POLYMERIZATION STUDIES OF
4-SUBSTITUTED -1 24-TRIAZOLINE-3, 5-DIONES
SYNTHESIS OF MODEL COMPOUNDS
RELATED TO TRIPLE STRAND POLYMERS
SAM RICHARD TULNm'R
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE
REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
This dissertation is dedicated to my parents.
My stay at the University of Florida has been a pleasant and
rewarding experience and I would like to offer my gratitude to the
many people who have made this experience possible.
I will be forever indebted to my advisor, Dr. George B. Butler,
for his patience, understanding, encouragement and guidance during
the course of this work.
I wish to thank Dr. W. R. Dolbier, Dr. J. A. Deyrup, Dr. M. T.
Vala and Dr. R. B. Bennett for giving of their valuable time to serve
on my supervisory committee. Also I wish to thank Dr. T. Hogen Esch
for his informing discussions with me on portions of this work.
The successful completion of this work would not have been
possible without the assistance of Dr. Lawrence J. Guilbault. His
enthusiastic and tireless research efforts have been and will always
be a great example for me.
Grateful thanks are extended to my fellow laboratory colleagues.
They have generated a pleasant and stimulating environment in which
I also wish to acknowledge the Air Force Office of Scientific
Research, the Petroleum Research Fund and the Tennessee Eastman
Company for providing financial support in the form of research
Finally, I wish to thank my wife, Pamela, for her love and under-
standing which have made the completion of this work an easier task.
TABLE OF CONTENTS
ACKNOWLEDGEMENTS . . . . . . . . . . iii
LIST OF TABLES . . . . . . . .... . . . . v
LIST OF FIGURES . . . . . ... .... . . vi
ABSTRACT . . . . . . . . . . . . vi
I. INTRODUCTION . . . . . . . . . . 1
A. Polymerization Studies of Azo Dienophiles .. 1
B. Triple Strand Polymer Model Compound Studies 9
II. POLYMERIZATION STUDIES OF 4-SUBSTITUTED-1,2,4-
TRIAZOLINE-3,5-DIONES . . . . . . . . 14
A. Copolymerization of Ethyl Vinyl Ether and Divinyl
Ether with 4-Phenyl-1,2,4-triazoline-3,5-dione 14
B. Copolymerization of 4-Phenyl-1,2,4-Triazoline-
3,5-dione with Other Monomers . . . . .. 46
C. Reactions and Attempted llomopolymierizations of
4-Phcnyl-1,2,4-triazoline-3,5-dione . . .. 53
D. Diels Alder Polymers . . . . . . . 58
III. TRIPLE STRAND POLYMER MODEL COMPOUND STUDIES . . 66
A. Attempted Synthesis of Tetracyclo[188.8.131.52'9.1 ]
Dodecane and Related Systems . . . ... 66
B. Attempted Synthesis of 5,5,6,6-Tetrasubstituted
Cyclohexadiene Structures . . . . . .. 82
IV. EXPERIMENTAL . . . . . . . . .. 86
A. General . . . . . . . . . . 86
B. Copolymerizations and Related Reactions of 4-
Substituted-l,2,4-triazoline-3,5-diones . .. 87
C. Syntheses Related to Triple Strand Model Compound
Studies . . . . . . . . . . 121
REFERENCES CITED . . . . . . . . . . . 153
BIOGRAPHICAL SKETCH . . . . . . . 157
LIST OF TABLES
Copolymerizations of EVE and PhTD . . . . . .
Copolymerizations of DVE and PhTD . . . . .
First Order Rate Constants in PhTD at 25C . . . .
Summary of Interception Reaction Results . . . .
Copolymerizations and Reactions of PhTD and Other
Monomers . . . . . . . . . . . ....
Summary of Catalyzed PhTD Reactions . . .
Catalytic Hydrogenation of Diazoquinone Adducts . ..
Reactions of Tetracyanoethylene with a-Pyrone . .
LIST OF FIGURES
1 Nmr spectrum of the copolymer of ethyl vinyl ether and
4-phenyl-1,2,4-triazoline-3,5-dione. . . . .. 16
2 Nmr spectrum of 3-phenyl-6-vinyloxy-l,3,5-triaza-
bicyclo[3.2.0]hepta-2,4-dione . . . . . .... 18
3 Nmr spectrum of the copolymer of divinyl ether and
4-phenyl-l,2,4-triazoline-3,5-dione . . . . .. 21
4 Nmr spectrum of the copolymer of divinyl ether and
4-methyl-l,2,4-triazoline-3,5-dione . . . . .. 22
5 Plot of M versus reaction time for EVE-PhTD in CH2,C
at 25C . . . . . . . . . . . 34
6 Nmr spectrum of 3-oxa-2,2-dimethyl-4-ethoxy-8-phenyl-
1,6,8-triazabicyclo[4.3.0]nona-7,9-dione . . . ... 38
7 Nmr spectrum of l-(formy]nethyl)-2-acetyl-4-phenyl-
1,2,4-triazoline-3,5-dione . . . . . . ... 49
8 Nmr spectrum of the Diels Alder, ene adduct of 4-phenyl-
1,2,4-triazoline-3,5-dione and styrene . . . ... 63
9 Nmr spectrum of the Diels Alder, ene polymer from styrene
and bis triazoline dione . . . . . . . .. 64
10 Plot of consumption of PhTD versus time for EVE-PhTD
in dioxane at 250C . . . . . . . .... ... 110
11 Logarithmic plot of [PhTD] /[PhTD or A /A versus time for
EVE-PhTD in dioxane at 25 .... . . . . 111
Abstract of Dissertation Presented to the
Graduate Council of the University of Florida in Partial Fulfillment
of the Requirements for the Degree of Doctor of Philosophy
POLYMERIZATION STUDIES OF
SYNTHESIS OF MODEL COMPOUNDS
RELATED TO TRIPLE STRAND POLYMERS
Sam Richard Turner
Chairman: Dr. George B. Butler
Major Department: Chemistry
The research described in this dissertation consists of two main
areas of investigation. The first area pertains to the use of azo
dienophiles, particularly 4-phcnyl-l,2,4-triazoline-3,5-dione, in
polymerization studies. The second area involves attempts to synthe-
size model compounds related to proposed triple strand polymers.
A. Polymerization Studies of 4-Substituted-l,2,/-triazoline-3,5-diones
The copolymerization and reactivity of the very potent cyclodien-
ophile, 4-phenyl-l,2,4-triazoline-3,5-dione (PhTD), with electron rich
coreactants were investigated. PhTD was found to spontaneously react
with vinyl ethers at room temperature in methylene chloride solution.
Ethyl vinyl ether (EVE) and isobutyl vinyl ether (IVE) were observed
to copolymerize spontaneously with PhTD to form 1:1 alternating co-
polymers. Divinyl ether (DVE) was observed to form a mixture of the
2+2 cycloadduct and copolymer at room temperature. At 60*C, only copol-
ymer was formed.
The structures of the copolymers were assigned from spectroscopic
and chemical data. Some physical property characterizations of the
copolymers were made.
A propagation mechanism involving the coupling of dipolar inter-
mediates was ascertained as the most probable mechanism of polymeriza-
tion.. Numerous experimental observations were in support of this pro-
posed method of polymerization.
When the spontaneous copolymerizations of EVE and IVE were effected
in acetone or cyclohexanone, the corresponding 1,3,4-tctrahydrooxadia-
zine compounds were obtained as well as the expected copolymers. These
new heterocyclic ring structures were fully characterized. The com-
pounds were believed to result from an interception of the initial 1,4-
dipole intermediate by the weakly dipolarophilic alkyl ketone. DVE was
noted to form only a trace of the corresponding oxadiazine structure.
The difference in the reactivity of EVE and DVE was explained in
terms of the stability of the positive center of the 1,4-dipole.
Other olefinic compounds were reacted with PhTD. Vinyl acetate
was observed to undergo a unique intramolecular rearrangement of the
initially formed 1,4-dipole. Divinyl carbonate only resulted in 2:1
copolymers. N-vinyl carbazole spontaneously copolymerized to yield
a 1:1 copolymer of significantly higher molecular weight than obtained
in the vinyl ethers. The electron poor olefins, divinyl sulfone and
acrylonitrile, were found to be unreactive with PhTD.
Some decomposition reactions of PhTD were studied. Attempts to
catalytically homopolymerize PhTD were unsuccessful. The reaction of
PhTD with nucleophiles like sodium cyanide in dimethylformamide and
triethylamine resulted in the formation of 3,7-diphenyl-l,5-diazabicy-
clo[3.3.0]octa-2,4,6,8-tetraone as well as an unidentified oligomeric
A new bis dionophile, 4,4'-(4,4 -diphcnylmethylene)-bis-l,2,4-
triazoline-3,5-dione, for use in cycloaddition polymerizations was syn-
thesized. It was found to spontaneously react with styrene to give a
high polymer believed to have been formed by first a Diels Alder re-
action and then an ene reaction.
B. Synthesis of Model Compounds Related td Triple Strand Polymers
Two dienes, 2-acetoxymethyl-l,3-butadiene and 2,3-di-(acetoxy-
methyl)-l,3-butadiene, were synthesized by a new sulfone pyrolysis pro-
cedure. The precursor acetoxy sulfones were prepared in good yields
from the corresponding bromides by use of silver acetate in acetonitrile.
The Diels Alder reactions of these acetoxy dienes and the corresponding
dihydroxy diene from the 2,3-diacetoxy diene were studied with p-benzo-
quinone. A facile aronatization or dehydrogenation of the adduct, de-
pending on the conditions, precluded its isolation.
2-Cyclohexene-l-one was found to react sluggishly with the 2-
acetoxymethyl diene. Attempts to force the reaction yielded polymeric
materials. Diazoquinone reacted with the disubstituted dienes to yield
the expected Diels Alder adducts. Various attempts at catalytic hydro-
genation of these adducts were not successful.
Tetracyanoethylene was found to be unreactive with a-pyrone.
Fumaronitrile and p-benzoquinone resulted in the expected adducts.
Bromination of 3,4-dimethylbutadiene sulfone and then dehydro-
bromination resulted in the corresponding 3,4-dimethylthiophene-l,l-di-
oxide. This compound was found to be unreactive with tetracyanoeth-
ylene, but it gave a double Diels Alder adduct with maleic anhydride.
A. Polymerization Studies of Azo Dienonhiles
Reactivity of azo dienophiles
Compounds containing dienophilic nitrogen to nitrogen double bonds
have been extensively studied as reactants in cycloaddition reactions.
Normally these azo compounds have been shown to be more dienophilic
than their carbon counterparts. For example azodicarboxylates, 1, have
a stronger dienophilic reactivity than the corresponding fumarates, 2.1
N ,NC02R H CO R
RO2C 2 RO2C H
However, at the time this research was initiated only one report of the
use of these compounds as monomers in polymerization studies was
recorded.2 This account involved the copolymerization of ethyl azo-
bisformate with the comonomers tetrafluoroethylene, acrylonitrile and
methyl methacrylate. In each case some incorporation of the azo com-
pound into the polymer was verified by elemental analysis. No other
structural characteristics were reported.
4-Phenyl-l,2,4-triazoline-3,5-dione (PhTD), 3, has been shown by
Sauer and Schroder to be the most reactive of the azo dienophiles and
perhaps the most reactive dienophile known. The authors compared the
reactivity of PhTD with tetracyanoethylene by utilizing competition
experiments with 2-chloro-l,3-butadiene. PhTD was observed to react
about one thousand times faster than tetracyanoethylene, which had been
previously described as one of the most potent dienophiles synthesized.4
With the same diene,maleic anhydride, 4, was found to react only about
one half as fast as the tetracyanoethylene. Hoi.ever, with other
dienes tetracyanoethylene reacted as great as 106 times faster than 4.
The structure of 4-substituted-l,2,4-triazoline-3,5 diones is very
similar to that of the carbon to carbon double bond dienophiles maleic
anhydride and N-substituted maleimides, 5. The latter compounds have
been utilized extensively as comonomers in copolynerization studies.
It was expected, then, "a priori" that these nitrogen analogues of maleic
H H H H
N--N Cb C /C'
0 ) 0 KA '0 0 Z0
3 4 5
anhydride and the maleimides would undergo similar copolymerizations
with the appropriate electron rich comonomer.
Cyclocopolymerization, originally reported by Butler,6 has evolved
as an extremely important research area in polymer chemistry, not only
because of the theoretical significance of the reaction but in the main
because some of the cyclocopolymers have been shown to be active anti-
tumor agents. Because of this importance, the mechanistic and prepar-
ative aspects of cyclocopolymerization have undergone a careful system-
atic study. The general mechanism proposed was that of an alternating
intra-inter molecular propagation6 as shown below for the divinyl ether-
maleic anhydride system which is the most thoroughly studied of all the
cyclocopolymerization-systems. Since the initial mechanistic proposal,
evidence consistent with the participation of a donor acceptor complex
0 R 0
R-' R '
R 00 0
0-- ( ,) Z: 0V O
as a comonomer has been presented.8 More recently the divinyl ether
maleic anhydride system has been observed to polymerize thermally in
the absence of a free radical initiator and this has been interpreted
as additional evidence for the participation of a donor acceptor
complex in the copolymerization.
Since 4-substituted-1,2,4-triazoline-3,5-diones are known to be
better dienophiles than the usual electron poor comonomers in cyclo-
copolymerization, it was thought that they might behave as electron
acceptors in the presence of electron rich 1,4-dienes and participate
in cyclocopolymerization. Furthermore, it seemed a distinct possibility,
because of their great reactivity, that a spontaneous cyclocopolymer-
ization might possibly be triggered.
