The Journal of Physical Chemistry, Vol. 83, No. 24, 1979 3075
Photochemistry of Silicon Compounds
References and Notes
I 2o
t
03
02
01
00
01
02
03
04
05
IT*
Figure 1. Plot of log k,e,values for pyrolysis of CH,=CHCH,CH(OAc)R against u*.
cules is of utmost importance for neighboring group part i ~ i p a t i o n .Since ~ the transition state for ester pyrolysis the occurrence is semipolar or semiconcerted3J3~21~25~26~34~35 of anchimeric assistance from substituents is therefore difficult. The participation found in reaction 1and 2 is, as is stated by the authors,lZ due to formation of a positive carbon atom stabilized by surface catalysis during the stirred flow pyrolysis of the system. An assessment of the effect of substituent R in the elimination of the ester type I11 toward the Co-H adjacent CH,=CHCH,CH-R OAc
IIIa, Z = H IIIb, Z = CH, IIIc, Z = (CH, ) 3 C to the vinyl group is plausible, since olefin formation is found to be produced by kinetic control (Tables IV and V). Therefore, the effect of R at the a carbon in I11 gives a good linear relationship with u* values (Taft equation)36 (Figure 1). Consequently the alkyl substituents at the a carbon in the butenyl acetates seems to exert a +I electron release and not steric acceleration.21
Hanack, M.; Schneider, H. H.; Schneider-Bernlohr, H. Tetrahedron 1967, 23, 2195. Thies, R. W.; Schick, L. E. J . Am. Chem. SOC. 1974, 96, 456. Kramer, D. J.; Smith, G. G. Int. J. Chem. Kinet. 1974, 6 , 849. Wertz, D. H.; Allinger, N. L. J . Org. Chem. 1977, 42, 698. Chuchani, G.; Martin, I.; Bigley, D. B. Int. J. Chem. Kinet. 1976, 10, 649. Herndndez, J. A.; Chuchani, G. Int. J . Chem. Kinet. 1976, 10, 923. Chuchani, G.; Martin, I.; Martin, G.; Bigley, D. B. Int. J. Chem. Kinet. 1979, I f , 109. Maccoll, A. Chem. Rev. 1969, 69, 33. Servis, K. L.; Roberts, J. D. J . Am. Chem. SOC.1964, 86, 3773. Rogan, J. B., J . Org. Chem. 1962, 27, 3910. Kosower, E. M. “An Introductionto Physical Organic Chemistry”; Wiley: New York, 1968; p 120, and references cited therein. Emovon, E. U.; Maccoll, A. J . Chem. SOC.1964, 227. Lum, K. K.; Smith, G. G. Int. J. Chem. Kinet. 1969, 1 , 401. Sarel, S.;Newman, M. S. J. Am. Chem. SOC.1956, 78, 5416. Golinov, V. P. Zh. Obshch. Khim. 1952, 22,2131. See Chem. Abstr. 1954, 48, 1240c. Wagner, G. Berichte 1694, 27, 2434. Vogel, A. I.“Practical Organic Chemistry”, 3rd ed;Longmans: London, 1956; p 372. Maccoll, A. J . Chem. SOC. 1955, 965. Maccoll, A.; Thomas, P. J. J . Chem. SOC. 1955, 979. Bridge, M. R.; Davies, D. H.; Maccoll, A.; Ross, R. A,; Banjoko, 0. J . Chem. SOC.B 1966, 805. Burgh Norfolk, S.de; Taylor, R. J. Chem. Sm.,Perkin Trans. 2 1976, 280. Scheer, J. C.; Kooyman, E. C.; Sixma, F. L. J. Recl. Trav. Chim. Pays-Bas 1963, 82, 1123. Ruzicka, V.; Cerveny, L.; Prochazka, J. Inst. Chem. Techno/.Prague, Org. Chem. Techno/. 1970, C15, 57. Benson, S. W.; O’Neal, H. E. Natl. Stand. Ref. Data Ser., Nat/. Bur. Stand. 1970, No. 21. Chuchani, G., Martin, I.; Avila, I. Int. J . Chem. Kinet. 1979, 11, 561. Taylor, R. J. Chem. Soc., Perkin Trans. 2 1975, 1025. Chytry, V.; Obereigner, B.; Lim, D. Europ. Polym. J . 1973, 9 , 493. Chuchani, G.; Martin, G.; Barroeta, N.; Maccoll, A. J . Chem. Soc., Perkin Trans. 2 1972, 2239. Chuchani, G.; Piotti de Chang, S.; Lombana, L. J. Chem. Soc., Pekin Trans. 21973, 1961. Emovon, E. U.; Maccoll, A. J . Chem. SOC.1962, 335. House, H. 0. “Modern Synthetic Method”, 2nd ed; Benjamln: Philippines, 1972; Chapter 9. Brown, R. S.; Taylor T. G. J . Am. Chem. Sac. 1973, 95, 8025. Johnson, C. D. “The Hammett Equation”, Cambridge: London, 1973; Chapter 2. Taylor, R. J . Chem. SOC.,Perkin Trans. 2 1972, 165. Cuenca, A.; Chuchani, G. Int. J . Chem. Kinet. 1977, 9 , 379. Taft, Jr., R. W. “Steric Effect in Organic Chemistry”, Newman, M. S., Ed.; Wiley: New York, 1956; Chapter 13.
