JOURNAL OF POLYMER SCIENCE: PART A-1 VOL.

6, 1671-1685 (1968)

Synthesis and Thermogravimetric

Analysis of Diol-Linked Tetrameric
HexaphenyldichlorophosphonitrilePolymers

A. J. BILBO, C. M. DOUGLAS, N. R. FETTER, and

D. L. HERRING, Naval Weapons Center, Corona Laboratories,
Corona, California 917’20

Synopsis
Several high molecular weight polymers have been prepared from the reactions of
hexaphenyldichlorophosphonitrile tetramer with p,p’-biphenol, 4,4’-oxydiphenol,
resorcinol, and benzidine. A simple fractionation procedure yielded samples of the
biphenol-linked polymer with molecular weights of approximately 500,000. DTA and
TGA studies of this polymer in a variety of environments indicate thermal degradation
is initiated in the phosphonitrile ring.

INTRODUCTION
The preparation of polymeric materials from the reactions of cyclic
phosphonitriles with aromatic dihydroxy linking agents has been described
by several authors. --‘I The reactions of hexaphenyldichlorophosphoni-
trile tetramer with several diols have been carried out in this laboratory
and have been reported.8 In this paper we wish to describe the prepara-
tion and fractionation of some of the more promising candidates from this
series, plus the synthesis of a polymer employing benzidine as a linking
agent. We also wish to present polymer stability data, based on thermo-
gravimetric analyses, which indicate the sources of degradation of polymers
in various enivronments.
The material of principal interest is the polymer I:

where -0- denotes

1671
 1672 BILBO, DOUGLAS, FETTER, HERRING

because it could be prepared from high-purity starting materials and the

crude product was easily fractionated by the procedure described below.
This polymer also has what was considered to be one of the simplest and
most stable diol-linking agents; hence, it would be the best candidate for
thermal stability studies. Polymers with the linking agents m-0-
Cdk-o-, p-NH-CsH4-CsH4--NH--, and p1p'-&c6H4-o-
CsH4-&, plus the model compound I1

I1
were also examined for a comparison of thermal stabilities.

EXPERIMENTAL
Apparatus
Molecular weights were determined on a Mechrolab PIiIodel502 membrane
osmometer at 37°C. with toluene as solvent. Solute concentrations ranged
between 1.0 and 4.0 mg./ml. for a set of four determinations for each
polymer sample. The instrument was calibrated regularly with standard
polystyrene polymers supplied by ARRO Laboratories, Inc., Joliet, Illinois.
Standards at 51,OOO, 97,200, 41l,OOO, and 860,000 number-average mo-
lecular weight were used.
Thermogravimetric analyses were carried out on a DuPont Model 950
thermal balance with a scan rate of 5"C./min. and gas flow of 1 ft.3/hr.
I n the one analysis done under vacuum, pressure was maintained at ap-
proximately 20 p throughout the run. I n the cases where wet atmospheres
were desired, the gas was passed via sintered glass bubbler through a water
column; otherwise, air, oxygen, and nitrogen were used directly from the
tank. All gases were determined to be dry and free of contaminants by
mass spectral analysis. Sample weights ranged from 3.5 to 65.0 mg.
Reagents
Hexaphenyldichlorophosphonitrile tetramer (m.p. 304-305"C.), I11 was
. prepared by the reaction of sodium azide with a 1 : l mixture of phenyldi-
chlorophosphine and diphenylchlorophosphinea and purified by recrystal-
lizing four times from dry toluene followed by vacuum drying a t 180°C.
for 24 hr. The p,p'-biphenol was purified by benzene extraction in a
Soxhlet apparatus for 24 hr. to remove p-hydroxybiphenyl, dried in vacu-
um for 18 hr. at 150°C. and then zone-refined through 75 passes on a
Sloan-McGowan zone melter. The 4,4'-oxydiphenol was recrystallized
twice from benzene and dried in vacuum at 150°C. for 24 hr. Benzidine
was recrystallized twice from anhydrous ethanol and vacuum-sublimed.
Biphenyl from Eastman Organic Chemicals was distilled twice under vac-
uum. Resorcinol (Mallinckrodt A. R. Grade) was sublimed twice in vucuo.
Xylene was distilled from lithium aluminum hydride.
 HEX APHENYLDICIILOROPHOSPHONITRILE 1673

