Title: Synthesis and some reactions of some diamino- and triaminophosphonium chlorides
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00097900/00001
 Material Information
Title: Synthesis and some reactions of some diamino- and triaminophosphonium chlorides
Physical Description: vii, 137 l. : illus. ; 28 cm.
Language: English
Creator: Frazier, Stephen E
Publication Date: 1965
Copyright Date: 1965
Subject: Phosphorus compounds   ( lcsh )
Nitrogen compounds   ( lcsh )
Polymers and polymerization   ( lcsh )
Chemistry thesis Ph. D
Dissertations, Academic -- Chemistry -- UF
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
Thesis: Thesis - University of Florida.
Bibliography: Bibliography: l. 131-136.
Additional Physical Form: Also available on World Wide Web
General Note: Manuscript copy.
General Note: Vita.
 Record Information
Bibliographic ID: UF00097900
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: alephbibnum - 000423878
oclc - 11022145
notis - ACH2283


This item has the following downloads:

PDF ( 4 MBs ) ( PDF )

Full Text







December, 1965


The author takes this opportunity to express his

sincere gratitude to his research director, Dr. Harry H.

Sisler. Although busy with administrative responsibilities,

Dr. Sisler has always provided enthusiastic encouragement

and offered valuable suggestions and advice concerning this

research. Dr. Sisler has also listened sympathetically to

the author's personal problems, and has been helpful in

obtaining the solutions to these problems.

The author wishes to recognize the interest given by

the other members of his committee and to thank them for

their assistance in preparing this dissertation.

A special note of appreciation should go to Dr. D. S.

Payne for the many valuable discussions and suggestions

shared during his visit at the University of Florida. The

author also wishes to acknowledge the help of his

colleagues, Dr. Donald F. Clemens and Mr. Robert L. McKenney.

A note of thanks goes to Mrs. Thyra Johnston who took a

personal interest in the typing of this dissertation.

Finally, the author wishes to thank W. R. Grace and

Company for the generous financial support of this work.




LIST OF TABLES. . . . . . .

LIST OF FIGURES . . . . . .


I. INTRODUCTION. . . . . .


Phosphonitrilic Derivatives.

The Chloramination Reaction.

* . .

* . .
. .

S 1

. 4

.. . . 4


Experimental . . . . . . .

Discussion . . . . . . . .

Summary . . . . . . . .

[(C2H5)2P(N2)-N=P(NH2)(C2H5)2C . .

Experimental . . . . . . .

Discussion . . . . . . .

Summary . . . . . . .. .


Experimental . . . . . .


. . . .


Discussion . . . . . . . 115

Summary. . . . .. . . . . 123


BIBLIOGRAPHY. .... . . . . . . . 151

BIOGRAPHICAL SKETCH . . . . . . . .. 137


Table Page

1. Average P-N Bond Energy in (PNC12)n . . 8

2. Base Strength of Phosphonitrilic Derivatives
Toward Protons in Nitrobenzene and Water. . 14

3. Variation of P-N Stretching Frequency with
Electronegativity of Exocyclic Groups . . 15

4. Phosphonitrilic Derivatives Synthesized from
[PNC123,4. . . . . . . . . . 18

5. Infrared Absorption Data (Cm- ) . . . 34
6. Infrared Absorption Data (Cm-1) . .. . . 70

7. Effect of Ammonium Chloride on Product
Distribution. . . . . . . . . 76

8. N. M. R. Spectral Data for
[(C2H5)22P )-N(NH)-N=P(N2)(C2H5)2 C1 . . 81
9. Infrared Spectral Data, Cm-1. .- 97
9. Infrared Spectral Data, Cm .. 97


Figure Page
1. Overlap of Npz Orbitals with Pdxz Orbital . 9

2. Pi-Electron Energy Levels of C6 6, (PNC12)3
and (PNC12)4 . . . . . . .. 10
3. Overlap of Pd -Orbitals with N -Orbitals . 11

4. Localized, Three-Center P-N-P Bonds in a
Phosphonitrile Segment. . . . . . 11
5. Minilab Reaction Flask. . . . . . 38
6. The Chloramine Generator. . . . . . 41
7. Infrared Spectrum of (C6H5)2P(O)NH2 (Nujol) 46
8. Infrared Spectrum of Intractable Chloramina-
tion Product of (C6H5)2P-P(C6H5)2 (Melt). .. 48
9. Infrared Spectrum of [(C6H5)2P(NH2)2]C
(Nujol) . . . . . . . . . 50
10. Infrared Spectrum of Intractable Chloramina-
tion Product of (C6H5P)4 (Thin Film). . . 53
11. Infrared Spectrum of [(C6H5P(NH2)3]C1 (Nujol) 55
12. Infrared Spectrum of C6H5P(O)(NH2)2 (Nujol) 56
13. Semimicro Sublimation Apparatus . . . 75
14. Infrared Spectrum of [(CH )2PN]n(KBr) . .. 77
15. H' N.M.R. Spectrum of
[(C2H5)2P(NH2)-N=P(N2)(C2H)2]Cl . . 80
16. Infrared Spectrum of [(n-C4H9)2PN]I (Melt). 85

Figure Page
17. Infrared Spectrum of CH3PC14 (Nujol) . . 102
18. Infrared Spectrum of (CH3)2PC13 (Nujol). . 103
19. Infrared Spectrum of (CH )2C14P3N3 (Nujol) . 105
20. Infrared Spectrum of [(C2H )2PN*C12PN]2 (KBr). 109
21. Infrared Spectrum of Suspected
[(CH )2PN.CH3C1PN]x (Thin Film). . . . 112




In recent years an area of study dealing with the

synthesis of phosphonitrilic derivatives and precursors

has been under investigation in our laboratories. Our

interest in this study was heightened by the discovery(1)

that diphenylchlorophosphine undergoes ammonolysis and

chloramination producing diphenylphosphonitrilic trimer

and a phosphonitrilic precursor,(2,'14

[(CH5)2P(NH2)-N=P(NH2)(C6H5)2]C1. On the basis of this
discovery a general reaction sequence was postulated in

which it was proposed that the chlorophosphine is

ammonolyzed with excess ammonia,

R2PC1 + 2 NH 3 R2PNH2 + NH4C1,

the aminophosphine reacts with chloramine to yield the

diaminophosphonium chloride,

R2PNH2 + NH2C1 [R2P(NH2)2]C1,

and the diaminophosphonium salt undergoes condensation

giving the observed products,


2[R2P(NH2)2]Cl [R2P(NH2)-N=P(NH2)R2]C1 + NH1Cl

[R2P(NH2)2]C1 + [R2P(NH2)-N=P(NH2)R2]C1 [R2PN]3

+ 2 NH4C .

Further evidence for this reaction sequence was

found in the study of the ammonolysis and chloramination

of dialkylchlorophosphines.) With dialkylchloro-

phosphines the chloramine-ammonia reaction produced di-

alkyldiaminophosphonium salts in which the alkyl group was

methyl, ethyl or n-butyl. In addition, the salt,

[(C2H )2P(NH2)-N=P(NH2)(C2H5)2]Cl, and the trimer,

[(n-C4H9)2PN](6) were produced. It was postulated that

these phosphonium salts should undergo self-condensation

reactions eliminating ammonium chloride and forming

phosphonitrilic derivatives. In addition, it was pre-

dicted that interesting phosphonitrilic derivatives should

result from the reactions of these salts with PCI RPC14,

or R2PC13 where R can be either an alkyl or an aryl group.

It was hoped that a study of these processes would not

only reveal a method by which cyclic derivatives could be

synthesized, but would also lead to high-molecular-weight

phosphonitrilic polymers.

Although it had been shown that diaminophosphonium

salts can be synthesized by chloramination and ammonolysis

of dialkylchlorophosphines, it was thought that these salts


might also result from the chloramination of substituted

diphosphines. An extension of this method to polyphosphines,
(RP)n, might also result in the synthesis of triaminophos-

phonium chlorides.

Therefore, the purpose of this study was (1) to

investigate the chloramination-ammonolysis of substituted

diphosphines and tetraphenylcyclotetraphosphine; (2) to

investigate the self-condensation reactions of diamino-

and triamino-phosphonium salts and

[(C2H5)2P(NH2)-N=P(NH2)(C2H5)2]C1; and (3) to investigate
the condensation reactions of some diaminophosphonium salts

with PC15, CH3PC14, and (CH )2PC13.



Phosphonitrilic Derivatives

The first phosphonitrilic derivatives to be synthe-

sized were the phosphonitrilic chlorides, [C12PN]n, formed

by the reaction of PC15 and ammonia. This reaction was

noted by Liebig in 18354.7 Although Liebig did not

correctly deduce the empirical formula of the product of

this reaction, it was not long before other workers(8)

confirmed the formula [C12PN]n. The most significant early

study in this field was carried out by Stokes who published

a series of papers(9) dealing with the reaction of PC15

with NH4C1. He confirmed the formation of P3N3C16 and

P4N4C18, obtained the first reliable physical data on these

compounds, and studied their ammonolytic and hydrolytic

reactions. Stokes perfected the method of forming these

derivatives and discovered higher members of the homologous

series, [C12PN]n, from n=3 to n=7. It was Stokes who

originated the name "phosphonitrilic" polymer and who first

postulated, for the lower members, cyclic structures with

alternating phosphorus and nitrogen atoms. He also isolated

a highly polymeric phosphonitrilic "rubber," a discovery


which has contributed to the renewed interest in phospho-

nitrilic derivatives in recent years.
Since the discovery of the phosphonitrilic

chlorides, a large number of derivatives have been synthe-

sized in which the substituents on the phosphorus atoms

vary widely. These substituents can be halogens, pseudo-

halogens, alkyl-, perfluoroalkyl-, aryl-, aryloxy-, and

alkoxy-groups, N-substituted primary and secondary amines,

OH and several others.
The phosphonitrilic structure, first suggested by

Stokes, has been investigated and confirmed by infrared

and ultraviolet spectroscopy, nuclear magnetic and

quadrupole resonance studies, and electron and X-ray dif-

fraction. References to these works are listed in the
various reviews available on this subject.(10,11,12) The

phosphonitrilic derivatives are now known to contain a

backbone of alternate phosphorus and nitrogen atoms in

which all of the P-N bonds are of equal length. The

trimer, the first member of the homologous series, has a
planar, six-membered cyclic structure

Cl /Cl Cl C

N. > .
Cl I | Cl ClJ |,C1
ClS P Z 1 C?1 N 1Cl
N *N

The tetramer and higher members of the (R2PN)n series

(R=C1,F) through n=17 are also cyclic.(15) They are not

planar but have puckered configurations which give improved

overlap of d-orbitals on phosphorus with the lone-pair

orbitals on nitrogen providing increased delocalization of

the lone-pairs and hence, greater stability. In a number

of derivatives for which measurements have been made, the
average length of the P-N bond is 1.56 A. This bond length

depends on the group attached to phosphorus and on the size

of the ring since it is affected by delocalization of the

unshared pair on nitrogen. The N-P-N bond angles are all

about 1200, however, the P-N-P angle varies from 1200 to

1500 depending on the ring size and the substituents on

phosphorus. The P31 magnetic resonance spectra of the

cyclic derivatives in which the phosphorus atoms have the

same substituents have a single peak indicating that the

equivalence of the phosphorus atoms is independent of ring


The bonding in phosphonitrilic derivatives is quite

interesting and a number of structural features can be

rationalized in terms of the bonding theories which have

been proposed. X-ray studies have shown that the phosphorus

atom is in an almost tetrahedral environment in the deriva-

tives studied. Thus its single bonds may be considered to

be formed using principally sp3 hybrid orbitals. In the

trimer the P-N-P angle is nearly 1200 and the nitrogen atom

is in a trigonal environment. The sigma-bonding electron

pairs and the lone-pair occupy approximate sp2 hybrid

orbitals. The fourth nitrogen electron pair occupies a 2-

orbital and can contribute to pi-bonding with phosphorus.

The pi-bonding in phosphonitrilics has excited a lively

controversy in the recent literature.(14,15) The struc-

tures shown above imply a resonance of the double bonds

and a bonding picture which is similar to that employed

for benzene. However, this analogy is complicated by the

necessity of using phosphorus d-orbitals in forming pi-

bonds. The d-orbitals are arranged in space in such a way

that considerable overlap with the nitrogen p-orbital, the

sp2-hybridized lone-pair orbital and with ligand orbitals

occurs. Several theoretical descriptions have been proposed

for this type of system.

