4241
which may be prepared in 7 G 9 0 % yield by reduction with LiAIH4: (C6H6),C4Si(H)CsH5 (VII), mp 198199", vsiH 2120 cm-l; (C6H5)4C4Si(H)CH3(VIII), mp 223-224", vsiH 2122 cm-I; (C6H&C4GeH2(IX), mp 192-193", V Q H 2060 cm-I; (C6H&C4Ge(H)C6H5 G. E. Ryschkewitsch, James M. Garrett (X), mp 187-188", Y Q H 2056 cm-I. Department of Chemistry, Uniaersity of Florida When wbutyllithium is added to X in T H F at -78", Gainescille, Florida 32601 a bright red color is produced. The addition of water Receiced Jurie 1 2 , 1967 to this red solution regenerates the starting hydride X. If trimethylchlorosilane is added in place of water, the silyl derivative, (C6Hj),C4Ge(C6Hj)Si(CH3)3, mp 178Functionally Substituted Sila- and 180°, vSiCH3 1250 and 850 cm-I, 7SiCH3 9.77, is formed Germacyclopentadienes. Anion Formation in 90% yield. Thus, the bright red color is ascribed to the germyllithium compound, (C6HJgC4Ge(Li)C6Hj Sir : (XII). There has been much interest in whether anions (11) On the other hand, when the silicon hydrides (VI1 and derived from sila- or germacyclopentadienes (I) would VIII) are treated with n-butyllithium in the same manbe stabilized by resonance since these anions have six ner, a very intense, dark purple color is generated as was .TT electrons and the resonance extremes would involve observed by Riihlmann.' The addition of water or DzO a silicon or germanium double bond to carbon. Riihlt o these solutions does not regenerate the starting hydrides (or deuterides) although the intense color is disR R\ R pelled. Nmr spectra of the resins obtained after removal of the T H F show broad unresolved absorptions Rin the region 7 10-7.7, and the area ratios indicate the presence of butyl groups in the resins. Furthermore, the splitting patterns of the H 2 0 - and DrO-quenched I I1 reaction products are different, suggesting that H or M =Si, Ge D is incorporated in the resins. N o Si-H or Si-D stretching frequencies are observed in the infrared specmannl recently prepared pentaphenylsilole (I, M = Si; tra of the products, so the hydrogen or deuterium must R = R ' = C6H6; Y = H) and tetraphenylmethylsilole be attached to the carbon skeleton. Unfortunately, the (I, M = Si; R = C6H5; R ' = CH3; Y = H) by C-D stretching region is obscured by a weak overtone allowing 1,4-dilithiotetraphenylbutadiene2to react with or combination band. phenyl- or methyldichlorosilane. Riihlmann also reT o determine if a hydrogen atom attached t o the ported that treatment of these siloles with phenyllithium heteroatom was necessary to form these intensely gave intense red-violet colors, but he could not decide if colored solutions, M-butyllithium was added to the prethe color was due to an anion radical, a ring-addition viously unreported (C6H&CdGe(C6Hj)r, mp 198product, or the silacyclopentadienide anion (11). Since 199". An intense purple color formed and this soluthe organodichlorisilanes, RSiC12H, used by Riihlmann tion exhibited a very weak electron spin resonance are difficult to prepare and not readily available, a syn(esr) signal at a g value near 2. The extreme weakness thetic procedure employing the easily obtained RMCI3 of the signal shows that radicals are not the predominant (M = Si, Ge) derivatives was sought. Such a procedure species in solution. Hence, in this case it is concluded would also have the advantage that the d o l e or germole that the butyllithium adds to the ring structure t o give a obtained would have Y = halogen in the structure (I), conjugated C-Li derivative which is responsible for the allowing for a wide variety of substitutions at the heterointense colors observed. atom (M). The above results may be compared with the reactions Although attempts to prepare halogen-substituted of triphenylsilane and triphenylgermane with n-butylsiloles and germoles by the reaction of (C6H&C4Li2 lithium. Triphenylgermane readily forms triphenylwith chlorosilanes and -germanes have been reported germyllithium, but triphenylsilane forms n-butyltrito be unsuccessful, it has been found that recerse addiphenylsilane and lithium hydride.6 Thus, there exists tion of a diethyl ether suspension of 1,4-diIithiotetrathe possibility that (C6H5),CgSi(C6H5>(C,H9) (XIII) phenylbutadiene2 to a diethyl ether or tetrahydrofuran and LiH are formed when butyllithium is added to VII, (THF) solution of R'MC13 or MC14( M = Si, Ge) gives and that the LiH then adds across a carbon-carbon in 5@-70z yield the corresponding heterocycle. Some double bond t o form a highly colored C-Li derivative. of the new chloro-substituted siloles and germoles (I, This possibility was tested by adding crushed lithium R = CgHj) are: ( C ~ H ~ ) , C J S ~ ( C I ) C (III), ~ H ~mp 181hydride to a T H F solution of XIII, which was prepared 183"; (C6Hj),C4Si(C1)CH3(IV), mp 194-195"; (C6H&from 111 and butyllithium in benzene. N o reaction was C4GeC12 (V), mp 197- 199" ; (C6Hj)4C4Ge(C1)C6H5 observed after stirring for 2 hr at room temperature. (VI), mp 210-21 1 ". The chloride may be displaced by However, the addition of LiA1H4 to a T H F solution of many reagents, e.g., RO-, R-, and transition-metal XI11 immediately causes an intense purple color. It is carbonyl anionsS4 Of interest here are the hydrides possible that LiH formed in situ is sufficiently more re(1) K. Riihlmann, 2.Chem., 5, 354 (1965). active than bulk LiH (which has a very low solubility in (2) E. H. Braye, W. Hubel, and I. Caplier, J . Am. Chem. Soc., 83, THF) that a direct comparison cannot be made. 4406 (1961).
by meaiis of optically active transition metal complex ani on s , Acknowledgment. The research was supported in part by National Institute of Health Grant G M 13650.
