816
PADWA AND DAS
The Journal of Organic Chemistry
and after each run with the probe of a d u d thermistor bolometer (YSI Model 65) placed inside the cell. Fluxes of cu. 3 X l W ergs/cm2 sec were employed. Unsuccessful attempts to photosensitize the reaction were carried out on nitrogen-purged equimolar solutions of I a and naphthalene or triphenylene in tetrahydrofuran. Pyrex-filtered light from a 550-W medium-pressure mercury lamp (Hanovia, Type A) was used. 5-Phenylamino-n's-3-hexen-Z-ol(V) .--Zinc dust, 7.5 g (0.12 mol) , was added slowly to a stirred solut,ion of 4.5 g (0.024 mol) of Ib in 40 ml of acetic acid. The temperature rose to 50" and stirring was continued overnight. The suspension was made basic with 28 g of sodium hydroxide in 100 ml of water and extracted with three 50-ml portions of benzene. The dried extracts were evaporated under reduced pressure to give 3.5 g of liquid. This material was distilled (short path, 0.1 mm) to give a pale yellow oil which was pure by vpc analysis. The analytical sample was collected from a 3-ft 10% Carbowax column at 220". Anal. Calcd for C I ~ H I ~ N OC,: 75.35; H, 8.96; N, 7.32. Found: C, 75.54; H, 9.07; N, 7.22. Oxidation of V.-A suspension containing 1.0 g of V and 1.0 g of activated manganese dioxidels in 20 ml of benzene was re(25) E.F. Pratt and T. P. McGovern, J . Org. Chem., 49. 1540 (1964).
fluxed for 17.5 hr. A sample was removed, filtered, and analyzed with a 3-ft Carbowax column. In addition to unreacted starting material, 2,5-dimethyl-l-phenylpyrrole and aniline (14:l) were the only products observed. The same reaction occurred more slowly a t room temperature. Control experiments showed that aniline and 2,5dimethyl-l-phenylpyrrole were essentially unchanged under the oxidation conditions. Low-Temperature Photolysis of 3,6-Dimethyl-2-phenyl-3,6dihydro-1,Z-oxazine (Ib) .-A film of I b between sodium chloride plates was irradiated in an infrared cell cooled to -180' by liquid nitrogen. Light from an unfiltered 800-W Westinghouse mercury arc lamp was employed. After 25 min, a peak appeared at 1695 cm-1. On warming above -46', it disappeared. Cooling the sample back to -180" did not regenerate the 1695-cm-' peak.
Registry No.-Ia, 19029-45-9; Ib, 19029-46-0; IC, 19029-47-1 ; Id, 19029-48-2; Ie, 19029-49-3; V, 19029-97-1. Acknowledgment.-This work was supported in part by a grant (GM 14305) from the National Institute of General Medical Science, U. S. Public Health Service.
Oxirane Radicals. The Thermal Decomposition of t-Butyl cis- and trans- a,6-Diphenylperglycidates' ALBERTPADWA~ AND N. C. DAS Department of Chemistry, State University of New York at Buffalo, Buffalo, New York 14214 Received June 94, 1968 The di-tbutyl peroxide initiated decarbonylation reactions of 2,3-&- and -trans-diphenyl-2,3-epoxypropanal and the thermal decomposition of t-butyl cis- and trans-a,p-diphenylperglycidateshave been investigated as sources of isomeric oxirane radicals. The products obtained from the decarbonylation reactions are exclusively derived from opening of the oxirane ring. &Butyl cia- and trans-a18-diphenylprglycidateaundergo thermal decomposition in cumene a t 70" a t identical rates and by a nonconcerted one-bond cleavage. The products (desoxybenzoin, trans-stilbene oxide, and 1,2,3,4-tetraphenyl-l14-butanedione)formed are the same from either perester. The results indicate that the oxirane radical can abstract hydrogen but that rearrangement to the desyl radical is facile and accounts for the major proportion of products.
