chemistry of cyclopropanols

7/29/2019 Chemistry of Cyclopropanols

1/9

ACCOUNTS OF CHEXICAL RESEARCHV O L U M E 1 N U M B E R 2 F E B R U A R Y , 1 9 6 8

The Chemistry of CyclopropanolsC . H. DEPUY

Department of Chemistry, University of Colorado, Boulder, Colorado 80308Received June 18, 1967

Cyclopropanols have been principally synthesized by (1 ) reaction of epichlorohydrins with magnesiumbromide followed by treatment with Grignard reagent and ferric chloride, (2 ) acid hydrolysis of cyclopropylvinyl ethers, and (3)hydrolysis of cyclopropyl acetates which are available via the Baeyer-Villiger oxida-tion of methyl cyclopropyl ketones or pyrolysis of acetoxypyrazolines. Cleavage of a carbon-carbon bondof cyclopropanols or cyclopropyl acetates occurs upon reaction with electrophiles, leading to incorporationof the electrophile in the C2 (or C,) position of the original cyclopropane in the final product. Protonicacids attack trans-2-phenyl-1-methylcyclopropanoleading to retention of configuration in the product.Reaction of this same alcohol with electrophilic halogen (t-butyl hypohalite, N-bromosuccinimide) givesinversion of configuration. When trans,trans- or cis,trans-2,3-dimethyl-l-phenylcyclopropanols cleavedwith brominating agents, bromoketones are formed, also with inversion of configuration. In base-catalyzedcleavage of cyclopropanols, protonation occurs with inversion of configuration. Solvolysis of cyclopropyltosylates and halides generally occurs with simultaneous ring opening and with t he ultimate formation ofallyl cations. The Woodward-Hoffmann rules, which predict a disrotatory mode of opening, are obeyed.In addition, a selection between the two allowed disrotatory modes is imposed by the relative configura-tional geometry of the leaving group; groups trans to the leaving group rotate outwardly and groups cisrotate inwardly. Cyclo-propanols are easily oxidized with ferric chloride, implying that homolytic cleavage of the 0-X bond isexceptionally easy.

Nitrite esters of cyclopropanols rearrange spontaneously to nitroso ketones.Cyclopropyl acetates generally rearrange upon pyrolysis and form allyl acetates.

HistoricalCyclopropanol was first synthesized (accidentally) byMagrane and Cottle in 1942 by the remarkable reactionof epichlorohydrin with magnesium bromide followedby treatment with ethyl Grignard reagent.' In thenext year attempts to reproduce the synthesis faileduntil it was recognized tha t an impure grade of mag-nesium was necessary for a successful reaction, or evenbetter that small amounts of ferric chloride had to beadded as a catalyst.2 Under the proper conditions,yields of impure cyclopropanol of greater than 40% wererealized; large amounts of ethane and ethylene wereliberated as by-products.

CHz/ \C H 2 w -w C H -O H (1 )Cyclopropanol prepared in this way was impossible topurify completely (87% maximum purity), because

(1) J. K.Magrane, Jr., and D. L . Cottle, J.Am. Chem. SOC.,64,(2) G. . tahl and D. L. Cottle, ibid., 65, 1782 (1943).484 (1942).

upon attempted distillation it isomerized to propional-dehyde. The structure of cyclopropanol was proven bythe preparation of several solid derivatives which werenot identical with those of any other simple alcohols andwhich had correct elemental analyses. Upon standingover solid potassium carbonate, cyclopropanol was con-verted to t he aldol condensation product of propional-dehyde.20PHZ\ KzCOaCH2---cHOH ---+C H , C H &H5 CHa

CHICHzCH=ACKO (2 )In 1951 Roberts and Chambers3 prepared smallamounts of cyclopropanol by the oxygenation of cyclo-propylmagnesium chloride and, more successfully,

by the Cottle procedure described above. Again, thepure alcohol was not obtained, but a tosylate derivativewas prepared and its solvolytic behavior investigated(see below). Although a few scattered references tomore complex cyclopropanols appear in the earlierliterature, in general the authenticity of their molecularstructures remains in doubt. It seemed to us that the(3 ) J. D. Roberts and V. C. C hambers, ibid., 73, 3176 (1951).