Photochemistry of Silicon Compounds. 6. The 147-nm Photolysis of Tetramethylsilane L. Gammle,’ C. Sandorfy,’ and 0. P. Sfrausz* Department of Chemistry, University of Alberta, Edmonton, Alberta, Canada T6G 2G2, and Dgpartement de Chimie, Universit6 de Montreal, Montreal, Quebec, Canada H3C 3Vf (Received December 28, 1978; Revised Manuscript Received August 10, 1979) Publication costs assisted by the University of Alberta
The 147-nm gas-phase photolysis of tetramethylsilane yielded ten measurable and several trace products along with a solid deposit. From the effect of pressure, exposure time, deuterium labeling, and added nitric oxide on the quantum yields of individual products, the following primary steps were postulated: CH3 + Si(CH3)3 (4 = 0.43); 2CH3 + Si(CH3I2(4 = 0.24); CH4 + CH2Si(CH3I2(4 = 0.17); CH3 + H + CH2Si(CH3)2(4 = 0.10); Hz + CHSi(CH3)3(4 = 0.02); and CH2 + (CH3I3SiH(4 = 0.04). Fluorescence could not be observed and the The secondary reactions of the principal silicon radicals Si(CH3),, CH2Si(CH3)2,and upper limit for +f is Si(CH3)2and the mechanism of the nitric oxide inhibited reaction are discussed and it is shown that the siloxy radical (CH3I3SiOcan displace CH3 from the substrate to give hexamethyldisiloxane. In earlier studies3+ on the 147- and 124.6-nm photolysis of monomethylsilane and dimethylsilane it has been shown *Address correspondence to this author at the University of Alberta. 0022-3654/79/2083-3075$01 .OO/O
that the photochemical decomposition of these molecules is characterized by a complex array of simultaneous, competing primary steps comprised of molecular and free-radical modes of splitting, Yielding two Or more fragments. The role of polyfragmentation is more important in the 0 1979 American
Chemical Society
3076
The Journal of Physical Chemistry, Vol. 83, No. 24, 1979
124.6-nm photolysis than in the 147-nm photolysis. The mechanism of the overall reaction taking place in these systems was deduced from product measurements and detailed kinetic studies of the effect of pressure, exposure time, wavelength of photolysis, added radical scavengers such as nitric oxide and ethylene, and deuterium labeling on the reaction. The intermediacy of the silylene diradicals SiH2, CH3SiH, and (CH&Si was demonstrated in the reactions and furthermore it was shown that all three silylenes readily insert into the Si-H bonds of the substrate molecule to yield disilanes but they are unreactive with respect to the C-H bonds or with nitric oxide or ethylene. Silyl radicals on the other hand are highly reactive with nitric oxide' and ethylene and therefore are readily scavengeable. Thus, the use of these free-radical traps makes differentiation between the reactions of monovalent and divalent silicon hydride radicals possible. Evidence in these studies was also found for the intermediacy of silaethylene8-11species, a highly reactive silicon analogue of ethylene. The photolysis of tetramethylsilane (TMS) was undertaken in order to extend earlier studies on methylsilanes to this species. A somewhat related study of this molecule has been reported several years ago, namely, its 50-feV X-ray radiolysis.12 In this case the major products were hydrogen, methane, ethane, and ethyltrimethylsilane (ETMS). The reaction mechanism was deduced from deuterium labeling and nitric oxide/ethylene scavenging experiments. The hydrogen and methane were produced by molecular elimination, atomic reactions, and, to a lesser extent, by fragmentation of "hot" intermediates. The ethane was formed almost exclusively by combination of methyl radicals and the principal silane products seemed to arise from reactions of the (CHJ3Si and (CH3),SiCH2 radicals. The authors estimate that -12% of the H2 and -40% of the methane are formed by molecular elimination and that 25% of the methane is formed from unscavengeable "hot" methyl radical abstraction reactions. Although to a large extent the reactions in this type of radiative excitation can be explained by free-radical mechanisms, the possibility of the involvement of ionic and superexcited species cannot be ignored.13 Usually molecules close to the particle tracks will be ionized while others further away will just be raised to an excited state. The thermal decomposition of TMS has also been investigated14in a flow system at 810-980 K and the decomposition was found to be first order. The major products were methane, ethane, trimethylsilane (Tri-MS), 1,1,3,3tetramethyl-l,3-disilacyclobutane, hexamethyldisilane (HMDS), and 2,2,4,4-tetramethyl-2,4-disilapentane (TetMDSP). The rate coefficient determined had the value, log k = (14.30 f 0.23) - (67600 f 800)/2.303RT. The activation energy is smaller than the reported D((CH3)?Si-CHJ of 76.6 kcal/mol, so the presence of a chain mechanism could not be ruled out. Silaethylene, (CH3)2SiCH2,was postulated as an intermediate because of the composition of the polymer formed, the presence of the disilacyclobutane dimerization product, and the possibility of radical decomposition (CH3)$3iCH2 (CH3)zSiCH2+ CH3 lowering the experimental activation energy. Disproportionation of trimethylsilyl radicals and cross disproportionation of trimethylsilyl and methyl radicals were also postulated as sources of the intermediate. The occurrence of secondary thermolysis of the radicals and the paucity of information on the stability and reactivity of the intermediates implicated renders the interpretation of the mechanism difficult.
-
Gammie, Sandorfy, and Strausz I
- 1 69
L u I
40'
I
I
n
Y
X
x rm
U
Y
x n
x /1
1
The Journal of Physical Chemistry, Vol. 83,No, 24, 1979 3077
Photochemistry of Silicon Compounds
-ae I
0
"
c c
."-.
0
T ' I
X
I
0 01
-
5 -
2
5 -
5 0M-k u-l >
0%
4 -
0 -
ci
15
2
0
-
L
3:
-
L1
8
6
Figure 4. Quantum yields of C,H, of time.
"
10
12
14
16
TIME (minutes)
" _
L
X
n
4
(0) and Tri-MS (X) as a function
U
W
IO
5
-
a_.,
0
0
" 0
1
0
0 J 1
2
>
30 -0
2
b
0
0
w
0
0
u
0
0
0
20-
5 w
X
D:
lo,;x
0
C,H, CH, Tri-MS H2 HMDS
0.35 0.33 0.13 0.12 0.09
EDMS ETMS Tri-MDSP Tet-MDSP DMDSH
::
10
5
IS
TIME ( m i n s )
0.05 0.04 0.04
Flgure 5. Variation in relative yields of C2H, (0) and H2 (X) with time.