The degradation approach to proving polymer structures is important.
For example, the cyclic structure resulting from the copolymerization
of diallyl quaternary aamronium salts was verified by the following
C -1. to hydroxide 31. 1E I (CH3) N
+ -2. to hydroxide
NBr 2. heat 3. heat
C!3 CH3 crosslinked
H3 3 C3 3 polymer
r J n
No such proof of the cyclic structure of the cyclocopolymer
systems has been possible to this time. If a 2:1 cyclocopolymer were
to form, the unique characteristics of the triazoline dione moiety
would offer a convenient handle to chemically probe the polymer
structure by the general degradation procedure shown below.
O N 0 base H H N2
0H low molecular
N weight products
For the preceding reasons,as well as the desire to build new
cyclic copolymer structures to be tested as anticancer agents, the
behavior of 4-substituted-l,2,4-triazoline-3,5-diones with electron
rich olefins was investigated.
Nitrogen backboned polymers
At the commencement of this research no nitrogen backboned polymer
had ever been reported. If homopolymerization of an azo dienophile
could be effected through the nitrogen to nitrogen double bond,
obviously such a polymer with a nitrogen backbone would result. A
report by Iuisgen of the participation of ethylazobisformate in a
substitution reaction with fluorene proceeding by a free radical chain
mechanism made a free radical polymerization of such an azo dienophile
an attractive possibility.
12 --- 21.
H R= --CO2C2CH3
By envisioning the results of such a successful polymerization
of a 4-substituted-l,2,4-triazoline-3,5-dione, one would have at hand
structure 6. Since such structures are readily hydrolyzable in strong
K t 6
base, a polyhydrazine structure, 7, would be distinctly possible.
These nitrogen backboned structures might possess practical value as
well as, assuredly, possessing theoretical importance. As another
K- r 7
goal of this research, the homopolymerization of 4-substituted-1,2,4-
triazolinc-3,5-diones, was investigated.
In the course of this work, Pirkle and Sticklerl2 reported the
homopolymerizarion of 4-n-butyl-l,2,4-triazo'ine-3,5-dione by initiation
with visible irradiation. These authors presented strong evidence that
the polymer obtained, indeed, possessed a nitrogen backbone.
Dials Alder polymers
Although the number of chemical reactions investigated in organic
chemistry is irmense, those that have been successfully adapted to
yield high polymers are few. Two types of cycloaddition reactions
have met with considerable success in producing high polymers, i.e.,
1,3-dipolar cycloadditions3 and Diels Alder reactions.14
The Dials Alder reaction has developed into the most profitable
application of the use of cycloaddition reactions to form polymers.4
To adapt the Diels Alder reaction to a polymer forming system whereby
the polymer is built up by stepwise Diels Alder cycloadditions requires
that the reactants be made difunctional. Two different approaches
have been utilized. The first involves the reaction of a bis diene
and a bis dienophile, generally referred to as an A-A, B-B system. One
example5 is the reaction of p-benzoquinone, 8, with the acetal prepared
from 2-hydroxymethyl-l,3-butadiene and acetaldehyde, 9.
Co H CH-2 O
The second general approach has utilized an A-B monomer or an
intramolecular dicne, dienophile. The reaction of the substituted
Ct-pyrone, 10, and p-phenylene bismaleimide, 11, is a good example of
0 i0 00
00 0 0
F 0 0 -
CHR -R I ---- polymer
q__. C O-- R -NI
A H 3 Ci2 0
L 0 0
The main limitation and drawback to the use of the Diels Alder
reaction has been the difficulty in obtaining high molecular weight
products. In a polymerization propagation that follows step growth
kinetics, as is the case in a Diels Alder polymerization,7 the degree
of polymerization DP follows the Carothers equationl8 DP = 1/l-p,
where p is the extent of reaction. This means that the reaction must
be practically quantitative before a high molecular weight product is
obtained. For example a reaction with a 98% conversion only yields
a DP of 50 or, in other words, a low average molecular weight.
Hence, any stepwise polymerization reaction that is subject to
side reactions of any significance will not be suitable for formation
of high molecular weight products.
Two main reasons have been put forth for the inability, in most
instances, of Diels Alder polymerizations to achieve high molecular
weights.19 The first involves the ease by which the retrodiene
reaction can occur and the second concerns the chain growth of the
diene under the reaction conditions. Another complicating factor in
some systems has been the precipitation of the rigid ladder type polymer
causing a premature termination.15
Since the Dials Alder reaction is a thermally catalyzed reaction
and these two complicating factors also are enhanced at higher temper-
atures, it is not surprising that the problem of obtaining high
molecular weight has plagued its use in producing high polymers.
Stille20 and coworkers have successfully circumvented the retro-
diene reaction by employing bis dienes such as bis a-pyrones and his
cyclopentadienones that lose carbon dioxide or carbon monoxide respec-
tively and hence prevent the degradative retrodiene reaction from
occurring. An example has been the successful use of 3,3'-(oxydi-p-
phonylene)bis(2,4,5-triphenylcyclopentadienone), 12, and p-phenylene
bismaleimide, 11.20 A polymer with an intrinsic viscosity of 1.01
was obtained in four hours in refluxing 1,2,4-trichlorobenzene.
O -0 + 11
0 Ph Ph 0
S ----Ph -- Ph
0 Ph Ph 0
Another possible route to circumvent the bothersome side reactions
would be to employ extremely reactive reactants that would not need to
be heated in order to obtain the necessary high conversions. Obviously,
a bis 1,2,4-triazoline-3,5-dione would be expected to serve in this
capacity because of its fantastic reactivity. Hence another objective
of this research was the synthesis, and polyn:erization with bis dienes,
of such bis dienophiles.
B. Triple Strand Polymer Model Compound Studies
Thermal stability of ladder polymers
The synthesis and study of thermally stable polymers have been
extremely active areas of polymer chemistry research in recent years.
The extensive accounts cf such research recorded in the chemical
literature attests to the theoretical and practical importance of this
work. No attempts will be made to review these accounts since a recent
book1 and several reviews222324 are readily available.
Physical polymer properties have generally been found to be
related to the molecular weight of the polymer. These structure-
property correlations have established that polymers with ladder or
double strand structures, 14, possess extremely high thermal stability.23
For example polydiallydiphenysilane, having structure 13 was observed
to have enhanced thermal stability over its non-cyclic counterpart
HC \C6H5 n
5 6 6 -13
prepared from the monoallyl derivative.. The reason advanced for the
thermal stability of such cyclic structures is that thermal bond
cleavage which occurs -within the cyclic polymer repeat units of the
backbone does not lead to a lower molecular weight polymer. It is
obvious that thermal bond cleavage in the non-cyclic structures results
in a lower molecular weight as depicted in 14a and 14b.
Therefore one goal that the polymer architect has striven for in
the design of materials possessing thermal stability has been the
synthesis of ladder or double strand polymers. Needless to say, many
different researchers have been successful in using the double strand
polymer concept in preparing materials resistant to thermal breakdown.
Possible triple strand polymers
As an extension of the ladder polymer concept of thermal stability
a triple strand polymer would be expected to possess thermal properties
reflected from a structure that requires three bond cleavages per ring
to cause a decrease in molecular weight. Two possible approaches to
constructing such a polymeric species are: 1) a formation of a tubular
polycyclohexane, 15, from a triply initiated chain reaction of benzene
and 2) a triple 1,4-polymerization of a monomer such as 16 to yield
Model compound studies
Generally, one of the objectives of this research was the
preparation and studies of model compounds related to these two
possible routes envisioned to triple strand polymers.
Much effort has been extended in the attempted synthesis of
tetracyclo[184.108.40.206'9.1 4'8dodecane, 18, the repeat unit of the
tubular polycyclohexane polymer.
The first approach has involved the homo Diels Alder reaction of
cis 4,5-diallylcyclohexene derivatives, 19.26a To date this approach
has not been successful.
The second approach has utilized 1,3,5-trisubstituted cyclobexane
derivatives 20 and 21. The tri ester 21 has been successfully prepared
and failed to undergo the desired cyclization to 18. A compound,
which would lead to a nitrogen analogue of 18, 22, has been successfully
prepared but also failed to yield the cyclic structure 23.27
R e 4 0 e R= -CHC02C2C H CH
base 2 2 2 3
O0CPNhIh NZOC :==0 _'0
The synthesis of compounds such as 24, which are functionally
capable of undergoing an intramolecular base catalyzed cyclization, is
the basis of the third approach to 18. One objective of this work
S 2CH s 2
was the preparation of compound 24, and other related structures as 25,
and cyclization of them to their respective cage structures 26.
A fourth approach to 18, which was also studied as a part of this
research, was the possible intramolecular Diels Alder reaction of 28
to 18. A route to 28 was envisioned by the well documented 1,4-
elimination of bromine by zinc in dimethylformamide28 of 27,29 the
Diels Alder adduct of p-benzoquinone and 2,3-di(bromoethyl)-l,3-buta-
In order for a monomer like 16 to be polymerizable by a triple
1,4-initiation it is immediately obvious that monomers such as 29 and
30, which are steric models for 16, would have to be subject to a 1,4--
polymerization. Hence another objective of this research was the
synthesis of 5,5,6,6-tetrasubstituted cyclohexadienes like 29 and 30
and the study of their polymerizations.
29 C 0 H3
N CN 0/ CH3
29 30 CH 3
Polymerization Studies of 4-Substituted-1,2,4-triazoline-3,5-diones
A. Copolymerization of Ethyl Vinyl Ether and Diviny Ether with
Structure of the copolymers
4-Phenyl-l,2,4-triazoline-3,5-dione (PhTD), 3, spontaneously
copolymerized with ethyl vinyl ether (EVE) in methylene chloride
solution at room temperature and the product was shown to have a 1:1
composition by elemental analysis and nmr analysis. The copolymer was
found exclusive of any 2+2 cycloadduct (1,2-diazetidine), 31a.
a: R= -CH2CH3
I b: R= -CR=CH2
0 =Z3 ) 0
Cycloadditions of ethyl azobisformate, 32, a similar azo dienophile,
have been shown to occur through both nitrogen to give 1,2-diazeti-
dines, 33, or through one nitrogen and one carbonyl oxygen to give
oxadiazines, 34. Therefore, structure 35, where propagation occurred
through one nitrogen and one carbonyl oxygen, and structure 36, where
propagation takes place through the nitrogen to nitrogen bond, appeared
to be the most likely candidates for the structure of the copolymer.
The nmr spectrum (Fig. 1) did not distinguish between the two
\ N 20_2CR R02 C N- N O2R
NZN 2 2
\ 2W C 33
possibilities since the observed resonance signals were consistent
with either structure. The signals were 67.50 (broad, aromatic
hydrogens, 5), 66.05 (broad, hydrogen adjacent to two electronegative
atoms, 1), 63.90 (broad, methylene hydrogens of ether, 4) and 61.20
(broad, methyl hydrogens of ether, 3).
Ph CH Ph
The infrared data, however, indicated that 35 was the predominant
repeat unit of the copolymer structure. A strong 1610 cm.1 absorbance
was observed and it was assigned to the -C=N- chromophore which is
only present in 35. Oxadiazine structures like 34 have exhibited
absorbances in the 1630-1680 m.-1 region and they were assigned to
the -C=N- chromophore.30 A recent report on the infrared spectrum of
N-cyclohexylacrylaldimine, 37, has assigned an observed 1608 cm.-
band to the -C=N- chromophore.31 Stickler and Pirkle12 synthesized
Stickler and Pike synthesized
I I I I I
7.0 6.0 5.0 4.0 3.0 2.0 1.0
Figure 1. Nmr spectrum of the copolymer of ethyl vinyl ether and 4-phenyl-1,2,4-triazoline-3,5-dione.
38, 39 and 40 as model compounds in a recent study of the homopolymer-
ization of triazoline diones. A strong 1605 cm.-i absorbance was
observed in 38 and not in 39 and 40.
C 11 C6 H CH C6 11 H
00 3 1 3
o0mQ -OCHi3 o0 300o),
CH3 CH3 CH3
38 39 40
Neither PhTD nor its precursor 4-phenyl urazole had an infrared
absorbance in the 1610 cm.1 region. This information, along with
that obtained from the model compounds, appears as solid evidence for
the assignment of 35 as the predominant repeat unit of the copolymer.
In contrast to the EVE system, divinyl ether (DVE) and PhTD were
observed to form a mixture of copolymer and an adduct at room temper-
ature in methylene chloride. When the reaction was done at 60C only
copolymer was isolated.
The adduct from the room temperature reaction was isolated in pure
form and was assigned as the 1,2-diazetidine, 3-phenyl-6-vinyloxy-1,3,5-
triazabicyclol3.2.0]hepta-2,4-dione, 31b. The infrared spectrum gave
no 1610 cm. absorbance, but gave strong vinyl ether absorbances at
1645 cm.- and 1625 cm. The nmr (Fig. 2) spectrum gave resonance
I I i 1 I
7.0 6.0 5.0 4.0 3.0 2.0 1.0
Figure 2. Nmr spectrum of 3-phenyl-6-vinyloxy-1,3,5-triazabicyclo[3.2.0]hepta-2,4-dione.
signals at-67.40 (singlet, aromatic hydrogen, 5), 66.60 (quartet,
vinyl hydrogen, 1), 65.90 (triplet, hydrogen next to vinyloxy on ring,
1), 64.50 multiplee, vinylic and ring methylene hydrogens, 4). The
mass spectrum yielded the correct molecular ion and the elemental
analysis agreed with the calculated value.
