Bond
Dissociatian
Energies of Silicon-Hydrogen Bonds
705
issociation Energies of Silicon-H ydrogen silanes as Estimated from bstraction Yields' Akio Hosaka and F. S. Rowland" Deparfment of Chemistry, University of California. lrvine, California 92664 (Received October 78, 7972) Publication costs assisted by the Division ot Research,
U.S. Atomic Energy Commission
The relative DT yields per Si-D bond for the abstraction of D by recoil tritium atoms have been measured as 4360 for (CH3)3SiD, 3790 for (CH&SiD2, and 3680 for CH3SiD3, as compared with yields per 6-D bond of 2380 for (CH3)sCD and 1190 for CH3CD3. The substantially higher yields from Si-D bonds than from C-D bonds suggest that all Si-D bond dissociation energies are appreciably less than 90 kcal/ mol. The Si-D bond dissociation energies in CH3SiD3 and (CH3)ZSiDz are each estimated to be approximately 2-3 kcal/mol higher than that in (CH&SiD. These differences are substantially smaller than those found with the corresponding C-D bonds in alkanes.
Introduction Energetic t r tium atoms are capable of abstracting hydrogen atoms from all carbon-hydrogen positions, as in ( I ) , and the yields of these reactions with alkanes and cyclanes correlate quite well with the accepted bond dissociation energies of the corresponding C-H bond^.^,^ Measurements with several amines have also shown a similar correlation for N-H bonds.4 While studies with CHF3 and CH3CD=CD2 have indicated that important deviations from the C-H correlation are found in comparisons involving rather dissimilar residual R radical^,^ ,6 comparison of the MT yields from molecules with similar residual radicals permits estimation of the relative bond dissociation energies involved. In connection with studies of all of the reactions of recoil tritium atoms with methylsilanes,7,8 we have measured the abstraction yields from several kinds of silicon-hydrogen bonds. For more accurate comparative measurements, we have now carried out abstraction experiments in the presence of excess moderators, using both CzH4 and NZin this role.
T* f R H - H T + R
(1)
The abstraction contributions from carbon-hydrogen and silicon--hydrogen bonds in the methylsilanes have been separated through the utilization as substrates of the partially deuterated compounds (CH3)3SiD, (CH3)zSiD2, and CH3SiD3, and measurement of the yield of DT, as illustrated for (CH3)2SiD2 in reaction 2. These measurements confirm that abstraction from silicon-hydrogen bonds occurs in higher yields than from most hydrocarbon C-H bonds. If the plausible assumption is made that abstraction from silicon-hydrogen and carbon-hydrogen positions occurs by Eimilar mechanisms, the higher yields suggest that the silicon-hydrogen bonds are normally weaker than citrbori-hydrogen bonds. However, much less variation in DT yield, and therefore in apparent bond strength, i s observed among the primary, secondary, and tertiary silicon-hydrogen bonds than among the corresponding carbon-hydrogen bonds.
Experimental Section Chemicals. The three partially deuterated methylsilanes were synthesized by the reaction of LiAlD4 with the corresponding methylsilyl c h l o r i d e ~ . ~The J ~ LiAlD4 was obtained from Merck Sharpe and Dohme and had a stated isotopic purity of 299% D; the isotopic purity of the Si-D positions was not checked. Each of the methylsilanes was purified by preparative gas chromatography using silicone oil columns. Preparation and Irradiation of Samples The samples were filled following the standard techniques used in our laboratories for gas-phase recoil tritium experiments, relying on neutron irradiation of 3He for the formation of the energetic tritium atoms.2-s Each sample bulb was about 12 ml in volume, and contained 10-18 Torr of 3He, 8-22 Torr of 0 2 , about 50 Torr of methylsilane, and about 700 Torr of CzH4 or N2. Groups of samples were irradiated simultaneously in the rotating specimen rack of a Triga nuclear reactor for about 12 min in a flux of 1012 n/cm2/sec. Complete details are given in ref 8. Radio Gas Chromatography. The isotopic hydrogen molecules were separated with a 1 2 4 activated alumina-Chromosorb P column operated at -196", and assayed by routine proportional counting procedures.ll ,I2 The macroThis research was supported by A.E.C. Contract No. AT(04-3)-34. Agreement No, 126, and constituted part of the work submitted by A. Hosaka in partial fulfillment of the requirements for the Ph.D. degree at the University of California, Irvine. J. W. Root, W. Breckenridge, and F. S. Rowland, J. Chem. Phys., 43, 3694 (1965). E. Tachikawa and F. S. Rowland, J. Amer. Chem. Soc., 90, 4767 11968). ?. Tominagaand F. S. Rowland,J. Phys. Chem., 72, 1399 (1968). E. Tachikawa, Y.-N. Tang, and F. S. Rowland, J. Amer. Chem. SOC., 90, 3584 (1968). E. Tachikawa and F. S. Rowland, J. Amer Chem. SOC., 91, 559 1\ .'1-QRQ\ - I I .