0.02 0.01
TABLE 11: Effect of Nitric Oxide on Photolysis of Tetrameth ylsilane
bilization of "hot" products and variations in the metathetical and combination reactions of free-radical intermediates. In a complementary fashion the yields of ethane, hydrogen, and ethyldimethylsilane decrease with pressure, suggesting that in the low-pressure regime fragmentation of hot primary products produces additional amounts of these molecules. Other products are not noticeably dependent on pressure. Above 100 torr, product yields become constant and consequently all subsequent experiments were carried out at pressures 2100 torr to avoid interference from pressure dependent fragmentations. The average relative yields of the pressure stabilized system are as follows: CzH6,30%; CH4, 31.0%; Tri-MS, 12.0%; H,, 9.0%; HMDS, 6.0%; EDMS, 3.0%; ETMS, 3.5%; Tri-MDSP, 2.0%; Tet-MDSP, 1.0%; and DMDSH, 1.0%. Since the polymeric deposit on the cell window had an attenuating effect on the light intensity during photolysis, product quantum yields were determined from exposure time studies by extrapolation to zero exposure time. Carbon dioxide was employed as the actinometric gas with @(CO)= l.0.15J6 Illustrative examples of the results are presented in Figure 4. Since these plots are nearly exponential in nature the error in the extrapolated values was minimized by using semilogarithmic plots of 4 vs. time. The product quantum yields so obtained are summarized in Table I. The exposure time study was also used to decide if any of the products were of secondary origin. As seen from the illustrative plots shown in Figure 5 which apply for all principal products, the relative product yields are essentially constant with time; consequently, they are of primary origin.
% added nitric oxide
producta
0.0
2.5
5.0
10.0
0.011 0.004 b b 0.091 0.138 b b ;hDS 0.000 0.000 EDMS 0.011 0.000 ETMS 0.050 0.032 Tri-MDSP 0.002 0.003 Tet-MDSP 0.000 0.000 DMDSH 0.000 0.004 A [0.817] [0.145] B 0.3221 [0.371] [0.428] HMDSO 0.022 0.288 0.328 0.985 C 0.0041 [0.010] [0.042] D 0.028] [0.028] [0.087J N2 0.450 1.780 2.200 b b b N,* a &mol. Not determined. Values in brackets indicated that the response factors are not known, and were taken the same as for HMDS. C,H, CH, Tri-MS
0.795 1.136 0.405 0.375 0.222 0.123 0.101 0.086 0.044 0.031
0.210 b 0.100 b 0.001 0.017 0.059 0.000 0.000 0.003
Next the effect of added nitric oxide was investigated. This molecule has been widely used in carbon free-radical systems as an efficient scavenger of monovalent carbon radicals, and some years ago it was shown to rapidly react with monovalent silicon hydride radicals as well.7 Since then it has been used as a scavenger to differentiate between products formed from monovalent carbon or silicon hydride and divalent carbon or silicon hydride free radicals since it is unreactive with respect to divalent species. Experiments were carried out with up to 10% added nitric oxide and the results are summarized in Table 11.
3078
The Journal of Physical Chemistry, Val. 83, No. 24, 1979
Gammie, Sandorfy, and S t r a w
Mole Fraction TMS
Mole Fraction TMS Figure 6. Variation in isotopic methane yield as a function of relative ( 0 )CH,, (A)CD,H, (0)CH,D, (H)CD,. amounts of TMS and TMS-d,,:
Ethane, ethyldimethylsilane, hexamethyldisilane, and the three other disilanes seem to be formed wholly by monoradical reactions since their yields are drastically suppressed by nitric oxide. The yields of trimethylsilane and ethyltrimethylsilane are both partially reduced indicating that each of these products is formed by at least two different reactions, and at least one of the contributing reactions is nonscavengeable. It is seen that 2.5% NO is sufficient to suppress HMDS formation and to scavenge all trimethylsilyl radicals. At this concentration of NO the C2H6yield is decreased to only about 25% and, therefore, the NO concentrations were increased to 5 and 10% which then caused the complete suppression of C2Hs. At 10% NO, however, secondary reactions such as the photolysis of NO, its alkyl free-radical induced disproportionation, R. + 3N0 R. + NO3 + Nz, oxidation reactions of NO3, scavenging of CHz and (CH&Si, and other not understood complications set in affecting the yields of Tri-MS and ETMS. The HMDSO and nitrogen yields are also affected, both seem to rise with increasing NO concentration. For these reasons the experiment with 10% added NO is disregarded and the scavenging limits for (CH3)3SiHand ETMS were taken from the 5% NO experiment. It will also be noted from Table I1 that several new products arose from the nitric oxide inhibited reaction. In addition to Nz and N20, hexamethyldisiloxane (HMDSO) and four other products (A-D) were also detected. These were tentatively identified as methoxytrimethylsilane (A), trimethylhydroxylamine (B), and two different trisiloxanes (C and D). The appearance of HMDSO clearly demonstrates that (CH3)3Siradicals are present in the system. This conclusion was further reinforced by experiments carried out in the presence of added SiH4which served as a hydrogen donor. The photolysis of a mixture of 200 torr of TMS 10 torr of SiH4 resulted in an approximately 50% increase in the yield of trimethylsilane owing to the metathetical reaction between (CH3)3Siand SiH4. Further experiments were carried out for the determination of the modes of formation of hydrogen and methane. These two molecules were previously found to be the chain products of the nitric oxide inhibited photolysis of rnethylsilane~~ and for this reason isotopic labeling was used for the derivation of the mechanism of their formation. As in previous ~ t u d i e s from ~ - ~ this laboratory, mixtures of light and deuterated silanes were photolyzed and the isotopic composition of hydrogen, methane, and ethane measured as a function of the isotopic composition of the substrate. The results of these experiments are presented in Figures 6-8. For the evaluation of these results it was
Flgure 7. Variation in isotopic ethane yield as a function of relative amounts of TMS and TMS-d,,: (0)C,H, (A)CH3CD3, (0) C2DB. IO
08
E
06
w
2
LL
a,
5 04
s
-
+
Mole Fraction TMS Figure 8. Variation in isotopic hydrogen yield as a function of relative amounts of TMS and TMS-d,,: (0)H, (A)HD, (0) D,.
