 7 SCH 2 CH,CBr=-CBrCH>CH,SC 7 H 7 ^

(VI) C 7 H 7 vSCH 2 CH 2 CE-CCH 2 CH 2 SC 7 H 7

KMnO 4

(VII) C 7 H 7 SO,CH 2 CH 2 COOH <

(VIII) C 7 H 7 S0 2 CH 2 CH 2 COCOCH 2 CH 2 S0 2 C 7 H 7

(IX) C 7 H 7 SO 2 CH 2 CH 2 C^CCH 2 CH 2 vSO 2 C 7 H 7

C 7 H 7 vSO 2 ClI 2 CH 2 CBr CBrCH.>CH 2 SO 2 C 7 H 7 C 7 H 7 vSO 2 CH 2 CH 2 COCH 2 CH 2 CH 2 vSO 2 C 7 H 7
(XI) (XII)

Di-(/>-tolylthio)-l,6-hexine-3 (VI). -M. p. 74.5-75.5. Anal. Calcd. for C 2 oH 2 >S,
C, 73.60; H, 6.79; S, 19.65; mol. wt., 326.3. Found: C, 73.50; H, 7.18; S, 20.1
niol. wt. (in freezing benzene), 291, 299, 302, 308.

Oxidation of Di-(p-tolylthio)-l,6-hexine-3 (VI) to -Tolylsulfonepropionic Acid
(VII). Ten grams of di-(-tolylthio)-l,6-hexine-3 suspended in 100 cc. of pure acetone
in a bottle was constantly stirred and maintained below 32 while 41 g. of powdered
potassium permanganate was added during two and a half hours. The permanganate
was rapidly reduced. The manganese dioxide was removed by filtration and washed
with acetone and water; the nitrates were evaporated, dissolved in water, treated with
decolorizing carbon, filtered and acidified. The semi-crystalline precipitate was dissolved
in a mixture of ether and chloroform, and the filtered solution was extracted with aqueous
sodium bicarbonate solution. The aqueous solution was acidified. The precipitated
solid (4 g.) melted at 102-113. After crystallization from alcohol and then from water
it was obtained in the form of needles melting at 110-1 13 (5).

XII ADDITION OF THIO--CRESOL TO DIVINYLACETYLENE 347

Anal. Calcd. for CioHi 2 O4vS: neutral equivalent, 228. Found: 234.

Partial Oxidation of Di-(/>-tolylthio)-l,6-hexine-3 (VI). Formation of Di-(/>-
tolylsulfone)-l,6-hexandione-3,4 (VIII) -To 3.2 g. of di-(/>-tolylthio)-l,6-hexine-3 in
30 cc. of carbon letrachloride, 100 cc. of 3 N sulfuric acid was added and then ice and
powdered potassium permanganate in small portions with constant shaking. About
0.5 g. of permanganate was consumed. The mixture was decolorized with aqueous
sodium bisulfite. A white solid, which was suspended in the mixture, was filtered off,
dissolved in chloroform, and precipitated with petroleum ether. After crystallization
from ethylene chloride, glacial acetic acid, and butyl acetate, it was obtained in the form
of fine, yellowish needles which softened at 197-200 and melted at 200-201.

Anal. Calcd. for C2oH 22 O 6 S 2 : C, 56.85; H, 5.22. Found: C, 50.37, 50.74; H,
5.23, 5.37.

This material was insoluble in ether, petroleum ether, ligroin, benzene, and carbon
tctrachloride; readily soluble in warm chloroform and ethylene chloride. It reacted
with phenylhydrazitie to form a crystalline derivative melting above 215.

Partial Oxidation of Di-(p-tolylthio)-l,6-hexine-3 (VI). Formation of Di-(/>-
tolylsulfone)-l,6-hexine-3 (IX) -Three grams of di-(p-tolylthio)-l,G-hexine-3 in -40
cc. of chloroform together with 100 cc. of 3 N sulfuric acid was vigorously shaken in a
stoppered bottle while ice and powdered potassium permanganate were added in small
portions until a purple color persisted in the mixture. A total of 0.7 g. of potassium
permanganate was applied. The mixture was treated with sulfur dioxide to dissolve the
precipitated manganese dioxide. The aqueous layer was discarded and the chloroform
layer was washed with dilute alkali and water, dried, filtered and evaporated. Crys-
tallization of the solid residue from dilute acetic acid gave 2.4 g. of white crystals melting
at 150-158; recrystallized from butyl acetate; needles, m. p. 157-158; insoluble in
boiling carbon tetrachloride.

Anal. Calcd. for C 20 H 2 2O 4 S 2 : C, 01.50; H, 5.08; S, 10.43. Found: C, 01.00
01.48; H, 5.85, 0.17; 8,16.54,16.68.

Action of Bromine on Di-(-tolylsulfone)-l,6-hexine-3 (IX). Formation of Di-
(/?-tolylsulfone)-l,6-dibromo-3,4-hexene-3 (XI). In 10 cc. of 13 M bromine in chloro-
form, 0.38 g. of di-(-tolylsulfoiie)-l,6-hexine-3 was dissolved. The mixture was illu-
minated by a 40-watt Mazda lamp for fifteen minutes. About 40 cc. of petroleum ether
was added and the crystalline precipitate was recrystallized from butyl acetate; m. p.
172.5-173.5.

Anal. Calcd. for C 20 H 22 O 4 S 2 Br 2 : Br, 29.00. Found: 28.87,28.92.

Action of Sulfuric Acid on Di-(-tolylsulfone)-l,6-hexine-3 (IX). Formation
of Di-(-tolylsulfone)-l,6-hexanone-3 (XII) - One gram of di-(/>-tolylsulfone)-l,0-
hcxine-3 was dissolved in 10 cc. of cold coned, sulfuric acid. The colorless solution was
allowed to stand for fifteen minutes and was then poured into 50 cc. of water. The re-
sulting white precipitate was crystallized twice from alcohol; m. p. 134-135.

Anal Calcd. for CaoHuOfiS*: 8,15.77. Found: S, 10.41.

Action of Bromine on Di-(p-tolylthio)-l,6-hexine-3 (VI). Formation of Di-(-
tolylthio)-l,6-dibromo-3,4-hexene-3 (X). Di-(-tolylthio)-l.G-hexme-3 in dilute chloro-
form solution was allowed to stand for a few minutes with an excess of standard bromine,
and the excess bromine was then titrated with standard thiosulfate. Approximately
1.05 moles of bromine was consumed by one mole of di-(-tolylthio)-l,G-hexine-3.

348 ACETYLENE POLYMERS AND THEIR DERIVATIVES

To 1.6 g. of di-(-tolylthio)-l,6-hexine-3 in 15 cc. of chloroform, 20 cc. of N bromine
in chloroform was added. The solution was allowed to stand for five minutes. It was
then shaken with 2 N thiosulfate solution, washed with water, dried with sodium sulfate,
filtered and evaporated. From the residual oil, colorless crystals gradually separated.
These were isolated and recrystallized from a mixture of alcohol and ether as trans-
parent columns, m. p. 46-47.50.

Anal. Calcd. for CaoHazSsB^: Br, 32.85. Found: Br, 32.98, 33.07.

Summary

The divinylacetylene described in a previous paper reacts with thio-p-
cresol to form a crystalline derivative whose structure as di-(^-tolylthio)-
1 ,0-hexine-3 is demonstrated by its oxidation to the known /3-p-tolyl-
sulfonepropionic acid. The following transformation products of the new
derivative are described : di-(/?-tolylthio)-l,6-dibromo-3,4-hexene-3, di-
(/?-tolylsulfone)-l ,0-hexaiidione-3,4, di-(/>-tolylsulfone)-l ,6-hexine~3, di-(p-
tolylsulfone)-! ,(i-dibromo-3,4-hexene-3, di-(/>-tolyl-sulfoiie)-l,G-hexanone-3.

Bibliography and Remarks

(1) Nieuwlanci, Calcotl, Downing and Carter, Journ. Am. Chem. Soc. t 53, 4197 (1931).

(2) This compound has perhaps been obtained in small amounts, though not completely char-
acterized, by Farmer, Laroia, Switz and Thorpe, /. Chem. Soc., 2948 (1927). See also Migaouac and
cie vSaint-Aunay, Compt. rend., 188, 959 (1929)

Lespieau and Guillemonat in a recent publication [Compt. rend., 195, 245 (1932)] have described
divinylacetylene under the title "A New Isomer of Benzene." Apparently they have entirely over-
looked the paper of Nieuwland, Calcott, Downing and Carter referred to above. They also state
that they have obtained about 12 cc. of (impure) vinylacetylene, of which hitherto only 1.4 g. has been
prepared (referring to Willstatter and Wirth), It may, therefore, be of interest to state that in the
laboratories and works of the du Pont Company many hundreds of kilograms of vinylacetylene and di
vinylacetylene has been prepared by the process of Nieuwland, Calcott, Downing and Carter.

(3) Posner, Ber , 38, 646 (1905).

(4) Ashworth and Burkhardt, /. Chem. Soc., 1791 (1928)

(5) Kohler and Reimer [Am Chem. J , 31, 175 (1904)] describe /3-/>-tolykulfonepropionic acid as
needles melting at 1 10-113. It is partially decomposed by repeated crystallization from boiling water.

XIII. The Action of Chlorine on Divinylacetylene*

Scarcely any information has been recorded concerning the behavior of
multiply conjugated enyne systems in reactions of addition. A system of
this kind is found in divinylacetylene, which has been made available
through discoveries described in the first paper of this series (1). In its
reaction with thio--cresol only the ethylenic bonds of divinylacetylene are
involved (2).

* D. D. Coffman and W. H. Carothers, Journ. Am. Chem. Soc., 55, 2040-7 (1933);
Contribution No. 116 from the Experimental Station of K. I. du Pont de Nemours and
Co.

Received October 22, 1932. Published May G, 1933.

XIII. ACTION OF CHLORINE ON DIVINYLACETYLENE 349

Chlorine behaves in a different manner. Reaction proceeds very
rapidly, and a point of apparent saturation is reached when six atoms of
chlorine have been absorbed. The product then consists of a crystalline,
volatile hexachloride (CeHeCle), and a pale yellow, very viscous sirup
which also has the composition CeHeCle. The sirup is not volatile and it
has a higher molecular weight than that required by the simple formula.
The crystalline hexachloride resists chlorination, ozonization and oxida-
tion by hot nitric acid; and the action of alkaline permanganate leads to
total destruction. Nevertheless, there can be no doubt that its structure
is correctly represented by the formula hexachloro- 1,2, 3,4,5, (>hexene-3
(IV), and a study of the di- and tetrachlorides throws some light on the
mechanism by which it is formed.

The di- and the tetrachlorides are both liquids, and they are not easily
obtained in good yields since the application of as little as one mole of
chlorine converts part of the reacting divinylacetylene to the hexachlo rides.
It appears on the other hand that only one dichloride and one tetrachloride
are produced: the same tetrachloride is obtained by the chlorination of
either divinylacetylene or its dichloride; and chlorination of the tetra-
chloride yields the hexachloride (IV) already referred to.

The dichloride still contains an open straight chain of six carbon atoms,
since it can be hydrogenated to w-hexane. Oxidation with permanganate
yields as the only identifiable product, chloroacetic acid; hence one of the
chlorines is contained in the grouping C1CH 2 C= and there is no terminal
CH.i group. The other chlorine is attached to a doubly bonded carbon,
since the action of boiling aqueous sodium carbonate liberates only one
chloride ion per molecule of the compound. The only reasonable structure
for the dichloride consistent with these facts is dichloro-1 ,4-hexatriene-2,3,5

(II).

I . CH2=CH C=C CH==CH 2

II. CICHr- CH=C=CC1CH=CH 2

III. C1CH 2 CH=CC1 CC1'-=CH CH 2 C1

IV. CICHsr-CHCl CC1== CC1 CHC1 CH,C1

Ha. C1CH 2 CHC1 C^C--CH=CH 2

lib. C1CH 2 CH=C=C=CH CH 2 C1

Ilia. CICHa CH=0=CC1 CHC1 CH 2 C1

Illb. C1CH 2 CH=CC1 CC1 2 CH=CH 2

IVa. C1CH 2 CHC1--CC1 2 CC1= CH CH 2 C1

V. CH 2 ==CH -CC1==CC1 CH=-CH 2

VI . CH 2 = CH CCl=O==C=CHs

VII. C1CH 2 CH=CH CCla CH=CH 2

VIII. CH 2 =-C CC1 CC1CH CHoCl

IX. CH 2 =C=CC1 CCl=C=CH a

:*f>0 ACETYLENE POLYMERvS AND THEIR DERIVATIVES

The tetrachloride of divinylacetylene when oxidized with potassium
permanganate also yields as the only identifiable product chloroacetic
acid, and hydrolysis with aqueous sodium carbonate shows that two of
the chlorine atoms are reactive and the other two inactive. This behav-
ior indicates that it has the structure tetrachloro-l,3,4,6-hexadiene-2,4

(in).