T. Tominaga, A. Hosaka, and F. S. Rowland, J. Phys. Chem.. 73, 465 11969). A. Hbsaka, Ph.D. Thesis, University of California, Irvine, 1971. A. E. Finholt, A. G. Bond, K. E. Wilzbach, and H. I. Schlesinger, J. Amer. Chem. SOC.,69,2692 (1947) C. Eaborn, "Organosilicon Compounds," Academic Press, New Vork - .., N.. V . , 1960 - - .. J. K. Lee, E. K . C. Lee, E. Musgrave, Y.-N. Tang, J. W. Root, and F. S. Rowland. Anal. Chem.. 34. 741 119621. E. Tachikawa,'Ph.D. Thesis, University of daiifornia, Irvine, 1967. The Journal of Physical Chemistry, Vol. 77, No. 5, 1973
Akio Hosaka and F. S. Rowland scopic content of 3He was measured for each sample through the difference in thermal conductivity between 3He and the 4He of the flow gas. The yields of H T and DT have all been normalized per unit amount of 3He, and are thus based on trequential thermal conductivity and radioactivity measurements on the same gaseous aliquot of each sample.
Results and Discussic.~rm LIT Yields in Excess CzH4. Since the absolute yields of hot reactions are dependent upon both the reactivity of the particulair bond involved and upon the specific energy loss processes in each individual molecular system, we have followed our standard procedure of making these comparisons in the presence of a large excess of some moderator-scavenger molecule combination.2-6 The data for DT yields from the methylsilanes are given in Table I for systems in which CzHa is the moderator, and for which both CzH4 and 0 2 serve as scavenger molecules for removing thermal atomo and radicals. Although H T was also routinely measured, its primary source in these mixtures lies in abstraction from CzH4 and the yield for abstraction from the CH3 groups in the methylsilanes represents a small, very impseciseiy measured increment to the HT formed by reaction with the moderator. For two of the methylsilane samples, the 0 2 concentration was reduced by about B factor of 2, with a measurable increase in the observed DT yield as a result. This sensitivity of yield to the precise 0 2 scavenger concentration is typical of very weak bonds from which abstraction occurs quite readily. In the comparisons made in Table I and Figure 1, these low 0 2 concentration experiments have been excluded. TABLE I: Specific Yields of DT from Recoil Tritium Reactions with Partially Deuterated Methylsilanes in Excess C2H4 Specific DS activity Substratea Per mcjiecule ________--.
(GH3)3SiD
(CH3)zSiDz
(CW3CD
c-C4Ds CH3CD3
4,290 f 4,430f 8,090f 7,588rb
70 80 80b 60 i1, m c f 906 71,13Q& 90 18,82C1i: 80 11,%10+110 2,4lC1f 90 2,350 f 80 12,530+ 90 3,58Ct 3~ 60
Per bond
* ”}
4290 4430 f 80 4050 f40 3790 f 3 0 3900 f 30 :37IO f 3 0 :3610 f 30 3730f40 :?410 f 90) 2350 f 8 0 1570 f20 1190 f 20
Average yield per Si-D bond
4360 f 80 3790f30
2380 1570f20 1190 f 20
aStandard sample pressures (Torr): 3He, 18; 02,15; RH, 50; C2H4, 700. *These samples contabned oniy 8 Torr of 02 and are not included in the averages.
