REARRANGEMENT OF EPOXY KETONES
Dec. 20, 1955
6525
ORGANIC AND BIOLOGICAL CHEMISTRY [CONTRIBUTION FROM THE
DEPARTMENT OF
CHEMISTRY,
MASSACHUSETTS ISSTITUTE
O F TECHNOLOGY]
The Rearrangement of a,P-Epoxy Ketones. 11. Migratory Aptitudes1 BY HERBERT0. HOUSEAND DONALD J. REIF RECEIVED JUNE 10, 1955 The boron trifluoride-catalyzed isomerizations of three a,P-epoxy ketone systems have been studied. 3,4-Epoxy-1phenyl-2-butanone yielded 2-phenyl-1,3-butanedione,a-methyl-P-phenylacrylophenone oxide yielded 1,2-diphenyl-1,3butanedione and a-phenyl-cis-benzalacetophenoneoxide yielded 1,2,2-triphenyl-1,3-propanedione.Under the same conditions or-phenyl-trans-benzalacetophenone oxide was isomerized to benzhydryl phenyl diketone. Possible interpretations of these results are discussed. R-C=O R The products obtained from the acid-catalyzed I \C-O isomerization of several substitued benzalaceto--C-&+m+ / 6 phenone oxides were best explained by the assump1 1 >c-c< tion t h a t a benzoyl group had migrated in prefers ence to the migration (or elimination) of a hydrogen I'l Q atom. These results, suggesting the preferential K & R migration of a benzoyl group, are not without prec'C=O \ edent. The decarbonylation of diphenyl triketone C=O was found to be explained most readily if the migra>c=c< f+ 5\ >c-c tion of a benzoyl group was a s ~ u m e d . ~The facile IC Ib migration of an acyl group to an adjacent electrondeficient center would appear to disagree with a the product of this reaction was actually l-phenylnumber of studies wherein the migratory tendency 2,3-butanedione (IV) which was formed via the inof an aryl group has been enhanced by the presence termediate chlorohydrin V.9 The same reaction is of electron-donating substituents and retarded by exemplified by the conversion of benzalacetophenthe presence of electron withdrawing substituents. one oxide to benzyl phenyl diketone in the presence However, there is no inherent reason why the mi- of hydrogen chloride. 2,10 Since the isomerization gration of an aryl group, a process which probably CsH,CH--CHCOCH3 CHjCOCH-CHO involves an intermediate phenonium should be I comparable with the migration of an alkyl or acyl \O' CsH, group. The transition state for the migration of I1 I11 an acyl group could be likened to the formation C H J C O C O C H ~ C ~ H ~ CbHbCH-CH-COCHj of a r-complex such as I between an olefin and a I / Lewis acid. The apparent high order of stability I\' C1 OH of acyl carbonium ions6 could be used to justify the v importance of contributions from resonance forms such as IC to transition states of the type I. X similar argument has been used to explain the products obtained from the cleavage of aliphatic ketones with peracids.' -1ccornpanying their discovery t h a t benzalacetoVI VI1 phenone oxide could be isomerized to formyldesoxybenzoin in the presence of a sulfuric acid-acetic acid of the epoxy ketone I1 to the keto aldehyde 111 mixture, Weitz and Scheffer described the conver- would presumably provide an example of the migrasion of benzalacetone oxide (11) to 2-phenyl-1,3-bu- tion of an acetyl group," the rearrangement of the tanedione (111)in the presence of hydrogen chloride.8 oxide I1 was reinvestigated. Treatment with an However, Moureu subsequently demonstrated that ethereal solution of boron trifluoride converted the epoxy ketone I1 to the borofluoride complex VI (1) Presented at t h e 128th Meeting of t h e American Chemical Society, Minneapolis, klinn., Sept. 11 t o 16, 1955. of 2-phenyl-1,3-butanedione. The same product (2) H. 0. House, THISJOTJRSAL, 76, 1235 (1954). was formed when an authentic sample of the keto (3) J. D. Roberts, D. R . Smith and C . C. Lee, i b i d . , 75, 018 aldehyde I11 was treated with boron trifluoride (1951). etherate. Treatment with phenylhydrazine con(4) (a) W. E . Bachman and J. W. Ferguson, i b i d . , 66, 2081 (1934); (b) J. G. Burr, Jr., and L. S. Ciereszko, i b i d . , 74, 5426, 6431 (1952); verted the complex VI to 1,4-diphenyl-Smethylpy(c) W. E. McEwen and N. B. Mehta, ibid., 74, 526 (1952); (d) J. D. razole (VII). Roberts and C. M. Regan, ibid.,76, 2069 (1953); (e) D. Y. Curtin and
+
hI. C. Crew, ibid., 76, 3719 (1954). ( 5 ) D. J. Cram, ibid., 71, 3863, 387.5 (1949); I). J. Cram and R . Davis, ibid., 71, 3871 (1949). (6) (a) R . J. Gillespie, J . Chem. Soc., 2997 (1950); (b) 13. Burton and P . F. G. Praill, i b i d . , 1203, 2034 (1950); ibid., 522, 529, 726 (19.51).