TABLE 111: Comparison of Relative Photolysis Yields for Tetramethylsilane and Tetramethylsilane-d,
C,H, CH, Tri-MS H, HMDS EDMS ETMS Tri-MDSP Tet-MDSP DMDSH
30.0 31.0 12.0 9.0 6.0 3.0 3.5 2.0 1.5 1.0
31.7 32.8 9.9 10.2 6.8 3.9 2.5 1.0 0.4 0.7
0.95 0.88
also necessary to determine the relative product yields from the photolysis of TMS-d12compared with those from TMS as listed in Table III. A search was also made for fluorescence from TMS; however, no fluorescence could be observed and an upper limit of 110-5was established for the fluorescence quantum yield. Discussion From the results presented above it is evident that the photolysis of TMS is complex and, as had been shown to be the case earlier for m ~ n o m e t h y l s i l a n eand ~ ~ dimethyl~ i l a n eseveral ,~ simultaneous primary steps must be operative. The nitric oxide scavenging experiments indicate that most of the products arise from monovalent free-rad-
The Journal of Physical Chemistry, Vol. 83, No. 24, 1979 3079
Photochemistry of Silicon Compounds
-
2(CH3)3Si
(CH3)3SiH 3- CHzSi(CH3)2
(10)
ical precursors with the notable exception of Tri-MS and ETMS (Table 11). Ethane, the highest quantum yield product with 4 = 0.35, is nearly completely scavengeable with nitric oxide. Therefore it can only form from combination of methyls. The possibility that it arises from ethyl radicals can be discounted because of the absence of n-butane, the combination product of ethyls. Hexamethyldisilane is formed by combination of two trimethylsilyl radicals 2(CH3),Si (CH3)3SiSi(CH3)3 (1)
with r#1(11) = 0.135 - 0.034 - 0.004 = 0.097. The nonscavengeable ethyltrimethylsilane could be obtained by insertion of methylene into a substrate molecule CH2 + (CH3)4Si C2H5Si(CH3)3 (12)
since HMDS is again completely scavengeable. 4(1) = 0.095 so the yield of trimethylsilyl radicals required is 0.19. Ethyldimethylsilane is most likely formed by the combination reaction CH3 + CH2Si(CH3), C2H,Si(CH3)z (2)
with $(12) = 0.042 X 0.5 = 0.021. The remainder of the methylene may also insert into the substrate but the chemically activated ETMS may decompose yielding the traces of ethylene found in the system: ETMS* CzH, (CH3)3SiH (13)
-
-
since dimethylsilaethylene is present in the system as will be shown below, followed by the abstraction reaction C2H5Si(CH3),+ Si(CH3)4 C2H5SiH(CH3)2+ CHzSi(CH3)3(3)
-
Thus, the yield of methyl radicals consumed via (2) and (3) is 0.050. Trimethyldisilapentane could form by two alternative combinations (CH3)3Si+ CHzSi(CH3)z (CH3)3SiCH2Si(CH3)2 (4)
-
or (CH3)3SiCHz+ Si(CH3)z
(CH3)3SiCH2Si(CH3)2(4a)
-+
followed by the abstraction step + (CHJ4Si (CH3)3SiCH2Si(CH3)2 Tri-MDSP
CH,Si(CH,), (5)
The abstraction reactions of these chemically activated silyl radicals should be significantly more efficient than those of the thermalized radicals and in view of the relatively minor importance of the end products, EDMS and Tri-MDSP (cf. Table I), the error due to the neglect of other possible reactions of these radicals should not be large. The formation of the Tet-MDSP and TMDSH can be accounted for in terms of the combination reactions (CH3)3Si CH2Si(CH3), Tet-MDSP (6)
+
-
-
2(CH3I3SiCH2 ((CH3)3SiCHz)2 (7) Dimethyldisilahexane formation would seem to require combination of two dimethylsilaethylenes
-
would be 0.046 X 0.095 = 0.004. The rest of the trimethylsilane is formed possibly via abstraction by trimethylsilyl from the substrate (CH3)3Si (CH3)4Si (CH3)3SiH+ CH2Si(CH3)3(11)
-
ETMS*
M
ETMS
+M
(14)
-
+
and thus, 4(15) = 0.042 - 0.021 = 0.021. The modes of formation of hydrogen and methane were determined from the results of the isotopic labeling studies. Both hydrogen and methane could be produced in molecular eliminations (CH3)4Si-% H2 + CHSi(CH3)3
-% CH4 + CH2Si(CH3)2 as well as by radical producing steps followed by abstraction from the substrate (CH3)4Si.-% CH3 + Si(CH3),
-?+ 2CH3 + Si(CH3& -kH + CH2Si(CH3I3 -% H + CH3 t CHzSi(CH3)z Kinetic schemes for the formation of the different isotopic hydrogen, methane, and ethane species were drawn up omitting the silicon fragment for clarity. Primary steps for hydrogen production are TMS-dlz + hv [TMS-d12]* (16)
---
[TMS-dlJ*
+
D2
(17)
[TMS-diz]* D TMS + hv [TMS]*
TMS
(CHJ4Si -?+ CH2 + (CH3)3SiH (9) Elimination of methylene has been found to occur in other methylsilane systems as well. @(Tri-MS)= 0.135, so @(9)= 0.034. As has been shown elsewhere the disproportionation of trimethylsilyl radicals is a minor process7J7with h d / k , N 0.046. Thus, the yield of trimethylsilane from disproportionation
-+ - +
The rest of the ethyltrimethylsilane must form by monoradical reactions, the major source being methyl and trimethylsilylmethyl combination CH3 CH2Si(CH3), C2H5Si(CH3)3 (15)
(CH3)~SiCH~CH~Si(CH3)2 ~(CH~)ZS~CH Z DMDSH (8) Formation of the disilanes requires a trimethylsilyl radical yield of $(Tri-MDSP) + 4(Tet-MDSP) I0.055. About 25% of trimethylsilane is unscavengeable in the presence of nitric oxide and this is formed in a primary step
-
+
(18)
(19)
H2
[TMS]"
(20)
[TMS]* H Abstraction then leads to the reactions H TMS Hz H + TMS-dlZ HD
+
(21)
-
-
(22) (23)
+
+ TMS D + TMS-di2 D
HD
(24)
Dz (25) It has been shown in the photolysis of dirnethyl~ilane~ that exchange reactions of the type H + TMS-d12 D; H + TMS-d12 CD3, etc. are unimportant at room temperature and were disregarded here. Primary steps for methane and ethane production are
-
+
-
3080
The Journal of Physical Chemistry, Vol. 83,No. 24, 1979
TMS
+ hv
[TMS]* [TMS] * TMS-dlz + hv
[TMS]*
(26)
CHI
(27)
CH3 [TMS-d12]*
(28)
-+
+
---
+
[TMS-d12]* [TMS-dlz]* and for abstraction CH3 + TMS CD3 + TMS
CD4
(30)
CD3
(31)
-+
CH3D
CD3H
(32) (33) (34)
CD3 + TMS-dlz -+ CD4
(35)
2CH3 -+ CzH6
(36)
CH3 + CD3 2CD3
-
+
CH3CD3
CzD6
(37)
CH3 + Si(CH3)3 CHI + CHzSi(CH3)z (40) may account for some of the methane and reaction 40 could be facilitated by the apparent stability of the silaethylene produced. From the results it is estimated that at least about 15% of the hydrogen arises from molecular elimination (CH3)4SiL Hz + CHSi(CH&
(41)
and the rest from atomic hydrogen via abstraction
+ CHzSi(CH3)3 2 H + CH, + CHzSi(CH3)2 H2 + CHzSi(CH3)3 H + (CH3)4Si--,