On warming an unpurified mixture of polymer and adduct to 600C,
a ring opening polymerization occurred. This was noted by the
disappearance of the 65.90 triplet in the nmr spectrum and the formation
of a product which gave identical spectral data as the original 60C
An elemental analysis of the copolymer was indicative of a 1:1
structure and for such a copolymer three structures were probable. The
first two, 41 and 42, were analogous to the EVE case in which only one
double bond is consumed and the third, 43, was precedented by cyclo-
copolymerization in which both double bonds are consumed.
N-N F7N-N --CH-CH-
0= -0Gl-CH-- O .:O0 9
N P N
Ph OH Ph 2
CH2 n 42 J
On the basis of spectroscopic and chemical evidence, repeat unit
41 appeared to predominate. First, the ab:senice of upfield resonances
in the nmr spectrum indicated that there were no methylene units
flanked by saturated carbon atoms. Since a requirement for the cyclic
structure 43 would be upfield resonances, this structure was eliminated.
The resonance signals observed (Fig. 3), 67.48 (broad, aromatic
hydrogens, 5), 66.30 (broad, vinylic hydrogen and hydrogen adjacent to
two heteroatoms, 2) and 64.20 (broad, methylene and vinyl hydrogens, 4),
were consistent with both structures 41 and 42.
The infrared spectrum of the copolymers, as in the EVE copolymers,
showed a strong 1610 cm.- band and this was assigned to the -C=N- unit.
Also present in the infrared spectrum were bands at 1640 cm. and
860 m.-1 due to the vinyl ether chromophore.3233
As insurance that the 1610 cm. band was not associated with
the aromatic moiety, a copolymer of DVE and 4-methyl-l,2,4-triazoline-
3,5-dione was prepared. This copolymer also exhibited the strong
1610 cm. absorbance. The nmr spectrum of this copolymer is given
in Fig. 4.
Catalytic hydrogenation of the DVE copolymer over palladium on
alumina at atmospheric pressure resulted in the absorption of the
theoretical amount of hydrogen for a structure like 41 or 42. The
resulting material exhibited a new 2770 cm.- band due to the saturated
carbon hydrogen stretch, and loss of 1640 cm.- and 860 cm.- bands
of the pendant vinyl group in the infrared spectrum. The 1610 cm.
band also disappeared. New resonances appeared in the nmr spectrum
at 63.90 and 61.20. These signals had the same chemical shifts as
the ethyl protons in the EVE copolymers.
I I I I I I
7.0 6.0 5.0 4.0 3.0 2.0 1.0
Figure 3. Nmr spectrum of the copolymer of divinyl ether and 4-phenyl-l,2,4-triazoline-3,5-dione.
I I I I I I I I
7.0 6.0 5.0 4.0 3.0 2.0 1.0 0.0
Figure 4. Nmr spectrum of the copolymer of divinyl ether and 4-nethyl-1,2,4-triazoline-3,5-dione.
One discrepancy in the hydrogenation results was the disappearance
of the 1610 cm. band in the infrared spectrum. This was thought to
be due to an isomerization over the alumina catalyst for the following
reasons. The -C=N- linkage has been shown to be resistant to
catalytic hydrogenation30 when it appeared in oxadiazine structures
like 34. The catalytic hydrogenation of the EVE copolymer was carried
out as a control and although no hydrogen was absorbed, the loss of
the 1610 cm.-1 band was noted. A plausible explanation is that the
catalyst promotes an isomerization from structure 35 to the structure
bonded through both nitrogens, 34. If this were the case, a decrease
in the molecular weight of the copolymers should be observed since
such a change would require bond breaking and then bond reformation.
This was exactly what was observed as a DVE-PhTD copolymer had a
molecular weight of 1590 before hydrogenation and 490 afterwards. An
attempt to use palladium on carbon as the catalyst only gave a
Physical characteristics of the copolymer
The copolymers of EVE-PhTD and DVE-PhTD were white, highly
electrostatic solids which were soluble in most organic polar solvents.
Maximum number average molecular weights (M ) were normally in the
1000 to 3000 range and were measured by vapor pressure osmometry CVPO)
in acetone solution. The polymeric materials were low melting and
both copolymers softened around 100-110C.
A molecular weight distribution (Mw/Mn) of 3.36 was obtained for
one DVE-PhTD sample from gel permeation chromatography CGPC). The
determination was made in dimethylformamide and a calculation of the
M from the GPC trace gave a value of 2500 versus 2750 from the VPO.
An attempted GPC analysis of an EVE-PhTD copolymer was unsuccessful due
to what was believed to be a degradation of the polymer in the highly
Since the copolyrnerization of EVE and PhTD was certainly a
spontaneous copolymerization, the literature was reviewed to obtain
examples of previous spontaneous copolymerizations. Spontaneous
copolymerizations are differentiated in this discussion from other
types of photoinitiated spontaneous copolymerizations and Lewis acid
catalyzed spontaneous copolymerizations35 by defining a spontaneous
copolymerization as one that occurs when two olefinic monomers are
mixed in bulk or in solution at room temperature or lower with no
additional initiator involved.
All observations of spontaneous copolymerizations reported in
the literature have involved an olefinic monomer pair in which one
partner was electron rich and the other electron poor. Apparently,
the first report of such a copolymerization was disclosed in a
Canadian patent authored by Miller and Gilbert.36 They reported the
spontaneous copolymerization of vinylidene cyanide and alkyl vinyl
ethers. No mechanistic interpretation of their results appeared.
Yang and Gaoni37 have prepared 1:1 copolymers from the spontaneous
reaction of trinitrostyrene, 44, as the acceptor or electron poor
olefin and either 4-vinyl pyridine,45, 2-vinyl pyridine or p-dimethyl-
aminostyrene as the donor or electron rich olefin. The authors
suggested that the monomers formed an initial donor acceptor complex
+ 6 copolymer
which initiated the radical copolymerization of the monomers.
Kosover38 has suggested that the copolymerization is initiated by
the formation of a ground state donor acceptor complex which goes
to an excited ion pair state which then copolymerizes the surrounding
ground state donor acceptor complexes according to the equation shown
above. Such a mechanism is designated by Kosower to be a "T-Class
Reaction" in which the electron transfer in the complex is brought
about by thermal energy. The general equation38 for this reaction,
DAtz ( (DA A.) D +A
proposed by Kosower, includes formation of a ground state complex (D,A)
which experiences a thermally induced electron transfer to the excited
ion pair form CD., A.) and then can go to products or can disassociate
to the separate ion pair excited state.
Butler and Sharpe939 have reported two other systems which fit
the definition of spontaneous copolymerization. One system in which
this type reaction occurs is the cyclocopolymerization of maleic
anhydride and DVE, and the other is the cyclocopolymerization of
divinyl sulfone and DVE.39 Both systems are believed to form a
donor acceptor complex which then goes to an excited state, couples
to form a diradical and thus initiates the copclymerization of
either completed or uncomplexed comonomers.
Mechanism of EVE and DVE copolymerizations
PhTD has been shown to undergo 2+2 cycloadditions with both
indene and p-dioxene.0 The cycloaddition with indene has been show,
to involve a dipolar intermediate since this 1,4-dipolar species has
been trapped with water. p-Dioxene formed the expected 2+2 adduct
plus a lot of discarded polymeric material.
In light of these results and in light of the polymer structures,
the most plausible propagation mechanism is believed to be the initial
formation of a stable 1,4-dipole. This intermediate, in the EVE
system, then went directly to copolymer, while in the DVE system it
formed the 1,2-diazetidine 31b and copolymer. At 600C, the diazetidine
is proposed to open to the initial 1,4-dipole and couple intermolecular-
ly to form the copolymer 41.
The difference in behavior between the EVE and DVE systems can
be rationalized in terms of the stabilities of the positive centers
of the respective 1,4-dipoles. The acid hydrolysis of EVE has been
observed to be 350 times faster than the acid hydrolysis of DVE.40
This difference is due to the stability of the corresponding oxonium
ion intermediates, 47 versus 48. The same positive center stabilization
occurs in the two 1,4-dipoles. Therefore it is believed that the more
stable 1,4-dipole (EVE) has a more pronounced dipolar character and
00 C/ 0 OCH2 H C
links intermolecularly, while the energetically higher DVE dipole
closes to the strained 1,2-diazetidine.
The exact nature of the transition state leading to the 1,4-dipole
is not known. Kosower has proposed that 2+2 cycloadditions which
are dipolar in nature are "T Class Reactions"and they go through an
excited charge transfer complex which then couples to the corresponding
dipolar species. This dipole then closes to a four membered ring.
The formation of the vinyl ether-PhTD intermediate 46 by this pathway
is shown below.
The other possibility is an electrophilic attack of the electron
poor PhTD on the electron rich olefin in which no complex is required
to arrive at 46,i.e.,
__/ CH- 0 R
N N- N-
The first observation that is consistent with the postulated
1,4-dipole is the non-cyclic structure of the DVE--PhTD copolymer. A
diradical intermediate would have been expected to cyclize. However,
cyclization of the 1,4-dipole 46 would be highly unlikely since vinyl
ethers are known to be resistant to anionic attack.
Copolymerizations of the two vinyl ethers with PhTD were found
to be insensitive to a free radical inhibitor and a free radical
initiator. The free radical initiator azobisisobutyronitrile (AIBN)
had no effect on the molecular weight of the EVE copolymers (Table 1,
exp. 11). The copolymer from the DVE case showed some increase in
molecular weight, which some random coupling of pendant vinyl groups
could easily explain (Table 2, exp. 3). The known free radical
inhibitor, m-dinitrobenzene, had no effect on the copolymerization
of DVE (Table 2, exp. 6).
An attempt to free radically copolymerize DVE and PhTD at -45C
by photolytic decomposition of benzoyl peroxide only resulted in a
product identical to the ambient copolymerizations. Since the conversion
was extremely low, the product was probably the result of a slow thermal
reaction, because at the same temperature the same product was obtained
without benzoyl peroxide.
In an attempt to exclude moisture from the system, a copolymer-
ization of DVE and PhTD was effected under more rigorous conditions
than the normal trials. It was felt that since water would terminate
the dipolar step-growing chain that the elimination of moisture would
lead to a higher molecular weight copolymer. The dry conditions
included flame drying of glassware, predrying of monomers and transfer-
ing monomers and solvents in an inert and dry atmosphere. Although
TABLE 1. COPOLYMERIZATIONS OF EVE AND PhTD
Exp. Reaction Conditions
No. Solvent Temp.
1 CH2C12 25
2 CH2C2 25
3 CH2C12 25
4 CH2C12 25
5 CI2C12 25
6 CH2C12 25
7 CH2C12 25
8 C CH2C12 60
10 CH2C12 60
11e CH2C12 60
n 1610 cm.
Exp. Reaction Conditions Product and
No. Solvent Temp. Time Conversion
12 DMF 25 0.5 50.3
13 THF 25 0.5 85.6
14 THF 25 0.5 67.0
15 CH2Cl2 60 118 71.5
16 CH2C12 60 166 77.5
17 CH2Cl2 60 211 76.2
s, (strong); m, (medium); w, (weak); -, (not present).
CSample 1 dissolved and heated.
Sample 1 hydrogenated.
0.5 wt. % AIBN added.
10 mol. % NaPh4B added.
TABLE 2. COPOLYMERIZATIONS OF DVE AND PhTD
Exp. Reaction Conditions Product and a
No. Solvent Temp. Time Conversion n
C (hr.) 1610
1 CH2C12 25 0.5 77.5 450 w
2 CH2C12 60 24 77.8 860 s
3c CH2Cl 60 24 80.8 1590 s
4 CH2C12 60 120 84.5 3130 s
5d CH2CI2 60 24 86.0 2540
6e CH2C12 60 24 65.8 2130 s
7 CH2C12 60 120 85.8 2760 s
8 CH2C2 25 22 85.7 w
9 Dioxane 60 24 66.4 1330 s
Exp. Reaction Conditions
No. Solvent Temp. Time
TABLE 2 (continued).
n (cm. )
1610 1640 860
s, (strong); m, (medium); w, (weak); -, (not present).
C0.5 wt. % AIBN added.
el m-dinitrobenzene added.
Polymer 6 hydrogenated.
a higher molecular weight material was obtained than in the normal
trials, a control in which the glassware was not treated yielded even
higher molecular weights (Table 2, exp. 7 vs. exp. 5).
Szwarc41 has discussed in great detail the effect of trace amounts
of water on ionic polymerizations and the conditions needed to conduct
polymerizations in the absence of water and other impurities. Since
the purification procedures needed for such experiments required
special techniques and equipment, this approach was abandoned.
Surprisingly, the EVE-PhTD copolymerization conducted at 60C
resulted in a lower molecular weight product than in the room tempera-
ture trials. The 1610 cm.-1 band in the infrared spectrum of the
copolymer was absent, although the nmr spectrum was the same. The
elemental analysis remained consistent with a 1:1 copolymer. The room
temperature copolymerization was repeated for a 24 hour period and a
molecular weight of 430 was obtained compared to 2440 for a thirty
minute polymerization time. The 1610 cm.- infrared band was absent
in the product from the 24 hour run. A study of molecular weight
versus time was conducted for the room temperature copolymerization
and the results are shown in Fig. 5. A decrease in intensity of
the 1610 cm.-1 band paralleled the decrease in molecular weight.
Table 1 gives the molecular weight versus time in tabular form. Co-
polymerizations conducted at 600C for extended times showed a rebuild-
ing of the chain length (Table 1, exps. 15, 16 and 17).
