Znd. Eng. Chem. Res. 1991,30, 1906-1911
1906
Model and UNIFAC. Ind. Eng. Chem. Process Des. Deu. 1984, 23, 194-209.
Thomas, E. R.; N e w " , B. A,; Long, T. C.; Wood, D. A.; Eckert, C. A. Limiting Activity Coefficients of Nonpolar and Polar Solutes in Both Volatile and Nonvolatile Solvents. J . Chem. Eng. Data 1982,27, 399-405. Trampe, D. M.; Eckert, C. A. Limiting Activity Coefficienb from an Improved Differential Boiling Point Technique. J . Chem. Eng. Data 1990,35, 156-162. Vernier, P.; Raimbault, C.; Renon, H. Proprietes Thermodynamique des Solutins Infinitiment Diluees D'Hydrocarbares dans les Solvants Polaires. J . Chim. Phys. Physicochim. Biol. 1969, 66, 429-436. Weidlich, U.; Gmehling, J. Extension of UNIFAC by Headspace Gas Chromatography. J . Chem. Eng. Data 1985, 30, 95-101.
Weidlich, U.; Berg, J.; Gmehling, J. Extension of UNIFAC by Headspace Gas Chromatography. 2. J. Chem. Eng. Data 1986, 31, 313-317.
Wilhoit, R. C.; Zwolinski, B. J. Physical and Chemical Properties of Aliphatic Alcohols. J . Phys. Chem. Ref. Data 1973, 2, Suppl. 1. Yang, Y.; Xiao, S.; Li, H.; Fu, Y. Measurement of Infinite Dilution Activity Coefficients by Nonsteady State Gas Chromatography. Chengdu Keji Dame Xuebao ( J . Chengdu Uniu. Sci. Technol.) 1988, No. 1, 35. Yin, C.; Hassett, J. P. Gas Partitioning Approach for Laboratory and Field Studies of Mirex Fugacity in Water. Enuiron. Sci. Technol. 1986,20, 1213-1217.
Received f o r review August 29, 1990 Accepted February 15, 1991
Reactivity of Technical Phosphorus Pentasulfide Michel C. DBmarcq Institut National des Sciences Appliquges, Laboratoire de Chimie Organique, 20 Avenue Albert Einstein, F69621 Villeurbanne, France
The reactivity of technical P4S10 with alcohols or water, in either heterogeneous or homogeneous processes, is a direct function of its solidification rate, which determines to what extent the dissociated state prevailing in the molten sulfide is retained in the solid. 31PNMR analysis reveals that quenched melts of P4S10 contain more redox products P4S9,P4S7,and phosphorus polysulfides P,S, ( y / m > 2.5) but seemingly less elemental sulfur than slowly cooled (unquenched) or annealed specimens. It has long been recognized that the reactivity of technical phosphorus pentasulfide with alcohols or water grows with the rate of its solidification(Robota, 1959; Moedritzer and Van Wazer, 1963; Cremer, 1965; Vincent, 1969; Cueilleron and Vincent, 1970; Childs, 1971) and is depleted by annealing (Robota, 1959; Roth and Taylor, 1960; Knapsack, 1965; Vincent, 1969; Niermann et al., 1979). Quenched grades of P4S10 are commonly produced by cooling the molten sulfide from 310-350 "C to room temperature on a revolving drum, so as to obtain flakes with a thickness under 1 mm, which may ultimately be ground to powder; chilling devices giving thicker flakes (e.g., 3-5 mm) afford much less reactive grades, hereafter referred to as "unquenched". Moedritzer and Van Wazer (1963) have shown that the reaction rate of a number of phosphorus sulfides-including P4S10, either pure recrystallized or quenched from the melt-with n-butanol is proportional to their surface area; patent data confirm these observations (Knapsack, 1965). The fact that the heterogeneous alcoholysis of P4S10 is sensitive to soluble nucleophilic catalysts (DBmarcq, 1972; see also supplementary material (see the paragraph at the end of the paper)) suggests that the above kinetics is not simply determined by the slow dissolution rate of the unreacted solid but is largely controlled by surface attack by the alcohol (or by the catalyst, affording soluble reactive intermediates). An alternative mechanism (Bencze, 19701, involving rapid dissolution of the phosphorus sulfide, followed by rate-limiting homogeneous alcoholysis, is not supported by our own experiments (unpublished). The enhanced reactivity of quenched P4S10 has been ascribed to its containing an amorphous phase (Cremer, 1965; Forthmann and Schneider, 1966; Childs, 1971). However, given the chemical inhomogeneity of technical P4Sl,-presence of redox dissociation products, especially P4S9(Dgmarcq, 19811-the question arose as to what ex-
