May 5 , 1962
COMMUNICATIONS TO THE EDITOR
usual, the crude product showed strong absorption a t 5.9 p and only a very weak band a t 6.1 p L Iindicative of ketol I11 as the major carbonyl component. Separation of the product mixture by gradient elution chromatography afforded 8% recovered 11, 56% of I11 and 36% of the diol (V) resulting from complete reduction. The structure of I11 follows from its infrared spectrum (>C=O at 5.9 p) and the absence of strong ultraviolet absorption in the 240-300 mp region, as well as its further reduction to the diol. Compound V, preparable from TI, 111, or IV by reduction with lithium aluminum hydride, was obtained as an oil, which could not be separated into individual racemates. It was, however, analytically pure (Anal. Calcd. for C17H2002: C, 79.65; H , 7.86. Found: C, 79.66; H , 7.73; infrared: 0-H stretching a t 2.7-3.0 p) and gave rise t o a crystalline bis-3,5-dinitrobenzoate,m.p. 70-74" (50yo ethanol), (Anal. Calcd. for C31H24N4012:C, 57.7'7; H, 3.75: Found: C, 58.18; H , 3.91). On the other hand, treatment of I1 with an equimolar quantity of sodium borohydride in ethanol and processing as above gave 46y0 of ketol IV 259, log E 4.3) and 54% diol V, but no I1 or 111. Clearly, further study of reaction conditions should lead to enhanced yields of I11 and IV, with less concomitant diol formation. 0 0 8 G C H 2 - C H , - - C - C I1H ,
OH
0 II
OHII1
0
IV
While it is somewhat premature to put forward specific mechanistic proposals in explanation of the above data, several points are quite clear. The highest reactivity toward I is observed in those ketones where an aryl group is forced out of conjugation with the carbonyl group,' thus allowing the -I effect to supersede the otherwise dominating resonance e f f e ~ t . ~ The abnormally high reactivity of phenyl t-butyl ketone toward sodium borohydride was rationalized in this manner.4 Dialkyl ketones, e.g., 2-octanone, are quite sluggish, because of the -tI substituents surrounding the carbonyl group, but are somewhat more reactive than aryl ketones where resonance interaction is not inhibited, e.g., acetophenone and 4-methoxyacetophenone. The phenomena reported here are of both theoretical and synthetic interest. Lye anticipate that a careful study of substituent effects on reactivity will help to elucidate the novel reactivity of complex I, particularly the differences compared with sodium borohydride toward substrates such as 11. This project is underway in our laboratory, as well as continued studies on the structure and selectivity of the reagent. Acknowledgment.--U7e gratefully acknowledge the generous support of this research by the National Science Foundation (Grant No. Gl5739). (7) R. N. Jones, % b i d ,69,
2141 (1945).
DEPARTMENT OF CHEMISTRY PETERT. LANSBURY UNIVERSITY O F BUFFALO JAMES0. PETERSON BUFFALO14, NEWYORK RECEIVED APRIL 2, 1962
I1
1
1757
1
\LiAIH4
Preliminary indications that the above selectivity is of appreciable scope have been gained by the conversions of 2-acetyl-9-fluorenone6 to 2-acetyl-9fluorenol and p-benzoylphenylacetone to 9-(ahydroxybenzy1)-phenylacetone in high yield by complex I. However, an attempted selective reduction of the &ketone function in 6-ketoestrone methyl ether was not successful, this being understandable in view of the efficient aryl conjugation of that group (vide infra). Compounds 111 and I V , obtained by column chromatography as described, were converted to the 2,4dinitrophenplhydrazones, m.p. 88-91' and 155-157', respectively (Anal. Calcd. for CzaHz?OoEd:C, 63.59; H, 5.10; N, 12.90. Found: for 111, 2,4-DKPH, C, 64.21; H, 5.22; S , 13.06; E 359 (22,400); for IV 2,4-DNPH, C, 64.01; H, 5.18; E, 12.80; e 381 (28,600). T h e absorption maxima of the dinitrophenylhydrazones are in good accord with those of model compounds, such as the corresponding derivatives of 2-butanone and 4-methylbenzophenone. (6) Y . Sprinzak, J . A m . Chem. SOC.,80, 5449 (19.58).
CONCERNING THE MECHANISM OF FORMATION O F PHENYL-(TRIHAL0METHYL)-MERCURIALS Sir:
Reutov and Lovtsova' recently described a useful synthesis of aryl-(trihalomethy1)-mercury compounds by the reaction of an arylmercuric halide with chloroform or bromoform and potassium tertbutoxide in benzene solution. The mechanism of this reaction was stated to involve insertion of a dihalocarbene into the mercury-halogen linkage, since the reaction conditions used paralleled those in the production of dihalocarbenes in the Doering-Hoffmann 1,l-dihalocyclopropane synthesis2 The interpretation of this synthesis of aryl-(trihalomethy1)-mercurialsin terms of a novel carbene reaction appears to have gained some ac~ e p t a n c e ,and ~ this prompts us to report results which contradict this assumed mechanism. The experiments noted here are pertinent. Reaction of phenylmercuric bromide with chloroform and potassium tert-butoxide in benzene a t 0" with high speed stirring, following the reported procedure, gave only phenyl- (trichloromethyl) mercury, m.p. 116.5-11s' (n-hexane), in 51% yield. Unreacted phenylmercuric bromide was recovered in 33y0yield. Furthermore, in a separate ( 1 ) 0. A . Reutov and A . N. Lovtsova, Izvesl. A k a d . .Vauk S . S . S . R . , Oldel. Khim. S a u k , 1716 (1960); Doklady A k a d . Y u u k S . S . S . R . , 139, 622 (1961). (2) W. van E . Doering and A. K. Hoffmann, J . A m . Chein. S o r . , 7 6 , 6162 (1954). ( 3 ) See for instance "A Brief Survey of Carbene Chemistry," R. C. de Selms, 01,~. Chem. Bull. (Easl?nan Kodak C o . ) , 3 4 , No. 1 ( 1 9 0 2 ) .
COMMUNICATIONS ro THE EUITVK
17%
experiment in which phenylmercuric bromide was treated with potassium chloride in a solvent mixture consisting of tert-butyl alcohol, chloroform and benzene for 4 hr. with high speed stirring, no halide exchange producing phenylmercuric chloride was observed. I n addition, phenyl-(bromodichloromethyl)-~iiercury,~ prepared by the above reaction in which bromodichloromethane was used in place of chloroform, was shown to be inert toward chloride ion under similar conditions. These results speak strongly against a dihalocarbene insertion mechanism for the reaction of phenylmercuric chloride with haloforms and potassium tert-butoxide. LVere such a mechanism to obtain, phenyl-(bromodichloromethy1)-mercury would be the product expected in the phenylmercuric bromide-chloroform-potassium tcrt-butoxide reaction in view of the demonstration t h a t neither phenylmercuric bromide nor phenyl-(bromodichloromethyl) -mercury undergo exchange with chloride ion under the conditions used. Our results indicate t h a t the formation of the phenyl(trichloromethy1)-mercurial isolated proceeded by simple nucleophilic displacement of bromide ion by the trichloromethyl anion. Hinej demonC6H6HgBr
+ CCla- --+-CGHjHgCCli - Br-
VO!.
h4
yields iii the case of the former mercurial. Phenyl(dibromochloromethyl)-mercurybserves excellently as starting material for the synthesis of l-bronio1-chlorocyclopropaiies by the same procedure. The mechanism of the phenyl-(trihalomethy1)mercury-olefin reaction is under investigation. Details concerning these and related experiments will be reported a t a later date. Acknowledgment.-'I he authors are grateful to the L.S.Xriny Research Office (Durham) for generous support of this work. (8) 1 l . p 110-111' ldec,), A n a l . Calcd. for C7HjCIDr,IIg: H A , 41 ; 3 0 ; CI, 7 :10, Br, ,12,U3, I'ound: HA. 4 L j f i ; C1, 7 . 0 7 ; Br, : 3 2 . i l L : E P A R T X E ~ T OF CHEMISIRY UIETMARS E T F ~ K T H
~ I A S A C I I U S EISSTITCTE ITS
TECHSOLOGY ~ A M E S51. UCKLITCH L I A R C H 24, 1962 OF
C A M B R I D G E , hIASSACHUSET'I'S KECEIVED
LARGE SECONDARY INTERMOLECULAR KINETIC ISOTOPE EFFECTS IN N ON-EQUILIBRIUM SYSTEMS. ENERGIZATION BY CHEMICAL ACTIVATION
Sir : The existence of very large, normal, secondary intermolecular kinetic isotope effects was pointed out recently by liabinovitch and Current, and the quantum statistical basis of the phenomenon described.l These eiiects may arise quite widely in uiiimolecular systems in which the energized species are produced in some non-eyuilibriulii distribution j ' ( c ) , i.c., one in which the poptlations of the various energy levels of interest are not governed by the ambient temperature si;d statistical thermodynamic equilibriurn considerations. The experimental techniques may in suitable cases include excitation by electrbil impact, light absorption, radiation, c$r. 111 Fact, thermal collisional C.P. ;activation may alsu be employed arid, under contlitions where noli-equilibrium populations I)re\,ail (the lower pressure region of thermal uiiiniolecular but the interpretation of these reactions is compli- reacticns), has been shown to give rise to very large cated by the fact t h a t exchange between bromide iizversc intermolecular secondary isotope e5ects.' ion (from the haloform-base reagent mixture, The above mentioned authors perforlned experiwhich is used in twofold excess) and phenylmercuric ments involving excitation of vibrational and active chloride does occur. rotational degrees of freedom by chemical activaThe use of phenyl-(trichloromethy1)-mercury tion,3 in which the rates of C-H rupture of enerand phenyl-(tribromomethy1)-mercury in the syn- gized ethyl-d, and ethyl-d:
