4 Reaction Mechanisms Part (ii) Polar Reactions By H. R. HUDSON Department of Chemistry The Polytechnic of North London Holloway Road London N78DB 1 Introduction This year’s report covers nucleophilic substitution (at saturated allylic and vinylic centres) carbocations and their rearrangements carbanions &elimination elec- trophilic addition and nucleophilic addition to carbonyl compounds. The aim has been to select material of a novel character where possible and not to give a comprehensive coverage of the literature which is dealt with elsewhere.’ Polar reactions in the gas phase have not been included as these were highlighted in last year’s report which referred to some of the more significant lines of research in this area. 2 Nucleophilic Aliphatic Substitution Substitution at Saturated Carbon.-The role of ion-pair intermediates in nucleo- philic aliphatic substitution has been referred to previously [Ann Reports (B),1974 71 114; 1976 73 551.As a further means of exploring this topic the reactivity- selectivity principle (r.s.p.) has been applied to studies of the solvolysis of n-octyl 1-methylheptyl and benzyl substrates (halides and/or arenesulphonates) in aqueous ethanol.* According to this prin~iple,~ highly reactive species such as ion-pair intermediates would be expected to show less discrimination in their reactions with competing reagents than would the less reactive neutral substrates. Selectivity values were determined and were expressed in terms of the ratio (kE/kw) of the rate constants for the reactions of each of the substrates with ethanol and with water respectively.Low selectivity values (0.63-3.61) coupled with the total absence of a leaving group effect were interpreted as possible evidence for product-determining attack on a highly reactive ion-pair intermediate not only for benzyl and 1-methylheptyl but more remarkably for n-octyl also. The nature of any ion-pair that might be involved was not however clear and the results were not claimed to be more than speculative. A similar use of the r.s.p. approach in the solvolysis in aqueous ethanol of a series of para-substituted benzyl derivatives has led to a more definitive identification of ’ ‘Organic Reaction Mechanisms 1978’ ed. A. R. Butler and M. J. Perkins Wiley London 1979.* A. Pross H. Aronovitch and R. Koren J.C.S. Perkin 11 1978 197. ’ A. Pross Adv. Phys. Org. Chem. 1917,1469. 48 Reaction Mechanisms the species involved and to the proposal of a new diagnostic tool for the identification of solvent-separated ion pairs in such reactions viz. that a leaving group effect on selectivity will be ob~erved.~ For p-chlorobenzyl and benzyl halides which solvolyse to yield products from intimate ion pairs a plot of log(k,/kw) against solvent ionizing power Y was linear and was essentially the same for both chlorides and bromides. In contrast the plots for p-methyl and p-methoxybenzyl halides deviated considerably from this pattern and indicated product formation from at least two solvolytic intermediates one being the solvent separated species (Scheme 1).R-X $ RtX-$ R+IIX-1 1 products products Scheme 1 In studies of the use of chlorine leaving group kinetic isotope effects (k3’/k3’)for the investigation of ion-pairing a series of benzyl chlorides p-XC6H4Cl (X = H Me or PhO) has been subjected to solvolysis in aqueous acetone with or without the presence of azide ion or thiophenoxide For processes thought to involve classical sN2 displacements values of ca. 1.0092 (X= H) 1.0094 (X = Me) and 1.0097 (X=PhO) were obtained. Smaller kinetic isotope effects were correlated with competing SN2 and ion-pair pathways and a lower limiting value of 1.0059 (X = Me) or 1.0058(X = PhO) was observed for solvolysis in 97 % trifluoroethanol. These figures are very close to the calculated equilibrium isotope effect (1.0057) obtained from i.r.vibrational frequency data and indicate rate-determining ion-pair dissociation in these cases. Complications in interpretation could however arise from compensating solvation of the incipient chloride ion in the transition state.6 To investigate the effect of solvent on transition state structure for a conventional SN2 process the displacement of chloride ion from n-butyl chloride by thiophenoxide [Equation (l)]was ~tudied.~ Although rate constants changed by three orders of PhS-+RCI + PhS--R--CI + PhSR +C1-(1) [ 6-6-1 [R = n-C4H9] magnitude over the range of solvents used (alcohols diglyme tetraglyme DMF and DMSO) the chlorine kinetic isotope effect (k3’/k3’) was essentially constant (1.0094f0.00012).This result is in contrast to that reported earlier’ for the effect of solvent variation on the displacement of dimethyl sulphide from tri-methylsulphonium by ethoxide ion [Equation (2)] for which k32/k34decreased from EtO-+MeSMe2 * EtO--Me--SMe2 + EtOMe+SMe2 (2) + [” 6+ 1 1.0095to 1.0035as the solvent was changed progressively from ethanol to a mixture of ethanol (35%) and DMSO (65 %). It has however been pointed out that the H. Aronovitch and A. Pross J.C.S. Perkin 11 1978 540. D. G. Graczyk J. W. Taylor and C. R. Turnquist J. Amer. Chem. SOC.,1978,7333. ‘G. W. Burton L. B. Sims J. C. Wilson and A. Fry,J. Amer. Chem. SOC.1977,99 3371. ’R. T. Hargreaves A. M. Katz and W. H. Saunders jun. J. Amer. Chem. SOC.,1976,98,2614.H. R.Hudson charge distributions at the transition states for these two systems are different.8 From a consideration of these observations and of other literature data on secondary a-deuterium kinetic isotope effects it has been concluded that there will be little or no solvent effect on SN2 transition state structure if the two nucleophiles involved are of like charge [Equation (l)]but that the transition state structure will be sensitive to solvent effects if one nucleophile is negatively charged and the other is neutral [e.g. Equation (2)].8 A novel approach to mechanisms of aqueous solvolysis has been based on the concept that water