J. Am. Chem. SOC. zyxwvu 1985, 107, zyxwvu 4415-4419 4415 to spectral data prior to irradiation. Direct Photolysis of 3 at 185 zyxwvutsr nm. VPC-purified 3 (342 mg, 2.76 mmol) in 100 mL of pentane was irradiated in a reactor with a Suprasil immersion well and Osram Hg lamp. Aliquots were monitored by VPC (50-120 OC program) as a function of the time of irradiation. At 10% conversion or less, (Z)-3-methylhex-3-ene was the only photofragmen- tation product observed. However, at conversions higher than 10% both the Z and zyxwvutsrqponml E isomers are observed. After 3 h of irradiation (50% con- version of 3 by VPC), the photolysate was concentrated by careful dis- tillation in a water bath (40 "C). Photoproducts were isolated by prep- arative VPC and identified. Recovered starting material was examined by 300-MHz proton NMR to determine the extent of geometrical isom- erization of 3 to 2. The photolysis was calibrated by a 0.01 M pentane solution of (Z)-3-methylhex-3-ene. Direct Photolysis of 2 at 185 nm. VPC-purified 2 (205 mg, 1.65 mmol) in 70 mL of pentane was irradiated in a reactor with a Suprasil immersion well and Osram Hg lamp. Aliquots were monitored as a function of the time of irradiation by VPC (50-120 OC program). At 10% conversion or less, (E)-3-methylhex-3-ene was the only photofrag- mentation product observed. However, at conversions higher than IO%, both Z and E isomers were present. After 3 h of irradiation (50% conversion of 2 by VPC) the photolysate was concentrated by careful distillation in a water bath (40 "C). The photoproducts were isolated by preparative VPC and identified. Recovered starting material was examined by 300-MHz proton NMR to determine the extent of geo- metrical isomerization of 2 to 3. The photolysis was also calibrated with a 0.01 M pentane solution of (E)-3-methylhex-3-ene. Quantitative Photolysis of 2 at 185 nm. VPC-purified 2 (181 mg, 1.46 mmol) in 50 mL of pentane was irradiated in a reactor with a Suprasil immersion well and Osram Hg lamp. The photolysis was conducted in the same manner as the preparative photolysis at 185 nm. After cali- bration by cyclooctene actinometry," a quantum yield of 0.29 was cal- culated for the photoreaction of 2 at 185 nm. (This calculated quantum yield does not account for the geometrical isomerism that is shown to occur, and, therefore, the actual quantum yield would be slightly higher.) (17) Schuchmann, H. P.; von Sonntag, C.; Srinivasan, R. J. Photochem. 1981, zyxwvutsrqpon 15, 159. Photolysis Products. (Z)-3-Methylhex-3-ene (5) was identified by comparison of its IR, PMR, mass spectrum, and VPC retention time with that of an authentic sample (Wiley 99%). The authentic compound was also used to calibrate the quantitative photolysis of 2 for the photofrag- mentation process. (E)-3-methylhex-3-ene (4) was identified by comparison of its IR, PMR, mass spectrum, and VPC retention time with that of an authentic sample (Wiley 98%). 1-Isobutylidene-2-ethylcyclopropane (8 + 9) was identified by the following: IR 3025, 1770, 1452, 1372, 1180, 1067, 955 cm-I; PMR 6 2.03 (2 H, m), 1.70 (br s, 3 H), 1.5-0.2 (1 1 H); MS, m/z 67 (base), 41, 27, 39, 29, 53, 55, 81, 109, 95; mol wt by mass spectrum 124. syn-I-Ethyl-I-methyl-2-propylidenecyclopropane (7) was identified by the following: IR 3030, 1458, 1372, 1128, 1105, 1083, 1053 cm-l; PMR 6 5.55 (m, 1 H), 2.02 (2 H, m), 1.28 (2 H, m), 1.1-0.7 (11 H); MS, m/z 67 (base), 41, 27, 39, 29, 53, 55, 81, 109, 96; mol wt by mass spectrum 124. anti- ]-Ethyl- 1-methyl-2-propylidenecyclopropane (6) was identified by the following IR 3030, 1760 (w), 1455, 1372, 1287, 1052, 1000,978, 942 cm-l; PRM zyxwvu 6 5.64 (m, 1 H), 2.01 (2 H, m), 1.22 (2 H, m), 1.05-0.7 (11 H); MS, m/z 67 (base), 41, 27, 39, 29, 55, 53, 81, 109, 96; mol wt by mass spectrum 124, and from its predominance in the pyrolysis re- action. Pyrolysis. Pyrolysis of 3 at 198 OC. VPC-purified 3 (200 mg, 1.61 mmol) was divided into three Pyrex tubes. The tubes were evacuated (with liquid nitrogen cooling), sealed with an oxygen-propane torch, and immersed in a heated silicon oil bath (198 "C). The pyrolysis samples were monitored by VPC (5C-120 OC program) at 30, 60, and 90 min and compared to a t = 0 sample of 3. The relative product ratios were calculated by VPC and were invariant (within experimental error) for each time interval of the pyrolysis. Pyrolysis of 2 at 198 OC. VPC-purified 2 (193 mg, 1.56 mmol) was pyrolyzed exactly according to the procedure given above. Acknowledgment. This work was supported by the National Science Foundation through a University-Industry Cooperative Grant to the University of Connecticut with Dr. Gary Epling as principal investigator. A Theoretical Study of Fluorine Atom and Fluoride Ion Attack on Methane and Silane Larry P. Davis,*+Larry W. Burggraf,+ Mark S. Gordon,' and Kim K. Baldridge' Contribution from the Department zyxwvu of Chemistry, United States Air Force Academy, Colorado Springs, Colorado 80840-5791, and the Department of Chemistry, North Dakota State University, Fargo, North Dakota 58105-5516. Received November 15, 1984 Abstract: We have performed MNDO and ab initio calculations for reactions of fluorine atom attack on methane and silane and, in addition, fluoride ion attack on the same molecules in the case of MNDO. We modeled both substitution and abstraction reactions in each case. Results were compared with experimental data, where available. Comparisons show that MNDO usually does as well as the ab initio methods in reproducing experimental values for hE's of these reactions, but M N D O predicts activation barriers too high in most cases. Nevertheless, MNDO does qualitatively agree with the ab initio result that, while carbon undergoes abstraction much more easily than substitution, silicon can undergo either substitution or abstraction quite easily. The analysis in the case of fluoride ion attack on silane is complicated by the predicted ease of formation of a stable trigonal-bipyramidal intermediate. Silicon chemistry is increasingly important to many products of materials science including catalysts, semiconductors, orga- nosilicon polymers, ceramics, glasses, and composites. Our special interest is silica surfaces and their interactions with organic molecules. In all of these applications an efficient, accurate, theoretical model for predicting silicon chemistry could produce 'US. Air Force Academy. 'North Dakota State University. a great economy of effort. In particular, a detailed model of the mechanisms and products for silica polymerization would be of great help in tailoring silica materials for specific purposes. To this end, we began a combined theoretical and experimental program to study silicon reactions and silica surfaces several years ago. Our goal is to perform theoretical calculations for rather large molecular systems without resorting to theoretical models which are expensive, time-consuming, and, thus, self-defeating for our 0002-7863/85/1507-4415$01.50/0 0 1985 American Chemical Society