Craig(16) has stated that because of the low local

symmetry of the phosphorus atom in phosphonitrilics, the d-

orbitals are completely non-degenerate. Four d-orbitals,

dy, dxz, dy, d 2 2, contribute to the overall pi-bonding
xy' xz yz x -y
(the d 2-orbital can be involved in exocyclic pi-bonding but

this is not considered to be of importance), and the contri-

butions need not be equal.

For pi-bonding, Craig considers the overlap of

phosphorus dxz-orbitals with nitrogen p -orbitals, shown in

Figure 1. The results of his molecular orbital calcula-

tions show that, unlike benzene, the highest energy bonding

orbital is non-degenerate, as shown in Figure 2. This

leads to an aromatic system of delocalized electrons which

is distinctly different from that encountered in carbon

compounds which utilize only p-orbital overlap. Any even

number of pi-electrons gives a closed shell. Huckel's

4n + 2 rule does not hold. This theory predicts that the

delocalization energy per electron increases steadily to a

limiting value as the number of pi electrons increases with

increasing ring size. Thus the trend in average P-N bond

energy (Table 1) is cited as a confirmation of this theory.

TABLE 1(16)


n in (PNC12)n 4 6 2

E(P-N)n-E(P-N)3 (Kcal) 0 0.59 0.54 0.60 0.62

Dewar,(15) on the other hand, takes advantage of

the favorable symmetry properties of the dy orbital on

phosphorus as well as the favorable overlap properties of

the dxz orbital. Thus he has hybridized these orbitals

to form two new pi-orbitals: da (dxz + dyz) and

d b -(dxz dy

Figure 5 shows a N-P-N segment of the phosphonitrile

structure and illustrates the approximate geometry of the



Fig. 1.-Overlap of NZ Orbitals with Pxz
pz dxz



1 _I _1_





Fig. 2.-Pi-ElectrP E6gy Levels of C6H6, (PNC12)3 and
(PNC12)4. 6,2


d /Ada

yz / 1
\ /


/ \

/ \

Fig. 3.-Overlap of Pd -Orbitals with N -Orbitals.

new pi-orbitals. These hybridized pi-orbitals overlap
nitrogen pz orbitals on both sides of the phosphorus atom
to form the pi-bonding system. Instead of an extensive
aromatic system, however, this method describes P-N-P units
connected by localized three-center bonds (Fig. 4). The

Fig. 4.-Localized, Three-Center P-N-P Bonds in a
Phosphonitrile Segment.

relatively simple molecular orbital calculations indicate

that the delocalization energy is on the order of 0.82 no

where B is the resonance integral and n is the number of

P-N units. This is considered a high value and therefore

the theory predicts high stability for the phosphonitrilic

structure. Unlike Craig's theory, this approach indicates

that the properties which depend on the pi-system alone

should not change with increasing ring size. Certain

spectral measurements are cited in support of this


Superimposed on the sigma- and pi-bonding in phospho-

nitriles is a third bonding system referred to as the pi'-

system. This system arises from the overlap of the d 2_y2

and d -orbitals of phosphorus with nitrogen lone-pair
orbitals. As mentioned above, the unshared pair of elec-

trons on nitrogen occupies an s-p hybridized orbital. In

the trimeric molecules where the P-N-P bond angle is about

1200 this orbital is an sp2 hybrid. However, in the

tetramer and larger derivatives the bond angle becomes

greater, the lone-pair orbital obtains more p-character,

and greater overlap with phosphorus d-orbitals can occur.

Electron density can therefore increase on phosphorus at

the expense of nitrogen in this second delocalized system.

In rationalizing the chemical and physical properties

of phosphonitrile derivatives it is difficult to separate

the several effects arising from interactions of the three

bonding systems and the inductive effects of the substitu-

ents on phosphorus. An example of this is presented by

Feakins et al.,(19) in a thorough study of the basicity of

phosphonitrilic derivatives. The observed trend in base

strength as the ligands on phosphorus were changed, or as

trimers and tetramers were compared, could not be explained

by either pi-bonding theory alone. The data could be

rationalized, however, by considering the various effects

mentioned above, and the conclusion was reached that the

base strengths depend in a complex way on a number of


It has been established that the base strength of(20)

phosphonitrilic derivatives is dependent upon the avail-

ability of nitrogen lone-pair electrons in the ring and

that protonation occurs on the ring. Thus, delocalization

of the nitrogen lone-pair into the region of the P-N bond

has the effect of decreasing the base strength and in-

creasing the P-N bond strength. Both of these effects are

sensitive to small changes in the ability of the substituent

on phosphorus to accept electrons and can therefore be

roughly correlated with the electronegativity of the sub-

stituent. Table 2 shows the increase in base strength of

certain derivatives with a decrease in the electronegativity

of the ligand.

TABLE 2(19)

R C1 OC2H5 C6H5 C2H5 NHC2H5 N(C2H5)2

pKa (NPR2)3 <-6.0 -0.2 1.5 6.4 8.2 8.5

pKa (NPR2)4 <-6.0 0.6 2.2 7.6 8.1 8.5

5.9 8.7
(water) (water)

6.5 8.7
(water) (water)

In the infrared spectra of phosphonitrilic deriva-

tives there occurs a strong band in the region of 1100-1400
cm1 (see Table 3) which is attributed to P-N ring stretch-

ing. The position of this peak is quite variable and the

variation can be correlated with the electronegativity of

the group attached to phosphorus. If the group has a high

electronegativity the nitrogen lone-pair electrons can be

delocalized into the region of the P-N bond, strengthening

that bond and increasing the frequency at which the infra-

red peak occurs. Conversely, if the group has a low

electronegativity or is an electron releasing group such as

an amine, the nitrogen lone-pair is localized on the nitro-

gen atom and the P-N stretching frequency occurs at lower

energy. This effect is illustrated in Table 3.



Group Electro- (a) P-N Stretching Frequency, Ref.
negativity Cm-1

Trimer, Tetramer,
(R2PN)3 (R2PN)4

F 4.0 1305 1435 23

Cl 3.0 1218 1315 24

C6H5 1.74 1190 1213 25

CH3 1.57 1180 1220 26

NH2 1175 1240 27

C2H5 1.2 1157 1280 25

n-C4H9 0.9 1155 6

(a)Electronegativities of F and Cl are Pauling's(21)
values; others are calculated from the
formula,(22) X. = XH + 0.183 (D(R-H)-[D(R-R)-D(H-H)]} .

Obviously the foregoing discussion is a simplifica-

tion. Such properties as the base strength and P-N stretch-

ing frequency also depend on a number of factors not

mentioned in connection with the bonding theory. In any

thorough analysis one must consider the steric requirements

of the ligands, hydrogen-bonding of the ligands, and ring

conformation. These factors affect the overlap of ligand

orbitals and ring-nitrogen orbitals with phosphorus d-

orbitals and thus may change the expected order of base

strengths or the expected P-N stretching frequency. In

addition, the electron releasing or accepting abilities of

the groups on phosphorus will not affect the pi'-electrons

alone but will also affect the sigma- and pi-electrons to

some extent.

This brief discussion indicates the advancements

which have been made since Stokes' pioneering work in

understanding the structure and bonding in phosphonitrilic

derivatives. Clearly, more experimental data are necessary

for a complete evaluation of the bonding theories.

Since the work of Stokes, a wide variety of phospho-

nitrile derivatives have been synthesized. A large number

of derivatives may be synthesized because of the wide

variety possible in the nature of the groups attached to

the phosphorus atom. In addition, the groups on phosphorus

in a given molecule can be different from each other. In

the cyclic trimer, for example, the groups on neighboring

phosphorus atoms can differ. The methods of synthesis of

phosphonitrilic derivatives can be considered in two general

classes: synthesis from non-phosphonitrilic starting

materials, and substitution of the desired group on a previ-

ously synthesized P-N backbone. An example of the first

method is the synthesis of the phosphonitrilic bromides from

PBr5 and NH4Br. This method is analogous to Stokes' synthesis

of the phosphonitrilic chlorides and gives the trimer and

tetramer in about 50 per cent yield.(28,29) If one uses

mixtures of PC5 and NH4Br or

NH4C1 + PC15 [NPC12]3,4,n + HC1

NH4Br + PBr5 [NPBr2]3,4 + HBr

PBr + Br2 + NH4Br [NPBr2]3,4 + HBr

PBr5 and NH4C1 the products are mixed chlorobromides such

as P3N3C1 Br, P3N3C1Br2 and P3N3C12Br4'(30)

Most syntheses of new phosphonitrilic derivatives

have involved the substitution of different groups on

previously synthesized phosphonitrilic backbones. The

usual starting material for such a synthesis is the

phosphonitrilic chloride trimer or tetramer. Thus, the

fluorides can be synthesized by reaction of the chlorine

derivatives with fluorinating agents. A number of examples

of substitution reactions of this type are presented in

Table 4. This method is suitable for the synthesis of a

large number of phosphonitrilic derivatives(48) although

it has not been found to be generally useful for the

synthesis of alkyl or perfluoroalkyl phosphonitriles. Only

one example of the substitution of such a group is known

in the chemistry of cyclic phosphonitriles,

P3N C13[N(CH3)2]3 + CH MgBr P3N3(CH3)3 N(CH 3)23



p C\j t J rC\ 0 rC\ K 0 0 -


1L F H

i- 0z
S i- K1 -

0 II p

E4 o P 4 0 r0
.^ 1 oouO 0

SC\j C\j 0 0 0 M
H rm P

c .1 r<1 . 11J K, 0N KN ,J
ri r-1 i M M
SE-i 0 Zr O O i-1 Z Z M Z

CO L. I I I i r C\ r\ L -i LJi L C\-
0a r-a K 0 -i -- P a a C r-a C
c 0 0 00 0Zp0

S r- 1m Fr- mO r

d -P Z Z Z rZ
oQ UU~U (t ^ o o + e i< o !z riw i
Wd- (D ^i cO CO (x h r c r o ^
^ W F U U U ii M ^ O O


0 Lt N

C- -
0- (N N j; ;

LC r
li X


c0 00
0 lA

Kn r\ tr\
p 01 P0
o o PA P PA

,o m oj apq
o bo r^ i. k
.< C' LCN rC 01'~

rC H (\i C0' tN p Q Cn
0M ) P LCN O

CO O\ 4- i
k PA PA PA [1' IpH ji M'


wI~~ u uu uio

The above discussion refers only to cyclic phospho-

nitriles and their reactions. A renewal of interest in

macromolecular phosphonitriles has occurred in recent years

because of the possibility of finding products for military

and space industry use.

As mentioned earlier, Stokes discovered the first

highly polymeric phosphonitrilic "rubber" in 1897. Stokes

found that by pyrolyzing the cyclic phosphonitrilic

chlorides at about 3000 C a rubbery material was formed

which could be formulated as [PNC12]n where n is a large

number. This material is thermally stable up to about

3500 C where it begins to depolymerize yielding cyclic
derivatives. In certain cases(13,50) polymeric species of

the type, (PNC12)n*PC15, or H(PNC12)nC1, were formed di-

rectly from the reaction of PC15 with NH4C1. It has been

postulated(51) that the first product of this reaction is

ammonium hexachlorophosphate

NH4C1 + PC1 NH4PC16

This salt could undergo a series of condensation reactions

eliminating HC1

NH PC16 H2N-PC1 + 2 HC1 ,

H2N-PC14 HN=PC13 + HC1 .

The new species, HN=PC13, has two possible courses available

for further reaction. It can undergo intermolecular


nHN=PCl H=t=NPC12--nCl + n-1 H1 .
3 2 n

Or it can react with more PC15 to form the polymer,

SC Cl Cl
Cl \ /
P-E-N P---n1 .
Cl' n

Polymers in both these series are oils with n usually having

values of from 10 to 20.

Strong evidence(52) provided by conductance and

capacitance dielectricc constant) measurements and by

electron paramagnetic resonance studies, indicates that

the thermal polymerization of the cyclic phosphonitrilic

chlorides takes place by ionization of a chlorine atom

followed by an electrophilic attack by the positive

phosphonium ion

Cl C1 Cl Cl

:N N: :N N: + 01 ,
Cl I|,cl Cl ICl

Cl >N C1 Cl N
** **

Cl Cl

:Nl N: + C ,C'
ci I l ci I ici >
Cl P N N:

Cl Cl


CI \C1 .1 C1 C1 [PNC12] 4
N N==P---N=P--N-"P- Cl- ,- >
Cl Cl (I

cll 11 Cl

C1"N"' P C11- x CI

The other known halogen or pseudohalogen substituted
phosphonitriles undergo a similar polymerization at high
temperatures. The cyclic fluorides can be polymerized to
rubbery materials by heating to 3500 C5) The trimeric
and tetrameric bromides give an elastomer when heated to
250-3000 C. 28'29) Likewise, the cyclic isothiocyanates
polymerize when heated to 1500 C.(36,37) The average
molecular weight of the polymeric chloride formed by thermal
polymerization has been estimated to be as high as 300,000.