tent its reactivity was dependent on its physical state (crystallinity, particle size, friability) or on its chemical composition. This prompted us to investigate the kinetics of the homogeneous alcoholysis and hydrolysis of this sulfide in relation to its NMR analysis. Experimental Section Materials. Technical phosphorus pentasulfide was obtained from Atochem and other commercial sources. Pure P4S10 and P,S9 were prepared as previously described (DBmarcq, 1990a). Pure P4S, was a commercial product, further recrystallized from carbon disulfide. CS2was purified and dried as in prior work (DBmarcq, 1987). Procedures. The homogeneous alcoholysis of phosphorus sulfides was performed in carbon disulfide, using Bu"0H as the alcohol; the concentration of the latter was monitored by measuring the IR absorptivity of the free OH vibration at 3640 cm-' (accuracy fca. 3%). The results are shown in Figure 1. For studying the homogeneous hydrolysis we used a two-phase procedure previously described (DBmarcq, 1990a), in which a CS2 solution of phosphorus sulfide is stirred with water and the extent of reaction determined by acidimetry of (water-soluble) hydrolysis products. Taking the solubility of phosphorus sulfides in water as negligible and the solubility of water in the organic phase as practically constant and equal to that in neat CS2-i.e., 142 ppm at 25 "C-the process may be regarded as pseudo-one-phase and pseudo-first-order. The accuracy is about &0.5%. The results are shown in Figure 2. 31PNMR spectra were obtained at 121.5 MHz (Bruker AC 300 spectrometer) or at 80.76 MHz (Jeol FX 200) by using 2.5-3 g/L solutions (no insoluble residue) of phosphorus pentasulfide in carbon disulfide. The pulse angle was generally set to 30 or 70°, with the relaxation delay varying from 2 to 20 s and the acquisition time from 0.4 to 3.3 s. Chemical shifts are referenced to external 85%
0888-5885191/ 2630- 1906$02.50/0 0 1991 American Chemical Society
Ind. Eng. Chem. Res., Vol. 30,No. 8, 1991 1907 NMR Composition (Percent of Total P ) of Table I. Quenched and Unquenched Commercial P,S,, (CS2 Solutions)o quenched unquenched 38-43 58-70 pISl0 36-44 28-36 p4sB PIS1 2-3 0 . 5 phosphorus polysulfides 12-20 0.5-2.5
P
p 4 s 9
f
"Small departure from the exact S/P stoichiometry, as occasionally encountered in commercial phosphorus pentasulfide, may, of course, affect the composition.
H3P04. No significant differences were observed between 'H decoupled or nondecoupled spectra. In some cases (especially for unquenched specimens of phosphorus pentasulfide), a small weighed amount of triphenylphosphine sulfide was added as a standard of phosphorus concentration to help integrate small signals. IR spectra were recorded on a Perkin-Elmer 577 apparatus and DTA diagrams on a Du Pont 900 thermal analyzer. Fusion enthalpies were measured with a Triflux microcalorimeter of the Thermoanalyse Co. Surface areas were measured by the BET method. Laboratory grinding was carried out in a Dangoumau (Prolabo) ball mill.
A
Figure 1. Kinetics of the homogeneous reaction'of various phosphorus sulfides (3.6 mequiv of P L-I) with n-butanol(7.65 mmol L-l) in carbon disulfide at 22 "C.A, unquenched commercial P4S10(PI& content, 44% PIP);B, quenched commercial P4S10(PISgcontent, 43% PIP).
/"/ ,
ps3 6
G A 4
0
0
unquenched
P4Slo
tech
0
F
0
CA
I
I-
p4s10
A -
0
I
30
,
m 1 n u tes
60
Figure 2. Kinetics of the homogeneous hydrolysis of various phosphorus sulfides (22.52 mequiv of P L-I) in water-saturated carbon disulfide at 25 OC. The extent of hydrolysis is calculated on the basis of 1 equiv of "hydrolyzed P" per equivalent of strong acidity. Labels C, F, CA,and G refer to quenched, unquenched, and quenched/annealed technical P4Slo and to the soluble phosphorus polysulfide PSs.Bdescribed in Table 11.
