Transfer of Sulfur from Arsenic and Antimony Sulfides to Phosphorus Sulfides. Rational Syntheses of Several Less-Common P4Sn Species

Article abstract of DOI:10.1021/ic9614881

It has been shown that triphenylarsenic sulfide and triphenylantimony sulfide rapidly transfer sulfur to a number of the known phosphorus sulfides. The reactions are performed at or below room temperature in carbon disulfide solutions. The transfers are neither highly selective nor random, making them useful but not ideal for synthetic purposes. The moderate selectivities of the reactions have been used in the assignment of structures to two new phosphorus sulfides, structures with molecular formulas P₄S₆ and P₄S₈. The reactions of P₄S₃ are unusual in that products with 7-9 sulfur atoms in the molecule are formed competitively with low sulfur products. The usefulness of triphenylantimony sulfide is limited by its tendency to undergo reductive elimination of sulfur. This reduction takes the form of a disproportionation to give triphenylantimony and elemental sulfur and has been shown to occur by a second-order process that appears to involve the formation of disulfur, S₂.

Full text of DOI:10.1021/ic9614881

Inorg. Chem. 1997, 36, 2641-2646
2641
Transfer of Sulfur from Arsenic and Antimony Sulfides to Phosphorus Sulfides. Rational
Syntheses of Several Less-Common P4Sn Species
Mark E. Jason
Monsanto Company, Performance Materials, 800 N. Lindbergh Blvd., T3W, St. Louis, Missouri 63167
ReceiVed December 13, 1996X
It has been shown that triphenylarsenic sulfide and triphenylantimony sulfide rapidly transfer sulfur to a number
of the known phosphorus sulfides. The reactions are performed at or below room temperature in carbon disulfide
solutions. The transfers are neither highly selective nor random, making them useful but not ideal for synthetic
purposes. The moderate selectivities of the reactions have been used in the assignment of structures to two new
phosphorus sulfides, structures with molecular formulas P4S6 and P4S8. The reactions of P4S3 are unusual in that
products with 7-9 sulfur atoms in the molecule are formed competitively with low sulfur products. The usefulness
of triphenylantimony sulfide is limited by its tendency to undergo reductive elimination of sulfur. This reduction
takes the form of a disproportionation to give triphenylantimony and elemental sulfur and has been shown to
occur by a second-order process that appears to involve the formation of disulfur, S2.
Introduction
The preceding paper1 discusses the reactions of elemental
phosphorus with elemental sulfur. In the history of this
chemistry, there have been relatively few tools in the craft of
assembling the products of these reactions. In fact, all synthetic
efforts can be collected into three categories: thermal reactions
(t),2-7 abstraction of sulfur with phosphines (a),8-10 and
construction of P-S bonds with sulfur equivalents (c).11-13
Examples of these three types of reaction are shown in Figure
1. An association can be made between each of the previously
isolated phosphorus sulfides and its methods of preparation, as
is done in Figure 2. Clearly, the most common method is the
equilibration of a mixture of phosphorus, or alternatively a lower
sulfide, and sulfur.
The nonthermal reactions in Figure 1 are mild and reasonably
specific. Triphenylphosphine apparently plucks off an exocyclic
sulfur from the outside of a phosphorus sulfide cage in distinct
preference to removal of an endocyclic sulfur. The oxidation
reactions shown in Figure 1 are quite specific: the intermediate
diiodide (or similar structure) is required in order to carry out
X Abstract published in AdVance ACS Abstracts, May 15, 1997.
(1) Preceding article: Jason, M. E.; Ngo, T.; Rahman, S. Inorg. Chem.
Figure 1. Examples of the known synthetic methods used to prepare
1997, 36, 2633-2640.
phosphorus sulfides.
(2) Hoffman, H.; Becke-Goering, M. In Topics in Phosphorus Chemistry;
Griffith, E. J., Grayson, M., Eds.; Interscience Publishers: New York,
1976; Vol. 8, pp 193-271.
the second step. It is evident there are few known ways in
which this class of compounds can be intelligently manipulated.
(3) Brylewicz, Z.; Rudnicki, R. Phosphorus, Sulfur Silicon, 1994, 89, 173-
179.
(4) Bues, W.; Somer, M.; Brockner, W. Z. Anorg. Allg. Chem. 1981, 476,
153-158.
It was in asking the question whether the reduction of
phosphorus sulfides with triphenylphosphine was a reversible
reaction that another tool in the synthetic chemistry of these
compounds was struck. Triphenylphosphine sulfide, under any
reasonable conditions, does not return sulfur to any of the easily
prepared phosphorus sulfides. But Baechler had already shown
that other, better, sulfur donors exist and that the rates of sulfur
transfer from triphenylarsenic sulfide or triphenylantimony
sulfide to triphenylphosphine were fast.14 It remained only to
put the phosphorus sulfides together with these more willing
donors.