33


2/9

34 C. H. DEPUY VOl. 1chemistry of such a simple molecule deserved furtherinvestigation, especially in view of the isomerizationreactions reported by Cottle, and we began several yearsago a systematic study of the synthesis and reactions ofwhat has proven to be a most remarkable group ofcompounds.Synthesis of Cyclopropanols

We began by assuming (wrongly, as we shall see)th at the synthesis of substituted cyclopropanols by theCottle procedure was not likely to be fruitful. We knewtha t in order to investigate a variety of cyclopropanolswe would need synthetic methods whose substitutionpatterns, both structural and stereochemical, could becontrolled and determined. Not too long before ourinitial investigations, Emmons and Lucas4 showed thatcyclopropyl acetate could be obtained in high yield bythe Baeyer-Villiger oxidation of methyl cyclopropylketone (eq 3).6 A large variety of cyclopropyl acetateswas potentially available, because methyl cyclopropylketones can readily be prepared from the correspondingacids, and because a large number of substituted cyclo-propanecarboxylic acids are known. Therefore, we

CHz/ \ 1 CFaCOaHCHz--- CH- CHs-CHz 0 CHz/ \ 11 OH / \CHz-- CH-OCCHa - Hz---CHOH (3 )

began by investigating the possibility of hydrolyzingcyclopropyl acetates to cyclopropanols.6 It was shownth at the hydrolysis rate of the ester is appreciablygreater than the isomerization rate of the alcohol, andthat it is therefore quite possible to synthesize cyclo-propanol by this method. Yields are improved, how-ever, and the isolation and purification of the alcohol issimpler, if the ester is treated with methyllithium in

ether or, less conveniently, with lithium aluminum hy-dride. As prepared by this series of reactions, cyclo-propanol can be purified by preparative scale gaschromatography or fractional distillation at somewhatreduced pressures. An analytically pure sample (bp100.5-102.0") is reasonably stable in the absence ofacids or bases.This general method of synthesis has since been ex-tended to the preparation of a large number of substi-tuted cyclopr~panols .~ he synthesis of trans-Zphenyl-cyclopropanol may serve as an example. A mixture ofethyl cis- and trans-2-phenylcyclopropanecarboxylate is(4) W. . Emmons and G . B . Lucas, J . Am . Chem. Soc. , 7 7 , 2287(1955).(5) Methyl cyclopropyl ketones give, on Baeyer-Villiger oxidation,cyclopropyl acetates almost exclusively. Side chains larger thanmethy l give mainly alkyl esters of cyclopropanecarboxylicacids [R. R.Sauers and R. W . Ubersax, J.Org. Chem., 30,3939 (1965) .(6)C.H. ePuy and L. R. Mahoney, J.Am . Chem. Soc., 86,2653(1964).(7) C.H'. DePuy, G. M. appen, K. L. Eilers, and R. A . Klein,J.Org. Chem., 29, 2813 (1964).

prepared by the thermal reaction of styrene and ethyldiazoacetate. The esters or their corresponding acidsmay be separated by distillation or crystallization.The trans acid is converted to the methyl ketone in 759.1,yield by reaction with methyllithium. Oxidation withperoxytrifluoroacetic acid gives the acetate in 75-80y0yield. Reaction of the acetate with two molar equiva-lents of methyllithium gives, after distillation and re-crystallization, the trans alcohol, mp 41.5-42.0", in 79Q/,yield. Because of the instability of cyclopropanols,they are best stored a t the acetate stage and generatedjust before use.C&,CH%HZ +NZCHCOOCpHS -C

H e 2 H 6 H f iC6H5 H CsH5 COOC2H,

H f i H C 0 O H + 2CH,Li+H&CH

After we had gained experience in the handling andpurification of cyclopropanols, we turned again to theCottle procedure for cyclopropanol synthesis and weresurprised to discover that substituted epichlorohydrinsalso react with ethyl Grignard and ferric chloride toproduce substituted cyclopropanols. The reaction with2-methylepichlorohydrin is typical (eq 6). Evidence0 CHz/ \ 1. MgBrz / \

ICH,---C-CHz-Cl _____t CH2---C--OH (6 )2. CzHsMgBr, CH ,FeClaCHP 207, yield-

was also obtained that 1- and 2-arylcyclopropanols canbe obtained in this way. In general, however, it isdifficult to prepare the requisite epichlorohydrin exceptin the few cases where the corresponding olefin is com-mercially available.Magrane and Cottle had already recognized that thefirst step in the transformation of epichlorohydrin tocyclopropanol involves the ring opening of the epoxidewith magnesium bromide (eq 7 ) . The correspondingOMgBr

(CzHs)z0 ICHz--CH-CH~Cl +MgBrz CH2-CH-CH2 ( 7 )' Ic 1Br

l-bromo-3-chloro-2-propanol can be isolated from thisreaction upon acidification, and it , in turn, can be con-verted to cyclopropanol upon reaction with ethyl Gri-gnard and ferric chloride. It occurred to us that ananalogous intermediate can be generated by the reactionof the commercially available 1,3-dichloroacetone with