contains an equilibrium concentration of ‘free’ hydroxyl and ‘free’ lone-pair groups as indicated by the overtone regions of the 0-H stretching band for water [Equation (3)].9It was suggested that these free groups are the reactive (H2O)bul~e (0H)free+ (LPhree (3) ingredients which may initiate hydrolyses involving for example C-H ionization [Equation (4)Jor C-Cl heterolysis (Scheme 2).The hypothesis gives a qualitative R-CI + (OH)fre $ Rf+ (Cl--H0)-HO (R-Cl--HO)+(OH)free $ R++Cl-:j ‘HO Scheme 2 explanation of the effect on reaction rates of the addition of an alcohol or basic aprotic co-solvent which will reduce the concentration of (OH), [Equation (5)] and RH + (LP)free + R-+ (H--LP)’ (4) Base + (0H)free + Base--(HO) (5) consequently increase that of (LP)free. SN2displacements which are relatively insensitive to the addition of co-solvents were explained on the basis of a two-stage process (Scheme 3) in which relative changes in the concentrations of (OH),, and (LP)freenearly compensate one another.R-hal+(OH)free T! Rhal--(HO) (LP)free+ Rhal--(HO) -+ Products Scheme 3 Investigations of the controversial role of a-participation in the solvolysis of exo-2-norbornyl systems” have continued unabated and further evidence has been produced in support of the view that high exolendo rate ratios are steric in origin. Comparisonsof solvolytic data for em- (1)and endo-2-norbornyl arenesulphonates (2) with those for adamantyl derivatives were made in aqueous ethanol or trifluoro- K. C. Westaway Cunud. J. Chem. 1978,56,2691. M. C. R.Symons J.C.S. Chem. Comm. 1978,418. lo H.C. Brown and P.von R.Schleyer ‘The Nonclassical Ion Problem’ Ch. 4 Plenum Press New York,N.Y. 1977. Reaction Mechanisms x (1) (2) ethanol,” and in a wide range of other media of differing nucleophilicities and ionizing powers ranging from methanol and ethanol through various aqueous solvents to hexafluoroisopropanol and trifluoroacetic acid. l2 The excellent cor- relations which were obtained indicated that there was no significantly larger solvent contribution to the rates of solvolysis for the 2-norbornyl systems than for 2-adamantyl a standard k substrate. The similarity in solvent effects on both the endo-and ex0 -2-norbornyl systems was also taken as strong evidence against participation in the latter by either solvent or neighbouring carbon.Apart from the semantic problem of whether such processes should be designated k unless there is absolutely no participation involved,12 it is clear that the results are in conflict with the usual interpretation of high exolendo rate ratios which is based on distinctly different kA and k processes. Increasing electron demand has also been used as a tool to show lack of significant electron release from exo-2-norbornyl to the a-carbon of aryl (em-2-norbornyl) methylcarbinyl p-nitrobenzoates (3) during solvolysis in aqueous acetone. l3 The endo-isomer (4) in this case solvolyses four times faster a result attributed to greater Ar (4) relief of steric strain on ionization. The importance of steric factors in other examples of this type has been demonstrated by comparison of reaction rates with those for various sterically hindered acyclic analogues.l4 Allylic Substitution.-A reinvestigation of certain aspects of the solvolysis of 1-and 3-substituted ally1 chlorides” has produced new evidence in support of the contro- versial sN2’ mechanism [Ann. Reports (B) 1971 68 2521. Product composition studies and in particular the results of work with optically active substrates (+)-RCHClCH=CH2 (R =Me or Bui) indicated that in anhydrous alcohols both the rearranged and unrearranged products were formed in concerted reactions. The stereochemical consequences of sN2’ reactions appear to cover the whole range of possibilities from syn-to anti-displacement according to the nature of the displacing and displaced groups counterions and solvent.l6 Preferential syn-attack (by a J. M. Harris D. L. Mount and D. J. Raber J. Amer. Chem. Soc. 1978,100,3139. I2 H. C. Brown M. Ravindranathan F. J. Chloupek and 1. Rothberg. J. Amer. Chem. Soc. 1978,100,3143. l3 H. C. Brown and M. Ravindranathan J. Org. Chem. 1978,43 1709. l4 H.C.Brown and M. Ravindranathan J. Amer. Chern. Soc.,1978,100,1865. l5 C.Georgoulis and G. Ville J. Chem. Research (S),1978 248. l6 T.Oritani and K. H. Overton J.C.S. Chem. Comm. 1978,454,and cited references. H. R.Hudson factor of 1.4-1.8) was demonstrated in the aminolysis of (R)-or (S)-[1-2H]oct-l- en-3-y1(5; X = n-C5H,,) or [1-2H]but-l-en-3-y12,6-dichIorobenzoate(5;X = Me) by (R)-or (S)-a-methylbenzylamine (Scheme 4). The results indicated a difference (5) Scheme 4 in activation energies for the two processes of not more than 0.5 kcalmol-'.The syn-geometry which has recently been demonstrated for the reaction of a-methyl- ally1 chloride with diethylamine" was therefore considered16 to reflect more effective hydrogen-bonding between the amine and allylic chlorine. In contrast intramolecular SN2'displacement of 2,4,6-trimethylbenzoate by an anionic nucleo- phile (RS-)proceeded entirely with anti-geometry (Scheme 5).Is Me Scheme 5 Vinylic Substitution.