Although these halogen and pseudohalogen substituted

phosphonitrilic high polymers are thermally stable, they

all share the same disadvantage of being susceptible to

hydrolysis. Thus, upon exposure to moisture, HX is evolved

with the subsequent loss of theologically

interesting properties.

When the high polymers are heated considerably above

the temperature of formation, a depolymerization reaction

occurs and the cyclic derivatives are regenerated. One

observes, therefore, the operation of a dynamic equilibrium,

n(Unit) --m (Unit)n x m

cyclic linear

in which the products, or the average molecular weight of

the system will depend on concentration, pressure and

temperature. Depolymerization is favored thermodynamically

at higher temperatures since, in the expression,


the enthalpy is constant because the number of bonds in

many small molecules is approximately the same as the number

of bonds in a few large molecules made of the same amount of

material, and the entropy of many small molecules is greater

than the entropy of a few large molecules of the same amount

of material.(54)

Thermal polymerization has not been observed with

the cyclic phosphonitriles which have groups other than

halogens or pseudohalogens on the phosphorus atoms. How-

ever, linear polymers with alkyl groups, aryl groups or N-

substituted amines on phosphorus have recently been

synthesized by several new procedures. It has been shown,

for example, that lithium azide and sodium azide react

with halophosphines to produce intermediates which undergo

thermal polymerization giving high-molecular-weight phospho-
nitriles.(5556) The following reactions are illustrative

PBr3 + NaN -3 [Br2PN]n + N2 + NaBr


(C6H5)PC12 + NaN3 170-175 [C6H5ClPN]n + N2 + NaCl

average molecular weight >5000

(CF )2PC1 + LiN3 (CF)2 PN3 50-60 [(CF)2PN]n + N
33 2 3 mm Hg )2 2
M.P. 90-940 C
Diphenylphosphinyl azide undergoes a similar reaction.(5)
(C6H )2P(0)N3 + (C6H5)2PCl (C6H5)2P---N=P(C6H5)2-- Cl
n = 3,4

Furthermore, recently published results(58) indicate that
the pyrolysis of hydrazinophosphines produces polymeric
phosphonitrilic species

n C6H5PNHN(CH3)2 2 160-2200 [C6H5P[N(CH)2]N]n

+ n H2NN(CH )2

n P[NHN(CH )2]3 180)> [P[N(CH )2][NHN(CH )2]N]n

+ n H2NN(CH )2

Halophosphines have been shown to undergo reaction
with tetrasulfur tetranitride, S4N4, giving phosphonitrilic
polymers and precursors:(59,606)

PC1l + S N PNC l PCl -->PNC12

C6H5PC12 + N (C6H PNC1) 4C6H PC14

(C6H5)2PC1 + S4N4 [(C6Hs)2PN]x'(C6H5)2PC1

The only alkyl-substituted high polymer reported in
the literature to date was synthesized by methylating poly-
meric phosphonitrilic chloride(62) by slurrying the chloride
with a solution of methylmagnesium bromide or methyl-
lithium. A complete characterization of the products of
this reaction has not yet appeared in the literature.
With the recent increase in interest in phospho-
nitrilic chemistry has come the development of several new
methods of synthesizing these derivatives from non-
phosphonitrilic starting materials. Some of these methods
involve the ammonolysis of chlorophosphoranes and are quite
similar to the original preparation using ammonium chloride
or ammonia and PC15

C6 PCI + :Cl [(C6H5):CI 13 + HC

(C6 ) 2PC13 + ::H3(:H4C1) [(C6H 5)2P],4 + IH4cl(HC1)4)

(C:) : + :HE4Cl [(CH )2T x + EC1
..... (26)
[(C3) ]x ( 2-5)3 (CH)2 PN]3,4

(C2H)2^! + 5 -* [(C2H )2 :]x + :NHC1
E (C2 220-2-DO H (25)
2[(C 52]x 2> [(C2 5 2 3,42

The phosohonitrilic precursor,
[(C6H-)2 3(:2)-:- P(:T2)(C6H )2]Cl, has recently been shown
to undergo rinc closure reactions(3'4) with a number of re-
-.r.ts producing a variety of cyclic derivatives

l(CE ) : -: 2 6 H]C1 + PC1 (CH ) Cl PcN
[(C6-5)2P(- 5)-: ':= (:2)(CET5)2]Cl + PCI5 (CH5)4CI2P3N3

+ (C6Hs)aCl 4'4N

[(C )2 2 ( E2) )2]Cl + CH C14 -

(CGH5)5C!P3 3 + (C6s)6Cl12P4NT

[(C )7 ( 2) '-P( 2)(C )2]C1 + (C6 5 2PC3 -

C(C6Il) -] 3,4

3ao, Dresdner and Youn(65) have reported a unique
ethod for the s -thesis of [? ] In this procedure
Sor .S' F- is passed over PE5 at 7100 C, and the cyclic
derivatives distill from the reaction zone.

Perfluoromethyl- and perfluoropropyl-derivatives
have been synthesized by the chlorination of bis(perfluoro-
methyl)aminophosphine and bis(perfluoropropyl)aminophos-
phine.(66) This reaction is carried out at -300 C and it
is postulated that the intermediate, [R2P(NH2)C1]C1, is
first formed. Trimethylamine is added to the reaction
mixture at the completion of the chlorination

R2PNH2 + Cl2-->[R2P(NH2)C1]C1

[R2P(NH2)C1]C1 + (CH3 )N--[R2PN]3,4,n + (CH3 )N-HC1

This type of reaction is apparently quite general be-
cause it has been demonstrated by Sisler and coworkers(1'67)
that diphenylchlorophosphine produces phosphonitrilic
derivatives and precursors when reacted with gaseous mix-
tures of anhydrous chloramine and ammonia,(68) with H2NNH Cl
and H3NNH C12, and with solutions of anhydrous, ammonia-free
chloramine. As pointed out on page 2, the reactive inter-
mediate postulated in the reaction sequence leading to
phosphonitriles was the phosphonium ion, [(C6H5)2P(NH2)2] ,
when excess ammonia was used, or, [(C6H5)2P(NH2)Cl]I when
ammonia-free solutions of chloramine were used. These ions
could undergo intermolecular condensation producing the pre-
cursors which were isolated


+ NH(,

2[(C6H5)2P(NH2)Cl]- [(C6H5) 2P(NH2)-N-P(Cl)(C6H5)2]
+ HC1
Indeed, the chloramination-ammonolysis of dialkylchloro-

phosphines(56) seems to follow a similar sequence and di-

alkyldiaminophosphonium chlorides were isolated from the

reaction mixtures.

The Chloramination Reaction

Chloramine, NH2C1, was first prepared and used as a

synthetic reagent by Raschig(69) in 1907. Raschig's method

consists of the addition of ammonia to dilute, aqueous

solutions of hypochlorite ion. Although this method has

the disadvantage that only dilute solutions of chloramine

can be obtained, and despite the difficulty in preparing

the completely anhydrous reagent, this process has been

widely used for many years for the commercial production

of hydrazine from chloramine and ammonia. In 1951, Mattair

and Sisler(70) found that anhydrous, gaseous chloramine

could be produced in excellent yields from the gas phase

reaction of ammonia and chlorine using an excess of NH3.

The first synthetic application of this method was the

production of anhydrous hydrazine

NH2Cl(g) + NH () NH2NH2() + NH4Cl(s)

Subsequently, chloramine was used to synthesize substituted

hydrazines from primary and secondary amines,(71) 1,1,1-
trisubstituted hydrazinium salts from tertiary amines,(72)
and aminophosphonium salts from tertiary phosphines.73)
These reactions may be generalized as follows
2 RNH2 + NH2Cl RNH-NH2 + [RNH 3C1 ,

2 R2NH + NH2C1 R2N-NH2 + [R2NH2]C1 ,

R3N: + NH2C1 [R3N-NH2]C ,

R3P: + NH2CI [R 3P-N H2]C .

Hart(74'75) examined thoroughly the chloramination
of aminophosphines and showed that chloramine attacks the
phosphorus atom instead of the nitrogen atom attached to
phosphorus, forming aminophosphonium salts instead of
hydrazinium salts
R2N-P(C6H5)2 + NH2C1 [R2N-P(C6H5)2NH2]1 ,

(R2N)2P(C6H5) + NH2C1 -C (R2N)2P(NH2)(C6H5)]CI ,

(R2N) P + NH2Cl [(R2N)3PNH2]C1 .

A similar study was undertaken by Clemens(76) using certain
aminophosphines and bis(phosphino)amines. The results were
analogous to those of Hart. With the bis(phosphino)amines
the reaction was postulated to follow the general course,

/ 2 / 2
RN + NH2Cl RN Cl

\ 2 2
RN Cl + NH3 RNT + NH4Cl

1 II
/ 2 / 2
RN + NH2Cl > RN Cl .

The chloramination-ammonolysis of monochlorophos-

phines has been mentioned previously (pages 2, 28).

Since halophosphines can be synthesized by cleavage
of diphosphines and polyphosphines with halogens,(7,8)

it seems reasonable to expect chloramine to cleave the P-P

bond in a similar manner. However, alkylation of diphos-

phines with alkyl halides, which is formally analogous to

the chloramine reaction,() produces different results

with different diphosphines. For example, tetramethyl-,

tetraethyl-, and tetra-n-butyldiphosphine react with methyl

and ethyl iodide to form diphosphonium salts

R2P-PR2 + R'I [R2P-PR2R']I.77'80)

R = CH3, C2H5, n-C4H9

R' = CH3,C2H5

On the other hand, tetra-cyclohexyldiphosphine is cleaved
by methyl iodide as is tetraphenyldiphosphine
(C6H11)2P-P(C6H11)2 + CH3I -
(C6H11)2PCH3 + (C6H11)2PI
[(C6H112P(CH )2]I [(C6H11)2P(CH3)I]I(77

(C6H5)2P-P(C6H5)2 + 3 CH3I [(C6H5)2P(CH3)2]I

+ [(C6H5)2P(CH3)I] .(81,82)

Burg(8) compares the R2P-group with a halogen or
pseudohalogen and states that diphosphines undergo many
reactions in a manner which is characteristic of halogens.
If one applies this idea to the chloramine reaction, one
would predict that phosphorus-phosphorus bonds would
undergo cleavage with chloramine.




Manipulation of materials.-The diphosphines used in

this study are extremely sensitive to oxygen, and detailed

precautions were taken to prevent contamination by oxygen

from the air. Similarly, some of the products formed in

this study are sensitive to moisture, and precautions were

taken to avoid contamination of solvents and other reagents

by moisture.

Benzene was either obtained as the reagent grade

product and stored over calcium hydride or obtained as the

technical grade product and distilled and stored over

calcium hydride. Petroleum ether and diethyl ether were

obtained as reagent grade products and were stored over

calcium hydride. Other solvents used were reagent grade.

Tri-n-butylphosphine was obtained from Food Machinery

and Chemical Corporation and used as obtained. Phenyl-

dichlorophosphine and diphenylchlorophosphine were obtained

from Victor Chemical Works and used as obtained. These

reagents were transferred by pipette under a stream of dry

nitrogen. Tetramethyldiphosphine and tetraethyldiphosphine

are liquids at room temperature and were purified by distil-

lation at reduced pressure. They were stored and transferred

by pipette under nitrogen in a D. L. Herring Model HE-43

Dri Lab equipped with a Model HE-93 Dri Train. Tetraphenyl-

diphosphine and tetraphenylcyclotetraphosphine are solids

and were stored and transferred under nitrogen in the dry


Infrared spectra.-Infrared spectra of the compounds

produced in this study were determined using a Perkin-Elmer

Model 337 grating infrared spectrometer. A summary of the

spectral bands of these compounds between 2.5 and 25

microns is presented in Table 5.

