THE PREPARATION AND POLYMERIZATION
OF CERTAIN DIUNSATURATED PHOSPHORUS
WILLIAM CARGILL BOND, JR.
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
This dissertation is dedicated to my grandfather,
Mr. Morton M. Boyd.
The author wishes to express his gratitude to
Dr. G. B. Butler for his counsel and encouragement in
the undertaking and completion of this work.
The author also wishes to express his appreciation
to his advisory committee and fellow graduate students
for their assistance, as well as understanding advice and
Special thanks are due Mrs. Thyra Johnston for
typing this manuscript.
The financial support of the United States Air
Force and the National Science Foundation are gratefully
TABLE OF CONTENTS
DEDICATION . . . . . . .
ACEKNOWLEDGMENTS. . . . . . .
LIST OF TABLES . . . .
LIST OF FIGURES. . . . . . .
I. INTRODUCTION . . . . .
Historical. . . . .
Development of the Problem.
Statement of the Problem. .
Discussion. . . . .
II. DISCUSSION OF RESULTS ..
Monomers. . . . . .
Polymers. . . . . .
Infrared Data . . . .
III. EXPERIMENTAL . . . . .
Source and Purification of M
Chemical Analyses . ..
Physical Measurements . .
Preparation of Intermediates
Preparation of Phosphines .
Preparation of Phosphonium S
. . . 1
. . . 1
. . . 6
. . . . ?
. . . . 8
. . . . 18
. . . 18
. . . . 25
. . . . 33
. . . . 35
materials. . 35
. . . . 35
. . . . 36
. . . 36
. . . 40
alts. .... 47
Polymerization of Phosphonium Salts. . 52
Preparation of Copolymers. . . . . 60
IV. SUMMARY ................... 62
BIBLIOGRAPHY. . . . . . . . . . 63
BIOGRAPHICAL SKETCH . . . . . . . . 66
LIST OF TABLES
1. Physical Properties, Analytical Data, and
Yields of Intermediate Unsaturated
Phosphines . . . . . . . . 10
2. Physical Properties, Analytical Data, and
Yields of Unsaturated Phosphonium Salts. . 12
5. Physical Properties, Analytical Data, and
Yields of Phosphonium Polymers . . ... .52
LIST OF FIGURES
1. Intrinsic Viscosity Determination of
Phosphorus Containing Monomers . . ... 30
Walling (1) noted as early as 1945 that the gel-
point for some dienes should occur later than calculated
by Stockmayer's (2) equation, as the equation failed to
account for occasional cyclization. Later Butler and
Ingley (3) found that diallyl quaternary ammonium salts
showed no gel-point and were water soluble, while under
similar conditions, triallyl salts gave crosslinked polymers
and monoallyl salts did not polymerize. These results
seemed in direct contradiction to Staudinger's (4)
hypothesis that polymers of non-conjugated dienes would
yield only crosslinked polymers. Further study of the
poly-diallylammonium salts revealed little or no residual
unsaturation; thus, it seemed that complete polymerization
had taken place and yet the polymer had not crosslinked.
Butler and Angelo (5) proposed an alternating intra-
molecular-intermolecular chain process to explain how both
double bonds could be polymerized and yet a linear polymer
produced. According to their mechanism, one double bond
reacted with the initiator and the resulting radical, anion,
or cation in turn attacked intramolecularly the double bond
at the opposite end of the molecule. This produced a cyclic
secondary reactive site which attacked an adjacent monomer
molecule intermolecularly to give a polymer with saturated
cyclic recurring units separated by methylene groups. The
process, termed "Cyclopolymerization," can be illustrated
P N B-CH B-CH-O
+ R" )
I II III
III H+ n
To prove the presence of the piperidinium ring,
Butler, Crawshaw, and Hiller (6) degraded poly-(diallyl-
ammonium bromide). The nitrogen was first benzoylated in
the presence of sodium hydroxide and then oxidized to open
the ring as shown:
2 IlHTTaOH "2
n 6 5 n
VI + 4n04 ) C OOH
o 5 n
The structure of the product was proven by elemental
analysis, potentiometric titration, infrared spectra, and
upon heating it slightly above its melting point benzoic
acid was sublimed and the residual polymer was found to be
crosslinked and insoluble.
Since the cyclopolymerization mechanism was proposed
it has been extended to systems more numerous than can be
listed herein. However, a few of the more important should
be mentioned. marvel and Vest (7) polymerized a,a'-
dimethylene pimelate derivatives to essentially saturated
linear polymers. marvel and Stille (8) used a Ziegler type
initiator to polymerize 1,5-hexadiene and 1,6-heptadiene
which were the first hydrocarbon cyclic polymers prepared
by this mechanism. 1,6-Heptadiene showed no crosslinking
and was 90 per cent or more cyclic polymer. 1,5-Hexadiene,
as would be expected, gave more crosslinking and showed
less cyclization. Acrylic (9,10,11) and methacrylic (12)
anhydrides were polymerized to give molecular weights as
high as 95,000. Marvel and Gall (15) and Field (14) re-
ported the cyclopolymerization of 2,6-diphenyl-l,6-
heptadiene, which had a melting point of 300C. Not only
1,6- and 1,5-dienes have been cyclopolymerized, but Marvel
and Garrison (15) even obtained polymers from higher a-
diolefins giving rings with seven to twenty-one members.
All types of initiation have been used to produce
cyclic polymers; for example, acrylic anhydride (11), by
free radical initiation; diisocyanates (16,17) by anionic
initiation; diallylsilanes by (18) Ziegler initiation, and
2,6-diphenyl-l,6-heptadiene (14) by cationic initiation.
Besides the homopolymers referred to above, 1,4-
dienes have been shown to undergo cyclic copolymerization
to produce copolymers in which there are two moles of comer
to each mole of diene (19). The mechanism below has been
suggested for the reaction of divinyl ether with maleic
CH2==CH O- H Z* + ZCH2---CH -H2
CHCO Z- O--CH C
-2 CCH-CRI -2
Monomer reactivity ratios (20,21) for certain of these co-
polymerizations have been calculated and are consistent
with the above mechanism.
One of the most recent and more interesting phases
of cyclopolymerization is that in which doubly stranded or
ladder polymers are produced. Angelo (22) reported the
cyclopolymerization of poly-3,4-isoprene; Overberger (23),
the preparation of a ladder polymer of vinyl isocyanate by
either of the two mechanisms shown below.
0 0 0
II II II
--C-CH2 H_ Co60 N N N N
0 0 0 0 0
CH2==CH-N=C=0 | II It I
I NeCN_ ATBN
-550 8 ^
Soluble, 90% ladder polymer
The above polymers show a much higher thermal stability than
cyclic polymers connected by methylenes, and the latter tend
to be more stable than regular straight chain polymers.
Development of the Problem
Previously, several 1,6-diunsaturated phosphonium
bromides were prepared and polymerized by Skinner (24) to
give polymers containing repeating six-membered rings in
the backbone of the chain. These were either diallyl or
dimethallylphenylphosphonium bromides with the fourth sub-
stituent being either phenyl or alkyl.
H CH 2
2 H2 BrR
R = H or CH ; R' = CH C2H5, C3H7, or C6,H
The polymers formed were soluble in ethanol and
dimethylformamide and their infrared spectra indicated
that they were free of unsaturation. Two of these were
converted to phosphine oxides and their infrared spectra
were found to be identical with phosphine oxide polymers
synthesized by another route (25).
Along with the six-membered rings it was decided to
determine if five-, unsymmetrical six-, seven-, and eight-
membered rings could be formed during polymerization. Thus,
the present project was initiated.
Statement of the Problem
The objectives of this investigation were as follows:
1. To synthesize a series of phosphonium salts which
could conceivably undergo cyclic polymerization to
give linear soluble polymers.
2. To determine the possibility of forming five-, six-,
seven-, and eight-membered rings from functional
groups other than diallyl or dimethallyl.
5. To isolate and study the properties of any soluble,
4. To show that cyclic copolymers of phosphines can be
formed and to study the copolymer obtained therefrom.
The method of preparation of phosphonium halides
used in this research was that which had been previously
proven successful by Skinner (24), who prepared the diallyl-
and dimethallylphosphonium halides. The procedure des-
cribed by Jones and Davies (26) was followed in the
preparation of the phosphines. To these phosphines was
added an alkyl halide in dry ether, which usually resulted
in a good yield of the phosphonium salt.
Some of the phosphines prepared were diphenyl
substituted; thus, for example, diphenylchlorophosphine
was reacted with the appropriate Grignard reagent to yield
the unsaturated phosphine. The phosphine was then reacted
with an olefinic halide in ethyl ether to yield the di-
olefinic phosphonium salt. A generalized synthetic scheme
is shown below.