Preparation of Hexaphenylbis(pheny1phenoxy)phosphonitriIe Tetramer

(11). A mixture of I11 (5.00 g., 0.007 mole) and phydroxybiphenyl
(3.93 g., 0.023 mole) was heated in a glass ampule attached to a vacuum
line. At 2O5-21O0C., the reaction was vigorous with the quantitative
amount of hydrogen chloride (0.014 mole) evolved in 6.5 hr. Excess
p-hydroxybiphenyl was separated by sublbation and the reaction product
was crystallized from acetonitrile to give 6.2 g. of 11, (m.p. 198-200°C.).
ANAL. Calcd. for C~HbsP4N402: C, 73.5%; H, 4.9%; P, 12.7%; N, 5.7%; 0,
3.3%; mol. wt. 980. Found: C, 73.4%; H, 4.9%; P, 12.6%; N, 5.7%; 0, 3.3%;
mol. wt. 967 (determined on a Mechrolab Model 301 vapor-phase osmometer).

PolymerizationReactions

Melt Reaction. I n a typical experiment I11 (14.26 g., 0.02 mole) and
p,p'-biphenol (3.72 g., 0.02 mole) were intimately mixed and loaded into
a heavy-walled glass ampule which was attached to a vacuum line. The
tube was heated by means of an oil bath a t 220°C. Evolution of hydrogen
chloride commenced after 1 hr. and the gas was continuously removed
from the reaction zone by condensation in a -196°C. trap. After 4 hr.,
when HC1 generation became slow, the temperature of the oil bath was
raised to 275°C. and heated 6 hr. more until gas evolution ceased. In-
frared analysis of the gas (435 ml. std. conditions 98%) indicated that it was
hydrogen chloride containing a trace of toluene. A small amount (35 mg.)
of white sublimate in the neck of the reaction ampule was identified as a
mixture of unreacted biphenol and ammonium chloride. The polymer
(IA), a pale yellow glass melting 185220°C. was soluble in common
organic solvents.
C, 69.7%; H, 4.6%; P, 15.0%; N, 6.8%i 0,
ANAL. Calcd. for CUIHS8P4N402:
3.9%. Found: C, 70.0%; H, 4.3%; P, 14.7%; N, 6.8%; 0, 4.0%; C1, 0. M , =
42,700 (crude product).
SolutionReaction-Biphenyl. A mixture of I11 (49.6532 g., 69.59 mmole),
p,p'-biphenol (12.9594 g., 69.60 mmole) and biphenyl (50 g.) was placed
in a 250-cc. round-bottomed flask with a 10-in. long neck fitted with a
detachable condenser. The apparatus was attached to a vacuum line
and provision was made for continuous purging by dry helium gas. The
reaction mixture was heated to the reflux temperature of biphenyl, 256"C.,
for 4 hr., during which time about 75% of the calculated amount of hy-
drogen chloride was evolved. The gas was collected in two liquid nitrogen
traps from which it could be passed to a measuring system on the vacuum
line. The entraining helium was passed through a calcium sulfate dry-
ing tube, a mercury bubbler, and a sodium hydroxide solution before it was
allowed to escape.
Heating at reflux was continued for 24 hr. ; at the end of this time 90%
of the total hydrogen chloride was evolved. Biphenyl was then distilled
off under vacuum after the detachable condenser was replaced by a receiver.
1674 BILBO, DOUGLAS, FETTER, HERRING

Although the reaction mixture was heated at 285°C. under vacuum for 5
hr. after the solvent had been removed, less than 1 mole-yo hydrogen
chloride was evolved. During this final heating period, a small amount of
material, which was identified as NH&l (70 mg.), together with a trace of
unreacted 111, sublimed from the melt.
The reaction flask was broken and the solid polymer was recovered, ground
in a mill and the resulting fine powder dissolved in benzene. The solution
was filtered to remove glass particles and then freeze-dried. The molec-
ular weight of the gross sample (IB) was 115,000.

