SilacyclobutaneImine Benzene (mmol) (mmol) (mmol) ,---------Products (% yield^-------(CHz)&Me2 (21.2) PhCH=NPh (14.5) (59) (MezSiNPh)z (41) P h C H d H z (32) (CHz)3SiMe2 (13.2) Ph2CbNPh (5.20) ( 5 5 ) (MezSiNPhX (30) PhzCbCHz (56) PhzC=CHZ (33)b (CH&SiPhz (6.33) PhzC=NPh (3.03) (62) (PhzSiNPh)z (25)*+ aThese are based on the amounts of material actually isolated (either by preparative glpc or crystallization) and are probably 10-207$ lower than the quantities actually present. All reaction products are known compounds and, except as noted, were completely characterized by their ir and nmr spectra and, where applicable, melting points. * Calculated after correction for a 24% recovery of limiting reagent. Material was impure (mp 335-343", lit.6 mp 355.5"). Further recrystallization did not afford a purer sample.benzophenone (4.73 mmol), and benzene (66.7 mmol) afforded a homogeneous, yellow pyrolysate which, after resolution by preparative glpc afforded benzophenone phenylimine (0.79 mmol), styrene (0.78 mmol), 1,ldiphenylethylene (1.95 mmol), octamethylcyclotetrasiloxane (0.04 mmol), hexamethylcyclotrisiloxane (trace), benzophenone (1.74 mmol), and N-phenylbenzaldimine (12.7 mmol). Glpc analysis of the pyrolysate did not detect any 1,3-diphenyl-2,2,4,4tetramethylcyclodisilazane, (MezSiNPh)2. Isolation of benzophenone phenylimine provides definitive evidence for the formation of MezSi=NPh and its reaction with benzophenone since pyrolysis of equimolar quantities of either benzophenone and N-phenylbenzaldimine or benzophenone and (Me2SiNPh)z did not produce detectable quantities of this material. An explanation which accommodates all of these observations is outlined in Scheme 11. Assuming quantitative formation Scheme I1 --t [Me2Si=CH?I + CH?=CH? 3 G1l " Me,Si PhN=CHPh/ 1ph2C-0 + Ph,C=CH, + [Me2Si=O] PhCH =CH, + 1 [MelSi=NPh] -x+ (Me,SiNPh)? I Ph,C=Pu'Ph + [Me,Si=Ol of Me&=CHz and correcting, on the basis of the isolated yield of 1, 1-diphenylethylene, for the fact that 38.5 % of the MezSi=CH2 produced reacted with benzophenone, then the 0.79 mmol of benzophenone phenylimine isolated represents a 25.4 % yield of trappedMezSi=NPh. This is a minimum value since isolated yields were used and factors such as decomposition of the trapping reagents and products and other side reactions were ignored. The absence of significant side reactions and the relatively high yield of benzophenone phenylimine indicate that reaction schemes more complex than I1 are unlikely.The absence of (Me2SiNPh)z and the comparable quantities of styrene and benzophenone phenylimine observed indicate that benzophenone is a very efficient trapping reagent for silicon-nitrogen doubly bonded intermediates. This is in accord with our observationle that benzophenone is an excellent trapping reagent for intermediates containing silicon-carbon double bonds.Acknowledgment.