Formally, the relation between the three chlorides and the parent
hydrocarbon is that each member of the series is derived from the pre-
ceding one by a process of 1,4 addition. It is impossible to demonstrate
conclusively that they actually originate by this mechanism. Addition
either 1,2 or 1,6 followed by an a,y shift would lead to the same result.
The dichloride (II) might arise thus from Ila or lib, the tetrachloride from
Ilia or II Ib and the hexachloride from IVa.

However, this dual mechanism appears highly improbable since it pro-
vides no visible reason for the fact that the observed products are without
exception compounds that would result from their precursors by a process
of 1,4 addition. There are moreover other independent facts to indicate
that 1,4 addition may be a favorite mode of reaction for enyne compounds.
Thus the action of hydrogen chloride on vinylacetylene yields chloro-4-
butadiene-1,2, CH 2 ^C^CH CH 2 C1 (3), and this must be a primary
product since in the rearranged halicle, CH2=CC1 CH=CH 2 , the chlorine
is attached to a doubly bonded carbon where its mobility is lost. It is
perhaps significant in this connection also that the dichloride (II) of
divinylacetylene shows no tendency to rearrange into the completely
conjugated triene (V) under experimental conditions that might be ex-
pected to favor especially such a transformation.

When the dichloride (II) is treated with alcoholic potash it rapidly loses
one molecule of hydrogen chloride. The terminal chlorine and its ad-
jacent hydrogen are involved in this reaction, and the product has the
structure chloro-3-hexatetraene-l,3,4,5 (VI). When hydrogenated it is
converted into w-hexane, and oxidation yields oxalic acid.

In the presence of cuprous chloride the dichloride (II) absorbs one
molecule of hydrogen chloride from aqueous hydrochloric acid. In the
resulting trichloro compound all three of the chlorines are reactive toward
aqueous sodium carbonate. It is therefore the central double bond bear-
ing the inactive chlorine atom of the dichloride which disappears, and the
trichloro compound must have the structure trichloro-l,4,4-hexadiene-2,5
(VII). When oxidized it yields chloroacetic acid.

The tetrachloride (III) of divinylacetylene also loses hydrogen chloride
when treated with alcoholic potash and yields a trichloro and a dichloro

XIII. ACTION OF CHLORINE ON DIVINYLACETYLENE 351

compound. In this case only two positions are available for the loss of
hydrogen chloride, namely, the 1,2 and 5,6. The trichloro compound
must therefore be trichloro- l,3,4-hexatriene-2,4,5 (VIII) and the dichloro
compound must be dichloro-3,4-hexatetraene- 1,2,4,5 (IX).

The chloro compounds described above furnish several novel and unusual
examples of triad systems potentially capable of allylic rearrangement.
One rather striking anomaly that appears in this connection has already
been referred to. In the dichloride, II, the terminal chlorine atom shows
no tendency to undergo a shift to the 7 position, in spite of the fact that
such a shift would result in a completely conjugated triene. It is true
that the 7 carbon is at the center of a pair of twinned double bonds, but
this in itself cannot be responsible for the absence of a tendency to re-
arrange since precisely the same configuration is found in chloro-4-buta-
diene-1,2, CH 2 =C=CH CH 2 C1, which undergoes the 01,7 transposition
with great facility. In the trichloro hexadiene, VII, also, a triad shift
of one of the central chlorines would bring the double bonds into the
conjugated configuration. No special attempts have been made to bring
about rearrangement in this case, but it has not been observed to occur
spontaneously. The absence of spontaneous rearrangements in the
compounds III and VIII is not surprising, since a shift of the mobile
chlorine atom in these compounds would destroy the conjugation of the
double bonds.

Four of the chloro compounds described above (II, VI, VIII and IX)
exhibit a structural feature that has, so far as we are aware, not been
exemplified in any compounds described hitherto. They contain a pair
of conjugated double bonds and at least one of the pair is the first mem-
ber of a series of two or more contiguous double bonds, thus,
- C=C=C C=C

I II- The question arises, can the carbon atom bearing

54321

the twinned double bonds function as one of the ends of the conjugated
system? The behavior of the dichloride II toward chlorine provides an
affirmative answer to this question. The chlorine here, almost certainly
adds 1,4 to produce the tetrachloride, III. On the other hand, the addition
of hydrogen chloride to the dichloride apparently involves only the 3 and
4 positions. The trichloro compound obtained (VII) is the compound
that would be produced by 3,4 addition, and it is not likely that it arises
by 1,4 addition followed by an allylic transposition, because the product,
C1CH 2 CH=CH CC1 CH CH 2 C1, which would be formed by 1,4
addition, is precisely analogous to dichloro- l,3-butene-2, CHa -CC1

352 ACETYLENE POLYMERS AND THEIR DERIVATIVEvS

CH CH 2 C1, and deliberate attempts to rearrange this compound have
given negative results (4).

With two exceptions (IV and VII) all of the above described chloro
compounds contain at least one pair of conjugated double bonds. Never-
theless, none of them reacts either with naphthoquinone or with maleic
anhydride at 100. This failure cannot be ascribed to the fact that in
several cases the conjugated system terminates in a series of contiguous
double bonds, since, as has already been pointed out above, this arrange-
ment does not preclude 1,4 addition; moreover, this arrangement is
not present in the tetrachloride III which also fails to react. Similar
failures have been reported before. Thus, X, XI and XII all react very

RCH=C CH=CH CH..=C C=--CHo CH*===C---C===CH CH 3

I I I " " I I

Cl Cl CHa Cl CH 3

(R = H, CH 8 . C 2 H 6 , C 4 H 9 , or C 7 H, 6 ) (XI) (XII)

(X)

smoothly with naphthoquinone (5) but XIII and XIV do not (6). It
C1CH=-CH CH-=CH, CH 2 =C C=CH 2

(XIII) (XIV)

is evident that the diene reaction can be regarded as diagnostic only if it
leads to positive results; comparatively trivial substitutions sometimes
cause it to fail.

Substitution similarly has a profound influence on the tendency of
dienes to polymerize. Of the above described chloro compounds only
the dichlorotetraene (IX) shows any great tendency to polymerize spon-
taneously. In the course of twenty-four hours at the ordinary conditions
it was converted into a dark brown, hard, brittle resin. The dichloride
II polymerized very slowly oti standing, and after three months it was
a viscous sirup. The polymerization was greatly accelerated by increased
pressure. At 6000 atmospheres and 25 it was changed in twenty- three
hours to a yellow, elastic, plastic mass. The chloro-3-hexatetraene-
1,3,4,5 at 6000 atmospheres and 25 was changed in twenty hours to a
hard coke-like mass. Pressure was not applied to the other compounds,
but none of them showed any signs of polymerizing when allowed to stand
for several weeks. We are indebted to Dr. H. W. Starkweather for the
experiments at high pressure.

XIII. ACTION OF CHLORINE ON DIVINYLACETYLKNK 353

Experimental Part
Preparation and Proof of Structure of Dichloro-l,4-hexatriene-2,3,5 (II)

Preparation of Dichloro-l,4-hexatriene-2,3,5 by Chlorination of Divinylacetylene.

Chlorine was passed into a solution of 400 g. (5 moles) of divinylacetylene in 300 g. of
dry carbon tetraehloride at 50. During five hours, 436 g. (6.1 moles) of chlorine
was absorbed. Fractionation of the reaction product gave 191 g. (1.28 moles) of di-
chloro-l,4-hexatriene-2,3,5 boiling at 45-50 (3 mm.) (25% of the theoretical amount).
Continued distillation of the residue yielded 57 g. of higher boiling material (b. p. 50-
115 (3 mm.)) which was a mixture of the tetrachloro and hexachloro derivatives. The
remainder consisted of undistillable material.

A chlorination similar to that described above was carried out by allowing only 332
g. (4.67 moles) of chlorine to be absorbed during five hours at 50. There was ob-
tained 145 g. (0.97 mole) of the dichloro derivative (19% of the theoretical amount)
and 61 g. of tctrachloro-l,3,4,6-hexadieiie-2,4. Chlorination of divinylacetylene at
(mole for mole) in carbon tetraehloride gave the dichloride in 20% yield.

The properties of dichloro-l,4-hexatriene-2,3,5 are: colorless mobile liquid with a
sharp characteristic odor; b. p. 38 at 1 mm., 45-46 at 3 mm.; d 1.1807; 2 D 1.5195;
MR calcd. 38.24; MR found, 38.34. On standing it very slowly polymerized, and
after three months it had changed to a viscous sirup.

The same compound was obtained in very small yield by the action of aqueous
hypochlorous acid on divinylacetylene.

Anal. Calcd. for C 6 H 6 C1 2 : C, 48.35; H, 4.03; Cl, 47.62; mol. wt., 149. Found:
C, 47.88; H, 4.00; Cl, 47.59; mol. wt. 155.

Reduction of Dichloro-l,4-hexatriene-2,3,5- to w-Hexane. The dichloro compound
(25 g., 0.168 mole) in 50 cc. of ethyl acetate (with 0.16 g. PtO 2 ) absorbed 0.484 mole of
hydrogen during three hours (calcd. 0.672 mole). A large amount of hydrogen chlo-
ride was formed during the reduction. The filtered ethyl acetate solution was frac-
tionated, and the fraction boiling at 50-75 was collected and allowed to stand over
aqueous alkali until saponification of the ester was complete. The water insoluble
layer was then separated and further dried over solid alkali. There was obtained
8 g. of w-hexane, b. p. 69-70; d 0.677; w 2 D 1.3812. There was also obtained 15 g. of
higher boiling material (b. p. range 30-80 at 21 mm.) from which no constant boiling
fraction was received.

The Reactivity of the Halogens in Dichloro-l,4-hexatriene-2,3,5. Hydrolysis by
Sodium Carbonate. The dichloro compound (50 g., 0.333 mole) was refluxed with
agitation during eight hours with 300 cc. of water containing 72 g. (0.67 mole) of sodium
carbonate. Analysis of the aqueous solution showed that 0.281 mole of sodium chloride
was formed. Hence only one active halogen atom is present in C 6 H 6 C1 2 . The hydroly-
sis product was a soft, sticky resin.

Permanganate Oxidation of Dichloro-l,4-hexatriene-2,3,5 to Chloroacetic Acid.
The dichloro compound 50 g. (0.33 mole) completely reduced 300 g. of potassium per-
manganate added with agitation in small portions to the aqueous solution (200 cc.) dur-
ing three hours at 35-10. Sulfur dioxide was passed into the reaction mixture; the
manganese dioxide was filtered, and repeatedly washed. Continuous ether extraction
of the strongly acidified aqueous solution gave 15 g. of acidic material, which was
identified as chloroacetic acid, b. p. 61 to 63 at 2 mm. ; m. p. 58-59.

354 ACETYLENE POLYMERS AND THEIR DERIVATIVES

To complete the identification the chloroacetic acid was converted, through its
chloride, into phenylglycocoll anilide, which showed the correct melting point and
mixed melting point.

Chlorination of Dichloro-l,4-hexatriene-2,3,5 to Tetrachloro-l,3,4,6-hexadiene-
2,4. Chlorine was passed into 100 g. (0.67 mole) of C 6 H 6 Cl2 in 54 g. of carbon tetra-
chloride at 5 to 10 until 42 g. (1.18 moles) of chlorine was absorbed. Fractionation
of the product gave 10 g. of tetrachloro-l,3,4,6-hexadiene-2,4; b. p. 85 to 92 at '3 rnm.;
d\ Q 1.4902; n 2 1.5458. Part of the original dichloro compound (21 g.) was recovered.
The remainder of the product consisted of higher boiling residue. The tetrachloro de-
rivative is further described below.

Preparation and Proof of Structure of Chloro-3-hexatetraene- 1,3,4,5 (VI)

Chloro-3-hexatetraene-l, 3,4,5 Obtained by the Action of Alcoholic Alkali on Di-
chloro-l,4-hexatriene-2,3,5. The dichloro compound (105 g., 0.67 mole) freshly dis-
tilled was added during one and one-half hours to a solution of 43 g. of potassium hy-
droxide (15% excess) in 200 cc. of absolute methanol with vigorous agitation at 10-15.
Stirring was continued at 15 during two hours. The insoluble potassium chloride was
filtered off, washed with methanol, and dried. The yield of potassium chloride was
47 g. (94% of the theoretical amount). The methanol filtrate was poured into 500 cc.
of water, and the water-insoluble material separated and dried. The yield of crude
material was 74 g.

The pure chloro-3-hexatetraene-l, 3,4,5 was obtained as a colorless liquid; b. p.
127 at 760 mm. (with decomposition), 82 at 163 mm., 55 at 54 mm.; dj 0.9997;
n 1.5280; MR calccl., 32.91 ; MR found, 34.64.

