Langmuir 1995,11, 2534-2538
2534
A n Infrared and Solid-state 31PN M R Study of the Adsorption of P(CHs)C12 and P(CH3)ClflC13 Mixtures on Silica S. J. Lang,? I. D. Gay,*,’ and B. A. Morrow*>+ Departments of Chemistry, University of Ottawa, Ottawa, Ontario, Canada K I N 6N5, and Simon Fraser University, Burnaby, British Columbia, Canada V5A 1S6 Received December 15, 1994@ The adsorption of PMeClz and of mixtures of PMeClz and Pcl3 on silica have been studied using infrared spectroscopy and 31P solid-state NMR. PMeClz gives a single phosphorus containing product (A) as a result of a reaction with surface SiOH groups, SiOP=O(H,Me); HC1 and surface Sic1 are byproducts of the reaction. A reaction mechanism has been devised. The rate of reaction is slow and high pressures of reactant (10 to 80 Torr)are required. Commercially available PMeClz samples are usually contaminated with PC13 at the 2 t o 5 mol % level. At the pressures needed to effect chemisorption of PMeC12, the contaminant PC13 has a significant influence on the course of the reaction. The major feature with such low levels of PCl3 as an impurity is the conversion of A to a second species B,(SiO)ZP=O(H), with minor amounts of SiOP=O(H,OH). A plausible sequence of reaction steps leading to this product has been elaborated, but the chemistry is complex.
Introduction
Experimental Section
The surface chemistry of organophosphorus compounds is of interest in several important areas.l For example, the catalytic and stoichiometric decomposition of toxic phosphorus compounds has been studied extensively, using techniques ranging from those of classic catalytic studies to those of surface In addition, surface phosphorus species are occasionally important components of industrial hydrotreating catalyst^.^ Several strategies for the immobilization of transition metal phosphine complex catalysts on oxide supports have been investigated.5 In the latter two examples, solid-state 31PNMR has been particularly valuable in elucidating the state and role of the surface phosphorus species. In t h e present work, solid-state NMR and FTIR have been used to investigate the reaction of PMeC12 with a pyrogenic silica. It is shown that, while the reaction of PMeCl2 with silica is relatively simple, PC13 (present in -5% amounts in commercial samples of PMeC12) reacts with the chemisorption products of PMeC12. This increases the apparent complexity of the reaction, and the combination of two spectroscopic techniques was essential to understanding the chemistry.
The silica used was a pyrogenic (fumed) material, Cab-0-Si1 HS5, having a surface area of 325 mVg. Its surface properties have been characterized previously by IR6,7and NMR.s The samples were pressed into 25 mm diameter disks containing 10 m&m2 unless otherwise specified. For IR studies the disk was placedin a previously describedgIR cell having an internal volume of 300 mL. This was connected t o a conventional vacuum system capable of obtaining a base pressure of about 10-6Torr, and the samples were generally activated under vacuum at 450 “C for 1h (this temperature will be assumed unless otherwise stated). For NMR studies of PMeC12/SiOz, a 5 mm 0.d. NMR tube was sealed to a pyrex vessel having a volume of 110 mL; pieces of the pressed disks about 