(CHJ4Si L H
(38)
In an earlier study3 on the photolysis of dimethylsilane the kinetics of an analogous mechanistic scheme was satisfactorily evaluated by iterative numerical integration, varying the relative rate constants for the individual reactions until a good fit was obtained with the experimental data. A similar procedure, taking tL = eHe and the relative rate constants kz7 = k30 = 1.5 X loW1,ha = k31 = 8.5 X lO-l, k32 = k34 = 5.0 X IO1, k33 = k35 = 5.0, and k36 = k3,/2 = k38= 3.0 X lo6, made it possible to reproduce the experimental isotopic distribution of the methane and ethane products (solid line curves in Figures 6 and 7). This gives 15% for molecular methane elimination. Reactions of methyl radicals with other radicals in the system were regarded as minor processes and were omitted to simplify the treatment. It is pointed out, however, that the computational results should not be significantly affected by the nature of the hydrogen abstraction steps. If the bulk of methane and Tri-MS arises from disproportionation with other radicals instead of abstraction from the substrate the steady state concentration ratios of undeuterated and deuterated radicals will be proportional to the concentration ratios of the substrate TMS/TMS-d12since the extinction coefficients at 147 nm and the individual product quantum yields in the photolysis of TMS and TMS-dlz may be considered identical to a first approximation. The experimental results on the isotopic distribution of the hydrogen product could not be reproduced possibly because the isotope effect on the molecular and atomic elimination steps is not negligible. Thus, Okabe and McNesbylRfousd an isotope effect on the Hz/D2elimination step in the photolysis of CH3CD3of 2.3 and Dees and Setzer on the HCl/DCl elimination from chemically activated ethyl chloride of 3.3.19 Nonetheless, the results clearly demonstrate the presence of both molecular and atomic hydrogens in the system and that the molecular elimination step is not negligible even in the TMS-d12 system. The minimum value for molecular elimination is at least 15% . One of the results of this simulation is that the rate constants for abstraction by methyl radicals are only five orders of magnitude less than the value for the combination of methyl radicals. This is four orders of magnitude larger than expected since the room temperature rate constant (1.4 X 10' L mol-l s-*)for methyl attack on tetramethylsilane from the measured20 rate coefficient is nine orders of magnitude less than that for methyl radical com-
-
bination.21 The implication is that hydrogen abstraction is much easier than expected possibly because the methyl radicals contain a considerable amount of excess energy or that hydrogen is abstracted from other radicals in the system. Cross-combination and disproportionation of methyl and trimethylsilyl radicals CH3 + Si(CHJ3 -+ (CH3)4Si (39) +
CH4
CH3 + TMS-dl2
(29)
Gammie, Sandorfy, and Strausz
(42) (43) (44)
It has been shown in the photolysis of simple alkanes22 and alkyl silane^^^ that single hydrogen atom loss is always a very minor process so reaction 42 has been omitted in favor of reaction 43. Thus, taking 4(41) = 0.122 X 0.15 = 0.018, we found that $443) will be equal to 0.122 - 0.018 = 0.104. The methane/ethane results show that molecular methane formation accounts for -15% of the total methane and methyl radical production. From the preceding steps we require a methyl radical yield of 24(CzH6)+ d(EDMS) d(ETMS) = 0.700 + 0.050 0.021 = 0.771 and we know that 4(CH4) = 0.330. Therefore
+
+
Thus, 4[CHlabs] = 0.165, and $[CHlmolecular] = 0.165, that is, approximately 50% of the methane is formed in a molecular elimination reaction (CHJ4Si -+ CH4 + CHzSi(CHJ2 (45) and 50% by methyl radical abstraction. The sum of methyl radicals in the system is now 0.771 + 0.165 = 0.936 and there are three major processes which can produce primary methyl radicals: (CH3),Si + hv CH3 + Si(CHJ3 (46)
--
2CH3 + Si(CH3)2 CH3 + H
(47)
+ CH2Si(CH3)2 (48)
It was not possible to differentiate between the first two steps since the silylene and possibly some of the trimethylsilyl radical will polymerize and be removed from the system. However, a probable scheme can be drawn up to account for all of the secondary products and most of the primary products and fragments. Since fluorescence does not occur, we assumed that the primary quantum yield for decomposition is unity and apportioned the methyl radical producing steps accordingly. So far we have a primary quantum yield of 9(9) + 4441) + q5(43) + 4(45) = 0.034 0.018 + 0.104 + 0.165 = 0.321. Thus, the remaining primary steps should account for 0.679 of the primary quantum yield, while at the same time they must give a methyl radical yield of 0.936 - 0.104 =
+
The Journal of Physical Chemistry, Vol. 83,No. 24, 1979 3081
Photochemistry of Silicon Compounds
TABLE IV: Summary of Primary Quantum Yields in the Photolysis of Tetramethylsilane excess @Aa @ B b energyC (CH,),Si + hu 1 -+
2 -+
3 -+
4 -+
5 -+
6 -+
CH,
+
Si(CH,),
2CH, t Si(CH,), CH, CH, H,
+ CH,Si(CH,), t H + CH,Si(CH,),
+ CHSi(CH,),
CH, t (CH,),SiH
0.51
0.43
110
0.16
0.24
53
0.17
0.17
117
0.10
0.10
13
0.02
0.02
100
0.04 0.04 -1.00
88
1.00
Assuming cross disproportionation does not occur. Assuming a k d / k , ratio of 0.1 for methyl and trimethylsilyl radical cross reaction. kcal mol". a
0.832. Therefore q5(46) + 4447) = 0.679 and 4(46) + 24(47) = 0.832. These equations give (b(47) = 0.153 and $446) = 0.526. However, it is possible to obtain somewhat different values for @(46)and 4(47) if we assume that cross combination and disproportionation does take place between methyl and trimethylsilyl radicals (reactions 39 and 40). This postulate might give more likely primary quantum yields for production of methyl radicals while simultaneously accounting for the trimethylsilyl radical yield. We first have to make a reasonable estimate of the k d / k , ratio for cross reaction of methyl and trimethylsilyl radicals and we have chosen a value of 0.1. Then, from the yields of products d(CH3)3Si)= 2441) + 4(4) + 403) + 24(10) + d11) + d439) + d 4 0 ) = 2(0.095) + 0.038 + 0.017 + 0.008 + 0.097 1.1$(39) = 0.350 + 1.14(39)