(5) Thamm, R.; Heckmann, G.; Fluck, E. Phosphorus Sulfur 1981, 11,
273-278.
(6) Bjorholm, T.; Jakobsen, H. J. J. Am. Chem. Soc. 1991, 113, 27-32.
(7) Blachnik, R.; Peukert, U.; Czediwoda, A.; Engelen, B.; Boldt, K. Z.
Anorg. Allg. Chem. 1995, 621, 1637-1643.
(8) Meisel, M.; Grunze, H. Z. Anorg. Allg. Chem. 1970, 373, 265-278.
(9) Barieux, J.-J.; Demarcq, M. C. J. Chem. Soc., Chem. Commun. 1982,
176-177.
(10) Griffin, A. M.; Sheldrick, G. M. Acta Crystallogr. 1975, B31, 2738-
2740.
(11) Griffin, A. M.; Minshall, P. C.; Sheldrick, G. M. J. Chem. Soc., Chem.
Commun. 1976, 809-810.
(12) Minshall, P. C.; Sheldrick, G. M. Acta Crystallogr., 1978, B34, 1326-
1328.
(13) Chang, C.-C.; Haltiwanger, R. C.; Norman, A. D. Inorg. Chem. 1978,
17, 2056-2062.
(14) Baechler, R. D.; Stack, M.; Stevenson, K.; Vanvalkenburgh, V.
Phosphorus, Sulfur Silicon 1990, 48, 49-52.
S0020-1669(96)01488-7 CCC: $14.00 © 1997 American Chemical Society

2642 Inorganic Chemistry, Vol. 36, No. 12, 1997
Jason
the chemical shifts and coupling constants for the species reported in
this study were obtained by fitting the spectroscopic data. A table of
these data can be found in the Supporting Information.
Typical Reaction of Triphenylarsenic or Triphenylantimony
Sulfide with a Phosphorus Sulfide. A 100 mL three neck flask, fitted
with a reflux condenser, nitrogen inlet, thermocouple, and magnetic
stirring bar, was charged, in order, with 30 mL of dry carbon disulfide,
0.185 g (5.3 × 10-4 mol) of R-P4S7, and 0.539 g (15.9 × 10-4 mol) of
triphenylarsenic sulfide, all at room temperature. A thin slurry of R-P4S7
was produced. At 30 min, the solution had become cloudy. After 2
h, an off-white precipitate had formed and the solution was no longer
cloudy. The newly formed precipitate had deposited on the unreacted
R-P4S7. Samples of the carbon disulfide layer were removed, added
to a nitrogen-purged NMR tube containing a lock capillary, and
analyzed by 31P NMR spectroscopy.
Synthesis of r-P4S6. To a Schlenk flask was added 0.224 g (7.88
× 10-4 mol) of R-P4S5 and 90 mL of dry carbon disulfide. Triphen-
ylarsenic sulfide, 0.265 g (7.84 × 10-4 mol), was added as a dry solid
under nitrogen. After 2 h, 1H and 31P NMR spectroscopy showed
greater than 90% reduction of the triphenylarsenic sulfide and slightly
more than 50% conversion of the R-P4S5 to other products. A second
addition of arsine sulfide, 0.204 g (6.02 × 10-4 mol), was made and
the reaction continued at room temperature for another 2 h. The
phosphorus sulfides were precipitated by the addition of 11 mL of dry
toluene and removal of the carbon disulfide under mild vacuum. A
light yellow solid was obtained that was filtered under argon, washed
with toluene, and dried under vacuum. This solid contained a small
amount of residual R-P4S5 along with R-P4S6 as the major product,
γ-P4S6, R-P4S7, and a very small quantity of â-P4S8. Most of the R-P4S7
could be removed by partially dissolving the crude solids in 90 mL of
hot carbon disulfide and filtering. A second crop of R-P4S7 was
removed by cooling the filtrate in CO2/acetone and filtering the slurry.
The product was obtained by evaporating the filtrate to about 5 mL
total volume under vacuum followed by filtration. The assay of this
material by 31P NMR showed 84% R-P4S6, 10% γ-P4S6, and 6% R-P4S7.
This mixture gave an M + 1 peak for P4S6 as the base peak in the
(isobutane) CI mass spectrum.
Figure 2. Known phosphorus sulfides, their trivial names, and the
literature methods used to prepare them. In the figure “t” refers to
thermal, “a” to abstraction, and “c” to construction methods of synthesis.
See text for discussion.
Synthesis of â-P4S8. R-P4S7, 0.142 g (4.071 × 10-4 mol), was added
to 250 mL of dry, distilled carbon disulfide in a round bottom flask.