(7) W. D. Emmons and G. D. Lucns, THIS JOTJRNAL, 77, 2287 (1855). ( 8 ) I?. Weitz nnri A. SrhPlirr. n p r . . 64, 2344 (1921).
(9) H. Moureu, Cumpt. r e n d . , 166, 350, 503 (15128); A n n . d i m . (Paris), (101 14, 339 (1930). (IO) E. P. Kohler and R. P . Barnes, T i m J O U R N A L , 56, 21 I (1934). (11) T h e conversion of t h e oxides of benzalacetone and benzalacetophenone t o the corresponding B-keto aldehydes does not permit one t u decide whether t h e phenyl group or t h e acyl group has migrated. Although our studies of analogous systems suggest t h a t the acyl group has migrated in each instance, definitive stiidirs employing isntrvir lnhrlinp terlinirllirs 1i:ivr not ~ " Iirrn t ilotie.
HERBERT 0.HOUSEAND DON.\LDJ. KEIF
0326
It was of interest to extend the observation that a benzoyl group would migrate in preference to a hydrogen atom by a comparison of the migratory tendencies of a benzoyl group and an alkyl group. A priori the apparent tendency for a hydrogen atom to migrate (or be eliminated) in preference to the migration of an alkyl group12would lead to the prediction that a benzoyl group would have a greater tendency to migrate than an alkyl group. X study of the isomerization of a-methyl-P-phenylacrylophenone oxide (VIII) substantiated this prediction. CHB Ce,H,CH-C-COC6Ha
\d
CHs C6I:>).
Dec. 20, 1955
REARRANGEMENT OF EPOXY KETONES
6527
phenone (XVIIc) was recovered unchanged after alcohol X X to the crystalline cis-epoxy alcohol X X I treatment with alkaline hydrogen peroxide. On followed by oxidation to the desired ketone XVI is the other hand a-phenyl-trans-benzalacetophenone analogous to the reaction sequence recently reported (XIV) could be readily converted to the epoxy by Wasserman and Aubrey.22 It is interesting to ketone XV under the same conditions2 and PIP- note that the trans-oxide XV, Xmax 252.5 mp diphenylacrylophenone (XVIIb) was converted to ( emax 16,400) absorbs ultraviolet light a t slightly the corresponding oxide XVIII in 477' yield by longer wave lengths and with slightly greater inthe use of a reaction time of 27 hours. These ob- tensity than the cis-oxide XVI Xmax 252 mp (Emax servations are in agreement with the suggestion15 15,100) as would be predicted on the basis of simit h a t the olefinic double bond in systems of type lar studies by Cromwell and his c o - w o r k e r ~ . ~ ~ XVII is attacked very slowly by nucleophilic reThe boron trifluoride-catalyzed isomerization of agents (in this case the hydroperoxide anion1*). the trans-epoxy ketone XV to benzhydryl phenyl Moreover, a study of molecular models of the corn- diketone ( X X I I ) , previously reported2 in ether pounds XVIIa and XVIIc where epoxidation was solution, has also been effected in cyclohexane unsuccessful revealed that molecular conformations solution. A consideration of possible transition of these compounds in which the olefinic double states X X I I I and XXIV which might be involved bond and the carbonyl group were coplanar were very unT ! Z C 6 H C5 H 3 C 0 3 H C :H !,, favorable sterically. These coplanar conformations would be " H required to permit conjugation C6H5 C6H5 between the olefinic double xx XIX bond and the carbonyl group with a consequent ease of