+
+
4(CH3) = 26(C2H6) 4(EDMS) + d E T M S ) + 0.165 + 4(39) = 0.700 + 0.050 + 0.021 + 0.165 + 4(39) = 0.936 + $439) Primary step (43) accounts for some of the methyl radicals so from primary steps (46) and (47) we require a methyl radical yield of 0.936 + 4(39) - 0.104 = 0.832 + 4(39) Since the sum of primary quantum yields (46) and (47) amounts to 0.679 and the yield of (CH3)3Siradicals is determined by (46) while the yield of methyl radicals is given by $(46) + 24(47) we have d(CHJ3Si) + 4(CH3) = d 4 6 ) + 4(46) + 2@(47)= 2(4(46) + 1$(47)) = 2(0.679) = 1.358 Le., 0.350 + 1.1$(39) + 0.832 + 4439) = 1.358 and 9(39) = 0.084. Then, $446) = 0.350 + 1.14(39) = 0.350 + 0.092 = 0.442, and d 4 7 ) = 0.679 - 0.442 = 0.237. So we can obtain a slightly different set of primary quantum yields (&J if we assume methyl and trimethylsilyl cross reaction. The primary quantum yields so obtained are summarized in Table IV. From the data presented in the foregoing, it can be seen that the 147-nm photolysis of TMS proceeds via at least six different competing primary steps, 98% of which involves C-Si bond cleavage and 33% C-H cleavage. The
free-radical mode decomposition (77% ) yielding CH3 and H-atom predominates over the molecular mode (23%) giving CHI, H2, or (CH3)3SiH. This particular set of primary steps was chosen on the basis of the principle of minimum bond cleavage and involves only one and two bond cleavages. Several additional primary steps can be considered on energetic grounds involving the rupture of three or more bonds which, however, statistically would be less probable. Also the other alternative one bond cleaving primary step (CH3)&3i CH2Si(CH3)3+ H may occur to some extent but since the total quantum yield of the H-atom yielding primary steps is small, -0.10, this was neglected in the kinetic treatment of the results. The three-fragment yielding primary steps could occur sequentially by pressure independent fragmentation of the primary fragments as was before in regard to the photolysis of monomethylsilane and dimethylsilane. From comparison of the data in Tables I and IV it can be estimated that, of the primary yield of $&) = 0.43 for the production of (CH3)3Siradicals, 0.33 appears in the retrievable products and 0.10 is lost to the polymer. The yield of retrievable products requiring hydrogen uptake is H H-R- H2 + R 0.10 CH4 + R 0.16 CH3 + H-R (CH3)3Si H-R (CH3I3SiH R 0.10 EDMS R 0.05 Et(Me)2Si + H-R Tri-MDSP. + H-R Tri-MDSP R 0.04 *DMDSH*+ 2H-R DMDSH 2R 0.02 which gives a sum of 0.47. Abstraction by hydrogen atom from TMS is fast and probably quantitative. This conclusion is based on the comparison of available kinetic data on the reactions of H atoms with (CH3)4C23 and (CH3)4Si,24 and the quantum yield of H2from the room temperature triplet mercury photosensitization of (CH3)4C.25Abstracand by tion by CH3 is slow, K = 1.4 X 10' L mol-l (CH3)3Siit should even be slower due to the endothermicity of this reaction. However, the CH3 and (CH3)3Si radicals produced in primary step 1carry a large excess of vibrational energy because of the high exothermicity of this step. The photonic energy at 147 nm is 195 kcal/ einstein and if it is assumed that the excess energy is equipartitioned statistically according to the number of internal degrees of freedom in the fragments then the methyl radicals carry on the average 17 kcal/mol and the (CH3),Siradicals 93 kcal/mol above their thermal energy. The propensity of these energized species for hydrogen abstraction should be considerably greater than for their thermalized states at room temperature especially in the case of trimethylsilyl. The methyl radicals from primary step 2 and the hydrogen atom from step 4 carry little excess energy and therefore should exhibit normal reactivity. The overall mechanism may be further complicated by the occurrence of disproportionation of trimethylsilyl radicals with other radicals present in the system R- + (CH3)&3i RH + CH2Si(CH3)2 The combined loss of (CH3),Si through this step and through polymer formation is 0.10. The driving force behind this reaction is the formation of the silaethylene structure. Silaethylenes have been postulated to be involved in numerous thermal and photochemical reactions of silicon h y d r i d e ~ ~ - ~ and , ~ - "shown by molecular orbital calculations to have a planar singlet ground state with an olefinic double For the room temperature dis-
-
+
--+
+
-+
-
+ + + +
3002
The Journal of Physical Chemistry, Vol. 83, No. 24, 7979
proportionation to combination ratio of trimethylsilyl radicals generated by the triplet mercury photosensitization of TMS7 and by the photolysis of bis(trimethylsily1)mercury17in the vapor phase, a value of -0.046 was obtained which is considerably smaller than the value for the tertbutyl radicals, 2.3-3.1.33 The k d / k , value for the cross disproportionation CH3 + (CH3)3C CH4 CHz=C(CH3)2
-
+
is 0.85" and the value assumed here for the CH3 + (CH3)3Si CH4 + CHZSi(CH3)' reaction is 0.1. As seen from the results in Table IV, inclusion of this step in the mechanism has a tendency to alter the ratio of primary steps 1 and 2 in favor of the latter. Dimethylsilylene being unable to attack TMS*% by and large ends up in the polymer and could presumably also undergo cross disproportionation R + CH3SiCH3 RH CHzSiCH3
-
+
to yield the silavinyl radical. This species still could be a relatively good hydrogen donor by forming silaallene: R + CHzSiCH3 RH + CHzSiCHz
-
Also, dimethylsilaethylene, like isobutene, could be an efficient hydrogen donor by forming the resonance stabilized silaallyl radical R + CH2Si(CH3)' RH + CH2SiCH3CH2
-