Solid samples were examined between KBr disks as

Nujol mulls, or, when the solid melted below 100C, as a

melt. In certain instances thin films could be obtained by

carefully evaporating a chloroform solution of the sample

on a KBr disk. Samples of substances which are sensitive

to moist air were prepared for infrared analysis in the dry


Elemental analyses.-Elemental analyses and molecular

weight determinations were performed by Galbraith Labora-

tories, Inc., Knoxville, Tennessee. Several nitrogen

analyses were carried out in these laboratories using a

Coleman Model 29 Nitrogen Analyzer.

TABLE 5(a)
[(CH3)2P(NH2)2]C1, Nujol Mul1(6)
3920(w), 3850(w), 3260(vs,b), 2940(vs), 2850(vs),
2620(m), 2560(w,sh), 2530(w,sh), 2400(m), 2260(vw),
2200(w), 2160(w,b), 2100(w,sh), 2060(w), 2000(w),
1940(m), 1625(w), 1570(s), 1460(s) 1420(s), 1410(m,sh),
1375(m), 1310(m), 1300(s), 1070(sh5, 1040(s), 990(s),
951(s), 885(m), 860(m), 835(m), 762(s), 695(s),
600(b), 527(b).
[(C2H5)2P(NH2)2]C1, Nujol Hull(6)
3940(w), 3160(vs), 3060(vs), 2940(vs,b), 2850(vs),
2640(w), 2550(w), 2330(w,sh), 2150(w), 2060(w),
1960(w,sh), 1920(w), 1870(w,sh), 1755(w,b), 1665(w,b),
1560(s), 1460(vs), 1395(s), 1380(s), 1275(s),
1240(m), 1160(w,b), 1075(s,sh), 1050(s), 1019(s),
980(s,sh), 968(s), 910(s), 761(s), 733(s), 721(s),
665(s), 625(s,b), 500(w,b), 447(m).
(C6H5)2P(0)NH2, Nujol Mull
3240(s,b), 3125(s), 2940(vs), 2860(vs), 1970(w),
1900(w), 1820(w), 1779(w), 1680(w), 1595(m),
1560(s), 1485(m), 1460(vs,b), 1440(vs,sh), 1439(vs),
1420(w,sh), 1380(s), 1365(m,sh), 1340(w), 1310(m),
1275(w), 1260(w), 1175(s), 1110(s), 1105(s),
1060(m), 1010(w), 990(m), 910(s), 858(w), 849(w),
752(s), 720(s), 695(s), 620(w), 532(s), 518(s),
488(m), 439(w).
[(C6H5)2P(NH2)2]C1, Nujol Mull
3170(s), 3075(s), 2980(vs,sh), 2940(vs), 2860(vs),
2710(w), 2540(w b), 2440(w), 1970(w), 1900(w),
1815(w), 1780(w), 1665(w,b), 1595(m), 1560(s),
1485(m), 1460(s), 1440(s), 1410(m), 1380(s),
1365(m,sh), 1340(w), 1320(w), 1280(w), 1190(w),
1160(w), 1110(s), 1075(w,sh), 1025(s,sh) 1010(s),
990(s,sh), 980(s), 922(m), 910(m), 855(w), 840(w),
756(s,sh), 751(s), 742(s), 720(s), 691(s), 638(w,b),
620(w), 549(w,b), 510(s), 505(s,sh), 472(m), 434(w),

Table 5 (cont'd)
6 / 2 C6 / 2
Suspected C- [ ---NP=N ---P NH, Melt

3400(s,sh), 3210(vs), 3060(vs), 2950(s), 2940(sh),
2635(w), 2600(w), 2300(m), 2250(w), 1960(m),
1900(m), 1820(m), 1770(w), 1720(w), 1670(m),
1640(m), 1590(s), 1550(vs), 1480(s), 1440(vs),
1280(s,sh), 1205(vs,vb), 1115(s), 1015(m), 910(m,b),
799(m), 742(s), 692(s), 685(sh), 615(w), 502(s,vb).
[(C6E5)P(NH 2)]Cl, Nujol Mull
3375(m) 3250(s,sh), 3170(s), 3075(s), 2940(s,sh),
2900(vs), 2850(s), 1590(w), 1559(m), 1460(s),
1440(s), 1415(m), 1380(s), 1135(s), 1084(w),
1070(w), 1025(w), 985(m), 925(-,sh), 875(m),
798(w), 763(s), 721(w), 705(w), 692(m), 620(w,b),
510(m), 474(m).
(C6H5P(0)(NH2)2, Nujol Mull
3350(s), 3275(s,sh), 5220(s), 3110(s,sh), 3055(m),
2940(vs), 2850(vs), 2740(w), 1955(w), 1900(w),
1770(w), 1590(m), 1560(m), 1460(s), 1440(s), 1410(m,b),
1580(s), 1330(w), 1310(w), 1182(m), 1155(vs),
1118(s), 1060(m), 1010(m), 959(s), 915(m,sh),
885(m), 855(w), 745(s,sh), 740(s), 720(m), 694(s),
620(w), 568(w,b), 520(s), 512(s), 494(w), 435(w,b).

(a)s,strong; m,medium; w,weak; b,broad; v,very;


oelting points were determined in a Thomas-Hoover

capillary mel` ": point a ratus and are uncorrected.

e Iaraion of sstituted diDhosohines and tetra-

. .he c. o tr .
tetraethydi oni wne were prepared by the desulfuration of

the corresoo_. 1- dipho ."'i disulfide with tri-n-butyl-

phosohine. The dipho -:..e disulf1ies were synthesized

by the Gri -.. -d method from PC1 and <_,-3r(R = CH C2H ).
This ethod has ben zhoroi. :ly discussed elsewhere. '

To the solid Adihosphine disulfide under a nitrogen atmos-

phere as x ed by pipette, tri-n-butylphosphine according

to the stoichio etry of the following general equation

)- )p + 2(C9) 3 R P-PR2 + 2(CH9)3P(S)

It was convenennt in his study to use from ten to fifteen

S- of the diphosphine sulfide in each run and to mix the

re ts in a 1l round-botto:ed distillation flask. The

mixture was then heated under reduced pressure. The tetra-

1 idi:ho>hine was o gained by fractional distillation

fro: he reaction mixture. The observed boilir.- point of

tet.a,.t, diphos3hine was 350 C at 16 mm ,-(Lit. :)

3o- = -0 C). -:e observed bill" ooint of tetra-

Shyldciphsphine was 67 C at 1.7 mm : -(Lit. : Bp74
S-o ). Yields of tetraaikyldiphosphine from this re-

action :ed between 3 and 80 per cent of theory based on

te amount of diphosphine disulfide used. Distillation was

slow because of the difficulty in removing the diphosphines

from (C4H9)3PS in which they are soluble.(84) At the

completion of the distillation the diphosphines were placed

in stoppered receivers under nitrogen and stored in the dry

box until needed. The infrared spectra and nuclear magnetic

resonance spectra of these tetraalkyldiphosphines have been

treated thoroughly elsewhere(84) and need not be considered


Tetraphenyldiphosphine was prepared by the reaction

of tri-n-butylphosphine with diphenylchlorophosphine(87)

which can be represented by the following equation

(C4H9)jP + 2(C6H5)2PCl (C6H5)2P P(C6H5)2

+ (C4H9)3PC12

In a typical experiment, 3.15 ml (14 mmoles) of tributyl-

phosphine was pipetted into 5.0 ml (27 mmoles) of diphenyl-

chlorophosphine in a Minilab reaction flask (Fig. 5) fitted

with a stirrer and a nitrogen inlet and outlet. The

mixture was stirred until it became homogenous. It was then

allowed to stand under a nitrogen atmosphere for about three

days. At the end of this time the reaction mixture had

become a solid, crystalline mass. This crystalline mixture

was extracted under a nitrogen flow with cool, freshly

boiled, distilled water until the washings gave no chloride

tes7 with aqueous silver nitrate. The white, crystalline




Fig. 5.-Minilab Reaction Flask.


residue was dried in vacuum and transferred to the dry box.

. P. 118-121 C (Lit.(78): M. P. 120.5 C). The average

yield of tetraphenyldiphosphine produced by this method of

synthesis was about 40 per cent of theory based on the

amount of diphenylchlorophosphine used. Although strict

precautions against oxidation of the diphosphine were

taken, a small amount of tetraphenyldiphosphine monoxide(88)

was probably present in each sample.

Tetraphenylcyclotetraphosphine was prepared by the

reaction of tri-n-butylphosphine with phenyldichloro-

phosphine 87) as indicated by the following equation

4(C4H9)3P + 4 C6H5PC12 (C6H5P)4 + 4(C4H9)3PC1

In a typical experiment, 5.0 ml (36.6 mmoles) of phenyl-

dichlorophosphine was pipetted into 8.24 ml of tri-n-butyl-

phosphine in a Minilab reaction flask (Fig. 5) under

nitrogen. Heat was evolved and after a few moments of

stirring the reaction mixture became a solid, crystalline

mass. When the mixture had cooled to room temperature, it

was extracted with copious amounts of absolute ethanol.

This extraction was continued until the acidified washings

gave no precipitate with aqueous silver nitrate. The white,

crystalline residue was dried in vacuum and transferred to

the dry box. M. P. 153-1550 C (Lit.(89): M. P. 155-1550 C).

The average yield of tetraphenylcyclotetraphosphine produced

- t his ethod of synthesis as 50 per cent of theory based

on amoun of ohenyldihlorophosphine used. Strict pre-

cautins .were ta en to avoid oxidation of the product.

7-e o an-d
e nrae s ra of tezraphenyldiphosphine and

tc-a~pn lc cloterao osine are available in the

_I~ ... . :;oer, a controversy exists in the cur-

ret te oce ni she nature of tetrapheny1cyclo-

to Cno' e e material produced in this study

s n:rj ana infrared spectrum reported for

"form A" ch is considered to be one stereoisomeric form

o. ( 7

? r- "".r... e-on 0o .lorai :e. -e preparation of

seous chlor -ine was carried out in a generator shown

cia a...: tical in y iure 6.( n In'. 'eous an onia,

clorine and nitro n --ere eeterea in rotareters and mixed

in s reaction tube. The roximate rate of intro-

duction hf ses was: 12, .1 mole/hr.; : 1.2

olh. oie, .. he approximate production rate

of c lora e as 0..1 mole/hr. Since anmonia was present

S 1, exc-ess of at -t -uired by the stoichiometry of

-I ecrc- 12L

+ 1 ( (s) 2'

consisted of c'bor -ine, anmonia and

r:;s ol iL were -laced in the reaction tube

0 *2 <

to filter out the finely divided ammonium chloride produced

in the reaction. The effluent gases were bubbled into a

receiver containing the solution of the substance to be


The Reaction of the Chloramine-Ammonia Mixture
with Tetramethyldiphosphine

(CH3)2P-P(CH3)2 + 3 NH2C1 + 2 NH 2[(CH3)2P(NH2)2]C1

+ NH4C1

In a typical experiment, 1.0 ml (7.1 mmoles) of

tetramethyldiphosphine was dissolved in 75 ml of dry benzene

and exposed to the effluent gases of the chloramine genera-

tor. A reaction took place immediately. A white precipitate

formed and the reaction mixture grew warm. After 10-15

minutes of chloramination the reaction appeared to be

complete and the temperature of the mixture dropped to about

200 C because of evaporation of the solvent. Chloramination

was continued for 5-10 minutes beyond this point to insure

exposure of the reaction mixture to an excess of chloramine.

At the completion of the chloramination the benzene was re-

moved by filtration. The solids remaining on the filter

were extracted with 60 ml of hot, 50/50 ethanol-acetone

mixture. White crystals precipitated from the solution as

it slowly cooled to room temperature. These crystals were

removed by filtration and dried in vacuum. M.P. 191-1940 C.

Anal. Found: N, 21.47. Calculated for [(CH3)2P(NH2)2]C1:

N, 21.79. The infrared spectrum of this material was
identical with that of dimethyldiaminophosphonium chloride

(M.P. 192-1940 C) produced by chloramination-ammonolysis

of dimethylchlorophosphine.(5'6)

Addition of petroleum ether to the mother liquor

caused a larger portion of dimethyldiaminophosphonium

chloride to precipitate. Yield: 1.08 g (63% of theory).