The mixture was maintained under a nitrogen atmosphere and stirred
until only a small amount of solid remained. A solution of 0.116 g
(3.42 × 10-4 mol, 84% of theory) of triphenylarsenic sulfide in 10 mL
of dry carbon disulfide was loaded into a syringe. The sulfide solution
was added through a septum to the R-P4S7 solution in 1 mL increments
over the course of 140 min. The mixture was allowed to stir overnight
at room temperature. The 31P NMR spectrum of the CS2 solution
showed only R-P4S7 and â-P4S8. After 24 h, 10 mL of dry toluene
was added and most of the CS2 distilled at atmospheric pressure. The
precipitate that formed was identified as R-P4S7, and the filtrate showed
a 1:3 ratio of R-P4S7 to â-P4S8. Attempts to further enrich this material
in â-P4S8 were not successful.
The reactions of phosphorus sulfides with triphenylarsenic
sulfide and triphenylantimony sulfide have provided the syn-
thetic entry into heretofore unavailable phosphorus sulfides as
well as deliver strong evidence for the structures of several
phosphorus sulfides that have only been observed as low
concentration products from the reactions of phosphorus and
sulfur. These reactive triphenylelement sulfides have the
potential to be remarkable reagents for both the transfer of sulfur,
to phosphorus and other acceptors, and the determination of
sulfide structures.
Experimental Section
Kinetics of Triphenylantimony Sulfide Decomposition. Solution
kinetics were performed by diluting 84.4 mg of commercial triphen-
ylantimony sulfide in a 10 mL volumetric flask with dry, distilled carbon
disulfide. The solution concentration was 0.022 M. Four NMR tubes
were prepared from this stock solution by volumetric dilution directly
in the tubes. The NMR tubes were evacuated and sealed. The NMR
spectrometer was thermostated at 35.0 °C. The 1H NMR spectroscopy
used a 30° pulse, a 10 s delay, and summation of 16 transients. The
aromatic signals for triphenylantimony sulfide and triphenylantimony
are baseline resolved, and quantitation was performed by integration
of these signals. The analysis of the kinetics was performed by
numerical integration using the program SIMUSOLV (Mitchell and
Gauthier).
General Information. P4S3 was obtained from Fluka Chemical Co.
It was recrystallized from CS2 prior to use. P4S10 was purified by
Soxhlet extraction. Carbon disulfide used for spectroscopy was used
as received (Fisher); the carbon disulfide used as solvent for chemical
reactions was dried by refluxing over P4S10. R-P4S5,15 â-P4S5,4 R-P4S7,5
3
and R-P4S9 were prepared according to literature procedures. Tri-
phenylphosphine, triphenylarsenic, and triphenylantimony are com-
mercially available from Aldrich Chemical Co., and triphenylantimony
sulfide is available from Strem Chemical Co. Only triphenylphosphine
was recrystallized (EtOH) prior to use. Triphenylphosphine sulfide
was harvested from the preparation of â-P4S5 and recrystallized twice.
Triphenylarsenic sulfide was prepared according to the literature.14
All of the synthetic and sample preparations were performed either
using Schlenk techniques, under nitrogen, or in a glovebag. Sulfides
were most commonly stored under vacuum in Schlenk tubes in the
dark. Spectroscopy was performed as described in the previous paper.1
Spin simulation and refinement calculations were performed using
the program gNMR (Cherwell, Ltd., available from SoftShell). All of
Reaction of Triphenylantimony Sulfide with 2,3-Dimethylbuta-
diene. To a 5 mL flask fitted with a stir bar and septum was added,
in order, 0.85 g (2.21 mmol) of triphenylantimony sulfide, 0.93 g (11.3
mmol) of 2,3-dimethylbutadiene (Aldrich), and 1 mL of f CS2. The
mixture was sealed with a septum and stirred overnight at room
temperature. After about 18 h, the material was a paste, primarily as
a result of loss of CS2 by evaporation. A small sample of the reaction
mass was dissolved in CS2 in an NMR tube. The 13C NMR peaks for
(15) Brauer, G. Handbook of PreparatiVe Inorganic Chemistry; Ferdinand
Enke Verlak: Stuttgart, Germany, 1960; p 506.

Less-Common P4Sn Species
Inorganic Chemistry, Vol. 36, No. 12, 1997 2643
Table 1. Reactant Ratios and Products from the Reactions of Phosphorus Sulfides with Triphenylarsenic Sulfide (19) and Triphenylantimony
Sulfide (21)
product composn
reacn
no.