Rssults The following conclusions emerge clearly from Figures 1 and 2: (i) As noted in previous reports (Meisel and Grunze, 1970; Neels et al., 1974; Barieux and DBmarcq, 1984; Bourdauducq and DBmarcq, 1987),P4Sgis extremely reactive with water or primary alcohols. (ii) Pure P4Slo is relatively inert under similar conditions. (iii) P4S7is moderately reactive. (iv) Technical phosphorus pentasulfide is much more reactive than pure P4Sl0-a consequence, in part, of its containing P4Sg. (v) Quenched grades of technical P4S10 are more reactive than unquenched ones, even at similar P4S9contents. (vi) Annealing considerably reduces the reactivity of quenched P4S10. These findings indisputably establish that the enhanced reactivity of quenched P4Slo is basically related to its chemical composition and not-or only incidentally-to its physical state. I t also appears that the P4S9content is a determining factor of the reactivity, though not the only one. That quenched grades of P4S10 differ in chemical composition from the unquenched (or annealed) product was confirmed by 31PNMR spectroscopy. As a general rule, the former contain less P4S10, some more P4Sgand P4S7 and much more phosphorus polysulfides (PPS) than their unquenched counterparts (Table I; Figures 3 and 4). Phosphorus polysulfides P,S ( y / m > 2.5) appear as small or very small NMR signds (up to 30 lines or so), scattered over a 100 ppm range. These comprise singlets at 6 18.2,56.2,and 89.2 ppm and (very weak and not always discernible) at 55.6, 55.75, 56.0,116.1,and 124.9 ppm, multiplets at 6 45.85 (dxt, J = 106.2 and ca. 4.9 Hz), 78.65 (dxdxd, J = 61.0,22.3,and 1.6Hz),and 103.2 (t, J 23.1Hz) ppm, and complex or unresolved massifs at 6 ca. 36.9,54.5, and 93.0ppm (Figure 3). Their assignment to phosphorus polysulfides relies on the fact that, allowance being made for inevitable slight variations of chemical shifts (these can be offset by taking PISloas an internal reference), most of them coincide with the NMR signals of soluble PPS found in P-S glasses with S/P between 2.5 and about 6 (DBmarcq, 1987 and 1990b;Tullius et al., 1990). As previously noted, such PPS are very sensitive to hydrolysis
1908 Ind. Eng. Chem. Res., Vol. 30, No. 8, 1991
93 A4
u'
14 '1 '0 I
''$'
116 0 V
---
!4S7
Y Scale
--_
78 7
36 6
'7
V
182
ppm
V
XlO
41 2
P Pm
A Ph3 PS
Figure 4. 31PNMR spectrum (80.76 MHz; 41 752 scans; pulse angle 30";pulse delay 2 s; Hz/Pt 0.732) of a specimen of unquenched commercial P4Sl0(solution in CSp). Phosphorus polysulfides are labeled with open triangles. The broad peak at 41.2 ppm corresponds to Ph,PS, added as a standard of phosphorus concentration.
(DBmarcq, 1987). Although no precise structural assignment is yet possible for individual PPS, general primary formulas 1 and 2 have been contemplated (Dgmarcq, S
f
-P=S
S=P-S
LSlJ
n i S,-P-S,
S
l
II
-P
U
n S
k
2
1
1990b); in the present case, 1 appears to be the more compatible with the estimated mean stoichiometry of PPS encountered in quenched P4S10, namely, S / P < ca. 6.7, as calculated from data in Table 11. In this connection, the detection by mass spectroscopy of species P2S6(the most
abundant) to PzS9in quenched melts of P4Sl0(DBmarcq, 1990a) appears significant. Curves in Figure 2, with the exception of P4S7,become quasihorizontal after 20-30 min, at which time practically all of polysulfides and P4S9has been hydrolyzed, along with 75-85'70 of P4S7(at any rate, a minor component), while most of P4S10 remains unreacted. This allows a convenient overall determination of P-containing redox impurities in technical P4S10, which matches up NMR data quite well (Table 11: compare the two right-hand columns). Upon annealing, even briefly, at 200-220 "C,quenched P4Sl0loses a substantial part of its homogeneous reactivity (Figure 2, curves C and CA). This is due to the more or less complete disappearance of PPS and P4S7and, to a
Ind. Eng. Chem. Res., Vol. 30, No.8, 1991 1909
NMR (CS2 Solutions) and Selective
Table 11. Composition of Phosphorus Sulfides (Percent of Total P ) As Determined by Hydrolysis
NMR analysis (P4S10 omitted) C CA D DA
E F
C
DhosDhorus sulfide . . quenched technical P4Slo flakes 0.8 mm thick, P, 27.4% w/w C, annealed at 210-220 "C for 5 mind quenched technical P4S10 flakes 0.6 mm thick, P, 27.7% w/w D, annealed at 200-217 "C for 2 mind quenched melt of pure P4S10: P, 27.9% w/w unquenched technical PISlo tablets 3-4 mm thick, P, 27.75% w/w soluble PS3.Q
P,S, -., 38 28 43 37 =46 35 21
PAS, 3 0 3 =l =2 =0.1 0 1
PPSa 119 22 214 =4 =13 -0.5 61
.
totd
160
230 260 =42 =61 -35.6 82
selective hydrolysis reactive sDecieab 62.25 32.8' 61.0 41.2 35.5' 82.3'
OGiven the number of PPS signals, their smallness and the fact that (at least) four of them fall in the B massif of P4Sp,their overall integral may be slightly underestimated. *Determinedfrom the extent of hydrolysis after 20-30 min (Figure 2). 'See Figure 2. dAnnealed in test tubes under dry nitrogen. *A glass capillary (inner diameter ca. 0.8 mm) was filled with pure P4S10, then sealed under reduced pressure, immersed for 5 min in a bath at 320 O C , and dropped in cold CHzCIP. ICs, extract of vitreous PS3.5.
and PPS contents. These findings are consistent with previous data on the relative rates (hereafter in parentheses) of surface reaction per unit area of phosphorus saldes, either with liquid n-butanol (Moedritzer and Van Wazer, 1963): pure
(1)< P4S7(3)