The same compound was obtained by the action of sodium methylate (in absolute
methanol) on dichloro-l,4-hexatriene-2,3,5.

Anal. Calcd. forC 6 H 6 Cl: C, 64.02; H, 4.45; Cl, 31.53; mol. wt., 112.5. Found:
C, 63.99; H, 4.61; Cl, 31.61; mol. wt. (cryoscopic in benzene), 110.

Reduction of Chloro-3-hexatetraene-l, 3,4, 5 to n-Hexane. The chloro compound
(13 g.) absorbed 87% (0.403 rnole) of the theoretical amount of hydrogen (catalyst 0.15
g. of PtCW in 25 cc. of ethyl acetate during three hours. A large amount of hydrogen
chloride was formed. The ethyl acetate solution was distilled and the distillate up to
80 was collected and allowed to stand over aqueous alkali during four hours. The un-
saponifiable material was separated and allowed to stand over solid potassium hydroxide
during fifteen hours. There was obtained 2 g. of w-hexane, b. p. 64-70; d? 0.691;
w 2 ,? 1.3766. A small amount of high boiling residue was obtained but not identified.

Permanganate Oxidation of Chloro-3-hexatetraene-l, 3 ,4,5 to Oxalic Acid.
The oxidation was made by adding in small portions 224 g. of potassium permanganate
to a vigorously agitated, aqueous solution (200 cc.)of 80 g. of potassium carbonate and
21 g. of chloro-3-hexatetraene-l, 3,4, 5 at 35-45. After ten hours the solution was de-
colorized with sulfur dioxide and the manganese dioxide filtered and repeatedly washed.
The combined filtrates were strongly acidified with sulfuric acid, while the reaction
mixture was cooled in an ice-bath. Continuous ether extraction of the acidified aque-
ous solution during fifteen hours gave 15 g. of oxalic acid dihydrate (calcd. 25 g.). The
acid melted at 101 (copper block).

Neutral equivalent. Subs., 0.2916 g. Required 41.10 cc. of 0.1117 N NaOH
Calcd. for H 2 C 2 O4-2H2O; neutral equivalent, 63.0. Found: 63.5.

XIIT. ACTION OF CHLORINE ON DIVINYLACETYLENE 355

Preparation and Proof of Structure of Trichloro- 1, 4, 4-hexadiene-2, 5 (VII)

Addition of Hydrogen Chloride to Dichloro-l,4-hexatriene-2,3,5. Dichloro-1,4-
hexatriene-2,3,5 (75 g., 0.5 mole) was shaken during twelve hours at 27 with 30 g. of
cuprous chloride and 75 cc. of hydrochloric acid (sp. gr. 1.18). The water insoluble
layer was separated, washed and dried. Fractionation gave 20 g. of the original di-
chloro compound and 28 g. of trichloro-l,4,4-hexadiene-2,5, which boiled at 100-103
at 4 mm.; d% 1.3036; w 2 D 1.5585; M n calcd. 43.58; M n found, 45.88.

Anal. Calcd. for C 6 H 7 C1 3 : C, 38.83; H, 3.78; Cl, 57.39; mol. wt., 185.5. Found:
C, 38.87; H, 3.83; Cl, 57.38; mol. wt. (cryoscopic in benzene), 185.

Permanganate Oxidation of Trichloro-l,4,4-hexadiene-2,5 to Chloroacetic Acid.- -
The trichloro compound (18 g.) reduced 84 g. of potassium permanganate (added in
small portions with vigorous agitation) in alkaline solution during four hours at 30-40.
The nitrates obtained after separating and washing the manganese dioxide were acidified
and subjected to continuous ether extraction during fifteen hours. From the ether ex-
tract there was obtained 10 g. of chloroacetic acid, which boiled at 65-68 at 4 mm. and
melted at 58-59. It was further identified by conversion through its acid chloride into
chloroacetanilide, which showed the correct melting point, 134-135.

The Reactivity of the Halogens in Trichloro-l,4,4-hexadiene-2,5. Hydrolysis by
Sodium Carbonate. -The trichloro compound (5 81 g , 0313 mole) was refluxed during
seven hours with agitation in 110 cc. of water containing 8.6 g. of sodium carbonate.
Analysis of the aqueous solution showed that 0.0763 mole of sodium chloride was formed.
Therefore three active halogen atoms were present.

Preparation and Proof of Structure of Tetrachloro-l,3,4,6-hexadiene-2,4 (III)

Chlorination of Divinylacetylene to Obtain Tetrachloro-l,3,4,6-hexadiene-2,4.
Chlorine was passed into divinylacetylene (234 g., 3 moles) at 40 to 50 with vigor-
ous agitation. During five and one-half hours 340 g. (4.8 moles) of chlorine was ab-
sorbed, corresponding to 80% chlorination to the tetrachloro derivative. Considerable
loss of hydrogen chloride occurred. Distillation gave 119 g. of crude dichloro-1,4-
hexatriene-2,3,5 and 96 g. of very crude tetrachloro derivative. Fractionation of 160
g. of the crude tetrachloro derivative (b. p. 50-120 at 3 mm.) gave 50 g. of tctrachloro-
1 ,3,4,6-hexadiene-2,4 boiling at 84-89 at 2 mm.; f/f 1.4013, 2 1.5465; Jl/ K calcd ,
48.43; 7lf R found, 49.71.

Anal. Calcd. for CeHeCli: C, 32.76; H, 2.73; Cl, 04.51 ; mol. wt., 220. Found:
C, 32.80; H, 2.49; Cl, 63.76, mol. wt. (cryoscopic in benzene), 222.

Permanaganate Oxidation of Tetrachloro-l,3,4,6-hexadiene-2,4 to Chloroacetic
Acid. The tetrachloroderivativc (58 g.) completely reduced 220 g. of potassium per-
manganate (added in small portions with vigorous agitation) during five hours at 35-40.
The filtrate obtained after separating and washing the manganese dioxide was strongly
acidified and subjected to continuous ether extraction during fifteen hours. The ether
extract gave 20 g.of chloroacetic acid which boiled at 64-66 at 3 mm. and melted at
58-59 after recrystallization from petroleum ether.

The Reactivity of the Halogens in Tetrachloro-l,3,4,6-hexadiene-2,4. Hydrolysis
by Sodium Carbonate.- The tetrachloro compound (5 g., 0.0228 mole) was refluxed dur-
ing eight hours with agitation in 100 cc. of water containing 7 g. (0.066 mole) of sodium
carbonate. Analysis of the aqueous solution showed that 0.0414 mole of sodium chloride
was formed. Therefore two active halogen atoms are present in the tetrachloridc.

356 ACETYLENE POLYMERS AND THEIR DERIVATIVEvS

The Chlorination of Tetrachloro-l,3,4,6-hexadiene-2,4 to Hexachloro-1,2,3,4,5,6-
hexene-3 (IV). Chlorine was passed into 27 g. (0.12 rnole) of tetrachloro- 1,3,4,6-
hexadiene-2,4, during eight hours at 60-70. The product was distilled and from the
portion boiling at 95-115 at 2 mm. was obtained 8 g. of pure hexachloro-1,2,3,4,5,6-
hexene-3 which melted at 57 to 58 after two crystallizations from petroleum ether;
yield, 23%. The remainder of the product was a viscous liquid not further investigated.
The hexachloride is described in more detail below.
The Behavior of Tetrachloro-l,3,4,6-hexadiene-2,4 with Alcoholic Alkali. Trichloro-

l,3,4-hexatriene-2,4,5 (VIII) and Dichloro-3,4-hexatetraene-l,2,4,5 (IX)

The tetrachloro compound (75 g., 0.34 mole) was added during two hours to a solu-
tion of 44 g. of potassium hydroxide (15% excess) in 200 cc. of absolute methanol with
vigorous agitation at 10-15. The reaction was allowed to proceed for one hour after
which the potassium chloride was filtered off (KC1, 36 g. or 48%). The methanol
filtrate was poured into water, and the water-insoluble layer was separated, dried and
distilled. From the crude material (50 g.) two fractions were obtained; one was tri-
chloro-l,3,4-hexatriene-2,4,5 (25 g.) which boiled at 50 at 1 mm.; d% 1.3132; 2 ,?
1.5517; Mn calcd. 43.11; MR. found, 44.61. The other was dichloro-3,4-hexatetraene-
1,2,4,5 (5 g.) which boiled at 38-40 at 8 mm.; 4 1.1819; w 2 D 1.5456; Mn calcd. 37.77;
Mn found, 39.41.

Trichloro-l,3,4-hexatriene-2,4,5. -Anal. Calcd for C 6 H 6 C1 3 : C, 3926; H, 2.73;
Cl, 58.01; mol. wt., 183.4. Found: C, 39.70; H,2.75; Cl, 57.52; mol. wt. (cryoscopic
in benzene), 190.

Dichloro-3,4-hexatetraene-l,2,4,5.~ vlwa/. Calcd. for C 6 H 4 C1>: Cl, 48.27
Found. Cl, 48.60, 48.32 (7).

The Chlorination of Divinylacetylene at 60-70 to Obtain Hexachloro-1, 2,3,4,5,6-
hexene-3 (IV). Divinylacetylene (100 g., 1.28 moles) in 160 g. of carbon tetrachloridc
absorbed 197 g. (2.8 moles) of chlorine during twelve hours. Distillation of the product
gave 30 g. (10% of the theoretical amount) of the hexachloride, which boiled at 110 to
112 at 2 mm. After recrystallization from petroleum ether the white crystals melted
at 58 to 59. The major part (150 g.) of the product from this reaction was an almost
colorless, viscous sirup. This is described in more detail below.

Hexachloro-l,2,3,4,5,6-hexene-3 is not attacked by ozone or by hot nitric acid, and
the action of alkaline permanganate leads to total destruction. It fails to chlorinate
further even at elevated temperatures in the presence of light.

Anal. Calcd. for C 6 H 6 C1 6 : C, 24.76; H, 2.06; Cl, 73.16. Found: C, 25.14; H,
2.18; Cl, 73.37.

The viscous sirup that formed the major part of the product of the above-described
reaction was obtained as a residue which remained after the removal of the hexachloro-
l,2,3,4,5,6-hexene-3 by distillation. This material could not be distilled even at 185 at
a pressure of 2 mm. although it darkened rapidly and showed a tendency to liberate
hydrogen chloride under these conditions. Its chlorine content was generally 1-3%
less than that of the pure hexachloride, and like the latter it showed no tendency to ab-
sorb more chlorine even at elevated temperature. This sirup probably results from the
coupling together of two or more of the six-carbon atom chains at some stage of the
Chlorination.

Molecular weight determinations were made on two specimens of the sirup both of
which contained a considerable proportion of the dissolved hexachloride. The values

XIV. THE DIHYDROCHLORIDE OF DIVINYLACETYLENE 357

obtained were (a) Cl, 68.90%; mol. wt. (cryoscopic in benzene), 360; (b) Cl, 72.90%;
mol. wt., 370.

Summary

Divinylacetylene reacts with chlorine to form a liquid dichloride and
tetrachloride, a crystalline hexachloricle, and a sirupy product having
approximately the composition of the hexachloride but a higher molecular
weight. The three monomeric products are each formed from their pre-
cursors by 1,4 addition and they have the formulas dichloro-l,4-hexa-
triene-2,3,5; tetrachloro-l,3,4,()-hexadiene-2,4; and hexachloro-1, 2,3,4, -
r),G-hexene-3. The following transformation products of these compounds
are described: chloro-3-hexatetraene-l, 3,4,5; trichloro-l,4,4-hexadiene-
2,5; trichloro-l,3,4-hexatriene-2,4,5; and dichloro-3,4-hexatetraene-l,2,4,5.

Bibliography and Remarks

(1) Nieuwland, Calcolt, Downing and Carter, Journ Am Chrm Soc., 53, 4197 (1931)

(2) Carothers, ibid., 55, 2008 (1933); this volume, page 344.

(3) Carothers, Berchet, and Collins, Journ Am Chem Soc , 54, 4000 (1932); this volume, page 300

(4) Unpublished results.

(f>) Carothers and Coffman, Journ Am Chem Soc , 54, 4071 (1932), this volume, page 384,
Jacobson and Carothers, Jouni Am Chem Sor , 55, 1024 (1933); this volume, page 331

(0) Carothers and Berchet, Journ Am Chem .Sor , 55, 2004 (1933); this volume, page 340.

(7) The analyst found great difficulty in the proper combustion of this compound The carbon
percentage was usually lower and the hydrogen percentage higher than the theoretical amounts.