Ph3MS/
P4Sn
sulfide
Ph3MS
2
3
5
6
8
9
10 11 12
14
15 16 17
18
E
F
H
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
18
17
19
P4S3 (2)
P4S3 (2)
19
21
19
19
19
19
19
19
19
19
19
19
21
21
19
21
19
19
19
1.2
0.7
0.2
1.0
2.0
1.0
1.0
0.9
4.5
0.6
1.0
2.0
1.7
3.0
3.0
4.0
0.5
2.0
1.0
93
95
2
1
2
1
1
2
11
25
6
1
0.5
3
3
2
1
0.5
14
33
47
61
37
a
R-P4S5 (5)
R-P4S5 (5)
R-P4S5 (5)
R-P4S5 (5)
R-P4S5 (5)
â-P4S5 (6)
â-P4S5 (6)
R-P4S7 (11)
R-P4S7 (11)
R-P4S7 (11)
R-P4S7 (11)
R-P4S7 (11)
R-P4S7 (11)
R-P4S7 (11)
R-P4S9 (16)
R-P4S9 (16)
P4S10 (18)
76
46
0
18
39
6
9
5
1
a
21
1
13
21
4
42
35
15
30
28
28
20
29
15
9
7
70
1
1
a
69
67
73
20
11
10
0
1
4
5
a
1
1
37
54
36 12
31
53
58
0
b
b
14
20
26
63
39
42
100
b
b
7
b
b
1
0
4
6
5
a Visible by NMR spectroscopy but too small to quantify. b Appears in filtrate after filtration and standing at room temperature for 1 day.
the adduct disulfide and tetrasulfide were observed as minor compo-
nents.16 These results are discussed below.
constants for transfer of sulfur to the phosphorus sulfides should
then be larger than for 19; the transfers should be more nearly
quantitative. However, the reactions between the phosphorus
sulfides and 21 were far less complete than with 19 at the same
stoichiometry.
Results
Reactions of Phosphorus Sulfides with Triphenylarsenic
Sulfide (19). The relative concentrations of the products of
many of the reactions performed in this study are collected in
Table 1. The relative percent figures in Table 1 are taken from
the 31P NMR spectra of the CS2-soluble material from the
reaction and therefore are not relative yields. The data in Table
1 should be used to separate major from minor products.
Because of the small scale (1-5 mg) of many of these
experiments that were performed directly in NMR tubes, it was
not possible to quantitatively recover and independently assay
the solids. The structures of the products, as well as they are
understood, are shown in Figure 2. The structures for com-
pounds E, F, and H are unknown at this time, though the
chemical shifts and coupling constants have been refined.
Compound E shows two triplets: 92.3 and 39.2 ppm with a
coupling constant of 12.3 Hz. Compound F shows three unique
phosphorus signals: P1, 47.6 ppm; P2,3, 42.7 ppm; P4, 26.5 ppm;
coupling constants J1,2-3 ) 32.3 Hz, J1,4 ) 29.0 Hz, and J2-3,4
) 88.8 Hz. Compound H shows an A3B pattern: P1, 44.9 ppm;
P2-4, 27.6 ppm; J1,2-4 ) 29.9 Hz. The mixtures in which these
sulfides appear, the chemical shift ranges, and the small coupling
constants all argue that these materials are phosphorus-sulfur
cages of high sulfur content. They are produced from reactions
with both triphenylarsenic sulfide and triphenylantimony sulfide,
making it unlikely they incorporate the Ph3M group in the
structure.
Attempts to prepare moderate concentration solutions of 21
in carbon disulfide resulted in disproportionation to give
triphenylantimony (22) and a noncrystalline form of sulfur.
These results are discussed below. Typically, cold carbon
disulfide would be added to a flask containing 21 and the
phosphorus sulfide cooled in an ice bath in order to minimize
the disproportionation. The reactions were allowed to proceed
cold for a few minutes and were then warmed slowly to room
temperature. In all cases, a precipitate of an uncharacterized
form of sulfur, often contaminated with phosphorus sulfides,
was noticed immediately upon addition of the carbon disulfide.
Because of these limitations on the use of 21, most of the
synthetic and mechanistic studies were performed with the
arsenic sulfide 19.
Reactions of Phosphorus Sulfides with Triphenylarsenic
and Triphenylantimony. During the investigation of the
possible mechanisms for these sulfide transfer reactions, it was
important to know the extent to which they were reversible or
participated in addition reactions with the intermediate sulfides.
In separate experiments, â-P4S5 and R-P4S7 were allowed to
react with an equivalent of 22 for 24 h in refluxing carbon
disulfide. No reaction to form other phosphorus sulfides was
observed by 31P NMR, and no formation of triphenylantimony
1
sulfide was observed by H NMR spectroscopy.
These reactions would often not go to completion, as
measured by the 1H NMR spectrum of the solution. The
chemical shifts of 19 and triphenylarsenic (20) are well separated
and easily integrated. The higher sulfides (R-P4S9, P4S10) appear
capable of transferring sulfur to 20. Thus, unlike the removal
of sulfur from higher sulfides with triphenylphosphine, an
apparently irreversible reaction, the transfer of sulfur from 19
to phosphorus compounds is reversible.