-The role of vinyl cations in solvolytic displacements has been reviewed" and the first examples of vinyl cations as intermediates in the solvolysis of vinyl fluorides have been reported.20 In buffered aqueous ethanol or trifluoro- ethanol the fluorides [(6)-(9)] gave the corresponding ketones together with (6) (7) (8) (9) Ar = p-MeOC6H4 Ar = p-MeC6H4 various elimination and/or rearrangement products.High negative entropies of activation dependence of reaction rates on solvent ionizing power and Winstein- R. M. Magid and 0.S. Fruchey J. Amer. Chem. SOC. 1977,99,8368. '' G. Stork and A. F. Knight J. Amer. Chem. SOC. 1977,99,3851. M. Hanack Angew. Chem. Internat. Edn. 1978 17 333; P.J. Stang Accounts Chem. Res. 1978 11 107. *' L. Eckes and M. Hanack Chem. Ber. 1978,111 1253. Reaction Mechanisms 53 Grunwald values of 0.53 [for (S)] and 0.86 [for (9)] were indicative of a mechanism which involved rate-determining formation of a vinyl cation as first intermediate.Additional information an the effect of bulky substituents on the properties of vinyl cations has been obtained by studying the solvolysis of 9-(a-chloroviny1)-anthracene In 80 YO ethanol 90 YO acetone and acetic acid the mechanism of reaction was entirely sN1 unlike that for the solvolysis of (R)-(a-chloroviny1)- anisole in which the AdE-E route contributes appreciably. The reaction showed a common ion rate depression not previously observed for a-arylvinyl systems devoid of P-substituents and it was thought that the relative stability of the sN1 intermediate (11) was a consequence of the fact that maximum conjugative interaction between the aromatic ring and the vinyl cation will occur when the ortho hydrogens and the &hydrogens are in perpendicular planes.In contrast stabilization of the AdE intermediate (12)by conjugation will be associated with steric interference between the ortho hydrogen atoms and the coplanar methyl and chlorine groups. HVC *=H CI Me II \/ CCI=CH2 H C+ H H C+ H m(yJ$pJyJ // // (10) (1 1) (12) The alkylation of aromatic compounds by vinyl derivatives is well known although the mechanisms of these reactions have received little attention. In a recent study of the reactions of a series of monosubstituted benzene derivatives CsHsX (X= MeO Me F H or C1) with vinyl triflates [R:C=CR20S02CF3 (R' = R2= Ph; R1 = Me R2= Ph; R' = R2= Me); cyclohexenyl cycloheptenyl and cyclo-octenyl triflate] an unusually low p value (-2.57) has been obtained.22 The result indicated the involvement of a highly reactive and intermolecularly non-selective electrophilic intermediate uiz.the vinyl cation and it was suggested that this intermediate may also be involved in alkylations with vinyl halides vinyl esters and alkynes. Vinyl substitutions which involve neighbouring group participation by sulphur iodine or phenyl etc. proceed with net retention of configuration as the result of a double inversion process (Scheme 6). It has been pointed out that each of these Y R' Y*+ Y+ 'I \/ / /\ -+ R~-C=C\-R~ + R~-C=C-R~ R 'X*-/\ Y= S I Ph etc. Nu*-R2 Nu Scheme 6 21 Z. Ramport P. Shulman and M. Thuval (Shoolman) J. Amer. Chem. SOC..1978,100,7041. 22 P. J. Stang and A. G. Anderson J. Amer. Chem. Soc.,1978,100 1520.H. R.Hudson displacement steps provides an interesting example of a transition state involving planar four-co-ordinate carbon.23 The stability of this transition state compared to that of the more common tetrahedral arrangement was attributed to enforced reduction in bond angle by the small ring and to delocalization of the lone pair by rr-conjugation over the molecular rr-orbital of the double bond. 3 Carbocations Preparations and Properties in Super-acid Media.-The combined use of 13C and 'H n.m.r. spectroscopy has continued to provide detailed information on the behaviour of carbocations in super-acid media. In the case of simple tertiary cations which undergo rapid degenerate rearrangement by 1,2-hydride or 1,2-methide shift sharp n.m.r.spectra have previously been observed down to -160 "C. The use of high field (67.9 MHz) 13C n.m.r. has now made possible the observation of line broadening at temperatures of -110 to -139 "C for the ions (13)-(19).24sRate~ of degenerate rearrangement have thus been measured and the free energies of activation have been calculated (AG*= 3.0-3.7 kcal mol-'). Ions (13)and (14) showed an inverse @-deuterium isotope effect which suggests that the methyl C-H force constants are greater (and that hyperconjugation is therefore less) in the transition state than in the ground state. The energy difference between a protonated cyclopropane and a proton bridge is reflected in the lower rate of methide shift in (18) than of hydride shift in (13). The still lower rate of rearrangement by hydride shift in (17) is correlated with the need for an accompanying change in conformation of the six-membered riag and in (15) with a steric barrier to rotation about the C3-C4 bond.The low temperature "C n.m.r. spectrum of the l-methylcyclobutyl cation has also been re-e~amined.~~ At -158 "C the ion (which at -80 "C undergoes rapid equilibration to give an averaged methylene signal) was previously claimed to be 'frozen out' and was assigned the unusual sp3 hybridized carbocation structure (20) on the basis of high 13C shielding.26 Comparison with other strained non-planar carbocations which do not show similar high shielding now suggests that the above 23 Z. Rapport Tetrahedron Letters 1978 1073. 24 M.Saunders and M.R. Kates J. Amer. Chem. Soc. 1978,100,7082. '' G.A.Olah G. K. S. Prakash D. J. Donovan and 1. Yavari J. Amer. Chem. Soc.. 1978,100,7085. 26 R.P.Kirchen and T. S. Sorensen,J. Amer. Chem. SOC.,1977,99,6687. Reaction Mechanisms interpretation is incorrect and that the ion has a a-delocalized structure containing two different methylene groups [S72.7 and -2.83].*' The high-field signal is indicative of five-co-ordinate carbon. It was not however possible to distinguish between a single symmetrical puckered species (21) and a degenerate set of rapidly Me &A .:+,* Me %; H-'H (20) (21) equilibrating bicyclobutonium-like ions (e.g. Scheme 7)with the symmetrical ion (21) as intermediate. It was emphasized that other a-delocalized ions (e.g.2-norbornyl) may not necessarily be static symmetrically bridged ions as has been claimed," but could be still equilibrating non-classical delocalized types. Scheme 7 A study of the conformations of t-cycloalkyl cations [(22) (23) (24)] in super-acid media has suggested that the chair conformations are relatively less stable than those (22) R = Meor n-C4H9 of the neutral analogues (i.e.the cycloalkanones and methylenecycloalkanes).27The twist-boat conformation (26)for each of the t-cyclohexyl cations (22;R = Me or Bun) is more stable than the corresponding chair form (25) by ca. 0.5 kcal mol-' (Scheme 8). In the 13Cn.m.r. spectrum of the methylcycloheptyl cation (23),a single peak was observed for each of the a-,p- and y-carbon atoms with no evidence of line broadening down to -130 "C.This result is similar to that for cycloheptanone and is ''R.P.Kirchen and T. S. Sorensen J. Amer. Chem. SOC.,1978,100 1487. 