XIV. The Dihydrochloride of Divinylacetylene*

The addition of thio-p-cresol (I) to divinylacetylene (2) involves only the
terminal ethylenic linkage. Chlorine on the other hand reacts by 1,4
addition to yield first a dichloride and then successively a tetra- and a
hexachloride (3). Study of the action of hydrogen chloride on divinyl-
acetylene has furnished further information concerning the behavior of
this multiply conjugated enyne system.

Divinylacetylene reacts fairly rapidly when shaken with aqueous hydro-
chloric acid at room temperature. As in the case of vinylacetylene (4),
the reaction is considerably accelerated by cuprous chloride. On the
other hand, the same product is obtained whether cuprous chloride is
present or not. Under ordinary conditions a point of apparent saturation
is reached when two molecules of hydrogen chloride have been absorbed.

* D. D. Coffman, J. A. Nieuwland and W. H. Carothers, Journ. Am. Chem. Soc.,
55, 2048-51 (1933); Contribution No. 117 from the Experimental Station of E. I dti
Pont dc Nemours and Co.

Received October 22, 1932. Published May 6, 1933

358 ACETYLENE POLYMERvS AND THEIR DERIVATIVES

The first step in the reaction undoubtedly involves the formation of a
monohydrochloride, but apparently this reacts more readily than the parent
hydrocarbon. Consequently the chief product is dihydrochloride even
when a deficiency of hydrochloric acid is used. The monohydrochloride
is produced in poor yields, and it is so difficult to separate from the un-
changed hydrocarbon that it has not yet been obtained in a state of purity.

Purification of the dihydrochloride to constant composition requires
rather sharp fractionation, and the boiling range (80 to 82 at 17 mm.)
is wide enough to suggest the presence of geometrical isomers. Only one
chlorine atom of the dihydrochloride is reactive. When heated with
naphthoquinone or maleic anhydride the only evidence of reaction is the
formation of a small amount of dark gummy material. However, as has
already been pointed out (3), negative results in the Diels- Alder reaction
are of no value in the demonstration of structure, since many 1,3-dienes
fail to react. The compound is stable. It shows no tendency to poly-
merize spontaneously, but a sample that had been submitted to a pressure
of 6000 atmospheres for ninety- three hours at 49 was changed to a dark,
somewhat elastic and only slightly plastic mass. When heated with
sodium acetate in acetic acid it yields a monoacetate, and the second
chlorine atom remains unaffected. Similarly the action of methyl alcoholic
potash results in a monomethyl ether. When oxidized with permanganate
the dihydrochloride yields acetic acid and chloro acetic acid.

These facts furnish a basis for a discussion of its structure. We assume
that the hydrogen atoms of the hydrocarbon retain their original positions
during reaction, and the results of oxidation then indicate that the di-
hydrochloride contains the residues CH 3 CH C and C1CH 2 CH 2 C .

Hence the chain has the formula CH 8 CH=C C=CH CH 2 C1. Only

I I
a hydrogen atom and a chlorine atom remain to be accounted for, and

one of these must be attached to the third carbon atom and the other
to the fourth carbon atom. No direct experimental method is available
for a decision between these two alternatives, but a satisfactory con-
clusion can be reached by a consideration of the probable mechanism
of the reaction. In the following chart the two alternative formulas,
dichloro-l,3-hexadiene-2,4 and dichloro-l,4-hexadiene-2,4 are designated
as A and B, respectively; the positions of the hydrocarbon chain are
marked with the numbers 1 to 6; and the H and Cl of the addendum are
called a and b, respectively. Addition at the first stage may be 1,2,3,4,
1,4 or 1,6 and it must bring the chlorine atom to position 1 or position
3 (or 4). The formulas I to IV therefore include all the reasonable possi-

XIV. THE DIHYDROCHLORIDE OF DIVINYLACETYLENE 35<)

bilities for the monohydrochloride. For further reaction the possibilities

are

ADDITION OF HYDROGEN CHLORIDE TO DIVINYLACETYLICNE

Mode of Monohydrochloride Mode of Dihydrochloride

No. addition produced addition produced

I la 4b CHi CH=C-CC1 CHCHi 3a, 6b CHt CH=CH CC1=CH CHjCl A

II 4a' Ib CHjCl CH-C- CH CHCHt 6a, 3b ClCHr-CH=CCl CH=CH CH, A

III 3a 4b CH,CH CHCCl CHCHt la, 6b CHr-CH-CH CCl-=CH CHiCl A

6a, Ib CICHr CH=CH CC1=CH CHs B

IV la 6b CHr CH=-CCCH CH 2 C1 3a, 4b CH CH=CH CC1CH CHjCl A

4a, 3b CHj CH=CC1 CH=CH CH:C1 B

I by 1,4 addition (3a, 6b) gives A III by 1,6 addition (6a, Ib) gives B

II by 1,4 addition (6a, 3b) gives A IV by 1,2 addition (3a, 4b) gives A

III by 1,6 addition (la, 6b) gives A IV by 1,2 addition (4a, 3b) gives B

Thus of the six pairs of reactions that might lead to A or B, four lead to
A and only two to B, and probability is two to one in favor of A as the
formula of the dihydrochloride. Further than this, however, there are
good reasons for rejecting both III and IV as possibilities. Evidence of
1,6 addition in the reactions of divinylacetylene has never been observed,
nor of 1,2 addition at the acetylenic linkage of either vinylacetylene or
divinylacetylene. Moreover, 1,2 addition followed by 1,6 addition and
1,6 addition followed by 1,2 addition both appear inherently improbable.
The addition of chlorine to divinylacetylene proceeds 1,4: CH2=CH
C C~CH CH 2 > CH 2 C1 CH=C=CC1 CH=CH 2 (3); and the
primary step in the action of hydrogen chloride on vinylacetylene is 1,4
addition: CH C CH=CH 2 -^ CH 2 =C=CHCH 2 C1 (4). It seems
certain therefore that the dihydrochloride of divinylacetylene is dichloro-
l,3-hexadiene-2,4 (A) and the mechanism by which it is formed is most
probably represented by II.

Experimental Part

The Addition of Hydrogen Chloride to Divinylacetylene in the Presence of Cuprous
Chloride.- Divinylacetylene (80 g ) was shaken al cither or room temperature during
one hour with 1.5 liters of 12 N hydrochloric acid containing 300 g. of cuprous chloride.
The reaction mixture was then subjected to steam distillation. By this process 80 g.
of crude dihydrochloride was obtained (53% of the theoretical amount). Fractionation
of 241 g. of material (b. p. 70-80 at 11 mm.) through an efficient column gave 170 g.
of analytically pure dichloro-l,3-hexadiene-2,4 which boiled at 8082 at 17 mm.;
t/f 1.1456; w 2 D 1.5271; M R calcd. 38.70; MR found, 40.42.

Anal. (Carius). vSubs., 0.1201 g.; AgCl, 0.2305 g. Calcd. for C 6 H,C1 2 : Cl, 46.99.
Found: Cl, 47.47.

The Addition of Hydrogen Chloride to Divinylacetylene in the Presence of Calcium
Chloride. Six pressure bottles each containing 150 cc. of hydrochloric acid (sp. gr.
1.187), 30 g. of anhydrous calcium chloride and 80 g. of divinylacetylene were shaken at

3(50 ACETYLENE POLYMERS AND THEIR DERIVATIVES

25 during one hundred and ten hours. The water-insoluble layer was separated,
washed with aqueous sodium carbonate and dried. Distillation gave 242 g. of material
boiling above 50 at 70 mm.

Fractionation through an efficient column gave 94 g. of material boiling at 50-60
at 52 mm., and 60 g. of dichloro-l,3-hexadiene-2,4 which boiled at 80-82 at 17 mm.
The residue weighed 00 g. The yield of dihydrochloride was 6.8% of the theoretical
amount.

Repeated fractionation of the low boiling material (b. p. 50-60 at 52 mm.) failed
to give a product having the composition of the monohydrochloride of divinylacetylene
(C 6 H 7 C1). The chlorine analysis of the samples collected was always less than the
theoretical amount for the monohydrochloride, indicating that hydrocarbon material
was present.

The Permanganate Oxidation of Dichloro-l,3-hexadiene-2,4 to Obtain Acetic and
Chloroacetic Acids. An analytically pure sample of the dihydrochloride (100 g.) was
introduced into an aqueous solution (400 cc.) containing 46 g. of potassium carbonate.
Potassium permanganate (486 g.) was added in small amounts during five hours with
continuous agitation. The solution was decolorized with sulfur dioxide and the man-
ganese dioxide filtered and washed. The combined filtrates were acidified with 100 cc.
of concentrated sulfuric acid and the solution subjected to continuous ether extraction
during fifteen hours. The ethereal extract was dried and the ether removed by evapora-
tion. The acidic residue gave by fractionation 15 g. of acid boiling at 105-120 (Fract.
C) and 10 g. of acid boiling at 62.5 at 3 mm. (Fract. A).

The determination of the Duclaux values of Fract. C indicated that acetic acid
was the material under examination. The Duclaux values observed are 7.3, 6.7, 7.4,
compared to 6.8, 7.1, and 7.4 for acetic acid, ,3.95, 4.40, 4.55, for formic acid, and 11.9,
11.7, 11.3, for propionic acid. The identification of Fract. C was completed by the
preparation of />-acet-toluididc, which showed the correct melting point and mixed melt-
ing point (149 to 150). Fraction A was similarly identified as chloroacetic acid by
conversion to chloroacetanilide, which showed the correct melting point and mixed
melting point (134 to 135).

The Action of Alcoholic Potash on Dichloro-l,3-hexadiene-2,4. Preparation of
Methoxy-l-chloro-3-hexadiene-2,4. Dichloro-l,3-hexadiene-2,4 (65 g.) in 200 cc. of
absolute methanol containing 38 g. of potassium hydroxide was rcfluxed with vigorous
agitation. Filtration of the reaction mixture gave 30 g. of dry potassium chloride
(calcd. 34 g.). Aqueous dilution of the alcohol solution was followed by separation of
the water-insoluble layer which was dried and fractionated. There was obtained 20
g. of low boiling material and 21 g. of the methyl ether which boiled at 8892 at 30
mm.; w 2 ,? 1.4928; d\ Q 1.0239; Jl/ u calcd. 40.10; Mn found, 41 .57.

Ami. Calcd. for C 7 H,,C1O: C, 57.35; H, 7.51. Found: C, 50.91; H, 774.

Action of Sodium Acetate on Dichloro-l,3-hexadiene-2,4. Preparation of Acetoxy-
l-chloro-3-hexadiene-2,4(5). One hundred grams of the dichloro compound, 100 g. of
fused sodium acetate, and 300 cc. of glacial acetic acid were refluxed together for four
hours. The acetic acid was neutralized with 20% sodium hydroxide and the mixture
was extracted with ether. The ether extract was dried with calcium chloride, and then
distilled. Three fractional ions gave 55 g. of colorless liquid boiling at 84 to 85 at 3
mm.; df 1.0915; n 2 ,? 1.4890; MR calcd., 44.61 ; M H found, 46.14.

Anal. Calcd. for CsHnO.Cl: Cl, 20.31. Found: 20.58,20.78.

XV. HALOGEN-4-BUTADIENEvS-l,2 361

Summary

The action of aqueous hydrochloric acid on divinylacetylene yields a
dihydrochloride. The compound is oxidized by permanganate with the
formation of acetic acid and chloroacetic acid, and this fact together with
a consideration of the probable mechanism of its formation, indicates that
it has the structure dichloro-l,3-hexadiene-2,4 and that it results from two
successive acts of 1,4 addition. When treated with sodium acetate its re-
active (terminal) chlorine is replaced by acetoxy, and the action of methyl
alcoholic potash similarly leads to the formation of a methyl ether.

Bibliography and Remarks

(1) Carothers. Journ Am Chem. Soc , 55, 2008 (1933); this volume, page 344

(2) Nieuwland, Calcott, Downing and Carter, Journ. Am. Chem Soc , 53, 4197 (1931).

(3) Coffman and Carothers, Journ Am Chem Soc , 55, 2040 (1933); this volume, page 348.

(4) Carothers, Berchet and Collins, Journ Am Chem Soc , 54, 4000 (1932) , this volume, page 300

(5) We are indebted to Dr. W F Talhot for this experiment.

XV. Halogen-4-butadienes-l,2. The Mechanism of
1 ,4-Addition and of ^-Rearrangement*

The reactions of vinylacetylene (I) and some of its derivatives exhibit
peculiar features which provide a unique opportunity for testing certain
assumptions concerning the mechanism of 1 ,4-addition and of a,7-rc-
arrangement. We have already shown (1) that the primary product of
adding hydrogen chloride to (I) is chloro-i-butacliene-1,2 (II) which is readily
rearranged to chloroprene (III) (2).

CH^C CH=CH, CH,=C CHCH,X CH, CX CH=CH,

(I) (II) (III)

We now present further analogous facts as a basis for discussion.

