ERXEST L. ELIELAND MARE;X. RERICK
1362
[cOh7TRIBUTION FROM THE CHEMICAL
LABORATORIES O F THE USIVERSI‘rk‘
Vol. 82 OF N O T R E DAME
I
Reductions with Metal Hydrides. VIP. Reduction of Epoxides with Lithium Aluminum Hydride-Aluminum Chloride B‘i ERNEST L. ELIELASD 51:i1i1~lethyleneoxide a1 o has been reduced Trith L.1H but tetraplieri?-lethylene oxide was resistant t u reduction in boiling ether. :kidition of alui inurn cliloride to the LhI-I solution brings about complete reversal of the outcome of reduction in the case of a-methylstyrene oxide arid a-tiiisobutylene oxide but little or no reversal in the case of epichlorohydrin, a - p h e n ~ l - l , 2 - ~ p o s y p r o p a nand e 3-plieiiox~--1,2-cpor?-pro~ane.The “reversal” seems to involve a hydride shift to a carbonyl intermediate. 8-Diisobutylene oxide similarly undergoes a l-butyl shift and is reduced t o 2,2,3,3-tetrarnethylbutanol-1, and tetramethylethylene oxide undergoes a partial methyl shift being reduced to a mixture of pinacolyl alcohol and ditnethylisopropylcarbinol. Tetraphenylethylene oxide is again inert under these conditions and cyclohexenc oxide is reduced to cj-clohexanol.
In a previous publication’ the reduction of certain epoxides with lithium aluminum hydride (LAH) and with LhH-aluminum chloride in ether was described. T h e results of this study are summarized in Fig. 1. Reduction with LAH alone involved attack of hydride a t the least substituted carbon whereas the “mixed hydride”3 (LAHdlCl3) involved, in inost cases, reduction at the R‘ R ” R’R ”
’
RC-CH
1
1 1
01-1 H
LAH f--
1
1
RC-PIT
LA13 ___z
\
\I
K‘ K”
1%
~
RC-C=O
1
H
Kl
1 l
7
1Z
I,
+ RC--Cl-I 1
€I O H
Fig. 1.
iiiost highly substituted carbon, formally speaking. Tracer studies indicated, however, t h a t this “reversed reduction” actually involved a hydride shift producing a carbonyl intermediate which was then further reduced to alcohol. The “reversal” occurred readily with isobutylene oxide, trimethylethylene oxide, styrene oxide, 1,l-diphenylethylene oxide and triphenylethylene oxide but only to a niinor degree with propylene oxide. I t was the objective of t h e preseut work t o study iurtlier reductions of epoxides with LXH in the presence and absence of aluminum chloride so as to determine exactly when “reversal” might or might not be expected to occur. Included in this study mere tetrasubstituted ethylene oxides which cannot undergo a hydride shift to give t h e carbonyl intcrmediate show-li in Fig. 1, b u t which could conceivably undergo shift of ari alkyl group under the influence of the Lewis acid, aluminum chloride. to give, after recluctioii, carbinols of the type IZ3CCHOHR. Results of t h e present work and the earlier study? arc summarized in Tables I and 11. Table I lists redsctions with LXH alone. T h e following general conclusions may be drawn : ( I j Reduction (1) This paper is bdscrl on the 1’h.D. dissertation o f 11. N.lierick. (‘7) Pager V I , 13. I>.Eliel and 11. JV. Delmonte, TIIISJ O U R N A L , 80. 1714 ( 1 9 , j X ) . ( 3 ) Ipor a review < i f reilnctions with the “nii\ed hydride” 6.:. I\I. X , Rerick, “Selective Reductions of Organic Compounds w i t h Comples Ietal Hydrides. Inc., J3 Congress Street, Beverly. llas.;,, 1 0 . j Y . T h e exact nature 111 tho reagent remains to he elucidated In the present w o r k , the LAH..IICl: ratio wa.; 1.4.
of epoxides with LAH, if it succeeds at all, involves exclusive or nearly exclusive hydride attack on the less substituted side of the epoxide; cases of epoxide reduction in the literature4 agree with this conc l ~ s i o n . ~( 2 ) As the epoxide becomes more highly substituted, the yield of reduction product decreases and unchanged epoxide is recovered, this being the case with trimethylethylene oxide, triphenylethylerie oxide, tetramethylethylene oxide and P-diisobutylene oxide (I, Fig. 2). The lastnamed oxide, surprisii~gly,docs undergo reductioii to the extent of 21%, although attack a t a neopentyl carbon is involved; in contrast, tetraphenylethylene oxide does not undergo reduction” in ether a t 35’. The reduction of epichlorohydrin (Table I) is noteworthy in that i t was carried out with a limited amount of LAH using the inverse additioii technique (hydride s o l u h n added to compound j ; the main product, uLider these conditions, is 1chloro-2-propanol. Pre\-iously, epichlorohydrin had Seen reduced to propaxoi-2 using an excess of L h H 7 ; the present result indicates t h a t attack a t the terniinal position of t h e epoxide precedes reduction of the primary chloride.8 Table I1 sumniarizes the reductions of epoxides with LAH-aluminum chloride. Previous work2 has indicated that iil tile case of styrene oxide and isobutylene oxide, this type of reduction involves a hydride shift. I t ma]- Le presuined, therefore, t h a t a hydride shift such as is shown in Fig. 1 occurs (4) Summarized by (a) >:. G . Gaylord, “Reduction with Complex Afetal Hydrides,” Interacience Publishers, Inc., S e w York, S. Y., pp. 64G-tj7:3; a n d (17) Y, hl. AIicovic and >I. L. hfihailovic, ride i n Organic Chemistry,” Belgrade, k n o w n . R I . llousserun, R . Jacquier, M. Zagdoun, Bidl. SOC. chiin. F ~ a r z c e , 1042 !1932), claim t h e rediiction of ethylidenecycluliexane oxide t o mcthylcyclohexylcarbinol, b.11 A Uxgstahler, P h . D . Dissertation, Harvard University, Cambridge, 3I:rss , 1932, indicates the product to be the expected ? - etli~~lcyclohexan~1. T h c reduction of the 3-hydroxy-58,GPepoxychole~tanes t o :3..i3-~:ihy[Irosq-cholestanes reported by PI. A. ttner, r t ai., HC,!L,.C h i i n A c t a , 32, 5 8 i ( l Y 4 Y ) , and 37, 238 (19541, d uti cirnformutional grounds. i n which rcrliiction of highly substituted epoxide fails are l i i t e d in rei 4 b , 11. 7 € , ( 7 ) I,. W. T r e v o y :ind i?
