STUDY OF CYCLOCOPOLYMERIZATION
YUAN CHIEH CHU
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UII'.VEF::ITY OF FLORIDA IN PARTIAL FULFILLtIF1T OF THE
PE..UiIF rlENTS FOR THE r,'EFFEE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
the mother of Lawrence Chu
The author would like to express his special gratefulness to
Professor George B. Butler for his encouragement and invaluable
guidance during the execution of this work.
The author also would like to express his appreciation to his
advisory committee and other faculty members in the Department of
Chemistry whose assistance and suggestions contributed much to this work.
He wishes to thank Mrs. Carol Albert for typing this thesis, Professor
G.B. Butler for checking the English composition, Dr. S.F. Tien for
his suggestions on the theoretical aspects, and T. Baugh for his work
on C nmr spectroscopy.
Finally, the author wishes to express great appreciation to his
wife, Rolan, whose forbearance and understanding made this work
The financial support for this work from the Department of
Chemistry and Professor G.B. -Butler's National Institutes of Health
grant is gratefully acknowledged.
TABLE OF CONTENTS
ACKNOWLE ME1 T . .
LIST OF TABLES . .
LIST OF FIGURES . .
I. INTRODUCTION . .
History of Cyclopolymerization and
Cyclocopolymerization . .. 1
Mechanism of Cyclopolymerization .. 11
Mechanism of Cyclocopolymerization ... 16
Statement of Problem . ... 22
II. THE STRUCTURE OF COPOLYMER OF DIVINYL ETHER-
MALEIC ANHi'L. FIDE SYSTEM . 24
Introduction . ... 24
Results and Discussion . ... 25
The Mechanism of Cyclization ... 47
III. THE COPOLYMERIZATION OF DIVINYL ETHER-
FUMARONITRILE . ... 52
Introduction . ... 52
Results and Discussion . ... 53
IV. EXPERIMENTAL . . 100
Materials . .
Equipment and Data . .
Synthesis Related to Monomer Preparation
Copolymerization . .
APPENDIX TEF E.LIHEMISTRY BACKGROUND FOR 5,5- AND
5,6-BICYCLIC SYSTEM IN THE COPOLYMER OF DVE-MAH ... .119
BIBLIOGRAPHY. . . ... 123
BIOGRAPHICAL SKETCH . . ... .. .127
LIST OF TABLES
I. The Relation Between Concentration and Cyclization. 12
II. The Energetical Parameters for Cyclopolymerization. 15
III. The Reactivity Ratio of Copolymerization .. 18
IV. The Comparison Between JHD and JHH .. 28
V. The Chi-mi:cal Shifts for the Methine Protons of
DVE-MAH Copolymer . .. 33
VI. Chemical Shift Differences Between cis and trans
Disubstituted Vicinyl Carbons in 13C nmr Spectra. 44
VII. Determination of Equilibrium Constant of FN-DVE in
Acetonitrile with Constant FN Concentration
(0.00101 m/l) . .... 57
VIII. Comparison of Copolymer Initiated by Different Methods
in Acetonitrile. ... . 61
IX. The Compositions of Copolymers Prepared in Aceto-
nitrile at Room Temperature within 10% Conversion 64
X. The Limiting Yield of Copolymerization with Excess
DVE in Feed Composition. . 67
XI. Comparison of 1C nmr Spectra between Homopolymer of
DVE and the Copolymer of-DVE and FN ... 70
XII. The Light Intensity Dependence on Quantum Yield and
Rate in Acetonitrile at Room Temperature .. 73
XIII. The Linear Relations of Intensity (I) to the Quantum
Yields (4) and Rates . 74
XIV. Comparison of Quantum Yield and Rate at 236 nm and
300 nm . . ... 75
XV. The Rate of Copolymerization of DVE-FN System in
Acetonitrile at 300 nm, at I = 1.52 x 1016
Photons/Sec. . . .. 82
XVI. The Rate of Copolymerization of DVE-FN System in
Acetonitrile at 236 nm, at I = 6.4 x 1014
Photons/Sec. . . .. 84
XVII. The Dependence of Total Concentration on Rate. .. 90
XVIII. The Determination of Reactivity Ratio rl ..... 97
XIX. The Copolymerization of DVE-FN System with Additives 111
XX. The Copolymerization of DVE-MAH System at Different
Temperature . .... 116
LIST OF FIC'FE[
1 IR spectrum of the copolymer of DVE-MAH prepared in
xylene at 1300C . .... 29
2 60 MHz nuclear magnetic resonance spectra of (a) DVE-MAH
copolymer, (b) DDVE-MAH copolymer in acetone d 32
3 100 MHz nuclear magnetic resonance spectrum of DDVE-MAH
copolymer ....... ... ..... .35
4 300 MHz nuclear magnetic resonance spectra of (a) DVE-MAH
copolymer, (b) DVE-MAH [r ".H copolymer in acetone-d6 .. 36
5 C nuclear magnetic resonance spectra of (a) DVE-MAH
copolymer, (b) poly(maleic anhydride), (c) hydrolized
DVE-MAH copolymer . . ... .40
6 C nuclear magnetic resonance spectrum of 2,3-dimethyl
succinic acid . . ... 42
7 C nuclear magnetic resonance spectra of DVE-MAH
copolymer prepared at (a) 130, (b) 100, (c) 72, (d) 250C 46
8 The absorption of the complex of DVE-FN in acetonitrile
(a) 0.6 m/l of DVE, (b) 0.6 m/l of FN, (c) (DVE) =
(FN) = 0.6 m/1. . . .. 54
9 The determination of the stoichiometry of the DVE-FN
complex in acetonitrile by continuous variation method
at 300 nm (DVE) + (FN) = 0.60 0.03 m/1. ... 55
10 Charge transfer absorption of DVE-FN complex in aceto-
nitrile. . . ... ..... 58
11 IR spectra of the DVE-FN copolymer initiated (a) at
300 nm, (b) at 236 nm, (c) by Alb!:. . ... 63
12 60 MHz nuclear magnetic resonance spectra of the DVE-FN
copolymer initiated (a) at 300 nm, (b) at 236 nm,
(c) by AIBN. . . ... .. ... 63
13 The compositions of the DVE-FN copolymer initiated at
(a) 236 nm, (b) 300 nm. . ..... 65
14 1C nuclear magnetic resonance spectra of the copolymer
initiated (a) at 236 nm, (b) by AIBN, (c) at 300 nm. 69
15 The dependence f feed composition on rate at 300 nm
(I 1.52 x 10 photons/sec) for total concentration
(a) 2.40 m/l, (b) 1.20 m/l, (c) 0.60 m/1. ... 83
16 The depencenc4 of feed composition on rate at 236 nm
(I 6.4 x 10 photons/sec) for total concentration
(a) 2.0 m/l, (b) 0.6 m/1. . ... 85
17 The dependence of total concentration on rate (x) at 300 nm,
fd = 0.5, (A) at 300 nm, fd = 0.6, (o) at 236 nm,
fd = 0.5, (*) at 236 nm, fd = 0.8. ... .' ..91
18 The determination of reactivity ratios -- (T) = 2.0 m/l,
at 300 nm ---- (T) = 2.0 m/l, at 236 nm, -- (T) =
1.1-4.5 m/l, with AIBN. . . ... 98
19 60 IH:: nuclear magnetic resonance spectra of (a) DVE,
(b) DDVE with 250 ppm sweep width, (c) DDVE with 100 ppm
sweep width. . . ... ... 105
20 60 MHz nucelar magnetic resonance spectra of (a) 2,2,2',2'-
tetradetueriodiethylene glycol, (b) diethylene glycol. 107
21 60 MHz nuclear magnetic resonance spectra of (a) bis-
(2-bromo-2,2-dideuterioethyl)ether, (b) bis(2-bromoethyl)-
ether. . . ... ....... 109
22 The viscosities of the DVE-MAH copolymer in DMSO prepared
in (a) benzene at 76-780C with 77% conversion, (b) in
benzene at 51-540C with 25% conversion, (c) in benzene
at 61-640C with 47% conversion, (d) in CH.i at 60-610C
with 43% conversion, (e) in benzene at 28-2AOC with 31%
conversion initiated by light. . .... .. 118
Abstract of 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
STUDY OF CYCLOCOPOLYMERIZATION
Yuan Chieh Chu
Chairman: George B. Butler
Major Department: Chemistry
The well-known alternating 1:2 cyclocopolymer of divinyl ether
(DVE) and maleic anhydride (MAH) has been found to possess interest-
ing anti-tumor and biological activities. Recent research on the
structure of the cyclopolymer has raised a question about the ring
size of this cyclocopolymer. In this research program the structure
of this copolymer was reexamined by use of spectroscopic methods.
A linear, soluble copolymer was obtained by the copolymerization in
solvents with active hydrogens, such'as chloroform and xylene. The
ir spectra showed the existence of both monomers in the copolymer
based on the strong bands at 1775, 1855, 1230 cm1 (cyclic anhydride),
and 1090 cm- (cyclic ether).
By use of deuterated copolymers, the H-nmr peaks at 6 2.31, 3.47,
4.06, and 4.49 ppm with an area ratio of 2:1:1:1 were assigned for
the hydrogens of methylenes, methines on the backbone anhydride unit,
methines on the ring anhydride unit, and methines adjacent to oxygen
on the cyclic ether ring, respectively. Through the examination of
the possible isomeric structures of the bicyclic ring, the splitting
of each peak group was further assigned for cis and trans di-substi-
tutions on the anhydride unit.
The splitting pattern on the 300 MHz nmr spectrum of divinyl
ether 2,3-dideuteriomaleic anhydride copolymer confirmed the unsym-
metrical six-membered ring structure. 1C nmr spectra was discussed
and was consistent with the conclusion from the H-nmr spectra.
A chair-form six-membered ring with trans isomer next to each
side of the ring oxygen with a small portion of cis isomer was assigned
for the structure of DVE-MAH copolymer.
Based on little or no change of the 1C nmr spectra of the
copolymers prepared at different temperatures, it was concluded that
there was no significant change of structure caused by temperature
effect. This led to the sole assignment of the six-membered ring
structure of the copolymer as an energetically favored product.
A mechanism for the cyclization was proposed based on the HOMO-
LUMO interaction of the comonomers and the radical intramolecular
addition on the preoriented double bond. This mechanism led to the
formation of six-membered ring structure of the copolymerization as
the sole product.
The participation of the charge transfer complex in the cyclo-
copolymerization was investigated with the divinyl ether-fumaronitrile
system by light initiation. The evidence and the composition of the
complex were obtained by the UV spectroscopic method. The structure
of the copolymer was proposed on the basis of elemental analysis and
C nmr spectrum, as a mixture of 1:1 cyclocopolymer and 1:2 cyclo-
copolymer structures, both with six-membered rings. The initiation
mechanism involving polar deactivation of excited molecules and the
propagating mechanism were proposed to explain the rate maximum
phenomena and the relation between copolymerization rate and total
concentration of the comonomers at 236 nm and 300 nm irradiation.
A consistent result was obtained by the proposal of the participation
of the CTC in both the initiation and propagation processes.
History of Cyclopolymerization and Cyclocopolymerization
It is well known that multifunctional monomers such as unconjugated
divinyl monomers are crosslinking agents in vinyl polymerizations.
However, certain compounds, such as diallyl phthalate and ethylene
diacrylate, undergo cyclization during polymerization, with the result
that 40-75% of the divinyl compound used is incapable of crosslinking.
Butler and coworkers3 found that several diallyl quarternary ammonium
salts polymerized to yield soluble and hence linear polymers containing
little or no residual unsaturation. Under the same conditions monoallyl
ammonium salts failed to polymerize. To explain the unusual results,
Butler and Ai4.1 o proposed an alternating intramolecular intermolecular
chain propagation mechanism to form a cyclic structure in the chain
(eq. 1). This type of process is now commonly termed cyclopolymerization.
