4 Reaction Mechanisms Part (ii) Polar Reactions ~~~~~~ By T. W. BENTLEY Department of Chemistry University College of Swansea Swansea SAZ 8PP 1Introduction This report consists of a selection of topics. Some important items are omitted for this year; in particular nucleophilic substitution at a saturated carbon atom the mainstay of this chapter for many years and which was covered in detail last year [Ann. Reports (B) 1974,71,112] this year makes way for its offspring nucleophilic vinylic substitution. Also the many extensive and continuing studies of carbocations and of neighbouring group participation are not discussed here but comprehensive coverage of the literature on organic reaction mechanisms is presented elsewhere.' 2 Nudeophdic Vinylic Substitution Although simple vinyl halides are unreactive towards nucleophilic substitution the field of nucleophilic vinylic substitution of simple vinylic substrates has developed rapidly since more reactive leaving groups were examined; e.g.one of the most suitable leaving groups is trifluoromethanesulphonate (triflate TfO CF3SO20-) which is displaced over lo4times more rapidly than toluene-p-sulphonate. It is now well established that many nucleophilic vinylic substitution processes proceed under solvolysis conditions by an SN1mechanism (Scheme l) and it has been possible to R' R3 R' R3 \ /R2 slow \ / c=c R2' -R' /c=c+ \x R2 X-rearranged products R1(') -CrC-R3 Scheme 1 'Organic Reaction Mechanisms' 1975 ed. A. R. Butler and M. J. Perkins Wiley London 1976.71 T. W.Bentley exclude plausible alternative mechanisms such as nucleophilic addition followed by elimination.2 The co-occurrence of products derived by rearrangement or elimina- tion in the vinyl cation (or ion pair) intermediate (1)also supports the proposed mechanism (Scheme l) and similar intermediates may be formed by addition to alkynes [see Section 3 below and Ann. Reports (B) 1974,71 1261. 1,2-Hydride shifts a characteristic process in trisubstituted carbocations have now been shown to occur in the disubstituted systems. Amongst the various solvolysis products of the triflate (2) was the ketone (3),which was probably formed by a 1,2-hydride shift across the double bond as shown in Scheme 2.3 Following 4 /Me /Me substitution H/c=c\ A 4/c=c+ + + elimination OTf H OTf 11.2-hydride shift Reagent i CF,CH,OH-H,O 80°C Scheme 2 previous work on 1,2-alkyl shifts towards the positively charged carbon atom in a vinyl cation [Ann.Reports (B) 1970 67 1271 it has now been shown that the corresponding 1,2-hydride shift also occurs (Scheme 3).4 In Schemes 1-3 the Me Me I substitution H-C-C=CH -& H-&-C=CH -+ + I1 I+ elimination Me OTf Me -0Tf 1,2-hydride shift Me Me I \+ I C-CH=CH -+ CF,CH,O-C-CH=CH, / I Me -0Tf Me Reagent i CF,CH,OH-H,O SO°Cb Scheme 3 counter-ion has been added to the authors’ original mechanistic interpretations so that the results are consistent with earlier stereochemical investigations showing preferential inversion of configuration during substitution [Ann.Reports (B) 1972 69 172 3831. These stereochemical investigations were extended to include cyclization processes for which it was suggested that inversion is again preferred 2 L. R. Subramanian and M. Hanack J. Chem. Educ. 1975 52 80. 3 K,-P. Jackel and M. Hanack Annulen 1975 2305; see also C. C. Lee and E. C. F. ICo,J. Org. Chem. 1975,40,2132. K.-P. Jackel and M. Hanack Tetrahedron Letters 1975,4295. Part (ii) Polar Reactions 73 because the (E)-hexadienyl derivative (4) yields more cyclized products [(7) (9) (lo) and (1l)]than the corresponding (Z)-isomer.' In the ion pair derived from the latter the leaving group and the participating double bond would be on the same side of the positively charged carbon atom and so cyclization would have to occur in the linear cation rather than in the ion pair.However both the ion pair derived from (4) and the linear cation are capable of cyclizing. The same four cyclized products [(7) (9) (lo),and (ll)]were also formed in the trifluoroethanolysis of the cyclohexen-4- yl tosylate (8); these results implicate the homoallylic cation (5) as a possible intermediate which may yield the allylic cation (6)by hydride shift.' elimination + substitution HH HH (4) H+ (7) (9) (10) Reagent :i CF,CH,OH 60 "C Scheme 4 When the vinyl group is substituted with appropriately placed aryl groups there is more direct spectroscopic evidence for vinyl cation intermediates. The 'H and 13C n.m.r.spectra of a solution of 1-fluoro-l-(p-methoxyphenyl)-2-methylprop-1-ene and SbFS in S0,ClF at -70 "C provides strong evidence for production of the vinyl cation (12) [Ann. Reports (B) 1974 71 2481 and a similar cation (13) has been generated from the corresponding chloride in a 1 :5 mixture of SbF and SO2C1F at -130°C.6 For both cations [(12),(13)]the n.m.r. spectra show that the methyl (12) (13) groups are equivalent (C2vsymmetry). From the 13C and 'H chemical shifts it was proposed that the favoured conformation of (13) was the bisected one in which the T. C. Clarke and R. G. Bergman J. Amer. Chern. Soc.,1974,96 7934; see also M.J. Chandy and M. Hanack Tetrahedron Letters 1975,4515. 