56 H.R. Hudson indicative of a fluxionally mobile system with low pseudorotational energy. In contrast line broadening was observed for the 1-methylcyclo-octyl cation (24) below -80 "C and at lower temperatures the system could be frozen out and assigned the chair-twist boat conformation (27),analogous to that of cyclo-octanone. In the bicyclic series the 9-methyl-9-bicyclo[3.3. llnonyl cation was unstable above -1 15 "C but below this temperature it appeared to consist of an equilibrating mixture of the two boat-chair conformers [(28) and (29)] in contrast to the 9-keto derivative which has a di-chair conformation.Me Me (27) (28) (29) The preferred ground state rotamer conformations of a series of bisected cyclo- propylcarbinyl cations in which the carbocation centre is conjugated to a second cyclopropyl group a vinyl group or a phenyl ring have been shown to be those which involve the most extended conjugation [(30) (33) (35)] unless significant steric factors are In suitable cases it was possible to investigate the equilibrium between each of these structures and the corresponding less stable conformer [(3 l) (34) or (36)] rotational barriers being mainly ca. 8-11 kcal mol-'. Steric hindrance played a significant role in determining the preferred conformation of the phenyldicyclopropylcarbinyl cation (32; R =Ph) in which interaction between the ortho-hydrogen of the phenyl group and one of the cyclopropane rings is minimized.In the less stable conformer (30; R=Ph) the phenyl group was shown to be non-coplanar with the carbocation centre. R R * -.+.-A R R (35) (36) Enthalpies of formation of simple alkyl cations from alkyl halides and antimony pentafluoride have been determined for the first time.29 The calorimetric measure- ments were made in SO2C1F usually at -55 "C. Values for the chlorides showed a very good linear correlation (of slope ca. unity) with the differential limiting free energiesof solvolysis for the same halides in ethanol. Support is thus given to the use of enthalpies of ionization and rate constants as guides to carbonium ion stabilities.28 N. Okazawa and T.S. Sorensen Canad. J. Chem. 1978,56,2355. 29 E. M. Arnett and C. Petro J. Amer. Chem. Soc. 1978,100,2563. Reaction Mechanisms Attempts to correlgte I3C chemical+shifts with c+constants for a ra9ge of substituted cumyl (ArCMe,) styryl (ArCHMe) and benzhydryl cations (ArCHPh) [Ar=p-XC6H4 (X=H MeO Me F C1 Br or CF,); m-XC6H4 (X=Me or F); 3,5-(CF3)&H3] have shown that for carbon at or adjacent to the ionic centre considerable deviations from linearity may A better correlation was obtained with an enhanced substituent constant (a")which described the effects of resonance-stabilizing substituents in super-acid media. Caution was therefore urged in the use of 13C n.m.r. chemical shift correlations with single substituent parameters to reach conclusions on the mode of stabilization of carbocations.Good correlations were however obtained31 in a range of 120 p.p.m. between I3Cchemicql shifts for C in substituted diphenylmethylcarbenium [(p-Xc$%)(p-YC6&)CMe] and diphenylhydroxycarbenium ions [( p-XC6&)( p-YC&C.OH](X and/or Y =MeO Me F C1 H CF, or NOz)(the latter in 98 % sulphuric acid) and INDO .rr-electron densities for C,. Rearrangements in Deamination and Solvo1ysis.-Em-and endo-bicyclo-[4.1.O]heptane-7-diazonium ions have been shown to favour synchronous disrotatory ring ~pening.~' Generated by diazotization of the corresponding amine and in the presence of sodium bromide the exo-isomer (37) gave mainly 7-exo-bromobicyclo[4.l.0]heptane (39; X =Br) whilst the endo-isomer (38)gave cyclohept-2-en-1-01 (40; X= OH).The latter was formed by hydrolysis of the corresponding bromide. When the same two ions were prepared by the action of weakly alkaline methanol on the corresponding nitrosoureas the same proportions of the products (39; X =OMe) and (40; X =OMe) were obtained in each case the cis-3-methoxycycloheptene being shown to arise by isomerization of the first formed trans-isomer (41). The results are consistent with a reaction scheme which involves a (37) (38) (39) (40) (41) partially opened cyclopropyl cation (42) as first intermediate in the formation of both products (Scheme 9). An analogous 'bent' cyclopropyl cation was reported previously on the basis of I3C n.m.r. studies in super-acid [Ann.Reports (B) 1977,74,78].Similar intermediates were thought to occur in the decompositions of exo- bicyclo[4.1 .O]hept-2-ene-7 -diaz~nium~~ and exo- bicyclo[S .1.O]oct-2-ene-8-diazonium ions.34 In contrast the em-bicyclo[S. 1.O]octa-2,4-diene-8-diazonium ion (43)appears to undergo a novel conrotatory ring-opening process to give the cyclo-octatrienyl cation in a configuration (44)which facilitates subsequent ring closure to yield bicyclo[3.3.0]octadiene derivatives (Scheme 10) instead of the energetically more favourable homotropylium ion.,' 30 D. P. Kelly and R. J. Spear Austral. I. Chem. 1977,30 1993; 1978,31 1209. 31 B.Ancian F.Membrey and J. P. Doucet J. Ore. Chem. 1978,43 1509. 32 W.Kirmse and H. Jendralla Chem. Ber. 1978 111 1857 and cited refs.33 W. Kirmse and H. Jendralla Chem. Ber. 1978,111 1873. 34 W.Kirmse and U. Richarz Chem. Ber. 1978,111.1883. 35 W.Kinnse and U. Richarz,Chem. Ber. 1978,111,1895. H. R.Hudson Scheme 9 (44) Scheme 10 Neighbouring group participation by the cyclopropane ring has been shown to constitute a significant reaction pathway in the solvolysis of 2-cyclopropylethyl tosylate in the weakly nucleophilic triflu~roethanol.