.&$'-Oxydiphenol
Melt Reaction. I11 (7.13 g., 0.01 mole) was reacted with 4,4'-oxydi-
phenol (2.20 g., 0.01 mole) in a manner similar to that described for melt
reactions of p,p'-biphenol, with the exception that the reaction mixture
was also heated a t 320°C. for 2 hr.; 92 mole-yo of the theoretical amount
of hydrogen chloride was evolved. The sublimate (35 mg.) was examined
by infrared analysis and was shown to be largely [(C6H5)2PN]3, together
with traces of NH&1 and unreacted diol. The product was a yellow,
brittle resin which started to soften a t 150°C.
ANAL. Calcd. for C U H ~ P ~ N ~C,O 68.5%;
~: H, 4.5%; P, 14.7%; N, 6.7%; 0,
5.7%. Found: C, 68.7%; HI 4.5%; P, 14.9%; N, 6.8%; 0, 6.1%; C1, 0; ii?.,5000.

This polymerization was repeated and the reaction temperature was

held at 180-185°C. for 42 hr. and at 225230°C. for 8 hr. During the heat-
ing cycle, a total of 86.4% of the theoretical amount of hydrogen chloride
was generated. No sublimation of reactants or by-products was noted.
The properties of the resinous product were similar to those described
above, although the molecular weight (8250) was slightly higher.
Solution Reaction. A slurry of I11 (14.26 g., 0.02 mole) and 4,4'-
oxydiphenol (4.04 g, 0.02 mole) in 100 ml. anhydrous xylene was heated
under nitrogen with vigorous stirring. After 1 hr. a t reflux, hydrogen
chloride evolution began. The solution became clear after two weeks,
and the reaction was continued for an additional week until the ei3uent
gas was no longer acidic. The solvent was then evaporated in vacuo
and the solid product was heated a t 275°C. for 2.5 hr. The light-tan,
brittle resinous product had a molecular weight of 29,000. Fraction-
ational precipitation of this polymer from a benzene-heptane solution
yielded a first fraction (15% of sample weight) with a molecular weight
of 135,000.

Resorcinol-Melt Reaction
A mixture of I11 (12.5 g., 0.017 mole) and resorcinol (2.0 g., 0.017 mole)
was loaded into a glass ampule and was heated in vacuo at 220°C. After
10 hr. the evolution of gas became negligible, and the reaction tempera-
ture was raised to 265°C. for 2 hr. The hydrogen chloride generated was
 HEXAPHENYLDICHLOROPHOSPHONITRILE 1675

691 ml. std. conditions, 91% of the theoretical amount. The light-brown
resinous product softened over the range 130-150°C.
ANAL. Calcd. for C4zHa4P4N40z: C, 67.2%; H, 4.5%; P, 16.5%; N, 7.5%; 2,
4.3%. Found: C, 67.3%; H, 4.7%; P, 16.2%; N, 7.5%; 0, 4.1%; C1, 0; M,,
25,000.

Fractional precipitation of this product yielded a polymer, 12% by

weight, which had a molecular weight of 104,500.
Benzidine-Melt Reaction
An ampule containing 3.68 g. of benzidine (0.02 mole) and 14.26 g. of
I11 (0.02 mole) was heated gradually over a 6-hr. period to 290°C. The
reaction temperature was maintained for 16 hr. During this period 87.5%
of the theoretical amount of hydrogen chloride was evolved. Benzidine
dihydrochloride (55 mg.) was recovered as a sublimate. The product
began to soften at 205°C. and was insoluble in benzene.
ANAL. Calcd. for C ~ H N P ~ N
C,~69.9%;
: H, 4.9%; P, 15.0%; N, 10.2%. Found:
C, 69.5%; H, 5.2%; P, 15.2%; N, 10.4%; C1, 0.24%.