The Reaction of the Chloramine-Ammonia Mixture
with Tetraethyldiphosphine

(C2H5)2P-P(C2H5)2 + 5 NH2C1 + 2 NH3 2[(C2H5)2P(NH2)2]C1
+ NH4C1

3(C2H5)2P-P(C2H5)2 + 9 NH2C1 + 6 NH3 2[(C2H5)2PN3
+ 9 NH4C1

Tetraethyldiphosphine (5 mmoles) was dissolved in

75 ml of dry benzene and exposed to the effluent gases of
the chloramine generator. A white precipitate formed

immediately and the solution became warm. After 10-15

minutes the temperature of the solution fell to about 200 C

because of evaporation of the solvent, and the reaction

appeared to be complete. Chloramination was continued,
however, for 5-10 minutes longer until a relatively large

excess of chloramine and ammonia had been added to the re-

action mixture. The benzene solution was removed from the

precipitate by filtration and evaporated to dryness. A trace

of oily crystals was observed. The crystalline material was

separated from the oil by dissolving it in 30-600 petroleum

ether. The oil was discarded. Evaporation of the petroleum

ether yielded crystals which melted at 113-1150 C and which

sublimed readily at 30-400 C and 0.1 mm Hg. This material

was identified by its infrared spectrum and melting point

(see Chapter IV for a more complete description of the

identification of this material) as diethylphosphonitrilic

trimer(25) (Lit: M.P. 117-1190 C). Yield: trace (1-5


The solids remaining on the filter were extracted

with three, 10 ml portions of absolute ethanol. This solu-

tion was evaporated to dryness and the resulting crystalline

extract was dissolved in 10-15 ml of hot, 50/50 ethanol-

acetone mixture. This solution was cooled with an ice bath

and to it was added approximately 100 ml of 60-1100

petroleum ether. The product crystallized from the solution

as well formed, white needles. These crystals were filtered

and dried in vacuum. M. P. 103-1050 C. Anal. Found: N,

17.9. Calculated for [(C2H5)2P(NH2)2]C1: N, 17.89. The

infrared spectrum of this material was identical with that

of diethyldiaminophosphonium chloride (M. P. 106-108.50 C)

produced by the chloramination-ammonolysis of diethylchloro-

phosphine.(5'6) Yield: 1.07 b (68% of theory).

The Reaction of the Chloramine-Ammonia Mixture
with Tetraphenyldiohosohine

In a typical experiment, 1.05 g (2.8 mmoles) of

tetraphenyldiphosphine was dissolved in 50-75 ml of dry

benzene and exposed to the effluent gases of the chloramine

generator. A white precipitate formed immediately and the

reaction mixture grew warm. Chloramination was continued

for about 20 minutes to insure exposure of the reactants

to an excess of chloramine. At the completion of the re-

action the temperature of the reaction mixture had fallen

to about 20 C because of evaporation of the solvent. The

benze-l, solution was removed from the white precipitate by

filtration and the solids were washed with 25 ml of fresh

benzene. The combined filtrate and washings were evapo-

rated to dryness yielding 0.39 g of a white powder which

melted in the range of 80-1100 C. This solid was redis-

solved in 10 ml of boiling benzene. Upon cooling, white

crystals precipitated from the benzene solution and were

filtered and dried in vacuum. M. P. 158-1600 C. Anal.

Found: N, 6.6. Calculated for (C6H5)2P(0)NH2: N, 6.48.

The infrared spectrum (Fig. 7) is consistent for the formu-

lation, (C6H5)2P(0)N2.(91) Evaporation of the benzene

solution to dryness yielded an oily foam. More crystalline

(C6H5)2P(0)NH 2 could be recovered from this foam by dis-

solving it in 2-4 ml of boiling benzene, adding hexane to




i TS~ c\



re I ;"
ri r: ~C

F-e L-i
i n
i a

e: o

Cj i;i
Z: :1







the hot solution until it became cloudy, and allowing this

mixture to stand for several hours. By repeating this

process several times, all but a trace of the sample was

recrystallized and shown by its melting point and infrared

spectrum to be (C6H )2P(0)NH2. The noncrystalline residue

was a yellow oil and was discarded.

In other experiments, evaporation of the benzene

solution yielded a dark foam which appeared to be air

sensitive, turning wet upon exposure. After recovery of

the first crop of crystalline (C6H5)2P(0)NH2 from this

mixture, further attempts at recrystallization resulted in

the formation of oily products. The dark foam produced by

evaporation of the benzene solution melted at 35-400 C.

Its infrared spectrum is shown in Figure 8.. Anal. Found:

C, 65.54; H, 6.03; P, 12.67; N, 6.06; Mol. Wt. (cryoscopic

in benzene), 450. The P:N ratio is 1:1. This implies the
C6H C,6H5
structural unit, -E-P=N-}- for which the calculated

composition is: C, 72.4; H, 5.1; P, 15.6; N, 7.0. The

analysis, however, is low and indicates that the material

may consist also of oxygen-containing material.

The benzene-insoluble solids remaining on the filter

were extracted with two, 10 ml portions of hot chloroform.

The extracts were combined and concentrated to about half

the original volume by evaporation. White crystals pre-

cipitated and were removed by filtration and dried in vacuum.

















. -

M.P. 230-2320 C. Anal. Found: N, 8.9. Calculated for

[(C6H )2PNH2P(NH)-N(NH2)(C6H5)2]C: N, 9.30. The infrared
spectrum of this material was identical with that of

[(C6H5)2P(NH2)-N=P(NH2)(C6H5)2]Cl produced by the
chloramination-ammonolysis of diphenylchlorophosphine.(1)
Yield: 0.42 g (33% of theory based on tetraphenyldiphosphine
used). If the vacuum drying process was not prolonged, the
1:1 chloroform adduct(2) of the substance was obtained.
Anal. Found: N, 7.3. Calculated for

[(C6H )2P(NH2)-N=P(NH2)(C6H )2]Cl CHC13: N, 7.36.
The chloroform-insoluble residue was extracted with

50 ml of hot, 50/50 acetone-ethanol mixture. Addition of
diethyl ether to this solution caused a small amount of
NH4C1 to precipitate. The solution was filtered. Upon
addition of a copious amount of diethyl ether more crystal-
line material precipitated. This material was filtered and
dried in vacuum. M.P. 204-206 C (dec.). Anal. Found:

C, 56.22; H, 5.84; P, 11.78; N, 10.69; Cl, 14.50. Calcu-
lated for [(C6H5)2P(NH2)2]Cl: C, 57.04; H, 5.58; P, 12.26;
N, 11.09; Cl, 14.03. This analysis indicates that the
material may be slightly contaminated with ammonium chloride.
The infrared spectrum of this material (Fig. 9) is con-
sistent for the formulation [(C6H5)2P(NH2)2]C1, and shows
that the amount of ammonium chloride present is too small
to be detected by this infrared technique. Attempts to



0<\ *

0 r
H4 H P-



synthesize the hexafluorophosphate and tetraphenylborate

derivatives of the diphenyldiaminophosphonium ion by meta-

thesis in water failed. Yield of diphenyldiaminophosphonium

chloride: 0.20 g (14% of theory based on tetraphenyldiphos-

phine used).

The remaining insoluble material was shown to be

ammonium chloride.

The 7 action of the Chloramine-Ammonia Mixture
with Tetraphenylcyclotetraphosphine

In a typical experiment, 0.98 g (2.27 mmoles) of

tetraphenylcyclotetraphosphine was dissolved in 50-75 ml

of dry benzene and exposed to the effluent gases of the

chloramine generator. A white precipitate formed immedi-

ately and the solution grew warm. The reaction was

apparently complete when the temperature of the mixture

dropped to about 200 C because of evaporation of the

solvent. Chloramination was continued for 5-10 minutes

longer to insure exposure of the reactants to an excess of

chloramine and ammonia. The benzene solution was removed

from the precipitate by filtration and the residue was

washed with 25 ml of dry benzene. The filtrate and wash-

ings were combined and evaporated to dryness yielding 0.40 g

of a white, air-sensitive solid. This solid melted to an

opaque liquid at 68-720 C and the melt did not become clear

when heated to above 2000 C. The analytical data for

several samples of this material are summarized in the

table below.

Found C H P N C1 Mol. Wt.

Sample 1 51.85 5.95 17.55 15.40 2.5 940 (in benzene)

Sample 2 20.51 17.15
20.41 17.18(Kjeldahl)

Sample 3 49.74 5.21 21.45 19.20 4.08 870 (in benzene)

The P:N ratio in these samples is 1:2. This implies the

structural unit C6H5 /NH2 and a likely structure

might be

might be

6 5\ 2 6 5\ / 2
Cl [ P-N ]- PP=-NH

where n 5 5 .

The calculated elemental analysis for such a formulation is:

C, 49.94; H, 4.97; P, 21.50; N, 19.42; C1, 4.16; Mol. Wt.

865. The infrared spectrum (Fig. 10) is consistent for

such a structure. Pyrolysis of the crude material at 2000 C

and 0.10 mm Hg produced ammonium chloride as a sublimate and

an intractable residue which had a nitrogen content lower

than the original material.