3/9

February 1968 THECHEMISTRYF CYCLOPROPANOLS 35Grignard reagents (eq 8) . In fact, this has proven to

0IICHZ-C H2 + RMgX-c 1 c 1I -7

be an exceedingly convenient way of synthesizing 1-substituted cyclopropanols.~ For instance, if 1,3-di-chloroacetone is allowed to react with phenylmagnesiumbromide and then with ferric chloride and ethyl Gri-gnard, 1-phenylcyclopropanol is obtained in 60y0 ield.The steps involved in the conversion of the 1,3-di-haloalkoxide to the cyclopropanol have not been provendefinitely, but a reasonable formulation is possible,based on the work of Kharasch and coworkers.* Gri-gnard reagents are known to react with ferric chloride togive intermediate, unstable organo-iron compounds.These compounds break down, forming, in the case ofethyl Grignard, ethyl radicals and, presumably, atomiciron (eq 9). Disproportionation of the ethyl radicalsleads to ethane $nd ethylene, observed to evolve from3C2HjMgBr+FeC13- (CzHS)aFe]- CzHa- +Fe (9)the reaction mixture, while the iron dehalogenates thedihalide. In this way, the iron is recycled, and only acatalytic amount is required.

R O MgX R O MgX\C +FeClz (IO)\ /CH1-C-CHp 4- Fe4Ic1

In the meantime, methods for th e synthesis of cyclo-propanols have been forthcoming from other labora-tories. One of the more versatile methods has beendevised by Schollkopf and coworker^.^ In this method,a carbene generated from chloromethyl 2-chloroethylether is added to a double bond, forming a 2-chloroethylether of a cyclopropanol. Dehydrochlorination to thevinyl ether followed by mild acid hydrolysis gives thecyclopropanol.

/ \CH2-CHzc 1

0C1CHzOCH2CHzCl+CH3Li%-CHOCHzCH,Cl -Finally, some methods are available for the synthesisof certain types of cyclopropanols, methods whichcompensate in convenience for what they may lack incomplete generality. Foremost among these is themethod developed by Freeman,lo in which a 2-pyrazo-line, formed by reaction of an a,@-unsaturated ketone(8) M. S. Kharasch, M. Weiner, 1%'.Nudenberg, A. Bhat tacha rya,(9) U. Schollkopf, J. Paust, A. Al-Azrak, and H. Schumacher,(10) J. P. Freeman, J. O r g . Chem., 29, 1379 (1964).

T. Wang, and N. C. Yang, J . Am . Chem. Soc., 8 3 , 3232 (1961).Chem. B e r . , 99, 391 (1966).

with hydrazine, is oxidized to a 3-acetoxy-1-pyrasolineby lead tetraacetate. This latter compound, upon heat-ing, forms cyclopropyl acetate by loss of nitrogen. Alarge number of cyclopropyl acetates, including somestereoisomeric ones, have been prepared by this method.

j P b ( C K W ,,OOCCH,

Wasserman" has also prepared cyclopropanols fromderivatives of cyclopropanone.Electrophilic Reactions on CyclopropanolsCottle112 ecognized th at cyclopropanol readily tau-tomerizes under the influence of either acid or base,yielding propionddehyde. The acid-promoted ringopening of cyclopropanes has, of course, been knownfor a long time, but the reaction is difficult to study fora number of reasons. In general, cyclopropanes openunder the influence of acidic reagents so as to form themost stable carbonium ion.12 The resulting carboniumion may rearrange, eliminate in one of several directions,react with solvent, or give rise to any one of a numberof products. Hydrogen exchange before ring openinglaand other complications are also being unco~ered.l~*~6Cyclopropanols, on the other hand, and especially 1-substitu ted cyclopropanols, share almost none of thesedisadvantages. The hydroxyl group, by virtue of itsability to stabilize an adjacent positive charge, controlsand facilitates the ring opening so that one, or atmost two, products are formed. The intermediatecarbonium ion, by loss of the hydroxyl proton, is con-verted completely to stable ketonic products. For ex-ample, 1-phenylcyclopropanol is converted quantita-tively to propiophenone in a matter of a few hours uponreaction with 1 N acid in dioxane-water a t 50". If adeuterio acid is used, a single p-deuterium atom is intro-duced into the ketone (eq 13). Kinetic studies have

OH CHzD50 ' I d+De- Hz-C--OH -

c6H5 CHiD(!!H%-C==O (13)

(11)H. H.Wasserman and D. C . Clogett, Tetrahedron Lettera,(12) A. Bhati, Perfumery Essent. Oil Record, 53, 15 (1962).(13) R. .Baird and A. A. Aboderin, J. Am . Chem. Soc., 86 , 252

(1964).(14) N.C. Den0 and D. N. Lincoln, ibid.,88, 357 (1966).(15) I . Hart and R. H. Schlosberg, ibid., 88, 030 (1966).