Polymer Fractionation
A 3.1-g. sample of polymer IA was dissolved in 325 ml. benzene. The
solution was heated at 55"C., and 75 ml. heptane was added slowly with
stirring. The solution was then allowed to cool overnight. The polymer
which had collected as an oil in the bottom of the container was recovered

TABLE I

Fract,ionation of +-m-s-4x Polymers

Fraction Cumulative
no. Weight, g. Weight % wt. % Bn
Fract,ionation of Polymer IA
1 0.64 20.6 20.6 259,000
2 0.46 14.8 35.4 183,600
3 0.52 16.8 52.2 77,700
4 0.70 22.6 74.8 .58, 300
5 0.43 13.9 88.7 24,800
6 0.08 2.6 91.3 16,400
7 0.15 4.8 96.1 7,460
8 0.02 0.7 96.8 3,720
9 0.01 0.3 97.1 1,600
Fractionat.ion of Polymer IB
1 0.495 9.5 9.5 541,000
2 1.22 23.5 33.0 233,000
3 1.22 23.5 56.5 65,900
4 0.665 12.8 69.3 23,300
-
1676 BILBO, DOUGLAS, FETTER, HERRING

by decanting the solution, dissolving the residue in 20 ml. of benzene, com-

pletely reprecipitating the polymer with 200 ml. heptane, and again de-
canting the solution. The fraction was dissolved in benzene, freeze-dried,
and weighed. The original benzene-heptane mother liquor was warmed
to 55OC. and additional heptane added to cause slight turbidity. The
second, and subsequent fractions were isolated in the manner described
above. The results of fractionations of two samples of p,p'biphenol-
linked polymer are given in Table I.

Vacuum Pyrolysis of 111-Biphenol Polymer

Samples of polymer IA, fraction 3, were pyrolyzed under vacuum to
recover and measure the amounts of volatile products. In a typical ex-
periment, the sample was placed in a 300 x 15 mm. tube and evacuated on
a vacuum system fitted with liquid nitrogen (- 196°C.) traps and Sprengel
pump for collecting noncondensable gases. The samples were heated
under static vacuum at about 550°C. for 15 hr., evolved benzene being
collected in the -196°C. traps. The hydrogen was pumped to a cali-
brated bulb and characterized by its mass spectrum. The sublimate was
collected from the reaction vessel above the heated zone and was char-
acterized as [(CeH&PN], by its infrared spectrum. The temperature of
550°C. was chosen for heating, because no further evolution of hydrogen or
benzene was observed above this point. The results of three pyrolyses are
shown in Table 11.

RESULTS AND DISCUSSION

Table I11 shows the results of temperature-scan pyrolyses of several
polymers in a variety of atmospheric environments. The arbitrary
criterion for stability was the temperature at which the initial sample weight
was decreased by 10%. This point was chosen rather than the initial
departure from the starting sample weight baseline, because this point was
not always easy to see and occasionally traces of solvent caused minor de-
partures from the baseline.
The data in Table I11 are presented to show the effects of polymer molec-
ular weight and gaseous environment on thermal stability. An examina-
tion of thermograms 5 , 6 , 7 , 8 , 9 , and 13 shows that the 10% decomposition
point of the same polymer varies only slightly with molecular weight. The
average 10% temperature was 473 5°C. The range of molecular
weights covered by these thermograms is from 9,200 (-10 monomer units)
to 50,000 (-600 monomer units). Thermogram 16 shows thermal sta-
bility of the dimer (V)