~~ Apart from the major products which were formed either by direct nucleophilic displacement by solvent or via 1,2-hydride shift cyclopentyl derivatives were obtained in which deuterium scrambling had occurred and it was concluded that an intramolecularly corner- alkylated cyclopropane intermediate (45) was involved (Scheme 11).A similar R =CF3CHz Scheme 11 route with cyclopropane participation enhanced by over 20 YO,was thought to be followed by the analogous 2-( 1-methylcyclopropy1)ethylt~sylate.~' An intermediate of this type would be analogous to the corner-protonated cyclopropanes which have been reported elsewhere [Ann.Reports (B) 1973,70 1361. '' I. M. Takakis and Y. E. Rhodes Tetrahedron Letters 1978 2475. 37 Y. E. Rhodes I. M. Takakis P. E. Schueler and R. A. Weiss Tetrahedron Letters 1978,2479 Reaction Mechanisms An unequivocal demonstration of sequential rearrangement involving a bridged phenonium ion has been provided by the solvolysis in aqueous dioxan of the tosylate of (-)-(R)-3-methyl-2-phenylbutan-l-o1(46; X =OTs) and by the deamination of the corresponding amine (46;X=NH2) with nitrous acid.38 In addition to the expected major product (47),arising from 1,2-phenyl shift products [(48)and (49)J were formed; these were explicable on the basis of further rearrangement in which either hydrogen or methyl acted as an internal nucleophile for displacement of phenyl (Scheme 12).In the course of deamination a 1,2-shift of isopropyl also occurred. HO ,H H' Ph (46) (47) Me H Ph + 1 1 4-m FPh OH OH C (49) Scheme 12 Side reactions in the preparation of cyclo-octa-2,7-dienone oia elimination from the trans-2,8-dibromocyclo-octanone ketal(50) provide what appears to be a novel example of homallyl-cyclopropylcarbinyl cation rearrangement under strongly alk- aline condition^.^' Heated under reflux with methanolic sodium hydroxide thc ketal (50)led to formation of minor products identified as the orthoesters (55) and (57).It was assumed that the elimination of one molecule of hydrogen bromide gave an intermediate ketal(5 1)which could ionize under the polar conditions used to give the homoallyl cation (52) and its cyclopropylcarbinyl counterpart (53). Rearrangement to the oxygen-stabilized cations (54) and (56),followed by capture by methoxide accounts for the products formed. 4 Carbanions The stabilities of carbanions in solution are generally expected to decrease with increasing alkyl substitution in accord with inductive electron release. The tri-t- butyl carbanion (59) would thus be highly unstable.Nevertheless evidence for its '* W. Kirmse and B.-R.Gunther J. Arner. Chem. SOC.,1978,100,3620. 39 H. 0.Krabbenhaft J. Org. Chem. 1978,43,4556. H. R. Hudson Br OMe (54) (55) MeOT WO] (53)- OMe (57) Scheme 13 formation in solution has been obtained in the interaction of trj-t-butylmethanol with dimsylate anion in DMSO (Scheme 14).40 The drivingforce for the formation of the carbanion from the first-formed alkoxide (58) was thought to be the release of steric strain. It was noted that the heat of reaction (-23.2 kcal mol-') was very close to the difference in calculated strain energies for tri-t-butylmethane (40.4 kcal mol-') and 1,l-di-t-butylethane (15.0 kcal mol-') which are sterically similar to the alkoxide (58) and its other decomposition product di-t-butyl ketone (60),respectively.BuiCOH + MeSOCH2-K' + C +Me2S0 Bu" Bu)iJ (58) 1 Me SO Me3CH 2Me& -K' + Bu$CO (59) (60) Scheme 14 Quantitative information on the effects of a-cyclopropyl substituents on carbanion stability has been obtained by studying base-catalysed hydrogen-deuterium exchange in a range of cyclopropylcarbinyl ketones [(61) and (62)].41Relative rates of exchange at 35.5 "C,compared to that for isovalerophenone fell within the range 0.76-50.6. The results showed that cyclopropyl groups exert little stabilizing effect 40 E. M. Amett L. E. Small R. T. McIver jun. and J. S. Miller 1. Org. Chern. 1978 43 815. 41 M. J. Perkins N. B. Peynircioglu and B. V. Smith J.C.S. Perkin II 1978 1025.Reaction Mechanisms $CHR2COPh (61) R'= R2= H; R' = H R2= Me; (62) R = Ph or p-02NC6H4 R' = Ph R2= H; R' = p-OzNC6H4 R2= H. on an adjacent carbanionic centre and that the transmission of substituent effects by cyclopropyl is poor. Vinyl groups on the other hand exhibited a considerable stabilizing influence in keeping with earlier observations on the effects of unsaturated groups. Stabilization by dipole interaction is believed to account for carbanion formation adjacent to oxygen nitrogen or sulphur when the heteroatom is conjugated to carbonyl (Scheme 15).42The importance of the carbonyl group in providing this I1 Base e.g. s-BuLi 1 -AAr-C,r-C, Ar-C \x/ Me ,CH in TMF or TMEDA X \v/ A X = 0,NR or S TMEDA = tetramethylethylenediamine Scheme 15 stabilization has been demonstrated in the case of thioesters by showing that the equilibrium obtained by adding methylthiomethyl-lithium to methyl 2,4,6-tri-iso-propylbenzoate (Scheme 16) is displaced to the right by at least one order of magnitude.43 4-P \/ \ SMe~2H8~HF,at-780c,~c40 \SCH Li,TMEDA + + MeSCHzLi,TMEDA MeSMe Scheme 16 pp-Diphenylacrylonitrile (63) has been shown to be a good model substrate for studyingthe formation of vinyl carbanions by reaction with strongbases.44 The steric effect of the two phenyl groups prevents the alternative possibility of nucleophilic addition at C, as occurs in the Michael reaction and the equilibrium with n-butyl-lithium (Scheme 17) is reached rapidly at -78 "C.Competing nucleophilic addition at C occurred however to give the carbanion intermediate (64) which 42 P. Beak and B. G. McKinnie J. Amer. Chem. Suc. 1977,99,5213; P. Beak B. G. McKinnie and D. B. Reitz Tetrahedron Letters 1977 1839; P. Beak and D. B. Reitz Chem. Rev. 1978,78 275. 43 D. B. Reitz P. Beak R. F. Farney and L. S. Helmick J. Amer. Chem. Soc. 1978 100 5428. 