nucleophilic attack a t the olefinic double bond. Similar coplanar conformations are much more favorable for the unsaturated ketones X I V and XVIIb which can be converted to the ( C g H 5 ) 2 C H c O C 0 C 6 H 5 corresponding oxides. The ultraviolet spectra of the unsaturated ketones are in accord XXII XXIIl XXIV with these suggestions, the compounds wherein coplanar conformations are in the isomerization of the trans(XV)-and cis(XV1)more favorable absorbing a t appreciably longer epoxides, respectively, suggested that some benwave lengthsI7: XIV, Xmax 255 mp (Emax 14,800) zoyl migration might be observed with the cis isoand 302 mp (emax 13,400); XVIIb, Xmax 233 mp mer XVI. I n this case the transition state XXIV for phenyl migration requires the energetically un(Emax 15,100) and 302 mp (Emax 9,500); XVIIa, desirable cis orientation16 of the phenyl and benzoyl Xmax 260 r n p (Emax 24,600) and 281 mp (Emax 23,000); XVIIC, Xmax 242 mp (emax 24,800) with a point of groups. No such orientation would be required in the corresponding transition state for the migrainflection a t 280 mp ( E 12,700). The synthetic procedure employed for the synthe- tion of a benzoyl group. However, the isomerizasis of the desired cis-epoxy ketone XVI was based tion of the cis-epoxy ketone XVI in the presence of on a-bromo-trans-stilbene ( X I X ) . The lithium boron trifluoride etherate, either in cyclohexane or compound derived from this bromo olefin has been ether solution, produced neither of the expected found upon carbonation to yield a-phenyl-cis- diketones X X I I or XXV but rather the keto aldecinnamic acid, the bromine atom having been re- hyde XXVI. The product was shown to be idenplaced by a carboxyl group without loss of the rela- tical with a sample of the keto aldehyde XXVI pretive configurations of the groups about the olefinic pared by the isomerization of P,P-diphenylacrylodouble bond.1g Similar results have been obtained phenone oxide (XVIII) as described previously. * with other systems.20 Our expectation that the The product readily lost formic acid to yield benzalcohol derived from the reaction of benzaldehyde hydryl phenyl ketone (XXVII). The product obwith the lithium salt obtained from the bromo olefin tained by the acylation of the sodium enolate of diX I X would have the cis-configuration X X has re- phenylacetaldehyde with benzoyl chloride has been cently been confirmed by Lutz and Rinker who pre- found not to be the keto aldehyde XXVI as repared the same alcohol by the reduction of the cis ported,24but rather the enol benzoate X X I X of diketone XVIIa with lithium aluminum hydride.21 phenylacetaldehyde. The enol benzoate XXVIII of desoxybenzoin, the The subsequent epoxidation of the unsaturated enol benzoate X X I X of diphenylacetaldehyde and (18) C. A. Bunton and G. J. Minkoff, J . Chem. Soc., 665 (1949).
3
(19) D. Y. Curtin and E . E. Harris, THIS J O U R N A L , 73, 4619 (1951). (20) (a) A. S. Dreiding and R. J. P r a t t , ibij., 76, 1902 (1954); (b) E. A . Braude arid J. A . Coles, J . C ' h e ? ? ~.S~JC.. . 2078 ( l 9 Z l j ; i c j 12. A . Rraude and C. J. Timrnons, iOid., 2000 (1930). (21) R. E. L n t z and E. 11. Rinker, Jr., l n r s J O I J K N A L . 7 7 , 36(i (1855).