These results can easily be accommodated by the proposed copolymer-
ization mechanism. The decrease in molecular weight is believed to
result from a depolymerization back to the 1,4-dipole structure and
then recombination through both nitrogen to give structure 36,
2 4 6 12 22 24
REACTION TIME (HOURS)
Figure 5. Plot of Mn versus reaction time for EVE-PhTD in CH2Cl2 at 25C.
possibly a more stable structure. The DVE copolymer once formed was
observed to be thermally stable at 60"C as noted by no decrease in
molecular weight or decrease in the intensity of the 1610 cm.-1 absorb-
ance. Since the activation energy to reformation of the vinyloxy 1,4-
dipole would be expected to be higher than the ethyloxy because of the
difference in stability of the positive center, this result is also
consistent with the polymerization mechanism.
It is generally thought that dipolar reactions should show pro-
nounced solvent effects. The results of a solvent study of the reaction
rate are shown in Table 3. This study was conducted by observing the
disappearance of the band in the visible spectrum of the PhTD. No
pronounced rate differences were observed in going from the fairly
non-polar solvent benzene to the strongly polar solvents dimethylform-
amide and acetonitrile. There was a distinctive rate difference between
EVE and DVE. This was probably another manifestation of the difference
in the stability of the positive center of the dipole.
Gompper42 has recently reviewed cycloaddition reactions involving
polar intermediates. He cites two cases in which significant solvent
effects may not accompany such reactions. The first case involves
reactants that are highly polar and thus experience significant solvation
effects to effectively negate the solvation of the intermediate. The
second case involves a system where the energies of activation of the
first step and the second step are nearly the same and a change in
solvent changes the rate determining step, but not necessarily the
overall reaction rate.
Since both monomers are highly polar molecules, it is reasonable
that they could experience enough ground state solvation to effectively
cause the solvation of the intermediate to be undetectable.
One criterion which can conclusively demonstrate the presence of
a dipolar intermediate is the interception of this intermediate before
cycloproduct formation.4 When EVE and PhTD were reacted at room
temperature in acetone 3-oxa-2,2-dimethyl--4-ethoxy-8-phenyl-1,6,8-
triazabicyclo[4.3.0]-nona-7,9-dione 49,was formed in 42% yield. Along
with this product was formed a copolymer which from the nmr spectrum
appeared to have about 12% acetone incorporated.
This result is believed to be conclusive evidence for the existence
for the 1,4-dipole since the product arises from the interception of
the 1,4-dipole with the weakly dipolarophilic acetone. Apparently some
2OC 2CH3 CH--0
CH2 OC N 3-A CH3
N- CH --- \
3 0= < o
Ph H OCH C 49
12 2 3
o -N CH3
N N 3
of the intercepted dipole did not close to the six membered adduct but
proceeded to link intermolecularly with other similar segments or with
other 1,4-dipoles to form the copolymcr containing about 12% acetone.
TABLE 3. FIRST ORDER RATE CONSTANTS IN PhTD AT 25C
A. Ethyl Vinyl Ether-PhTDa
Solvent k(sec)-l (D)
Acetone 5.36 x 10-2 20.7
Dimethylformamide 6.02 x 10- 37.6
Dioxane 6.83 x 10- 2.2
Acetonitrile 2.22 x J0-1 37.5
Methylene chloride 6.04 x 10-1 9.1
Benzene 2.85 x 10-1 2.3
B. Divinyl Ether-PhTDa
Dioxane 3.05 x 10-3 2.2
Acetone 1.13 x 10-2 20.7
Methylene chloride 4.20 x 10-1 9.1
al0:1 vinyl ether to PhTD
The infrared spectrum of 49, which is a new 1,3,4-tetrahydrooxa-
diazine ring structure, showed strong absorbances at 2980-2880 cm.-
(saturated carbon hydrogen), 1770 and 1710 cm.-I (double carbonyl) and
strong bands in the 1200-1000 cm.1 region due to the acetal linkage.
The nmr spectrum (Fig. 6) gave resonance signals at 67.40 multiplee,
aromatic hydrogens, 5), 65.02 (quartet, hydrogen on acetal carbon, 1),
63.73 multiplee, methylene hydrogens, 4), 6.1.85 (two equivalent
singlets, nonequivalent methyl hydrogen from incorporated acetone, 6)
and 61.25 (triplet, methyl hydrogens from ether, 3). The elemental
analysis agreed with the calculated value.
8.0 7.0 6.0 5.0 4.0 3.0 2.0. 1.0
Figure 6. Nmr spectrum of 3-oxa-2,2-dimethyl-4-ethoxv-8-phenyl-l,6,8-triazabicyclo[4.3.0]nona-7,9-dione.
In previous studies42,43,44 of the 1,4-dipoles, many different
reactive dipolarophiles have been utilized to intercept the dipolar
intermediates, i.e., isocyanates, acetylene dicarboxylate esters and
ketenes, among many others. However, apparently only one case has
been reported of a ketone performing this function and this involved
the reaction of perhaloacetones with cyanamides to form 1,3,5-oxa-
CH3 CF CF3 01 /CH3
1N-CEN + = 0 --- CF3 C-NH
CH CF3 N3 3
3N-CN possible intermediate CF3 CC3
CH3 C_ NCH3 + CH
CH CH CH N 3
3 3 3
To investigate the scope of this novel interception reaction,
both the vinyl ether and the ketone solvent were varied. When isobutyl
vinyl ether was substituted for ethyl vinyl ether 3-oxa-2,2-dimethyl-
4-isobutoxy-8-phenyl-l,6,8-triazabicyclo[4.3.0]nona-7,9-dione, 50, was
formed in 47% yield. The infrared spectrum was practically identical
to that of 49 with only minor differences noted in the C-H stretching
frequencies and in the fingerprint region of the spectrum. The nmr
spectrum was similar to 49 except a doublet for the methyl groups of
the isobutyl group was present at 60.95 (6), and the multiplet due
to the methylene protons was less complex because of the isobuty]
H2CH(CH3)2 OCH CH3 CHiCH
9 1 23 0 2
C CH--0 3H-
H2 CH CH CI2 C 3
Nh'N o )N-N 3
0 N =O 0= N 0 O < 00
50 51 52
Changing the ketone solvent -from acetone to cyclohexanone
resulted in formation of 3-oxa-2-spirocyclohexyl-4-ethoxy-8-phenyl-
1,6,8-triazabicyclo[4.3.0]nona-7,9-dione 51 in 12% yield. Apparently
the low yield resulted from the more complicated procedure for product
isolation. The infrared spectrum appeared practically identical to
49 and 50. The nmr spectrum differed from that of 49 in that the
nonequivalent methyl resonances at 61.70 and 61.85 were replaced by
a broad multiple resonance from 61.50-62.80 (10) from the cyclohexyl
Divinyl ether, PhTD and acetone resulted in a small amount of the
tetrahydrooxadiazine 52 and a trace of 1,2-diazetidine 31b. Both were
identified by nmr with the oxadiazine spectrum resembling 49, 50 and 51,
while the 1,2-diazetidine resonances were identical to those of the
room temperature product of DVE-PhTD in methylene chloride 31b.
The mass spectral fragmentation patterns offered strong corrobora-
tive evidence for the proposed new 1,3,4-tetrahydrooxadiazine ring
structures. The fragmentation patterns were very similar to known
cyclic acetal fragmentations.6 They were different in that the
molecular ion appeared. This could be due to the unique structure
in which one ether linkage was inside the ring and the other was
outside. Schemp I shows the fragmentation pattern for 49. The
molecular ion m/e 305 (1%) is believed to fragment by two pathways,
a and b.
b CH -0
\ C 3 a
C- .CH C / N 3 0
23 0 0 -CH3CC
O-- ----- 0
= -0 N 0
CH =0CH2 CH3
O=< )=0 m/c247
N base peak
-CH= CH2 /0
-II0 N N-
Pathway "a" involves loss of acetone to form m/e 247, the base
ion. This is followed by the loss of ethylene to m/e 219 (8%) which
then appears to rearrange and lose a hydrogen atom to m/e 218 (22%).
Ion m/e 190 (59%) subsequently follows by loss of carbon monoxide.
Pathway "b" involves loss of the ethoxy radical to form ion m/e 260
(7%), a general cracking course for both acetals46 and glucosides.47
Comparison of the fragmentation of 50 and 51 (Schemes II and III
GCR CH(CH )
CH3 H>0 C
/c H \ + 32
CH b CH2 /'-CH a
3 -- -- N ----
-0 -6CH2oCH(CH3)2 o 0- \ -C3 31
CH, 2 H
2 -H -CO NN
m/e275 -CHCH(CH )
2\ H 3 2
N N -0
respectively) with that of 49 shows similar ions resulting from the
same double fragmentation pathway.
From these results (summarized in Table 4) it appears that the
reactivity of the 1,4-dipole is quite sensitive to stabilization of
its positive center. The high yields of oxadiazine when the carbonium
ion stabilizing substituents, ethoxy and isobutoxy, are present and
the low yield with vinyloxy are believed to mean that the intermediate
has more pronounced dipolar character when its positive center is
stabilized and hence can add across the weak dipolarophilic alkyl
ketones. This is the same reasoning used to explain the earlier
copolymerization results. In agreement with the above observations
is the PhTD-indene30 reaction in which no oxadiazine was reported even
though the reaction was performed in acetone.
TABLE 4. SUIIARY OF INTERCEPTION REACTION RESULTS
Vinyl Ether Ketone % Yield M.P.
CH2=CH-O-CH2CH3 Acetone 42% 149-151C
CH2=CH-OCH2CH3 Cyclohexanone 12%a 170-171C
CH2=CH-O-CH2-C-C3 Acetone 47% 125-126'C
CH2=CH-0-CH=CH2 Acetone trace
(a) Yield lower possibly because of difficulty in work up.
(b) Identified from nmr. Also some 1,2-diazetidine formed.
By employing this unique property of these 1,4-dipoles, it was
felt that information on the degradation of the EVE-PhTD copolymer
could be obtained. When a high molecular weight sample of this
copolymer was refluxed in acetone, the usual decrease in molecular
weight was observed along with a trace amount of oxadiazine 49. This
was believed to result from degradation back to the initial 1,4-dipole
which in turn was trapped by the acetone. This is interpreted to mean
that, at least in part, the degradation is a reverse reaction of the
original polymer formation.
Soluble salts are generally known to exhibit significant effects
in either anionic or cationic polymerizations. It was felt that the
ions of a soluble salt would effectively "tie up" some of the "living
ends" of the propagating dipole and prevent their intermolecular cou-
pling and hence lower the molecular weight. Mhen the EVE-PhTD copoly-
merization was effected in tetrahydrofuran solution in the presence of
10 mole % of sodium tetraphenylboron, a significant change in the mol-
ecular weight was observed. A control in tetrahydrofuran yielded an
average molecular weight of. 2040 versus 940 for the sample containing
Another mechanistic pathway which merits consideration is that of
a ring opening polymerization resulting from attack of the zwitterionic
intermediate on an unopened 1,2-diazetidine. Although the data do
not allow this mechanism to be disregarded, especially in the DVE case,
the observation of molecular weight increase with time for both the DVE
and EVE cases is not consistent with such a propagation mechanism.
This evidence is particularly persuading for the EVE-PhTD copolymeri-
zations because of two observations. The first observation resulted
from an experiment conducted in an nmr tube. This experiment showed
that when the reactants were mixed and inserted directly into the nmr
cavity only copolymer was observed. The second observation was that
even though all the PhTD was consumed after 10 minutes in the PhTD-
EVE copolymerization, the molecular weight increased over a 0.5 hour
time period (Fig. 5). Since no adduct was present as indicated by the
first experiment, the molecular weight growth is attributed to dipolar
coupling. Wilson and Beaman48 have recently observed a system in which
the molecular weight increased after an addition polymerization was
complete and attributed the phenomenon to a coupling of residual poly-
Another polymerization in which some dipolar coupling was postu-
lated was the catalytically initiated polymerization of 1,1,3,3-tetra-
The vinyl ether-PhTD copolymerizations are believed to be the
first reported cases of a dipolar coupling being the sole propagation
Mechanism of termination
In all of the copolymerizations the highest molecular weights ob-
tained were only in the vicinity of 3000 (Tables 1 and 2). These low
molecular weights are thought to be a result of facile termination re-
actions. As was stated earlier, ionic polymerizations are readily sub-
ject to termination reactions by impurities such as water. At least
two other termination reactions could occur. The first would be an
intramolecular dipolar coupling to a macrocyclic structure and the
second is a type of disproportionation to an inactive species as
shown below. Another possible consideration is that the dipoles are
S TN-T-N H,- N v
2 + N
0 0= 0 ---)- CHZCH 0( .O0
R N 0
so stabilized to result in a "living" structure. Unfortunately no
experimental evidence was obtained supporting any of the possible
B. Copolymerization of 4-Phenyl-l,2,4-triazoline-3,5--dione with-
Vinyl acetate and divinyl carbonate
To broaden the scope of the copolymerization and reaction character-
istics of PhTD, its reactivity was investigated with several other
olefinic systems. Vinyl acetate (VAC) and Divinyl carbonate (DVC)
were chosen for investigation because these monomers were expected to
be somewhat less reactive than the vinyl ethers.