1to 3 mm in size were inserted into the large volume reactor where the thermal activation and chemical reactions were carried out; finally the reactor was detached from the vacuum line and tilted so as to slide the disk fragments into the small volume NMR tube. The NMR tube was then sealed and detached from the reactor with a glass blowing torch. Chemically doped NMR samples were activated directly in the NMR tube. Spectrometers. NMR spectra were recorded using a homebuilt instrument, either at 1.4T producing a resonance frequency of 24.3 MHz for 31P,or at 3.5 T160.5 MHz (see caption to Figure 1). Proton decouplingfields of 45-60 kHz were used, with magic angle spinning (MAS) ofthe sealed samples at 2000-3000 Hz.l0 IR spectra were recorded using a BomemMichelsonMBlOO FTIR instrument at a resolution of 4 cm-l. All reactions (unless otherwise stated) and spectral measurements were done at room temperature, 22 f 1 “C. Chemicals. For the initial work, two PMeCl2 samples were purchased (6 months apart) from an established commercial supplier. Although both samples were apparently from different “lots”,a 31PNMR analysis of the neat liquids revealed that both samples contained about 5% Pc13 as the only P-containing impurity. In the text we will nonetheless refer to this as PMeCl2 where it is understood, unless explicitlystated, that this material was contaminated by about 5% Pc13. As will be detailed below, the presence of PC13even as a minor contaminant, proved to have a significant effect on the course of the reaction. Therefore, we synthesized two samples of PMeClz which were free of PC13. The first of these involved the synthesis of 13CPMeCl2 (99%labeled)frompure Pc13 and 13CH31according
University of Ottawa. Simon Fraser University. Abstract published in Advance ACS Abstracts, June 1, 1995. (1)(a) Ekerdt, J. G.; Klabunde, K. J.; Shapley, J. R.; White, J. M.; Yates, J. T., Jr. J . Phys. Chem. 1988,92,6182. (b) Ramsier, R. D.; Henriksen, P. N.; Gent, A. N. Surf. Sci. 1988,203,72. (c) Templeton, M. K.; Weinberg, W. H. J . Am. Chem. SOC.1986,107,774. +
@
(2)(a)Henderson, M.A.; Jin, T.; White, J. M. J . Phys. Chem. 1986, 90,4607. (b) Templeton, M.K.; Weinberg, W. H. J.Am. Chem. SOC. 1986,107,97.(c) Lee, K. Y.; Houalla, M.; Hercules, D. M.; Hall, W. K. J . Cutul. 1994,145,223. (d) Tzou, T. 2.; Weller, S. W. J . Catul. 1994, 146,370. (3)(a)Rao, L.-F.;Yates,J. T.,Jr. J . Phys. Chem. 1993,97,5341.(b) Zhang, X.;Linsebigler, A.; Heiz, U.;Yates, J. T., Jr. J . Phys. Chem. 1993,97,5074. (c) Paul, D.K.; Rao, L.-F.; Yates, J. T., Jr. J . Phys. Chem. 1992,96,3446. (4)(a) Decanio, E. C . ; Edwards, J. C.; Scalzo, T. R.; Storm, D. A.; Bruno, J. W. J . Cutul. 1991,132,498.(b)Han,0. H.; Lin, C. Y.; Sustache, N.; Mcmillan, M.; Carruthers, J. D.; Zilm, K. W.; Haller, G . L. Appl. Cutul. A 1993,98,195. (5) (a)Shinoda, S.; Nakamura, K.; Saito, Y. Chem. Lett. 1983,1449. (b) Choudry, B. M.; Kumar, K. R.; Kantam, M. L. J.Cutal. 1991,130, 41. (c) Blumel, J. Znorg.Chem. 1994,33,5050.
(6)Morrow, B.A.;McFarlan, A. J. J . Phys. Chem. 1992,96,1395. (7)Morrow, B. A.;McFarlan, A. J. Lungmuir 1991,7, 1695. (8)Morrow, B. A.; Gay, I. D. J . Phys. Chem. 1988,92,5569. (9)Morrow, B.A.; Ramamurthy, P. J . Phys. Chem. 1973,77,3052. (10)Gay, I. D. J.Mugn. Reson. 1984,58,413.