No.7,341 (1964).


4/9

36 C. H. DEPUY Vol. 1shown16 that the reaction is bimolecular, t ha t is firstorder with respect to reacting alcohol and to acid. If wereduce it to its essentials, the transformation is one inwhich a carbon-carbon single bond is broken by a proton(eq 14). Such a reaction would be classified according to

I 1 I II I I I

the Ingold scheme as S E ~nd has previously been in-vestigated only among organometallic compound^.^^

(14)C-C- +H e- C-H +@C-Stereochemistry of Ring Opening in Acid and Base

We turned first to an examination of the stereochem-istry of the ring opening of cyclopropanols. As asubstrate, trans-2-phenyl-1-methylcyclopropanolwaschosens and was prepared in optically active form.lgRing opening was accomplished by treatment with 1 Ndeutelium chloride (DC1) in dioxane-water at 90.Two products are formed, the desired 4-phenyl-2-butanone and the undesired 3-phenyl-2-butanone.Racemization of the latte r compound by base and care-ful separation of the isomers gives optically active4-phenyl-Z-butanone with retention of configuration (eq15) .19 Since a kinetic study of 1-arylcyclopropanols has

shown that appreciable positive charge develops on the1-carbon in the transition state,16 a reasonable mecha-nistic picture is one in which the proton attacks theCI-C2bonding electrons, which are postulated to bulgeout around the edge of the cyclopropane ringz0 eq 16).

In dilute basic solution, trans-2-phenyl-1-methyl-cyclopropanol is isomerized exclusively to 4-phenyl-2-butanone. The reaction is rapid at room temperature.gThe exclusive formation of the ketone resulting fromcleavage of the Cl-Cz bond, plus the fact that the prod-uct arises by inversion of configuration a t the benzyliccarbon atom, led us initially to suppose that an SE1reaction, similar to those investigated by Cram and co-workers,21was involved (eq 17). These stereochemicalresults (inversion in base, retention in acid) were ele-

(16) R . A. Klein, unpublished results.(17) D. J. Cram, Fundamentals of Carbanion Chemistry, Aca-demic Press, New York, N . T.,965, Chapter 3. See, however, H.Hoeeveen and A . F. Bickel. Chem. Commun.. 635 (1967). for a recentreport of an 832 reaction involving a saturated hydrocarbon.(18) Th e phenyl and hydroxyl groups are trans.(19) C. H. DePuy, F. R. Breitbeil, and K. R. DeBruin, J. Am.Chem. Soc., 8 8 , 3347 (1966).(20) C. A. Coulson and W-.Moffitt, Phil.Mag., 0, 1 (1949);K. B.Wiberg, Phvsical Organic Chemistrv, John Wilev and Sons, Inc.,Ne w York, N.Y., 9@, p 124.(21) Reference 17, Cha pter 4.

Js?O

C8.5gantly demonstrated in the all-aliphatic system ofKickon22 eq 18).

In an SEl reaction, a carbanion intermediate is formedin the rate-determining step. The very high degree ofstereospecificity observed in these base-catalyzed open-ings argues against free carbanions. Besides, in thesolvent-base system used by Xck on (t-butyl alcohol,&butoxide), Cram had found predominant retention ofconfiguration in his open-chain systems in which carb-anions definitely seem to be involved (eq 19). Re-CH3 O e

\ I (CHI)ICOHCsHs


5/9

February 1968 THECHEMISTRYF CYCLOPROPANOLS 37the center at which the proton must attack is quater-nary, as i t is in the compounds studied by Wharton and

HISO t D-C-CH2 (a)

R p\Bair, a carbanion intermediate may be involved. It isknown that attack at such centers is slow relative toattack at centers containing hydrogen atoms. Alter-natively, Whartonza has suggested th at Nickons resultsmay not be general in tha t ex0 attack in the norbornylsystem may be so much favored as to force inversion.Direction of Ring Opening in Acid and Base

Additional information about the transition states inacid- and base-catalyzed ring openings may be gainedfrom studies of the direction of opening in unsym-metrically substituted cyclopropanols. In every caseinvestigated so far , base-catalyzed opening occurs morereadily toward the carbon (C2 or C,) which can betterstabilize a negative charge. As already mentioned,CH3 CH2, pI I \I 11- -C-C- f R-C-H R (22)IRR R O