Reactions of Phosphorus Sulfides with Triphenylantimony
Sulfide (21). The work of Baechler and his students14 predicted
that triphenylantimony sulfide (21) would react even more
efficiently than 19 with phosphorus sulfides. The equilibrium
Excess P4S10 was observed to rapidly oxidize 20 to give 19,
with R-P4S9 as the only phosphorus sulfide produced. Similarly,
R-P4S9 was very slowly reduced to give predominately R-P4S7
along with a small quantity of R-P4S8. Because of the slow
self-reaction of 19 to give sulfur and 20, the calculation of
equilibrium constants from the data was not possible.
Stability of the Reagents. One of the first observations made
in working with triphenylantimony sulfide (21) was its rapid
reaction with itself. Although dilute solutions of this reagent
could be prepared and kept for short periods of time at low
temperature, high concentrations of triphenylantimony sulfide
in carbon disulfide at room temperature rapidly led to the
precipitation of sulfur. The reaction was observed in a number
(16) Williams, C. R. Ph.D. Thesis, McGill University, 1991.

2644 Inorganic Chemistry, Vol. 36, No. 12, 1997
Jason
of common NMR solvents (CS2, CDCl3, acetone-d6, and CD3-
CN) making it unlikely the decomposition requires the partici-
pation of the solvent.
Kinetic analyses of the reduction of triphenylantimony sulfide
in carbon disulfide at 35 °C were performed by 1H NMR
spectroscopy. The second-order rate constant was 0.014 (
0.002 s-1 M-1. Some of the reactions displayed precise second-
order kinetic behavior while others were better fit by the addition
of a companion first-order process.
The self-reaction of triphenylarsenic sulfide, 19, was signifi-
cantly slower than that of the antimony sulfide. The dispro-
portionation of 19 does not to go to completion. At 25 °C,
approximately 17% of the sulfide is reduced to triphenylarsenic,
20, as measured by 1H NMR spectroscopy. A kinetic analysis
was not performed.
Discussion
Figure 3. Examples of reaction paths for the addition of sulfur to
R-P4S5 and R-P4S6.
The sulfur transfer reactions of triphenylarsenic sulfide and
triphenylantimony sulfide with the phosphorus sulfides have
provided a much needed tool for the manipulation of these
unique chemical structures. The first question evaluated was
the potential of these two materials as synthetic reagents.
Because of the limitations experienced with the use of 21
mentioned above, all the truly synthetic experiments were carried
out with 19. The easy syntheses of R-P4S5 and R-P4S7, as well
as the importance of assigning the structures of the products
expected from the single addition of sulfur to them, led to the
use of these two sulfides as examples for judging synthetic
utility.
From small-scale reactions it was clear the predominant
product from the oxidation of R-P4S5 with 19 was R-P4S6,
making it a reasonable target for isolation and purification. After
removal of all of the triphenylarsenic species from the mixture,
the crude mixture had the following composition: R-P4S5, 4%;
R-P4S6, 71%; γ-P4S6, 16%; R-P4S7, 7%; â-P4S8, 1%; R-P4S9,
1%. A series of precipitations was used to remove all of the
residual R-P4S5 and the two highest sulfides, â-P4S8 and R-P4S9.
The analysis of the material at the end of three stages of
purification was: R-P4S6, 84%; γ-P4S6, 10%; R-P4S7, 6%. The
similar solubilities of these three prevented continued recrys-
tallization from improving the purity of R-P4S6.
The stoichiometric oxidation of R-P4S7 with 19 gives one
major product, â-P4S8, contaminated with unreacted starting
material and small amounts of higher sulfides as shown in
reaction 12 of Table 1. Attempts to purify the â-P4S8 by the
precipitation of R-P4S7 led to the disproportionation of â-P4S8
with the formation of R-P4S7, both P4S9 isomers, P4S10, and
two of the unknown sulfides, compounds F and H. This
behavior is consistent with the thermal sensitivity of â-P4S8
discussed in the preceding article.1
Figure 4. The four C1 isomers of P4S6.
phosphorus, are observed in the reactions of 19 and 21 with
the various phosphorus sulfides. The reaction of R-P4S5 with
less than 1 equiv of 19 (experiment 3, Table 1) gives rise to
two major products, R-P4S6 and γ-P4S6. In no reaction of
R-P4S5 with 19, regardless of stoichiometry or reaction condi-
tions, is â-P4S6 observed. Since the structures of R-P4S5 and
â-P4S6 are known from X-ray diffraction studies, path c of
Figure 3 plays no part in the reaction of R-P4S5 with 19.
The work of Bjorholm has shown the mother-daughter
relationship between R-P4S5 and R-P4S6, although the exact
placement of the extra sulfur atom in R-P4S6 was left in doubt.