0743-746319512411-2534$09.00/0 0 1995 American Chemical Society
Langmuir, Vol. 11, No. 7, 1995 2535
Adsorption of P(CH3)C12 on Silica
A
I 3800
3600
3400 cm-l 3200
3000
3800
3600
3400 c['
3000
n 100
50
0
-50
-100
PPm
Figure 1. 31PCPMAS-NMR spectra. Spinning rates 2 t o 2.5 kHz. A to C were recorded at 24.3 MHz, D and E at 60.5 MHz. The spinning sidebands that would have been present in D and E have been suppressed by a TOSS sequence: (A) 70 mgof Si02 activated at 175 "C, allowed to react with 50 Torr of PMeCl2 for 45 min, evacuated for 20 min; (B)100 mg of Si02 doped to about 500 pmol/g with MeP=O(H,OH) from HCCls solution, evacuated 3.5 h; (C) 70 mg of Si02 activated at 175 "C, allowed to react with 50 Torr of PMeCl2 for 19 h, evacuated 20 min; (D) 100 mg of Si02 doped to 310 ,umol/g with aqueous HsP03, evacuated 2 h; (E) 100 mg of Si02 doped t o 310 ,umol/g with aqueous H3P03, evacuated 2 h at 200 "C. to the procedure described by Colquhoun and McFarlane." Sufficient material for two NMR experiments was obtained. In
a second procedure,a C12MeP=S sample having no phosphoruscontaining impurities at greater than 0.1 mol % was purchased from Alfa Inorganics. This was reacted with tri(n-butyl)phosphine according to the procedure of Ulmer et aZ.12to give PMeCl2 which was free of Pc13 (0.1 mol % detectability). Phosphorus acid was Fisher Certified grade and no Pcontaining impurities were detected by 31PNMR. Methylphosphinic acid was prepared by the methanolysis ofPMeCl2 according to the procedure of Fiat et al.I3 In the descriptions below, we will write SiOP=O(X,Y)for the structure X
I
SiOP=O
I
Y
Results The 31P NMR spectrum observed after 45 min of adsorbing PMeC12 on Si02 which had been previously activated at 450 "C showed a broad peak a t 22 f 2 ppm (Figure 1A) whose width, using two different magnetic fields, was constant when expressed in ppm. Therefore, the peak width is probably due to a wide dispersion in isotropic chemical shifts rather than to dipolar coupling with quadrupolar C1 nuclei. If the proton decoupler was (11)Colquhoun, I. J.;McFarlane,W. J.Labeled Compd.Radiopharm. 1977,13, 535. (12) Ulmer, H. E.; Groenweghe, L.C. D.; Maier, L. J.Znorg.Nucl. Chem. 1961,20,82. (13)Fiat,D.;Halmann,M.;Kugel,L.;Reuben, J.J.Chem.Soc.1962, 3837.
3200
Figure 2. Infrared spectra (3800-2800 cm-l) as a function of time following the addition of 73 Torr (A) or 44.6 Torr (B)of PMeCl2 to silica. For each set of spectra, the arrows indicate that there is a continuous decrease in the intensity of the IR bands at 3747 and 3650 cm-l, the first curve (maximum intensity)correspondingto the spectrum observed immediately after addition of PMeC12. The remaining spectra were recorded at the following times after addition of PMeC12: (A) 15 min, 75 min, 2.5 h, 3.5 h, 5 h; (B)30 min, 1 h, 2.5 h, 3.5 h. The intensity at 3250 cm-l increases steadily for the solid line curves, then starts to decrease for the dashed line curves.
turned off for 70 ,us before the start of data acquisition, the signal a t 22 ppm disappeared. This delayed decoupling shows that the species responsible for this peak bore a t least one directly bonded proton. The same NMR spectrum could be generated, Figure lB,by doping Si02 with methylphosphinic acid, MeP=O(H,OH). When PMeClz ws allowed to react with silica for 19 h, a second NMR peak appeared near - 18 ppm which was asymmetric to high frequency (Figure IC). Figure 1D shows the 31Pspectrum of H3P03 doped Si02. Aqueous H3P03 has a chemical shift of 5 ppm; as a solid we found a doublet at 8 and 13.6 ppm indicative of the presence of two nonequivalent sites in the crystal. For a H3P03 doped Si02 sample evacuated a t room temperature, the peak maximum occurred near -8 ppm, with a shoulder a t more negative shift. When this sample was degassed a t 200 "C (Figure lE),the peak maximum occurred at -18 ppm, with a shoulder to positive shift. The second peak which was observed from the PMeCldSi02 reaction appears to be due to a combination of these two species. Delayed decoupling showed that the second group of species obtained from the PMeCl2 reaction (-8 to -18 ppm) had directly bound protons. Delayed decoupling on the H3POdSi02 samples showed that no decomposition of PH had occurred under these mild degassing conditions; all of the P was still protonated. The NMR evidence shows that two chemisorbed or immobile PH containing species were created during the PMeC12 reaction on Si02. For convenience in the following, we will designate the first species A (+22 ppm), and those created for longer reaction times, B (-8 to -18 ppm). The progress of the reaction could be monitored continuously more easily by infrared spectroscopy. Figure 2
Lang et al.