2-phenyl-1-methylcyclopropanol gives exclusively 4-phenyl-2-butanone. In addit ion, 1,2,2-trimethylcyclo-propanol gives pinacolone exclusively upon treatmentwith basez4 (eq 23).CH3 CHz OH CH I 0\ C CH3--t!--!+-CHa (23)/ I ICH3 6H a

The effect of structure upon the direction of acid-catalyzed opening is not yet clear. The nature of sub-sti tut ion a t C1 as well as a t Cz affects the product ratio,as the following two examples s h o ~ ~ ~ ~ ~eq 24). I nthe second example, substitution of the benzene ringa t Czby either electron-attracting or electron-donatinggroups has little effect on the rate of acid-catalyzedC1-C2 bond breaking. Furthermore, a change in stereo-chemistry (making the phenyl groups trans to eachother) also has lit tle effect on the rate. Acid treatmentof 1,2,2-trimethylcyclopropanol ives a mixture of 75%pinacolone and 25% methyl isobutyl ketone (eq 25),perhaps an indication of the operation of steric hin-drance.Z4J6

(24)C.H.DePuy, W . C. Arney, Jr., and D. H. Gibson, sub-(25) J. P. Clark, unpublished results.(26)B. R. avis and P.D. Woodgate, Chem. Commun., 65 (1966).

mitted to J . Am . Chem. SOC.

XH3 saH6CHzCHz8OHa +CeH6 H- -CHa60%0%CsHsI \CeH6 0 CHa 0It I II

93% 7 %CeH6CHzCHzCCeHs +CeHsCH-C-CaHa

CH3 \CH3CHa 0 CHZ 0/ \ /C-CH, (2 5 )C H I ~ - - - E - C H ~ +CHa-YH6H a

75 % ICH3 25%Ring Opening with Electrophilic Halogen

Considering their great reactivity toward protonicacids, it seemed reasonable to suppose that cyclo-propanols would undergo ring opening with other elec-trophiles, and we have recently completed studies withseveral halogenating agents.24 The results are ex-tremely clear-cut, although mechanistic ambiguitiesstill remain. 1,2,2-Trimethylcyclopropanol reactsrapidly and quantita tively with sources of positivehalogen (hypochlorous acid, t-butyl hypochlorite, t-butyl hypobromite, N-bromosuccinimide) in a varietyof solvents (water, &butyl alcohol, chloroform). Asingle haloketone results from cleavage of the C1-C3bond. Reaction a t the hydroxyl bond, with either an+Xe- H3-C- HzX 8CH3 (2 6 )CHz OHCHad ICH3IAH, CH3

electrophile or a radical, should have given rise to CI-C2bond cleavage (see below for the case of t he cyclo-propoxy radical), We assumed, therefore, th at elec-trophilic att ack was occurring. We next examinedthe ring opening of optically active trans-2-phenyl-l-methylcyclopropanol.27 The results were surprising inthat , in contrast to acid-catalyzed opening of the samecompound, only a single bromoketone was formed, withinversion of configuration28and with C1-C2 bond break-ing.

(27)W. .Arney, Jr., unpublished results.(28) This stereochemical conclusion must be qualified to this ex-ten t: the absolute configuration of the product has not been deter-mined chemically and its assignment rests on close analogies in theliter ature and also on Brewsters rules [J. H. Brewster, J . A m . C h e m.SOC.,1,5475 (1959)l. Experiments are underway to prove this pointconclusively. Our present opinion is th at the reaction is completelystereospecific, proceeding with 100% inversion. If this is true, thismethod may afford a unique and useful way of synthesizing opticallyact ive halides of known configuration.


6/9

38 Vol. 1. H. DEPUY0//H CH2 OH CHI-C

We next turned our attention to a system in whichthere was not an aryl group at the 2 position. Becauseof the possible complications in the norbornyl system(see above), we chose to examine the 2,3-dimethyl-1-phenylcyclopropanols, compounds which could be syn-thesized fairly easily. If ring opening with positivehalogen were to occur with retention of configuration,then the trans,trans-dimethyl alcohol would give rise toerythro bromoketone and the cis,trans-dimethyl com-pound to the threo bromoketone. If inversion were tooccur, these results would be reversed. Although theconfigurational assignment to the threo and erythrobromoketones had not been made previously, it seemedpossible to do so on the basis of a stereospecific transelimination under mild conditions. A mixture of thetwo bromoketones was prepared by the zinc bromidecatalyzed addition of benzoyl bromide to 2-butene and,as the nmr spectrum shows (Figure l), the methyldoublets of the two diastereomers appear a t distinctlydifferent positions, making stereochemical analysis easy.Each isomeric bromoketone was identified by its exclu-sive conversion to an unsaturated ketone by base-