The bromide II is obtained by the action of liquid hydrogen bromide on

I at low temperature and by the action of sodium bromide on the chloride

II at room temperature. The second method has also been applied to the
preparation of the iodide II, and in both cases the product is free from the
isomeric halide III.

The chloride and the bromide II react readily with hot water (preferably
in the presence of sodium carbonate) and yield the corresponding carbinol,
a strongly lachrymatory liquid, which can then be converted back to the

* W. H. Carothers and G. J. Berchet, Journ. Am. Chem. Soc., 55, 2807-13 (1933) ;
Contribution No. 120 from the Experimental Station of E. I. du Pont de Nemours and
Co.

Received December 20, 1932. Published July (, 1933.

362 ACETYLEN^ POLYMERS AND THEIR DERIVATIVES

chloride or the bromide II by the action of phosphorus trihalide in the pres-
ence of pyridine. The carbinol II is also converted to the chloride II by
boiling dilute hydrogen chloride (18%). The action of sodium acetate on
the chloride or bromide in boiling acetic acid solution yields the acetate II,
The latter is also obtained from the carbinol by the action of acetic an-
hydride plus sulfuric acid. These reactions all proceed normally, i. e.,
none of the isomeric compounds III has been detected in the products.
In this respect they differ from those involving Grignard reagents de-
scribed in the next paper (3). Phenylmagnesium bromide acts on the
chloride II yielding a mixture of phenyl-4-butadiene-l,2 (normal) and
phenyl-2-butadiene-l,3 (abnormal). Methyl- and heptylmagnesium
halides gave only abnormal products, while benzylmagensium chloride gave
only the normal product. The isomeric chloride III does not react with
Grignard reagents.

The bromide II, like the chloride (4), is rapidly converted to the isomeric
compound III by dilute HX containing CuX. The iodide isomerizes still
more readily: when heated alone to about 130 it suddenly evolves heat
and the lower-boiling iodide III rapidly distils. It may be inferred that
this transformation also occurs spontaneously at the ordinary temperature
since, within one or two days, the iodide II is completely converted to a
granular polymer almost certainly derived from III. Deliberate attempts
to rearrange the acetate and the carbinol II have been unsuccessful.

In r6sum6 the facts are :

Normal Metathetical Reactions of II : RC1 > RBr (I); RCl(Br) >ROH(OAc);

ROH > RCl(Br, OAc).

Rearrangements: RCl(Br, I) II > III.

Abnormal Reactions: RC1 (II) with Grignard reagents gives both normal and ab-
normal products.

The halides of series III polymerize rapidly: the speeds for chloroprenc
and bromoprene (5) are roughly 700 and 1100 times that of isoprene, and
iodoprene polymerizes so quickly as to interfere with the determination of
its physical properties. None of the compounds of series II shows any
tendency to polymerize spontaneously. (The apparent polymerization of
the iodide II is accounted for by prior rearrangement.)

Theories (6) concerning the mechanism of 1,4- addition and of a,7-rear-
rangements are too numerous to review but, with some concession for
differences in terminology, the following propositions (1-4) will be recog-
nized as pertinent to various suggestions that have from time to time been
published.

XV. HALOGEN-4-BUTADIENES-l,2 363

1. That all apparent examples of 1 ,4-addition really involve (com-
pleted) 1,2-addition followed by a,7-shift. This once attractive hypothesis
appears to have been generally abandoned; the facts now show that it
cannot possibly be true. The rearrangement of the halides II > III is
irreversible, hence II, the 1,4-product, must be a primary product.

2. That addition at conjugated systems is a two stage process: the
cation of the addendum first attacks one end of the system yielding isomeric
ions (or intermediate fragments) corresponding to the two isomeric prod-
ucts. The ions arising from attack of (I) by hydrogen ion would be CH 2 =
C+ CH=CH 2 (III) and CH 2 C=CHCH 2 + (II). Thus in the addition
of HX to (I), the hydrogen ion attacks the acetylenic (positive?) end of the
molecule; the resulting ions are those involved in the isomerizations dis-
cussed under 4 below.

3. That the probable nature of the predominating hypothetical inter-
mediate or of the predominating product can be inferred from conclusions
based on fact or on speculation as to which is the more stable.

The facts show that this is not possible, since in the addition of hydrogen
chloride or hydrogen bromide to (I) the unstable isomer (II) is a primary
product, and is in all probability the only primary product.

4. (a) That in triad systems it is the ion R+ not the molecule RX which
rearranges, (b) That R + is a necessary intermediate in metathetical
reactions of RX involving the formation of X". (c) That the ion R + ,
whether formed directly by dissociation of RX or as an intermediate in
metathetical reactions of RX, will immediately assume the forms corre-
sponding to the equilibrium products.

The practical implication of (c) is that metathetical reactions of triad
systems (involving X) must yield an equilibrium mixture of the isomers.
This has been taken so much for granted that it forms the basis of a sup-
posed demonstration (7) of the intimate mechanism in the hydrolysis of car-
boxylic esters. It is now obviously directly opposed to the facts, since a
whole series of metathetical transformations in compounds of type II has
been carried out without any rearrangement to the corresponding isomers
III. Not only does no isomerization occur under these conditions where,
according to (b), ionization must be assumed to occur, but the iodide II
rearranges under conditions that appear to be especially unfavorable to
ionization. This throws considerable doubt on assumption (4a).

Abnormal reactions of triad systems do, nevertheless, exist. The action
of RMgX on the chloride II furnishes an unambiguous demonstration of
this fact. The isomeric chloride III is (experimentally) incapable of
reacting with RMgX; hence completed rearrangement cannot precede

304 ACKTYLENE POLYMERS AND THEIR DERIVATIVES

reaction. The isomeric hydrocarbons II and III cannot rearrange under
the conditions used. The isomerization is therefore an integral part of the
metathesis. Three explanatory assumptions appear to be possible:

(a) The abnormal reaction consists in addition of RMgX at the double
bond followed by elimination of MgX 2 . This has the disadvantage of being
ad hoc since addition of RMgX at simple ethylenic linkages has never been
realized.

(b) The initiating step results in the formation of a cation (or some other
fragment) which adjusts itself before entrance of the anion of RMgX.
This is the currently popular assumption (4a and b above). It presents
this difficulty as shown in the comment under 4 : the facts now require that
these intermediate fragments (if they exist at all) must be capable (con-
trary to 4c) of resisting or escaping rearrangement under some conditions
and not under others. In either case the conditions cannot be specified in
advance and the assumption of intermediate ions contributes nothing to
the explanation of the facts, but, at best, serves merely as a pictorial method
of rationalizing the facts after they are known.

(c) The initiating step is the formation of a complex (coordination
compound). This assumption also, in the present state of knowledge
concerning complexes, is too vague to furnish a basis for detailed predic-
tions. It is, however, unquestionably closer to reality than the assumption
of free ions or radicals as intermediates. The closest approach to a simple
free alkyl anion is probably found in the alkali alkyls. These materials
are extraordinarily reactive; the ethyl anion, for example, is capable of
reducing the sodium ion to sodium hydride at temperatures below 100
(8). Experimentally, simple alkyl cations are completely unknown; it is
reasonable to suppose that they would be still less stable than alkyl anions.
Metathetical reactions of organic compounds that can be formulated as
proceeding through the intervention of ions generally involve media or
catalysts capable of giving rise to coordination compounds (9).

Johnson (10) has shown that coordination and chelation are probably in-
volved in the mechanism of many types of organic reactions including ab-
normal reactions and o:,7-rearrangemerits. Our application of his ideas to
the present case is illustrated below.

In a triad system the usually accepted valence angles are capable of
placing the unshared electrons of X fairly close to the 7-carbon atom,

R~CH==CH CH 2 -X- Polarization at ft 7 favored by the proximity
of these unshared electrons permits still closer approach of X toward
7 and consequent loosening at a; these tendencies finally bring the

XV. HALOGEN-4-BUTADIENEvS-l,2 365

two forms into equilibrium. This mechanism is in accordance with the
observed order I > Br > Cl ^> OAc, where the effective diameters of the
atoms progressively diminish and the constraints on the electrons in-
crease, while in OAc the Ac group may besides act as a steric obstacle.
This may be supposed to apply to the rearrangement of the halides II
themselves at elevated temperatures. In fact, for the chloride and the
bromide II rearrangement at ordinary temperature appears to require the
presence of salts having great coordinating power conditions merely
favoring ionization, e. g. y water plus hydrogen chloride, do not suffice.
_ CH=CH ^ n ^ e t ner hand, water plus hydrogen chloride plus

x< / \CH cu P r us chloride, or dry ferric chloride alone, induces very
-- X rapid rearrangement. An easily conceivable function of

| the salt in this case is symbolized in the figure, where the

^ etc * steric factors are analogous to those involved in chelation

while the metal salt acts as a carrier of the entering group. In reactions
of metathesis the chloride II yields abnormal products R ,OEt

only with RMgX. Reaction with Grignard reagents ){Mg

generally undoubtedly requires entrance of the reactant
into the complex, RMgX(OEt) 2 . Coupling of a halide
probably involves some intermediate step such as (A). Obviously, in
view of the facts, such a purely "dative" electron shell as that which sur-
rounds the magnesium is unusually vague and mobile. This mobility
would provide the opportunity for a sufficient approach of R and R' in
forms that are essentially cation and anion without ever requiring the inter-
mediacy of free alkyl ions. If R' is a substituted allyl group the steric
factors referred to above will permit the coupling to occur at the 7 as well
as the a carbon. Reaction of the chloride II with sodium iodide to give
the corresponding iodide might appear to require intervention of the
cation II indicated in 2 above, but we reject this idea also, and prefer to
assume that some kind of coordination precedes reaction and that the
cation is never liberated as &free entity. On the other hand, the coordinat-
ing power of the iodide ion is certainly relatively feeble, and there is nothing
in its structure or known behavior to indicate that it could be very effective
in the kind of coordination that is here suggested to account for rear-
rangement or abnormal reaction. The observed absence of abnormality in
other reactions of compounds II can be similarly explained.

Experimental Part

Bromo-4-butadiene-l,2 (II). Viiiylacelylene (otic mole) was passed into 3 moles of
liquid hydrogen bromide during five to seven hours. The product, washed and dis-

366 ACETYLENE POLYMERS AND THEIR DERIVATIVES

tilled, gave two fractions: crude bromo-4-butadiene-l,2, 53.5% of calcd., and dibromo-
2,4-butene-2 (11), 36%. Careful redistillation yielded no other products. Under other
conditions, bromoprene (11) was found in the reaction product. For example, liquid hy-
drogen bromide (2.18 moles) added to vinylacetylene (4 moles) at ca. 50 after three
hours gave the following products : bromoprene, 10%; bromo-4-butadicne-l,2, 19%; di-
bromo-2,4-butene-2, 12%. The bromide II was also obtained from chloro-4-butadiene-
1 ,2 and sodium bromide in acetone plus water. Conversion was far from complete, re-
peated distillation was required to obtain a pure product, and the isolated yield was poor
(25%); but there was no indication of any bromoprene in the product. Better yields
of the bromide II (69%) were obtained from the carbinol II plus phosphorus tribromidc
in the presence of a little pyridine at 10-20 (ca. one hour); a colorless, lachrymatory
liquid; b. p. 64-66 (181 mm.), 109-111 (760 mm.); df 1.4255; w 2 D 1.5248.

Anal. Calcd. for C 4 H 5 Br: C, 36.09; H, 3.76; Br, 60.15. Found: C, 36.61 ; H,
3.91; Br, 60.13.

Iodo-4-butadiene-l,2 (II). Chloro-4-butadicne-l,2 with one mole of sodium
iodide was allowed to stand in 80% alcohol or in acetone at room temperature for three
hours. The heavy yellow oil precipitated by dilution with water was distilled; yield,
46%; b. p. 43-45 (38mm.); estimated b. p. at 760mm., 130; d% 1.7129; n 2 D 1.5709.

Anal. Calcd. for C 4 H 5 I : C, 26.66; H, 2.77; 1,70.55; mol. wt., 180. Found: C,
27.00; H, 2.97; 1,70.40; mol. wt. (in freezing benzene), 184.

It reacted instantly with aqueous silver nitrate. When it was allowed to stand at
ordinary temperature solid particles began to separate within twenty-four hours; within
forty to fifty hours the transformation was usually complete and the product was a
granular mass of amorphous, insoluble particles similar to the granular polymers ob-
tained from chloroprene and bromoprene (5). The polymerization is doubtless preceded
by isomerization.

Anal. Calcd. for (C 4 H B I) n : I, 70.55. Found: I, 69.49, 69.44.

lodoprene (III). When the iodo compound (II) was heated at 125 130 a short
but lively reaction took place; the product then distilled at 111-1 13 (760 mm.), while a
large amount of viscous residue remained in the flask. The change in physical proper-
ties indicated that the distillate was the isomeric compound, iodo-2-butadienc-l,3;
n 2 1.561. It polymerized completely within forty-eight hours.