44 U. Melamed and B. A. Feit J.C.S. Perkin I 1978 1228. H. R.Hudson Ph,C=CHCN + BuLi $ Ph2C=CCN Li' + BuH (63) Scheme 17 formed products by loss of either cyanide or hydride ion (Scheme 18). Elimination of hydride under these circumstances was unexpected in view of the better leaving ability of cyanide and it was assumed that a suitable acceptor e.g.a molecule of alkene (63) must be involved. Ph2C=CHCN + BuLi S Ph2CCH(Bu)CNLi' (63) (64) 7I-.- Ph2C=CHBu Ph2C=C(Bu)CN Scheme 18 Kinetic isotope effects (kH/kD)for the interactions of carbanions with MeOH or MeOD have been obtained in terms of product isotope effects (pie.) in the reactions of substituted benzyltrimethylsilanes with methanolic sodium methoxide (Scheme 19)."' Although an overall increase in kH/kDwas observed with increasing reactivity MeO-+ Me3SiR -* MeOSiMe3+ R-R-+ MeOH (orMeOD) + RH (orRD) + MeO-R = XC6H4or XYC6H3 Scheme 19 of RSiMe3 (and thus with pK for the corresponding carbon acid RH) the values remained effectively constant at 1.2-1.3 for X or XY = H C1 rn-CN m-NO, or 3,5-C12. Intermediate values were observed for X = p-CN (2.0) and p-SO,Ph (2.9) followed by a steep increase to 7.0 for X = p-COPh and to 10 for X or XY = #-NO2 p-N02 2-Me-4-N02 or 2-Me-6-N02.These results together with redetermined p.i.e. values for Ph2CHSiMe3 and Ph3CSiMe3 of 1.3 * 0.2 are inconsistent with an earlier literature k,/k value of 4.2 for proton abstraction from triphenylmethane (and hence by calculation of ca. 8.1 for the reverse reaction). The reason for this discrepancy is however not clear. Other aspects of carbanion chemistry are referred to in the following section on Elimination Reactions. 5 &Elimination For E2 eliminations from @-aryl-activated alkyl halides ArCH2CH2X(X = C1or Br) values of p (kH/kD)B,and k,,/k,- are very similar whether the attacking base is phenoxide or the much bulkier 2,6-di-t-butylphenoxide The steric require-45 D.Macciantelli. G. Seconi and C. Eaborn J.C.S. Perkin 11,1978 834. 46 S. A. E. Baciocchi and P. Perucci J. Org. Chem. 1978 43 2414. Reaction Mechanisms ments of the base appear therefore to have little effect on transition state structure. Some of the dangers inherent in relating kinetic isotope effects to transition state geometries have been reported as a result of studies of methoxide-induced elimina- tions from a number of 1,2-diaryl-l-chloroethanes,p-RC6&CH2CHPhC1 (R = NO2,C1,H or Me) substituted with deuterium at the a!-and/or p-po~ition.~~ Certain features of the reactions were in agreement with expectation from previous work. Thus Hammett plots confirmed the carbanionic character of CB in the transition state primary kinetic isotope effects (kH/kD)B increased from 2.1 to 8.8 with change of R in the order Me<H<C1<N02 (i.e.as the transition state became more reactant-like) and secondary isotope effects (kH/kD) changed in the opposite direction from 1.375 (R = Me) to 1.08 (R = NO2). However high zero-point energy differences (E -ED)B == 13 kJ mol-' and abnormally low pre-exponential factors AH/AD= 0.001 to 0.08 indicated that proton tunnelling or an internal-return mechanism may be complicating this E2 elimination. Furthermore the energies of activation showed different reactant-like relationships for the transition states from those based on the kinetic isotope effects. Further information has been obtained on the factors controlling elimination from carbanion intermediates (Scheme 20; R' = R2= H).The effects of the activating group G (PhS02 CN or Bz) on leaving group ability are seen to be c~mplex.~' G.CHR'.CHR~.Z+B- kl k2G.CR'-CHR~.Z+GR'C=CHR* + :Z k-1 Scheme 20 Although similar ranking orders for the leaving ability of Z were obtained in some cases e.g. PhAMez>OPh >SPh >OMe > CMe2N02 for G = CN or PhS02 and OMe >NRAc >CN CMe2N02 for G = Bz or PhS02 there were variations in others and it was not possible to make simple correlations of orders with any molecular parameters. The results suggest that there is little cleavage of the bond to the leaving group in the transition state in which Z is expelled; this conclusion is supported by a lack of sensitivity to changes in solvent and to ion association.Substituent effects in these reactions (Scheme 20; Z = OPh OMe S02Ph; R' R2= Me and/or Ph) have also been determined and are in accord with the same general For nitriles and sulphones (G= CN or PhS02) the (ElcB)R mechanism operates (k2 is rate-determining) and both methyl and phenyl substituents at C or Cpgive rise to mild acceleration of the expulsion of the leaving group. Ketones (G= Bz) react by the irreversible (ElcB)I route and the effect of substitution is on the rate-determin- ing deprotonation only. The reactions in these cases are sensitive to steric inter- ference by substituents at CBbut not at C,. Markedly different effects of P-phenyl substitution on the rates of formation of carbanions from nitriles ketones and sulphones have been discussed in terms of steric interference with the formation of a planar carbanion.47 F. M. Fouad and P. G. Farrell Tetrahedron Letters 1978,4735. 48 D. R. Marshall P. J. Thomas,and C. J. M. Stirling,J.C.S. Perkin 11,1977,1898; P. J. Thomas and C. J. M. Stirling ibid. 1978 1130. 