9 7Figure 1. Methyl peaks from the nmr spectrum of a mixture oferythro- and threo-a-methyl-p-bromobutyrophenone.

trans, trans t hreo

CH, (28)base-

CH3 OHcis, trans e r y t h r o

FH3

catalyzed elimination (potassium acetate or potassiumhydroxide in ethanol at room temperature). The un-saturated ketones, presumed to arise via a trans elim-ination process, were identified by comparison of theirspectral properties with those of authentic samples.The results were as follows: both the trans,trans andcis,trans isomers react with bromine in t-butyl alcoholor in acetic acid, with N-bromosuccinimide in t-butylalcohol, or with t-butyl hypobromite in t-butyl alcoholexclusively with inversion of configuration. No trace ofthe isomer resulting from retention could be detected.The stereochemical results were not unique to the freealcohol, since the corresponding acetates also undergocleavage with brominating agents exclusively with in-version of configuration.It seems clear, then, th at the stereochemistry of elec-trophilic reactions of cyclopropanols depends in somenot yet understood way on the nature of the electrophileand the substrate. With bromine, inversion seems tobe the rule. With hydrogen, retention is observed inacid solution, while both inversion and net retentionhave been observed in basic solution, depending uponthe substrate used. Clearly additional work is neededbefore this interesting and important problem is clari-fied.Solvolysis of Cyclopropyl Tosylates

Roberts and Chambersz9 were the first workers tostudy the solvolytic behavior of cyclopropyl compoundsquantitatively. They observed th at cyclopropyl tos-ylate is extremely inert, requiring 180" fo r acetolysis.The product of the reaction is not cyclopropyl acetate,but rather allyl acetate. These results are consistentwith the rate-determining formation of the cyclopropylcation followed by rearrangement to the allyl cation,although, as Roberts and Chambers pointed out, thisdoes not constitute a proof of this mechanism. The(29) . D. oberts and V. C . Chambers, J. Am. Chem. Roc., 73,6034 (1951).


7/9

February 1968 THECHEMISTRYF CYCLOPROPANOLS 39CHz CHz

/ CH-HsCOOHCHz- CH-OTs - Hz- 8CHzCHrCOOHCH - HZ=CH-CH~OOCCH~ (29)/CH28

very high activation energy could be due in par t t o theincrease in angle strain in going to the intermediatecation, as well as to possible ground-state effects.In our own investigations30 we discovered tha t, asexpected, 1-arylcyclopropyl tosylates react much morerapidly than cyclopropyl tosylate itself. More surpris-ing was the observation t ha t either cis- or trans-2-aryl-cyclopropyl tosylate is also more readily solvolyzed thanthe parent compound. To accommodate these and re-lated results we postulated that the cyclopropyl cationis not an intermediate in these solvolyses, but that ringopening occurs simultaneously with loss of tosylate,leading to a partial positive charge on the benzyl carbonatom in the transition state. In this way a 2-arylgroup, either cis or trans, can stabilize the transitionstate and thereby enhance the solvolysis rate.31

The rearrangement of a cyclopropyl to an allyl cationis the simplest example of an electrocyclic transforma-tion, and hence is susceptible to treatment by the Wood-ward-Hoffmann In this case the prediction isextremely easy to make; a cyclopropyl cation is pre-dicted to rearrange in a disrotatory fashion so as togenerate th e required symmetry of the lowest molecularorbital of the allyl cation (eq 31). For a cyclopropyl

rdl \ /

(30)C.H. ePuy, L. G. Schnack, and J. W. Hausser, J . Am . C h m .Soc., 88, 3343 (1966).(31) oote and Schleyer had earlier postulated, on th e basis of theircorrelation of solvolysis r ate with bond angle, tha t cyclopropyl com-pounds solvolyze with simul taneous ring opening [C. 5. oote, ibid.,86, 1853 (1964); P. von R. chleyer, ibid., 86, 864 (1964)l.(32)R. .Woodward and R. Hoffmann, ib id . , 87, 395 (1965).

cation substitu ted cis in the 2,3 positions, two differentdisrotatory transformations are allowed, still observingthe orbital symmetries demanded for these transforma-tions; the two substituents may both move outward,leading to a trans,trans-allyl cation, or both inward,leading to a &,cis-allyl cation (eq 32). It should beemphasized that both transformations are allowed on

symmetry grounds; the conrotatory transformationleading to the cis,trans-allyl cation is disallowed by theWoodward-Hoffmann rules (eq 33). Given a freechoice between the two disrotatory modes, the first

would appear to be more favorable on steric grounds.We speculated, however, that the molecules mightnot have a free choice between the two disrotatorymodes, but rather that the mode of rotation is dependentupon the stereochemistry of the leaving group. Cal-culations predicteds2 and experiments showeda3 th atsubstituents cis to the leaving group rotate inwardly andsubstituents trans to the leaving group rotate outwardy(eq 34). This postulate was based initially on our feel--ing that the electrons of the C2-C3bond would interactwith th e rear of the developing orbital on C1. Laterkinetic and stereochemical work has seemingly fully con-firmed this postulate. The most extensive kinetic in-vestigations are those of Schleyer, Schollkopf, et ~ 1 .