(3) F. C. Leavitt, T. A. Manuel, F. Johnson, L. U. Matternas, and D. S . Lehman, ibid., 82, 5099 (1960). (4) M. D. Curtis, unpublished results.
( 5 ) H. Gilman and C. Gerow, J . Am. Chem. Soc., 78, 5435 (1956). (6) H. Gilman and H. W. Melvin, {bid.,71, 4050 (1949).
Communications to the Editor
4242
Although the nature of the species responsible for the intense colors has not been conclusively established, these results d o show that silacyclopentadienide anions are not formed from treatment of the hydrides with nbutyllithium at low temperatures. Further studies are i n progress to determine if the silacyclopentadienide anions can be prepared by other methods. The germanium hydride X does form the germyllithium derivative XII, but triphenylgermane also forms triphenylgermyllithiuni under the same conditions. It thus appears that the ring structure does not confer enhanced acidity to the hydrides (I, Y = H). M. David Curtis Deparmierii of Cliemisii,y, Nortliwesteix C'nicersity Ecorisroti, llliriois 60201 Receird M(ry 19, 1967
Aziridines. X. The Influence of Steric Factors on Nitrogen Inversion Rates and Ring Proton Shifts in Aziridines Sir: The study of nitrogen inyersion in aziridines has become an active field of interest in recent years. In many aziridines, the rates of nitrogen inversion are of suitable magnitude for study by nmr spectroscopy.' At cu. 25", the ring proton spectrum of I-alkylaziridines (I, R = Me, Et, i-Pr) gives rise t o an AzBz pattern indicating that the rate constant (ki) for the nitrogen inversion process (Ia $ Ib) is small compared to the chemical
In this communication we present new data concerning nitrogen inversion in 1-t-butylaziridine (11) and discuss the factors which govern the magnitude of vAB in 1-alkylaziridines. In contrast to the study of Bottini and Roberts,Ib we hake found that the ring protons in the 40- and 60Mc/sec spectrum of I1 as the neat liquid or a 5 0 x CCl, solution form f b t v peaks separated by -2 and 3 cps, respectively, at room temperature. The magnitude of vAB was found to be solvent and concentration dependent. For example, the proton spectrum of I1 as a 2 0 x benzene solution (w/v) exhibits an AzBz pattern with vAB = 11 cps at 34" and lower. On heating a benzene solution of I1 the ring proton spectrum broadens and then collapses to a single band at about 52", the coalescence t e m p e r a t ~ r e . ~The thermodynamic parameters for the inversion process in l-tbutylaziridine-2,2-& (Tc = 52") are currently being assessed from the deuterium-decoupled spectra in various solvents. The present study clearly demonstrates that steric factors d o indeed accelerate the nitrogen inversion process in 1-alkyla~iridines.~However, the rate of nitrogen inversion in the case of 1-t-butylaziridine (11) is far less than that prekiously claimed by Bottini and Roberts.'" Of particular relevance is the similarity of the present data and those reported for l-alkyl-2,2-dimethylaziridines (iii,lb III1c~e)and rrans-l-ethyl-2,3-dimethylaziri(IV). Such similarities strongly imply that the degree of steric assistance to the inversion process in these molecules is comparable.G
R'" Ia
CH,CH,
Ib
shift, vAB. At higher temperatures, the ring proton spectrum of I (R = Et), for example, broadens and then at -108" the coalescence temperature (T,) collapses t o a singlet. At T,, the inversion rate constant becomes equal to 2.22va4B, or approximately 60 sec-1.lb Replacing the ethyl group by methyl, benzyl, phenethyl, or cyclohexyl does not greatly affect the latter value. On the other hand, the bulky t-butyl group appears to increase the rate of nitrogen inversion dramatically. Thus, Bottini and Robertslb report that the 40-Mc/sec nmr spectrum of 1-t-butylaziridine (11) even at - 77" possesses only a single sharp line for the ring protons, "indicating that inversion occurs too rapidly for measurement." The lower inversion barrier was ascribed1" to nonbonded interactions between the ring protons and the t-butyl group in the ground state of 11.2 (1) (a) H. S. Gutowsky, A i m N . Y. Acad. Sci., 70, 786 (1958); (b) A. T. Bottini and J. D. Roberts, J . Am. Chetn. SOC.,80, 5203 (1958); (c) A. Loewenstein, J. F. Neumer, and J. D. Roberts, ibid., 82, 3599 (1960); (d) T. J. Bardos, V. Szantay, and C. I