S5 65 R'X C65
C, CH CH R+
P--C1 + RI-gX ---- P--R P
CH CH C,HZ R'
o 5 o 5 o 5
R = C-=CK- or CH=CHCH2CH-
X = Cl or Br
R' = CH HC-, CH =-CH2, or CH2=CHCHCR
It was found in the literature that diphenylvinyl-
phosphine (27) had been previously synthesized but that
3-butenyldiphenylphosphine had not. The phosphonium
compounds prepared were all new compounds, none of which
were found in the literature. The monomers were used as
obtained from the last reaction in the scheme. Those
isolated were characterized by use of infrared spectra,
elemental analyses, and melting points. The values ob-
tained may be found in the experimental section, and in
phosphines, which were not isolated in previous
isolated and characterized by boiling point,
index, density, infrared spectra, and elemental
They were found to be pure and fairly stable.
obtained may be found in Table 1.
(e 0 p
= H h
rI C C
Fd (Ta 01
P-4 \ 0
r(u t U\ Lr\
V) 0 0
flu l -
1-1 0-% 01%
pi u p
0 b CP4 '
~~ P C
to C o
m \0 \0
*i c ^
*" t *
A free radical initiator was used to start the
polymerization of the salts; tert-butylhydroperoxide in
water solution or a : '-azodi-iso-butyronitrile in dimethyl-
formamide was used respectively with the ammonium, and the
diallyl substituted phosphonium salts. However, neither
of these gave good results with the monomers prepared.
Benzoin and an ultraviolet light source gave more effective
initiation and a polymer which was easily recovered.
The mechanism for the polymerization of these
compounds would be expected to be similar to that proposed
by Butler and Angelo (5) in the polymerization of the
diallylammonium compounds. It can be represented, for
example, by the diphenyl substituted phosphonium compounds
in the following equations:
CH =C CH
(CH(c) (C ) + I. -..
(C2 ) (C )
R = H or CH XXIII
X = Br or CI
y or z = 0, 1, or 2
The study of the polymers produced consisted of an
elemental analysis, softening and melting range, infrared
spectra, and the intrinsic viscosity. The viscosities of
polyelectrolytes are difficult to determine due to ioni-
zation; however, usually this can be suppressed by using a
common ion. Even under these conditions, the viscosities
of the ionic polymers were ambiguous. To overcome this
difficulty, one of the phosphonium salts was converted to
the phosphine oxide which gave a more accurate viscosity
A number of 1,4-dienes have been cyclocopolymerized
with a variety of comers. For this reason, an attempt was
made to copolymerize divinylphenylphosphine with acrylo-
nitrile. According to the mechanism described earlier,
Butler (19) suggests that two moles of comer polymerize
with each mole of diene. This, however, was with maleic
anhydride which does not tend to homopolymerize. Using a
comer such as acrylonitrile, a mole percent greater than
two times the diene concentration is always obtained in
the polymer, due to the pronounced tendency of it to homo-
polymerize. The copolymer prepared had approximately three
times the mole percent of acrylonitrile and no residual
unsaturation. Thus, it is felt to be consistent with the
cyclocopolymerization mechanism but with a slightly greater
percent of acrylonitrile than predicted, due to its homo-
A sample of diallyldimethylphosphonium chloride was
prepared to determine if possibly in the aryl substituted
phosphonium compounds the phenyl groups were so large that
they restricted chain growth due to steric hindrance. The
method of synthesis of this monomer was entirely different
from that previously demonstrated. The procedure described
by Burg and Slota (28) was followed in which dimethylamino-
dichlorophosphine was reacted with a Grignard reagent,
producing the disubstituted dimethylaminophosphine. This
product was then reacted at a higher temperature and for
a longer period with another Grignard reagent which cleaved
the phosphorus-nitrogen bond and produced the trisubstituted
phosphine. This phosphine was then reacted with an alkyl
halide to yield the phosphonium salt desired. The total
synthesis is shown below.
+ 2CH3MgBr --
XXVI + CHC -H-C1 -
Xay XVI 3
C H2==CH-CH CH
C=CH- -C CH3
DISCUSSION OF RESULTS
In order to obtain the 3-butenyl substituted monomer
it was necessary to synthesize 4-bromo-l-butene. The
bromination of allylcarbinol was the only feasible route
found; however, the synthesis of this alcohol gave unde-
sirable by-products which at times constituted the entire
In the synthesis of allylcarbinol through the
reaction of vinylmagnesium halides with ethylene oxide,
it was observed that under certain conditions the pre-
dominant product was the halohydrin derived from ethylene
oxide rather than the expected alcohol. A further study
of this reaction indicated that a specific halogen ion
effect existed. Use of vinylmagnesium chloride leads to
better yields of allylcarbinol and to none or to minor
amounts of the ethylene chlorohydrin, while under the same
conditions, use of vinylmagnesium bromide leads to yields of
ethylene bromohydrin varying from 10 to 40 per cent, with a
corresponding reduction in the yields of allylcarbinol. In
both cases, the yields of allylcarbinol increased with
ethylene oxide/Grignard ratio, and with increasing
temperature of the reaction. The mechanism of this re-
action was not determined, as the results obtained were
not consistent. Two runs under apparently the same
conditions produced different yields of both halohydrin
The allylcarbinol obtained from the reaction could
be recovered in about 97 per cent purity, as analyzed by
vapor phase chromatography. It was converted to the
bromide by a modification of the reaction proposed by
Gaubert, Linstead, and Rydon (29). The bromide obtained
was also found to be about 97 per cent pure; thus, very
little had isomerized to crotyl bromide, as might be
expected. The 4-bromo-l-butene was then converted to the
Grignard reagent and reacted with the appropriate chloro-
phosphine to produce the 3-butenyl substituted phosphines.
Phosphines, as a group, are very reactive substances,
which is characteristic of all derivatives of trivalent
phosphorus. The lower aliphatic substituted derivatives
are, in fact, spontaneously combustible when exposed to
atmospheric oxygen. The tri-aryl phosphines are, on the
other hand, stable to air oxidation and can be stored for
long periods of time with no special precautions. The
mixed aryl-alkyl phosphines prepared lie between these two
extremes. Usually, the monophenyl substituted phosphines
were transferred and all open manipulations were carried
out in a nitrogen flushed dry box. The diphenyl substituted
derivatives were transferred under a flow of nitrogen and
did not oxidize appreciably. At elevated temperature both
types of phosphines oxidized readily. For this reason,
distillations were performed under nitrogen. Even small
amounts of oxygen present during distillation caused hazy,
impure distillates which were contaminated by the solid,
insoluble phosphine oxide.
After distillation, the aryl substituted phosphines
were usually reacted immediately with an alkyl halide dis-
solved in ether. In order to run refractive indices and
density measurements on the phosphines a small sample was
distilled and sealed under nitrogen. This fraction was
then placed in a nitrogen filled dry box where the refrac-
tive index was taken and a pychnometer, for density measure-
ments, could be filled without oxidation.
The aryl substituted phosphonium salts produced were
not extremely hygroscopic; however, presence of the smallest
amount of the very hygroscopic phosphine oxide impurity
caused deliquescence of the phosphonium salt. Therefore,
if a sample of phosphonium compound was contaminated to any
extent with the phosphine oxide it would form an oil and
could not be recrystallized, as the solubility of both the
salt and oxide were very similar. In the infrared region,
the phosphorus-oxygen bond in the phosphine oxide gave a
broad, strong absorption at 1170 cm- and was found in a
number of the impure salts. Usually, when this peak was
present, it was also possible to see a strong absorption
at 3500 cm- due to water. When pure phosphonium salts
were produced, however, they showed neither absorption
and could be opened to the atmosphere and still remain
relatively free of water.
The dimethylphosphines prepared were extremely
sensitive to oxygen, and any exposure to the atmosphere
caused immediate oxidation of the product. The samples
were sealed in ampoules after distillation and neither the
refractive index nor density of the pure phosphine were
obtained. An infrared spectrum was run on allyldimethyl-
phosphine and it was found to contain a large amount of
oxide; however, the polymer of diallyldimethylphosphonium
chloride does not show the characteristic peak at 1170 cm-1
Thus, the phosphine must have oxidized during the spectral
analysis. The diallyldimethylphosphonium chloride produced
from the phosphine above was deliquescent and had to be
kept in a drying pistol over phosphoric anhydride. Even
with these precautions the elemental analysis was consistent
with the hypothesis that there was water in the sample and
the melting point, taken in a sealed tube, showed a wide
melting range and water condensed in the tube.
Discussion of the phosphines and phosphonium compounds
produced will be divided into three groups dependent upon
the number of phenyl substituents in the compound. Exami-
nation of the chemical structures, boiling points, and
reactivity of the phosphines suggests placing them in these
groups. Considering the diphenyl series first, allyl,
methallyl, and 3-butenyldiphenylphosphine were readily pro-
duced by the method described by Jones and Davies (26).
Their physical characteristics were as noted and their
stabilities were as described previously. The properties
of diphenylvinylphosphine were surprising because it was the
least stable of all the diphenyl substituted phosphines
prepared. One would expect the vinyl group to increase the
stability of the phosphine in a manner somewhat analogous
to an aryl substituent rather than to an alkyl substituent
since it has the possibility of stabilizing the pair of free
electrons on the phosphorus atom by forming a resonance
hybrid similar to that of a phenyl group (30). In fact,
both nuclear magnetic resonance and infrared spectra show
that there is some type of interaction between the pi cloud
of the double bond and the pair of electrons on the
phosphorus atom. A similar interaction, but to a much
greater extent, was found in the ultraviolet spectra of
vinyl substituted amines (31). The vinyl substituted
phosphines, unfortunately, did not show an absorption
maxima within the reliable range of the instrument when
their ultraviolet spectra were run. Thus, they could not
be studied by this means.