2536 Langmuir, Vol. 11, No. 7, 1995
t
I
1412
E .3
I
z Q m
A
LY
0 (n
m
1306
Q.
2
2500
2450 cm-l
2400
2350
.1
C
2500
2450 c(l
2400
1
2350
Figure 3. IR spectra in the PH stretching region as a function of time for the addition of 73 Torr of PMeClz to silica (correspondinginitially to the spectra shown in Figure 2A).(A) The intensity at 2390 cm-l increases with time, the bottom spectrum being that observed immediately after addition of PMeC12, the remaining spectra being recorded after 15 min, 75 min, 2.5 h, 3.5 h. (B)The arrows indicate the direction of the spectral changes after 3.5 h, 4 h, 5 h, 6 h, 7 h, 8 h, 9 h, 10 h, 12 h, 15 h, and 20 h. shows two examples of the IR spectra recorded in the spectral region between 3800 and 2800 cm-l, during the first several hours (see figure caption for details) of the reaction of PMeClz with 450 "C activated silica. Following addition of PMeClz there was a n immediate decrease in the intensity of the 3747 cm-l peak due to isolated silanols [not shown in Figure 2, but the intensity of the 3747 cm-' peak was about 1.5 absorbance units initially]. This was accompanied by the appearance of a broad band a t 3650 cm-1. Over the next several hours, both the 3747 and 3650 cm-l bands decreased in intensity as a very broad band a t 3250 cm-l grew in intensity and then decreased in intensity. Throughout all of these changes, the spectrum of gaseous HC1 was observed and its intensity increased with the time of reaction (the HC1 bands have been subtracted from the curves in Figure 2). If a lower pressure of PMeClz was added, the decrease in the intensity of the 3747 cm-l was lower initially, but the accompanying spectral changes were qualitatively similar. Accompanying the above spectral changes, a new band a t 2390 cm-l appeared in the spectral region associated with PH, stretching vibrations. This reached its maximum intensity after about 3 to 5 h, depending on the pressure used (generally from 10 to 80 Torr), then slowly started to decrease in intensity with longer reaction times as a second PH band developed a t 2490 cm-l (Figure 3B). The latter band was asymmetric to low wavenumber. Upon reacting PMeClz with a deuterated silica, the 2390 and 2490 cm-' bands appeared a t 1705 and 1805 cm-', a low wavenumber shift by a factor of 1.40 and 1.38,respectively, demonstrating that both bands were indeed due to P W PD stretching modes. Finally, the extent of growth of either PH band did not exactly correlate with the increase or decrease in the intensity of the broad 3250 cm-l feature. Therefore, in agreement with the NMR results, at least two proton containing species were formed following
bo
-1
1350
1300
cm
Figure 4. IR spectra (1450-1360 cm-l) for addition of 44.6 Torr of PMeClz to silica as a function of time. The bands at 1284 and 1412 cm-l are due to the parent PMeC12. The only significant spectral changes correspond to the growth of a new band at 1306 cm-l (as indicated by the arrow; see text), and the spectra are recorded after initial addition of reactant (zero intensity at 1306 cm-') and after 30 min, 1 h, and 3.5 h. adsorption of PMeC12 on SiOz: speciesAfor shorter periods of reaction, and then species B. Evacuation a t room temperature did not remove species A or B from the surface, but the growth of the PH band of species B ceased. In the Discussion it will be shown that species A is SiOP=O(Me,H) and species B is mainly (SiO)ZP=O(H). Some additional