The symmetry of R-P4S6 is known to be C1. Among all of the
structural isomers of P4S6 built upon the expansion of a P4
tetrahedron, there are four with C1 symmetry (Figure 4).
Structures 8 and 26 are the two structures already proposed for
R-P4S6;6,7 structures 27 and 28 are far less likely. These last
two structures utilize a phosphorus-sulfur cage that is not seen
in any other known phosphorus sulfide. In addition, all three
In order for these reactions to be useful for the assignment
of structure, the first formed products must be predictable from
the structure of the starting phosphorus sulfide. There are a
number of potential reactions that are easily drawn for these
additions. The most likely are shown in Figure 3. Reactions
a and b simply transfer a sulfur from arsenic to phosphorus,
though the paths are quite different. Either reaction running in
reverse could represent the reduction of phosphorus sulfides by
triphenylphosphine, and it would be reasonable to expect that
oxidation and reduction reactions are mechanistically connected.
Unfortunately, there has been neither experimental nor theoreti-
cal evaluation of the mechanism of either the oxidation or the
reduction reaction.
1
of the JPP values that one would measure in 27 and 28 would
be large (>150 Hz), rather than only two.1 Finally, one can
use a plausibility argument to reinforce the distinction. The
reaction of R-P4S5 with 19 to give either 27 or 28 would require
the formation of a new PP bond. A sulfur atom donor is not
expected to cause the reduction of a phosphorus-sulfur cage
under the mild conditions of these reactions (carbon disulfide
and 0-25 °C). Therefore, structures 8 and 26 are the only two
that need be considered for R-P4S6, and the conclusion one must
draw is that the reaction of R-P4S5 to give R-P4S6 is the oxidation
of a trivalent phosphorus.
Both types of reactions shown in Figure 3, oxidation of a
trivalent phosphorus and ring expansion by attack at a tetravalent

Less-Common P4Sn Species
Inorganic Chemistry, Vol. 36, No. 12, 1997 2645
R-P4S6 is structure 8 and not 26, and the major source of the
R-P4S7 in these reactions involves the oxidation of R-P4S6 and
not γ-P4S6. The oxidation of γ-P4S6 by the same mechanism
should lead to â-P4S7.
The oxidation of â-P4S5 provides a little more insight into
the dominant mechanisms and the stability of the phosphorus
sulfides. The major product from the reaction of â-P4S5 with
10
19 is â-P4S6, structure 9. The structures of both â-P4S5 and
â-P4S67 have been determined by X-ray crystallography and are
not in question. The major product of the second oxidation is,
not surprisingly, R-P4S7, the result of the oxidation of the other
PIII center in â-P4S5. â-P4S7 is produced by the oxidation of
1
1
â-P4S6 at a rate between /8 and /4 that for the formation of
R-P4S7 (statistically corrected), indicating the rate of oxidation
in this sulfide observes the order PIII > PII. The structure for
â-P4S7 given by Blachnik is consistent with this chemistry as
well as the 31P NMR spectroscopy expected of structure 12.
The dominant reaction of R-P4S7 with 19 under stoichiometric
conditions is formation of â-P4S8, structure 14. The structure
assigned by Blachnik to this compound is consistent with the
oxidation of one of the PII phosphorus centers in R-P4S7. This
chemistry provides strong support for both the stoichiometry
and the structure of â-P4S8.
There are two minor components of the reaction of R-P4S7
with 19 that are difficult to explain. The first is a material that
has been labeled γ-P4S8. In carbon disulfide it has the following
parameters for an A2BC coupling system: P1,2, 95.4 ppm; P3,
57.2 ppm; P3, 44.2 ppm; coupling constants J1-2,3 ) 20.1 Hz,
J1-2,4 ) 30.9 Hz, and J3,4 ) 156.0 Hz. The structure of this
compound was assigned to the isomer of P4S8 shown as
compound 15. There are only two structures with stoichiom-
etries between P4S8 and P4S10 that have the symmetry dictated
by the NMR spectrum and no S-S bonds: structures 14 and
15. The assignment of 14 to â-P4S8, derived from this and
Blachnik’s7 effort, is reasonably secure, and that of 15 to γ-P4S8,
more tenuous.
The observation of γ-P4S8 in this reaction mixture poses a
problem: There is no straightforward way to get from R-P4S7
to γ-P4S8 without either sulfide exchange or a series of
rearrangements. The fact that this material appears as a minor
component of the reactions of phosphorus and sulfur detailed
in the preceding article gives support to the notion that this NMR
pattern corresponds to a binary phosphorus sulfide of at least
moderate stability and longevity.