The benzene-insoluble precipitate was extracted with

50 ml of 50/50 ethanol-acetone mixture. The solution was

removed from the insoluble ammonium chloride by filtration

and a product was precipitated by the addition of copious

amounts of diethyl ether. Upon filtration and drying in

vacuum the crude material was found to melt with decompo-

sition in the range 160-170 C. After recrystallization



PC\ 0




0 o
(^- P

o 0 -

HI 0
o ed
a) r-H .-

,- -t --

a o





rc1 *
K1- *
.-^-^^ .--- oj ?-
~~~ -- ---- __________-- ^ 4 ->
-- == r-\
/-- -- M r
-- ~~~c --- ~-c*

^- --- ---- ----

~~^ 0
c^,C 1-1
^"*^- r
_^- ro>

from 50/50 acetone-ethanol, the observed melting point was

164-1650 C. Anal. Found: C, 38.24; H, 6.15; P, 16.37; N,

21.96; Cl, 17.35. Calculated for [(C6H5)P(NH2)3]Cl: C,

37.61; H, 5.79; P, 16.17; N, 21.93; Cl, 18.50. The infrared
spectrum (Fig. 11) is consistent for the formulation

[(C6H5)P(NH2)3]Cl. Attempts to synthesize the hexafluoro-
phosphate and tetraphenylborate derivatives of the phenyl-

triaminophosphonium ion by metathesis in water failed.

Yield: 0.24 g (14% of theory).

The phenyltriaminophosphonium chloride so prepared

was quite sensitive to hydrolysis. Slow recrystallization

of the product from alcohol exposed to the atmosphere pro-

duced the hydrolysis product, C6H5P(O)(NH2)2(92) M.P.

180-1810 C (dec.) (Lit.(91): M.P. 1890 C). Anal. Found:

C, 45.47; H, 5.81; P, 19.86; N, 17.95; 0 (by difference),

10.43. Calculated for C6H5P(O)(NH2)2: C, 46.16; H, 5.81;

P, 19.84; N, 17.94; 0, 10.25. The infrared spectrum of

phenylphosphinic diamide is shown in Figure 12.


The results of this study show that under the condi-

tions of our experiment chloramine cleaves the phosphorus-

phosphorus bond in tetramethyldiphosphine, tetraethyldi-

phosphine, tetraphenyldiphosphine and tetraphenylcyclotetra-



I o





co o

I 0



f K\ i-i
\^ CaJ
_^_^---- 't^
C.r 'K'


_- -
~~ *~ ~- r






-" I
-I *
-4 0


N >














phosphine. The chloramination-armonolysis of tetramethyl-

diphosphine yields dimethyldiaminophosphonium chloride as

the only important product. Thus, this reaction appears to

be quite analogous to the chloramination of dimethylchloro-


Similarly, the chloramination-ammonolysis of tetra-

ethyldiphosphine gives diethyldiaminophosphonium chloride

in yields which average between 60 and 70 per cent of theory.

Unlike in the chloramination of diethylchlorophosphine,

however, the intermediate, [(C2H5 2P(NH2)-N=P(NH2)(C2H5)2]Cl,

is not produced. On the other hand, a trace of diethyl-

phosphonitrilic triner is observed. It is curious that

only a trace of diethyldiaminophosphonium chloride undergoes

complete condensation,

[(C02 P(2 )]Ci [(C2H5)2PN]3 + 3 NH4C,

while no intermediate product of condensation,

[(C2H5)2P('2)-N=P(NH2)(C2H5)2]C1, is observed. One might

speculate that at some time during this exothermic reaction

the complete condensation could momentarily become

kinetically favored.

The chloramination-ammonolysis of tetraphenyldiphos-

phine is .lot so straightforward and simple as the chlorami-

nation-ammonolysis of tetraalkyldiphosphines. However, it

is clear that cleavage of the phosphorus-phosphorus bond

occurs. The principal product which could be isolated and

identified was the substance,

[(C6H5)2P(NH2)-Y=-P(IH2)(C6H5)2]C1. This was also an impor-
tant product of the chloramination-ammonolysis of diphenyl-
chlorophosphine.(1) A more interesting result is the

identification of [(C6H5)2?P(H2)2]C1 as a product of the

reaction. Diphenyldiaminophosphonium chloride or the di-
phenyldiaminophosphonium ion is the postulated reactive

intermediate(1) in the formation of

[(C6H5)2P(NIH2)-N=P(HH2)(C6H5)2]C1 and [(C6g5)2PN] from

(C6H5)2PC1, NH2C1 and lH3. Unfortunately, the yield of
this material was small and it could not be purified from

traces of ammonium chloride. The substance is soluble in
polar solvents such as alcohol and water but is insoluble

in chloroform and non-polar solvents such as benzene and
n-hexane. Dissolved in water, it gives an immediate test

for chloride ion with Aga. However, it must hydrolyze
rapidly in water for no metathetic derivative of PF6 6 or

(C6H5)4B3 6 could be detected.
A third product which was isolated and identified
was the amide of diphenylphosphinic acid, (C6H5)2P(0)NH2.
This material was characterized by its melting point,

nitrogen analysis and infrared spectrum. It is quite
possible that during its preparation, the starting material,
tetraphenyldiphosphine, became contaminated with the
monoxide, (C6H5)2P-?(O)(C H5)2. This material might undergo

chloramination-ammonolysis giving the amide as one product,

(C6 5)2-P(O)(6H 5)2 +

[(C6H5)2P(Hm2)2]C1 + H2N-P(O)(C6H5)2

This is an interesting possibility and should be investigated

Alternatively, the amide may arise from the hydrolysis

of diphenyldiaminophosphonium chloride,

[(C6H5)2P(H 2)2]C1 + H20 (C6H5)2P(O)NH2 + NH Cl

Finally, in some experiments an intractable, benzene-

soluble tar was isolated from the chloramination-ammonolysis
of tetraphenyldiphosphine. This material melted at just

above room temperature and contained no chlorine, but still

appeared to be air sensitive, turning wet after a few
minutes exposure to the atmosphere. Some structural

features are evident from examination of the infrared

spectrum (Fig. 8). This spectrum shows N-H stretching and

deformation peaks in the region from 3380 cm-1 to 3180 cm-1

and at 1575 cm-1, respectively. Phenyl C-H stretching
peaks and "C=0" stretching peaks are observed at 3040 cm1
and at 1590 cm-1, 1480 cm-1 and 1439 cm-. A strong, broad
peak is observed at 1260 cm-1 which is characteristic of

"P=N" stretching in phosphonitrilic derivatives. A strong

peak occurs at 1175 cm- a region which is sometimes

associated with the phosphoryl group, P--0. This peak
appears as a shoulder on a strong peak at 1160 cm which

might be attributed to phenyl C-H in-plane bending. A
sharp, strong peak occurs at about 1117 cm This peak is

usually observed in compounds in which the phosphorus atom

is tetracoordinate and bonded to at least one phenyl

group.(74,75) At 921 cm- a peak of medium intensity

occurs which might be assigned to the P-NH-P grouping.(9

A similar peak occurs in the spectrum of

[(C6H 5)2 P(?2)-N=P(NTH)(C6 H 5)2]C.(1) Several sharp peaks
appear in the region of 800-695 cm Peaks in this

region are usually attributed to C-H out-of-plane bending

vibrations in che phenyl groups and to P-phenyl bonds.

The observed molecular weight (450) of this substance is

in the right range for the grouping, [(C6H5 )2-N-P(C6H5)2];

however, the analytical data obtained so far are not help-

ful in completing the characterization of this substance.

The infrared spectra of [(C6H5)2P(NH2)2]C1 and

(C6H )2P(0)RH2 (Figs. 7 and 8) resemble each other closely.
In fact, in the region from 4000 to 1200 cm- very little
difference is noted in the placement and number of peaks

in the two spectra. At 1175 cm- however, the spectrum

of (C6H5)2P(0)THp2 has a strong absorption which is attributed
to the phosphoryl group. This peak is absent in the spectrum

of [(C6H5)2P(NH2)2]C7 The two spectra also differ

considerably in the region from 1105 to 695 cm This

region contains the C-H bending peaks as well as the peaks

attributed to P-phenyl bonding. In general, we can say

that the infrared spectra confirm the identity of these


The chloramination-ammonolysis of tetraphenylcyclo-

tetraphosphine produces a material which is apparently a

linear phosphonitrilic derivative. An examination of its

infrared spectrum (Fig. 10) shows a broad N-H stretching
band which peaks at 3210 cm a strong N-H deformation

band which peaks at 1550 cm- a very broad band at 1250-
1155 cm which is very similar to the "P=N stretching
band in phosphonitrilics, and a sharp peak at 1115 cm

which is associated with tetracoordinate phosphorus bonded

to phenyl. In addition the infrared spectrum contains

peaks usually attributed to phenyl C-H stretching and

bending vibrations, "C=C" stretching vibrations and P-phenyl

bonds. Thus, the infrared spectrum is consistent for the

unit, /P The analytical data compiled on several

samples of the same material indicate the same unit since,

in all cases, the P:N ratio was almost exactly 1:2.

The average molecular weight of this material, de-

termined ,ryoscopically on several samples, varied from

750 to 950 indicatin- an average of 5-7 P-N units. The

complete solubility of this substance in solvents such as

benzene and chloroform indicates that the small amount of

chlorine present is not caused by impurities such as

ammonium chloride. Therefore, we postulate that the

chlorine is bonded to the phosphorus atom of the molecule,

perhaps as an end-stopping group. This postulate is sup-

ported somewhat by the broad, infrared peak at 510 cm-

which is in the region of absorptions attributed to the

P-Cl bond. The postulated structure is then,

C6H5 NH C6H5 NH2
6 5 / 2 6 5 / 2
Cl [ --P N -n--- P =NH where the average value of n is 5.

This formulation is supported by infrared and analytical

data as well as by consideration of the physical properties

of the material. Clearly, however, the structure of this

material is still a subject of speculation. Additional

physical and chemical information should be obtained on

this interesting substance. At the time of this writing

a nuclear magnetic resonance investigation was being under-

taken. However, the results of such a study are not yet


The second compound obtained from the chloramination-

ammonolysis of tetraphenylcyclotetraphosphine was identified

as phenyltriaminophosphonium chloride, [(C6H5)P(NH2)3]Cl.

Although this material could not be isolated in absolute

purity from ammonium chloride, there is little question of

its identity. The infrared spectrum (Fig. 11) contains the

usual N-H stretching and deformation peaks and the peaks

associated with phenyl C-H bonding, phenyl "C=C" bonding

and P-phen;l 'onding. A strong peak is observed at

1135 cm-1 which is attributed to tetracoordinate phosphorus

bonded to phenyl. The absence of peaks in the region from

1300 cm1 to 1150 cm1 indicates that the material has

neither a condensed P-N=P structure, nor is it contaminated

with the amide, C6H5P(0)(NH2)2. The amide was synthesized

easily from phenyltriaminophosphonium chloride by exposing

a solution of the chloride in alcohol to moist air. Indeed

the infrared spectrum of the amide (Fig. 12) is quite

similar to that of phenyltriaminophosphonium chloride.
However, it contains a sharp peak at 1155 cm attributed

to the phosphoryl group.

Although portions of this study are incomplete,

there are certain results which are significant. We can

now state that chloramine cleaves the phosphorus-phosphorus

bonds in substituted diphosphines and in tetraphenylcyclo-

tetraphosphine. Although other reaction sequences may be

possible, we feel that the first step in the chloramination

of diphosphines involves the formation of an aminophosphine

and a chlorophosphine,

R2P-PR2 + :`2 01 R2PNH2 + R2PCl .

This step is analogous to the first step in the reaction of

certain diphosphines with methyl iodide (Chapter II), and

is analogous to the first step in the halogenation of di-

phosphines. The aminophosphine and the chlorophosphine

could then undergo further chloramination and ammonolysis

to produce the observed products.

This study has provided a convenient new method for

the synthesis of dialkyldiaminophosphonium salts. Previ-

ously, these derivatives were synthesized from dialkyl-

chlorophosphines which are obtained with some difficulty.

For example, instead of using dimethylchlorophosphine which

has a short shelf-life and must be used as soon as it is

obtained, we can now utilize tetramethyldiphosphine which

can be stored indefinitely and used when desired. The

convenience of this method for the preparation of dialkyl-

diaminophosphonium chlorides will be further appreciated

in the discussion of Chapter IV which deals with the

condensation reactions of these salts.

This method seems to favor the formation of diamino-

phosphonium salts. Thus, using tetramethyldiphosphine and

teoraethyldiphosphine, the corresponding diaminophosphonium

salt was either the only product or the major product of

the reaction. Chloramination-ammonolysis of diethylchloro-

phosphine, on the other hand, produces only a 20 per cent

-ield of the diaminophosphonium salt. Tetraphenyldiphosphine

gives diphenyldiaminohsphophonium chloride, a material which

was a postulated product in the chloramination of diphenyl-

chlorophosphine but was never isolated. Likewise, tetra-

phen'lcyclotetraphosphine produces phenyltriaminophosphonium


In view of the possible value of diamino- and tri-

aminophosphonium salts as phosphonitrilic precursors

(see Chapter IV), we feel that this reaction should be

applied to a number of other diphosphines and polyphos-

phines. This would indicate the generality of the method

and possibly make available a variety of diamino- and

triaminophosphonium salts for further investigation. Some

poly hosphines which should be chloraminated are (CF-P)

and (2P)x where R is an alkyl group. It would also be

interesting to examine the reaction of chloramine with