Hydroxy-4-butadiene-l,2 (II). Chloro-4-butadicnc-l,2 (6 moles) was stirred at
60-90 with 6 moles of sodium carbonate in 1500 cc. of water for fifteen hours. Ether
extraction and distillation gave the carbinol in 50% yield (losses were due to solubility
in water and formation of higher-boiling products) ; colorless liquid miscible with water
and most organic solvents; sharp pungent odor, lachrymatory, strongly vesicant;
b. p. 68-70 (53 mm.), 126-128 (756 mm.); 4 0.9164; w 2 ,? 1.4759. The carbinol was
also obtained in 50% yield by similarly hydrolyzing the bromide II.

Anal. Calcd. for C 4 H 6 O: C, 68.57; H, 8.57. Found: C, 68.65; H, 8.68.

Hydrogenation of the carbinol gave w-butyl alcohol identified through its 3-nitro-
phthalic ester (mixed m. p. 145).

The carbinol was recovered unchanged after being refluxed with sodium ethylate in
ethanol, 25% aqueous sulfuric acid, or 2% aqueous hydrogen chloride.

When the carbinol (0.5 mole) was refluxed with 100 cc. of 18% hydrogen chloride
for one hour, a 9% yield of chloro-4-butadiene-l ,2 was obtained ; the rest was unchanged

XV. HALOGEN-4-BUTADIENES-l,2 367

carbinol. By the action of phosphorus trichloride on the carbinol in the presence of a
little pyridine a 62% yield of ehloro-4-butadiene-l,2 was obtained. No chloroprene
was found in the product.

Acetate of Hydroxy-4-butadiene-l,2. To hydroxy-4-butadiene-l,2 with excess
acetic anhydride was added a drop of sulfuric acid. The lively reaction was moderated
by cooling and the mixture was finally refluxed for one-half hour. Dilution with water,
extraction and distillation gave a 75% yield of the acetate; b. p. 85-86 (125 mm.),
140-140.5 (780 mm.); 4 0.9641; w 2 D 1.4504. The same compound was obtained by
refluxing for seven hours one mole of chloro-4-butadiene-l,2 with 2 moles of sodium
acetate dissolved in 200 cc. of glacial acetic acid; yield 59%. Similarly from bromo-4-
butadiene-1,2 a 73% yield of pure acetate was obtained. There was no indication in
either case of the presence of any isomeric compound.

Anal. Calcd. for C 6 H 8 O,: C, 64.28; H, 7.14. Found: C, 64.30; H, 7.16.

Hydrogenation of the acetate gave pure n-butyl acetate. vSaponification gave pure
hydroxy-4-butadiene-l,2.

Rearrangement of Chloro-4-butadiene-l,2 (II). To the observations recorded
previously (1) the following may be added. The chloride II (30 g.) with 50 cc. of 18%
hydrogen chloride after two hours at 70-80 gave 2 g. of crude III. The rest was un-
changed II. The chloride II (50 g.) with 150 cc. of 18% hydrogen chloride and 10 g. of
cuprous chloride shaken at 20 for sixteen hours gave a 70% yield of III. The rest was
polymer; no II was recovered. When 3% of dry ferric chloride was added to the
chloride II heat was evolved and distillation set in at once. The distillate was a mixture
of II and III. We are indebted to Dr. D. D. Coffman for the last two of these experi-
ments.

Summary

Compounds of the formula CH 2 C~CHCH 2 X in which X is Cl, Br,
I, OH and OAc undergo a series of metathetical reactions without yielding
abnormal (rearranged) products. The chloride and the bromide are
formed by the 1,4-addition of HX to vinylacetylene. They are easily and
irreversibly rearranged to CH 2 =CX CH=CH 2 . The iodide rearranges
spontaneously. The direct bearing of these facts on theories concerning
the mechanism of 1,4-addition and of ^^-rearrangement is discussed.

Bibliography and Remarks

(1) Carothers, Berchet and Collins, Journ. Am. Cheni Soc , 54, 4000 (1932), this volume, page 306.

(2) Carothers, Williams, Collins and Kirby, Jonrn Am Cheni Soc , 53, 4203 (1931) , this volume,
page 281.

(3) Carothers and Berchet, Journ. Am. Chem. Soc., 55, 2813 (1933); this volume page 308.

(4) Carothers, Berchet and Collins, Joufn. Am Chem. Soc., 54, 4066 (1932) , this volume, page 300.

(5) Carothers, Collins and Kirby, Journ. Am. Chem. Soc., 55, 789 (1933), this volume, page 314.

(6) See for example, Gillet, Bull soc. chim. Belg., 31, 366 (1922); Prevost, Bull soc chim., 43,
990 (1928); Burton, /. Chem. Soc., 1651 (1928); Burton and Ingold, ibid., 904 (1928); Whitmore,
Journ Am. Chem. Soc., 54, 3274 (1932).

(7) Ingold and Ingold, J. Chem. Soc., 756 (1932)

(8) Carothers and Coffman, Journ. Am. Chem. Soc , 51, 588 (1929); 52, 1254 (1930).

(9) Cf. Meerwein, Ann., 455, 227 (1927).

(10) Johnson, Journ. Am. Chem. Soc , 55, 3029 (1933).

(11) Carothers, Collins and Kirby, ibid , 55, 789 (1933), this volume, page 314.

368 ACETYLENE POLYMERS AND THEIR DERIVATIVES

XVI. The Preparation of Orthoprenes by the Action of
Grignard Reagents on Chloro-4-butadiene-l,2*

The term orthoprene is here presented as a designation for derivatives of
butadiene-1,3 having a single substituent and that in the 2-position (III).
Isoprene is the historically important member of this class and (except
for a single reference (1) to ethylbutadiene) it was the only one known until
the discovery of chloroprene (2) and bromoprene (3). The extraordinarily
superior properties of these compounds from the standpoint of rubber
synthesis prompted the preparation and examination of other dienes (4).
The results (5) suggested that similarly desirable properties are not likely
to be found among any other types of dienes than the orthoprenes as such,
and it became important to obtain further members of this class. The lack
of any satisfactory general method for this purpose led to the development
of the method described here, which is based on the observation that chloro-
4-butadiene-l,2 (I) like other substituted allyl halides (6) reacts with
Grignard reagents to produce abnormal (III) as well as normal (II) prod-
ucts. The theoretical implications of this fact have been discussed in
the preceding paper (7).

CH.r=O=CHCH 2 Cl (I) CH 2 -=C==CHCH,R (II) CH2=CR CH===CH 2 (III)

For the purpose in view the new method leaves much to be desired.
Yields are rather low, and separation of the desired product from by-
products is rather laborious. No doubt further study would lead to con-
siderable improvement. In its present state, however, the method has
sufficed for the isolation of two orthoprenes especially wanted : w-hepto-
prene and phenoprene. As precursors of rubber neither of these com-
pounds approaches chloroprene; they are in fact probably inferior to iso-
prene. Heptoprene polymerizes rather more rapidly than isoprene but the
product obtained under most conditions appears to be softer and weaker.
Phenoprene polymerizes ten to one hundred times as rapidly as isoprene.
Under most conditions the predominating product is the dimer, a crystal-
line solid which has, of course, no rubber-like properties. The high polymer
formed, for example, at very high pressure appears to have a relatively low
molecular weight and, although slightly rubber-like, it is soft and deficient
in strength and elasticity. The experiments here reported dealt with
reagents in which R was methyl, w-butyl, w-heptyl, phenyl and benzyl.

* W. H. Carothers and G. J. Berchet, Journ. Am. Chem. Soc., 55, 2813-7 (1933) ;
Contribution No. 121 from the Experimental Station of E. I. du Pont de Nemours and
Co,

Received December 20, 1932. Published July 6, 1933.

XVI. PREPARATION OF ORTHOPRENES 369

From methylmagnesium chloride or iodide or from w-heptylmagnesium
bromide only the abnormal products (III) were isolated in a state of purity.
The product from w-butylmagnesium bromide contained a considerable
amount of ^-octane from which the butyl-2-butadiene-l,3 could not be
completely separated. The presence of the latter compound was, how-
ever, demonstrated by the Diels-Alder reaction. Phenylmagnesium bro-
mide gave both the normal and the abnormal products which were sepa-
rated and isolated in a state of purity. From benzylmagnesium chloride
only the normal product was isolated. The structures of the orthoprenes
were established by their reaction with naphthoquinone to form the crystal-
line addition products (IV), which were readily oxidized to the correspond-
ing anthraquinones (V). The phenyl-4-butadiene-l,2 was identified by
hydrogenation to n-butylbenzene. Benzyl-4-butadiene-l,2 was identified
by its conversion to phenylpropionaldehyde by ozonization.

R i

Experimental Part

Preparation of Isoprene. A C^rignard reagent from 312 g. (2.2 moles) of methyl
iodide and 53 g. of magnesium (2.2 moles) in -butyl ether was treated slowly with 177
g of chloro-4-butadiene-l,2 (2 moles). The mixture was re fluxed for one-half hour, acidi-
fied, and the material boiling below 120 (74 g.) removed from the butyl ether by dis-
tillation. In a similar manner methylmagnesium chloride from 40 g. (1.65 moles) of
magnesium in butyl ether with 132 g. of chloro-4-butadicne-l,2 (1.5 moles) yielded 35
g. of liquid boiling below 110. The products from the two experiments were combined
and distilled. Except for a considerable dark residue having the odor of dibutyl ether,
only a single fraction was obtained. This was isopreiic (47 g.) boiling at 34.5 to 35.

In alcohol solution it reacted readily with naphthoquinone yielding methyl-2-
tetrahydro-l,4,4a,9a-anthraquinone-9,10 (IV, R = CH 3 ) which crystallized from alcohol
in white needles melting at 86 (copper block) (8). When suspended in alcoholic potash
it was readily oxidized by air to /3-methylanthraquinone, m. p. 177 (9).

-Butyl-2-butadiene-l,3. Six moles of butylmagnesium bromide in ethyl ether
was treated with 5 moles of chloro-4-butadiene-l,2. After acidification, distillation of
the ethereal layer yielded a series of fractions boiling between 33-45 at 49 rnm. and a
residue composed of a viscous, somewhat elastic mass. The largest fraction (126.5 g.)
boiled at 44-45 (29 mm.) (121-123 (760 mm.)). This is very close to the boiling point
of w-octane (125) and analysis (C, 84.8; H, 13.5) showed that its composition lay be-
tween that of octane and butylbutadienc. Bromine titration indicated the presence

370 ACETYLENE POLYMERS AND THEIR DERIVATIVES

of about 57% of the latter compound, and this was demonstrated to be butyl-2-buta-
diene-1,3 by the fact that the mixture when heated for two hours with an equal weight of
napthoquinone yielded butyl-2-tetrahydro-l,4,4a,9a-anthraquinone-9,10, (IV, R =
n-C4H 9 ), white microscopic crystals from 80% alcohol, m. p. 63-64 (copper block).

Anal. Calcd. for CwHsoOj: C, 80.59; H, 7.46. Found: C, 79.40, 80.31; H,
7.21, 7.53.

Oxidation with air in dilute alcoholic potassium hydroxide gave /3-w-butylanthra-
quinone (V, R = w-C 4 H 9 ); yellow crystals from alcohol, m. p. 89 (copper block).

Anal. Calcd. for C 18 H 14 O 2 : C, 81.81; H, 6.06. Found: C, 81.08, 80.83; H,
6.43, 5.69.

n-Heptyl-2-butadiene-l,3 (Heptoprene). One mole of w-heptylmagnesium bromide
in ethyl ether was treated with one mole of chloro-4-butadiene-l,2. After acidification,
distillation of the ethereal solution yielded (a) 13 g. boiling at 47-48.5 (5 mm.), (b) 32
g. at 52-54 (5 mm.), and (c) 42 g. boiling chiefly at 99-101 (3 mm.). Fraction (c)
crystallized on being cooled, and it was apparently tetradecane (calcd : C, 84.84; H,
15.15. Found: C, 84.05; H, 15.19); m. p. 5-7. Analysis of fraction (a) showed that
it contained considerable amounts of material other than hydrocarbon. Fraction (b)
was w-heptyl-2-butadiene-l,3; w 2 ,? 1.4511; 4 0.7796; M K calcd., 52.05; MR found,
52.52.

Anal. Calcd. for C,,H 20 : C, 86.84; H, 13.15. Found: C, 85.43, 85.44; H,
12.96, 13.40.

When heated with naphthoquinone at 90-100 for two hours it gave w-heptyl-2-
tetrahydro-l,4,4a,9a-anthraquinone-9,10 (IV, R = w-C 7 Hi 5 ); white needles from ace-
tone; m. p. 81.