(22) H . H. Wasserman and N. E. Aubrey, ibid., 77, 590 (1955). (23) (a) N. H. Cromwell a n d M. A. Graff, J . Org. Chem., 17, 414 (1952); ib) N. H. Cromwell and co-workers, THISJ O I J K N A I . , 73, 1044 (1951); ( c ) N . IT. Croniwell aiid R . A. Settrrri~~iqt, ibid., 76, 6752 (1!154),
(24) W. Schlenk, 11. Ilillcrn;mr~ nnrl I. lR, A u i i . , 287, I35 (1931).
HERBERT 0. HOTXE AXD DONALD J . REIF
G52X
Yol. 77
the diketone X X V have all been excluded as possiExperimentalz7 ble intermediates in the rearrangement of either 3,4-Epoxy-4-phenyl-Z-butanone(II).-The keto oxide, oxide X V or XVI. When the enol ester XXVIII prepared by the epoxidation of trans-benzalacetoneZ8 as by IVeitz and Scheffer,2Qcrystallized from a was subjected to the reaction conditions employed described methanol-water mixture as white plates, m.p. 5 4 4 5 . 5 " for the above isomerizations only the unchanged (lit. 52-53°,9 54-56OZy). The infrared spectrum30 of the ester XXVIII and desoxybenzoin could be isolated; oxide exhibits a n absorption band at 1710 attributable similarly, the enol benzoate XXIX was recovered to a carbonyl group. The ultraviolet spectrum has a maxia t 221 m p (emila 10,200). unchanged. IVhen 1,2,3-triphenyl-l,3-propanedi-mum Rearrangement of 3,4-Epoxy-4-phenyl-Z-butanone.--A one (XXV) was subjected to the same reaction con- solution of 0.50 g. (0.0031 mole) of the keto oxide and 2.0 ditions the unchanged diketone XXV was recov- ml. (0.016 mole) of boron trifluoride etherate in 25 ml. of ered accompanied by a high melting, yellow, fluo- ether was stirred for 30 minutes and then washed with two of water. After the solvent had been removed rescent material whose cornposition suggested that portions from the organic layer, the residue was crystallized from it has the structure of the borofluoride complex petroleum ether (b.p. 90-100'). The borofluoride complex
xxx.
CHO SSVI
SS\'
SXIX
I t would thus appear that the rearrangement uf aphenyl-cis-benzalacetophenone oxide (XVI) represents a case wherein the oxirane ring of an a,P-epoxy ketone has been cleaved to produce a t least a partial positive charge on the carbon atom alpha to the carbonyl group (i.e., X X X I ) in violation of Pauling's adjacent charge rule.*5 It will be noted that a transition state XXXII of the type previously discussed does not require the unfavorable cis-orientation of
CgH5
XXXI
-
XXY
5 F3
XXXII bulky groups on adjacent carbon atoins. In addition the positive ion XXXI may be stabilized by resonancc with the adjoining phenyl ring.2fi (2.;) I,. I'aiiling. "The r a t i t r e of t h e Chemical Bond," Corncll Iini vcrsity Press, Ithaca, N. Y., 1948, p. 199. (26) This stabilization would probably also be available t o t h e ion derived from t h e trans-epoxy ketone X V . A s t u d y of molecular models suggests t h a t t h e conformation required f o r such stabilization (coplanxrity of t h e Iihenyl r i n g and t h e incipient carlioniiini ion) i h s l i x h t l y 11.~5 i:~vctr:~l>le liir 1 lit, Ii-iitis-osi~lt-S V t h u n for tlir c i s i,c)nier
x\.r