The fate of the trimethylsilylmethyl radical in the system is not understood and the bulk of it seems to contribute to the formation of the polymer. The trimethylsilylmethylene formed in primary step 5 cannot be accounted for either, but its yield is small and it could be the precursor of one of the trace products which was tentatively identified as tetramethyldisilahexane. Comparing the results from the 147-nm photolysis of the three methylated silanes, monomethyl-, dimethyl-, and tetramethylsilane, we can see a more or less gradual change in the pattern of decomposition. Quantum yields for carbon-silicon bond cleavage increase from 0.36 in monomethyl to 0.56-0.64 in dimethyl to 0.98 in tetramethylsilane and, for the molecular mode of decomposition, they decrease from 0.63-0.70 to 0.57-0.70 to 0.23, respectively. The quantum yields of two fragment processes change from 0.56 to 0.58 to 0.66-0.74 with increasing methylation of the silane molecule. These changes can be correlated with the size, the ratio of the Si-H to Si-C bonds present, and with the spectroscopic properties of the molecule. The far-UV spectra of methylsilanes have been investigated recently by Sandorfy et al.39 The lowest lying excited species all have mixed Rydberg-valence shell character and correspond to transitions to 4s, 4p, and 3d orbitals. Sandorfy40 suggested that Rydberg and valence shell excited states exhibit different reactivities because Rydberg orbitals are largely nonbonding while valence shell excited orbitals are typically antibonding in one or more bonds in the molecule. As a consequence, if no crossing between states of these two different types takes place, valence shell excited orbitals favor simple bond cleavages whereas Rydberg orbitals lead mainly to molecular mode of decomposition. The spectra of silanes as those of paraffins can be divided into two categories, the round field for molecules whose spectra conform to approximate tetrahedral symmetry like tetramethylsilane, and long field, like monomethyl and dimethyl silanes. The significance of this in the present context is that the spectrum of TMS exhibits a
Gammie, Sandorfy, and Strausz
pronounced minimum around 150 nm while no such minimum is apparent in the spectra of monomethyl or dimethylsilane. In all three molecules 147-nm excitation can lead to promotion of an electron from one of the highest lying ground state orbitals with ionization potential differences of several tenths of an electronvolt to either the 4s,4p, or 3d Rydberg orbitals having various degrees of Rydberg character. Therefore decomposition may occur simultaneously from more than one electronic state leading to different reaction channels. This could explain the multitude of primary steps. From accumulated experience some general rules can be deduced for the photolysis of methylsilanes which can be summarized as follows: (a) The importance of Si-C cleavage increases with the number of Si-C bonds present in the molecule. (b) Loss of hydrogen occurs preferentially from the silicon atom and from the carbon atom it' is always of minor importance. (c) Geminal elimination of molecular hydrogen from the silicon atom is suppressed and vicinal elimination from the Si and C atoms is favored with increasing numbers of methyl groups. (d) In molecular elimination of methane, hydrogen is preferentially abstracted from the silicon atom. (e) Molecular elimination of ethane does not occur. (f) As a consequence of b, d, and e molecular elimination in tetraniethylsilane is less important than in mono- or dimethylsilane. (g) The two-fragment mode of decomposition is favored with increasing size of the molecule owing to the larger number of internal degrees of freedom and resultant longer lifetime of the vibrationally excited primary fragments. Turning to the inhibited reaction, nitric oxide has been shown previously to react with silyl radicals7 via an inverse addition H3Si + NO H3SiON
-
-
The resultant radical may undergo disproportionation 2H3SiON 2H3Si0 + N2 combination 2H3SiON
+M
-
H3SiON=NOSiH3
+M
-
or can react with nitric oxide41 H3SiON + NO H3Si0 + NzO The resultant siloxy radical can attack the Si-H bond by hydrogen displacement7 H3Si0 + SiH4 H3SiOSiH3+ H
-
-
which then leads to a chain process H + SiH4 Hz + SiH, The major products of the chain are the disiloxane, hydrogen, nitrogen, and nitrous oxide. An alternate displacement step has also been postulated involving molecular hydrogen elimination41 where instead of hydrogen atoms the disiloxyl radical is the chain carrier, according to the sequence H3Si0 + SiH4 H3SiOSiH2+ Hz H3SiOSiH2+ NO H3SiOSiH20N,etc
--
In the present study the major silicon product of the nitric oxide inhibited reaction was hexamethyldisiloxane indicating that elimination of a methyl radical from TMS is taking place.
Photochemistry of Silicon Compounds
(CH3)3Si0t TMS
-+
HMDSO t CH3
This has not been observed before. The four other unidentified products detected and designated in Table I1 as A, B, C, and D were not identified. C and D were thought to be trisiloxanes, A, trimethylhydroxylamine, a known product of the reaction of CH3 with NO, and B, methoxytrimethylsilane. The formation of HMDSO clearly shows that the trimethylsilyl radical is a principal intermediate in the photolysis of TMS and, unlike disiloxanes from the reactions of not fully methylated silanes, cannot be produced in a chain reaction. Therefore from the combined yield of HMDSO and the two trisilanes the approximate value of the quantum yield of trimethylsilyl radicals can be estimated to be -0.45 in agreement with the 4B value predicted from the mechanism (Table IV). In regard to the low observed yields of trimethylsilyl radicals in uninhibited photolysis it should be pointed out that trimethylsilyl radicals add to ethylene over 102-fold faster than methyl radicals42whereas their combination rate constant is about a factor of 4 smaller than that of methyl radicals, this suggests a loss mechanism involving addition to dimethylsilaethylene (CH3)&4i+ CH2Si(CHJ2 which ultimately can lead to polymer formation. From the results presented in this article, it is seen that the 147-nm photolysis of TMS leads to similar primary steps as was found in its 50-feV X-ray radiolysis12 and it also bears close resemblance to the 147-nm photolysis of ne0pentane.4~The main processes in the latter system are molecular methane elimination, and CH3 t H and 2CH, elimination. Interestingly, the elimination of CH4 is significantly higher here than with TMS, comprising 75% of the methane formed.