amino-substituted diphosphines such as C6H5(R2N)P-P(R2N)C6H5

and (R2N)2P-P(R2N)2.


Chloramination-ammonolysis of tetramethyldiphosphine

and tetraethyldiphosphine produces dimethyldiaminophos-

phonium c' ride and diethyldiaminophosphonium chloride in

good yields, and produces traces of diethylphosphonitrilic

trimer. These materials were identified by chemical

analysis and by comparison of their infrared spectra and

melting points with those of the authentic materials pro-

duced by oth.- methods.

The chloramination-ammonolysis of tetraphenyldi-

phosphine produces a new compound, diphenyldiaminophos-

phonium chloride, and its postulated condensation product,

[(C6H5)2P(NH2)-N=P(NH2)(C6H )2]Cl. In addition, the amide,

(C6H5)2P(0)IH2, and an unresolved tar were produced in the


Chloramination-ammonolysis of tetraphenylcyclotetra-

phosphine produces a material tentatively identified as
C -H NH CF ?,-K
o6 / 2 o6- / 2
Cl- [ --N ]n P--NH where the average value of n is 5.

This reaction also produced a new compound, phenyltriamino-

phosphonium chloride. The identity of this material was

established by elemental and infrared analysis and by

synthesis of a derivative, C6H5P(O)(NH2)2.

Infrared data are presented and interpreted in terms

of the structures of the compounds synthesized in this


The chemical evidence leads to the conclusion that

the chloramination of diphosphines and polyphosphines causes

cleavage of the phosphorus-phosphorus bonds. Further

chloramination and ammonolysis of the fragments produces

the same materials which arise from the chloramination of

2PC12 or R2PCl compounds. This reaction constitutes a new


method for the production of d.amino- and triaminophos-

phonium chlorides. In this connection, the reaction is

more convenient than th. one which employs chlorophosphines

because diphosphines and polyphosphines are easier to

prepare and store.


CELORIDES AOD CF L(C2H5)2P(Ih 2)-N=P(NH2)(C2 H5)2]Cl


Nanioulation of materials.-The aminophosphonium

salts used in this study were sensitive to atmospheric

moisture, deliquescing and subsequently hydrolyzing after

only a few minutes exposure. Therefore, the starting

materials were stored and transferred in a D. L. Herring

Hoael HE-43 Dri Lab equipped with a Model HE-93 Dri Train.

The products of the condensation reactions were usually

unreactive to moist air. However, in certain cases the

products were hygroscopic in air. These materials were

dried in a vacuum desiccator and stored in sealed vials in

the desiccator.

Benzue and petroleum ether were obtained as reagent

grade materials and were dried and stored over calcium

hydride. Tri-n-butylphosphine was obtained from Food

Machinery and Chemical Corporation and used as obtained.

Phosphorus trichloride was obtained as the reagent grade

material and used as obtained. All solvents and liquid

reagents were transferred by pipette to minimize exposure

to moisture.

Met-hos of analysis.- elemental analyses were per-

formed by Galbraith Laboratories, Inc., Knoxville,

Tennessee. Some nitrogen analyses were also obtained using

a Coleman :Iodel 27 nitrc-en analyzer.

:Molecular weights were determined either by the

cryoscopic method in benzene, or in benzene or chloroform

solution using a :Kecrolab Vapor Pressure Osmometer.

The infrared spectra of materials synthesized in

this study were recorded on a Beckmann Model IR-10 spectro-

photometer. Solid samples were either mulled with Nujol,

pressed into KBr disks, or melted on KBr plates and

examined as thin films when the melting point was below

1000 C. A summary of the spectral bands of new materials

produced in this study between 3.4 and 30 microns is found

in ;.1ble 6.

Kelting points were obtained in capillary tubes

using a Thomas-Hoover capillary melting point apparatus.

Preparation of dialkyldiaminoDhosphonium chlorides

and [(C 2H52P( 2)-N=P(I'H2)(C2H5)2]Cl.-In the early stages

of this investigation, dimethyldiaminophosphonium chloride

and diethyldiaiinophosphonium chloride were prepared by the

chloramination-ammonolysis of dimethylchlorophosphine and

diethylchlorophosphine,() respectively. However, soon

after this investigation was begun, a more convenient method

was found, namely, the chloramination-ammonolysis of

TABLE 6(a)


[(CH )2::]n, KBr disk

2975(m), 2900(m), 1410(m), 1290(s), 1270(s), 1200(s,sh),
1160(vs), 1010(w,sh), 965(m), 940(m), 915(s),
880(s), 855(vs), 751(m), 720(s), 685(m),
650(w), 560(m), 490(m), 420(s), 368(w).

[(n-CqH9)2 P]3 melt

29'5(s), 2920(s), 2855(m), 2300(w,b), 1459(w),
1450(w), -GO(w), 1362(w), 100(w), 1260(w),
1224(:i), 1155v"-s), 1085(w), 1045(w,sh),
c1000(,b), 95-(wI ), 913(w), 889(w), 785(w),
719(m), 660(m,b), 505(w), 280(vs,b).

s(a) strong; m,medium; w,weak; b,broad; v,very;
sh, shoulder.

tetramethyldiphosphine and tetraethyldiphosphine. This

method of preparation is discussed in detail in Chapter III.

Substance, [(C2H5)2P( N: 2)-N=P(NH2)(C2H5)2]C1,

was prepared by the chloramination-ammonolysis of diethyl-

chlorophosphine.) It was also obtained as a product of

the condensation of dieth:ldiaminophosphonium chloride.

Di-n-butyldiaminopospsphonium chloride was synthesized

by the chloramination of di-n-b ylchlorophosphine.5)

This chlorophosphine was conveniently prepared by the method

of V. t. Plets(4) which involves the decomposition of tri-


(C H)PC2 (C4H9)2PC1 + C ,Cl

Tributyldichlorophosphoranc can be readily synthesized by

the reaction of triburylphosphine with PC1:(87)

3(C4 9g)P + 2 PCl -3(C4H9) PC12 + 2 Px

In a typical experiment, 2.91 ml (33.3 mmoles) of phosphorus

trichloride was dissolved in about 50 ml of dry benzene and

placed under nitrogen in a Minilab reaction flask (Fig. 5).

To this solution, 11.3 ml (50 mmoles) of tri-n-butylphosphine

was added dropwise with stirring under a slow flow of dry

nitrogen gas. The solution rapidly became deep, orange-red

in color arn heat was evolved. Near the end of the addi-

tion an oran- -red, amorphous solid suddenly precipitated

from the mixture. At the end of the addition the mixture


was stirred vigorously until it had cooled to room tempera-

ture (250 C). The benzene solution was removed from the

precipitate by filtration and the precipitate was washed

with 1-2, 25 ml portions of dry benzene and filtered, care

being taken not to expose the solution to atmospheric

moisture. The filtrate and washings were combined and

evaporated to dryness in vacuum. The white, crystalline

product, (C4H9)3PC12, was obtained in almost quantitative

yield. This material was transferred in the dry box to a

200 ml round bottomed distillation flask and heated under

an atmosphere of nitrogen at about 1400 C. Butyl chloride

distilled from the mixture (b.p. 780 C) and some HC1 was
evolved. When the distillation of butyl chloride had

ceased (1-5 hours) the mixture was cooled and distilled

under vacuum. The product distilled at 570 C and 0.1 mm Hg

(Lit.,6): b.p. 0.25 = 39-470 C). The yield of dibutyl-

chlorophosphine by this method was usually over 50 per cent

of theory based on the amount of (C4H9)3PC12 used.

The Pyrolytic Condensation of Dimethyldiaminophosphonium

n[(CH )2P(NH2)2]Cl [(CH3)2PN]n + n NH4Cl

n = 3,4, higher
In a typical experiment, 1.15 g (8.9 mmoles) of

dimethyldiaminophosphonium chloride was placed in a semi-

micro sublimation apparatus (Fig. 15) and heated at about

Fig. 13.-Semimicro Sublimation Apparatus.

2000 C and 0.2 mm Hg for three days. The resulting subli-

mate was removed from the cold finger under nitrogen and

extracted with several portions of hot, 30-600, petroleum

ether. The residue from this extraction was identified as

ammonium chloride. Evaporation of the petroleum ether

yielded a white, crystalline material with a melting range

of 160-1700 C. Anal. Found: C, 31.83; H, 7.98; P, 41.15;

N, 18.73. Calculated for [(CH )2PN]: C, 32.01; H, 8.06;

P, 41.28; N, 18.66. Crude yield: 0.55 g (82% of theory).

The mixture was separated into almost equal amounts of

dimethylphosphonitrilic trimer and tetramer(26) by

fractional crystallization from petroleum ether. The

molecular weights (cryoscopic in benzene) were 220 (calcu-

lated: 225) and 311 (calculated: 300), respectively. The

trimer melted with sublimation at 187-1900 C (Lit:(26) M.P.

195-196 C) and the tetramer melted with sublimation at 157-
1600 C (Lit:(26) M.P. 163-1640 C). The infrared spectra of

these compounds agree with the assignments reported by


The Synthesis of Highly Polymeric Dimethylphosphonitrile

In several pyrolyses of dimethyldiaminophosphonium

chloride, described above, a trace of a black, glassy

residue was observed in the sublimation pot at the com-

pletion of the sublimation of ammonium chloride and

[(CH )2PN] ,4. This material melted in the range of 156-

1460 C and was not visibly sensitive to moist air. Anal.

Found: C, 51.81; H, 8.16; N, 18.79; P, 41.14; Mol. Wt., 7640

(Mol. Wt. of a (CH3)2PN-unit, 75). Thus, it was discovered

that the pyrolytic condensation of dimethyldiaminophos-

phonium chloride produces trace amounts of [(CH )2PN]n where

n is a large number. The yield of high polymer could be

increased considerably by mixing the starting phosphonium

salt with finely divided ammonium chloride produced by the

gas phase reaction of chlorine and ammonia. In a typical

experiment the phosphonium salt and a weighed amount of

ammonium chloride were thoroughly mixed in the dry box,

placed in a semimicro sublimation apparatus and pyrolyzed

as before. The dimethylphosphonitrilic trimer and tetramer

and ammonium chloride sublimed to the cold finger and were

separated as before. The high polymer remained in the

sublimation pot and was either scraped from this vessel or

removed by dissolving it in benzene. The effect of ammonium

chloride on the distribution of products of this reaction

is shown in Table 7. The infrared spectrum of the crude

polymer is shown in Figure 14.

The high polymers so formed were usually soluble in

hot benzene and chloroform. In some cases a residue was

observed which swelled in benzene but did not dissolve.

Extraction(95) of the crude material with cyclohexane pro-

duced, upon evaporation of the solvent, a white powder

melting at 139-1460 C and having a molecular weight of from



Wt. % of NH Cl in



Starting Material Yield of Polymer, Yield of [(CH )2PN]5 4
_%% 5> d 4












20 11.2 68











3500 to 9000 depending on the sample used. The residue from
the cyclohexane extraction was soluble in benzene, melted

at 136-1380 C and had a molecular weight of 12,500.

Thermal gravimetric analysis(95) (using an Aminco

Thermograv, American Instrument Co.) of a sample with an

average molecular weight between 8000 and 9000 was obtained

using a 100 cc/min flow of helium at a rate of temperature

increase of 3/min. This analysis showed no weight loss up

to 300 C, and a rapid loss of weight from 500 to 5000 C
during which essentially all of the sample sublimed from
the apparatus as low-molecular-weight cyclic material,

principally trimer and tetramer.
Differential thermal analysis was carried out on

this material(95) in helium using a heating rate of 3/min

with A1203 as a reference. A sharp endotherm was observed

at 1420 C (M.P. of sample: 143-1450 C) and a broad exotherm

was observed starting at about 300 C, the latter presumably

resulting from the onset of thermal decomposition.

The Pyrolytic Condensation of Diethyldiaminophosphonium

2[(C2H )2P(NH2)2]Cl [(C2H5)2P(NH2)-N=P(NH2)(C2H5)2]Cl
+ NH4Cl

[(C2H5)2P(NH2)2]Cl + [(C2H5)2P(NH2-N=P(NH2)(C2H5)2]Cl

- [(C2H )2PN]3 + 2 NH C1

In a typical experiment, 0.99 g (6.5 mmoles) of di-

ethyldiaminophosphonium chloride was placed in a semimicro

sublimation apparatus and heated to about 1800 C and 0.2 mm

Hg for five days. This process produced a white sublimate

as well as a substantial amount of light-colored residue.

The residue was dissolved in boiling benzene and filtered.

Upon evaporation of the benzene a white, crystalline

material was recovered which melted at 61-64 C; yield,

0.30 g. This material is hygroscopic, water-soluble and
gives a Cl- test with aqueous AgNO,. Anal. Found: N, 16.5.

Calculated for [(C2H5)2P(NH2)-N=P(NH2)(C2H5)2]C: N, 16.18.

The infrared spectrum of this material was identical with

that of [(C2H5)2P(NH2)-N=P(NH2)(C2H5)2]C1 (M.P. 58-610 C)

produced by chloramination-ammonolysis of diethylchloro-

phosphine.5) The nuclear magnetic resonance data for

[(C2H5)2P(NH2)-N=P(NH2)(C2H )2]C1, not included in the
earlier reference, 5) provides some interesting additional

information about the structure of this substance. The

proton magnetic resonance spectrum (Fig. 15) was run on a

Varian high resolution spectrometer at 56.4 Mc using a

CDC13 solution. The spectrum was obtained by sweeping

slowly through the field and interchanging the reference,
acetaldehyde, with the sample. Peak A refers to the NH2

protons, peak B to the methylene protons and peak C to the

methyl protons. The usual methylene quartet and methyl






_ __ _ **"==; --

^ -^. 0

----- : -z= ^ *
-- .-^ ^-

'-^ m ^-


triplet are further split by spin-spin coupling with the

phosphorus atoms. The spectrum is consistent with the

S2 2
structure(1) C2H5 P N PC2HH5 j C.
C2H5 C2H5

Approximate P values and average chemical shift values are

listed in Table 8.