V
 TABLE I1 z
Vacuum Pyrolyses of 111-Biphenol Polymer m2
Evolved Evolved Evolved Mole ratios
Sample wt., benzene, g. trimer, g. hydrogen, ml.
Sample no. g. (mmole). (mmole) (mmole) (mmole) CsH,/sample H2/sample Hz/CeHe Trimer/sample
1 4.335 0.876 0.706 10.88
(5.24) (1 1.24) (1.19) (0.48) 2.15 0.091 0.043 0.23
2 0.713 0.116 0.050
(0.863) (1.49) (0.084) 1.72 0.10
3 0.531 0.118 0.043
(0.643) (1.51) (0.072) 2.33 0.11
8 Formula wt. of monomer = 826.
 TABLE I11
Temperature Scan Thermograms of Phosphonitrile Polymers
10% wt.
Polymer Sample loss, Total
No. Polymer strnctiire fraction an Atmosphere wt., mg. "C. wt. loss, mg. Remarks

980 N2 15.0 400 15.0 Model compound

(monomer)

2 1st 105,000 Wet air 17.3 390 17. 3

3 1st 10.5,ooo Nz 20.4 460 10.4

4 Crude Not. deter. NP 40.9 470 17.0

H

28.6 477 Not meas. a

Crude SO, 300 N, 21.4 478 Not meas. a

L Jx
 1679

a a a a

z
i

C
m
00

c * dl '9 b 31
m 5
m
a? 3 2 m

i
.I
d
+
c

2 z d z d d

4- c c
a
3 .li
3
3

pi
8
0

4-
p
3
d
1680 BILBO, DOUGLAS, FEN'EH. HERRING

'9 @? p:
t- P- s
0 hl m
u5
I0 C D
-8 M M

9 "! m
rn '9
M
2 m

.-
I-

+
d z s

42 *
n
r( 1

uP

2 2
 HEXAPHENYLDICHLOROPHOSPIIONITRILE 1681

( M , = 1600) whose 10% point was 450"C., considerably below the average
of the higher molecular weight samples. Thermogram 1 is a monomeric
model compound containing the P 4 - C bonds of the diol-linked poly-
mer,

The 10% point of this compound is lower (400°C.) than that of the dimer
V.
These results show that, for polymers of about 10 monomer units and
higher, decomposition is independent of chain length, suggesting random
cleavage rather than chain uruipping from a labile end group. The pres-
ence of [(CJI&PN]4, [(CaH&PN]a, and benzene in the volatile residues of
vacuum pyrolyses (Table 11) suggests that rupture of the phosphonitrile
tetramer ring may be the primary source of degradation.
Similar observations have been made by Korshak et al.,9who studied the
pyrolysis of octaphenyltetraphosphonitrile (VI)

in sealed ampules up to 500°C. Mixtures of pentamers, hexamers, and

higher polymers, plus benzene were obtained between 460 and 500°C.
At temperatures below 420"C., trbners were also observed.
The effect of environment on the 10% point may be seen by examination
of thermograms 9, 11, 12, 13 and 14. High molecular weight samples
of polymers were examined in dry and wet nitrogen, dry oxygen, and dry
and wet air. The 10% points in dry nitrogen and dry air were identical
(475°C.) and in dry oxygen slightly lower (457°C.). In wet air the value
decreased to 380°C. and in wet nitrogen a similar value was obtained
(400°C.).
These results show a trend which suggests that moisture and perhaps
oxygen lower the temperature at which decomposition begins. The low
value of the 10% point (310°C.) in thermogram 15 suggests that at atmo-
spheric pressure, volatile products form which are not swept away, but in
vacuum these materials sublime or are condensed in traps. This result
indicates that the onset of decompositionmay be nearer 300°C. than 475"C.,
as indicated in atmospheric pressure thermograms.
Under the much harsher conditions described in Table IV, where samples
are exposed for several hours at constant temperature, a more realistic pic-
ture of the thermal stability of the diol-linked polymers emerges. Even
under the inert conditions of dry nitrogen, decomposition occurs between
300 and 350°C. and in ambient air, evidence of decomposition
 E
G0
TABLE IV
Constant-Temperature Thermograms of Phosphonitrile Polymers C