Theoretical arguments3t4 predict that the most favorable hybridization for an alkyl radical is sp3, although electron spin resonance investigations of methyl and other alkyl radicals are best interpreted by postulating a planar g e ~ r n e t r y . ~ -It~ seems that the energy difference between planar and pyramidal conformations of free radicals are much smaller than the energy differences between the similar geometries of carbanions and carbonium ions. Support for this contention stems from studies on the configurational stabilities of vinyl radicals which indicate that these radicals equilibrate with great facility.8-'2 In contrast to the above, (1) Epoxidation Studies. V. For IV, see A. Padwa and L. Hamilton, J . Org. Chem., 81, 1995 (1966). (2) Alfred P. Sloan Research Fellow, 1968-1970. (3) J. W.Linnett and A. J. Poe, Trana. Faraday Soc., 47, 1033 (1951). (4) A. D. Walsh, J . Chem. Soc., 2296 (1953). (5) H. M . McConnel and R. W. Fessenden, J . Chem. Phys., S i , 1688 (1959). (6) M. Karplus and G. K. Fraenkel, ibid., 86, 1312 (1961). (7) M. C. R. Symons, Aduan. Phys. Org. Chcm., 1, 283 (1963). (8) J. A. Kampmeier and R. M. Fantzier, J . Amer. Chcm. Soc., 86, 1959, 5219 (1966). (9) J. A. Kampmeier and G . Chen, ibid., ST, 2608 (1965). (10) L. A. Singer and N. P. Kong, Tetrahedron Lett., 2089 (1966);J . Amer. Chem. Soc., 88, 5213 (1966);89, 5251 (1967). (11) 0.Simma, K.Tokumaru, and H. Yui, Tetrahedron Lett., 5141 (1966). (12) G. D. Sargent and M. W. Browne, J . Amer. Chem. Soc., 89, 2788 1967).
an electron pair in a trigonally hybridized orbital seems to be able to maintain its spatial configuration to a greater extent.13 The stereochemical fate of an electron in a nonbonding orbital that is part of a three-membered ring has also been ~tudied.'~-'~Applequist concluded that the cyclopropyl radical is either planar or inverts its configuration very rapidly even though a large barrier, due to the extra I strain in the rehybridized state, exists.14 This was demonstrated by showing that in the Hunsdiecker degradation of both cis- and trans2-methylcyclopropanecarboxylic acid the identical mixture of cis- and trans-1-bromo-2-methylcyclopropane was obtained. Similar conclusions were reached by Walborsky from studies on the thermal decomposition of the diacyl peroxide of (+) (R)-l-methyl-2,2-diphenylcyclopropanecarboxylic acid in hydrogen donating solvents.15J6 The formation of optically inactive (13) D. Y. Curtin and J. W. Cramp, ibid., 80, 1822 (1958). (14) D. E.Applequist and A. H. Peterson. ibid., 89, 2372 (1960.) (15) H.M.Walborsky, Rec. Chem. Prom., 38, 75 (1962). (16) H. M.Walborsky and A. E. Young, J . Amer. Chem. SOC.,86, 3288 (1964). (17) H. M.Walborsky and C. J. Chen, ibdd., 89, 5499 (1967). (18) T. Ando, F. Namigata, H. Yamanaka, and W. Funaaaka, ibid., 89, 5719 (1967).
Vol. $4,No. 4,April 1969 l-methyl-2,2-diphenylcyclopropanewas taken to mean that the l-methyl-2,2-diphenylcyclopropy1radical was incapable of maintaining its configuration. When the decomposition of the diacyl peroxide was carried out in pure carbon tetrachloride a small amount of 1methyl-2,2-diphenylcyclopropanewas isolated. The observation that the cyclopropyl compound retained some of its activity suggested that the bent cyclopropyl radical was able to maintain its configuration to a significant extent in a, radical disproportionation within a solvent cage.14 Recently, Ando and coworkers reported that the stereospecificity observed in the reduction of some gem-halofluorocyclopropanes with trin-butyltin hydride can readily be interpreted by assuming that the fluorocyclopropyl radical is pyramidal and abstracts hydrogen from tri-n-butyltin hydride much more rapidly than it inverts its configuration.'* Although some related small ring heterocyclic systems have been studied in terms of radical phenomena,19g20little attention has been directed to the question of the stereochemical capabilities of such radicals. I n particular the possibility that the electron pair on heteroatom retards the inversion process due to an unfavorable dipoledipole interaction between the ?T orbital of the ra,dical and the electron pair of the heteroatom in the rehybridized state has not been studied. I n order to ascertain whether radicals that are localized on an epoxide ring have a short stereochemical memory of their origin, we have investigated the di-t-butyl peroxide initiated decarbonylation reactions of 2,3-cis- and -trans-diphenyl-2,3-epoxypropanal (I and 11) and the thermal decomposition of tbutyl cis- and trans-a,/?-diphenylperglycidate(I11 and IV). Both of these methods have been extensively employed to introduce a radical center a t a predetermined position in molecules in a host of free radical investigations.21 These compounds are convenient since the radicals may be formed a t relatively low temperatures, the lxoduct epoxides are stable under the reaction conditions, and the stereochemical results can readily be determined by analysis of the known cis- and trans-stilbene epoxides.
OXIRANERADICALS 817 SCHEME I ph
"XPh H CHO
alkaline
I
1
II
NaBH,
Ph
H