(33) S. .Cristol, R. M. Sequeira, and C. H. DePuy, ibid., 87,4007(1965).(34) P. von R. Schleyer, G . W. Van Dine, U. Schiillkopf, andJ. Paust,, ibid., 88,2868 (1966).(35)U. chollkopf, K. ellenberger, M. Patsch, P. von R. Schleyer,T. Su, and G. W. van Dine, Tetrahedron Letters, lo . 37,3639 (1967).


8/9

40 C. H. DEPUY Vol. 1These workers showed that 2-trans-3-trans-dimethylcy-clopropyl tosylate, which on the basis of the above pos-tulate should give the stable trans,trans-dimethylallylcation, undergoes acetolysis approximately 5000 timesfaster than its C1 pimer, which is predicted t o give riseto the cis& cation. Notice, in these cases, th at it is

C A o T s - 1

the isomer with the alkyl groups trans to the leavinggroup which reacts the more rapidly. This order of re-activity can be reversed by simply joining the two sub-stituents into a ring.33,35 If the ring is small, a trans,trans allyl cation cannot be accommodated, but a cis,ciscation can.

relative rate

(37)H H OTs

The results above are kinetic, and the stereospecificityis only to be inferred from the relative rates. However,it has been shown recently that isomeric olefins areformed from the solvolysis of 8-bromobicyclo[5.1.0]-octane.37 The endo isomer reacts readily and gives riseto a cis-cyclooctenol. The ezo isomer is much less reac-tive, but when reaction is forced the highly strainedtrans-cyclooc enol is formed.cb, @OH (39)

Ah r l r HIt appears then that this highly stereospecific rear-

rangement of cyclopropyl derivatives can profoundlyinfluence both th e reactivity of t he starting materialand the struc ture of the products. Questions concern-ing this reaction still exist, however. In particular, itwould be interesting to know how much energy is re-quired to carry out a solvolysis which violates the Wood-ward-Hoff mann rules. An example of this violationwould be the solvolysis of the norcaranyl tosylate inwhich the six- and three-membered rings are fused trans.If rearrangement occurred at all, it might lead, via aconrotatory transformation, t o the cycloheptenyl cation,rather than to a trans cycloheptene derivative. The

starting tosylate does not appear to be prohibitivelystrained. sH 0%(----y4Q (40)

"H

Other Reactions of CyclopropanolsIn the course of our investigations of the electrophilicand solvolytic reactions of cyclopropanols, a number ofother transformations of these compounds and theirderivatives have been uncovered and are currentlyunder investigation in these and other laboratories.As mentioned earlier, we had to worry about whethercyclopropoxy radicals might be involved in the reactionof various halogenating agents with cyclopropanols.We set out to generate such radicals by photolyzing th enitr ite ester of 1,2,2-trimethylcyclopropanol. Nitriteesters are customarily prepared by the reaction of alco-hols with nitrosyl chloride in pyridine; the purifiednitrite esters are photolyzed to give a nitroso carbonylcompound in systems in which a Barton reaction cannotoccur39 eq41). When an a ttempt was made to prepare

a nitrite ester of several cyclopropanols, the nitroso-ketone was formed directly, even at -5040 (eq 42).

Apparently the cyclopropoxy radical, or its ring-openedanalog, is formed with great ease. The fact that open-ing occurred between C1 and Cz (as one would predict)is to be contrasted with the exclusive cleavage betweenC1and C B n the halogenative openings (eq 26).It has also been observed*l that , in contrast to otheralcohols, cyclopropanols are easily oxidized by ferric(38) It is interesting to speculate about whether similar stereo-specificity might not occur in other carbonium ion rearrangements.Specifically, the postulation that ci s groups rotate inwardly, transgroups outwardly in the cyclobutyl tosylate-homoallyl cation re-arrangement allows one to rationalize the enormous (108) endoleu,rate ratio observed in t he acetolysis of the bt osyla tes of bicyclo-

[a.l.l]hexane [K . B. Wiberg and R. Feroglio, Tetrahedron Letters,No. 0 , 1273 (1963)l. If one considers that cyclobutane is merelyhomocyclopropane, the same rules may be applied. In these casesthe products are again of no help, since further transformations areinvolved.n

(39) P. Kabasakalian and E. R. Townley, J. Org. Chem., 2 7 , 2918(1962): Barton reaction: D. H. R. Barton. J. M . Beaton. L. E.(36) Not on the same scale as above.(37) G. H. Whitham and M. Wright, Chern. Comm un., 294 (1967).