Average "$
Total wt. loss Tem-
q
Sample time, Total wt. rate, perature,
No. Polymer structure Fraction a,, Atmosphere wt., mg. hr. loss, mg. mg./hr. OC. 3t!
1 F - 9 - q 1st 259,000 N2 43.0 18.5 0.0 0.0

2 259,000 N2 25.5 5.0 5.6 1.12 350

 HEXAPHENYLDICHLOROPHOSPHONITRILE 1683

8
M

0
1 0
'?
I4
8
0
c15

x c?
2
9
x
rl

2
9
Q)
M
c?
2 z
Lo

0
0
0.
Q,
co
10

4
P
rl
Ym
l-4 m
rl
1684 BILBO, DOUGLAS, FETTER, HERRING

0
//
is observed at 225°C. with the isolation of (C6H6)2P . It is apparent
\
OH
that long-term stability under ordinary atmospheric conditionsis consider-
ably lower (-250°C.) than short term stability under inert conditions
(temperature scan TGA in dry Nz).
An examination of thermograms 3, 4, and 17 in Table I11 shows the
stability of polymers with linking agents other than p,p'-biphenol. The
compounds of 3 and 4 have essentially the same 10% points as the diol-
linked polymer, but compound of thermogranl 17 has a much lower value
(362°C.). I n this polymer it may be the -C6H4-~--c&- group,
which the other polymers do not have, that degrades at a lower tem-
perature than the tetramer ring. This hypothesis is supported by ther-
mogram 18 (Table 111) which shows that the 10% point (355°C.) is only
slightly affected by moisture and oxygen, suggesting that the linking agent
degrades before the phosphonitrile ring, which is affected by moisture.
Although no work has been published on the kinetics and mechanism
of decomposition of the type of polymer described above, Gimblettlo has
reviewed work on polydichlorophosphonitriles and has proposed a first-
order mechanism for the formation of chain polymer from dichlorophos-
phonitrile timer.

The diradial undergoes a series of propagation, crosslinking, and termina-

tion reactions resulting in the formation of a rubbery polymer.
If the tetramer ring of the diol-linked polymers, described above, de-
graded in a similar manner by the elimination of N=P(C6H6), groups, a
kinetic treatment from weight loss measurements could be carried out and
our observation of [N=P(C6H6)~]3formation and those of Korshak et al.g
would be rationalized. However, our observation of the simultaneous
formation of benzene and diphenylphosphonitrile trimer suggests that the
degradation of I may be more complex, and weight loss measurements
(TGA) may not yield data interpretable by the first-order kinetics, un-
less N=P(CeH& is the initial degradation product.

References
1. S. M. Zhivukhin, V. B. ToLtoguzov, and V. V. Kireev, Plasticheskie Massy, 1963,
24; Chem. Abstr., 60,1848 (1964).
2. S. M. Zhivukhin, V. B. Tolstoguzov, and V. V. Kireev, Vysokomolekul. Soedin.,
6,1111 (1964).
3. S. M. Zhivukhin and V. V. Kireev, Zh. Obshch. Khim., 34, 3126 (1964).
 I~EXAPITENYJ~DICIILOROPI-IOSPIIONITRII~E 1685

4. C. A. W a r n , U. S. Pat. 2,866,773 (1958).

5. R. G. Rice, B. H. Geib, L. A. Kaplan, and J. R. Hooker, U. S. Pat. 3,108,989
(1963).
6. R. G. Rice, B. H. Geib, and L. A. Kaplan, U. S. Pat. 3,121,704 (1964).
7. R. Pornin, Bull. SOC.Chim.France, 1966,2861.
8. D. L. Herring and C. M. Douglas, Znorg. Chem., 4,1012 (1965).
9. V. V. Korshak, I. A. Gribova, T. V. Artamonova, A. N. Bushmarina, Vysoko-
nwlekul. Soedin., 2, 377 (1960).
10. F. G. R. Gimblett, Trans. J . Plastics Zmt., 28,65 (1960).
Received October 25, 1967