Anal. Calcd. for C 2 iH 26 O 2 : C, 81.29, H, 8.38. Found: C, 81.02; H, 8.64.

Oxidation with air in the presence of alcoholic potash gave /3-w-heptylanthraquin-
one; pale yellow crystals from alcohol; m. p. 87 (copper block).

Anal. Calcd. for C 21 H 22 O 2 : C, 82.35; H, 7.18. Found: C, 81.77; H, 6.96.

Action of Phenylmagnesium Bromide on Chloro-4-butadiene- 1,2. This reaction
was complicated by the fact that considerable amounts of phenol were always formed
even when attempts were made to exclude air by passing a stream of nitrogen into the re-
action flask. The following experiment is typical. Seven moles of phenylmagnesium
bromide in ethyl ether was treated with 6 moles of chloro-4-butadiene-l,2. The mix-
ture was washed first with dilute acid and then with dilute alkali to remove the phenol,
dried and distilled. A certain amount of benzene due to the excess of the Grignard re-
agent distilled first, and then the following fractions were collected: (a) 52-60 (17
mm),23.5g.; (b) 60-61 (17 mm.), 72 g.; (c) 61-72 (17 mm.), 8.5 g.; (d) 72-73 (17
mm.), 31 g.; (e) residue, 209 g.

The fractions (b), (d) and (e) were further purified as described below and the
following compounds obtained: phenyl-4-butadiene-l,2 (47.2%), phenyl-2-butadiene-
1,3 (8.4-9.2%), dimer of phenyl-2-butadiene-l,3 (25.3-26.7%). The yields are based on
the chloro-4-butadiene-l,2 applied, and the two sets of figures result from two separate
experiments. By a slight modification of the conditions the yield of phenyl-2-butadiene-
1 ,3 was greatly increased. The reaction product was decomposed very rapidly with ice
and dilute acid, the ethereal solution was kept at low temperature until it could be dis-
tilled, the ether was evaporated in vacuo, and the distillation was carried out at a lower

XVI. PREPARATION OF ORTHOPRENES 371

pressure (2.5 mm.) than was used in previous experiments. The yield of pure phenyl-
2-butadiene-l,3 was 24%.

Fraction (b) was redistilled and then showed w 2 ,? 1.5489; d 0.9226; MR calcd.,
43.85; MR found, 44.93. It was identified as phenyl-2-butadiene-l,3 (phenoprene).

Anal. Calcd. for C, H 10 : C, 92.31; H, 7.69. Found: C, 91.60; H, 7.69.

When heated with an equal weight of naphthoquinone at 90-100 it gave phenyl-
2-tetrahydro-l,4,4a,9a-anthraquinone-9,10 (IV, R = C 6 H 6 ); crystallized from acetone,
m. p. 146-147 (copper block).

Anal. Calcd. for C 20 H 16 O 2 : C, 83.33; H, 5.55. Found: C, 83.00; H, 6.00.

Oxidation with air in the presence of alcoholic potash gave -phenylanthraquinone,
identified by mixed melting point (163-164) (10).

Anal. Calcd. for C 2 oH, 2 O 2 : C, 84.50; H, 4.22. Found: C, 84.14; H, 4.43.

Fraction (d). After redistillation this showed n 1.5460; d 0.9220; MR calcd.,
43.85; MR found, 44.64. It was identified as phenyl-4-butadiene-l,2 by its analysis and
by hydrogenation to n-butylbenzene. It failed to react with naphthoquinone.

Anal. Calcd. for C] Hi : C, 92.31; H, 7.69. Found: C, 92.44; H, 7.77.

Hydrogenation in alcohol with PtOa catalyst was rapid and complete. The product
had b. p. 178-179; d\l 0.863; n 1.4895. w-Butylbenzene has b. p. 180; </ 15 0.864;
n l S 5 1.494 (11).

Fraction (e). Distillation of fraction (e) gave first some diphenyl, and then the
bulk of the material distilled at 220-225 (10 mm.) with only a very small residue.
The distillate solidified when cooled, and after crystallization from ethanol gave white
needles melting sharply at 62. The same product was slowly formed from phenyl-2-
butadiene-1,3 on long standing or more rapidly under the action of heat, but not from
phenyl-4-butadiene-l ,2. Analysis shows that it is a dimer of phenyl-2-butadiene-l,3.

Anal. Calcd. for C 2 oH 2 o: C, 92.31; H, 7.69; mol. wt. 260. Found: C, 92.92;
H, 7.71 ; mol. wt. (in freezing benzene), 232, 232.

Benzyl-4-butadiene-l,2. One and one-half moles of benzylmagnesium chloride in
ethyl ether was treated with 1.13 moles of chloro-4-butadiene-l,2. The reaction had a
tendency to proceed by spurts. Distillation of the reaction product through an efficient
column gave three fractions: (a) 30.5 g. 72-73 (7 mm.); (b) 73 g. 76-77 (7 mm.);
and (c) 30 g. of residue. Fraction (c) was probably mostly dibenzyl. Neither (a) nor
(b) reacted with naphthoquinone. Fraction (b) was identified as benzyl-4-butadicnc-
1,2 and (a) was apparently a less pure specimen of the same material. Fraction (b)
showed w 2 D 1.5400; d\ Q 0.9169; MR calcd., 48.45, M R found, 49.28.

Anal. Calcd. for C n H^: C, 91.66; H, 8.33. Found: C, 91.12; H, 8.87.

When oxidized with potassium permanganate it yielded benzoic acid. Ozoniza-
tion in chloroform solution followed by hydrolysis gave an oil having the odor of phenyl-
propionaldehyde. Its oxime melted at 95-97.

Anal. Calcd. for C 8 H n ON: C, 72.48; H, 7.38; N, 9.39. Found: C, 71.68; H,
7.32; N, 8.83.

Summary

Diene hydrocarbons of the formulas CH 2 : ^CR CH=CH 2 and/or
arc obtained by the action on chloro-4-butadiene-l,2

372 ACETYLENE POLYMERS AND THEIR DERIVATIVES

of RMgX where R is methyl, w-butyl, n-heptyl, phenyl and benzyl.
Some polymers and derivatives are described.

Bibliography and Remarks

(1) Ipatiew, J. prakt. Chem., [2] 59, 534 (1899).

(2) Carothers, Williams, Collins and Kirby, Journ, Am. Chem. Sue , 53, 4203 (1931), this volume,
page 281.

(3) Carothers, Collins and Kirby, Journ. Am. Chem. Soc., 55, 789 (1933); this volume, page 314

(4) Carothers and Coffman, Journ. Am. Chem. Soc , 54, 4071 (1932), this volume, page 384,
Jacobson and Carothers, Journ Am. Chem. Soc , 55, 1G24 (1933); this volume, page 331.

(5) See also Whitby and Gallay, Caw J. Research, 6, 280 (1932).

(6) Prevost and Daujat, Bull. soc. chim , 14] 47, 588 (1930).

(7) Carothers and Berchet, Joutn. Am. Chem. Soc., 55, 2807 (1933); this volume, page 301

(8) Diels and Alder [Ber., 62, 2357 (1929)] report 81.

(9) Diels and Alder report 175.

(10) vScholl and Neovius [Ber., 44, 1075 (1911)] give 160 to 161. The melting point recorded
above is observed in a capillary tube on slow heating. On a copper block the substance first melts at
about 145 with the evolution of some gas; it then solidifies and melts again at 1(51 to Ifi5. If the
substance is first heated in vacua at 125, only a single melting point of 103-104 is observed 0-
Phenylanthraquinone from another (commercial) source showed a similar behavior

(11) Beilstein, IV ed., Vol. V, page 413.

XVII. Mercury Derivatives of Vinylacetylene*

As a means of identifying true acetylenic compounds the derivatives
formed by the action of K 2 HgI 4 are useful since they are generally crystal-
line and have definite melting points (1). When this reagent is applied to
vinylacetylene (I) it furnishes di-vinylethynyl-mercury (II), which sepa-
rates from alcohol in the form of white leaflets melting at 144 to 145.
The compound is readily soluble in chloroform, but on standing in the air
for twenty-four to forty-eight hours it becomes yellow in color and insol-
uble in chloroform. Analysis indicates that considerable amounts of
oxygen are absorbed during this transformation. The oxidized product
does not melt, but sometimes explodes on being heated or subjected to
mechanical shock.

Di-vinylethynyl-mercury undergoes the following transformations.
Cold dilute hydrochloric acid regenerates the vinylacetylene; chlorine and
bromine yield the a-chloro- and bromo-vinylacetylenes (III) ; and metallic
sodium in dry benzene produces sodium vinylacetylide.

Di-vinylethynyl-mercury is also produced by the action of mercuric ace-
tate in acetic acid on vinylacetylene at ordinary temperatures, but if the

* W. H. Carothers, R. A. Jacobson and G. J. Berchet, Journ. Am. Chem. Soc., 55,
4(565-7 (1933); Contribution No. 125 from the Experimental Station of R. I, du Pont
de Nemours and Co.

Received July 21, 1933. Published November 7, 1933.

XVII. MERCURY DERIVATIVES OF VINYLACETYLENE 373

temperature is raised to 60 or 70 reaction proceeds further with the forma-
tion of a compound whose composition and chemical behavior agree with
that required for 1 , 1 -di-acetoxymercuri-2-acetoxymercurioxy- 1 ,3-buta-
cliene (IV). The same product is obtained by treating di-vinylethynyl-
mercury with mercuric acetate in acetic acid at 60 to 70 (2).

^=CH C=C) 2 Hg CH 2 =CH C

(I) (II) (III)

CH,~-CH- (CH 3 COOHgO)C=C(HgOOCCH 3 ) 2 CH 2 =CH

(IV) (V)

CH2=CH CO CBr 3 CHjp^CH (OHgX)C=C(HgX) 2

(VI) (VII)

This compound is a white crystalline solid somewhat soluble in water
and more soluble in dilute acetic acid. It is infusible. By the action of
hydrochloric acid it is decomposed with the formation of methyl vinyl
ketone (V). Similarly by the action of bromine it is converted into tri-
bromomethyl vinyl ketone (VI). The action of potassium iodide, bro-
mide or chloride on IV causes replacement of the acetyl groups by halogen
and gives insoluble products corresponding in composition with the general
formula VII.

Di-vinylethynyl-mercury.- The addition of an alcoholic solution of vinylacetylene
to an excess of the alkaline mercuric iodide reagent (1) yielded a copious precipitate.
This was filtered off and crystallized from boiling absolute alcohol. It separated in the
form of white leaflets melting at 144 to 145.

Anal. Calcd. for C 8 H 6 Hg: Hg, 66.25. Found: Hg, 66.46, 66.35.
The same compound was obtained by adding vinylacetylene (30 g.) to a cold solu-
tion of 5 g. of mercuric oxide in 250 cc. of acetic acid.

Anal. Calcd. for C 8 H 6 Hg: C, 31.72; H, 1.99. Found: C, 32.00; H, 2.46.

Upon standing for forty-eight hours the di-vinylethynyl-mercury became yellow
in color and insoluble in chloroform. It had also acquired the ability to explode under
the action of heat or mechanical shock. Analysis of this explosive product showed the
presence of only 27.70% carbon and 2.23% hydrogen.

Freshly prepared di-vinylethynyl-mercury reacted with bromine or iodine to give
the corresponding bromo- and iodo-vinylacetylenes. In one experiment 23.5 g. of the
mercury compound in 177 cc. of chloroform was treated slowly with a 10% solution of
bromine in chloroform; 22.64 g. of bromine was absorbed before decolorization ceased.
The chloroform was removed, and distillation of the residue (11 g.) then gave 3 g. of 1-
bromo-2-vinylacctylene, b. p. 50 to 52 at 210 mm.

l,l-Di-acetoxymercuri-2-acetoxymercurioxy-l,3-butadiene (IV). A solution of 5.2
g. of vinylacetylene in 100 cc. of acetic acid was added to a solution of 86.6 g. of mercuric
oxide in 700 cc. of acetic acid surrounded by a water-bath at 45. After one-half hour
the temperature of the bath was increased to 60 and maintained between 60-70 for
four hours. During this period a white crystalline solid separated. The mixture was

374 ACETYLENE POLYMERS AND THEIR DERIVATIVES

filtered and the solid (79 g.) crystallized from dilute acetic acid; yield 93.4%. The prod-
uct was somewhat soluble in water, more so in dilute acetic acid, but insoluble in the com-
mon organic solvents.