Since these initial investigations, a large number of 1-6 non-
conjugated dienes of type I have been cyclopolymerized using appropriate
catalysis to yield soluble, most-ly saturated polymers with cyclic struc-
tures in the main chain.
In cyclocopolymerization, the comonomer contributes to formation of
the cyclic structure along with the nonconjugated diene. Butler reported
the first example of cyclocopolymerization. By copolymerizing divinyl
ether (DVE) and maleic anhydride (MAH), using a free radical initiator,
a soluble polymer was obtained. A mechanism similar to the cyclopolymer-
ization was proposed which is shown in equation 2.
This cyclocopolymer has been found to possess interesting
X CH Si
P 0 :S
/\ 0 / '
+ R* Initiation>
+ R -------;
Intermolecular \t agation'
- Alte > >r
The proposed mechanism was based on the following facts:6
(a) a soluble copolymer was formed, (b) little or no residual unsaturation
was detected in the copolymer and characteristic absorption bands for
cyclic anhydride and ether were observed in infrared spectra, (c) the
composition of the copolymer was close to a diene-olefin molar ratio
of 1:2 over a wide range of comonomer feed composition, (d) quantitative
conversion to copolymer was obtained, (e) the cleavage by hydriodic acid
and the incorporation of iodine into the copolymer indicated the pres-
ence of cyclic ether groups, and (f) neither monomer homopolymerizes
readily under the same condition as copolymerization.
Following Butler's discovery, the copolymerization reactions for
a wide variety of 1,4-dienes and olefins have been reported. Butler
and coworkers further studied some mono-olefin-1,4-diene systems to
produce 1:2 (1,4-diene:monoolefin) copolymer with generalized structure II.
X X = 0, SO, Si
/ /2 / oCH3
Y n O0
y Y = -CN or Y-Y = 0
Stille and Thomson prepared, with a variety of free radical initia-
tors, a soluble copolymer of a nonconjugated diene, 1,5-hexadiene, and
sulfur dioxide. Elemental analysis of the copolymer samples substantiated
the existence of a material containing exactly two sulfur dioxide portions
to one of diolefin. The six-membered ring structure was confirmed by
comparing the ir spectra of the copolymer sample with pentamethylene
Price and coworker prepared the copolymers of N,N-divinyl
aniline with diethyl fumarate in bulk and solution with azobis-isobutyro-
nitrile (AIBN) at 600C. Diethyl fumarate, having widely different
polarity from that of divinyl aniline, and being reluctant to homo-
polymerize, polymerized with divinyl aniline in widely different monomer
concentration ratios to give copolymers of nearly constant 2:1 molar
ratio composition. The colubility of this copolymer and the negative
results of analysis for residual double bond by infrared absorption in
the 6.0-6.2 p region, supported the cyclic copolymer structure IV.
C 6H COOEt
The copolymerization of divinylphenylphosphine with acrylonitrile
was studied by Butler, et al. The copolymer produced was soluble in
DMF and showed no residual unsaturation. Elemental analysis indicated
the copolymer contained 0.265 mole fraction of divinylphenylphosphine
and 0.732 mole fraction of acrylonitrile. The structure of this
copolymer and the reaction for its formation are as follows (eq. 3):
+ nCH 2=CH -->
'p ---------CH -CH --
.27 CN 0.73
Rigid or multicyclic systems have been reported. Meyersen and
Wang0 prepared several copolymers of fused ring systems such as bicyclo-
pentene with sulfur dioxide or maleic anhydride in solution by free
radical initiation. Copolymer was obtained in the ratio of 1:2 as in
structure V in equation 4.
X = SO or
X :SO2 or
Yamaguchi and Onoll reported the copolymerization of sulfur dioxide
and dicyclopentadiene (DCPD) in liquid sulfur dioxide with AIBN or other
radical initiators at o or 200C. Soluble copolymer was obtained with a
composition ratio of DCPD to sulfur dioxide of 1:2, and containing
^/ \ ^
virtually no residual double bond.
was proposed as VI.
The structure of this polysulphone
Butler and Pledger2 proposed 1:1 alternating copolymer of maleic anhydride
with 5-ethylidene-, 5-methylene-, and 5-vinyl-bicyclo(2,2,l)-2-heptene.
Evidence supported a tricyclic structure which incorporated maleic
anhydride as part of a six-membered ring repeating unit, for example
The essential structure of the cyclocopolymers were proposed based
on the following methods.
1. Elementary analysis Most of the copolymers which have been studied
have a 1:1 or 1:2 comonomer composition in the copolymer regardless of
the wide range of feed composition, solvent, or initiator concentration.
2. Little or no residual absorption in carbon carbon double bond -
This indicates that all the double bonds in the diene were incorporated
into the copolymer. Only high degree of cyclization and crosslinking
can explain this observation. The residual unsaturation can be
detected by the absence of, or little ir absorption in the 6.0 to 6.2 p
region and nmr. Negative results of catalytic hydrogenation has been
used to show the absence of residual carbon carbon double bond in the
3. The solubility of copolymer and the absence of gelation at high
conversion rule out the crosslinked structure.
4. Existence of both monomers in the resulting polymer This can be
shown by the characteristic absorption bands of the comonomers in the
infrared spectra. For example, the 1100-1085 cm strong, broad band
for cyclic ether and the 1860, 1795 cm- bands for cyclic anhydride in
the ir spectra of DVE-MAH copolymers indicating the existence of both
comonomers in the copolymer.
5. The alternating ratios are always supported by elementary analysis.
Chemical method may support the proposed structure. Butler and
Guilbaultl3 hydrolyzed and dehydrohalogenated the copolymer of DVE-
chloromaleic anhydride, followed by the treatment with KMnO to get the
vic-diol copolymer. The degradation result of this vic-diol copolymer
upon periodic acid cleavage supported the 1:1 structure with only DVE
units on the backbone.
i. KOH 1. KMnO
SC1n 2. HC L H n 2.
V ---0 I
COOH n C=O C=0
'H n COOH
Both five- and six-membered rings can be formed during the cycliza-
tion step. Unfortunately not much work has been done along this line
to distinguish between these possibilities.
Stille and Thomson7 used the ir absorption spectra of model compounds,
VIII and IX, to compare with the 1,5-hexadiene-sulfur dioxide copolymer.
They concluded that the six-membered ring structure (III) was present in
This method has also been used by Meyersen and Wang0 in their
fused ring system-sulfur dioxide copolymer as mentioned in the last
Butler and Fujimori14 studied'ring size by the readiness of the dehy-
drohalogenation of the cyclocopolymer. The quantitative elimination
of only half the chlorine content of dichloro-maleic anhydride-DVE
cyclocopolymer provided the first support for the proposed six-membered
cyclic repeating unit. There would have been no such elimination if the
repeating unit were five-membered. If both the halogens and the
hydrogens at the 2-position to the oxygen were eliminated, the product
would be a furan derivative which should be detectable due to its
O7 _-- -HC
Cl Cl n / n Eq. 6
O '\0 0 CO CO
aq. NaOH i
0 0 n -.2HC I nN Eq. 7
On the other hand, the ring structure of the cyclopolymer has been
intensively investigated by chemical and spectroscopic methods. Butler,
CrawshaS and Miller5 conclusively proved the existence of cyclic struc-
ture of cyclopolymer prepared by free radical initiation for diallyl
quaternary ammonium salts. A nondegraded polymer was obtained by
treatment with KMnO4. However, the ring size was not determined by
these authors. Later, by 13C nmr spectroscopy,6 both the five- and
six-membered rings were found in these cyclopolymers by comparing with
model compounds. Dehydrogenation of the cyclopolymer followed by the
identification of the resulting aromatic rings enabled Marvel and Vest17
to confirm the six-membered ring structure in the cyclopolymer of
2,6-dicarbomethoxy-l,6-heptadiene prepared by free radical initiation.
Me OMe Me
C OO 0e
C r R- KC10 Eq. 8
Spectroscopic methods have been applied to studying the ring
size for the cyclopolymer. Butler and Myers8 used both ir and nmr
spectroscopy to analyze the cyclopolymer obtained from dimethacrylamide
and its N-methyl and N-phenyl derivatives. They found that the polymers
were composed of both five- and six-membered rings. Electron spin
resonance studies have shown that the radicals from N-substituted
N,N-diallylamine in the presence of TiCl /H 02 or TiCl /N2NOH initiation
are five-membered ring cyclic species. It is possible that the radicals
involved in polymer formation have very short lifetimes (thereby not
detectable by ESR) and those radicals which are detectable are stable
but non-propagating species.
The extended chemical shift range, the absence of the simplicity
of proton decoupled 1C nmr spectra, in which each carbon resonates as a
singlet, makes the 13C nmr spectroscopy extensively useful for study of
polymer structures. Johns, Willing, Middleton,and Ong21 used the 13C
natural abundance pulsed Fourier transform nmr spectroscopy to distin-
guish the different structure features of the polymers formed by radical
induced cyclopolymerization of a series of N,N-diallyl amines by
comparing with the model compounds. The polymers of N,N-diallylamines
all contained cis and trans substituted pyrrolidine rings with the ratio
5:1. The polymers of N,N-bis(2-allyl)amine gave complex spectra due to
the presence of both cis and trans pyrrolidine and piperidiene rings.
Mechanisms of Cyclopolymerization
Since the discovery of cyclization in the cyclopolymerization, many
studies have been made on the nature of the cyclization step by theoreti-
cal consideration and experimental methods.
A statistical approach to cyclopolymerization was taken by Butler
and Raymond.22 They concluded that to explain the high degree of cycliza-
tion at high monomer concentration (Table I), a more favorable pathway
from 1,6-diene to cyclic polymer might exist than would be predicted
on a purely statistical basis.
The Relation Between Cyclization
Monomer Cone. m/l Cyclization % Conc. m/1 Cyclization %
Diallyl quaternary pure
ammonium salts >5.0 96-100 monomer <50
Diallyl silanes 1.2-2.3 >95 1.00 70
Acrylic anhydride %8.0 9P-100 0.10 96
Diallyl phosphine bulk 100 0.001 100
Gibbs and Barton23 explained this by considering the presence of the
large pendant group which would tend to prevent intermolecular reaction
and will frequently be presented to the reactive species in a conforma-
tion which is favorable for polymerization (cyclization). Butler
et al. studied the effect of the stability of the cyclized radical on
the rate of cyclopolymerization. The overall rates of divinyl monomers
were considerably greater than that of the corresponding monovinyl
monomers, the ratios varying from 2 to 10. They estimated the effective
concentration of the intramolecular double bond with respect to the
radical propagation, and values greater than 20 M at 500C were obtained.
This incredible effective concentration indicated that a considerable
preorientation prior to reaction exists. This favorable preorientation
may be related to an electronic interaction between the developing
radical site and the intramolecular double bond or in the ground state
before the initiation as proposed by Butler.25
In an effort to explain the strong polymerization of 1,6-heptadienes
relative to their monoolefinic counterparts, Butler25 proposed that
an electronic interaction between the unconjugated double bonds of
1,6-dienes or between the intramolecular double bonds and the reactive
species after initiation might exist.
These interactions would reduce the entropy change in going from the
ground state to the activated state required for the intramolecular
pr.1.: .tion. The ground and activated state energies both would be
reduced. Mikulasova and Hvirik26 calculated the total activation energy
for radical polymerization of diallyldimethylsilane and found it to
be ca. 9 Kcal/(mole double bond) less than that for allyltrimethylsilane.
The electronic interaction is supported by UV absorption evidence,27
but it is not necessary for cyclopolymerization as shown by Gibbs.28
They found that in the case of methacrylic anhydride versus methacrylic
acid, there was essentially no difference in total activation energies
in the polymerizations. Marvel and Stille29 obtained a cyclic polymer
from 2,5-dimethyl-l,5-hexadiene, and suggested an unusual driving
force from diene monomer to cyclic polymer.