49 R. P. Redman P. J. Thomas and C. J. M. Stirling.J.C.S. Perkin IZ 1978 1135. H. R.Hudson The first quantitative information on the contribution of ring strain to an elimina- tive ring opening has been reported for the reaction of the sulphone (65) with ethanolic sodium ethoxide (Scheme 21).50Ring strain accelerated the elimination to Scheme 21 such an extent that 6-deprotonation became rate-determining i.e.the mechanism was (ElcB), as shown by a primary isotope effect (kH/kD)of 2.5. Compared to the acyclic analogue P-methoxypropyl ethyl sulphone which undergoes elimination by the (ElcB) mechanism the rate was 2.46X lo6times faster. Elimination reactions of all types in which expulsion of a leaving group is accompanied by ring fission have been reviewed.51 6 Electrophilic Addition The most common topic for investigation with respect to both alkenes and alkynes has been that of open versus bridged transition state structures and methods for their differentiation. As one approach to this problem in the alkene series it has been proposed that acid-catalysed hydrations [Equation (6)]and the addition reactions of arenesulphenyl chlorides [Equation (7)] should be taken as models for these two \/ HO I1 c=c S \C+-C-H AHO-C-C-H+H+ (6) / \ slow / I II Ar I.Ar S extreme reaction types.52 From these it should then be possible to obtain informa- tion on the structures of rate-determining transition states for other electrophilic additions e.g. bromination by comparison of structure reactivity profiles. The method was illustrated by showing that the logarithms of rate constants for bromina- tion in methanol (kgBr2)gave a poor correlation with those for acid-catalysed hydration (kZH+). On the other hand a good correlation was obtained with log kZArSC' for the addition reactions of 4-chlorobenzenesulphenyl chloride with a wide range of alkenes (excepting a few in which steric or other special features were present).The result is consistent with the usually accepted bromonium-type struc- ture for the rate-determining transition rate for brominations and the method was advocated for the testing of other controversial mechanisms such as those of oxymercuration and thallation. R. J. Palmer and C. J. M. Stirling J.C.S. Chem. Comm. 1978,338. C. J. M. Stirling Chem. Rev. 1978,78 517. 52 G. H. Schmid and T. T. Tidwell J. Org. Chem. 1978,43,460. Reaction Mechanisms Further confirmation of the symmetrical nature of the transition state in alkene brominations has been obtained by the use of a new procedure of internal comparison which avoids resorting to external structural scales and substituent constant^.^^ This involved the measurement of rate constants for the brominations in methanol of four isosubstituted series of alkenes RCH=CH2 trans-RCH=CHMe RMeC=CH2 and RCH=CMe2 (R=Me Et Prh PhCH2 CH3C02CH2 or C1CH2).Plots of log k for each series against log k for the first series (RCH=CH2) gave sets of parallel lines. The result is most significant for those alkenes (RMeC=CH2 and RCH=CMe2) which are the ones most likely to involve carbo- nium-ion-like pathways. It was argued that since the carbocatio? intermediates which eacp of these two series would involve would be RMeC-CH2Br and RCHBr-Me2 respectively they would be expected to respond differently to the polar influence of substituent R. The parallel plots showed however that this was not so and it was concluded that the transition state was a symmetric entity.A slight secondary effect that was detected was attributed to hyperconjugation. The electrophilic bromination of alkynes is generally considered to follow a course (Scheme 22) which is similar to that for alkenes. Recent of solvent and AaH products Scheme 22 structural effects on rates of bromination in acetic acid formic acid trifluoroacetic acid and methanol-water mixtures show that nucleophilic solvent assistance cas be important in the reactions of alkylacetylenes e.g. hex-l-yne and hex-3-yne whereas arylacetylenes are virtually unaffected by solvent changes. The alkylacetylenes were considered to react through a bridged transition state with assistance from hy- droxylic media e.g. as shown [(66) or (67)].A bridged intermediate (68)was also SH ‘0’ s R-C=,C-R ,‘a+.\ I R-C=C-R ‘d?-:Bk Brh-X-__. I, . / Br --Br*-H (66) (67) indicated in the addition of bromine monochloride to hex-l-yne which yielded only the trans-isomers of both the Markownikoff and anti-Markownikoff addition products (Scheme 23).55 Structural effects on the rates of bromination in methanol of a series of substituted diarylacetylenes pointed to the involvement of a highly 53 E. Bienvenue-Gotz and J. E. Dubois Tetrahedron 1978 34 2021. 54 G. Modena F. Rivetti and U. Tonellato J. Org. Chem. 1978,43 1521. 55 V. L. Heasley D. F. Shellhamer J. A. Iskikian and D. L. Street J. Org. Chern.,1978 43 3139. 66 H. R. Hudson R-C~C-H BrCl -+ +Br /’ R-cLC-H (68) R Br Br H a-\ / \/+ c=c H R’ \Cl/c=c\-+ c1 R = n-CdH9 (90 Yo) (10 Yo) Scheme 23 asymmetrical vinyl-cation-like transition state in which the positive charge developed essentially on the carbon which was attached to the better substituted benzene ring (69)[(-a&) > (-at)].’“ (69) Evidence for the intervention of a bridged phenonium ion has been obtained in the bromination in acetic acid of 3-(p-methoxyphenyl)propyne which yielded mainly the product of syn addition (Scheme 24) unless lithium bromide was present when the AdE3 mechanism operated and anti addition occ~rred.’~ The intermediate (70) H Br2 Br -/ _I* p-MeOC6H4CH2CrCH CH,-C=CHBr -p-MeOC6H4CH2C=C I (70) Br ‘Br Scheme 24 was described as an aryl-stabilized vinyl cation.With less electron-attracting groups in the ring (H p-Me rn-CF,) the products were almost entirely trans-isomers and were presumably formed via cyclic bromonium intermediates. 7 Nucleophilic Addition to Carbonyl Compounds The first systematic study has been made of the effect of structure upon equilibrium constant for the aldol condensation [Equation (8)].” The number of data that can be CH3COR*+ R2COR3 $ R2R3C(OH)CH2COR’ (8) obtained directly for this reaction are limited by several factors such as easy dehydration in certain cases of the aldol product to the corresponding enone and the fact that cross-product formation from two reactants might not occur in the desired direction. Experimentally accessible equilibrium constants have therefore been ’‘ J.A. Pinock and C. Somawardhana Canad.J. Chem. 1978,56 1164. ’’ J. P. Guthrie Canad.J. Chem. 1978 56 962. Reaction Mechanisms augmented by values calculated from thermodynamic data for the enones and their precursors equilibrium constants for the dehydration reaction and estimated free energies of formation based on hypothetical disproportionation reactions. In this way a full set of equilibrium constants was obtained for the condensations of