VAC was observed to react with PhTD at room temperature, but at
a considerably slower rate than the vinyl ethers. The infrared spectrum
of the product exhibited a medium intensity 1610 cm.- band. The nmr
spectrum gave broach resonances at 67.48, 64.32 and 62.16. Superimposed
over these resonances were sharp singlets. A reaction conducted at
60C in a sealed tube yielded as the major product an adduct which was
assigned structure 55, (foyl hyl)-2-acty-2,4-tiazoline-3,5-
assigned structure 53, l-Cfo-a-mylmeichyl)-2-acetyl-l,2 ,4-triazoline-3,5-
TABLE 5. COPOLYMERIZATIONS AND REACTIONS OF PhTD AND OTHER MONOMERS
b0.5 wt.%AIBN added
CCalculated from GPC
CH 2 Cl2
60 24 73.5
-- PhTD oligomeric species
-- PhTD oligomeric species
-- Possibly some copolymer
from elemental analysis
PhTD oligomeric species
PhTD oligomeric species
No sulfur present from
,000c 1:1 copolymer
-- 1:1 copolymer
-- Copolymer and 55
-- Not identified
370 Mostly 55
-- Mostly 55
-- 2:1 copolymer
1240 2:1 copolymer
done. This compound apparently arose from an internal trap of the
intermediate 1,4-dipole 54.
CIE 0 0 C-H
C 2 / CH3 -C C-CH
0:= ( 0 0M =0
The infrared spectrum of 55 showed two weak bands characteristic
of an aldehyde carbon hydrogen stretch at 2860 cm.-1 and 2750 cm.-1 A
double carbonyl absorbance at 1800 cm.- and 1730 cm.- was also char-
acteristic of the urazole type structure. The nmr spectrum (Fig. 7)
gave resonance signals at 69.56 (singlet, aldehyde hydrogen, 1), 67.50
(singlet, aromatic hydrogens, 5), 64.80 (singlet, methylene hydrogens,
2) and 62.60 (singlet, methyl hydrogens, 3). The mass spectrum yielded
the correct parent peak at m/e 261 and the elemental analysis agreed
with the 1:1 structure.
The broad nmr resonances and the 1610 cm.- infrared band were
probably indicative of the formation of some copolymer in the room
temperature reaction. The small amount of polymeric material is be-
lieved to be indicative of the low activation energy of the intramolec-
ular six menbered transition state rearrangement.
An attempt to trap the initial dipole before the intramolecular
reaction occurred was made utilizing a large excess of phenyl isocy-
anate as the dipolarophile. Only 55 was isolated. The reaction was
also performed in acetone so that the initial dipole could be trapped
~i --------------------- i --- ---i-- ------
9.0 8.0 7.0 6.0 5.0 4.0 3.0
Figure 7. Nmr spectrum of 1-(formylmethyl)-2-acetyl- ,2,4-triazoline-3,5-dione.
to form an oxadiazine as was accomplished in the vinyl ethers. The ma-
jority of the reaction was 55. A few milligrams of material were iso-
lated which gave an nmr spectrum with resonance signals at 66.30 (mul-
tiplet), 63.75 doublett) and 61.72 (singlet). The spectrum was
complicated by the presence of 55. The above resonances appeared in
the ratio of 1:2:6. This could be evidence for the formation of a
small amount of the oxadiazine 56.
DVC was found to react slowly at room temperature and at 60C to
give a material shown by elemental analysis to be a 2:1 structure. A
strong 1610 cm.-1 band was observed in the infrared spectrum and the
nmr spectrum gave broad resonances at 67.38 (singlet, broad, 6) and
65.31-63.31 multiplee, broad, 1). A molecular weight of 1220 was ob-
tained for the 6000 polymerization. No observations were indicative
of the occurrence of a double internal trapping reaction analogous to
the vinyl acetate reaction.
Divinyl sulfone and acrylonitrile
The normally electron poor olefins divinyl sulfone (DVS02) and
acrylonitrile CAN) were chosen for investigation to see if they would
act as electron donors with PhTD.
DVS02 appeared to give some reaction since more solid was obtained
in an attempted copolymerization at 600C than was obtained with a con-
trol of PhTD. No sulfone absorbances were noted in the infrared spec-
trum and an elemental analysis only yielded a trace of sulfur.
AN showed no indication of reaction at room temperature or at
600C for six days. At 60C, with a free radical initiator present,
some product appeared to result. An analysis of the infrared spectrum
showed a new medium intensity 1610 cm.- band. An elemental analysis
yielded a higher nitrogen content than would result from a reaction
only involving PhTD. No structure assignments were feasible.
N-vinyl carbazole CNVC) was observed to spontaneously copolymerize
at 25C to yield an 86% yield of a 1:1 copolymer. The composition was
established by elemental analysis. Only a weak-medium 1610 cm.-I ab-
sorbance was observed in the infrared spectrum. The nmr spectrum yield-
ed two large broad resonances centered at 68.50 and 65.80. Because of
the weak 1610 cm.-1 absorbance the polymer structure was thought to
N-N---- CH --CH
Insolubility of the polymer in acetone precluded the usual vapor
pressure osmometer molecular weight determination. However, calculation
of the number average molecular weight from gel permeation chromatog-
raphy, calibrated with polystyrene, gave a value of 21,000. The
weight average was calculated to be 54,000 and the molecular weight
distribution (M /M ) was 2.57. In making the calculations a very
dangerous assumption was necessary. It was assumed that polystyrene
and the copolymer had similar coil dimensions in the solvent used
for the determination. Although this assumption appeared to be valid
in the DVE-PhTD copolymer, it was not necessarily valid in this case.
The molecular weight obtained in this system is significantly higher
than in the vinyl ether copolymers. No reasons are apparent for the
change in structure and the change in molecular weight of the NVC
When the polymerization was conducted in acetone, a similar
polymeric structure was obtained. Only a minute amount, if any, of
the corresponding oxadiazine was observed. An nmr spectrum from
a small amount of apparently non-polymeric material showed multiplets
at 64.50 from methylene hydrogens and 62.20 from methyl hydrogens.
Not enough material was isolated to identify and characterize this
product. The failure of acetone to trap the probable intermediate
dipole is probably due to the large steric effect of the vinyl
Stille and Aoki 50 have initiated the homopolymerization of
electron donors by addition of catalytic amounts of strong electron
acceptors. The homopolymerizations are proposed to be initiated by
an excited ionic s'ate of a charge transfer complex. Since NVC is
known to readily homopolymerizc, it was thought that the addition of
a small amount of PhTD to a solution of NVC could possibly trigger a
cationic homopolymerization of the NVC in which the initiation would
be effected by the positive end of the dipole. An attempted polymer-
ization was unsuccessful.
C. Reactions and Attempted Homopolymerizations of 4-Phenyl-1,2,4-
4-substituted-l,2,4-triazoline-3,5-diones have the potential to
yield polymers possessing a nitrogen backbone if an initiation could
be effected through the nitrogen to nitrogen double bond. This promise
made the investigation of the behavior of the triazoline diones in
the presence of radical and ionic initiating species an attractive
PhTD has been described51 as decomposing instantly in basic
solution, while in the presence of acid, water or alcohol the rate
of decomposition was noted to be slower. Even in the presence of
light or in solution at room temperature some decomposition was
reported. No one has studied the decomposition products and their
mechanisms of formation.
The thermal decomposition of PhTD at 60OC in a scaled tube was
studied to determine if there was any appreciable decomposition that
could compete with the copolynerizations. During the course of the
reaction the color remained red and a few milligrams of a white
crystalline solid formed in the bottom of the tube. Precipitation
of the red solution yielded a small amount of tan-w:hite solid 59.
The white cry-talline solid did not melt when the temperature
was raised to 300"C. The infrared spectrum gave no saturated carbon-
hydrogen stretching absorbances and gave a double carbonyl at 1785 cm.
and 1755 cm.-1 The mass spectrum yielded a parent peak at m/e 322
and 1755 cm. The mass spectrum yielded a parent peak at ale 322
and a base peak at m/e 119 (phenyl isocyanate ion). Structure 58,
3,7-diphenyl-1,5-diazabicyclo[3.3.0]octa-2,4,6,8-tetraone, was assigned
from this data. An elemental analysis perfectly agreed with the
Two reports of 58 were found in the literature. Snyder52 reported
the formation of a bright orange solid, melting point 203-2040C,
during a refluxing reaction of PhTD. In this case PhTD was generated
by oxidation with t-butyl hypochlorite in acetone. The only structure
proof presented was infrared data which included a double carbonyl
and an aromatic absorbance. Koch and Fahr reported the formation
of 58 in a thermal reaction of PhTD in an apolar solvent. No structure
proof or physical characteristics were presented.
The tan-white solid 59 isolated by precipitation into hexane
had 3500 cm.-1 and 1715 cm.-1 absorbances in the infrared spectrum.
The nmr spectrum gave broad resonances at 67.50 (broad multiple, 6)
and 64.48 (broad multiple, 1). The number average molecular weight
from vapor pressure osmometry was found to be 600. The elemental
analysis did not correspond to any regular structure resulting from
a homopolymerization of PhTD.
When PhTD was heated in a sealed tube in methylene chloride
solution in the presence. of azobisisobutyronitrile, results identical
to reaction without the free radical initiator were obtained. An
attempt to initiate a free radical homopolymerization of PhTD by
photolysis of benzoyl peroxide at -45C only resulted in a small
amount of the oligomeric PhTD decomposition product 59.
In attempting to explain the mechanistic pathway for the thermal
decompositiolof PhTD to 59, the first reaction is probably the forma-
tion of phenyl isocyanate. Phenyl isocyanate could then in turn
0 0 Ph-N C= 0
react with water to form aniline. When a catalytic amount of aniline
H 0 -CO
Ph-N=C=0 + H20 -> [Ph-N-c-0-H] -- PhNH2
was added to a methylene chloride solution of PhTD, some of the
oligomeric decomposition product 59 resulted.
Another related experiment was to see what products were formed
in the decomposition of PhTD by water, since PhTD had already been
shown to readily decompose in water.5 When PhTD, in dioxane solution,
was added to water a vigorous reaction insued and a large amount of
oligoneric product 59 was obtained. A small amount of N,N'-diphenyl
urea was also obtained. One of the products from the reaction is
most assuredly phenyl isocyanate which then goes to aniline and then
reacts with more phenyl isocyanate to yield the diphenyl urea.
Since aniline served as a catalyst for for:ration of the oligomeric
product 59, an investigation using other nucleophilic species was
initiated in hopes of effecting the formation of high polymer.
Triethylamine was used as the initiator and a similar product 59
was obtained. Surprisingly approximately 10% of 58 was
The strong nucleophilic reagent, sodium cyanide in dimethylforma-
mide was then used as the initiator and a large yield 70-80 per cent
of the tetraone 58 was obtained. A small amount of 59 was also obtained.
The rationalization of the formation of 58 from attack of a
nucleophile on PhTD presents an interesting mechanistic problem that
was not covered in the scope of this research. However, the following
explanation is put forth as a possible mechanistic pathway of the
Kealy54 has observed the formation of 61 from the decomposition
of diazoquinone 60. He did not propose a mechanism for the formation
of 61, but only referred to earlier work on the decomposition of azo
compounds by radical pathways. For example, Leffler and Bond5 have
studied the radical decomposition of dibenzoyl diimide 62 where 63 is
one radical intermediate.
0 0 0
t 2 PhC.
0 0 0
For the diazoquinone, a probable diradical intermediate would
be 64 which could couple with an intact molecule of diazoquinone 60
to yield 61.
If such a diradical. intermediate 65 was involved in the formation
of the tetraone 58 from a nucleophilic catalyzed reaction of PhTD
an unusual carbenoid mechanism could be involved as shown below. No
proof has been gathered in support of this mechanism, although in the
studies conducted on the radical decomposition of 6255 much higher
rates of decomposition were observed in strongly nucleophilic solvents
like aniline. The reason for this could be a nucleophilic catalyzed
reaction as was discussed for the formation of 58.
TABLE 6. SUMMARY OF CATALYZED PhTD
Most electron poor comonomers
CH2Cl2, 600C, 70 hrs., sealed tube
CH2C12, 60C, 24 hrs., sealed tube
water, dioxane, 25'C, 1 hr.
aniline, CH2C12, 25C
triethyl amine, CH2C12, 25'C, 1 hr.
sodium cyanide, dimethylformamide
2.5 hrs., 25"C
(n = 1000)
(27%), 58 (1%)
(10.4%), 58 (trace)
(22.0%) (Mn = 470)
(24.5%), 58 (9.5%)
A summary of all the reactions and decompositions of PhTD is
presented in Table 6.
While this research was being conducted, Stickler and Pirkle12
reported the first synthesis of a nitrogen backboned polymer. Their
o=< )o +- -o J \ -
N I N N + N N
N N +NI
Ph Ph Ph
success resulted from a visible light irradiation of 4-n-butyl-l,2,4-
D. Diels Alder Polymers
Synthesis of 4,4'-(4,4'-diphenylmethylene)-bis-l,2,4-triazoline-3,5-
Since the diisocyanate 66, 4,4'-diphenylmethane diisocyanate,was
readily available from commercial sources, the bis triazoline dione
selected as a target for synthesis was 69, 4,4'-(4,4'-diphenylmethyleno)-
bis-1,2,4-triazoline-3,5-dione. The synthesis of the diurethane
precursor 67 was accomplished in a near quantitative yield by the
addition of two moles of ethyl carbazate to 66. The infrared .spectrum
yielded strong N-H stretches at 3380 cm. and 3305 cm.-. Strong
carbonyl absorbances vere observed at 1735 cm.- and 1685 cm.- The
nmr spectrum gave resonances at 68.77 (singlet, hydrogens on nitrogen,
2), 67.91 (singlet, hydrogens on nitrogen, 2), 67.20 (A2B2 quartet,
aromatic hydrogens, 8), 64.05 (quartet, methylene hydrogens, 4),
63.79 (singlet, methylene hydrogens, 2) and 61.19 (triplet, methyl hy-
drogens, 6.). The product gave the correct elemental analysis.