insights into the nature ofthe products formed can be gleaned from the low wavenumber spectral region. Sharp IR bands near 1410 and 1300 cm-l are diagnostic of the presence of a PCH3 containing species1* (CH3 antisymmetric and symmetric angle deformation modes respectively). For the parent compound, these bands are at 1412 and 1284 cm-', respectively. During the early course of the reaction (1 to 3.5 h; see Figure 4) the growth of a new IR band a t 1306 cm-l could be correlated with the growth of the PH stretching band of species A a t 2390 cm-'. These bands (1306 and 2390 cm-l) were also observed when silica was doped with methylphosphinic acid. At low wavenumber, a pair ofweak IR bands was always observed near 700 and 635 cm-l (Figure 5 shows several examples). IR bands a t these wavenumbers have been shown to be due to the direct chlorination of silica15(Sic1 and SiC12, respectively). On the other hand, the P-CH3 stretching mode is also expected to lie near 700 cm-' and the intensity at 700 cm-l many also have a contribution from this modes1* Finally, the reactions described above proceeded a t a reasonable rate only if the silica had been activated at temperatures below about 500 "C. For higher temperatures of activation, the OH concentration diminishes,16 and is very low when the activation is carried out a t 800 ~
~
~~
~~
(14)Thomas,L. C.Interpretation of the Infrared Spectra of Organophosphorus Compounds; Heydon: London, 1974. (15)Lang, S.J.;Morrow, B.A. J.Phys. Chem. 1994,98,13314. (16)Zhuravlev, L. T.Langmuir 1987,3, 316.
Adsorption of P(CHdC12 on Silica
Langmuir, Vol. 11, No.7,1995 2537
+ PMeC1, - MeP=O(H,Cl) + Sic1 SiOH + MeP=O(H,Cl) - SiOP=O(Me,H) + HCl SiOH
T
(1) (2)
the complete reaction being
2SiOH
'/
635
700
/
Figure 5. IR spectra (740-550 cm-') recorded as follows: (A) 2 mg/cmzdisk, 47 Torr of PMeClz after 3.5 h; (B) 6 mg/cmz disk, 150 "C activation, 43 Torr of PMeClz after 18 h; (C) 4 mg/cmz disk, 37 Torr of PMeClz after 20 h. In all cases the gas phase had been evacuated for 15 min.
"C or higher. We found t h a t the reaction rate was slower by a factor of -3 if the silica had been activated a t 700 "C, whereas for 1100 "C activation there was insignificant reaction with the few residual surface silanol groups after 3.5 h.
Discussion (a) Identity of Species A. The NMR and IR results have shown t h a t species A can be produced by the adsorption of PMeC12 or methylphosphinic acid on silica. This species is immobile and is probably: H
I
SiO- P=O
I
Me A
This is assignment is supported by the following: (1) IR bands at 1306/1412cm-l suggest that there is a PMe group; (2) a PH band at 2390 cm-' for A is close to 2386 cm-l found for methylphosphinic acid, MeP=O(H,OH), and the PH frequency is not expected to change significantly for replacment of the OH group by "OSi"; (3) the 31Pshift for MesSiOP=O(H,Me) is 30.4 ppm17compared with 22 ppm found for A. The generation of this species requires the presence of SiOH. The reaction rate is very slow for a highly dehydroxylated surface which contains few SiOH species and is much faster when the SiOH concentration is higher. There is evidence that there is chlorination of the surface to yield SiC1, and HC1 is generated in the gas phase. Therefore, we postulate the following basic two-step process: (17)Livantsov, M. V.;Prishchenko. A. A.; Lutsenko, I. F. J . Gen. Chem. USSR (Engl. Transl.), Zh. Obshch. Khim. 1985,55,1976.