acetaldehyde acetone acetophenone and acetic acid (acting as nucleophiles) with formaldehyde acetaldehyde benzaldehyde acetone and acetophenone (acting as carbonyl acceptors). The results showed that a previously determined free-energy relationship for the additions of alcohols thiols amines hydrogen cyanide and bisulphite to carbonyl compounds (E.G. Sander and W. P. Jencks J. Arner. Chern. SOC., 1968,90,6154) can be extended to include the aldol reaction. Prediction of the equilibrium constant for any aldol condensation therefore becomes possible from a knowledge of that for the addition of another nucleophile such as water or HCN to the acceptor and the y-value of the carbonyl nucleophile. The equilibrium and kinetics for cyanohydrin cleavage and formation in aqueous solution at 25 "C and at ionic strength 1.0 M have been determined for a range of substituted benzaldehydes (4-N02; 3-C1,4-C1; 4-C1; 4-H; 4-Me; 4-Me0; 4-MezN).'* The rates of formation and breakdown showed no significant general-base-catalysed contribution in the pH range 2.5-7.4 and it was concluded that neutral reactants and neutral products were involved.In the breakdown of the cyanohydrin complete removal of the proton to form the oxyanion was necessary for the CN-group to depart. The major contribution of the substituent effect to the equilibrium constant was its effect on the stability of the reactant aldehyde. However the anionic cyanohydrins were less sensitive to electronic effects than were the neutral cyano- hydrins the a-carbon atom of which is relatively more electron-deficient. There have been a number of recent investigations into the mechanism of carbonyl reduction by borohydride. The overall process is first order in each of the reactants and is generally considered to proceed by a stepwise series of hydride transfers to yield R2CHOBH3- (R2CH0)2BH2- (R2CHO),BH- and (R2CHO),B- followed by ,&ydrolysis of the tetra-alkoxyborate.Previous evidence for this mechanism and for other possibilities has been s~mmarized.~~ Strong support for the stepwise process has been obtained by the isolation of sodium tetrakisbenzyloxy- borate from the reduction of benzaldehyde in DMS0.59 Reductions of acetone pivaldehyde and benzaldehyde in DMSO and DMSO-water mixtures all showed a marked fall in reaction rate as the water content of the medium was decreased indicating the importance of hydrogen-bonding from a hydroxylic solvent in the transition state (71). Further detailed information on the mechanism was obtained 6"- ,' ; .> /H H\ '* Wei-Mei Ching and R.G. Kallen J. Amer. Chem. Soc. 1978 100 6119. s9 C. Adams V. Gold and D. M. E. Reuben J.C.S. Perkin ZZ 1977 1466. 6o C. Adams V. Gold and D. M. E. Reuben J.C.S. Perkin ZZ 1977 1472. H. R. Hudson by showing that the reduction of benzaldehyde by tritiated sodium borohydride in DMSO and DMSO-water mixtures resulted in the incorporation of tritium into unchanged aldehyde.60 This implies reversibility of the hydride-transfer process which must therefore precede that in which B-0 bond formation occurs. As there was no evidence for the formation of free borane in the reaction medium it was believed that the initially formed benzyl oxide ion and borane must react competi- tively within a solvent cage (74) without diffusing apart to give either starting materials with hydride exchange or the alkoxyhydridoborate ion (Scheme 25).Alternative formulations for the reaction involving four-centre (72) or six-centre .. (72) (73) (73)transition states are not compatible with this result although the six-centre process had been suggested in order to account for the formation of the alkoxyborate corresponding to solvent alcohol rather than that corresponding to the alcohol derived from the carbonyl compound when reductions were carried out in alcoholic PhCTO +B&-f PhCHO +BH3T-S {PhCHTO- BH3} L (74) PhCHTOBHj-Scheme 25 media.61 Solvolysis of the first-formed alkoxyborate should however be considered as a possibility in this context.” The possible disproportionation of reaction intermediates (Scheme 26) has been excluded by carrying out the reduction of cyclohexanone or 3,3,5,5-tetramethylcyclohexanonewith mixtures of NaBa and NaBD in propan-2-01 when no isotope exchange between the reagents occurred.62 2ROBH3-+ BH4-+(R0)2BH2-3(R0)2BH2-+ BH4-+2(R0)3BH-4(R0)3BH-+ BH4-+3(R0)4B-Scheme 26 This also confirms the stepwise nature of the reduction by successive alkoxyborate intermediates and shows that no more than 25% of the carbonyl compound is reduced by BH,-itself a factor relevant to selectivity.The influence of alkali cations on the mechanisms of nucleophilic additions to carbonyl compounds including metal hydride reductions has been reported by a D. C. Wigfield and F. W. Gowland Tetrahedron Letters 1976,3373;J. Org. Chem. 1977,42 1108.62 D. C. Wigfield and F. W. Gowland Canad. J. Chem. 1978,56,786. Reaction Mechanisms number of workers in recent years.63 Such reactions are considered to occur either under ‘complexation control’ which involves loose ion-pair association with carbo- nyl oxygen [Equation (9)],or ‘association control’ in which tight ion-pairing with the attacking nucleophile is involved [Equation (lo)]. As an experimental criterion to \ I NU-+ C=O--M++ NU-C-0-M’ (9) / I \ I M’NU-+ c=o + NU-C-0-M’ / I determine the dominant factor controlling the reactivity of a carbonyl compound under given conditions the kinetic effect of adding cryptand or crown-ether is These reagents will trap the cation with a consequent decrease in reaction rate as the effect of carbonyl complexation is removed or an increase in rate as the effect of association with the nucleophile is removed.The principle has also been applied to regioselectivity of attack in a-enones but with the added compli- cations of hard or soft character at the two possible sites of attack and in the attacking nucleophile. J.-M. Lefour and A. Loupy Tetrahedron,1978 34 2597 and cited references.