CH -N=C-0 24- 2H2NNHCOCH2CH3- CHH0 --NHCNHNHCOCII2CH3
67 was quantitatively cyclized to the bis urazole 68 with 4M po-
tassium hydroxide. The infrared spectrum yielded a broad band front
67 4M1 KOH CH2 H
11 H 1 2
3320 cm. to 2760 cm., a characteristic noted in the simple urazoles,
and gave a double carbonyl absorbance at 1765 cm.- and 1680 cm.- The
nmr spectrum gave resonance signals at 67.80-67.20 (broad with a sharp
spike at 67.38, aromatic hydrogens and hydrogens on nitrogen, 12) and
64.03 (singlet, methylene hydrogens, 2). The elemental analysis
agreed with the calculated value.
The oxidation of 68 to the bis triazoline dione 69 was accomplished
in approximately 80% yield using dinitrogen tetroxide in the same man-
ner as employed by Stickler and Pirklel2 in the synthesis o[ the simple
triazoline diones. The infrared spectrum showed an absence of any N-H
stretching frequencies and yielded the characteristic double carbonyl
at 1790 cm.-1 and 1760 cm.-1 The nmr spectrum yielded, surprisingly,
only two singlets, on, at 67.48 (aromatic hydrogens, 8) and 64.18
(methylene hydrogens, 2).
69 reacted readily -with reactive dienes. The reaction with 2,3-
dimethyi-l,3-butadiene yielded 70. 70 was identified by its nmr spec-
II 1 0 Ci
7 CH CH 3
C2 NC i 3 -3 CH-2 N
2 0 0 3
trum which showed resonance signals at 67.33 multiplee, broad, aromatic
hydrogens, 8), 64.00 (broad singlet, allylic next to nitrogen and benzy-
lic hydrogens, 10) and 61.75 (broad singlet, methyl allylic hydrogens,
Diels Alder polymerization of styrene and 69
Cookson and coworkers51 have investigated the Diels Alder reactions
of triazoline diones and found that PhTD spontaneously reacted with
styrene to yield the double Diels Alder adduct 71. Maleic anhydride56
and ethyl azobisformate57 have been shown to react with styrene in a
Diels Alder fashion, but the final products in both cases were 72 and
73. These products resulted from an ene reaction after the first Diels
Alder reaction. Cookson's results meant that styrene, potentially,
would be an attractive bis diene to investigate with 69.
When styrene was reacted with 69 at room temperature in dimethyl-
formamnrde solution, the deep red initial color faded to a much lighter
red after two hours. After seven and one half hours, the red color,
characteristic of unreacted 69, persisted. The solution was divided
and the first half was precipitated into ether to yield a tannish,
highly electrostatic powder. The second half was in turn divided into
two portions and a discharge of the red color was noted after a few
minutes. Precipitation into ether yielded an identical polymer (in-
H H -C02CH2CH3
0 dO2 CH2H3
PhNN I I
S -0 ,
NN & N
71 0 0 75 0/ NPh
frared) as the portion precipitated while still red. The last portion
was heated to reflux and immediately it began to darken. It was pre-
cipitated into ether and a much darker solid resulted.
The repeat structure of the polymer was assigned to 74, which re-
sults from a Dicls Alder adduct which rearomatizes via an ene reaction
as 72 and 73. This is contrary to what was expected a priori. The
assignment was made from the infrared and nmr data and comparison to
the model compound 75.
When PhTD and styrene were reacted in methylene chloride at room
temperature 75 was formed rather than the reported Diels Alder adduct
71. The infrared spectrum gave an N-H stretch at 3280 cm. and the
usual double carbonyls at 1765 cm. and 1720 cm. The nmr spectrum
(Fig. 8) gave resonance signals at 68.38 doublett, hydrogen on nitrogen,
1), 67.35 and 67.38 doublett and multiple respectively, aromatic
hydrogens, 4), 65.70 (triplet, benzyl hydrogen, adjacent to nitrogen,
1), and 64.22 multiplee, methylene hydrogens, 2). The elemental
analysis agreed with the 2:1 structure.
The polymer gave the double carbonyl in the infrared spectrum at
1780 cm. and 1715 cm.-. The N-H stretch was very weak. The nmr
spectrum (Fig. 9) showed resonances at 68.40 doublett, broad, hydrogen
on nitrogen, 1), 67.37 (singlet, broad, aromatic hydrogens, 12), 65.67
(singlet, broad, benzyl hydrogens adjacent to heteroatom, 1) and
64.50-63.80 (muitiplet, broad, two sets of methylene hydrogens, 4).
The polymer wis soluble in both dimethylformamide and dimethyl-
sulfoxide. The thermal decomposition was recorded using a Differential
Scanning Calorimeter and.was found to start at 3070C. An intrinsic
viscosity determined in direthylformamide at 280C was 0.12. Calculation
I I I 1
8.0 7.0 6.0 5.0 4.0 3.0 2.0
Figure 8. Nmr spectrum of the Diels Alder, ehe adduct of 4-phenyl-1,2,4-triazoline-3,5-dione and styrene.
I I I I I I 1
9.0 8.0 7.0 6.0 5.0 4.0 3.0 2.0
PPM f s a
Figure 9. Nmr spectrum of the Diels Alder, ene polymer from styrene and bis triazoline diane.
from gel permeation chromatography showed a molecular weight distri-
bution of 3.34 with the weight average molecular weight, An = 36,000.
These values were calculated from a calibration with a polystyrene
standard, and again the assumption is made that the coil sizes are
similar for the two polymers.
An attempted polymerization of 69 using sodium cyanide in
dimethylformamide yielded a light brown polymeric material. The
synthesis of a polymer with structure 76 was the object of this exper-
iment. The infrared spectrum of this material was not similar to
the monomeric tetraone 58 so apparently the desired decomposition and
coupling reaction did not take place.
Triple Strand Polymer Model Compound Studies
A. Attempted Synthesis of Tetracyclo [220.127.116.11 '9.1 48] Dodecane and
The synthetic pathway for the attempted synthesis of the cage
structure 18 is shown below in Scheme IV.
CH20R OH20R (1) ORCH O21s
JCH OR Y CHOR (2) C- COAs
2 0 -- Q2 2
77a,b 83a,b 24
R=b= -H (1) catalytic hydrogenation
Scheme V is a proposed synthetic pathway to 25 which is the precur-
sor to 26, a simplified cage structure re-lated to 18.
IC CHOR 0 0
< >2 O CH2OR (1) H O20Ms
+-2 (2) K
89 0 8a,
R=a= -OCH3 25
R=- (1) catalytic hydrogenation
25 -- -------->- 26
Synthesis of dienes
The first step in Scheme IV required either 2,3-di(acetoxymethyl)-
1,3-butadiene or 2,3-di-(hydroxymethyl)-1,3-butadiene, 77a and 77b.
Both of these dienes had previously been synthesized by Bailey and
Sorenson.58 These authors prepared 77a by pyrolysis of 2,3-di-(acetoxy-
methyl)-l,4-diacetoxybutane, 81, at 480C. Although the reported
CH CH Br CH OCCH
NB0LS /- S2 A' / >'2 77a3
2 C\ fCHJ\ r CHCH2lC 0.1 mm.
CH!3 2 2 3
82 83 84
yield was 84%, this was based on the recovery of unreacted starting
material and the formation of a mono elimination product. Actually
the conversion to 77a was only 16%. It was felt that the method of
Butler and Ottenbrite29 used in the preparation of 2,3-di-(bromo-
methyl)-l,3-butadione would result in a more efficient preparation
The first precursor to the desired dines, 2,5-dihydro-3,4-
dimethylthiophene-l,l-dioxide or commonly referred to as dimethyl-
(CH3COCH2) 2CHC CH2OCCH3) 2
CH3 / I Br //- CHCHOCCH3
02S I NBS 02S As ICH3 -O- 02S- 79a
85 86 87
butadiene sulfone, 82, was prepared by the method of Frank and Seven59
in 90% yield. This was then brominated2 with N-bronosuccinimide in
38% yield to 83. Treatment of this dibromide with silver acetate in
acetonitrile resulted in a 73% yield (after recrystallization) of the
diacetoxy sulfone 84. The elemental analysis of this compound agreed
with the calculated and the infrared spectrum showed a strong carbonyl
absorbance at 1720 cm.- for the ester group. The nmr spectrum yielded
resonances consistent with the structure and they appeared at 64.75
(singlet, broadened, allylic hydrogens adjacent to acetoxy, 4), 63.86
(singlet, broadened, allylic hydrogens adjacent to sulfone, 4), and
62.08 (singlet, methyl protons, 6).
Pyrolysis of 84 was accomplished smoothly at 200"C and 0.1 mm
pressure in 56.6% yield using a Hoskins furnace as the external heat
supply. The white solid diene was purified by recrystallization and
yielded the reported melting point.58 Saponification5 to 77b was
accomplished in 90% yield.
By using the same general procedure as used in the preparation of
84, the sulfone precursor, 87, to the acetoxymethyl diene 79a was
prepared in 74% yield. The bronmcmethyl sulfone 86 was prepared by
the method of Krug and Yen60 by bromination of isoprene sulfone 85
with N-bromosuccinimide. When 86 was reacted with silver acetate in
acetonitrile 87 was formed as a yellow-brown oil.. It showed a 1735 cm.
ester carbonyl in the infrared spectrum and yielded the correct elemen-
tal analysis. The nmr spectrum exhibited resonance signals at 66.07
(singlet, broad, vinylic hydrogen, 1), 64.69 (singlet, broad, allylic
hydrogen next to acetoxy, 2), 63.81 (singlet, broad, allylic hydrogen
next to sulfone, 4) and 62.10 (singlet, methyl hydrogens, 3).
Pyrolysis of this oil was accomplished at 2000C at 0.5 mm. and an
84% yield of 79a was obtained. This was purified by distillation and
the infrared characteristics were identical to those reported by
Bailey and coworkers6 who had prepared 79a from the pyrolysis of 88.
CH COACH CH CHCH o0CH
CH 2 CCH3
Attempted Diels Alder reactions with p-benzoquinone
When the diacetoxymethyl diene 77a and the dihydroxymethyl diene
77b were reacted with p-benzoquinone, 89, the normal Diels Alder adducts
were not isolated. Reaction of 77a with 89 in acetic anhydride resulted
in the formation of the dehydrogenated adduct 90. The infrared spectrum
of this reddish brown solid exhibited carbonyl absorbances at 1725 cm.-
(ester) and 1655 cm.- (quinone). The nmr spectrum yielded resonances
at 66.75 (singlet, quinodial hydrogens, 2), 64.78 (singlet, allylic hy-
drogens adjacent to acetoxy, 4), 63.23 (singlet, doubly allylic hydro-
gens, 4) and 62.12 (singlet, methyl hydrogens, 6). A correct elemental
analysis was obtained. The same reaction carried out in benzene
yielded a crude oil possessing the same nmr resonances as 90 with
additional resonances at 66.77, 66.75, 64.64 and .62.43. The 63.23
resonance was broadened considerably. This was apparently a mixture of
unreacted quinone (66.77), 90 and the desired Diels Alder adduct 78a.
On addition of ethanol to the oil 90 was formed as a precipitate and no
indication of 78a was found in a spectroscopic analysis.
The reaction of 77a and 89 was then performed in refluxing chloro-
form and the progress of the reaction was followed by nmr spectroscopy.
Disappearance of resonances attributable to starting quinone and diene
was observed and the appearance of signals at 66.68 (singlet, quinodial
hydrogens, 2), 64.65 (singlet, allylic hydrogens adjacent to acetoxy,
4), 63.27 multiplee, broad, bridgehead hydrogens, 2) and 62.44 (multi-
plet, broad, allylic hydrogen, 4) was noted. After 135 hours a con-
version of 80% was calculated from the nmr spectrum. Removal of the
chloroform yielded a yellow oil. This was chromatographed on silica
gel and a white solid resulted that showed a strong 3400 cm.- band
and a disappearance of the 1650 cm.- in the infrared spectrum. This
indicated that 78a had rearranged to the hydroquinone derivative 91.
The nmr spectrum gave resonances at 66.64 (singlet, aromatic hydrogens,
2), 64.48 (singlet, allylic hydrogens adjacent to acetoxy, 4), 63.47
(singlet, doubly allylic hydrogens, 4) and 62.88 (broad, hydroxyl
0 CHR CH2R
6 12 IR CE 2 R 90
2 CH 2 CUR
89 77a 0
R= -0CCl3 C 9 CH2R
The reaction of the dihydroxy diene 77b with 89 failed to pro-
duce the desired Diels Alder adduct as an isolable material also. In
refluxing chloroform a brownish solid formed whose nmr spectrum gave no
resonances that could be assigned to a Diels Alder adduct.
When the reaction was performed in tetrahydrofuran, a brownish-
yellow oil was obtained which was initially chloroform soluble, but
precipitated within a short time as a white solid. The absence of a
quinone carbonyl in the infrared spectrum and the absence of bridge-
head hydrogen resonances in the nmr spectrum was indicative that the
desired Diels Alder adduct 77b was not present. The presence of a
strong 1710 cm. absorbance in the infrared spectrum quite possibly may
be due to some double Diels Alder adduct 92. There was, however, no
11OH9 C .!y< CH OH CH OH
2 H 2 O 2
92 H 93
other indication of this. More likely, since the bridgehead protons did
not appear in the nmr spectrum, 93 was formed.