Geller,'and M. M. Pechet, J. Am . Chem. SOC., 2 , 2640 (1960).(40) H. L. Jones, unpublished results.(41) S. E. Schaafsma, H. Steinberg, and T h. J. De Boer, Rec. Trav.Chim. , 8 5 , 7 3 (1966).


9/9

February 1968 THECHEMISTRYF CYCLOPROPANOLS 41chloride in ether. When this reaction is applied to1,2,2-trimethy1cyclopropano1,40 chloroketone is formed(eq 43). These results indicate that homolytic cleavage

of the -0-X bond in cyclopropane derivatives is easilyaccomplished. It has not yet been determined whetherthe cyclopropoxy radical is an intermediate (if so , itmust have some special stability, like the cyclopropyl-carbinyl cation), or whether the reactions proceedreadily because of ring opening in the transition s ta tesuch that the activation energy is lowered by relief ofring strain. Electron spin resonance studies4' favor thelat ter interpretation, since ring-opened radicals have sofar been the only ones detected.These results may have a bearing on the reactions ofthe dimethylphenylcyclopropanols with chlorinatingagents. The same mixture of threo- and erythro-chloro-ketones is obtained upon treatment of trans,trans-2,3-dimethyl-1-phenylcyclopropanolwith ferric chlorideand of the cis,trans or trans,trans compound withchlorinating agents (eq 44). We feel that this lack of

CH3I

(44)

50% er y thr o 50% threostereospecificity may be due to attack at the -0-Hbond, and not electrophilic at tack at either C-C bond.42Finally, among miscellaneous reactions of cyclo-propanol derivatives might be mentioned the rearrange-

(42) We a t one time suggested tree radical attack a t the -0-H bondt o account fo r the rearrangement of cyclopropanols when heated incarbon tetrachloride solution in the presence of oxygen [ C .H. DePuy,G. M. Dappen, and 3. W . Hausser, J. Am. Chem. SOC.,83, 3156(1961)1. We later withdrew this suggestion [C . H. DePuy and F. W.Breitbeil, ibid., 85, 2176 (1963)] for the cases in which we had pro-posed it. The results reported above on the ease of breaking of the-0-X bond make it seem likely that under someasyet to be discoveredconditions such a mechanism is possible.

ment of cyclopropyl acetates upon pyrolysis.43 We firstnoticed this reaction when, in view of our earlier interestin pyrolytic elimination^,^^ we wondered if cyclopro-penes could be formed in this manner. Indeed, uponpyrolysis, l12,2-trimethylcyclopropylcetate smoothlyeliminates a mole of acetic acid at 475", but the productis 2,3-dimethylbutadiene (eq 45). All speculationsCH, OOCCHs CHs

(45)4750 \ /CHa+\$CHzFC\Ha-C- CHzI ICHI CH,

about the possibility of cyclopropene intermediates inthis transformation were pu t t o rest when it was shownthat 1-methylcyclopropyl acetate rearranges exclusivelyto 2-methylallyl acetate under the same conditions (eq46) . In the trimethyl case such rearrangement would4750- H2=C--CH200CCHa (46)

AH *CHr--cH\~00ccH3ICHa

lead to an acetate which would pyrolyze to dimethyl-butadiene under the reaction conditions.Initially we favored a concerted, cyclic mechanismfor this rearrangement] in line with those postulatedin other pyrolytic eliminations. With more exten-sive work, however, it appears that homolytic C2-C3bond breaking in the cyclopropyl ring is the initialreaction. In agreement with this idea, l-phenylcyclo-propyl acetate (in which homolytic CI-C2bond breakingwould be more favored over C2-C3 bond breaking) givesa new set of products.

CsH5 n (47)CH3CH2CCGHS + +other products

I a m greatly indebted to the Natio nal Science Foun datio n for i tsgenerous support of this work, and to a fellowship from the Alfred P .Sloan Foundation w hich allowed us to initiate our studies in cyclo-propane chemistry. I also wish to ezpress m y sincere appreciationto the graduate students and postdoctoral associates to whose work Ihave referred in this article.(43) D. Zabel, unpublished results.(44) C.H. DePuy and R. W . King, Chem. Rev., 60, 31 (1960),

chemistry of cyclopropanols

Documents