+ PMeC1, - SiOP=O(Me,H) + Sic1 + HC1 A
(3)
The intermediate methylchlorophosphine oxide was not detected via IR or NMR spectroscopy; the other products were. We assume that it is only present as an intermediate in low concentration and that it readily reacts with SiOH. The immediate decrease in the intensity of the free SiOH peak a t 3747 cm-l and the appearance of a broader new peak at 3650 cm-l following addition of PMeC12 are due to its physical adsorption and its weak perturbing influence on the free SiOH groups. A similar effect was found for the physical adsorption of Pcl3 on silica.18 As the reaction proceeds, the 3650 cm-l band slowly decreases in intensity as a very broad 3250 cm-' band appears, accompanied by the growth of the PH band of species A at 2390 cm-l. In the early stages of the reaction, and particularly for lower pressures of PMeC12, the 3250/ 2390 cm-l peaks grew in unison, but after about 2 to 3 h, the rate of growth ofthe 3250 cm-l band decreasedrelative to the PH band, and eventually started to decrease in intensity (Figure 2A,B). As will be discussed below, the decrease in the intensity of the 3250 cm-' band was not related to the formation of species B. A similar broad band was found from the adsorption of trimethyl phosphite or dimethyl methylphosphonate on silica,l9 but its intensity was about eight times greater. The 3250 cm-l band is due to a strong H-bonding interaction between A and some OH groups. The eventual decrease in the intensity of the 3250 cm-l band occurs because the SiOH groups are continually being consumed by the reactant. (b)Identity of Species B. The assignment of species B is more problematic. The IR and NMR results suggest that B can arise from the deposition of phosphorus acid, H3P03,on silica. The species is protonated, a s evidenced by 31Pdelayed decoupling, and IR spectroscopy has shown that the 2490 cm-l PH stretching band could also be produced by doping silica with phosphorus acid.18 Further, we have previously shown t h a t the reaction of PC13 and Si02 also gave species B but never in such a large quantity and only on a longer time scale.18 The two peaks a t -8 and -18 ppm must be due to SiOP=O(H,OH) and (Si0)2P=O(H), respectively. Because the parent compound has a P-CH3 group and this is never removed in any chemical reaction a t room temperature,20 we were forced to assume that species B must have been generated from an interaction between PC13 and the products of the SiODMeC12 reaction. In order to address this problem, we synthesized two PMeCl2 samples which were free of PCl3. The first was 99% enriched in 13C, but only enough material for two NMR experiments was obtained. At short reaction times, 31PMAS-NMR showed that only species A was formed, as anticipated. For long reaction times species A was also formed, but species B was not. The 13C MAS-NMR of both these samples showed a doublet centered a t 15 ppm with a J p - c coupling constant of about 85 Hz. These observations further support the assignment of the structure SiOP=O(H,Me) to species A, since the analogous (18)Morrow, B. A.; Lang, S.J.; Gay, I. D. Langmuir 1994,10,756. (19)Gay, I. D.;McFarlan,A. J.; Morrow, B. A. J . Phys. Chem. 1991, 95,1360. (20)Weissermel, K.;Kleiner, H.-J.; Finke, M.; Felcht, U.-H. Angew. Chem., Znt. Ed. Engl. 1981,20, 223.
Lang et al.