NormalizaLioii points were measured for several deuterated hydrocarbons, irradiated simultaneously and measured under closely normalized conditions: CH3CD3 for primary C-D; c-C+l>a for secondary C-D; and (CH3)sCD for tertiary C-D. The relative yields of DT per bond summarized in Table I show that the DT yields per bond from Si-D bonds are all much higher than from any of the C-D positions. Further, the relative increase in DT yield from progressively more substituted central silicon atoms is The Journal of Phys,’ca/ Chemistry, Yo/. 77, No. 5, 1973
a -1
I
i
w>. 3,000 I-
a
0 !k
y2,ooo
-
a v)
W
L
4 1,000W IY
C-D
or Si-D BOND DISSOCIATION ENERGY (e.v.1
Figure 1. Relative specific DT yields per bond vs. C-D or Si-D bond dissociation energy. All deuterated bond dissociation energies assumed as 0.12 eV higher than the Corresponding Si-H or C-H bond dissociation energies. Si-D in (CH3)3SiDfrom ref 16.
Measured DT yields are indicated for (CH3)2SiD2 and CH3SiD3 for which no bond dissociation energy values are available.
markedly less than for the comparison hydrocarbons (even the absolute increments in yield are less for Si-D than for C-D), Previous experiments have established a correlation curve between H T yields for abstraction from RH in CzD4, as well as the fact that deuterated positions routinely give lower yields for abstraction than the corresponding protonated positions (e.g., HT/DT from CH2D2 = 1.32 f 0.O1).l3 Under our standard conditions, the DT yields from C-D positions show essentially the same correlation found in the RH-CzD4 systems but shifted to higher energies as shown in Figure 1. The data of Table I have been graphed us. the bond dissociation energies of the individual C-D and Si-D bonds in Figure 1. Very few C-D bond dissociation energies have actually been measured, so that data have been plotted us. the accepted bond dissociation energies, with an estimated correction for d e ~ t e r a t i 0 n . l ~ The bond dissociation energies of DC1, DBr, and DI are 0.139, 0.126, and 0.108 eV higher than those for the respective protonated halides, and the few hydrocarbon values seem consistent with a spread of this magnitude. Accordingly, we have adopted 0.12 eV (2.8 kcal/mol) as an approximate added correction for calculation of the bond dissociation energies of the deuterated hydrocarbons from the best estimates for the corresponding protonated species. It should be noted that the yields for abstraction from C-D bonds fit rather well to the curve established for abstraction from C-H bonds (lower DT yields and higher bond dissociation (13) J . W. Root and F. S. Rowland, J. Amer, Chem. Soc., 85, 1021 (1963). (14) J . A. Kerr, Chem. Rev., 66, 465 (1966). (15) See J. W. Root, Ph.D. Thesis, Universityof Kansas, 1964.
Bond Oissocnation Energies of Silicon-Hydrogen Bonds
7 07
energies). The implication is that there is very little isotopic influence c1n the mechanistic process of abstraction other than those differences already reflected in the increased bond dissociation energies of the deuterated species. Substantial disagreement exists concerning the bond dissociation energies of Si-H (and Si-D) bonds, with 216 to values for Si-‘I-I in (CH3)3SiH ranging from 81 8 8 l 7 kcal/mol, while the more recent values for Si-H in SiH4 are 94 3 Is and 95 kcal/mol. The concave upward curve through the hydrocarbon data in Figure 1 would fit smoothly through a value of about 87-88 kcal/ mol for Si--D in (CI13)3SiD, corresponding to about 85 kcal/mol for Si-H in (CH3)aSiH (a value midway between the two discrepant literature numbers for the Si-H bond). However, no information exists concerning the relative yields of D?‘ to be expected from (hypothetical) C-D, Si-D, N-D, or 8-D bonds of the same bond dissociation energies. Therefore, we do not know how to extend a hydrocarbon curve to f i t silicon-hydrogen bond yield data, and can make no legitimate estimate of the absolute bond dissociation energies of thew Si-D bonds. Estimates o!’ relative bond dissociation energies can be obtained by asmrning that the slope of the yield us. bond dissociation energy curve for Si-D will be similar to that for C-D bonds. The general uncertainty in this slope leads to the estimate that the Si-D bonds in (CH3)zSiDz and CH3SiD3 are 1.5-2.0 and 1.8-2.4 kcal/mol stronger, respectively, than in (CH3)3SiD. The total spread in Si-D bond dissociation energies from tertiary to primary is indicated as