of 2-phenyl-1 ,J-!utanedione separated as white needles, m.p. 100-101.5 , yield 0.30 g. (46%). The ultraviolet spectrum of the product exhibits maxima a t 227 mp (emax 8,500), 268 m p (emax 5,900) and 276 mu. (emaX 5,800); the no band in the 3 fi region infrared s p e c t r ~ m ~ 0 exhibits .3~ dttributable to a hydroxyl function. I n the 6 u region i i strong band is found a t 1625 cm.-' which may be ascribed to a carbon-carbon double bond (or a carbon-oxygen double bond) in a chelate ' l n u l . Cdlcd. fo; CI0HyO2BF~: C, 57.19; H , 4.32; I;, 18.10; B, 5.1c5. Fouiid: C, 57.16; H , 4.21; F, 17.81; B, ,528. The same compound, identified by a mixed melting point determination, was formed when an authentic sample33of 2phenyl-l,3-butanedione, m.p. 71-73' (lit. 73-74°,34 7 6 ° 3 3 ) , wds treated with boron trifluoride etherate. In another experiment the crude rearrangement product WAS treated with a boiling solution of 1.0 ml. of phenylhydrazine and 0.1 g. of sodium acetate in 25 ml. of ethanol. After 20 minutes the boiling solution was diluted with water, decolorized with Sorit and cooled. The crude 5-methyl-1,4diphenylpyrazole crystallized as light yellow plates, m.p. 157-160", yield 0.29 g. (40%). Recrystallization afforded the pure pJ-razole as white plates melting a t 159-160.5". The ultraviolet spectrum of the pyrazole has a maximum .it 253 mu ( t l n z L18,500j . A n u l . Cdlcd. for ClsHlrS1: C, 82.03; H , 6.02; S,11.96. Found: C, 81.99; H, 6.09; X, 12.08. An authentic sample of the 2-pheny1-1,3-butanedione was converted to the same pyrazole, the two samples being identified both by a mixed melting point determination and by comparisori of their infrared spectra. a-Methyl-p-phenylacrylophenone Oxide (VIII).-ahIetliyl-~-phen~-lacrylophenone, a light yellow liquid boilingat 151-~1.59"(0.3 mm.), n% 1.6163, was prepared by the method of Kohleras who reported the boiling point to be 190192" (28 xnm.). The ultraviolet spectrum of the unsaturated ketone has maxima a t 224 mp (e,,,,, 12,000) and 290 m p t m R x17,800). A solution of 5.00 g. (0.0225 mole) of the unsaturated ketone in 100 ml. of methanol was treated with 10.0 ml. of 30% hydrogen peroxide and 5.0 nil. of 6 iV .iqueous sodium hydroxide. After the mixture had been stirred for 20 hours a t room temperature, it was poured into 350 ml. of water and extracted with ether. The ether extract was washed with water, dried over magnesium sulfate and the ether mas removed. m'hen a solution of the residue, a viscous oil, in a methanol-water mixture was chilled in Dry Ice, 3.05 g. (,57%) of the crude oxide, m.p. .
~
( 2 7 ) All melting points are corrected and all boiling points are uncorrected. T h e infrared spectra were determined with a Baird double beam infrared recording spectrophotometer, model B, fitted with a sodium chloride prism. T h e ultraviolet spectra were determined in 95'3 ethanol (except where noted) with a Cary recording spectrophotometer, model 11 MS. T h e microanalyses u-ere performed by Dr. S. X I . S a g y and his associates. ( 2 8 ) G. Gamboni, V. Theus and H. Schinz, ffelw. Chim. d c f a , 38, 2%?.5 (19.551, (29) E. Weitz and A. Scheffer, Ber., 5 4 , 2327 (1921). (30) Determined in carbon tetrachloride solution. (31) Determined as a N u j o l mull. (32) I,. J. Bellamy, "The Infrared Spectra of Complex hioleculcs," John Kiley and Sons, Inc., iYew York, N. I-., 1934, p. 126. (W)\ I . R t x h , C " I I I Er e~n~d.. , 220, 322 (1045). (41) K . Y C > I I Arirvrrs and I T , I,lirlewifi, .411!1., 626. 130 ( l ! l 3 l i j . [:?2) 1;. 1'. K c ~ l l l r r ,,AJI