The second minor component has been labeled compound E
and, by virtue of its coupling pattern, has four phosphorus atoms
in the structure as two chemically equivalent pairs (or a multiple
of this arrangement). The only structures with stoichiometries
between P4S8 and P4S10 that have the necessary symmetry have
already been ascribed to R-P4S8 (13) and â-P4S9 (17). Com-
pound E is also a transient species, being observed within the
first several hours of the reaction and being absent from a
spectrum taken at 24 h. Because there is no reaction between
R-P4S7 and triphenylarsenic, the stoichiometry of compound E
should be P4S7 or higher. Compound E, and other unassigned
species in these mixtures, may be cage disulfides.
The reactions of R-P4S9 with 19 were observed to be either
simple or complex, depending on the manner in which they
were performed. If the R-P4S9 is completely dissolved in carbon
disulfide prior to the addition of a carbon disulfide solution of
19, the only materials observed are R-P4S9, P4S10, R-P4S7 (the
sample of R-P4S9 contained 5-10% R-P4S7), and â-P4S8, made
from the residual R-P4S7. The relative quantity of the sum of
the R-P4S7 and â-P4S8 concentrations remained roughly constant
through the course of the reaction and well afterward, indicating
no reduction of R-P4S9 by the triphenylarsenic formed in the
Figure 5. The seven Cs isomers of P4S6.
The chemistry of â-P4S5, as discussed below, gives further
weight to the assignment of structure 8 to R-P4S6. In both of
the first two reactions of â-P4S5 the faster (or only) oxidation
is on the phosphorus of higher formal oxidation state. Thus,
only the PIII phosphorus atoms are oxidized in â-P4S5, and the
PIII site is more rapidly oxidized than the PII sites in â-P4S6.
This ordering would be upheld for R-P4S5 only if R-P4S6 has
structure 8. The reaction of R-P4S5 to give 26 requires the
oxidation of a formally PI center.
What about γ-P4S6? Studies involving both short reaction
times and substoichiometric oxidations of R-P4S5 argue that
γ-P4S6 is a first-formed product of the oxidation of R-P4S5. There
are seven isomers of P4S6, shown in Figure 5, that have the
necessary symmetry to give rise to the observed AB2C 31P NMR
spectrum: P1, 169.5 ppm; P2,3, 103.1 ppm; P4, 58.4 ppm;
coupling constants J1,2-3 ) 38.7 Hz, J1,4 ) 46.7 Hz, J2-3,4
)
298.8 Hz. It is difficult to draw straightforward conversions
that transform 5 into any of the products 30-33. Structure 9
is already known to be â-P4S6. It is also well-known that the
substructure PsPdS displays large coupling constants, typically
larger than 200 Hz.1,6 Of the remaining isomers, 10 and 29,
only 10 is consistent with the 31P NMR spectrum. Therefore,
the transformation of R-P4S5 into γ-P4S6 involves only the
oxidation of a PII phosphorus atom to give structure 10.
There are three possible products from the oxidation of a
trivalent phosphorus in R-P4S5. The third member of the group,
structure 26, is not found in this mixture of products. Unlike
R-P4S6, extensive rearrangements would be required to transform
26 into R-P4S7, making it unlikely that 26 is formed competi-
tively with 8 and 10. If it were produced in a quantity similar
to the observed isomers of P4S6, it would have to react rapidly
either by decomposition or with another 1 mol of triphenyl-
arsenic sulfide.
The last common product of the reaction of R-P4S5 with 19
is R-P4S7. This reaction must involve a P-P bond cleavage
step. The oxidation of R-P4S6 (8) as shown in reaction c of
Figure 3 is one possible mechanism for the insertion of a sulfur
into the cage with formation of R-P4S7 (11). This reaction is
an association-rearrangement pair that leaves the attacked
phosphorus tetravalent and simply places the sulfur from 19
between a pair of phosphorus atoms. The mechanism drawn
is analogous to the Baeyer-Villiger reaction of organic
chemistry. If the ring expansions of the phosphorus sulfide
cages involve simple rather than multistep reactions, the new
sulfur must be placed adjacent to an existing tetravalent
phosphorus. Thus, structure 8 should readily lead to R-P4S7. It
is also the only structure among the group 8, 10, and 26 that
has a bond between two trivalent phosphorus atoms, a necessary
feature of 11. Both 10 and 26 would require either reduction
of a phosphorus or PP bond formation during the oxidative
transformation. Assuming the above discussion is correct, then

2646 Inorganic Chemistry, Vol. 36, No. 12, 1997
Jason
Figure 7. Disproportionation of 21 and the products of the reaction
of S2 with 2,3-dimethylbutadiene.