The activation energy for cyclization is not necessarily smaller
than the intermolecular addition for same radical as shown by several
authors based on the kinetic scheme developed by Mercier and Smets.30
They derived the kinetic relationships between intramolecular and
intermolecular propagation for the free radical polymerization of
(CH 2CHCO) 0
The rate ratio of the intermolecular propagation (R.) to the intra-
molecular propagation (R ) was derived;
R./R = 2 [monomer] l -
i c -E /RT
where the difference in activation energies (E -E.) can be calculated.
Several E -E. values are listed in Table II. These results indicated
that the intramolecular propagation step requires greater energies
than the intermolecular step (E >E.). However, the rate of cyclization
is considerably larger than for intermolecular propagation (k /k.>l).
The values for the ratio A /A. indicated a high entropy factor favoring
cyclization. The decrease in entropy for a cyclization step would
perhaps be expected to be smaller than that for addition of a new
monomer unit. Only rotational motion will be lost in cyclization; on
adding a new monomer molecule to the chain, the loss of translational
and rotational degrees of freedom will result. Therefore, as far as
entropy is important, cyclization would be favored over intermolecular
The Energetic Parameters for Cyclopolymerization
Monomers E -E. Kcal/mole A /A. mole/l k /k. mole/l
C 1 c 1 c i
Acrylic anhydride 2.4 167 5.9
o-Divinyl benzene 1.9 50 2.8
Diallyl phthalate 0.3
Methacrylic anhydride 2.6 256 2.4
Guaita31 has studied the temperature independent factor of
cyclization parameters for the free radical copolymerization of acrylic
anhydride and divinyl ether. The results indicated that the high frac-
tions of cyclization in cyclopolymers from symmetrical unconjugated
dienes can be thermodynamically accounted for by an entropic effect
largely exceeding the energetical one. The entropy decrease was smaller
in the intramolecular reaction than in the intermolecular reactions.
These entropic effects are consistent with Butler's25 postulation of
the preorientation of two double bond by electronic interaction in both
ground and activated states. The nature of this preorientation between
the double bonds has been though to be due to a charge transfer complexa-
tion in the copolymerization, especially in the cyclocopolymerization to
produce alternating copolymers.
Mechanism of Cyclocopolymerization
Barton and Butler32 described a general
equation of 1,4-dienes and monoolefins where
is bimolecular in construction. The kinetic
in equation 10.
mi + M
m + M2
the cyclic repeating unit
scheme considered is shown
m + M ---2
m + M -----
me + M2 -2
m2 + M1 ---j
mS + M2
M is the diene CH 2=CH-X-CH=CH2 where X is CH2, O, SO2, etc. M2 is
the monoolefin, CHY=CHZ. The m' is the radical,-QCH CHCHHCH2, m2 is the
radical --CHYCHZ, m3 is the uncyclized radical
and m" is the cyclized radical
A five-membered ring structure for m" is possible.
The derived equation which related copolymer composition to monomer
feed composition in terms of five reactivity ratio parameters was given
as equation 11.
1 1 x
(1 + rx) + ( -)]
1 M a r
n = (E
1 x 3 1 r2 -
S[ + + 2] + + ( 1 + + )(1 + r x)
a r x M2 x c
where x = M /M2 is the mole fractional ratio of monomers in the feed,
n = m /m2 is the mole fractional ratio of monomers in the copolymer at
low conversion, rI = kll/kl2' r2 = k22/k21' r3 = k32/k31' rc = kcl/kc2'
and a = k /k32. The equation may be approximated to simpler forms in
the following cases.
a. If kc>>k32 so that a is very large and cyclization is the predominan
reaction of the radical, m3, the equation (11) gives
n = (1 + rlx)(l + r x)/[r x + (r2/x) + 2]
1 c 2
This is equivalent to considering the addition of monoolefin to diene
radicals to be a concerted bimolecular step proceeding through a cyclic
transition state and producing the cyclic repeating units.
b. If in addition there is a strong alternating tendency so that
r lr2,r =0 then equation (11) reduces in the limit to n= 1/2. This
predicts an alternating copolymer composition of 2:1 molar in contrast
to 1:1 for the similar limiting case of the classical binary copolymer
composition equation and that for the cyclopolymer composition equation.3 34
c. If the diene has a negligible tendency to add to its own radicals and
rl=r c0 and there is also predominant cyclization, then equation (12)
n = 1/[r,/x) + 2]
All three cases have been found as in Table III.3
The Reactivity Ratio of Copolymerization
Case a Case b Case c
System r r2 r3 System System r2
DVE-AN 0.024 0.938 0.017 1,4-PD-MAH 1,4-PD-AN 1.13
DVS-AN 0.364 0.067 0.067 DVE-MAH DM-1,4-PD-AN 3.31
DVE-PMI DVE-4-VP 32.0
DVE = Divinyl ether; AN = Acrylonitrile; 1,4-PD = Dimethyl-l,4-penta-
diene; MAH = Maleic anhydride; PMI = N-Phenyl maleic imide;
4-VP = 4-Vinyl pyridine: DVS = Divinyl sulfone; DM-1,4-PD =
In cases (b) and (c) very low values for r have been reported.
This low value is remarkable in the case of monomers such as acryl-
nitrile, and indicates that almost all the AN radicals in m3 react
to form a ring. The actual ring formation may be either a stepwise
reaction or a concerted reaction to form the product. Orientation
of monomers via a charge transfer complex (CTC) prior to free radical
reaction explains this unusual cyclic structure and also accommodates
the kinetic data (eq. 14).
O R R "
.--- Concerted __
/ \0 n, (Eq. 14)
Evidence for the participation of a CTC in the copolymerization
between styrene and MAH was presented by Tsuchida and Tomono.3 They
concluded that the CTC and uncomplexed MAH took part in the copolymer-
zation. Evidence for participation of a CTC in the cyclocopolymeriza-
tion of 1,4-dienes with monoolefin was presented by Butler and Joyce36
on the comonomer pairs, DVE-MAH, DVE-MI, DVE-FN and dihydropyran
(DHP)-MAH. Butler and his group have intensively studied the partici-
pation of CTC in cyclocopolymerization. The evidence of the existence
of the CTC was confirmed by both UV and nmr spectroscopies. The
equilibrium constant of the complexation can be obtained by these
spectroscopic methods. The complex is formed by interaction of an
electron-rich donor (D) and an electron -deficient acceptor (A).
D + A ----- [D, A--]
The composition of all the complexes formed by 1,4-diene and monoolefins
were 1:1 complexes. The alternating copolymer compositions found were
2:1 in olefin to 1,4-diene ratio for most of the 1,4-dienes with MAH or
FN. This is consistent with the postulation that the CTC undergoes a
1:1 alternating copolymerization with the electron acceptors such as
MAH and FN, which can thus account for the structure of the copolymer.
Some 1:1 copolymer has been obtained which can be considered as the
product of the homopolymerization of the 1:1 complex.
One necessary requirement for these alternating copolymerizations
is that neither of the comonomers should be homopolymerizable to a
significant extent under the same condition of the copolymerization. If
the acceptor is homopolymerizable such as AN, MMA and 4-VP, acceptor will
be incorporated in the copolymer to a greater extent than the expected
value for 2:1 copolymer, i.e. 67%.
The contribution of complex can be further demonstrated by
terpolymerizations. The terpolymerization of styrene, MAH, and
2-chloroethyl vinyl ether (ChEVE) studied by Tsuchida and Tomono35
can be explained by treating the system as a copolymerization of two
complexes, styrene-MAH and ChEVE-MAH. Butler and Campus studied the
terpolymerization of DVE-MAH-AN system. The DVE-MAH ratio in the
terpolymer was always less than 1:1 and had a lower limit of 1:2
regardless of the feed ratio of the termonomers. These results were
interpreted in terms of the participation of the CTC of DVE-MAH in the
copolymerization process with either MAH or AN.
The MAH reacted with DVE in the absence of normal radical initia-
tion to form cyclic 2:1 copolymer.3 It was postulated that initiation
via a molecular complex occurred (eq. 15).
D + A --- [D!, A7] D + A7 (Eq. 15)
The initiation of CTC can be demonstrated by photo-initiation of this
system and the DVE-FN system. Zeegers and Butler3 photo-initiated the
DVE-FN system with different wavelengths. They showed that both the
complex formed between DVE and FN and the noncomplexed species were
able to initiate the polymerization by light initiation.
Miller and Gilbert40 observed that vinylidine cyanide spontaneously
copolymerized with vinyl ethers when the two comonomers were mixed at
room temperature. Yang and Gaoni41 observed that 2,4,6-trinitrostyrene
as the acceptor monomer spontaneously copolymerized with 4-VP as the
donor monomer. Butler and Sharp42 reported the spontaneous copolymeriza-
tion of DVE and DVS.
The concerted cyclization has been argued by Butler and
Guilbault. They found that chloromaleic anhydride copolymerized
with DVE to form soluble copolymers of 1:1 composition with no residual
unsaturation. The ease with which the copolymer underwent dehydro-
halogenation indicated that the'hydrogen and chlorine atoms on the
anhydride unit are in a trans configuration as a result of a stepwise
The steric effect of highly substituted acceptors, tetrahydro-
naphthoquinone (THNQ) and dimethyltetrahydronaphthoquinone (DMTHNQ) on
the copolymerization with DVE was studied by Fujimori and Butler.43
They found that the copolymers was in constant 1:1 composition regard-
less of the feed composition. A terpolymerization of these two acceptors
with DVE was studied. Both the copolymerization and terpolymerization
and the composition can be explained by assuming that competition
between an acceptor monomer and the CTC towards the cyclized DVE radical
in the propagation step appears to favor the CTC in CTC mechanism.
These authors4 studied the steric effect of substituted MAH on the
copolymerization with DVE. They found that (i) a strong complex gave
1:1 cyclocopolymer having a copolymer backbone consisting of only DVE
units, (ii) a sterically hindered acceptor would produce 1:1 cyclocopolvmer,
and (iii) a weak CTC and reactive acceptor would produce 1:2 cyclocopolymer.
They did not mention the reactivity change of the acceptor due to the
substitution. The alternating tendency and the rate increased by using
a large amount of ZnCl2 with the DVE-FN system. A 1:2 alternating
copolymer was obtained spontaneously. This system studied by Butler
and Fujimori was consistent with the participation of a CTC in the
Solvent effects have been studied by Butler's group. The K
value (equilibrium constant of CTC formation) decreased with increase
of the dielectric constant of the solvent for MAH-DVE system. The
rate of the copolymerization and number average molecular weights
decreased in more polar solvents. In all cases, 2:1 copolymer resulted.
The study of the initial rate as a function of the feed composition
made it possible to determine the relative value of the different pro-
pagation reaction rate constants consistent with a mechanism by successive
and selective addition. However, participation of the CTC in a competing
mechanism with the above cannot be completely excluded.
In conclusion, a large portion of the evidence favors the partic-
ipation of CTC in both the initiation and the propagation steps in the
alternating cyclocopolymerization. A complete explanation or mechanism
to fit all known data has not been reported. It is reasonable to say
that the reactivity, the complexation and the steric hindrance of the
comonomers all take part in the alternating tendency, rate, and the ring
size in cyclocopolymerization. The solvent and initiator may also
determine the rate profile of this copolymerization.
Statement of Problem
The objective of the present research has been to study cyclo-
polymerization of donor 1,4-dienes with acceptor monoolefins and hope-
fully learn more about the role of charge transfer complex (CTC) formed
between the comonomers in the cyclocopolymerization mechanism and to
develop a method to elucidate the structure of the cyclocopolymer by
various nuclear megnetic resonance spectroscopic methods. The following
research was conducted with this purpose in mind.
The Structure Analysis of Cyclopolymer of Divinyl Ether-Maleic Anhydride
(DVE-MAH) Comonomer Pair.