Group Chemical Shift,T Coupling,

NH2 5.18 H-H 8

CH2 7.98 (ave.) P-CH2 14

CH3 8.75 (ave.) P-CH3 18

The ratio of areas under the NH2 peak to those under

the CH2 and CH3 peaks is approximately 1:5. When the

nuclear magnetic resonance spectrum was run at 19.3 Mc

a single, broad peak for phosphorus was observed corres-

ponding to a chemical shift of 41 p.p.m. to low field in

comparison to 85 per cent phosphoric acid.

The sublimate was dissolved in benzene and filtered

to remove ammonium chloride. The benzene was removed by

evaporation and the resulting white, crystalline material

was carefully resublimed at about 500 C and 0.1 mm Hg. The
sublimate melted at 109-1120 C. Anal. Found: C, 46.39; H,
9.92; P, 30.30; N, 15.49; Mol. Wt. 305. Calculated for
[(C2H5)2PN] : C, 46.60; H, 9.78; P, 30.04; N, 13.59; Mol.
Wt. 309. Yield: 0.21 g (33% of theory). The infrared
spectrum of this material agreed with the assignments listed
in the literature(25) for diethylphosphonitrilic trimer.
The literature value for the melting point of the trimer is
117.5-1190 C.
The residue from this sublimation was recrystallized
from benzene and identified as the intermediate,
[(C2H5)2P(NH2)-N=P(NH2)(C2H5)2]C1. The total yield of this
intermediate was 0.54 g (41% of theory).

The Pyrolytic Condensation of
[(C2H5)2P(NH2)-N=P(NH2) (C2H5)2]C1

[(C2H5)2P(NH2)-N=P(I2)(C25)2]C1 -[(C2H5)2PN],4 + NH4C

In a semimicro sublimation apparatus, 0.85 g (3.5
mmoles) of (C2H5)2P(NH2)-N=(NNH2)(C2H5)2]Cl was heated at
about 2000 C and 0.15 mm Hg for 4-5 days. This process
produced a white, oily sublimate and a trace of black re-
sidual ash. The sublimate was extracted with several
portions of petroleum ether. The residue from this extrac-
tion was again extracted with several portions of hot

benzene. From the benzene solution, 0.25 g (0.96 mmoles) of

unreacted starting material was recovered. From the

petroleum ether solution, 0.36 g (75% of theory based on the

starting material actually reacted) of a mixture of di-

ethylphosphonitrilic tetramer and trimer was recovered by

evaporation. The final residue of the solvent extractions

consisted of ammonium chloride.

The mixture of diethylphosphonitrilic tetramer and

trimer was separated by repeated fractional sublimation at

530 C and 0.1 mm Hg. By this method the mixture was found
to consist of about 20 per cent trimer (M.P. 110-1140 C)

and 80 per cent tetramer (a viscous oil 25). Anal. Found:

C, 46.36; H, 9.76; P, 29.84; N, 13.30; Mol. Wt. 401.

Calculated for [(C2H5)2PN]4: C, 46.60; H, 9.78; P, 30.04;

N, 13.59; Mol. Wt., 412. The infrared spectrum of this

material agrees quite well with the assignments listed in

the literature(25) for diethylphosphonitrilic tetramer.

The Pyrolytic Condensation of Di-n-butyldiamino-
phosphonium Chloride

3[(n-C4H9)2P(NH2)2]C1 [(n-C4H9)2PN] + 3 NH4C1

Di-n-butyldiaminophosphonium chloride (1.08 g or 5

mmoles) was placed in a semimicro sublimation apparatus and

heated at 1900 C and 0.1 mm Hg for four days. A white

sublimate was observed. Only a trace of black residue

remained in the pot at the completion of the reaction. The

sublimate was extracted with 30-600 petroleum ether. The

insoluble residue was shown to be ammonium chloride. Evapo-

ration of the petroleum ether yielded a white wax which

melted at 45-500 C and sublimed readily at 70 C and 0.1 mm

Hg. Anal. Found: C, 60.09; H, 11.37; P, 19.23; N, 8.90;

Mol. Wt., 470. Calculated for [(C4H9)2PN]3: C, 60.35; H,

11.40; P, 19.45; N, 8.80; Mol. Wt., 478. Yield: 0.77 g

(95% of theory). The infrared spectrum of dibutylphospho-
nitrilic trimer is shown in Figure 16.


The results of this study show that diaminophosphonium

chlorides can be easily converted by pyrolytic condensation

to phosphonitrilic derivatives. The yields observed in such

a process are essentially quantitative.

The pyrolytic condensation of dimethyldiaminophos-

phonium chloride produces amounts of dimethylphosphonitrilic

trimer and tetramer. These compounds were separated from

the byproduct, ammonium chloride, by their solubility in

petroleum ether, and the trimer/tetramer mixture was sepa-

rated by fractional crystallization from petroleum ether.

The materials were identified by elemental analysis, molecu-

lar weight determinations and comparison of their infrared

spectra with published infrared data.(26) When finely






o0 I

O 1
0H 0

r- 0 ~

S0 -P
-H 01 0

0 0)
O cd
S0 nI



.r 1

divided ammonium chloride was mixed with the starting

material, pyrolysis produced a new compound identified as

highly polymeric dimethylphosphonitrile. This material is

r 3, / 3
a polymer of the formula,[ P-----N--n where n varies

from 50 to 150. It was characterized by elemental analysis,

molecular weight determinations, infrared analysis, thermal

gravimetric analysis and differential analysis. The infra-

red spectrum of this material (Fig. 14) is somewhat similar

to that of the cyclic trimer or tetramer having C-H stretch-

ing and deformation peaks at 2970 cm1 and at 1410 cm-1
respectively. Two peaks occur at 1290 and 1270 cm1 which

are attributed to P-CH3 bonding. (26) The "P=N" stretching

band exhibits a broad, strong shoulder at about 1200 cm1

and a strong peak at 1160 cm- There are several peaks in

the region from 850 cm1 to 965 cm1 and from 685 cm- to

755 cm-1. Peaks in this region are generally attributed to

"P=N" elongation and "P-N=P" deformation.(24'25) Definite

assignments for these peaks cannot be made at the present

time. Yields of the polymer as high as 32 per cent of

theory were obtained when mixtures of about 10 per cent

finely divided ammonium chloride and 90 per cent dimethyl-

diaminophosphonium chloride were pyrolyzed.

The pyrolytic condensation of diethyldiaminophos-

phonium chloride produces the primary condensation product,

[(C2H5)2P(NH2)-N=P(NH2)(C2H5)2]Cl which was recovered in
about 40 per cent yield, and diethylphosphonitrilic trimer,

[(C2H5)2PN]3 which was recovered in about 30 per cent yield.
The first condensation product was shown by elemental

analysis and infrared analysis to be identical with

[(C2H5)2P(NH2)-N=P(NH2)(C2H5)2]Cl produced by the chlorami-
nation-ammonolysis of diethylchlorophosphine. 5 Diethyl-

phosphonitrilic trimer was identified by elemental analysis,

molecular weight determination and comparison of its infra-
red spectrum with published infrared data. 25

The pyrolytic condensation of the intermediate,

C(C2H5)2P(NH2)-N=P(NH2)(C2H5)2]C1, produces a mixture of
diethylphosphonitrilic trimer and tetramer which was re-

covered in about 75 per cent yield. The phosphonitrilic

mixture, consisting of about 80 per cent tetramer and 20

per cent trimer, was separated from the byproduct, ammonium

chloride, by its solubility in petroleum ether. The trimer

and tetramer were then separated from each other by repeated,

fractional vacuum sublimation. In this case, the trimer was

slightly more volatile than the tetramer and a separation

could be made.

The pyrolytic condensation of di-n-butyldiamino-

phosphonium chloride produces a nearly quantitative yield

of a new material identified as di-n-butylphosphonitrilic

trimer, [(n-C4H9)2PN]3. This material was characterized by

elemental analysis, molecular weight determination and infra-

red analysis. The infrared spectrum of this material (Fig.

16) has the usual C-H stretching and deformation peaks in

the region of 2950-2850 cm1 and 1450-1370 cm- respec-

tively.. Bands attributed to skeletal -CH2- vibrations occur

at 1400 cm-1 1500 cm1 and at 1260 cm-1. A sharp peak is

observed at 1215 cm-1 which may be attributed to P-C bonding.
The "P=N" stretching peak occurs at 1155 cm Of all the

phosphonitrilic derivatives for which this peak has been

measured, it occurs at the lowest energy in this derivative.

This is to be expected because the butyl groups have little,

if any, electron attracting ability. Thus, they are unable

to cause appreciable delocalization of the nitrogen lone-pair

electrons into the region of the P-N bond. The remainder of

the spectrum between 1155 cm-1 and 650 cm-1 contains several

peaks of weak intensity, some of which may be attributed to

skeletal vibrations of the butyl groups. Others lie in the

region assigned to "P=N" deformation peaks, "P-N=P" stretch-

ing peaks and ring breathing vibrations. We can make no

definite assignments in this region at this time.

This study establishes the fact that dimethyldiamino-

phosphonium chloride, diethyldiaminophosphonium chloride,

di-n-butyldiaminophosphonium chloride and

[(C2H5)2P(NH2)-N=P(NH2)(C2H5)2]C1 are convenient phospho-

nitrilic precursors, giving excellent yields of cyclic

phosphonitrilics under pyrolytic conditions. Combined with
the findings of Chapter III, this study constitutes an
important part of a new method of synthesis of such deriva-
tives from substituted diphosphines, chloramine and ammonia.
The results of this investigation have provided very
convincing evidence in favor of the step-wise reaction se-

quence postulated(1) earlier in which the then unknown
diaminophosphonium ion, R2P(NH2)2(, was considered to be
a reactive intermediate leading to phosphonitrilic
derivatives. The case in point is the pyrolysis of di-
ethyldiaminophosphonium chloride. The only products of
this reaction are [(C2H5)2P(NH2)-N=P(NH2)(C2H5)2]Cl and
[(C2H5)2PN]3. Indeed, these are the expected products if
one assumes the following sequence

2[(C2H5)2P(NH2)2]Cl [(C2H5)2P(NH2)=N-P(NH2)(C2H5)2]Cl

+ NH4Cl

[(C2H5)2P(NH2)2]1C + [(C2H5)2P(NH2)-N=P(NH2)(C2H5)2]Cl

[(C2H5)2PN]3 + 2 NH4Cl

The pyrolysis of the intermediate,
[(C2H5)2P(NH2)-N=P(NH2)(C2H5)2]C1, produces chiefly diethyl-
phosphonitrilic tetramer. Some cleavage of the intermediate
does occur, however, resulting in the formation of diethyl-
phosphonitrilic trimer and some decomposed material.

One of the most interesting results of this study

was the discovery and characterization of highly polymeric

dimethylphosphonitrile. This discovery has attracted

interest for several reasons. First of all, despite all

the previous work in phosphonitrilic chemistry, there is a

paucity of information concerning the synthesis of thermally

stable and unreactive high polymers with the repeating unit

---PP=N--- (See Chapter II). This method of synthesis is

one of the few which results in a phosphonitrilic high

polymer which is soluble in common solvents, which can be

molded mechanically and which is relatively unreactive

toward moisture and oxygen.

In addition to exciting academic interest, this

method may also prove valuable in the synthesis of com-

mercially useful polymers. Although the polymeric material

synthesized in this study has some disadvantages in its

ready solubility, low melting point and thermal instability

above 3000 C, films of the substance can be cast which have

reasonable strength.(95) The dimethylphosphonitrile is
isoelectronic with dimethylsiloxane, --- Si--- 0-4-, the

basic unit of an inorganic polymer system which has enjoyed

such wide use in recent years that its parent, (CH )2SiC12,

is now produced commercially in large quantities. If the

isoelectronic relationship is any indication of chemical

and physical similarities which may exist in these two

systems of inorganic polymers, then there is a possibility

that the process discovered in this study may become com-

mercially important in the future.

It is interesting to note that the yield of high

polymers from the pyrolysis of dimethyldiaminophosphonium

chloride seems to depend on the amount of ammonium chloride

mixed with the starting material (Table 7), under similar

conditions of temperature, pressure and time. The postu-

lated first step of the condensation process,

2[(CH )2P(NH2)2]C1 [(CH )2P(NH2)-N=P(NH2)(CH3)2]C1

+ NH4Cl ,

has been confirmed in the pyrolysis of diethyldiaminophos-

phonium chloride. Since the actual mechanism of condensa-

tion is not clearly known, any statement concerning the

mechanism of the effect of ammonium chloride on the first

step or subsequent intermolecular condensations would be

highly speculative.

However, in view of the fact that the cyclic di-

methylphosphonitriles are weak bases,(19) it is conceivable

that in the molten reaction mixture a proton transfer could

occur between ammonium chloride and a ring nitrogen. This

could result in the breaking of a P-N bond and the formation

of a =P species which could attack other ring-nitrogen

atoms and lead to linear products. For instance, the attack

could occur on a cyclic trimer

:N N:
Rj ,R + H @


:N N:H


HN===P-N= =P 0

HN=P=P-N- -P ( + :N N:


R \ / R \ /
- HN=P--N P ,


This is similar to the mechanism postulated by Allcock and

Best(52) for the polymerization of [PNC12]3 4 (see page 21).

The polymer formed is not highly cross-linked and therefore

undergoes almost complete depolymerization to cyclic deriva-

tives in the 300-5000 range.

There are two suggestions which we can make for

further work in this area. Both suggestions are concerned

with the formation of high polymers. The effect of ammonium

chloride on the pyrolytic condensation of diaminophosphonium

chlorides was investigated only in the case of dimethyl-

diaminophosphonium chloride. This effect should be studied

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs