J. Org. Chem., Vol. 37, N o . 12, 1072 1899
tranS-2,3-DI-tert-BUTYLCYCLOPROPANONE HEMIKETALS
Stereochemistry of Opening of Cyclopropanols. trans-2,3-Di-tert-butylcyclopropanon~ Hemiketals P. S. WHARTON* AND A. R. FRITLBERG Department of Chemistrv, Wesleyan University, Middletown, Connecticut 06467 Received September 9, 1972
trans-2,3-Di-tert-butylcyclopropanone was converted by methyl alcohol-sodium methoxide and ethylene glycol with its sodium salt to the corresponding 2-tert-butyl-4,4-dimethylpentanoateesters, tert-BuCHzCH-tert-BuCO2R. The stereochemistry of proton substitution a t C 3 was determined to be no less than 97 and 93% retention by use of the combination of deuterium-labeled solvents and nmr analysis of the diastereomeric deuterium-labeled products. Because similar but more general systems have previously yielded substantial amounts of inversion with these dissociating solvents, the results indicate the operation of a very strong factor favoring retention, and an explanation is advanced in terms of the specific and unusual geometry of the system studied.
Various remlts have been reported for the stereochemistry of the electrophilic substitution involved in base-catalyzed ring opening of cyclopropanols.’ Almost complete inversion (with respect to carbon as the leaving group) was found in the opening of trans-2phenyl-l-methylcyclopropanolzwith sodium hydroxide in dioxane-water, the only solvent system used. Opening of truns-2,3-dibutyl-2,3-dimethylcyclopropan01,~ endo- and exo-7-hydroxy-l,6-dimethyl [4.1.O]bicycloheptane^,^ and tricyclo[4.4.1.01~6]undecan-ll-013 proceeded with almost complete retention with potassium tert-butoxide in tert-butyl alcohol, but considerable amounts of inversion were observed with ethylene glycol and its sodium salt. In all these results the influence of a previously recognized solvent influence is discernible: low dielectric, nondissociating solvents, like tert-butyl alcohol, favoring retention and high dielectric, dissociating solvents, like water and ethylene glycol, affording substantial amounts of inversi0n.j However, it should not be surprising to find cyclopropanols possessing unusual features which do not conform to such behavior, and indeed l-hydroxynortricyclene6 opens with almost complete inversion in tert-butyl alcohol (as well as in methyl alcohol). The present study counterbalances this example with the trans-2,3-di-tert-butylcyclopropanone hemiketals (3), representing another unusual system which also fails to conform, but, in the opposite sense, affording almost complete retention in methyl alcohol and ethylene glycol. The henniketals were not isolated as such but were formed in situ by mixing trans-2,3-di-tert-butylcyclopropanone (2) with the appropriate alcoh01.~ I n the presence of alkoxide and excess alcohol, the conjugate bases of the h.emiketals (4) suffer ring opening to 2-tertbutyl-4,4-dimethylpentanoate esters ( 5 ) ; this conversion constitutes the second half of the normal Favorskii reaction* which occurs upon base treatment of an a-halo ketone, in the present case a-bromodineopentyl ketone (1). (1) Earlier results with discussion are presented b y C. H. DePuy, Accounts Chem. Res., 1, 33 (1968), and T. D. Hoffman and D. J, Cram, J. Amer. Chem. Sac., 91, 1009 (1969). (2) C. H. DePuy, F. 7.V. Breitbeil, and K. R . DeBruin, ibid., 88, 3347 (1966). (3) A. R.Fritzberg, Ph.D. Thesis, Wesleyan University, 1971. (4) P. S. W h a r t o n a n d T. I. Bair, J . Oro. Chem., 81,2480 (1966). (5) D. J. Cram, “Fundamentals of Carbanion Chemistry,” Academic Press, New York, N. Y.,1965,Chapters 3 and 4. A useful summary is presented on p p 153-155. (6) A. Nickon, J. L. Lambert, S. J. Lambert, R. 0. Williams, and N. H. Werstiuck, J. Amer. Chem. Sac., 88, 3354 (1966). (7) S.F.Pazos a n d F . D. Greene, ibid., 89, 1030 (1967). ( 8 ) A. S. Kende, O w . React. 1 1 , 261 (1960).
2
3
ROH
tert.BuCH2CH-tert-BuCOlRC 5
ROD
tert-BuCHDCH-tert-BuCO,RC6
tert-Bu H
p i tert-Bu 4
Although obtention of unlabeled ester 5 cannot reveal the stereochemistry of electrophilic substitution, the diastereotopic nature of the geminal protons on C 3 does allow this determination, in principle, by generation, in deuterated solvents, of the two diastereomeric monodeuterated esters represented by 6, one corresponding to retention and the other to inversion. Analysis and Results. -In practice the determination of stereochemistry of hemiketal opening by using deuterated solvents was considered to be applicable providing that (a) nmr spectroscopy provided a viable analysis of the diastereomers represented by 6 and (b) randomizing exchange did not obscure the anticipated stereochemical differentiation, most obviously by enolization of the product esters ; both conditions were met.g The all-proton ester 5 , R = CHa, was prepared by the Favorskii reaction on a-bromo ketone 1, best by using a sample of commercial potassium tert-butoxide in tertbutyl alcohol, which yielded directly, not the tertbutyl ester, but the corresponding acid, presumably from the action of adventitious hydroxide in the tertbutoxide. Diazomethane treatment afforded the methyl ester. The 100-MHz spectrum of the methyl ester was somewhat clearer than the 60-MHz spectrum, displaying the expected ABC pattern in almost AMX form, with chemical shifts and coupling constants readily assigned to the individual protons of 5 in a conformation essentially frozen about the C 2,3 bond with the (9) This was not found to be the case in the base-catalyzed opening of trans-2,3-di-tert-butylcyclopropanol, which was simply prepared by lithium aluminum hydride reduction of a: the 2-tert-butyl-4,4-dimethylpentanal generated in deuterated solvents was found to have completely exchanged the a hydrogen for deuterium.$
WHARTON AND FRITZBERG
1900 J. Org. Chem., Vol. 37, No. 1W, 1979
effectively monolabeled (8), with one hydrogen and one deuterium atom attached to the ring, and subsequent opening, with an almost equal probability of breaking the two possible ring bonds, yielded a mixture of 9 and 6. 61.25
t&-t-Bu61.70 7
tert-Bu threo-6
tert-Bu erythro-6
tert-butyl groups trans related (see 7):1° 6 1.25 (Ha), 1.70 (Hb, deshielded relative to Ha by the proximate ester function) and 2.10 ppm (He) with Jab = 14.0, Jac= 1.2, and J b o = 10.7 Hs (corresponding to geminal, vicinal-gauche, and vicinal-trans relations, respectively). Negligible exchange of the acidic hydrogen of the ester (H,) occurred when it was treated with base in deuterated solvent under conditions which generated the ester from the cyclopropanone. Thus, on opening the cyclopropanone in methyl alcohol-0-d with sodium methoxide, both threo- and erythro-6, R = CHs, could be formed by electrophilic substitution, with each diastereomer exhibiting its own distinct nmr spectrum. The spectrum of the ester experimentally obtainedll consisted most obviously of two large and approximately equal peaks, each with a small coupling of ca. 1 Hs centered at 6 1.25 and 2.10 ppm corresponding to Ha and H, of threo-6, R = CH, (and a retention mechanism). From integrations of the regions of absorption corresponding to the BC pattern of erythrod, R = CH3, an upper limit of 3% was placed on this species, showing that the stereochemistry of proton addition had occurred with no less than 97% retention of stereochemistry.lz Opening of the cyclopropanone in ethylene glycol-0dz in the presence of its sodium salt yielded 6, R = CH2CH20H1with a spectrum for the a, b, and c hydrogens which similarly revealed that no less than 93% of proton addition had occurred with retention of stereochemistry. One other slightly unusual inverse labeling experiment yielded stereochemical results. Dineopentyl ketone was repeatedly dissolved in methyl alcohol-0-d containing sodium methoxide until 98% of the four a! hydrogens had been exchanged for deuterium. By treatment with cupric bromide in chloroform-ethyl acetate13 the ketone was converted to a-bromo ketone 1 containing 95% of deuterium in the three a! positions. However, the Favorskii reaction on the deuterated bromo ketone, carried out with the same sample of commercial potassium tert-butoxide in tert-butyl alcohol, as described for the all-proton compound, afforded, after esterification, a methyl ester in which only one deuterium atom per molecule remained. The nmr spectrum suggested the sequence of events which had taken place, starting with exchange of the deuterium atom on the methine carbon prior to closure to the cyclopropanone.14 The cyclopropanone thus became (10) Diagrams indicate the absolute stereochemistry of only one of the two enantiomeric sets of racemic systems. (11) All speotra were subjeoted t o deuterium decoupling. (12) No lower limit can usefully be attributed t o inversion product because of the difficulty of separating hny real signals from noise and adventitious or satellite peaks. (13) L.C.King and G. K. Ostrum, J. O w . Chem., 29, 3456 (1664). (14) For examples of exchange in the Favorskii reaction see F. G. Bordwell, R. R. Frame, R. G. Scamehorn, J. G. Strong, and S. Meyerson, J . Amer. Chem. Soc., 89,6704 (1967).
tert-Bu H tert-BuCHpCD-tert-BuC0,R 0 2 ! y 9
P=
D
tert-Bu 8
tert-BuCHDCH-tert-BuC02R 6
Although there is no stereochemical consequence observable from the formation of 9, there is in the formation of 6, with the difference that threod now corresponds to inversion and erythro-6 to retention. Experimentally, the nmr spectrum of the methyl ester obtained consisted most obviously of almost equal intensity AB and BC patterns arising from 9 and erythro-6 (retention), and, by integrating over the region of absorption of the a and c protons corresponding to threod (inversion), it was concluded that no less than 95% retention had occurred.
Discussion In the opening of cyclopropanone 2 under Favorskii conditions in tert-butyl alcohol with (presumably) hydroxide ion, the observed retention result does not require a special explanation, and the operation of any special effect is masked by the nondissociating character of the solvent which alone could be responsible for the stereochemical result.6 However, the stereochemical result of retention in methyl alcohol and ethylene glycol indicates the operation of a strong effect opposing the natural dissociative forces of the solvents, and an explanation can be sought for in the unusual and specific geometry of the system. Considering the substitution to be of the HE1 category, opening of the hemiketal 10 presumably occurs with rotation of the carbanion terminus of the breaking bond such that the most stable, all-staggered conformation 11 is reached directly (rotation in the opposite sense yields a conformation in which the tert-butyl groups are gauche-related), assuming a pyramidal configuration for the carbanion. The most stable con-
10
tert-bu
tert-Bu
11
12
figuration of the inverted carbanion is similarly likely to be 12 and, although there is no formidable or lopsided torsional barrier apparent in the simplest inversion process which converts 11 to 12 by sweeping a
trC&nS-2,3-DI-tert-BUTYLCYCLOPROPANONE HEMIKETALS
hydrogen atom across the vicinal tert-butyl group as the carbanion hybridization changes, the two configurations, in conflormations 11 and 12, must be considerably different in energy because of differential solvation. Carbanion 11 is effectively solvated by the gauche ester function and its accompanying solvent shell, whereas carbanion 12 is surrounded by a relatively hydrophobic region which is intensified by the bulk of the tert-butyl groups.1j The stereochemical consequence of these relationships is clearly retention : uninverted carbanion is more ntable than inverted carbanion, and it is surrounded by favorably disposed solvent molecules on its open face. In general, it may be anticipated that a system with an intrinsic and extreme molecular asymmetry16 will open with a strong stereochemical preference. The recent literature provides other examples in this category, and rationalization of the results presents an interesting exercise: cyclopropoxide 136 opens with inversion whereas caged cyclobutoxide 1417and strained cyclopentoxide 1518 open with retention.
13
ROD
014
ROD
15
Experimental Section Physical Data.-Melting and boiling points are uncorrected. The spinning band distillation was carried out on a Nester-Faust 24-in. Teflon column, NFT-50, fitted with an automatic reflux ratio control. Gas-liquid phase chromatography was performed on a Varian Aerograph unit, Model A-90-P, using two 5 ft X 0.25 in. stainless steel columns packed with 60/80 firebrick coated with (1)20% Apiezon L and (2) 20% SF-96. Infrared spectra were recorded using a Perkin-Elmer 137 infrared spectrometer. Nmr spectra were recorded using a Varian A-60A spectrometer except for one 100-MHz spectrum of methyl 2-tert-butyl-4,4dimethylpentanoate which was generously run on a Varian HA100 unit a t Yale University. Chemical shifts are reported with (15) The other conformations available t o inverted carbanion are comparable t o 11 with respect t o solvation b u t are disfavored by a gauche interaction of the tert-butyl groups. (16) Intrinsic asymmetry of the type under discussion can be defined with respect t o the formation of diastereomeric products upon protonation as distinct from intrinsic symmetry which yields enantiomeric products. Other unsymmetrical carbanion systems are discussed in ref 5 . (17) A. J. Klunder and B. Zwanenburg, Tetrahedron Lett., 1721 (1971). (18) W. T. Borden, V. V. M. Cabell, and T. Ravindranathan, J . Amer. Chem. f l o c . , 93,3800 (1971).
J . Org. Chem., Vol. 37, No. 12, 1972 1901 reference to an internal tetramethylsilane standard. Deuterium decoupling was performed with a Nuclear Magnetic Resonance Specialties Heteronuclear Spin Decoupler, Model HD-60, and multiscan averaging was carried out with the assistance of a Varian 620i computer. Analyses of small concentrations of protons were made by comparing integrations of their signal intensities with those of appropriate 13C satellite signals. a-Bromodineopentyl Ketone (3-Bromo-2,2,6,6-tetramethyl-4heptanone) (l).-Bromination of dineopentyl ketone was carried out by following the procedure of King and Ostrum.la To a three-necked flask equipped with a paddle-stirrer, reflux condenser attached to a gas trap, and dropping funnel were added 73.2 g (0.328 mol) of cupric bromide and 160 ml of ethyl acetate. To the ethyl acetate, heated t o reflux, was added dropwise, over a 30-min period, a solution of 30.0 g (0.176 mol) of dineopentyl ketone in 165 ml of chloroform. The reaction mixture was stirred and maintained a t reflux for an additional 6 hr. At that time essentially all of the solid dark green cupric bromide had been replaced by solid white cuprous bromide, although the green color of the solution persisted. The reaction mixture was filtered, washed several times with water, and dried. Solvent was removed by evaporation and the residue was subjected to a spinning band distillation, with monitoring of the distillate by glpc on column 1 a t 150' (retention times of 5.1 and 17.8 min were observed for dineopentyl ketone and the a-bromo ketone, respectively.) Complete separation from starting material gave 30.4 g (69%) of a colorless liquid: bp 66-67" (2.5 mm); ir (film) 5 . 8 2 ~ ;nmr (cc14)6 4.11 (s, l ) ,2.50 (s, 2), 1.13 (s, Q ) , and 1.03ppm (s, 9). Favorskii Reaction of a-Bromodineopentyl Ketone. Methyl 2-tert-Butyl-4,4-dimethylpentanoate(5 = 7).-To a solution of 300 mg of a-bromo ketone in 1.O ml of tert-butyl alcohol was added 400 mg of potassium tert-butoxide (MSA Research). The mixture was stirred a t 40' for 12 hr. Work-up afforded 40 mg of a neutral oil and 195 mg (87%) of a white solid acid which could be crystallized from methanol-water: mp 69.5-71.5'; ir (CClr) 5.80 p ; nmr (CC1,) S 12.28 (9, l ) , 2.25-1.17 (ABC pattern, 3), 0.98 (s, Q ) , and 0.90 ppm (s, 9). Treatment of the carboxylic acid with diazomethane yielded, after work-up, an oil which was purified by preparative glpc on column 2 a t 135': ir (film) 5.76 p ; nmr (cc14)6 3.58 ( 6 , 3), 2.10, 1.70, 1.25 (Ha, Hb, and H, of ABC pattern, J a b = 14.0, J,, = 1.2, J b o = 10.7 Hz), 0.89 (s, 9), and 0.84 ppm (s, 9). Anal. Calcdfor C&aOz: C, 71.89; H, 12.13. Found: C, 72.02; H , 12.24. Favorskii Reaction of a-Bromodineopentyl Ketone-a,a' ,a'-d3. -To 25 ml of methanol-0-d (QQOJ, di) in which 200 mg of sodium had been dissolved was added 10.0 g ,of dineopentyl ketone. The solution was maintained a t 45' for 20 hr and then cooled and quenched by the rapid addition, with stirring, of 20 ml of 15% aqueous acetic acid. Work-up afforded ketone with 73% a-d4 deuterium incorporation (theoretical, 75%). Three repetitions of this treatment afforded 8.24 g of ketone with 98% a-d4 deuterium incorporation. a-Bromodineopentyl ketone-a,a',a'-d3, with 95% a-deuterium incorporation, was prepared from the tetradeuterated ketone by the procedure already described for the undeuterated compound, with, however, precaution to remove traces of protonic contaminants in the solvents: ethyl acetate was distilled from phosphorus pentoxide and chloroform was filtered through active alumina and then distilled. To a solution of 130 mg of deuterated a-bromo ketone in 1.0 ml of tert-butyl alcohol was added 135 mg of potassium tert-butoxide (MSA Research). The reaction mixture was stirred a t room temperature for 4 hr. Work-up afforded 20 mg of a neutral oil and 68 mg (71%) of a white solid acid. Esterification of 65 mg of the acid with diazomethane afforded 55 mg of a neutral oil which was purified by preparative glpc on column 2 at 135". The nmr spectrum of the methyl 2-tert-butyl-4,4-dimethylpentanoate so obtained is described in the text. Opening of trans-2,3-Di-tert-butylcyclopropanone.A.-Into an nmr tube were placed 80 mg (0.48 mmol) of di-tert-butylcycl~propanone~ and 0.80 ml of methanol-0-d (99yo d l ) , in which had been dissolved 10 mg (0.48 mmol) of sodium, and the sealed tube was placed in a bath maintained a t 61' for 13 hr. The reaction mixture was then diluted with 10 ml of pentane, washed with water, and dried. Removal of solvent gave 75 mg of an oil whose nmr spectrum corresponded to that of undeuterated methyl 2-tert-butyl-4,4-dimethylpentanoate except for the region S 2.7-1.1 ppm, which contained two major broad signals a t
1902 J. Org. Chem., "01. 37, No. id, 1972
FRIEDMAN AND GRABER
6 2.10 and 1.25 ppm which sharpened to doublets (JE 1 H a ) upon pentane solution was washed with water and then dried. Redeuterium decoupling. The ester was purified by preparative moval of solvent and preparative glpc of the residue on column glpc and its nmr spectrum was then multiscanned through the 2 a t 172' gave a colorless liquid with a retention time of 13 min: region 6 2.7-1.1 ppm with deuterium decoupling to determine ir (CCL) 2.79, 2.89, and 5.75 M ; nmr (CCl,) 6 4.2-3.6 (A2Bz the relative amounts of diastereoisomers 8 and 9 present, with pattern, 4), 2.14 (broad, 1, changed to a doublet, J E 1 Ha, results as discussed in the text. upon deuterium decoupling), 1.80 (broad, 1, changed to a douB.-Into an nmr tube were placed 80 mg (0.48 mmol) of diblet, J 1 Ha, upon deuterium decoupling), 0.92 (s, 9), and lert-butylcyclopropanone and 1.20 ml of ethylene gly~ol-O-d~1~ 0.86 ppm (9, 9). The spectrum was then multiscanned through in which 5.0 mg (0.25 mmol) of sodium had been dissolved. the region 6 2.7-1.1 ppm with deuterium decoupling to deterSome solid material separated after a few minutes and the mixmine the relative amounts of diastereoisomers 8 and 9 present, ture was periodically agitated during immersion of the tube for 14 with results as discussed in the text. hr in an oil bath maintained a t 61'. The reaction mixture was Registry No.-1, 33712-48-0; 2, 14743-58-9; 5, transferred to a separatory funnel with 15 ml of pentane and the (19) D.J. Cram and B. Rickborn, J . Amer. Chem. Soc., 83,2178(1961).
33712-50-4; 33712-51-5.
a-bromodineopentyl
ketone-a,a',a'-d3,
Stereospecific Rearrangement during the Piperidine-Catalyzed Condensation of Benzaldehyde and Bis(ethylsulfony1)methane. A n Abnormal Knoevenagel Condensation ALANR. FRIEDMAN" AND DAVIDR. GRABER Agricultural Research Laboratories, The Upjohn Company, Kalamazoo, Michigan Q9001 Received December 9, 1971 The piperidine-catalyzed Knoevenagel condensation between benzaldehyde and bis(ethylsulfony1)methane gives not the reported P,P-bis(ethylsulfony1)styrene (3) but stereospecifically rearranged (E)-a,p-bis(ethylfony1)styrene (1). An independent synthesis of 1 and its stereostructurally related isomer (Z)-alp-bis(ethy1sulfony1)styrene (7) is presented. When the condensation is carried out with excess piperidine, a-(ethylsulfony1)8-piperidinostyrene (10) is isolated in good yield.
During the course of another study, we were interested in preparing the previously reported p,p-bis(ethylTwo sets of workers, Leonard' sulfony1)styrene (3). and Oftedahl, et reported a synthesis of 3 by a Knoevenagel-type condensation between benzaldehyde and bis(ethylsulfony1)methane catalyzed by piperidine and piperidine acetate, respectively. Earlier, Rinzema, et aLj3 had reported the synthesis of 3 by the perphthalic acid oxidation of p,p-bis(ethy1thio)styrene (2). Although the melting points reported for 3 are coincidentally close,4 our investigation of the reaction reported by Leonard and by Oftedahl showed that the compound reported by them as 3 is actually (E)-+ bis (ethylsulfony1)styrene (1).
1
2
3
A mechanism with its stereochemical implications is presented for the formation of 1 and other products observed during this reaction. An independent synthesis of 1 and its stereostructurally related isomer (Z)-a,Pbis(ethylsulfony1)styrene (7)is also presented. Results and Discussion When either the synthesis reported by Leonard or by Oftedahl was repeated, we obtained a 33-3773 yield of 1. Although 1 has the approximate melting point (1) E.C. Leonard, J . Org. Chem., 80, 3258 (1965). (2) M. L. Oftedahl, J. W.Baker, and M. W. Dietrich, ibid., SO,296 (1965). (3) L.C. Rinzema, J. Stoffelsma, and J. F. Arens, Reel. Trav. Chim. PaysBas, 78, 359 (1959). (4) Melting points reported for 1: Leonard, 93-95"; Oftedahl, 96-97': Rinzema, et al., 92-93O.
(94-96°)3 reported for 3, it lacks the strong ir band near 1620 em-' reported by R i n ~ e m a . ~A sample of 3 was prepared by the method of Rinzema, et and was found t o be different from the Leonard-Oftedahl product 1. As reported by Rinzema, et al., 3 reacts rapidly and exothermically with phenylhydrazine, generating benzaldehyde phenylhydrazone and bis(ethy1sulfony1)methane. Under the same conditions, 1 remains unchanged. Independent Synthesis and Stereochemistry of 1. An independent synthesis of 1 was desired. The basecatalyzed addition of mercaptans t o acetylenic compounds in general, and acetylenic sulfones specifically, has been studied by several workers. Among them, Truce6 and Stirling' have shown the additions to be stereospecifically trans. The addition of ethyl mercaptan t o ethyl phenylethynyl sulfone (5)s was ex(ethylsulfony1)-a- (ethy1thio)stypect ed to give (2)-/3rene (6),which could subsequently be oxidized to (2)a$-bis(ethylsulfony1)styrene (7), the stereoisomer of 1. Treatment of 5 with sodium ethanethiolate in ethanol at 5" gave 6 (23%) and (E)-a-(ethy1thio)-@-(ethylsu1fonyl)styrene (8) (2%). Surprisingly, the major product (42y0 yield) was p-(ethylsulfony1)-P-(ethylthio)styrene (4). The structure of 4 was assigned from its perphthalic acid oxidation t o 3. Oxidation of the major p-(ethyl(5) Neither Leonard's nor Oftedahl's paper presented ir data. However, the ir spectrum of the compound Oftedahl reported as 3 can be found in "Sadtler's Standard Spectra," Sadtler Research Laboratories, Philadelphia, Pa., 1966,ir spectrum no. 28346. ( 6 ) W. E. Truce and J. A. Simms, J . Amer. Chem. Soc., 78, 2756 (1956). (7) J. M. Stirling, J . Chem. Soc., 5856 (1964). (8) ( a ) Prepared b y the m-chloroperbenzoic acid oxidation of the previously reported ethyl phenylethynyl (h) L. Brandsma, H. E. Wljers, and C. Jonkers, Recl. Trau. Chzm. Pays-Bas, 83, 208 (1964). (0) Compound 5 has recently been synthesized b y a n alternate procedure: W. E. Truce and G. C. Wolf, J . Org. Chem., 36, 1727 (1971).