The Diels Alder adducts of 89 and both 2-bromomethyl-1,3-butadiene,
94, and 79a were not isolable under the reaction conditions employed.
94, the bromine substituted diene, reacted with 89 in acetic
anhydride to give a 60% yield of the aromatized and acetylated product
95. All spectral characteristics were consistent with the structure
CH Br I 3 CH2Br
89 2 acetic
4 1 95
proposed. The infrared spectrum gave a strong 1750 cm. carbonyl from
the ester functions. The nmr spectrum had resonances at 67.00 (singlet,
aromatic hydrogens, 2), 66.05 multiplee, vinylic hydrogen, 1), 64.08
(singlet, broad, allylic hydrogens adjacent to bromine, 2), 63.30 (sing-
let, broad, allylic hydrogens, 4) and centered at 62.30 (two singlets,
nonequivalent methyl groups, 3 each). The mass spectrum gave the molec-
ular ion and the molecular ion +2 in the correct percentage for one
bromine atom, m/e 339 and m/e 341. The base peak m/e 176 corresponded
to the loss of both acetoxy groups and the bromine atom.
In the reaction of the acetoxymethyl diene, 79a, with 89 in benzene
at room temperature, quinhydrone,62 the molecular complex of hydroquinone
and quinone, was formed as an insoluble precipitate. It was identified
by its melting point, nmr spectrum and mixed melting point with an
authentic sample. On evaporation of the benzene a dack oil was formed,
which on extraction with hot petroleum ether gave a bright yellow solid
which quickly darkened. Column chromatography on silica gel resulted
in a yellow oil which quickly darkened also. A thin layer chromatogram
yielded two spots. The infrared spectrum gave a strong 3500 cm.1
absorbance indicative of a hydroxyl group. The nmr spectrum was very
complex and from the methyl signal at 62.10, three spikes were apparent
indicating a mixture of three compounds. The bridgehead protons for
the Diels Alder adduct appeared to be present as a multiple at 63.25.
The downfield region centered at 66.80 appeared as a double. From
these data it is believed that the initially formed Diels Alder adduct
80a aromatized to 96. From the color and the apparently three differ-
ent methyl groups, some dehydrogenated material 97 was possibly present
also. All attempts to separate the mixture were unsuccessful. When
the reaction was carried out in ethanol a similar oily solid was
obtained that yielded an infrared spectrum identical to the previous
reaction in benzene.
CH2CH OC C0&
11- 2 3 2 3 C CN
IO II 0C CN
0 0 CH CN
H 96 97 3 98
The acetoxydiene 79a readily reacted with tetracyanoethylene to
yield the Diels Alder adduct 98. Spectroscopic data and the elemental
analysis were all consistent with the normal Diels Alder structure.
At the present time, there appears to be no plausible explanation
why the Diels Alder adducts of p-benzoquinone, 89, and the dienes 77a,
77b, 79a and 94 isomerize so facilely to the hydroquinone derivatives
or dehydrogenate to the quinone like structures. Since quinhydrone
appeared as a byproduct in the formation of 90 and 96, 89 probably ahts
as an oxidizing agent in the reactions.
Although these reactions appear to be the first time such reactions
have been observed to occur spontaneously, Ansell and Knights6 have
reported the facile aromatizations of some Diels Alder adducts of
1,1'-acetoxyvinylcyclohexene, 99, and 89. They found that in ethanol-
R= -0CCH3 0
0 0 0
99+ 8 --100. 101 102
99 100 101 102
at room temperature the normal Diels Alder adduct 100, formed in 63%
yield. In refluxing methanol, however, the authors observed the form-
ation of 101 and 102. They offered no reasons for their observations
but did report that most aromatizations of this nature are catalyzed
by acid or by base. A previous synthesis of 10064 revealed that melt-
ing the adduct and then resolidifying resulted in the formation of 101.
2-Cyclohexene-l-one and diazoguinone as dienophiles
As an alternative approach to the synthesis of a precursor
capable of undergoing the cyclization reaction to the cage structures,
it was decided to employ dienophiles whose adducts were incapable of
undergoing the facile aromatization reaction. The two dienophiles
selected for this study were 2-cyclohexene-l-one 103 and diazoquinone
If the normal Diels Alder adduct, 105, of 103 and 79a were to
form then it would not be capable of aromatizing. Hence it was believed
that the reactions shown in Scheme VI could be followed.
The synthesis of 103 was accomplished by bromination of cyclo-
hexanone with N-bromosuccinimide65'66 followed Ly dehydrobromination
with colliding. The bromination reaction was accomplished in a
53% yield, while the dehydrohalogenation gave a 58% yield of the
desired a,3-unsaturated ketone, an improvement over the 42% yield
103 + 79a IH2 OCH3 1) catalytic
separate 2) Ms Cl
j Y -CH20Ms base as
reported by the original investigators.67 An attempted dehydrobromin-
ation using lithium bromide in dimothylformamide68 resulted in a
yellow viscous oil.
From earlier experiments with 103 by Bartlett and Woods, this
dienophile was known to be quite sluggish in the Diels Alder reaction;
therefore forcing reaction conditions were employed.
An autoclave attempt to form the adduct from neat reactants at
2000C resulted in a waxy, gummy, polymeric substance. This material
gave a sharp carbonyl absorbance at 1710 cm.- The nmr spectrum
yielded broad resonances centered at 67.00 (1) and 61.00 (13).
In benzene solution, in a sealed tube at room temperature for
four days,starting materials were recovered. 103 and 79a were then
reacted neat, in the presence of a small amount of hydroquinone, in
a sealed tube at 190C for three days. As in the autoclave reaction,
a polymeric substance was isolated. It had the identical spectral
characteristics of the former polymer.
Aluminum chloride had been found to be'an effective catalyst
for sluggish Diels Alder reactions. When the Diels Alder reaction
of 103 and 79a was attempted in the presence of a catalytic amount
of anhydrous aluminum chloride, a vigorous exothermic reaction was
noted and could only be controlled by use of an ice bath. On work up,
an oil was obtained that yielded a double carbonyl absorbance at
1680 cm.- and 1710 cm.- Chromatography on silica gel resulted in
an oil with a strong 1705 cm.- absorbance and no 1680 cm.-1 absorbance.
The nmr spectrum exhibited resonances consistent with structure 107
with 65.85 multiplee, vinyl hydrogen, 1), 64.05 (singlet, allylic
hydrogens next to chlorine, 2) and 62.20 (multiplet, broad, ring
hydrogens, 12). No resonances that could be assigned to the acetoxy
methyl hydrogens were present.
The mass spectrum gave a parent peak at m/e 198 with a P+2 peak at
m/e 200 which was 35% of the parent. The theoretical P+2 intensity71
for one chlorine atom is 32.6%. The base peak of the spectrum was
m/e 163 corresponding to loss of the allylic chlorine atom.
The mechanism is not known for the formation of 107 but one
could speculate that the initial Diels Alder adduct, which appears to
be present in small yield from the infrared spectrum, reacts with
adventitious hydrogen chloride present. It is generally known that a
trace amount of this impurity is contained in aluminum chloride.
The use of diazoquinone, 104, as a dienophile was believed to
be an attractive course of action to pursue to obtain an adduct incapable
of aromatizing. Scheme VII shows the intended reaction sequence.
Kealy54 had reported the synthesis of 104 and had reported it to be
an extremely potent dienophile.
SI0 OR 1) catalytic
S / CH 20R hydrogenation
+ 77a,b _
,N CIH 2R 2) MsC1
The synthesis of 104 was accomplished by oxidation of the potassium
salt of maleic acid hydrazide with tert-butyl-hypochlorite at -77C
in acetone. Reaction of 104 with the diacetoxydiene 77a at -55C for
twelve hours yielded 45% of a light yellow crystalline solid which
was consistent with structure 108a in every respect. The infrared
spectrum gave the ester carbonyl at 1730 cm.- and the quinone amide
carbonyl at 1650 cm.-I The nmr spectrum gave resonances at 66.92
(singlet, vinyl hydrogens, 2), 64.80 (singlet, allylic hydrogens
adjacent to nitrogen, 4), 64.57 (singlet, allylic hydrogen adjacent
to acetoxy, 4) and 62.09 (singlet, methyl hydrogens, 6). The elemental
analysis agreed with the calculated value.
The reaction of 104 with the dihydroxy diene 77b yielded the
Diels Alder adduct 108b. Again all spectral characteristics were
consistent with the assigned structure. The infrared spectrum gave
absorbances for the hydroxyl groups at 3480 cm.- and 3400 cm.- The
amide like carbonyl appeared at 1630 cm.-1 The nmr spectrum had
resonances at 66.98 (singlet, vinylic hydrogens, 2), 64.51 (singlet,
allylic hydrogens adjacent to hydroxyl, 4), 64.11 (singlet, allylic
hydrogens adjacent to nitrogen, 4) and 63.50 (broad, probably hydroxyl
hydrogens, no integration). The elemental analysis agreed with the
Attempted catalytic hydrogenation of Diels Alder adducts
The catalytic hydrogenations of the adducts 108a and 108b to the
desired structure for the cyclizations were unsuccessful. Two catalyst
systems were employed and both were chosen for their efficient reduction
of double bonds and their inactivity toward hydrogenolysis of reactive
The catalyst system, 5% rhodium on carbon, has been successfully
used in the reduction of compound 110 to 111.72 Hydrogenation using
5% palladium on carbon has been found to be efficient in systems
[ 1-- ~ \\ >-
0--- =o0 - -'C=o
containing allylic acetoxy groups. Compound 112 was reduced to 113
using this system.73
0 OCH 1.
S3 0 CCH3
The results of the hydrogenation attempts are shown in Table 7.
In all cases, a mixture of products appeared to be present from the
complex nmr spectra. Attempts at separation by column chromatography
were almost unanimously unsuccessful.
Two distinct trends, however, were apparent from these experiments.
The first was that the quinone like double bond was easily hydrogenated
and the tetrasubstituted double bond was highly resistant to reduction.
In all atmospheric hydrogenations almost always approximately one-half
of the theoretical hydrogen uptake was recorded in the first fifteen
minutes. In experiment 3, the hydrogenation of 108b with 5% rhodium
on carbon in ethyl acetate, a compound was isolated and purified and
assigned structure 114. The nmr spectrum showed an absence of the
quinodial protons of the starting material and gave resonances at
64.29 (singlet, allylic hydrogens adjacent to hydroxyl, 4), 64.10
(singlet, allylic hydrogens, 4), 63.87 (singlet, broadened, hydroxyl
hydrogens) and 62.60 (singlet, hydrogens adjacent to carbonyl, 4).
In a hydrogenation attempt under more forcing conditions, i.e.,
5% rhodium on carbon at 35 psi., a white crystalline solid was isolated
that melted at 151-40C. The nmr spectrum was indicative of structure
115 which could result from the hydrogenolysis of a carbon nitrogen
bond. The resonance signals recorded were 64.31 and 64.13 (two over-
lapping broad singlets, allylic hydrogens, 4), 62.65 (singlet, hydrogens
adjacent to the carbonyls, 4) and 61.75 (broad singlet, allylic methyl
hydrogens, 3). The mass spectrum gave the correct parent peak at
m/e 228 and the base peak at m/e 210.
H CH 3
The resistance to hydrogenation of the tetrasubstituted double
bond was not surprising since at least one report of a similar case
is documented in the chemical literature. Gillis and Beck attempted
to hydrogenate 116 with Pd/C, Rainey nickel and Pd/Al2)3 at 59 psi
and were unsuccessful.
xJ-qOCH 2 Ctl3
Hydrogenation of the double bonds of compound 90 was recognized
as another route to 24. At atmospheric pressure hydrogenation with
5% rhodium on carbon in ethyl acetate resulted in the absorption
TABLE 7. CATALYTIC HYDROGENATION OF DIAZOQUINONE ADDUCTS
Rh/C, ethyl acetate
2250 cm.-1 ir
2250 cm.-1 ir
of one third the theoretical amount of hydrogen. On work up a white
solid was obtained with all spectral characteristics identical to its
hydroquinone derivative 91.
Attempt to prepare precursor to intramolecular Diels Alder pathway
27, the Diels Alder adduct of 2,3-di(bromomethyl)l,3-butadiene
and p-benzoquinone, was smoothly prepared by the method of Butler and
Ottenbrite.29 The 1,4-elimination of bromine from this compound by
the use of zinc in dimethylformamide28 to yield 28 was unsuccessful.
Attempts to trap the intermediate diene with maleic anhydride and
tetracyanoethylene were unsuccessful. In each case unidentifiable
soluble polymeric material was obtained. Since Alder and Fremery28
had postulated diradical-intermediates in their investigation of the
zinc dimethylformamide system, it is possible that the diene formed
but then quickly polymerized.
B. Attempted Synthesis of 5,5,6,6-Tetrasubstituted Cyclohexadiene
Two synthetic routes were investigated as possible means to
generate 5,5,6,6-tetracyanocyclohexadiene, 29. Earlier work75 had
shown that the Diels Alder adduct of butadiene and tetracyanoethylene,
117, was resistant to halogenation reactions. The routes investigated
CN CH3 CH3
117 118 119
were the reaction of tetracyanoethylone with a-pyrone, 118, and with
3,4-dimethylthiophene-l,l-dioxide, 119, and then subsequent loss of
carbon dioxide or sulfur dioxide to the desired tetrasubstituted diene.