The ir, H-nmr, and C nmr spectroscopies are employed for this well
known 1:2 alternating copolymer. The 100 MHz and 300 MHz H-nmr spectra
were investigated in order to analyze the ring size and hopefully the
cis and trans content of the bicyclic ring. The 1C nmr spectrometer
should give simpler spectra and by comparing with literature values,
it should be possible to determine the ring size of the copolymer.
A partially deuterated divinyl ether (DDVE) was prepared which should
simplify the H-nmr spectral analysis. The 100 MHz and 300 MHz H-nmr
spectra of this DDVE-MAH copolymer should give more information on the
copolymer structure and help in the assignment of the respective signals
of the spectra which has been shown to be informative based on the liter-
Rate Maximum Analysis
The study of the copolymer rate copolymerization rate as a function
of the feed composition made it possible to determine the participation
of CTC in the cyclocopolymerization. An irradiation at the wavelength
where only complex absorbed should confirm the initiation through CTC.
The participation of CTC in propagation can be supported by an analysis
of the proposed kinetic scheme. In order to compare the rate at the same
light intensity, the quantum yield was measured right after each irradia-
tion. The structure of the cyclocopolymer of divinyl ether-fumaronitrile
monomer pair was determined by elemental analysis and 13C nmr spectroscopy.
monomer pair was determined by elemental analysis and C nmr spectroscopy.
THE STRUCTURE OF COPOLYMER OF DIVINYL ETHER-MALEIC ANHYDRIDE SYSTEM
The field of cyclopolymerization has been explored extensively. A
variety of monomers containing two isolated double bonds have been found
to polymerize to form linear polymers containing cyclic units and little
or no residual unsaturation. -High degrees of cyclization are obtained
when five- or six-membered rings can form and when all double bonds have
the same reactivity such as in diallylquaternary ammonium salts and
N-substituted diallylamines.21 The polymers of N,N-diallyl amines all
contained cis and trans-substituted pyrrolidine rings with ratio 5:1.
With a substituent on the 2-position such as N-methyl-N,N-bis(2-alkylallyl)-
amines, a complex spectra showed the presence of both pyrrolidine and
piperidine rings. The 1C nmr spectra of poly(diallyldimethylammonium)
chloride showed a predominant content of five-membered ring linked
mainly in a 3,4-cis configuration. Several works have studied the polymers
obtained from N-substituted dimethacrylamides.1' 4 In most cases five-
membered ring was found predominant with a small amount or no six-
The radical cyclization reaction involved in the polymerization has
been studied by means of model compounds. Julia reviewed the works
on the cyclization of the 5-unsaturated radicals.47 Without substitution
on the radical carbon, only five-membered ring product was obtained.
With electronegative substituents on the radical carbon, the six-
membered ring product predominated.
Smith studied the cyclization of several 2-(allyloxy)ethyl radicals
with tributyltin hydride. The 2-allyloxy ethyl radical gave the tetra-
hydrofuran derivative as the only cyclic product. The 2-methyl- and
2-phenyl-allyloxyethyl radicals cyclized to give both five- and six-
membered ring products; in the latter radical the pyran derivative was
the predominant product.
Smith also found that with higher temperature the percentage of the
unfavored product increased; this indicated that the ring formation is
temperature dependent. It can be concluded that a five-membered ring is
energetically favored in the case of symmetrical non-conjugated dienes.
A steric effect would change the direction for radical attack on the
double bond. For unsymmetrical non-conjugated dienes with opposite
polarization on each double bond, the six-membered ring is predominant
as explained later. A head to tail cross-propagation has been found for
several oppositely polarized vinyl monomer pairs, such as maleic anhydride-
vinyl ether pairs, which introduced an alternating copolymer. The
cyclocopolymerization of DVE-MAH is similar to unsymmetrical non-conjugated
dienes, in which the cyclization step involves a favorable cross-propaga-
tion reaction of oppositely polarized units.
It seemed worthwhile, therefore, to investigate the structure of the
copolymer of DVE-MAH and to study the temperature dependence of the five-
and six-membered ring distribution in the copolymer.
Results and Discussion
Synthesis and Copolymerization of Bis(2,2-dideuteriovinyl)ether
Due to the small amount of materials available, the structures of
the products in each synthesis step were determined by ir and nmr spectro-
scopy. A comparison of these spectra with spectra reported in the
literature and the spectra of the corresponding non-deuterated products
prepared by the same procedure confirmed the structures of the deuterated
compounds. In all cases the ir and nmr spectra of the non-deuterated
products were exactly the same as reported in the literature. Therefore,
the procedure for preparation of the non-deuterated ompounds were
applicable for the deuterated analogues.
A reaction scheme is shown in the following route:
O LiAlD /
P D\ Br
/' \D KOH -D PBr3
0 -- 0 D 3
S D Triethanol ,D Pyridine
The 2,2,2',2'-tetradeuteriodiethylene glycol was prepared essentially
according to the method given by Bloomfield and Lee with only minor
modification.50 For complete reduction of anhydride, a prolonged reflux
period was required. In order to take the most advantage of lithium
aluminum tetradeuteride (LiAlD ), only a little excess of the deuteride
was used. The resulting glycol was soluble in water to a large extent,
hence, to isolate it from water solution was difficult. Even a salting
out process did not succeed. The reduction in ether solution was not
successful because of the low solubility of the anhydride. The reaction
seemed not to go at all. The best result was by using tetrahydrofuran
(THF) as solvent, with a small amount of water to destroy the aluminum
salt and release the glycol. The addition of 9 N sulfuric acid to dissolve
the aluminum salt did not improve the yield significantly. Therefore,
the complete dissolution of aluminum salt was not necessary. Water, fol-
lowed by dilute acid was used to bring the glycol into THF solution.
The addition of excess anhydrous potassium carbonate, K2C02, neutralized
the acid by evolution of carbon dioxide and absorbed water present in
the THF solution.
The diethylene glycol prepared by the same procedure showed exactly
the same ir and nmr spectra as reported in the literature. The ir
spectrum of 2,2,2',2'-tetradeuteriodiethylene glycol showeJ two
absorptions at 2220 and 2110 cm of C-D stretching. The structure was
further confirmed by the nmr spectrum, in which a singlet was observed
instead of the multiple in the spectrum of non-deuterated diethylene
glycol. The peak ratio of 1:2 instead of 1:4 (for diethylene glycol) also
indicated that the product obtained was tetradeuteriodiethylene glycol.
The bromination of deuterated ethylene glycol was straightforward.51
The same method when applied to non-deuterated ethylene glycol gave a
product with exactly the same nmr and ir spectra as reported in the
literature. The ir spectrum of the resulting liquid for bis(2-bromo-
2,2-dideuterioethyl)ether showed an absorption at 2170 cm which was
assigned to the C-D stretching. The nmr spectrum clearly confirmed the
structure with a singlet at 6 3.80 ppm. On the contrary, non-deuterated
dibromoethyl ether showed an AA'BB' multiple at 6 3.70 ppm.
Finally the synthesis of bis(2,2-dideuteriovinyl)ether was prepared
by dehydrobromination from the corresponding dibromo compound. Only
the nmr spectrum was analyzed. The disappearance of ABX system which
showed up in the spectrum of divinyl ether, was evidence of the replace-
ment of the four terminal hydrogens by deuteriums. A clear spectrum
was obtained by using larger sweepwidth. The constants for the hydrogen-
deuterium coupling (JHD) was obtained by analyzing this spectrum.
By multiplying JHD by 6.5, the corresponding hydrogen coupling constants
(JHH) were obtained and were found to be close to the reported value.
The Comparison Between JHD and JHH
Coupling J (Hz) Multiplied Reported
by 6.5 JHH (Hz)
trans 2 13.0 13.8
cis 1 6.5 6.2
The synthesized terminal deuterated divinyl ether (DDVE) was
copolymerized with MAH at 720C in cyclohexanone by AIBN initiation.
The polymeric product was isolated and purified as a white powdery
solid which was soluble in acetone and dimethyl sulfoxide.
A series of copolymers of DVE and MAH were prepared at different
temperatures. The ir and nmr spectra were obtained and are discussed in
the next section.
The Structure of Copolymer
Infrared (ir) spectra. The ir spectra were shown in Fig. 1. The
two strong peaks at 1775 and 1855 cm1 correspond to the reported
absorption of succinic anhydride at 1782 and 1865 cm for symmetric
and antisymmetric carbonyl stretching, respectively. The strong peaks
at 1230 and 950-920 cm were assigned for the C-O-C absorptions for
cyclic anhydride unit. The strong peak at 1090 cm-1 with a shoulder
between 1060-1020 cm was assigned for the C-O-C stretching for pyran
structure. The five-membered ring structure was ruled out by the fact
that the C-O-C stretching absorption for tetrahydrofuran is at 1062 cm
The structure of the copolymer shown as structure X, with a 2,6-disubsti-
tuted tetrahydropyran ring and an anhydride unit on the 3,4-positions.
The 1:2 composition of DVE to MAH has been reported by Butler based on
the facts discussed in Chapter I. The spectra for 2,5-disubstituted
tetrahydrofuran has been reported by Mihailovic,et al.5 They observed
strong absorption at 1100 cm .for both cis and trans-2,5-dimethyl-
The structure of the copolymer supported by ir spectra is likely
to have a six-membered ring. A conclusive result cannot be reached
because a suitable model compound for the comparison of ir spectra is
Hydrogen nuclear magnetic resonance (H-nmr) spectra.
The H-nmr spectra of polymers are usually broad. Therefore,
structure determination by H-nmr spectra is difficult. With the help
of the spectra of deuterated compounds and high resolution nmr spectro-
scopy, it is possible to simplify the spectra and separate the overlapping
peaks. The 60 MHz nmr spectrum for DVE-MAH system copolymer was shown
in Fig. 2a. A four peak pattern was observed at 6 4.49, 4.06, 3.47 and
2.31 ppm with an area ratio of 1:1:1:2. The H-nmr spectrum of the
copolymer prepared with DDVE and MAH under the same polymerization
conditions exhibited the disappearance of the strong peak at 6 2.31 ppm
(Fig. 2b). Only the weak peaks corresponding to the residual solvent
were observed. Hence, the peak in the spectrum of nondeuterated copolymer
at 6 2.31 ppm was assigned for methylene protons which were B to the
ring oxygen. It has been reported that in the spectrum of the copolymer
prepared with DVE-3,4-dideuteromaleic anhydride (DMAH), the two peaks
centered at 6 3.47 and 6 4.06 ppm disappeared.5 Hence, these two
peaks are due to the four methine protons linked on the anhydride forms.
The peak at lower field was assigned to the proton on the rir., based on
the fact that the proton in a fixed position within a bicyclic ring will
experience a downfield shift caused by the electronegative oxygen nearby.
The methine protons next to oxygen are expected to experience a
deshielding effect to shift tb lower field at 6 4.49 ppm. The methine
protons next to oxygen have been reported to absorb between 6 4.00 and
6 3.30 ppm.57 Strong electron withdrawing effect of the two succinic
anhydride groups apparently shifts the methine proton to even lower field.
According to this analysis based on peak assignment and area inter-
gration, it can be concluded that the repeating unit of the copolymer has
two to one ratio of maleic anhydride to DVE which are arranged in an
exactly alternating manner. A random distribution of bicyclic and succinic
The Chemical Shifts for the Methine Protons of DVE-MAH Copolymer
Spectraa Next to 0 On the Ring On the Backbone
60 MHzb 4.49 4.06 3.47
60 MHzc 4.38-4.44 4.08 3.50
100 MHzc 4.18-4.44 3.88 2.96, 3.30
300 MHzb 4.29, 4.57 3.92, 4.14 3.01, 3.30
300 MHzd 4.34, 4.46
A6e 0.22 0.29-0.34
All chemical shifts (6) are in terms of ppm relative to the
cDeuterated copolymer of DDVE and MAH.