2538 Langmuir, Vol. 11, No. 7, 1995 species A
3HC1+ 3SiOP=O(Me,H) 3MeP=O(H,OH) 2PCl,
+ 3MeP=O(H,OH)
f
+ 3Si-C1 (5a) 3PMeC1, + 2H3P0, (5b)
2H,PO, + 3Si-Cl(SiO),P=O(H)
+ SiOP=O(H,OH) + 3HC1 (5c)
the complete reaction being
3SiOP=O(H,Me) (SiO),P=O(H)
2500
2400
2450
2350
cm’ Figure 6. Dashed (- - -) curve: IR spectrum observed in the PH stretching region after exposure to 2 1 Torr of pure PMeClz (freeof PC13 as an impurity). The gas phase was then condensed in liquid nitrogen in a separate compartment from the sample, and Pc13 equivalent to 7%oftotal P was added to the condensate. This was then warmed to room temperature and was re-added to the reaction cell. The subsequent spectra showing a decrease at 2390 cm-l and an increase at 2490 cm-l were recorded at the following times after admission of the “mixture”to the cell: 0.3, 2.3, 4.3, 6.3, and 12.3 h. values were found to be 14.6 ppm and 90 Hz in the model compound MeOP=O(H,Me).21 The second sample, not 13C enriched, was used for IR experiments. Again only species A was generated, even after 20 h ofreaction time. As with the PC13contaminated sample, the broad 3250 cm-l band initially increased in intensity and then decreased. When PC13 was added to the gas phase after 2 h of reaction, species B formed with a concomitant decrease in the amount of species A, as was found with commercial PMeClz samples (Figure 6). The above results show clearly that the formation of species B is an artifact due to the presence of PC13 in the “as received” samples of PMeC12. There is a n isosbestic point in Figure 3B showing that there is a correlation between the disappearance of species A and the creation of B. Now it is known that PC13 reacts with phosphinic acids as follows:22
2PC13 + SRP=O(H,OH)
-
3PRC1,
+ 2H3P03
(4)
When silica was doped with MeP=O(H,OH), only the parent acid and species A were observed by NMR and IR. When PC13 was allowed to react with the doped SiOz, species B formed rapidly, and species A was consumed. In this reaction the formation of PMeCl2 was observed by both NMR and IR. This suggests that a reaction similar to eq 4 can occur on silica. If it is assumed that gaseous HC1 present in the ambient gas can participate in equilibria involving the chemisorbed species, the following scheme can rationalize the formation of species B from (21) Synthesized by us a s a by-product ofthe methanolysis ofPMeCl2,
natural abundance I3C NMR. (22) Frank, A. W. I n Organic Phosphorus Compounds; Kosolapoff, G. M., Maier, L., Eds.; Wiley-Interscience: New York, 1972; Vol 4.
+ 2PC13 * + SiOP=O(H,OH) + 3PMeC1,
(5d)
Equation 5d would lead one to predict equal amounts of mono- and difunctionally chemisorbed phosphorous acid should be formed. In fact, 31P NMR shows that the difunctionally adsorbed form, (Si0)2P=O(H), predominates a t least a t the late stages of the reaction between Si02 and PMeClflC13 mixtures. The scheme outlined in eqs 5 does not account for all of the subtleties of the observed reactions, and it cannot be expected to do so. There are simply too many possibilities for competing reactions to occur. For example, since SiOH groups are low in concentration a t the late reaction times during which species B appears, it is tempting to invoke the participation of siloxane bridges in the formation of species
B:
SiOP=O(H,Me)
+ PC1, + SiOSi (SiO),P=O(H) + Sic1 + PMeC1,
(6)
Unfortunately, a large number of reaction equations can be written to explain the observed products, but, since distinguishing between them experimentally is very difficult, further speculation is not warranted. Whatever the exact mechanism, the generation of species B is an artifact due to the presence of PC13. In spite of being present in only 5 mol % amounts, the presence of PC13 manifested itself due to the high pressures of PMeClz required for the reaction to proceed a t a reasonable rate. Thus, a 25 to 50 Torr pressure of reactant in a 300 mL volume would contain 20 to 40 pmol of Pcl3. A 50 mg silica disk of area 325 m2/gand containing 1.2 OH groups/ nm2 contains 32 pmol of OH.’ Therefore, the quantity of PC13 present is, in fact, comparable to the number of OH groups available. We have verified that with PC13 as an intentionally added impurity to PMeC12, a t levels between 2 and 15 mol %, the same product distribution is observed. Conclusions The reaction of pure PMeC12 with Si02 produces SiOP=O(H,Me), SiC1, and HC1 and a likely reaction mechanism has been deduced. If the reactant is contaminated with small quantities of PC13in the range from 2 to 15 mol %, then the above species reacts with the mixture after long (several hours) reaction to produce two secondary phosphorus containing chemisorbed products, SiOP=O(H,OH) [minor product] and (Si0)2P=O(H) [major product], and some plausible reaction sequences have been discussed. Acknowledgment. We are grateful to the Natural Sciences and Engineering Research Council of Canada, and to Esso Petroleum Canada, for financial support. LA941008N