Adducts of a-pyrone
118 was synthesized by the excellent preparation of Zimmerman and
coworkers. Surprisingly the cisoid diene failed to form the desired
Diels Alder adduct with tetracyanoethylene under a variety of conditions.
Upon mixing the two reactants, a strong reddish-brown color was always
observed; however either starting materials or black tars were re-
covered from each reaction as summarized in Table 8.
The failure of tetracyanoethylene to form a Diels Alder adduct
with 118 is surprising since this diene has been observed to be quite
reactive with rather unreactive dienophiles. Fieser and Haddadin77
observed 118 to react with the unusual dienophile 120 to yield 121.
0 1180 0-
Diels and Alder78 observed the formation of the Diels Alder adduct
with maleic anhydride. This reaction was successfully repeated.
Successful Diels Alder additions of 118 with both p-benzoquinone
and fumaronitrile to yield 122 and 123 were observed. 122 was
122 0 123
characterized by the infrared spectrum which gave both a 1750 cm.
ester carbonyl and a 1655 cm.- quinone carbonyl. The nmr spectrum
yielded resonances at 66.82 (singlet, quinone hydrogens, 2) 66.57
(triplet, vinylic hydrogens, 2), 65.69 (quartet, allylic hydrogens
next to oxygen, 1), 64.21 (quartet, hydrogen next to carbonyl, 1) and
63.60 (two doublets, bridgehead hydrogens, 2). The elemental analysis
was satisfactory for 122.
The fumaronitrile adduct, 123, gave weak infrared absorbances
for the nitrile groups at 2245 cm.-1 and 2200 en.-I The ester
carbonyl absorbance was at 1755 cm.- The nmr spectrum gave resonances
at 66.92 multiplee, vinylic hydrogen, 2), 65.80 multiplee, hydrogen
adjacent to oxygen, 1), 64.08 multiplee, unassigned, 2) and 63.63
multiplee, unassigned, 1). A correct elemental analysis was obtained.
Both dimethyl maleic anhydride and tetraethylethylenetetracarboxyl-
ate failed to form the Diels Alder adduct also. It is probable that
tetracyanoethylene along with these two dienophiles failed to react
because of steric interference. Under the forcing conditions, where
the tars were obtained, it is possible that the adduct formed, lost
carbon dioxide and then the resulting diene thermally polymerized
under the severe reaction conditions.
Adducts of 3,4-dimethylthiophene-l,l-dioxide
3,4-Dimethylthiophene-l,l-dioxide, 119, was prepared by first
the bromination of 2,3-dimethylbutadiene sulfone, 82 to the dibromide
124 and then dehydrohalogenation with sodium methoxide in tetrahydro-
furan. The results of this reaction were not reproducible. Sometimes
119, the desired product, was formed in good yield, while at other times
125 was formed. The literature preparation of 119 involved oxidation
Br -Br YH3
of the corresponding thiophene to the thiophene dioxide. A reported
dehydrobromination80 of 124 with potassium hydroxide only yielded 125.
119 was found to be unreactive with tetracyanoethylene in the
same manner as was a-pyrone. However, maleic anhydride reacted with
119 in refluxing toluene'to form 126.
This compound was identified by its infrared spectrum with an
TABLE 8. REACTIONS OF TETRACYANOETHYLENE WITH a-PYRONE
anhydride carbonyl 1850 cm.-L
anhydride carbonyl 1350 cia.
and 1780 cm.-1
and 1780 cm.
The nmr spectrum exhibited
resonance signals at 63.55 (siOglet, broad, hydrogen adjacent to
carbonyl, 4), 63.20 (singlet, broad, hydrogens at bridgehead, 2) and
61.62 (singlet, allylic methyl hydrogen, 6). The mass spectrum
yielded a molecular ion at m/e 276.
tube 600 188
All melting points and boiling points are uncorrected and reported
in degrees centigrade. Melting points were determined in open capil-
lary tubes on a Thomas-Hoover capillary melting point apparatus.
Pressures are reported in millimeters of mercury. Elemental analyses
were done by Galbraith Laboratories, Inc., Knoxville, Tennessee,
Peninsular ChemResearch, Inc., Gainesville, Florida or Atlantic Micro-
lab, Inc., Atlanta, Georgia.
Infrared spectra were recorded with either a Beckman IR 10 or
a Beckman IR 8 Spectrophotometer. Visible spectra were recorded on
a Beckman DK-2A Spectrophotometer. Proton nuclear magnetic spectra
were obtained by use of a Varian A-60 Spectrometer and all resonances
are given in the chemical shift parameter 6 and are measured from
tetramethylsilane (TMS) as an internal standard. Mass spectral data
were obtained using a Hitachi Perkin-Elmer RMN mass spectrometer.
All solvents were commercial reagent grade and used as received
unless specifically noted. Polymer nonsolvents were technical grade
and were filtered before use. All chemicals used as reactants were
obtained commercially and used as received unless specifically
designated as otherwise.
The comonomers divinyl ether, ethyl vinyl ether, isobutyl vinyl
ether, styrene, acrylonitrile, divinyl sulfone and vinyl acetate were
obtained from commercial sources and distilled immediately prior to use.
Divinyl carbonate and N-vinyl carbazole were obtained in high purity
from commercial sources and used without further purification.
Number average molecular weights were obtained from a Mechrolab
Model 302 Vapor Pressure Osmometer. Molecular weight distributions
were obtained from a Waters Associates GPC 300 Gel Permeation Chromato-
graph. Thermal characteristics of the polymers were recorded by a
Perkin-Elmer DSClB Differential Scanning Calorimeter. Intrinsic
viscosities were measured employing a Cannon-Ubbelohde semimicro
dilution viscometer using standard procedures for operation and
B. Copolymerizations and Related Reactions of 4-Substituted 1,2,4-
1. Synthesis of 4-phenyl and 4-methyl-l,2,4-triazoline-3,5-diones
Ethyl carbazate -Diethyl carbonate (2Q00 g., 1.80 mol.). and
88.0 g. (1.80 mol.) of 99% hydrazine hydrate were shaken together
for approximately 20 minutes. Some exothermicity was observed and
after this time the original two phase system blended into one phase.
The solution was left standing overnight. Distillation at 95C and
12 mm. yielded 147.7 g. (83.5%) of a clear liquid which solidified
on standing to a white solid, m.p. 43-45.5 (lit.81 44-45.5).
l-Ethoxycarbonyl-4-phenylsemicarbazide -Ethyl carbazate (70.0 g.,
0.67 mol.) was placed into a 500 ml. round-bottomed, three-necked
flask with 350 ml. of benzene. The flask was equipped with a magnetic
stirrer, a reflux condenser guarded by a calcium chloride drying tube,
a dropping funnel and a thermometer. After refluxing to get the
ethyl carbazate into solution, the solution was cooled to 10. Through
the dropping funnel 70.0 g. (0.59 mol.) of phonyl isocyanatewere intro-
duced at a rate -.hich kept the temperature between 100 and 200. After
all the isocyanate was added, the resulting white solid slurry was
refluxed for 20 minutes and then cooled and filtered. After drying
under vacuum 124 g. (89%) of the desired product m.p. 152-3' (lit.82
m.p. 154') were obtained.
4-Phenyl urazole82-To 275 ml. of hot 4M potassium hydroxide, 124 g.
(0.55 mol.) of l-ethoxycarbonyl-4-phenylsemicarbazide were added in
small portions while shaking the potassium hydroxide solution on a
steam bath. After the addition was completed, the solution was main-
tained on the steam bath for an additional 20 minutes. The solution
was cooled and carefully acidified with concentrated hydrochloric
acid. The precipitate was removed by vacu-um filtration and extracted
with 95% ethyl alcohol via a Soxhlet extractor. On cooling and then
treating the mother liquor 69.2 g. (70.3% yield) of the urazole m.p.
204-7' (lit. 82.p. 206-7') was obtained.
4-Phenyl-l,2,4-triazolina-3,5-dione, PhTD -Anhydrous sodium
sulfate, 25.0 g., was placed into a 500 ml. Erlenmeyer flask with
300 ml. of methylene chloride and 6.0 g. (0.034 mol.) of 4-phenyl
urazole. The slurry was stirred by a magnetic stirrer and cooled to
0-5' with an ice-water bath. By means of a pipette attached to a gas
cylinder with a tube, dinitrogen tetroxide was bubbled slowly through
the solution. The temperature was kept below 5'. after a few minutes
a deep red color developed. After all the urazole appeared to have
been consumed, usually in about thirty minutes, the sodium sulfate
was removed by vacuum filtration and the deep red solution was
evaporated using lukewarm water on a rotary evaporator. 4.8 g. (80%
yield) of dark red crystalline PhTD was obtained. Purification was ac-
complished by sublimation at 70-750 under a vacuum of less than 0.5 nn.
The red solid was stored in the absence of light in a freezer until it
4-Methyl-l,2,4-triazoline-3,5-dione, MTD -The same procedure used
in the synthesis of PhTD was employed in the synthesis of MTD. This
synthesis had been previously reported in the literature.83 ethyl iso-
cyanate was added to the ethyl carbazate and a quantitative yield of the
corresponding semicarbazide derivative was obtained. This was cyclized
with 4M potassium hydroxide to the corresponding 4-methyl urazole in a
yield comparable to the phenyl derivative. Oxidation with dinitrogen
tetroxide was accomplished in approximately 90% yield. The light red
fluorescent appearing powder was purified by sublimination at less than
0.1 mm. and a temperature between 50 and 60.
2. General copolymerization procedures
a. Spontaneous copolymerizations
The following experimental procedure was typical of all spontaneous,
room temperature ccpolymerizations unless specifically noted otherwise.
All glassware, i.e., Erlenmeyer flasks, syringes and volumetric
flasks were scrupulously cleaned, rinsed with acetone, dried in an 80*
oven and then allowed to cool to room temperature in a large desiccator.
On removal from the desiccator the flasks were immediately capped with
serum caps. Solvents were reagent grade and dried for a minimum of 24
hours over 3A molecular sieves.
Ethyl vinyl ether, EVE, PhTD copolymerization-A 50 ml. Erlenmeyer
flask equipped with a magnetic stirring bar and a rubber serum cap was
charged with 0.499 g. (2.84 mmol.) of PhTD by removing the cap and adding
as a solid. Next, 15 ml. of methylene chloride were introduced using a
syringe and adding through the serum cap. This was followed by the add-
ition of EVE (0.203 g., 2.83 mmol.) in the following manner. The vinyl
ether was carefully weighed into a 10 ml. volumetric flask through a
serum cap and then the flask was brought to volume with methylene
chloride. A careful quantitative transfer to the reaction vessel was
made using a syringe. The reaction mixture was stirred and within two
minutes the dark red colored solution, characteristic of PhTD, was
discharged to yellow. After 30 minutes the reaction mixture was pre-
cipitated by dropping slowly through a coarse filter frit into cold
hexane. The hexane solution was filtered to yield 0.615 g. (86% yield)
of a white-yellowish granular appearing solid. The solid appeared to
soften at 1000 and above 1500 it darkened and decomposed. The number
average molecular weight was determined to be 1660 by use of a Vapor
Pressure Osmometer in acetone solution. The infrared spectrum (KBr)
showed absorbances at 2980 (w), 1770 (w), 1715 (s), 1610 (s), 1500 (m),
1470 (m), 1420 (n), 1310 (m), 1130 (n), 1070 (w), 900 (w), 820 (w),
760 (m) and 700 (m) cm.I The nmr spectrum (CDC3 ) gave resonance
signals at 67.45 (s, broadened, 5), 66.00 (m, broad, 1), 63.83 (m, broad)
and 61.16 (t, broad, 3). The spectrum is shown in Fig. 1.
Anal. Calcd. for a 1:1 copolymer structure, C12 H3N303: C, 58.30;
H, 5.26; N, 17.00. Found: C, 58.08; H, 5.14; N, 17.16.
Table 1 contains a summary of experimental results of all EVE-
b. Copolymerizations at 600
The following experimental procedure was typical of all copolymer-
izations at 60 unless specifically noted otherwise.
The copolymerizations were effected in heavy walled 13 cm. x 3 cm.
Pyrex tubes. All glassware i.e., Pyrex tubes, syringes and volumetric
flasks was thoroughly cleaned, rinsed with acetone, dried in an 80"
oven and then allowed to cool to room temperature in a large desiccator.
On removal from the desiccator the volumetric flasks and the polymer
tubes were sealed with serum caps. Solvents were reagent grade and
were dried for a minimum of 24 hours over 3A molecular sieves. Copolymer-
izations were conducted in a 600 oil bath regulated by a Sargent NS1-
12 controller to 0.1.
Copolymerization of divinyl ether, DVE, and PhTD at 60-A heavy
walled Pyrex polymer tube, capped with a serum cap, was charged with
0.508 g. (2.81 mmol.) of PhTD by removing the cap and adding the solid.
This was followed by the addition of 15 ml. of methylene chloride by.
means of a syringe. Next DVE (0.200 g., 2.82 immol.) was carefully
weighed into a 10 ml. volumetric flask using a syringe and adding
through a serum cap. The flask was brought to volume with methylene
chloride and then carefully transferred to the polymer tube by means of
a syringe. The polymer tube was then placed on the high vacuum line
and carried through a freeze-thaw degassing cycle using a 10 mm.
vacuum. The tube was sealed and warmed to room temperature before
placing in the 600 bath. After 24 hours at 600 the originally red
colored solution had changed to yellow. This color change normally
took place within a few minutes after placing the tube in the bath. The
tube was removed from the oil bath after 24 hours, cooled to room temp-