Deuterated copolymer of DVE and DMAH.
The chemical shift differences between the assigned cis and
trans methine protons.
See reference 59.
anhydride units would give a more complex spectrum than the one reported
The 60 MHz H-nmr spectrum is not able to distinguish between the
five- and six-membered ring due to the broad peaks caused by vicinal
protons and the different conformations and configurations.
For the purpose of distinguishing the cis and trans disubstitutions
and the five- and six-membered ring structures, a high resolution 100 MHz
H-nmr spectrum was performed for the copolymer of DDVE-MAH in acetone-d6
at 500C (Fig. 3). The broad peak at 6 3.50 ppm with a shoulder at 6 3.16
ppm in 60 MHz spectrum was separated into peaks at 6 3.30 and 2.96 ppm.
The shoulder between 6 4.38 and 4.44 ppm was still unseparated and broad.
A 300 MHz nmr spectrum of DVE-MAH copolymer (Fig 4a) was performed by
Butler's group, in which a further separated pattern was observed.58
The three peaks for methine proton groups are summarized in Table V.
Bode and Brockmann reported that the cis- and trans-2,3-disubsti-
tuted succinic anhydride showed 0.47-0.28 ppm difference in chemical
shift and the difference decreased with larger substitution.59 Therefore,
we can assign the methine groups of highest field with 0.34-0.29 ppm
splitting as the anhydride unit in the copolymer backbone. The one at
the lower field with less splitting (0.22 ppm) is then assigned as the
anhydride unit on the ring, which is deshielded by the neighboring pyran-
oxygen as discussed in the last section. In disubstituted succinic
anhydride, the chemical shift for the trans form is less than that for
cis form, 6 >6 t thus the four peaks in the MHz spectra for anhydride
protons can be assigned as: 3.01, 3.20, 3.92 and 4.14 for the protons on
trans-backbone, cis-backbone, trans-ring and cis-ring anhydride units,
respectively. On the backbone, the population of cis anhydride units
o3 c a
0~ 0 N
and the trans anhydride units are more or less the same. In contrast, more
trans form was present in the ring. The latter can be explained by the
possible bicyclic ring conformation, where a trans configuration is more
favorable. Butler and Guilbault prepared the copolymer of chloromaleic
anhydride with DVE having cyclized 1:1 composition with only DVE units
in the backbone.3 They investigated dehydrohalogenation of the copolymer.
From the ease with which the copolymer underwent dehydrohalogenation, they
suggested that the hydrogen and chlorine atoms on the anhydride unit in
the ring were in a trans configuration. However, H-nmr spectra showed
the presence of both cis and trans conf1,irations with the latter in
By examination of the model of the bicyclic ring with large
substitutions on the carbons next to the oxygen, the six-membered ring
with chair form and trans junction is more favorable (trans isomer). In
this structure (XI), the two protons at the junction are in trans configura-
tion. A chair form with cis junction is another alternative configuration
(cis isomer). However, the latter configuration (XII) experiences more
ring strain. On the contrary, a bicyclic ring with two five-membered
rings experiences much strain and only cis junction is possible, especially
when one of the rings is an anhydride unit (XIII). The heat of combustion
of trans bicyclo(3.3.0)octane is greater than the cis isomer by ca.
6 Kcal/mole.6 [hu:., a trans junction pyran bicyclic structure with the
presence of cis isomer explains the analysis of 300 MHz nmr spectrum of
the DVE-MAH copolymer, based on the analysis of methine protons on the
For a disbustituted cyclic ether, the chemical shift difference
between the cis and trans of the methine protons next to oxygen in the
literature have been reported to be between 0.11 and 0.51 ppm.6 62,
In the case of the DVE-MAH copolymer, the methine proton signal was
split into two peaks by 0.28 ppm as shown on the 300 MHz spectrum. It
is reasonable to assign this group of peaks as a mixture of the two
configurations with more trans-isomer present because of the larger
area in the higher field corresponding to the cis form which was more
populated for a favorable trans isomer.
cis isomer 0 0
This splitting pattern was more clearly observed in the 300 MHz
spectrum for the copolymer of DVE and DMAH (Fig. 4b). The two separated
methine proton (next to oxygen) peaks were further split into doublets.
The doublet indicated the nonequivalency of the two protons on each side
of the oxygen in the pyran ring. In contrast, the two protons for the
five-membered ring structure are more or less equivalent.
The methylene peaks on the higher field can be analyzed as a
mixture of two doublets; the larger doublet is assigned for trans isomer
which if further split by the fact that the methylene proton groups are not
equivalent. The splitting peak at the lower field with less intensity can
be assigned to the less populated cis isomer. In fused five-membered ring
structure, it is possible that the two methylene groups are linked in
either cis or trans configurations.63 However, a trans configuration
would introduce some strain in the bicyclic system with the two bulky sub-
stitutions on the pseudo equatorial positions of the half-chair conforma-
tion.6 A cis configuration with the possible conformation is shown in
structure XIII.64 In this structure both the two methine hydrogens and
the two methylene groups on the 2- and 5-position are equivalent. There-
fore, the splitting on Fig. 4b is most likely due to a chair form six-
membered ring with trans junction (trans isomer) with some cis isomer.
Further confirmation could be obtained by using bicyclic model compounds
and/or studying copolymers of additional monomers, e.g., CH2=CDOCD=CH2.
C3 Nmr Spectra
The line broadening inherent in H-nmr has severely limited the poten-
tial of this technique for polymer structure analysis. In contrast, the
extended chemical shift range, the absence of significant dipolar line
broadening and the simplicity of proton decoupled 1C nmr spectra, in
which each carbon resonates as a singlet, makes these spectra extremely
useful for the study of polymers.
The spectrum of the copolymer of DVE-MAH was shown in Fig. 5a.
The spectrum consists of five major peak groups. From comparison with
published data the following general assignments can be made: the broad
peaks centered at 6 31.4 ppm with a shoulder at 6 35.8 ppm for methylene
carbon, the broad peak centered at 6 44.2 and 51.5 ppm with a shoulder
at 6 53.7 ppm for the carbon adjacent to the carbonyl groups, broad peak
; 1 ^
centered at 6 78.7 ppm for methine carbon adjacent to the oxygen, the peaks
at 6 171.7 and 174.3 ppm for carbonyl carbons (Fig. 5a). The spectrum for
the copolymer hydrolyzed in D 0 does not change the pattern, but shifts
the peak to lower field (Fig. 5c). A comparison between the hydrolyzed
copolymer and the diacid model compound can be used for the assignment
In a large series of compounds containing carbonyl groups, the
shielding is mainly influenced by the local electrons on the carbons.
The two carbonyl peaks Lui,.~-t that two types of carbonyl groups are
present corresponding to the two anhydride carbonyls proposed by H-nmr
apectra, where one of the anhydride units is in the backbone and the other
forms the bicyclic ring. The two carbonyl carbons in each anhydride unit
are not expected to be greatly different from each other as far as the local
electron density is concerned. Therefore, the broadening of the peaks can
be explained either by the two different carbonyl carbons in each anhydride
unit or by the mixing of both cis and trans anhydride forms. A broader
peak is observed at the lower field, which may indicate the equal mixing
of both the cis and trans forms.
The assignment of each carbonyl peak can be made by considering the
chemical shift difference between cis and trans configurations (A cis a),
which would broaden the carbonyl peaks. The carbonyl carbon absorptions
of succinic anhydrides have been reported to be between 6 171.7 and 175.3
ppm. The C nmr spectrum for a dl-meso mixture of 2,3-dimethyl succinic
acid is shown in Fig. 6. The chemical shift differences between the
two carbonyl carbons, the two methine carbons and the two methyl carbons
are 0.7, 0.9 and 1.4 ppm respectively. This small chemical shift differ-
ence for carbonyl carbons between the isomers of dimethyl succinic anhydride
is expected for the carbonyl carbon peaks for the copolymer where only
a broadening effect is observed. The peak at the lowest field can be
thus assigned for the carbonyl carbons on the backbone because it is
broader than the other carbonyl carbon peak. This is reasonable because
the sharper peak then assigned for the anhydride on the ring, can be
explained by the less population of cis form. This sharper peak is
broadened by the hydrolysis of the copolymer as expected by the loss
of the rigidity of the bicyclic structure.
A comparison between the spectrum of the copolymer and poly(maleic
anhydride)(Fig. 5b)8 showed clearly that the broader carbonyl carbon peak
can be assigned to the anhydride unit on the backbone, and also, the methine
carbon peak at the higher field is for the anhydride unit on the backbone.
On the other hand, the two peaks at 51.5 and 53.7 ppm with a difference of
2.2 ppm can be assigned for the two non-equivalent methine carbons on the
six membered ring by comparing with the $- and y-carbons on the tetra-
hydropyran which has a difference of 2.8 ppm.
The peaks at the lower field of the methine carbon region (6 51.5-
55.5 ppm) are then assigned to the methine carbons on the ring anhydride
unit; the larger peak from trans isomer and the weak peak from the smaller
contribution of cis isomer. The methylene carbon peak is assigned to
the trans isomer which is broadened by the non-equivalency of the two
methylene groups in the copolymer. The shoulder at 6 35.8 ppm can be
assigned for the methylene carbons of the less populated cis isomer.
By comparing the methine carbons of the copolymer with poly-
(diallyldimethylammonium)chloride, (XIV), 1,1,3,5-tetramethylpiperidinium
iodide, (XV), and 1,1,3,4-tetramethylpyrrolidinium iodide, (XVI), in
Table VI, the chemical shift difference between the cis- and trans-
methine carbons in the copolymer is closer to the six-membered ring
structure than the five-membered ring structure.
Chemical Shits Differences Between cis and trans Disubstituted
Vincinal Carbons in 13C nmr Spectra
Structure A6cis-trans (ppm)
On the Backbone On the Ring
Kunitake and Tsukino66 suggested an all five-membered ring structure
for the copolymer of DVE-MAH system by the fact that a highly symmetric
structure is involved, the two singlets observed for the carbonyl carbons
and the comparison of the estimated chemical shift values to the
experimental values. This suggestion can be argued against by (1)
in the spectrum of the copolymer, the difference between the two
carbonyl carbons in each anhydride unit is so small that a splitting
causes only the broadening of the peaks. Hence, the absence of the
doublet for each peak cannot be explained as the absence of the non-
equivalent carbonyl carbons, such as those in the six-membered ring
in the copolymer. (2) The analysis of H-nmr spectra was best represent-
ed by six-membered ring copolymer structure, although the unlikely
mixing of both five- and six-membered ring structures cannot be ruled
The Temperature Dependence on the Copolymer Structure
Due to the possibility of the mixing of both six- and five-
membered ring structures in the copolymer, temperature effect experi-
ments were carried out to investigate the temperature dependence on the
contribution of these two structures.
The copolymers were prepared at different temperatures, 250, 720,
1000, and 1300C. The results are shown in the experimental section. The
spectra for each copolymer are shown in Fig. 7. No major change was
observed. The small side peak a.t 6 35.8 and 51.5 ppm both gradually
disappeared or flattened with the decreasing of temperature. This small
change cannot be considered to be real because of the accidental experi-
mental errors such as lower resolution and the different concentrations of
the samples. Even the different solvents used for preparation should be
It is reasonable to conclude that no significant change of structure
caused by the temperature effect is obtained. This fact can be explained
-co o---/- 0
/ H-CH CH
171.7 78.7 51.544.2 29.2
Fig. 7 1C nuclear magnetic resonance spectra of DVE
MAH copolymer prepared at (a) 130, (b) 100, (c) 72,
(d) 25 OC
by the production of only one energetically favored product which is
the six-membered ring structure as revealed by all the spectroscopic
The Mechanism of Cyclization
The structure and temperature analysis concluded that a highly
energetically favored six-membered ring copolymer was the product of
the copolymerization of DVE-MAH system. In cyclization of the sym-
metric non-conjugated dienes, the ring size is controlled by the
entropy effect of the two ring formation processes. A five-membered
ring product is favored because of the less entropy change involved.
With substitution, a sterically favored six-membered ring formation
is able to compete with the former process. Also, the substitution on
the radical carbon of the uncyclized radical makes both the processes
higher activiation energies and hence, less selective. At the extreme,
with two electronegative and steric substituents on the radical carbon,
a six-membered ring formation will be highly favored.47
The energetical factor is much different for a cyclization involving
non-symmetric non-conjugated dienes such as the cyclization in the
cyclocopolymerization of DVE-MAH system. During the cyclization step,
the vinyloxy double bond is attacked by a radical on the anhydride unit:
XVII / N'/-
This process can be shown to be highly enthalpy-controlled by
considering the closer energy gap between the HOMO of vinyloxy double
bond and the singly occupied molecular orbital (SOMO) of the anhydride
radical than that for the corresponding symmetrical diene cyclization.
HOMO radical destablized
double double bond
The HOMO of the vinyloxy double bond is polarized to have higher
orbital density on the terminal position.6 Therefore, a fast radical
addition on the terminal carbon of the double bond leads to a six-
membered ring radical. A copolymer with six-membered ring structure X
is thus obtained.
As discussed in chapter I, a charge transfer complex was proposed to
explain the fast cyclization. The charge transfer complex can be applied
here with the help of HOMO-LUMO concept to predict the ring structure
of the cyclization step.
The Milliken theory of overlapping and orientation principle predicts
that stabilization in the molecular complex formation should essentially
be determined by the overlap of the donor HOMO and the acceptor LUMO.68
In the examination of ir and Raman spectra of DVE, Claugue and Danti
proposed the presence of two rotational isomers.9 The more stable isomer
has C symmetry, in which the two vinyl groups, although coplanar are
non-equivalent. Hirose and Curl examined the microwave spectrum and
assigned the C conformer. They found a small nonplanarity caused
by H-H repulsion between the 3-hydrogen of the cis vinyl group and the
a-hydrogen of the trans vinyl group-(XVIII).
The charge distributions in vinyl ether and vinyl methyl ether were
calculated by CNDO/2 method by Fueno, et al.67 It was found that a large
electron density was on the terminal position as in sturctures XIX and
This charge density of orbital actually describes the orbital density of
the HOMO of DVE. The LUMO of MAH has been described by Fukui as
structure XXI with higher orbital densities on the double bond carbons,
but antisymmetrical to the plane of symmetry of this compound.71
Therefore the most stable conformation for a DVE-MAH complex can
be expected as XXII, based on the conformer structure of DVE and the
molecular densities of both.comonomers.
0 path a
R* path b
.ii-n this complex is initiated by a radical, a six-membered ring
radical will be formed concerted (Path a) or stepwise through an anhydride
radical addition on the terminal carbon of the vinyloxy unit (path b).
This complexation would reduce the energy gap between the complex and
the propagating anhydride radical, thus, a radical addition on the
complex occurred and the reaction is supposed to be fast. This special
complexation and/or interaction would significantly reduce the activa-
tion enthalpy for the formation of six-membered ring. In the range
of the temperature sutdied, a five-membered ring formation cannot compete
with it at all, which explains the temperature independence on the
structure of the cyclocopolymerization.
In conclusion, on the mechanism of the cyclization and copolymer-
ization of DVE-MAH system, it is reasonable to be stated as follows.
(1) The intramolecular cyclization is favored over the intermolecular
addition due to the lower entropy change of the former process than the
latter one. This explains the high degree of cyclization.
(2) The entropy preference cannot be explained on the base of activation
energies and the statistical probability. A preorientation either
through the delocalization of the radical with the intramolecular double
bond or the formation of complex is proposed.
(3) This preorientation would lead to a six-membered ring structure
by a favorable energy factor based on the !IlH',) orbital density of DVE.
For a symmetrical nonconjugated diene the five-membered ring cyclization
is favored by the entropy factor.
(4) A faster rate of this cyclocopolymerization than the copolymerization
of the corresponding monoolefin -pairs can be explained by the closer
energy of the anhydride radical to the complex.
The proposed cyclization mechanism can be applied on other comonomer
pairs and is worthy of further investigation.
THE COPOLYMERIZATION OF DIVINYL ETHER-FUMARONITRILE
It was pointed out in Chapter 1 that a donor-acceptor pair of
comonomers could produce alternating copolymer through a charge transfer
complex (CTC). In the cyclocopolymerization of donor 1,4-dienes with
acceptor monoolefins, alternating cyclocopolymers having 1:2 composi-
tion were obtained for several systems. The participation of CTC formed
between the donor and acceptor was proposed to explain the alternating
copolymerization. However, it has been known for sometime that the
compositions for the copolymers are 1:2 while the stoichiometry of the
CTC are always 1:1. An alternating copolymerization of a CTC and a free
monoolefin was proposed as an explanation. It has been shown that
when an acceptor monoolefin is highly sterically hindered and hence much
less reactive, according to the above explanation, a 1:1 cyclocopolymer
was obtained apparently through the homopolymerization of CTC., 43
With a less sterically hindered monoolefin and less reactive monoolefin
it is possible to form a cyclocopolymer with composition between 1:1
and 1:2. The cyclocopolymer of DVE-FN has been reported having the FN
content between 0.55 and 0.63 mole fraction which is in the range of
0.50-0.67 for the 1:1 and 1:2 composition.32 With dilution, less FN
content was reported and a contribution of either homopolymer structure
of DVE and/or 1:1 comonomer unit in addition to the regular 1:2
comnomer unit was proposed as an explanation.37
comonomer unit was proposed as an explanation.
A further structural analysis and the study of the participation of
CTC in the copolymerization are discussed in this chapter
Results and Discussion
Study of the DVE-FN complex in Acetonitrile
On the basis that the complex formed between DVE and FN may
participate in the mechanism of initiation of the photocopolymerization
in acetonitrile solution, the characteristics of the complex were
studied in the same solvent. The existence of the complex was estab-
lished by UV spectrophotometry. A mixture of 0.6 m/l of DVE and 0.6 m/1
of FN showed a large absorption between 250 nm and 350 nm (Fig. 8)
although a distinguishable new absorption band was not observed.
This large enhancement of absorption indicated the presence of complex.
A 1:1 stoichiometry was determined by the continuous variation method
at 300 nm- (Fig. 9). The maximum of the absorption for different
compositions of DVE and FN, while their total concentration was kept
constant was found for equimolar composition.
The charge transfer complex of an acceptor-donor pair is in
equilibrium with the free components. The charge transfer complex exists
in resonance between the no-bond state and the dative state; thus the
wave function of the charge'transfer complex (TCT) can be expressed as a
linear combination of wave functions of the no bond state [Y(D,A)] and
the dative [W(A*, DT)] (Equation 1 and 2).
A + D --- [(A,D) + (A*, D.)] (Eq. 17)
CT = a'-(A,D) + b.'(A7 DI)
- O >J o
i O O o 0
C X II
-H 4 -)
m r'- o
For a regular loose complex, a2>>b2 in the ground state of the com-
plex. The dative structure corresponds to an ionic-radical-like pair.
There must be also an excited state (41CT) which can be called a charge
transfer state given by
CT = b o'(A,D) + a Y,(AD,D*) (Eq. 19)
The excited state is mostly dative (a 2>>b ); excitation of an electron
from TCT to 'CT essentially amounts to the transfer of an electron from
donor to acceptor. Spectroscopic absorption would occur with this excita-
tion (charge transfer absorption). A charge transfer absorption is
possible for any pair of molecules if in contact, even if they do not
form a stable complex.
A complete absorption spectrum of a complex consists of absorption
to (1) locally excited state (states of donor or of acceptor, more or
less but usually not greatly modified in the complex.) (2) charge
transfer states [(YCT in Eq. 19, and other charge transfer states includ-
ing the excited dative structures, for example 4(D* ,A*)].
The equilibrium constant (K ) of the complex can be measured by
using Merrifield and Phillips method.74
(D) -K A + K (A)c (Eq. 20)
(D)o c Ac c ( Ac
(D)o = The initial donor concentration
(A)o = The initial acceptor concentration
Ac = Absorbance of complex at certain wavelength
K = The equilibrium constant for the formation of weak complex
hc = The extinction coefficient of the complex at wavelength.
For a series of solutions containing different concentrations of
DVE but with constant concentration of FN, with condition (DVE)>>(FN),
a plot Ac/(DVE) against Ac should be linear. From the gradient of the
line, K may be evaluated directly without recourse to an extrapolated
intercept. The absorption of DVE-FN complex and the resulting plot in
acetonitrile are shown in Fig. 10 and Table VII. The equilibrium con-
stant was small (K = 0.10) and cannot be evaluated exactly, but it is
compatible with the equilibrium constant measured in methanol solution
(K = 0.12 to 0.20).
Determination of Equilibrium Constant of rEN-DVE in
Acetonitrile with Constant FN Concentration (0.00101 m/1)
(D)0 280 nm 300 nm
m/c A A/(D)0 A A/(D)0
0.846 0.031 0.0520 0.013 0.0154
1.69 0.081 0.0479 0.029 0.0172
2.12 0.100 0.0472 0.031 0.0146
2.96 0.137 0.0463 0.042 0.0142
3.81 0.152 0.0399 0.040 0.0105
K 0.09 0.11
The equilibrium constant of complexation can be determined by nmr
spectroscopy using the Hanna-Ashbaugh equation.75 The attempted nmr
method failed because one of the quartet absorptions of the a-vinyl
L_ (NJ CH
0 C C C) C
0 ) 0
proton of DVE covered the peak of the protons of FN, whose chemical shift
was to be used to determine the equilibrium constant in acetonitrile.
The Structure of the Complex
The spectroscopic determination of stoichiometry and the equilibrium
constant of the charge transfer complexes of monoolefins with 1,4-dienes
have been thoroughly discussed previously.37 'i lr ly, in all cases
studied the stoichiometry is 1:1. The structure of the complex has not
been established. Either one or both of the double bonds of the 1,4-diene
can be completed with acceptor. Take DVE and FN as an example:
A, / 1:1 complex
C-N, CN CN -- -CN
The first structure is not likely because with a free double bond
available, a second acceptor would be completed more or less as easily
as the first acceptor molecule to form a 1:2 complex.
CN 1:2 complex
As pointed out in Chapter II, considering the conformation of DVE
and the orbital densities of both HOMO of DVE and LUMO of acceptor, the
most stable conformation of the complex can be predicted as structure X:III.
The proposed relation between no bond state and the dative state of this
complex is shown in equation 21.
.--/-N "(Eq. 21)
+ 0 __
Dative state No bond state
At ground state the complex can be represented by the no bond state,
the excited state (CTC)* after the absorption of appropriate light energy,
can be represented by dative state. In general an excited state would be
radiationlessly deactivated without associating with external environment.
In the case of (CTC)*, the state is so polar that the environmental polar
molecules (acceptor, donor and ground state complex) would participate in
its deactivation processes.
At 236 nm, while most of the light is absorbed by the free monomers
to form excited monomers, the excited state of DVE would interact with
the ground state of FN to form an exciplex which possesses higher energy
than the previously mentioned c rge transfer state [(CTC)* formed by
irradiation at 300 nm where only the complex absorbs].76 The exciplex is
polar enough to be deactivated with the participation of ground state
molecules. (Scheme I)
300 nm environmental
(DVE -- FN) -- -- (DVE -- FN)* en DVE + FN
236 nm FN environmental
DVE n DVE* -F-- (DVE**FN) Dr -ome DVE + FN
The Structure of the Copolymer of DVE-FN System
Ir and H-nmr spectra. The ir, H-nmr and 1C nmr spectra were record-
ed for the copolymers prepared by irradiation at 300 nm, 236 nm and with
AIBN, in acetonitrile with feed composition fd = 0.5 (Table VIII).
Comparison of Copolymers Initiated by Different
Methods in Acetonitrile
Feed Composition Initiation Conversion md'
0.6 m/1 0.6 m/1 AIBN 44% 0.408
0.6 m/1 0.6 m/l 236 nm 40% 0.402
0.6 m/1 0.6 m/1 300 nm 55% 0.486
*Molar fraction of DVE in copolymer calculated from nitrogen
and carbon content
All the copolymers showed the same characteristics in ir and H-nmr
spectra, thus indicating that the same propagation processes are employed
for all methods of initiation.
The copolymers absorbed in the infrared region: 2250 cm- (s)
(CN stretching) and near 1100 cm-1 (broad)(ether group), showing the
existence of the comonomer units. There was also the presence of
vinyloxy double bond absorption at 1630 cm It has been observed32
that only a small amount of residual unsaturation (from 2.0 to 3.5%) was
in the copolymer prepared by the initiation of AIBN in dimethylformamide.
The infrared spectra of copolymers initiated by light at both 236 nm
and 300 nm had the same characteristics as the one initiated by AIBN in
acetonitrile (Fig. 11). It is reasonable to assume that the same small
amount of unsaturation was in the photocopolymers. This conclusion was
also shown in the H-nmr spectra (Fig. 12) where no absorption contributed
by vinyloxy double bond was observed. Therefore, there is no significant
contribution of the structures with the pendant vinyl group in the copoly-
\ CN / n
The composition of the copolymer. The composition of the copolymers
has been determined over a wide range of monomer feed compositions. The
copolymers of DVE-FN are quite hygroscopic. The elemental analysis showed
higher hydrogen and oxygen weight p-r>c:nrtjr, than calculated from nitrogen
Wave number (cm- )
Fig. 11 IR spectra of the DVE-FN copolymer initiated
(a) at 300 nm, (b) at 236 nm, (c) by AIBN.
7.0 6.0 5.0 4.0 3.0 2.0 l.O 0
Fig. 12 60 MHz nuclear magnetic resonance
spectra of the DVE-FN copolymer initiated
(a) at 300 nm, (b) at 236 nm, (c) by AIBN
and carbon. The nitrogen content is only from FN monomer, the calculation
of the number of nitrogen atoms permitted the determination of the
no. of moles of FN = no. of moles of nitrogen atoms/2
Sno. of moles of carbon atoms no. of moles
no. of moles of DVE = o F
of FN x 4
md = the molar fraction of DVE in the copolymer
no. of moles of FN
no. of moles of DVE + no. of moles of FN
The results are shown in Table IX and Fig. 13.
The Compositions of Copolymers Prepared in Acetonitrile at
Room Temperature within 10% Conversion
Concentration (m/l) 300 mn 236 nm
(FN) (DVE) fd md m
0.4 1.6 0.80 0.410 0.408
0.8 1.2 0.60 0.372 0.436
0.6 0.6 0.50 0.486a 0.402a
1.2 0.8 0.40 0.412 0.400
1.6 0.4 0.20 0.351 0.436
Average 0.406 0.416
More than 40% conversion.
Molar fraction of DVE in feed.
Molar fraction of DVE in copolymer calculated from
nitrogen and carbon weight percentage.
I I I
0 0 0 0
There was no apparent trend for compositions of copolymer. The
copolymers irradiated at different wavelengths have similar compositions
(average md = 0.406-0.416). It indicated that the same propagation process
was employed for both wavelengths. For a typical 1:2 copolymer and a 1:1
copolymer, md is 0.33 and 0.50 respectively. The compositions of both
copolymers fell in the range from 0.351 to 0.486 within a wide range of
feed compositions. Together with the spectroscopic data this is the basis
of assuming a copolymer containing both the 1:1 and 1:2 copolymer structures
XXIV. Note that is is not necessarily a block copolymer in the sense of
1:1 and 1:2 monomer combinations. The 1:1 repeating unit has been found
in the copolymerization of p-dioxene-MAH,77 DVE-DMTHNQ and DVE-THNQ
systems, in which the homopolymerization of a 1:1 CTC was considered.
S CN x
When the reaction time of copolymerization was long enough so that
the reaction was almost completed, it was observed that the yield at 236 nm
for large excess of DVE in feed composition was more than the corresponding
theoretical maximum of conversion of a perfect 1:2 copolymer but less than
or close to the 1:1 copolymer (Table X). At 300 nm the yield was less
than the theoretical 1:2 copolymer.
The excess yield at 236 nm can be explained by the involvement of
either the homopolymerized DVE or the 1:1 copolymer structure in the 1:2
overall copolymer structure. When FN is about used up at the end of the
reaction, homopolymerization of DVE has a chance to compete with the
copolymerization. The fraction of DVE in copolymer (md) supports the
involvement of homopolymerization of DVE. What is important is that the
close value of nd at both 300 nm and 236 nm (md = 0.560.03) indicating
that the involvement of homopolymerization is in the propagation process
instead of the initiation process, in other words, in the early stage
of the copolymerization, the homopolymerization is not involved.
The Limiting Yield of Copolymerization with
Excess DVE in Feed Composition
1:2 Copolymer 1:1
The fraction of DVE in copolymer calculated from nitrogen and carbon
Based on the reacted amount of FN.
Average = 0.56.
The different initiating wavelength would finally form the same
intermediate or intermediates which initiates the same propagating radicals.
The structure of this copolymer can be further confirmed by 1C nmr spectro-
The 1C nmr spectra. The 1C nmr spectrum of the copolymer is
shown in Fig. 14. The four peaks at 6 118.5, 117.9, 116.6, and 116.1
ppm from TMS in the CN region supports the six-membered ring structure
where two of them correspond to the two unsymmetrical CN's on the ring
and the other two correspond to the CN's on the skeleton.
It was pointed out in Chapter II that the small difference between
the two carbonyl carbons would only cause a broadening of the peak
instead of splitting it. This assumption can be applied here that the
doublet of each CN peak does not indicate two non-equivalent CN's in
the ring or on the backbone. The chemical shift difference (0.5-0.6 ppm)
between each peak in the doublet is close to the difference of the carbonyl
peaks of dl-meso mixture of 2,3-dimethyl succinic acid (0.9 ppm). We
can assign the doublet as the consequence of the mixture of cis- and trans-
dicyano substitutions on each fumaronitrile unit. These assignments will
lead to a reasonable conclusion that in backbone, cis- and trans-disubstitu-
tions are more or less equally populated, by considering the equal
intensities of the doublets as shown on the peaks at 6 117.9 and 118.5 ppm.
The CN's at 6 116.6 and 116.1 ppm indicated the different populations of
cis- and trans-disubstitutions, whichare possible during the ring formation.
It is difficult to assign the chemical chifts of the CN groups
without comparing with model compounds, but at least a copolymer with
the 1:1 and 1:2 copolymer composition is consistent with the spectra.
The two C-O-C peaks in the C-O-C region indicated the existence of
two different C-O-C linkages in the copolymer. Comparing the 1C nmr
spectra of homopolymers of DVE with the copolymers enabled us to clarify
e exclusion of homopolymer structure in the copolyr.78 (Table XI)
the exclusion of homopolymer structure in the copolymer. (Table XI)
Fig. 14 13C nuclear magnetic resonance
spectra of the copolymer initiated (a) at
236 nm, (b) by AIBN, (c) at 300 nm.
of 1C nmr Spectra Between
the Copolymer of DVE
Homopolymer of DVE and
(ppm from TMS)
No. of CN Substitutions
a, 0; B, 1; Y, 2
a, 0; l, 1; Y, 1
*The special carbon in consideration.
82.3, 83.4, 151.1
Comparing the C*'s in trimethyleneglycol (two oxygens are three carbons
apart) with pentamethyleneglycol (two oxygens are five carbons apart),
there is a 2.9 ppm difference in chemical shift due to the distance of
the two oxygens.79 The 2.9 ppm difference of the two C-O-C absorption
in the copolymer can then be explained by comparison between the C*'s
in structure XXVII and then XXVI.
HO OH HO OH
59.2 ppm 62.1 ppm
In conclusion, at low conversion, the copolymer contained a mixture
of structures XXVI and XXVII, a small amount of residual unsaturation and
probably some homopolymer structure which can be neglected for kinetic
consideration. The average md = 0.41 corresponding to a %". of structure
XXVII. This ratio of structure XXVI to structure X'".'11 was reflected
by the almost equal intensity of the two C-O-C absorption in 1C nmr
To distinguish between pure five- and six-membered ring structures
is very difficult because of the lack of information. As discussed
in Chapter II, it is unlikely to have a mixture of both six- and five-
membered rings structures in the copolymer because of the simplicity of
the spectrum. Furthermore, a favorable conformation of the complex and
the orientation of the cyclization makes the six-membered ring structure
Quantum Yield Study
Quantum yields for formation of copolymer were measured in aceto-
nitrile and at different wavelengths. The intensities are different at
different wavelengths for monochromatic light source. It will be pos-
sible to compare the quantum yields at different wavelengths by knowing
the dependence of intensity on quantum yield.
The data obtained for 0.6 m/l FN and 0.6 m/l DVE in acetonitrile
at-300 nm showed small dependence on light intensities. The intensity
was altered by using different slits.
The quantum yields were not exactly independent of the intensity
for nonequal molar solutions with 0.47 m/l of FN and 0.71 m/l of DVE
in acetonitrile at 300 nm. The intensities were altered either by
the different slits or with a copper screen in front of the sample tube
in this case. Also, the quantum yields were changed with different
intensities at 236 nm, with 0.6 m/l of DVE and 0.6 m/l of FN. The data
of the light intensity dependence on qauntum yeild and rate are listed
in Table XII.
The linear dependence of intensity on quantum yield are shown in
Table XIII. The linear correlation coefficients were close to one
indicating that the equations can be used for comparing the quantum
yields at the same intensity for different wavelengths. The results are
shown in Table XIV. Apparently the quantum yield at 236 nm is larger
than at 300 nm.
Ferree and Butler observed that the quantum yield was constant or
declined slightly as the wavelengths decreased, until part of the divinyl
ether band is excited at 236 nm, at which point the quantum yield in-
creased drastically. This fact suggests that only excitation in either
CO r-H LO C0 D m D CO
Cm =f r-C H D CO t- @l in N ln m El- Cc m H
o n C c- oc ri o N H C o Co o o
C) 0 CD H-, C CD C) C) C) C) C) C) C) C) CI
4o O O 4 CO -o cO c O LO O on o4 O
0 C) C) C) 0 0 0 C0 0 CD CD 0) 0 -0 0H
C) C) C) C) C) C C) C) C) C) 0D C) C) C) C)
M C E
in co 4-
co CO in N N co
0o oC (D CD C04
o co C CO CD CO
+1 +1 +I +1 +1 +1
LO CDO O LO LO LD
CO CO CO CO CO CO
C@ CM @4 @4 @4 @4
,--i ,-D LO -" (D (D --) I ,--t
I H r-i H H H- H H
C-lD C) C C) CD C) C) C)D
0 0 0 0 0 0 0 0 0
roo CM CO Ln .d- ilL
0 CO 4--
+1 +1 +1
0 0 0
C 0 0C
CO .CO CO
H O c
O I--t 0
co 0 C)n C
cm; O O
-- I H-
I 0 I
O r-4 CO O
-I C, r-i-
-H + I-I
I + b
bj D O M O
In 0 rH 0
r r-1 r I H
I I H 0
O O C r-1 <
--I n -i co c-
x C; x
o o o
0 ) O C o
IH 4 r- H
0 I + r- I C- I
I II 0 II 0 II
11 CN I1 C II C1
o o o
0 0 0
0( CO C\N
- r- r-H r
0t f0 0
S00 0 00 0
H-1 H-I H H-- i H r-H
the charge transfer complex absorption band or the divinyl ether
absorption band leads to initiation of copolymerization.
Comparison of Quantum Yield and Rate at 236 nm and 300 nm
Compound Concentration Wavelength Intensity Quantum Rate
(nm) Photons/sec Yield mg/min.
FN 0.47 m/l 236 8.86 x 1014 0.118 0.479
DVE 0.71 m/1 300 8.86 x 1014 0.076a 0.304a
aCalculated from equations in Table XIII.
It has been suggested that an exciplex was responsible for the
copolymerization while no ground state charge transfer complex was
observed,76 such as the photocopolymerization of vinylcarbazol and
acrylonitrile. Since at 236 nm the only effective photoabsorbing
species was DVE, the initiation process must proceed via the excited
state of DVE. Although there was no evidence of exciplex of DVE and FN,
no significant homopolymerization was obtained with DVE alone, this
fact indicating that the excited DVE initiated copolymerization possibly
through the interaction with the ground state of FN.
DVE -236 > (DVE)*
(DVE)* + FN (DVE* -> FN)
At 300 nm the photoabsorbing species was complex, therefore the
initiation process proceeded via the excited state of the complex.
DVE + FN ---< (DVE FN) ---- (DVE-FN)*
<---- 300 nm
At both wavelengths, the excited species dissociated into initia-
ting radicals. Although the initiating radicals have not been identified
they could be a paired cationic radical and anionic radical.
(DVE+FM)* -- 2R*
The ionic radicals initiated the copolymerization through free radical
processes because both air and a small amount of diphenylpicryl-
hydrazyl free radical (DPPH) retarded the reaction.
R* + (DVE -- FN) --
A small amount of FN (1%) did not initiate
excess DVE (99%) to a significant extent.
ether radical formation can be excluded at
R* + =-0-= -- R-
the homopolymerization of large
This indicated that the divinyl
least in the early stages of
The initiation through free fumaronitrile cannot be excluded but probably
is not able to compete with the low energy pathway of the complex
At 300 nm the light was totally absorbed by the complex for the
solution of 1.2 m/1 and of FN and 1.2 m/1 of DVE.
FN + DVE -- (DVE -- FN)
(DVE -- FN)* is the excited complex.
(DVE FN)* (DVE)
300 nm d
30 nm (DVE -- FN)* 2R*
Ground state molecules
I = The light intensity absorbed by the complex.
=I (light intensity absorbed by actinometer).
c = The quantum yield with only radiation deactivation considered.
The excited complex may be deactivated by collision processes.
As discussed before the exact structure of the excited state is not
known, but it may still be very polar and very likely to be deactivated
by the polar monomers and the ground state complex. The radiationless
deactivation process may be proposed as in Scheme II.
The rate of formation of the primary radicals (R*) can be derived by
applying the steady state assumption and assuming that the radiationless
rates of the deactivation of excited complex is proportional to the total
concentration of monomers, (T).
-d(DVE -- FN)* (T)DE
( dt )rl = kr(T)(DVE --, TN)*
dt rl rl
d(DVE -- FN)*
d(D I-c Fc (DVE -> FN)"(kd + k r(T)) 0
dt cc d rl
From Scheme II
R. = k (DVE -P FN) =. (kdI c )/(kd + K ,(T))
i d dcc d rl
At 236 nm the complex absorbed insignificantly but DVE absorbed part
of the light, Id = absd(DVE), and is responsible for the initiation
through the following equation.
236 nm FN d
DVE I DVE* ---) (DVE* -> FN) 2R*
d d df
Dd is the quantum yield with only radiation deactivation considered.
In the same way as proposed at 300 nm, the rate of formation of the
primary radicals (R-) can be derived by assuming the rate of the
d(DVE* --> FN)
deactivation of exciplex, (- d( ----) is proportional
to the total concentration of monomers (T).
d(DVE* > FN)
dtV- = k (DVE*)(FN)-(DVE*-- FN)(k' + k' (T)) = 0
dt df d rl
k'dIabs d d(DVE)
R. = k' (DVE* --: FN) = d abs d d
i d k' + k'rl (T)
Dependence of Feed Composition on Rates
It was of considerable interest to measure the rate of copolymeriza-
tion obtained for different feed compositions. As explained later, the
study of the rate as a function of the feed composition when the total
monomer concentration is kept constant should give information on the
role of the charge transfer complex in the mechanism of copolymeriza-
The effect of the variation of the concentration of the reactants
on the rate of a photochemical reaction is difficult to determine
unless the rates are corrected to a constant absorbed light intensity
from known variation of rate with the absorbed light intensity.
The rate of copolymerization was measured in acetonitrile. The
samples were prepared as indicated in the experimental section. The
solutions were irradiated at room temperature for a measured period
of time and the reaction was stopped by immediately opening and roto-
vapping to dryness. Methanol was added, and the insoluble polymer was
filtered and washed with methanol, and dried in an oven at 500C at least
for 4 hours. The actinometer solution was irradiated in the same tube
right after the reaction. The rate was obtained by measuring the
weight of the polymer after drying. The dependence of the light
intensity on rate is reported in Table XII. The linear relation is
shown in Table XIII.
Rates of products formed in secondary reactions (including the
polymerization) usually show some other than first order dependence
on intensity. In general, bimolecular termination by a reaction
involving active chain carrying species results in rates proportional
to the square-root of intensity. Termination which occurs from a first
order reaction leads to a rate which is dependent on the first power
of intensity. It was observed that the rates were dependent on an
order between 0.72 1 of intensity. It was close to first order for
the equal molar solution at 300 nm. Possibly both the bimolecular and
unimolecular termination are operating in photocopolymerization. The
first order termination is more possible because the propagating radical
may be terminated by chain transferring to solvent and the polar monomers.
The dependence of feed composition on corrected rate was listed in
Table XV and XVI and Fig. 15 and 16.
The rates were compared at I = 1.52 x 10 photons/sec and
6.4 x 1014 photons/sec for irradiation at 300 nm and 236 nm respectively.
It was observed that at 300 nm, the maximum rate was at fd = 0.5
(equal molar solution), but at fd > 0.5 at 236 nm for both total
concentration [(T) = 2.40 m/l and 0.60 m/l]. The composition of copoly-
mers were in the same range (md = 0.35-0.48), and not much different
from each other for both wavelengths. Interestingly, at 236 nm, the
rate maximum fell on fd = 0.66 and 0.80 for higher total concentration,
(T) = 2.00 m/l, but on fd = 0.55-0.80 for lower total concentration,
(T) = 0.60 m/l. The ir and nmr spectra were almost identical. These
facts indicated that the same propagation process was employed. The
different positions of rate maxima were then due to the different
The kinetical derivation of the overall rate of copolymerization
can be done by assuming the simplest propagation as follows.
\ CN k
/ \ \CN
I + (DVE F N) k -2
+ (DVE-> FN) -
The Rate of Copolymerization of DVE-FN System in Acetonitrile
at 300 nm, at I = 1.52 x 10 Photons/Sec.
Total Ratea Total Ratea
Concentration fd (mg/min) Concentration fd (mg/min)
0.60 1.00 nil 1.20 0.80 0.179
0.60 0.90 0.051 1.20 0.60 0.313
0.60 0.82 0.117 1.20 0.50 0.376
0.60 0.80 0.098 1.20
0.60 0.70 0.131 2.40 0.98 0.087
0.60 0.67 0.154 2.40 0.70 0.475
0.60 0.60 0.150c 2.40 0.60 0.604
0.60 0.55 0.154 2.40 0.50 0.749
0.60 0.52 0.185 2.40 0.33 0.580
0.60 0.50 0.198 2.40
0.60 0.45 0.161 2.40
0.60 0.40 0.190 2.40
0.60 0.40b 0.053 2.40
0.60 0.30 0.167 2.40
0.60 0.20 0.147 2.40
0.60 0.00 --- 2.40
aThe rates were corrected to I = 1.52 x 106 photons/se with equation
log rate a /rate = log 1.52/I x 10.
b cal exp exp
Sample was open to air.
0 H II 0
O *rl *H
Y / ", P 4r-
S\ 0 0
\ X 0
\ \ ,C-
S 0 4 -
o o 4-Co
0 0 I 000 0 4 -1
(~uI/rEC) uoT 4- ) ( q-
The Rate of Copolymerization of DVE-FN System in Acetonitrile
at 236 nm, at I = 6.4 x 10 Photons/Sec.
Total Ratea Total Ratea
Concentration fd (mg/min) Concentration fd (mg/min)
0.60 1.00 nil 2.00 0.90 0.0413
0.60 0.90 0.0217c 2.00 0.80 0.0692
0.60 0.80 0.0274c 2.00 0.80b 0.0267
0.60 0.70 0.0231c 2.00 0.66 0.0686
0.60 0.60 0.0229c 2.00 0.50 0.0644
0.60 0.55 0.0303 2.00 0.40 0.0448
0.60 0.50 0.0246c 2.00
0.60 0.48 0.0235 2.00
0.60 0.45 0.0170c 2.00
0.60 0.41 0.0197 2.00
0.60 0.40 0.0179 2.00
0.60 0.31 0.0209 2.00
0.60 0.00 --- 2.00
The rates were corrected to I- 6.4 x 101 photons/sec with the equation
log rate /rate = 0.72 log 6.4/I x 10
b cal exp exp
Sample was open to air.
0 0 0 0 0 0 0
0 0 0 0 0 0 0
(uyTm/S) uoTqezTjGwa m odoo jo Ga'H
In this copolymerization, the cyclized DVE radical (XXXI)
attacked complex and monomers to produce polymers with 1:1 and 1:2
compositions, respectively. The 1:1 composition cannot be explained
by the polarity of the radical and complex, because the negatively
polarized radical (XXXI) would attack the positively polarized carbon
carbon double bond of an acceptor monomer since polymerization of DVE
under the same condition is negligible. The propagation of the copoly-
merization can be interpreted as a competition between the acceptor
monomer FN and the CTC toward the cyclized DVE radical (XXXI).
Although the cyclization of the DVE radical (XXXI) could react
preferably with the acceptor such as MAH because the reaction of a
radical and a monomer of opposite polarities would stabilize the
transition state,0 the FN may not be so preferable because of the
less polarization then MAH. Thus, the reactivities of FN and the complex
are compatible. As a result, a competition between FN and CTC control
the structure of the copolymer of DVE-FN system. On the other hand,
the complexation of the HOMO of DVE and LUMO of FN would increase the
energy of HOMO of DVE71 and make the CTC more reactive toward the
destabilized cyclized radical (XXXI) than the uncomplexed DVE.
In equations 22 and 23; the cyclization may proceed by both
concerted and stepwise processes. A fast cyclization has been observed
for this cyclocopolymerization, which indicated a concerted mechanism.
But some results showed the stepwise mechanism, especially for the
fact that both cis and trans disubstituted ring units were obtained in
the analysis of spectra. A stepwise cyclization cannot be ruled out.
However, by assuming a fast stepwise cyclization (probably just a little
slower than the cis-trans rotation around the disubstituted single bond)
of radical XXXII, the same result will be concluded as with the
concerted process as far as rate is concerned.
By a;zuinir, only the cross-termination
mi + m2 -- dead polymer
and the steady state:
kl2(ml')(M2) = k21(m2.)(M)
k (M )
Ri = R = 2k 2(ml.)(m ) = 2k 21(M (m.)2
Then, by assuming (DVE --- FN) = K (FN)(DVE)
i 21 1/2 1/2 1/2
Rate = ( )/2(DVE) /2(FN)K [2k2 + k lK (DVE)]
2k k c 12 11 c
By definition (FN) = (1 fd)(T), (DVE) = fd(T)
where (T) = total concentration= (FN)+(DVE), f = (DVE)/(T)
The rate maximum is a function of (T) and will be at fd = 0.5 by keeping
The rate maximum is a function of (T) and will be at f = 0.5 by keeping