The simplest explanation for the second-order decay of 21 is
shown in Figure 7. Two molecules of triphenylantimony sulfide
are simultaneously reduced with the formation of a molecule
of S2. The reaction could occur in one step as drawn or the
intermediate formation of triphenylantimony. The rate at which
the disulfide decomposes to S2 or some other species would
have to be fast relative to its formation because it is not observed
by NMR spectroscopy. Disulfur, S2, is a transient species
isoelectronic with singlet oxygen and has been the object of
some interest for the preparation of organic disulfides.16-19 The
products of the reaction of S2 with 2,3-dimethylbutadiene have
served as markers for its formation.16 At high concentration in
a mixture of CS2 and the butadiene, the expected disulfide and
tetrasulfide products, 36 and 37, were formed. They were
Figure 6. Possible intermediates in the reaction of P4S3 with 19. See
text for details.
reaction. If the R-P4S9 was present as a solid during the reaction,
several other species were formed, not as spurious products but
rather as the consistently observed outcome of reactions run
with solid sulfides present. These experiments were not
included in Table 1.
The reactions of P4S10 with 19 were decidedly dull. The small
amount of R-P4S9 present in the sample was converted to P4S10;
no other reactions were apparent. In this system, it did not
matter whether the phosphorus sulfide was completely dissolved
or not. None of the unusual products mentioned above were
observed, indicating that the unusual species from the oxidations
of solid R-P4S9 are not the result of oxidation of P4S10 either as
a solid or in solution.
1
identified by their H and 13C NMR spectra. There are other
possible explanations for the second-order kinetics and the
formation of 36 and 37, but the formation of S2 is consistent
with both the kinetics and the formation of products believed
to arise from the chemistry of S2.
The reaction of P4S3 with 19 bears little resemblance to the
calm and seemingly rational chemistry displayed by the higher
sulfides. The first difference is the rate of reaction. P4S3 reacts
so slowly that within the first 30 min at room temperature no
reaction is visible by 31P NMR spectroscopy. By contrast, all
of the other sulfides show extensive or complete reaction within
this same time period. Second, there are no products of
intermediate sulfur content. The lowest sulfide observed is
R-P4S7. Though there are countless paths that link P4S3 with
R-P4S7, the formation of either 34 or 35, as shown in Figure 6,
is consistent with the chemistry of the higher sulfides. Only
one more oxidation by 19 is required to transform 34 into
R-P4S5. Similarly, 35 can be converted to R-P4S6 by two
successive oxidations at the same phosphorus. The very slow
reaction between P4S3 and 19 dictates that R-P4S5 and R-P4S6
should not be observed in these oxidations; these intermediate
sulfides would be produced in the presence of a large excess of
19, leading to continued and extensive oxidation. There are
observed in these oxidations only low concentrations of R-P4S7,
â-P4S8, R-P4S9, and P4S10, as shown in Table 1.
The reactions of two phosphorus sulfides with triphenylan-
timony sulfide (21) were also investigated, but the results were
not as straightforward as those obtained with triphenylarsenic
sulfide. The principle reason for the difficulty derives from
the self-reaction of 21. In all solvents studied, 21 undergoes a
second-order disproportionation reaction to form triphenylan-
timony (22) and sulfur. The sulfur product of this dispropor-
tionation is expected to be an oxidizing species in its own right.
Since the effective sulfur donor changes from triphenylantimony
sulfide to a form of sulfur during the reaction, a great deal of
mechanistic insight is lost.
Conclusions. The triphenylarsenic and triphenylantimony
sulfides have been shown to rapidly oxidize most of the known
binary phosphorus sulfides. Oxidation of a cage phosphorus
by the attachment of an exocyclic sulfur is the predominant
mode of reactivity. The relative rates of these oxidations
observe the order PIII > PII > PI. Oxidation of a PP bond to
make a P-S-P group is slower but well within the capabilities
of the reagent. Mechanisms for these oxidations are easily
drawn from the atom transfer and addition-elimination mech-
anisms found in classical inorganic chemistry. These oxidations
were used to prepare several phosphorus sulfides (R-P4S6 and
â-P4S8) and to aid in the assignment of structures to two others
(γ-P4S6 and γ-P4S8). The unusual disproportionation reaction
of triphenylantimony sulfide was observed and a case made for
the formation of disulfur, S2.
Acknowledgment. I thank Claude Jones for his assistance
with the NMR spectroscopy. I also need to thank Ed Griffith
and Frank Pacholec for allowing me the time to pursue this
remarkable chemistry. Finally, I thank Darryl Fee for teaching
me the “P2S5” business.
Supporting Information Available: Listings of 31P NMR data,
representative NMR spectra, and plots of kinetic data (13 pages).
Ordering information is given on any current masthead page.
IC9614881
(17) Steliou, K.; Salama, P.; Yu, X. J. Am. Chem. Soc. 1992, 114, 1456-
1462.
(18) Tardif, S. L.; Williams, C. R.; Harpp, D. N. J. Am. Chem. Soc. 1995,
117, 9067-9068.
(19) Bosser, G.; Paris, J. New J. Chem. 1995, 19, 391-399.