The phosphonium salts obtained from diphenylvinyl-
phosphine were very difficult to isolate in any degree of
purity. Allyl bromide, which alkenyl bromide reacts most
rapidly with the phosphines, formed the purest salt. As
the rate of reaction decreased, so did the purity of the
monomer obtained directly from the reaction mixture.
Methallyl chloride, which is quite slow in its reaction,
did not give any isolatable salt in numerous attempted
The diaryl substituted phosphonium salts prepared
for polymerization were all synthesized starting with one
of the four diphenyl phosphines mentioned above. The
phosphonium compounds produced from diphenylvinylphosphine
have already been mentioned; those from the other phosphines
were easily prepared in good yields and high purity. The
alkyl chlorides were always much slower to react with the
phosphines than the bromides but often gave a better yield
In the second group, the monophenyl substituted
derivatives, divinyl-, diallyl- and di-5-butenylphenyl-
phosphines were prepared. They all showed similar
characteristics: they were more reactive toward oxygen and
all had lower boiling points than the corresponding diphenyl
substituted phosphines. Divinylphenylphosphine, similar to
the diphenylvinyl derivative, showed much less stability
and colored rapidly on standing, even under nitrogen. The
diallylphenyl-, and di-3-butenylphenylphosphines were
prepared for use in the preparation of boron-phosphorus
intermediates in other work (32), and also as alternate
synthetic intermediates in this work. Diallylphenylphos-
phine had been previously prepared by Jones and Davies (26)
but di-3-butenylphenyl was a new compound and its physical
characteristics were determined.
The last group of phosphines prepared were the
dimethyl substituted. It was decided that perhaps the
large phenyl groups were causing steric hindrance during
polymerization and by replacing these with methyl groups
higher molecular weight polymers with superior character-
istics might be obtained. The procedure described by Burg
and Slota (28) had not been applied to the synthesis of
unsaturated methyl phosphines. The method did yield a
satisfactory phosphine which was readily converted to the
phosphonium salt. The phosphine and the salt produced from
it are both new compounds; however, their physical
characteristics could not be measured due to their extreme
sensitivity to the atmosphere.
Since the phosphonium compounds are ionic in
character, the use of anionic or cationic initiation is
precluded. It is also known that Ziegler-Natta catalysts
are poisoned by this type of system (33). Therefore, free
radical initiation was chosen for the polymerization of
the monomers. This method of initiation presents the
complication that hydrogens a to the double bonds are
easily extracted by free radicals causing degradative chain
transfer and essentially termination of the polymer chain.
The choice of which free radical initiator to use
was dependent on previous work. The first choice was ac '
azodi-iso-butyronitrile in dimethylformamide. This had
been found to give polymers by Skinner (24) in his work
with diallyl and dimethallyl substituted phosphonium
bromides. He found that polymers obtained using this type
of initiation were difficult to isolate and none had an
intrinsic viscosity ([E]) above 0.04. Tert-butyl hydro-
peroxide had been used effectively with the quaternary
ammonium salts but was found to give much poorer yields of
phosphonium polymer. In an attempt to produce higher
molecular weight polymers which would be easier to isolate,
benzoin and ultraviolet irradiation were used. This
catalyst was indeed found to give higher molecular weight
polymers, ([q] = 0.025 for five-membered rings to 0.08 for
six-membered rings) which were easily isolated from acetone
as white amorphous solids. The infrared spectra of some of
the polymers indicated that they contained up to 10 per cent
residual unsaturation. This was difficult to explain since
the polymers described by Skinner had been reported to be
saturated. The difference between the polymers presented
here and those produced in Skinner's work is the result of
the difference in initiator concentrations used. His
polymerizations were all run using 4 per cent initiator,
while the benzoin concentration was 0.1 per cent in the
polymerizations which left residual unsaturation. Polymers
produced using 1 per cent or more benzoin as initiator were
found to be essentially saturated as determined by absorption
in the infrared spectra at 935 cm-1. Therefore, the majority
of polymerizations in this work were carried out using an
initiator concentration of 2 per cent.
Solution polymerization was used throughout this
study. In the benzoin initiated polymerizations, almost
saturated aqueous solutions (50%, by weight) were used. No
bulk polymerizations were attempted, as the phosphonium
monomers were very high melting.
The polymers obtained from the phosphonium monomers
discussed earlier were soluble in ethanol and dimethyl-
formamide. Those produced with more than 1 per cent
initiator showed little or no absorption in the infrared
region corresponding to residual unsaturation. The
softening points of the polymers were much higher than the
melting points of the corresponding monomers; the results
of the analyses indicated that the polymers were isolated
with one molecule of water per phosphonium unit.
The properties of these aryl phosphonium polymers
were consistent with those predicted by the cyclic mechanism
proposed by Butler and Angelo (5), since they were
essentially saturated and soluble. Their tendency to form
monohydrates 'was! consistent with the observations made by
Skinner (24). The conversion of poly-(diallyldiphenyl-
phosphonium bromide) to poly-diallylphenylphosphine oxide)
was an additional proof of cyclization since the infrared
spectrum of the phosphine oxide obtained was identical with
the spectrum obtained by Berlin (25) who polymerized
diallylphenylphosphine oxide to yield a cyclic polymer.
All of the diolefinic monomers synthesized produced
polymers, except vinyl-3-butenyldiphenyl, and di-3-butenyl-
diphenylphosphonium bromides. It is quite surprising that
the former compound did not polymerize, if not by way of the
cyclic mechanism to produce the six-membered ring structure,
by way of the vinyl group. The fact that the latter com-
pound did not polymerize is not too surprising since cyclic
polymerization would require formation of an eight-membered
ring. Also, the 3-butenyl double bond is not very reactive
toward free radical initiators.
The specific viscosities of the polymers were calcu-
lated using this equation (54):
rsp = t
where t is the flow time of the sample and t is the flow
time of the solvent. The intrinsic viscosity [q] was de-
termined by graphically plotting specific viscosity divided
by concentration versus concentration and extrapolating to
zero concentration. Intrinsic viscosity is defined by the
Sc = 0o
The viscosity data for the poly-phosphonium salts
were typical of the data for polyelectrolytes in general.
It was reported (35) that the viscosity curves ('sp/C vs. C)
of polyelectrolytes were strongly concave upward, in con-
trast to the behavior of uncharged linear polymers. This
was believed to be due to the dissociation of the ionic
bond in the solution which lead to large repulsive forces
between the positively charged groups remaining on the
chain. These forces gave rise to greatly expanded configu-
rations and very large intrinsic viscosities. At high
concentrations, the molecules in the polyelectrolyte had not
been significantly ionized and tended to partially overlap;
thus, they were not appreciably expanded. As the solution
is diluted, the molecules no longer filled all the space
and some of the halide ions left the regions of the chains.
This caused a development of charge and extension of the
chains. The addition of a strong electrolyte suppressed
the loss of halide ion and the viscosity behavior became
more normal. However, the intrinsic viscosity is dependent
on the volume of the polymer in the solution and it was
found that added electrolyte compressed the hydrodynamic
unit corresponding to the polyelectrolyte. The compression
changed the shape of the polymer and caused a marked de-
crease in the viscosity of the solution. Thus, the term
intrinsic viscosity, for a polyelectrolyte can not be used
in the same sense as for an uncharged polymer since it is
dependent entirely on the concentration of added salt.
This type of behavior was exhibited by the poly-phosphonium
compounds and is illustrated in Figure 1. In line I the
poly-phosphonium bromide was dissolved in ethanol, and as
was expected, sp/c tended to increase with dilution.
Addition of a strong electrolyte, such as a potassium halide,
provided a common ion effect which repressed the ionization
of the polymer. At this time the viscosity behaved similar
to a linear uncharged polymer, as shown by line III. Line II
0 0.4 0.8 1.Z
Plots of nsp/ against c for:
I Poly-(diallyldiphenylphosphonium bromide) in 95% ethanol.
II Poly-(diallyldiphenylphosphonium chloride in a 1.0 M 1:1
ethanol-water solution of potassium chloride.
III Poly-(diallyldiphenylphosphonium bromide) in 1:1 ethanol-
water 0.1 M in potassium bromide.
IV Poly-(diallylphenylphosphine oxide) in 95% ethanol.
Fig. 1.-Intrinsic Viscosity Determinations of Phosphorus
represents poly-diallyldiphenylphosphonium chloride rather
than the bromide; but demonstrates an intermediate change
between line I and III where insufficient common ion was
present to completely repress ionization. Line IV represents
the viscosity of the phosphine oxide, which was produced from
the same sample of diallyldiphenylphosphonium bromide used
for the determination of line III. The intrinsic viscosity
for the phosphine oxide polymer was probably the most
reliable, as it was not a polyelectrolyte and can not be
affected by ion concentrations.
The intrinsic viscosity measurements for the phos-
phonium compounds will be given in the experimental section
and in Table 5.
Copolymers of divinyl compounds have been known for
some time. Divinyl ether (56) and divinylaniline (20) were
both copolymerized to yield cyclic copolymers with a variety
of comers. Since divinylaniline is very similar to divinyl-
phenylphosphine it was felt that copolymers could be easily
obtained. However, previous experiments with divinylphenyl-
phosphine and maleic anhydride had been attempted and no
polymer was produced (37). The copolymerization with
acrylonitrile gave a brilliant red, brittle, solid which
contained approximately three moles of acrylonitrile for
each mole of divinylphenylphosphine. The percent of
acrylonitrile in the polymer increased as the concentration
\, cY 0 o 0 co t-
H---. H CN Cj ON\D 0 t-
C( J1 H O 0 01 H H 0
1OJ OJ C\J C'J H CJ J1 H
- *. - *, 0. 1
-ri m r l u
CO 1 H co co o oo oo o\ H
CC H H 01 LA (1 LO -4
O i H **
0 0 0
N PC01 O C O CO
2O H 1 H-
SII II II *rII H c
O 0 o 0 0 0 0 0 0 0
d GJ4 4-3 4- 0)
P-i EP cvj o- P4 0 0U 200
0n ) m m
0 I I II *
P H 0 0 P- O 0
gu t" o a c t co s co f u
m o o
j 0 a C 0 0 A 0
S)P *m *H *H *H *H *H
S 0 0 0 0 0
N (U O
,Co (U a H H H
,, O Hj >' i-i 'dp ( C I dl
of that comer increased in the initial charge. The co-
polymer produced was soluble in dimethylformamide and showed
no residual unsaturation. Thus, it had. apparently undergone
cyclocopolymerization. The excess acrylonitrile in the
polymer was most likely due to its strong-tendency to under-
go homopolymerization. The copolymer also exhibited a
broad absorption in its infrared spectrum at 1170 cm- which
corresponds with the phosphine oxide; thus, the phosphine
probably was oxidized during the work-up of the polymer.
Using infrared techniques it was possible to deter-
mine the presence of phosphine oxide in the monomers, the
amount or absence of residual unsaturation in the polymers,
and also the purity of the phosphines. These determinations
were mentioned earlier and also the use of this technique
was discussed in great detail by Skinner (24). The
spectral data obtained for the vinyl substituted phosphines
weredifferent, in that, the non-conjugated carbon-carbon
stretching vibration at 1640 cm-1 present in all the other
phosphine spectra was absent in the vinylphosphines. This
was probably due to the conjugation of the double bond with
the pair of electrons on phosphorus. The nuclear magnetic
resonance of these compounds did, indeed, show that there
was conjugation of this type present. The CH and CH2 out of
plane deformations at 990 and 910 cm- respectively, were
present in the spectra of all the compounds, thus, it was
felt that the vinyl substituted phosphines merely did not
show the expected peak at 1640 cm-1 due to its conjugation.
In allyldimethylphosphine, a medium band was found
at 1290 cm-1 which corresponded to an asymmetric CH3 de-
formation mode. This peak had shifted in the polymer to
1305 cm-1. This shift was expected, due to the change from
trivalent to pentavalent linkages in the phosphorus molecule.
This had been described by Bellamy (38). A strong band at
1630 cm-1 was found for the vinyl stretching vibration in
the dimethyl substituted phosphine. This peak disappeared
in the polymer; thus, it was felt that the polymer was
Source and Purification of Materials
The diphenylchlorophosphine and dichlorophenyl-
phosphine were generously donated by Victor Chemical
Company. They were distilled under nitrogen before use.
From Peninsular ChemResearch Incorporated was
purchased ethylene oxide, phosphorus tribromide, vinyl-
magnesium chloride in tetrahydrofuran, and methylmagnesium
bromide in ethyl ether, all of which were used as received.
Also purchased from the same source were dimethylamino-
dichlorophosphine, tetrahydrofuran, and allylbromide which
were redistilled before use.
The majority of the chemical analyses were determined
by Galbraith Laboratories of Knoxville, Tennessee. The
analyses of the phosphonium salts and polymers were
generally low in carbon; therefore, a method was devised by
the university analytical department which in many cases
gave superior results. Where these are reported, it will
be so noted.
All temperatures reported are uncorrected and
reported in degrees centigrade. Pressures are expressed
in millimeters of mercury, having been determined by means
of either a Zimmerli or McLeod gauge.
Infrared spectra were obtained with a Perkin-Elmer
Infracord Double-beam Infrared Recording Spectrophotometer,
while Ultraviolet spectra were run on a Bausch and Lomb
Spectronic 505 Recording Spectrophotometer.
Refractive indices were determined on a Bausch and
Lomb Abbe 34 Refractometer equipped with an anchromatic
Nuclear magnetic resonance spectra were obtained on
a Varian U-4302 High Resolution Nuclear Magnetic Resonance
Melting point determinations were carried out in
open capillary tubes in a Thomas-Hoover Melting Point
Intrinsic viscosities were calculated from efflux
times of solution through a Cannon-Ubbelohde Semi-Micro
Dilution Viscometer at 250 in a constant temperature bath.
Preparation of Intermediates
Allylmagnesium bromide.-This reagent was prepared
by the method of Grummitt, Budewitz, and Chudd (39), To
a five-liter, three-necked flask, equipped with a reflux
condenser, addition funnel and mechanical stirrer, was
added 85 g. (3.5 g.-atoms) of magnesium turnings and 1.75
liters of dry ethyl ether. The flask was flushed with dry
nitrogen. Two hundred and twelve grams (1.75 moles) of
allyl bromide was weighed out and a few milliliters added
to the flask, with stirring. When the ether began to
reflux, the mixture was cooled with an ice bath, and the
remainder of the allyl bromide, diluted with an equal
volume of ether, was added dropwise over a period of nine
hours. During this time the flask was cooled in the ice
bath. When the addition was complete, the mixture was
stirred for an additional hour and then two 1 ml. aliquots
were titrated. The solution was found to contain 1.65
equivalents of the allylmagnesium bromide; yield, 94 per
Nethallylmagnesium chloride.-Using the same method
(59) as was used for the preparation of allylmagnesium
bromide, 48 g. (2 g.-atoms) of magnesium turnings were
placed in a one-liter, three-necked flask fitted with a
stirrer, condenser, and constant pressure addition funnel.
The flask was flushed with dry nitrogen and a constant
flow was maintained during the reaction. Dry ethyl ether
(500 ml.) was added and the flask placed in an ice bath.
One ml. of ethyl bromide was added and the reaction mass
was stirred for ten minutes. Ninety-one grams (1 mole) of
methallyl chloride in an equal volume of ether was added,
with stirring, over a four-hour period. The smooth
suspension produced was too thick to analyze by titration.
Allylcarbinol.-This compound was prepared by a
modification of the procedure of Ramsden, Leebrick, et al.
(40). To 382 ml. (1 mole of 2.62 molar) of vinylmagnesium
chloride in tetrahydrofuran was slowly added 44 g. (1 mole)
of ethylene oxide dissolved in 100 ml. of tetrahydrofuran.
The temperature rose immediately and was maintained'at
55-600 during the addition, after which the reaction
mixture was refluxed for one hour. The tetrahydrofuran
was then removed at water aspirator pressure until a thick
paste formed which was poured into a solution of 100 g. of
ammonium chloride dissolved in 500 ml. of water. The
hydrolysis was quite exothermic. The layers failed to
separate until 100 ml. of 37 per cent hydrochloric acid
was added. The organic layer was removed and the water
layer extracted with two 150 ml. portions of ether. The
organic portions were combined, dried over anhydrous sodium
sulfate, and fractionated. The alcohol was removed by
distillation at 112-1150 through an 80 cm. Vigreux at
atmospheric pressure to yield 48.2 g. (67%) of allylcarbinol,
nD 1.4255, lit. (41), n2 1.4234, b.p. 112-1140.
4-Bromo-l-butene.-The method of Gaubert, Linstead,
and Rydon (29), was used in this synthesis. To a nitrogen
flushed one-liter, four-necked flask, in an ice bath,
equipped with a stirrer, condenser, thermometer, and con-
stant pressure addition funnel was added a solution of
145 g. (2 moles) of allylcarbinol and 60 g. of pyridine.
To the solution was slowly added 300 g. (1 mole) of
phosphorus tribromide. The addition required three hours,
while the temperature was kept below 150. The reaction
was refluxed one-half hour and distilled until thick white
fumes appeared. The distillate was washed with 1001ml.
of 10 per cent sodium hydroxide and then two 100 ml.
portions of water. The product was distilled through an
80 cm. Vigreux column to yield 185.6 g. (69%) of 4-bromo-
1-butene, b.p. 98-1000, nD 1.4628, lit. (42), b.p. 99,
3-Butenylmagnesium bromide.-To a nitrogen flushed
two-liter, four-necked flask, in an ice bath, equipped as
in the previous experiment, containing 36.5 g. (1.5 g.-
atoms) of magnesium and enough dry ether to cover it was
added 1 ml. of 4-bromo-l-butene. The reaction was started
and 150 ml. of dry ether added. A solution of 135 g. (1
mole) of 4-bromo-l-butene in 250 ml. of dry ether was added
over a three-hour period keeping the temperature below 200.
After complete addition the reaction was stirred another
three hours. The solution was titrated and found to be 95
per cent 5-butenylmagnesium bromide.
Preparation of Phosphines
A general method, developed by Jones and Davies
(26), was used in all the aryl substituted phosphine
syntheses. The phosphine preparations were carried out in
a one-liter, four-necked flask, equipped with a stirrer,
thermometer, condenser, and constant pressure addition
funnel allowing nitrogen to flow slowly through the system
at all times. The temperature during the reaction of the
Grignard reagent with the phosphorus halide was maintained
at 10-150. The phosphines were always prepared, stored,
and reacted under an atmosphere of dry nitrogen. If any
reactions were carried out under conditions which varied
from the above description, these changes will be noted
in the specific experiment.
Diphenylvinylphosphine.-One mole of vinylmagnesium
chloride in tetrahydrofuran was cooled in an ice bath and
88 g. (0.4 mole) of diphenylchlorophosphine in 100 ml. of
dry ether was slowly added. After complete addition, the
mixture was refluxed for two hours then cooled to 100. A
solution of 100 g. of ammonium chloride in 500 ml. of water
was slowly added (very exothermic). To the emulsion formed
was added 300 ml. ethyl ether and the layers separated.
The aqueous layer was extracted with two 100 ml. portions
of ether, and the organic portions were combined and dried
over anhydrous sodium sulfate. The solvents were removed
and the residual oil distilled through an 80 cm. Vigreux
column at 0.25 mm., to yield 56.2 g. (67%) of diphenyl-
vinylphosphine, b.p. 1040, n19 1.6289, d0 1.054 g./cc.,
lit. (43), b.p. 105.5-104.8/0.24 mm., n27 1.6229.
Allyldiphenylphosphine.-To 1 mole of allylmagnesidm
bromide in 250 ml. of dry ether, at ice temperature, was
added 88 g. (0.4 mole) of diphenylchlorophosphine over a
two hour period. The mixture was refluxed one-half hour
and again cooled in an ice bath. The reaction mass was
hydrolyzed using a solution of 100 g. of ammonium chloride
in 500 ml. of water. The organic layer was separated and
combined with two 100 ml. ether extractions of the water
layer. The combined ether solutions were then dried over
anhydrous magnesium sulfate. The liquid was decanted into
a distilling flask fitted with a nitrogen bubbler and a
vacuum-jacketed Vigreux column. The ethyl ether was
stripped off at reduced pressure and the product distilled
at 0.4 mm. to yield 60 g. (66%) of allyldiphenylphosphine,
b.p. 106-108, n21 1.6197, d0 1.033 g./cc., lit. (44),
b.p. 114/0.7 mm.
Diphenylmethallylphosphine.-To a solution of 1 mole
of freshly prepared methallylmagnesium chloride in ethyl
ether, at ice temperature, was added 88 g. (0.4 mole) of
diphenylchlorophosphine in 250 ml. of dry ether over a two-
hour period. The mixture was then refluxed for one-half
hour and again cooled in an ice bath. It was hydrolyzed
with a solution of 100 g. of ammonium chloride in 500 ml.
of water. The organic layer was removed and the aqueous
phase extracted with two 150 ml. portions of ethyl ether.
The ethereal extracts were combined with the organic
portion and this solution was dried over anhydrous sodium
sulfate. The ether was removed at atmospheric pressure
and the product distilled through an 80 cm. Vigreux column
at 0.25 mm. to yield 61 g. (63%) of diphenylmethallyl-
phosphine, b.p. 110-111, n20 1.6089, d0 1.051 g./cc; lit.
(24), b.p. 118-1200/0.45 mm.
Anal. Calcd. for C16H17P: C, 79.98; H, 7.13;
P, 12.89. Found: C, 79.78; H, 7.25; P, 12.81.
3-Butenyldiphenylphosphine.-To 0.7 mole of freshly
prepared 3-butenylmagnesium bromide in ethyl ether, at ice
temperature, was added 88 g. (0.4 mole) of diphenyl-
phosphinious chloride in 200 ml. of dry ether. The reaction
was very exothermic and the addition required four hours.
The solution was allowed to come to room temperature and
was stirred overnight under nitrogen. It was then
hydrolyzed with a solution of 100 g. of ammonium chloride
in 500 ml. water. The organic layer and three 100 ml.
extracts of the water layer were combined and dried over
anhydrous sodium sulfate. The ether was stripped off at
reduced pressure and the product fractionated through an
80 cm. Vigreux column. The yield of 3-butenyldiphenyl-
phosphine was 79.5 g. (85%), b.p. 115-1150/0.3 mm., nD
1.5991, d0 1.028 g./cc.
Anal. Calcd. for C16H17P: C, 79.98; H, 7.13;
P, 12.89. Found: C, 79.80; H, 7.02; P, 12.95.
Divinylphenylphosphine.-To a solution of 1 mole of
vinylmagnesium chloride (2.56 molar in tetrahydrofuran)
was added 45 g. (0.25 mole) of dichlorophenylphosphine in
50 ml. of dry benzene; the addition required six hours.
At the end of this time the solution was refluxed one hour
and then cooled to ice temperature. The reaction mass was
decomposed by adding a solution of 100 g. of ammonium
chloride in 500 ml. of water. An emulsion and a large
quantity of yellow solid were formed. The solid was
filtered off and washed with 150 ml. of ether. The layers
were separated, and the water layer extracted with two
150 ml. portions of ether; the combined organic portions
were dried over anhydrous magnesium sulfate. The solvents
were removed at water aspirator pressure and the residue
was distilled under vacuum. The product was a water white
liquid boiling at 520/0.8 mm., giving a yield of 19 g. (47%)
of divinylphenylphosphine, n20 1.5822; lit. (27), b.p.
54.5-550/1o5 mm., n2 1.5832.
Diallylphenylphosphine.-To 1.75 moles of freshly
prepared allylmagnesium bromide, at ice temperature, was
added 87 g. (0.486 mole) of dichlorophenylphosphine in
150 ml. of dry ether. The temperature was kept at 20-250.
The addition required two hours. The reaction was then
refluxed two hours, cooled in an ice bath, and hydrolyzed
with a solution of 100 g. of ammonium chloride in 500 ml.
of water. The layers were separated and the aqueous layer
extracted with two 200 mli portions of ether. The organic
portion and ether extracts were combined and dried over
anhydrous sodium sulfate. The ether was distilled and the
residue fractionated at 0.7 mm. through an 80 cm. Vigreux
column to yield 41.2 g. (44.6%) of diallylphenylphosphine,
bp. 790, n25 1.5676; lit. (45), b.p. 1180/7-8 mm.,
Di-3-buteniylhenylphosphine.-To a solution of 0.38
mole of 3-butenylmagnesium bromide in ether, cooled in an
ice bath, was added 27 g. (0.15 mole) of dichlorophenyl-
phosphine. The temperature was held at 5-100 during
addition and then raised to reflux for two hours. At this
time, there was a large quantity of white pasty solid,
which had precipitated out of the solution. The reaction
was cooled to ice temperature and 250 ml. of 20 per cent
ammonium chloride solution was slowly added. The layers
separated and the water layer was extracted with two 100 ml.
portions of ether. The organic portion was combined with
the ether extracts and dried over anhydrous sodium sulfate.
The product was fractionated through an 80 cm. Vigreux
column at 0.3 mm. to yield 13 g. (535) of di-3-butenyl-
phenylphosphine, b.p. 85-88, nD2 1.5448, d 0.9467 g./cc.
Anal. Calcd. for C H- P: C, 77.00; H, 8.78;
-- 14 -9
P, 14.20. Found: C, 76.39; H, 9.34; P, 14.07.
Dimethylaminodimethylohosohine.o-This compound was
prepared by the method described by Burg and Slota (28).
To 97 g. (0.66 mole) of dichlorodimethylaminophosphine in
1 liter of dry ether at -500 was added 460 ml. (1.32 moles
of 2.87 molar) of methylmagnesium bromide in ethyl ether.
A white solid formed immedia-ely upon addition. After
about three-fourths of the Grignard solution had been added,
the stirrer was stopped by the solid material which formed.
When the addition was complete, the flask was allowed to
warm to room temperature and th. solid suspension could be
stirred. After stirring one day, the ether was removed by
use of a filter stick and the solid washed with three 250 ml.
portions of ether. The solutions were combined, dried, and
the ether removed by distillation. The product which was
distilled through a Vigreux column at 990 resulted in a
yield of 14 g. (18.5%) of dimethylaminodimethylphosphine.
Lit. (28), b.p. 99.40.. Densities and refractive indices
were not run for the non-aromatic phosphines, as they were
spontaneously ignited in air.
Allyldimethylphosphine.-Following the general method
described by Burg and Slota (28), 11.5 g. (0.10 mole) of
dimethylaminodimethylphosphine and 20 ml. of dry ether were
added to a nitrogen flushed flask and cooled to 100. To
this solution was added 60 ml. (0.102 mole) of allyl-
magnesium bromide. There was no exothermic reaction, so
the flask was allowed to warm to room temperature. After
about five hours, a dark colored oil had separated from
the solution and the ethereal layer was clear. After one
day of stirring, the oil had become solid and the ethereal
solution was removed by use of a filter stick which was
dipped into the solution. The ether was distilled and the
residual oil fractionated on a spinning band column. The
fraction boiling from 91-990 gave 3.6 g. (35.2%) of dimethyl-
allylphosphine. The infrared spectrum indicated that the
desired product was obtained; however, a strong band at
1170 cm-1 indicated that some oxidation to the oxide had
Anal. Calcd. for C 5H P: C, 58.80; H, 10.86.
Found: C, 58.42; H, 11.58.
Preparation of Phosphonium Salts
The general method for preparation of phosphonium
compounds, as described by Skinner (24), was followed. A
solution of the appropriate alkyl halide was added to a
100 ml. three-necked flask containing the phosphine.
Following this, the flask was removed from the distillation
apparatus and stoppered under an atmosphere of nitrogen.
All analyses referenced with an (*) were determined by
C. D. Miller of the Analytical Division, Department of
Chemistry, University of Florida.
3-Butenyldiphenylvinylphosphonium bromide.-To 10 g.
(0.047 mole) of diphenylvinylphosphine in a nitrogen flushed
flask was added 20 ml. of dry reagent grade acetone and
21 g. (0.16 mole) 4-bromo-l-butene. The flask was stoppered
and allowed to stand at room temperature for three days, at
which time the solid which had been produced was filtered,
washed with acetone and collected. A yield of 10 g. (61%)
of 3-butenyldiphenylvinylphosphonium bromide, m.p. 163-170,
was obtained. The sample was slightly impure but further
attempts at purification were unsuccessful.
Anal. Calcd. for C18H20PBr: C, 62.26; H, 5.81;
P, 8.72. Found: C*, 61.07; H, 6.12; P, 8.85.
Allyldiphenylvinylphosphonium bromide.-To 24.6 g.
(0.11 mole) of diphenylvinylphosphine in a nitrogen
flushed, ice cooled, flask was added 15 g. (0.123 mole)
of allyl bromide in 40 ml. of dry ether. The reaction
was exothermic and even in an ice bath was very rapid.
Within two hours, the flask was filled with a white solid.
The solid was removed and washed in an Erlenmeyer flask
with a large excess of acetone. A yield of 12.5 g.
(32.3%) of allyldiphenylvinylphosphonium bromide, a white
powdery solid, m.p. 169-1710, was collected.
Anal. Calcd. for C17H18PBr: C, 61.28; H, 5.44;
P, 9.23. Found: C, 61.18; H, 5.87; P, 9.18.
Diphenylmethallylvinylphosphonium chloride.-To 10 g.
(0.047 mole) of diphenylvinylphosphine in 50 ml. of dry
ether was added 17 g. (0.16 mole) of methallyl chloride.
The flask was sealed and allowed to stand twenty days,
under nitrogen. At this time, a small amount of white
crystalline deposit had formed, was collected and washed
with dry ether. In this and three other attempts it was
not possible to isolate a pure sample of the phosphonium
Allyl-3-butenyldiphenylphosphonium bromide.-To 30 g.
(0.155 mole) of freshly distilled 3-butenyldiphenyl-
phosphine in 50 ml. of dry ether was added 48.5 g. (0.4 mole)
of allyl bromide. The flask was placed in a beaker of ice.
After the ice melted the reaction was allowed to come to
room temperature. It was left three days, filtered, washed
with ether and collected. The product was dried in a
vacuum desiccator and gave a yield of 37 g. (87.8%) of
allyl-3-butenyldiphenylphosphonium bromide, melting at
Anal. Calcd. for C19H22PBr: C, 63.17; H, 6.14;
P, 8.57. Found: C, 63.15; H, 6.24; P, 8.57.
24 g. (0.1 mole) of 3-butenyldiphenylphosphine was added
36.2 g. (0.4 mole) of methallyl chloride in 50 ml. of dry
ether. The flask was sealed under nitrogen and allowed to
stand twelve days, at which time long needle like crystals
were collected and washed with ether. The mother liquor
and washings were allowed to stand another twelve days and
more crystals were collected. The total yield was 32 g.
(97%) of 3-butenyldiphenylmethallylphosphonium chloride,
melting at 175-1770.
Anal. Calcd. for C20H24PCl: C, 72.60; H, 7.31;
P, 9.36. Found: C, 72.06; H, 7.07; P, 9.60.
Di-3-butenyldiphenylphosphonium bromide.-To a
nitrogen flushed flask containing 22.5 g. (0.093 mole) of
3-butenyldiphenylphosphine was added 13.5 g. (0.1 mole)
4-bromo-l-butene in 25 ml. dry ether. The flask was sealed
and allowed to stand at room temperature for fifteen days,
after which the crystalline solid was filtered, washed,
and dried. The product was 9.3 g. (35%) of di-3-butenyl-
diphenylphosphonium bromide, m.p. 115-117.
Anal. Calcd. for C20H24PBr: C, 63.85; H, 6.45;
P, 8.25. Found: C, 63.36; H, 6.74; P, 8.57.
Diallyldiphenylphosphonium bromide.-To 22.6 g.
(0.1 mole) of allyldiphenylphosphine in 100 ml. of dry
ether, at 0, was added 48 g. (0.4 mole) of allyl bromide.
The reaction was quite rapid and fairly exothermic. After
two hours, it was allowed to come to room temperature.
Three days later the precipitate which formed was filtered,
washed with ether, collected and dried in a vacuum desic-
cator. A yield of 30.5 g. (88%) of diallyldiphenyl-
phosphonium bromide, m.p. 173-1760; lit. (24), 155-1570,
Anal. Calcd. for C18H20PBr: C, 62.25; H, 5.80.
Found: C*, 62.29; H*, 5.78.
Diallyldiphenylphosphonium chloride.-To 22.6 g.
(0.1 mole) of freshly distilled allyldiphenylphosphine,
under nitrogen, was added 31 g. (0.4 mole) of allyl
chloride in 100 ml. of dry ether. The salt did not
crystallize at first; instead, a thick oil formed. After
a week, fine white crystals began to form and in approxi-
mately a month the white needle like crystals were
collected. These crystals, which smelled strongly of
phosphine, were shaken for three hours, with a large
excess of dry ether. They were recollected and dried,
yielding 12.5 g. (53%) of fine white crystals, m.p. 152-
1600. Attempted recrystallizations were unsuccessful.
Anal. Calcd. for C18H22PBr: C, 71.46; H, 6.64;
P, 10.25. Found: C*, 71.87; H', 6.96; P, 10.19.
Ally ldiphen.,ylropy7lphosphonium bromide.-To 22.6 g.
(0.1 mole) of freshly distilled allyldiphenylphosphine in
a nitrogen flushed flask was added 48.8 g. (0.4 mole) of
n-propyl bromide in 100 ml. of dry ether. The sealed flask
was allowed to stand for ten days. The white crystals
which had separated were collected, washed with 200 ml. of
dry ether, and dried in a vacuum desiccator at 600. A
yield of 10 g. (28%) of long white crystals, m.p. 134-1570
Anal. Calcd. for Co8H22PBr: C, 61.90; H, 6.51;
P, 8.88. Found: C, 61.67; H, 6.53; P, 9.01.
Diallyldimethylphosphonium chloride.-In a nitrogen
filled dry box 2.0 g. (0.019 mole) of allyldimethyl-
phosphine was added to 5.7 g. (0.075 mole) of allyl
chloride dissolved in 5 ml. of dry ether in a flask. The
flask was sealed and removed from the dry box. After four
days the white crystals which formed were removed by
filtration and washed with 25 ml. of dry ether. The salt
was very hygroscopic and was dried in a drying pistol over
phosphoric anhydride at 550. A yield of 2.9 g. (82%) of
a solid, melting at 76-820 in a sealed tube, was obtained.
It was found to be impossible to remove the last traces
of water from this compound under the conditions used.
Thus, the analysis indicates the molecule crystallized with
one mole of water of hydration.
Anal. Calcd. for C8H16PC1: C, 53.88; H, 9.03;
P, 17.34. Calcd. for C8H16PC1-H0: C, 48.85; H, 9.23;
P, 15.40. Found: C, 48.27; H, 9.42; P, 15.40.
Polymerization of Phosphonium Salts
All of the phosphonium compound polymerizations were
run in solution in 5 ml. sealed ampoules. The solution
in the ampoule was always frozen and thawed twice under a
flow of dry nitrogen to remove dissolved oxygen from the
reaction. The ampoule was then sealed and the polymeri-
zation completed. Three different free radical initiators,
designated A, B, and C in the discussion which follows,
were used in the case of certain of the monomers. The
analyses of all the polymers indicated that they retained
one mole of water per phosphonium unit. These results are
consistent with the work reported by Skinner (24).
1 g. of 5-butenyldiphenylvinylphosphonium bromide dissolved
in 1 g. of water was added 2 per cent (0.02 g.) of benzoin,
and the solution was placed in front of an ultraviolet
light source, in a sealed ampoule, in an oven at 600 for
a period of six days. The ampoule was opened and the
contents poured into 200 ml. of stirred reagent grade
acetone. No polymer was produced. The polymerization was
repeated using ethanol as solvent, but again no polymer
Poly-(allyldiphenylvinylphosphonium bromide).-To a
solution of 1 g. of allyldiphenylvinylphosphonium bromide
dissolved in 1 g. of water, in an ampoule, was added 1.8
per cent (0.018 g.) of benzoin. The ampoule was sealed
and placed in front of an ultraviolet light source in an
oven at 600 for five days. The contents of the ampoule
were then poured into 200 ml. of stirred reagent grade
acetone and the solid which separated was recovered. The
product was dried in a vacuum desiccator at 600. After
drying, 0.42 g. (42%) of a white powder, softening at 2200
and melting at 5100, was obtained. The intrinsic viscosity
[n], run in 1.0 M potassium bromide in ethanol-water (1:1)
Anal. Calcd. for (C17H18PBr-H20)n: P, 8.82; Br,
22.76. Found: P, 8.71; Br, 22.45.
A-.-To a solution of 1 g. of allyl-5-butenyldiphenyl-
phosphonium bromide in 1 g. of water, in an ampoule, was
added 2 per cent (0.02 g.) of benzoin. The ampoule was
then sealed and placed in an oven at 600, in front of an
ultraviolet light source, for a period of three days.
The solution produced was poured into 200 ml. of stirred
reagent grade acetone and the solid which separated was
recovered. The product was dried in a vacuum desiccator
and gave 0.58 g. (58%) of a solid which softened at 2000,
and melted at 2600. The intrinsic viscosity [t] run in
0.5 M potassium bromide in ethanol-water (3:1) was 0.046.
Anal. Calcd. for (C19H22PBr-H20)n- P, 8.18;
Br, 21.10. Found: P, 8.12; Br, 20.90.
-B.-To a solution of 1 g. of allyl-5-butenyldiphenyl-
phosphonium bromide in 1 g. of dimethylformamide, in an
ampoule, was added 2 per cent (0.02 g.) of a, c~-azodi-iso-
butyronitrile (AIBN). The ampoule was then placed in an
oven at 620 for seven days. During this time the solution
darkened considerably and did not gain noticeably in
viscosity. The solution was decanted into stirred ethyl
ether and a very small amount of tan solid was produced.
It was recovered, dried, and gave a yield of 0.02 g. (2%).
-C.-To a solution of 1 g. of allyl-3-butenyldiphenyl-
phosphonium bromide in 1 g. of water, in an ampoule, was
added 1 drop of 60 per cent tert-butylhydroperoxide. The
solution was allowed to stand at 620 for seven days. The
slightly orange solution produced was decanted into 200 ml.
of stirred acetone. The flocculent precipitate was re-
covered and dried in a vacuum desiccator. While drying,
the product became very brittle. It gave a yield of 0.1 g.
(10%;) of polymer, softening at 1800 and melting at 2200.
A-. A solution of 1 g. of 5-butenyldiphenylmethallyl-
phosphonium chloride was treated as in part A above. After
seven days the sample was poured into 200 ml. of stirred
acetone. The product was a voluminous white solid which,
upon recovery, turned to an oil. The oil was redissolved
in ethanol and reprecipitated twice from acetone. A yield
of 0.60 g. (60S) of brittle white solid, softening at 2100
and melting at 3000 was recovered after drying in a vacuum
desiccator. The intrinsic viscosity [I], run in 1.0 H
potassium chloride in ethanol-water (1:1) was 0.058.
-B.-A solution of 1 g. of 5-butenyldiphenylmethallyl-
phosphonium chloride was treated as in part B above. After
seven days in the oven at 620 the temperature was raised
to 800 for two additional days. The dark orange solution
produced was then poured into 250 ml. of dry ether and the
precipitate recovered. During its drying in a vacuum
desiccator, however, it became very brittle. A yield of
0.15 g. (15%) of polymer, softening at 2600 and melting
at 3000, was obtained.
-C.-A solution of 1 g. of 3-butenyldiphenylmethallyl-
phosphonium chloride was treated as in part C above. After
seven days in the oven at 620 the temperature was raised
to 800 for two additional days. The light orange solution
produced was then poured into 200 ml. of dry acetone. The
white amorphous solid produced was collected and on drying
acted in the same manner as those described above. A
yield of 0.08 g. (8%) of polymer, softening at 2750 and
melting at 3030, was obtained.
solution of di-3-butenyldiphenylphosphonium bromide in 1 g.
of water was treated as in A above. After seven days in
front of an ultraviolet light source at 620, the solution
was poured into 200 ml. of acetone; but, no precipitate
formed. Impure monomer, melting at 97-1040, was recovered
by evaporation of the solvent and decanting the oil formed
-B.-A solution of 1.2 g. of di-3-butenyldiphenyl-
phosphonium bromide was treated as in B above. The sample,
after nine days, seven at 60 and two at 800, was decanted
into dry ether and no precipitate formed. The solvent was
evaporated, the residue dissolved in ethanol, and then
dropped slowly into stirred ether. The oil formed was
stirred until an impure dark solid formed, melting at 88-
950. This dark solid was probably impure monomer.
-C.-A solution of 1 g. of di-3-butenyldiphenylphos-
phonium bromide was treated as in C above. Again, no
polymer was recovered; however, the impure monomer could be
Poly-(diallyldiphenylphosphonium bromide)-A.-A solu-
tion of 1 g. of diallyldiphenylphosphonium bromide was
treated as in A above, except that only 0.1 per cent of
benzoin initiator was used. After seven days in front of an
ultraviolet light source, at 600, the water and oil had
separated into two layers. The water layer was decanted and
the oil dissolved in ethanol. The solution was then dropped
slowly into dry ether. A viscous sticky oil was formed,
which upon continued stirring became a white flaky solid.
The product was washed with 100 ml. of dry ether and dried
giving 0.80 g. (80%) of white solid, softening at 190-2000
and melting at 3050. The intrinsic viscosity [q], run in
0.1 M potassium bromide in ethanol-water (1:1) was 0.052.
-B.-A solution of 1.44 g. of diallyldiphenylphos-
phonium bromide was dissolved in three times its weight of
dimethylformamide. It was placed in an ampoule, charged
with 1 per cent (0.0144 g.) AIBN, and heated in an oven at
700 for four days. The solution was then poured into 200 ml.
of dry acetone and a white flocculent precipitate was ob-
tained. The solid was dried and a yield of 0.28 g. (19%) of
polymer, softening at 2500 and melting at 3200, was obtained.
-C.-A solution of 2.35 g. of diallyldiphenyl-
phosphonium bromide was dissolved in a similar amount of
water. The solution was treated as in C above, using 2 per
cent initiator. The sample was left at room temperature
for one day and then at 700 for three days. After this time
it was poured into 200 ml. of acetone and no precipitate was
recovered. Monomer was recovered by the previously described
method, m.p. 170-178.
Poly-(diallyldiphenylphosphonium chloride).-To a
solution of 1 g. of diallyldiphenylphosphonium chloride,
dissolved in 1 g. of water, in an ampoule, was added 1 per
cent (0.01 g.) of benzoin. The ampoule was sealed and ir-
radiated at 600 by an ultraviolet light source for seven
days. The solution produced was poured into 200 ml. of dry
acetone and the solid precipitate collected. A yield of
0.74 g. (74%) of polymer, softening at 2700 and melting at
3000 was obtained. The intrinsic viscosity [is], run in 1 M
potassium chloride in ethanol-water (1:1) was 0.079.
Anal. Calcd. for (C18H20PC1*H20)n: P, 9.67; Cl,
11.08. Found: P, 9.41; Cl, 10.77.
Polyr-(diall7ldimethylohorphonium chloride) .-To a
solution of 1 g. of diallyldimethylphosphonium chloride,
dissolved in 1 g. of water, in an ampoule was added 1 per
cent (0.01 g.) of benzoin. The ampoule was sealed and
irradiated at 500 by an ultraviolet light source. After
two hours the solution was still very fluid. One day
later, the sample was a transparent gel. The gel was
insoluble in ethanol. It was ground and extracted with
ethanol in a soxhlet extractor. The solid was then removed
and dried. Eighty per cent (0.80 g.) of crosslinked polymer
was obtained. The extraction solvent was evaporated and 5
per cent (0.05 g.) of soluble polymer was obtained. The
soluble portion softened at 1500 and melted at 270.
Poly-(diallylphenylphosphine oxide).-To 7.35 g. of
poly-(diallyldiphenylphosphonium bromide) dissolved in
25 ml. of methanol was added a solution of 4 g. of potassium
hydroxide in 25 ml. of water. The colorless solution
immediately changed to yellow and, on heating, to orange,
then red, and after three days at reflux a gummy brown
solid separated out of the clear solution. The solution
was decanted and the solid extracted in a soxhlet extractor
with ethanol. The solution was evaporated until approxi-
mately 10 ml. was left. This solution was then dropped
slowly into stirred ether and a fine precipitate was col-
lected. An infrared spectrum showed strong absorption at
1170 cm-1, thus, indicating that poly-(diallylphenyl-
phosphine oxide) was the product. A yield of 4.2 g. (96%)
of a light tan solid, softening at 2600 and melting at
3200 was collected. The intrinsic viscosity [rk], determined
in ethanol, was 0.109.
Anal. Calcd. for: (C12H17PO-H20): C, 64.39; H,
7.64; P, 13.80. Found: C, 64.89; H, 7.39; P, 12.67.
This analysis was run by Schwarzkopf Microanalytical
Preparation of Copolymers
Divinylphenylphosphine and acrylonitrile-A.-To a
solution of 1.34 g. (0.0083 mole) of divinylphenylphosphine
and 0.456 g. (0.0087 mole) of acrylonitrile dissolved in
5 g. of dimethylformamide, in an ampoule, was added 1 per
cent (0.0183 g.) of AIBN. The ampoule was sealed and
allowed to stand in an oven at 600 for four days. The
very fluid red solution was poured into 250 ml. of dry
ether. An orange solid precipitated, was filtered, and
dried in a vacuum desiccator. A yield of 0.6 g. (34%) of
a brittle red solid softening at 1100 and melting at 2000
was obtained. The infrared spectrum indicated that the
phosphine had been largely converted to the phosphine oxide.
The intrinsic viscosity [k], determined in dimethylformamide,
Anal. P, 9.535 = 0.265 mole fraction of divinyl-
phenylphosphine oxide; i, 11.89S = 0.732 mole fraction of
-B.-To a solution of 1 g. (0.00625 mole) of divinyl-
phenylphosphine and 0.67 g. (0.0128 mole) of acrylonitrile
dissolved in 5 g. of dimethylformamide, in an ampoule, was
added 1 per cent (0.0161 g.) of AIBI. The ampoule was
sealed and allowed to stand in an oven at 600 for three
days. The brilliant red solution which formed was then
poured into 200 ml. of stirred ether and the precipitate
collected and dried in a vacuum desiccator. A yield of
1.10 g. (65?') of an orange solid, which softened at 1200
and melted at 1950 was recovered. The infrared spectrum
showed that the phosphine had been converted to the
phosphine oxide during the work-up. The intrinsic
viscosity, determined in dimethylformamide, was 0.108.
Anal. P, 8.54% = 0.215 mole fraction divinylphenyl-
phosphine oxide; IT, 14.40% = 0.785 mole fraction of
Monomers functionally capable of forming five-, six-,
seven-, and eight-membered rings, when polymerized by the
cyclic mechanism, were synthesized. A free radical
solution polymerization of the monomers gave polymers which
were soluble, and essentially saturated in all cases except
for the case of an eight-membered ring. This, as might
have been expected, did not form a polymer. The solubility,
saturation and resemblance to the polyammonium compounds
seemed to indicate that these polymers contained cyclics
which were produced during polymerization by the intra-
molecular-intermolecular mechanism. The copolymer of
acrylonitrile with divinylphenylphosphine showed that the
phosphines, like the related ethers and amines could be
used in cyclocopolymerization. Conversion of the poly-
phosphonium compound to the phosphine oxide showed that
the term "intrinsic viscosity" of the salt is an ambiguous
statement and actually has little correlation with the term
used for non-electrolyte polymers.
1. C. Walling, J. Am. Chem. Soc., 67, 441(1945).
2. H. Stockmayer, J. Chem. Phys., 12, 125(1944).
>. G Butler and F. L. Ingley, J. Am. Chem. Soc., 73,
4. H. Staudinger and W. Heuer, Ber., 67, 1159(1934).
5. G. B. Butler and R. J. Angelo, J. Am. Chem. Soc., 79,
6. G. B. Butler, A. Crawshaw, and W. L. Miller, ibid., 80,
7. C. S. Marvel and R. D. Vest, ibid., 79, 5771(1957).
8. C. S. Marvel and J. K. Stille, ibid., 80, 1740(1958).
9. J. F. Jones, Italian Pat., 563,941, June 7, 1957, to
B. F. Goodrich Company.
10. J. F. Jones, J. Polymer Sci., 33, 15(1958).
11. A. Crawshaw and G. B. Butler, J. Am. Chem. Soc., 80,
12. G. B. Butler and M. D. Barnett, U. S. Dept. Com., Office
Tech. Serv., P. B. Report, 145, 435(1959).
13. C. S. Marvel and E. J. Gall, J. Org. Chem., 25,
14. N. D. Field, ibid., 25, 1006(1960).
15. C. S. Marvel and W. E. Garrison, J. Am. Chem. Soc., 81,
16. W. L. Miller and W. B. Black, Division of Polymer
Chemistry Preprints, 142nd National Meeting of the
American Chemical Society, Atlantic City, N. J.,
September, 1962, p. 345.
17. C. King, J. Am. Chem. Soc., 86, 437(1964).
18. G. B. Butler and R. W. Stackman, J. Org. Chem., 25,
19. G. B. Butler, J. Polymer Sci., 48, 279(1960).
20. E. Y. Chang and C. C. Price, J. Am. Chem. Soc., 83,
21. J. M. Barton, G. B. Butler, and E. C. Chapin, Division
of Polymer Chemistry Preprints, American Chemical
Society, Philadelphia, Pa., April, 1964, p. 216.
22. Chemical and Engineering News, Dec. 16, 1963, p. 42.
23. Ibid., Dec. 30, 1963, p. 33.
24. D. L. Skinner, Ph.D. Dissertation, University of
Florida, January, 1961.
25. K. D. Berlin and G. B. Butler, J. Am. Chem. Soc., 82,
26. W. J. Jones, W. C. Davies, et al., J. Chem. Soc., 1947,
27. M. I. Kabacknek, Chung-Yu Chang, and E. N. Tsvetkov,
Proceedings of the Academy of Sciences USSR. (English
Translation), 135, 1309(1960).
28. A. B. Burg and P. J. Slota, J. Am. Chem. Soc., 80,
29. P. G. Gaubert, R. P. Linstead, and H. N. Rydon,
J. Chem. Soc., 1937, 1921.
30. J. R. Van Wazer, Phosphorus and its Compounds, Inter-
science Publishers, Inc., New York, 1958, Vol. I, p. 43.
31. N. J. Leonard and D. M. Locke, J. Am. Chem. Soc., 77,
32. G. B. Butler, G. L. Statton, and W. C. Bond, unpublished
33. J. K. Stille, Chem. Revs., 58, 541(1958).
34. F. \W. Eillmeyer, Tex:tbook of Pol:ymer Chemistry, Inter-
science Publishers, Inc., iew York, i;7, p. 128.
35. R. I. Fuoss and U. P. Strauss, J. Polv-er Sci., 3,
36. G. E. Butler, Abstract of Paper, 153rd alstional
meeting c the Armeric.an Cnemical Society, San
Francisco, California, April, 1958, p. 6R.
37. G. E. Butler, C. F. Hauser, et al., Technical Report
io. ASD-TR-61-237, part I, September, 1961.
38. L. J. Bellamy, The Infra-red Spectra of Complex
Molecules, John Wiley and Sons, Inc., ilew Yor:, 1950,
39. 0. Grummitt, E. P. Budewitz, and C. C. Chudd, Org. Syn.,
40. H. E. Ramsden, J. R. Leebrick, et al., J. Ore. Chem.,
41. D. J. Foster and E. Tobler, ibid., 27, 834(1962).
42. N. S. Johary and L. N. Owens, J. Chem. Soc., 1955,
43. D. N. Paisley and C. S. Marvel, J. Polymer Sci., 56,
44. M. C. Browning, J. R. Mellor, et al., J. Chem. Soc.,
45. H. H. Sisler, H. S. Ahuja, and N. L. Smith, J. Org.
Chem., 26, 1819(1961).
William C. Bond, Jr. was born on August 4, 1938,
at Bethesda, Maryland. He attended elementary school at
Bethesda, Maryland and Evanston, Illinois. He received
his high school education while in attendance at Admiral
Farragut Academy at St. Petersburg, Florida. The author
registered in Wheaton College, Wheaton, Illinois, in
September, 1956, and was graduated in 1960 with a Bachelor
of Science degree in Chemistry. He entered the graduate
school at the University of Florida in September, 1960.
The author is a member of Gamma Sigma Epsilon and
the American Chemical Society.
Donna Allen Bond is the wife of the author and they
have two sons, William C. Bond, III, and Harold Bruce Bond.
This dissertation was prepared under the direction
of the chairman of the candidate's supervisory committee
and has been approved by all members of that committee.
It was submitted to the Dean of the College of Arts and
Sciences and to the Graduate Council, and was approved as
partial fulfillment of the requirements for the degree of
Doctor of Philosophy.
April 18, 1964
Dean, College of Arts;a~d Sciences
Dean, Graduate School