Anal. Calcd. for CioHi 2 O 7 Hg3: C, 14.18; H, 1.43; Hg, 71.15. Found: C,
14.14; H, 1.95; Hg, 70.72.

l,l-Di-iodomercuri-2-iodomercurioxy-l,3-butadiene. To a hot solution of 4 g.
of l,l-diacetoxymercuri-2-acetoxymercurioxy-l,3-butadiene in 100 cc. of 50% acetic
acid was added slowly with stirring 2.8 g. of potassium iodide in 20 cc. of water. The
yellow precipitate was filtered, washed several times with hot 50% acetic acid, then with
water, and finally with acetone. The pale yellow solid was insoluble in water and in
the common organic solvents.

Anal. Calcd. for C4H 3 OHg 8 I 3 : Hg, 57,34; 1,36.28. Found: Hg, 55.22, 55.47;
I, 36.27, 36.45.

l,l-Di-bromomercuri-2-bromomercurioxy-l,3-butadiene was obtained in a similar
manner from potassium bromide as a white solid insoluble in water and in the common
organic solvents.

Anal. Calcd. for C 4 H 3 OHg 3 Br3: Hg, 66.23; Br, 26.37. Found: Hg, 67.82,
68.14; Br, 25.18, 25.29.

l,l-Di-chloromercuri-2-chloromercurioxy-l,3-butadiene was obtained in a similar
manner from potassium chloride as a white solid insoluble in water and in the common
organic solvents.

Anal. Calcd. for C 4 HaOHg 8 Cls: Cl, 13.72. Found: Cl, 14,38.

Hydrolysis of l,l-Di-acetoxymercuri-2-acetoxymercurioxy-l,3-butadiene.--To
423 g. of the mercury compound (0.5 mole) in 200 cc. of water was added 300 cc. of hydro-
chloric acid (37%). After three hours, a small amount of hydroquinone was added,
and the solution distilled until the distillate amounted to 300 cc. From the latter,
pure methyl vinyl ketone was isolated by ether extraction and distillation. With phenyl-
hydrazine it gave the known derivative phenyl-l-methyl-3-pyrazoline melting at 76 (3).

Action of Bromine on l,l-Di-acetoxymercuri-2-acetoxymercurioxy-l,3-butadiene.
Bromine (239 g.) was slowly added to a suspension of 285 g. (0.337 mole) of the
mercury compound in 700 cc. chloroform until decolorization was complete. Cooling
was required to maintain the temperature below 60. After standing, the mixture was
filtered, washed with 10% hydrochloric acid and with water and distilled. At 0.5 mm
44 g. of a yellow viscous liquid boiling at 128-130 was obtained. On standing for forty-
eight hours it solidified. Three crystallizations from petroleum ether gave white
rectangular plates melting at 73-75. Although additional crystallizations did not
alter the melting point, the product was not analytically pure. However, the ease with
which bromoform was liberated when the crystals were warmed with dilute alkali indi-
cates that the product was mainly tribromomethyl vinyl ketone contaminated perhaps
with dibromomethyl vinyl ketone.

Anal. Calcd. for C 4 H 3 OBr3: C, 15.64; H, 0.98; Br, 78.15. Found: C, 16.42,
16.85; H, 1.89, 1.75; Br. 72.18, 71.89.

Summary

Vinylacetylene when treated with potassium mercuri-iodide or with mer-
curic acetate at the ordinary temperature yields di-vinylethynyl-mercury, a

XVIII. l-HALOGEN-2-VINYLACETYLENES 375

crystalline solid melting at 144 to 145. The action of mercuric acetate on
vinylacetylene at 60 to 70 yields l,l-di-acetoxymercuri-2-acetoxymercuri-
oxy-l,3-butadiene. Some reactions of these compounds are described.

Bibliography and Remarks

(1) Johnson and McEwen, Journ. Am. Chem. Soc., 48, 469 (1926).

(2) Myddleton, Barrett and Seager, ibid., 52, 4405 (1930), have presented evidence favoring the
general structure C(OHgOAc) = C(HgOAc)z for the products obtained by the action of mercuric
acetate on monosubstituted acetylenes.

(3) Maire, Bull. Soc. Chim., (4) 3, 272 (1908).

XVIII. l-Halogen-2-vinylacetylenes*

By the action of alkaline hypohalites, true acetylenic hydrogens are
generally replaced by halogen (1). This reaction has now been applied to
vinylacetylene, and the l-halogen-2-vinylacetylenes whose properties are
listed in Table I have been prepared. In a general way they resemble
other halogen acetylenes. They are liquids with highly characteristic
repulsive sickening odors. Under diminished pressure in an atmosphere
of nitrogen they can be distilled, but dangerous explosions occur if air is
present, or if heating of the residue is carried too far. A sample of the
chloro compound in one instance inflamed spontaneously when a specimen
was being removed for analysis.

The compounds when freshly distilled are colorless but they darken on
standing and are finally transformed into black, brittle solids. These
solids are sensitive to heat and percussion and they explode with con-
siderable violence. The product from the iodo compound is the most
sensitive and violent; that from the chloro compound the least.

PHYSICAL PROPERTIES OF CH 2 =CH C=C X

Nature of X B. p., C.

Cl 55 to 57 at 700 mm.
Br 52 to 5,'} at 217 mm.
J 78 at 125 mm.

The bromo compound was formed rather rapidly when vinylacetylene
was treated with potassium hypobromite; hypochlorite acted very much
more slowly and the yields of the chloro compound were rather low. In

* R. A. Jacobson and W. H. Carothers, Journ. Am. Chem. Soc., 55, 4667-9 (1933) ;
Contribution No. 126 from the Experimental Station of E. I. du Pont de Nemours and
Co.

Received July 21, 1933. Published November 7, 1933.

a l?

rf a j

MRn
calcd.

MRn
found

Exaltation

1 . 4603

\ 0032

22 94

23 . 89

0.95

1.5182

1 . 4804

25 81

26.80

.99

1 5948

1 8968

31.07

31.88

.81

376 ACETYLENE POLYMERS AND THEIR DERIVATIVES

preparing the iodo compound a solution of iodine in potassium iodide was
used. The iodo and bromo compounds have also been obtained (2) by the
action of iodine and bromine on di-vinylethynyl-mercury. The iodo com-
pound was also formed when vinylethynylmagnesium bromide was treated
with iodine. Since a deficiency of iodine was used, it appears that the
tendency of the iodo compound to react with the Grignard reagent is very
slight.

In alcoholic solution the iodo compound adds one molecule of hydro-
gen chloride. This reaction was carried out in the expectation that the
product would be the substituted chloroprene, CH 2 =CH CC1 -~-=CHI.
The product did indeed polymerize spontaneously, but it yielded only a
sticky black tar and no attempt was made to confirm its structure.

Experimental Part

Preparation of l-Bromo-2-vinylacetylene. Bromine (80 g.) was added at to
180 g. of potassium hydroxide in 800 g. of water. Then 30 g. of vinylacctylene was
added during one-half hour with stirring under nitrogen. After two hours a heavy oily
layer separated. This was dried with calcium chloride and distilled under reduced
pressure of nitrogen. Some vinylacetylene was recovered, and then 36.4 g. (55%) of 1-
bromo-2-vinylacctylene was collected between 52 and 53 at 21 7 mm. A small amount
of liquid remaining in the distilling flask exploded when further distillation was at-
tempted. The bromovinylacetylene though colorless at first, quickly became yellow
and then progressively darker upon standing. Its odor was nauseating and exposure to
its vapors caused headaches. A specimen stabilized with hydroquinone did not poly-
merize within a month. When examined several months later it was a highly explosive
black solid.

Anal. Calcd. forC 4 H 3 Br: C, 36.GG; H, 2.27, mol. wt., 130.9. Found: C, 36.72;
H, 2.46; mol. wt., 131, 136 (cryoscopic, benzene).

Preparation of l-Iodo-2-vinylacetylene. To a solution of 84 g. (1.5 mole) of potas-
sium hydroxide in 1000 g. water at 0, 65 g. of vinylacetylene was added and then, with
vigorous stirring during two hours, a solution of 140 g. of potassium iodide and 127 g. of
iodine in 110 cc. of water. After standing overnight, the l-iodo-2-viriylacetylene (95 g.)
which had separated to the bottom was removed, dried with calcium chloride, and dis-
tilled under reduced pressure of nitrogen; yield, 49 g. or 27.5%. The compound was at
first colorless, but on standing it became reddish-brown. After a month at 10 it had
polymerized to a jelly, and after several months to a black solid similar in appearance to
charcoal. This polymer was extremely explosive.

Anal. Calcd. for C 4 H 3 I: C, 26.97; H, 1.69; mol. wt., 177.95. Found: C, 27.30;
H, 1.75; mol. wt., 181 (cryoscopic, benzene).

The iodo compound was also obtained by the action of iodine on vinylethynyl-
magnesium bromide. To an ethereal solution containing one mole of the reagent, 127
g. (one gram atom) of iodine was slowly added. A smooth reaction occurred and the
iodine instantly dissolved with decolorization. A small test sample was removed and
found to take up considerably more iodine. However, the main portion was distilled

XVIII. l-HALOGEN-2-VINYLACKTYLENES 377

in vacuum and a liquid boiling at 82 to 83 at 150 mm. was collected. This agreed in
odor and physical constants with the l-iodo-2-vinylacetylene described above.

Preparation of l-Chloro-2 -vinylacetylene. Some difficulty was at first experienced
in obtaining this compound, but with the following procedure it was isolated in yields
of about 10%. To a solution of 500 g. of potassium hydrioxde in 1800 cc. of water at
were added 225 g. of chlorine and then 300 g. of vinylacetylene. The mixture was
stirred for eight hours. After standing overnight the upper layer (47 g.) was separated,
dried with calcium chloride, and distilled under nitrogen. Two distillations gave 23 g.
of l-chloro-2- vinylacetylene boiling at 55 to 57. It had a nauseating odor and, while
colorless at first, it darkened on standing and ultimately polymerized to a black, brittle
resinous solid.

Anal. Calcd. for C 4 H 3 C1: C, 55.49; H, 3.49; mol. wt. 86.5. Found: C, 55.28;
H, 4.06; mol. wt., 87, 89 (cryoscopic, benzene).

Addition of Hydrogen Chloride to l-Iodo-2-vinylacetylene. A solution of hydrogen
chloride (38 g.) in 95% alcohol was mixed in a pressure bottle with 49 g. of l-iodo-2-
vinylacetylene, 10 g. of ammonium chloride, and 15 g. of cuprous chloride. The bottle
was shaken for twenty-four hours at 25 and the oily layer was distilled first with steam
and then under nitrogen. The distillate was red in color and boiled at 73.574.5
at 35 mm.; yield, 26 g. (44%) ; <J a 2, 1.9161; w'i? 1.6073.

Anal. Calcd. for C 4 H 4 IC1: C, 22.38; H, 1.88; total silver halide, 0.1704 g.; mol.
wt., 214.4. Found: C, 22.34; H, 2.07; total silver halide, 0.1609 g.; mol. wt., 225,
228 (cryoscopic, benzene).

Upon standing for several months the compound polymerized to a soft black tar.

The addition of hydrogen chloride to l-bromo-2-vinylacetylenc under similar con-
ditions also gave a volatile product which polymerized on standing.

Summary

The acetylenic hydrogen of vinylacetylene has been replaced by bromine,
iodine and chlorine and the corresponding l-halogen-2-vinylacetylenes
obtained. These compounds are unstable liquids which polymerize
upon standing to black solid polymers. The latter are explosive when
heated or submitted to mechanical shock. Addition of hydrogen chloride
to the iodo compound yielded an addition product which also polymerized
spontaneously.

Bibliography and Remarks

(1) Straus, Kollek and Heyn, Bcr., 63, 1868 (1930).

(2) Carothers, Jacobson and Berchet, Jaurn. Am Chem Soc , 55, 406.5 (1933), this volume, page
372; Vaughn and Nieuwland, Journ Am Chem Soc , 55, 21.50 (1933), report that iodovinylacetylenc
is obtained from iodine and vinylacetylene in liquid ammonia.

378 ACETYLENE POLYMERS AND THEIR DERIVATlVEvS

XIX. The Structure of Divinylacetylene Polymers*

Divinylacetylcne, when allowed to stand in contact with a little air, is
transformed to a soft, transparent, oxygen-containing jelly which is danger-
ously explosive (1). The action of heat in the absence of air yields a quite
different type of polymer, a yellow oil which dries to hard, chemically
resistant films (1, 2).

We now record some observations concerning the nature of this oily
polymer. The reaction involved in its formation is a typically thermal
polymerization; the rate is but slightly affected by oxidants or anti-
oxidants. In a typical case about 8% of a sample of divinylacetylene was
polymerized in three hours at 80. If the reaction is carried too far (some-
where between 20 and 50%) the mixture sets to a gel. Before this stage
is reached, the polymer is easily isolated by evaporation in a vacuum of
the unchanged monomer. The properties of the residual oil vary con-
siderably depending upon how far the reaction has progressed. In one
experiment a sample taken when 13.4% of the monomer was polymerize