Acknowledgment. The authors are indebted to the National Research Council of Canada for continuing financial support, to Dr. A. Hogg for the mass spectrometric analysis, and Drs. E. M. Lown and I. Safarik for helpful suggestions.
References and Notes (1) L. Gammie, Ph.D. Thesis, Univeristy of Alberta, Edmonton, Alberta, 1975. Buffalo Pound Filtratlon Plant, P.O. Box 1587, Moose Jaw, Saskatchewan. The material of this paper was presented in part at the CIC Meeting, Regina, 1974 and in full at the CIC Meeting, Toronto, 1975. (2) Ddpartement de Chimle, Universitd de Montreal. (3) A. 0. Alexander and 0. P. Shausz, J. Phys. Chem., 80, 2531 (1976). (4) 0. P. Strausz, K. Obi, and W. K. Duholke, J. Am. Chem. Soc., 90, 1359 (1968).
The Journal of Physical Chemistry, Vol. 83, No. 24, 1979 3083
(5) K. Obi, A. Clement, H. E. Gunning, and 0. P. Strausz, J . Am. Chem.
SOC., 91, 1622 (1969). (6) A. G. Alexander, K. Obi, R. Roberge, C. Sandorfy, and 0. P. Strausz, to be submitted for publication. (7) M. A. Nay, G. N. C. Woodall, 0. P. Strausz, and H. E. Gunning, J . Am. Chem. Soc., 87, 179 (1965). (8) L. E. Gusel'nikov, N. S. Nametkin, and V. V. Vdovin, Russ. Chem. Rev., 43, 620 (1972). (9) L. E. Gusel'nikov, N. S. Nametkin, and V. V. Vdovin, Acc. Chem. Res., 8, 18 (1975). IO) P. Jutzi, Angew. Chem., 14, 232 (1975). 11) C. M. Golino, R. D. Bush, P. On, and L. H. Sommer, J . Am. Chem. Soc., 97, 1957 (1975). 12) G. J. Mains and J. Dedinas, J. Phys. Chem., 74, 3476 (1970). 13) R. A. Shaw, Pure Appl. Chem., 13, 297 (1966). 14) R. P. Clifford, 6.G. Gowenlock, C. A. F. Johnson, and J. Stevenson, J . Organometal. Chem., 34, 53 (1972). (15) L. F. Loucks and R. J. Cvetanovic, J . Chem. Phys., 58, 321 (1972). (16) T. G. Slanger, R. L. Sharpless, and G. Black, J . Chem. Phys., 61, 5022 (1974). (17) L. Gammie, I. Safarik, and 0. P. Strausz, submitted for publication. (18) H. Okabe and J. R. McNesby, J . Chem. Phys., 34, 668 (1961). (19) K. Dees and D. W. Setzer, J . Chem. Phys., 49, 1193 (1968). (20) N. L. Arthur and T. N. Bell, Rev. Chem. Intermediates, 2, 37 (1978), and references therein. (21) F. C. James and J. P. Simons, Int. J . Chem. Kinet., 8, 887 (1974), and references therein. (22) P. Ausloos and S. G. Lias in "Chemical Spectroscopy and Photochemistry in the Vacuum-Ultraviolet", C. Sandorfy, P. J. Ausloos, and M. B. Robin, Ed., Reidel, Dordrecht, Holland, 1974, p 465. (23) R. R. Baker, R. R. Baldwin, and R. W. Walker, Combust. Flame, 27, 147 (1976). (24) M. A. Contineau, D. Mihelcic, R. N. Schindler, and P. Potzinger, Ber. Bunsenges. Phys. Chem., 75, 426 (1971). (25) R. J. Norstrom, 0. P. Strausz, and H. E. Gunning, Can. J . Chem., 42, 2140 (1964). (26) M. D. Curtls, J . Organometal. Chem., 60, 63 (1973). (27) R. Damrauer and D. R. Williams, J . Organometal. Chem., 66, 241 (1974). (28) R. Walsh, J . Organometal. Chem., 38, 245 (1972). (29) M.J. S. Dewar, D. H. Lo, and C. A. Ramsden, J. Am. Chem. Soc., 97, 311 (1975). (30) 0. P. Strausz, L. Gammie, G. Theodorakopoulos, P. G. Mezey, and I. G. Csizmadia, J . Am. Chem. Soc., 98, 1622 (1976). (31) 0. P. Strausz, M. A. Robb, G. Theodorakopoulos, P. G. Mezey, and I.G. Csizmadia, Chem. Phys. Lett., 48, 162 (1977). (32) D. M. Hood and H. F. Schaefer, 111, J . Chem. Phys., 68, 298 (1978). (33) M. J. Gibian and R. C. Corley, Chem. Rev., 73, 441 (1973). (34) B. Cox and H. Purnell, J . Chem. Soc., Faraday Trans. 7 , 71, 859 (1975). (35) M. D. Sefcik and M. A. Ring, J. Am. Chem. Soc., 95, 5168 (1973). (36) P. John and J. H. Purnell, J . Chem. Soc., Faraday Trans. 7 , 69, 1455 (1973). (37) 0. F. Zeck, Y. Y. Su, G. P. Gennaro, and Y. N. Tang, J. Am. Chem. Soc., 96, 5967 (1974). (38) P. S. Skell and P. W. Owen, J . Am. Chem. Soc., 94, 5434 (1972). (39) R. Roberge, C. Sandorfy, J. I. Matthews, and 0. P. Strausz, J. Chem. Phys., 69, 5105 (1978). (40) C. Sandorfy, J . Mol. Struct., 19, 183(1973); Z. Phys. Chem., 101, 307 (1976); in "Chemical Spectroscopy and Photochemistry in the Vacuum-Ultraviolet", C.Sandorfy, P. J. Ausloos, and M. B. Robin, Ed., Reidel, Dordrecht, Holland, 1974. (41) E. Kamaratos and F. W. Lampe, J. Phys. Chem., 74, 2267 (1970). (42) K. Y. Choo and P. Gaspar, J . Am. Chem. SOC.,96, 1284 (1974). (43) S. G. Lias and P. Ausloos, J . Chem. Phys., 43, 2748 (1965).