6 S. Masamune M. Sakai and K. Morio Canad. J. Chem. 1975,53,784. T.W.Bentley ring conjugates with the empty p-orbital of the vinyl cation rather than with the double bond; apparently very little charge resides at the @-carbon of the cation. Generation of both cations was also substantiated by quenching experiments the more convincing one being the quenching of (13)in methanolic sodium methoxide at -80 "Cto give the appropriate methoxystyrene in over 50% yield.6 As might be expected evidence for the formation of even more highly stabilized cations can be obtained from the classical kinetic technique common-ion rate depression. During acetolysis of the bromide (14) there is a strong common-ion rate depression within the run or by added bromide ion and it was estimated that >93% of the acetate product arose from the dissociated cation (15).' An An \ /An \+ C=C C=C-An + Br-An = p-methoxyphenyl / Ph ' 'Br Ph (14) (15) Thus the behaviour of vinyl substrates in SN1reactions appears to be very similar to that of other substrates and perhaps the most surprising feature is the thirty year delay before vinyl substrates began to be studied extensively.In contrast to the S,l reactivity discussed above vinyl substrates appear to be much less susceptible to bimolecular nucleophilic substitution than other substrates. 3 Addition There has also been considerable interest in the production of vinyl cations by electrophilic addition to alkynes. In an extension of work on the formation of cyclopent-2-enones by acylation of alkynes[Ann.Reports (B) 1974,71,254] it has been shown that the corresponding reaction (Scheme 5) of the cycloacyl system (16) (19) Reagent i AgBF, CH2CI,-C2H,CI, -60°C Scheme5 2.Rappoport and Y.Apeloig J. Amer. Chem SOC.,1975,97,821 836. Part (ii) Polar Reactions leads to the fluoride (20) rather than to the cyclopentenone (19). These results suggest that the vinyl cation (17) undergoes a novel 1,5-hydride shift to give the secondary cation (18).8 Further work is needed to establish the mechanism of the cyclization to give cyclopentenones which occurs when the developing cation [marked * in (17)] would be primary and part of an acyclic system. Despite their frequent appearance in the more speculative literature there is no satisfactory evidence that primary aliphatic carbocations are formed as reactive intermediates and as they appear to be so unstable it seems likely that this is the source of the divergence of mechanisms.One of the ‘development areas’ of physical organic chemistry is the effect of increased pressure on organic reactions. Recent work on the gas-phase addition of hydrogen chloride to propene at pressures of up to 30 atm suggests that the reaction is first order in propene and at least third order in hydrogen chloride. The results were explained by rate-determining attack of a dimer of HCl on an HC1-propene complex [equations (1)-(3)] HCI +CH3CH=CH2 K2ecomplex (2) (HCI):!+complex A 2-chloropropane+2HC1 (3) The overall rate is then given by equation (4) d[RCl]/d t = k3K1K2[HC1l3[CH3CH=CH2] (4) and the overall rate constant (kobs)by equation (5) kobs = k K K -AI~(AS,+AS~)/R~-(AH,+AH~+E:)/RT3 1 2- (5) where subscripts 1 and 2 refer to equations (1) and (2) respectively and EL is the activation energy for reaction (3).The rate of reaction decreases with increased temperature and this can be accounted for if EL< IAHl +AH2( and evidence that AHl and AH2 were both -2.5 kcal mol-’ was presented. Thus in terms of the simplest form of the Arrhenius equation (kobs=Ae-Ee/RT) the sum of the energy terms in equation (5) corresponds to a negative activation energy. The reaction was envisaged to occur when the alkene complexed on one side by one molecule of hydrogen chloride is attacked on the opposite side by the dimer of hydrogen chloride.’ A six-centre transition state [e.g.possibly (21)]was proposed by analogy with earlier work.” H Me \ HCI / H/,c=c*\H H’ Cl ‘Cl__ -H’ (21) * A.A. Schegolev W. A. Smit V. F. Kucherov and R. Caple J. Amer. Chem. SOC.,1975,97,6604. M. J. Haugh and D. R. Dalton J. Amer. Chem.SOC.,1975,97,5674; see also Ann. Reports (B),1974,71 126. 10 D. L. King D. A. Dixon and D. R. Herschbach J. Amer. Chem. Soc. 1974,% 3328. 76 T.W.Eentley 4 Elimination Many aspects of p-elimination continue to be actively investigated. It is now widely recognized that there is a whole spectrum of mechanisms from the carbanionic ElcB to the E2C which involves some interaction between the base and the a-carbon attached to the leaving group.Movement within the spectrum of mechanisms depends on reactants and reaction conditions (Scheme 6) and thus the various y 6-c-c (22) (23) (24) Electron withdrawal at C -B Electron donation at C 6 Changing to poorer leaving group + Changing to more electronegative leaving group -P Changing C from primary to secondary or tertiary + Increase in base strength + Scheme 6 mechanistic classifications can be merged. However neither the detailed mechan- isms of each classification (ElcB E2H E2C) nor the detailed mechanism of merging seems to be generally agreed. For the E2C reaction the following new hypothesis for the nature of the bonding between the attacking base and the substrate has been put forward." In the transition state (22) the good leaving group X has departed from C but the weak base B-has not 'gained control' of the p -proton to any great extent.Therefore the partial .n-bond was visualized to have a bond order of less than 0.5 and the electron deficiency at C was assumed to be substantial Then assuming that the B....H...-Csystem is non-linear there could be an electrostatic interaction between the base B-(hardly neutralized by bonding to the H) and the electron-deficient C,. This proposal adopts some of the features of Bunnett's E2H spectrum and some of the Winstein-Parker E2C-E2H spectrum. The vital difference between the new proposal and the Winstein-Parker E2C mechanism is the nature of the interaction between B-and C, and among the evidence claimed to favour the electrostatic interaction was a correlation of the rates of elimination and the rates of substitution for a wide range of nucleophiles.Weak polarizable neutral bases (e. g. thiourea) gave more substitution than predicted by the correlation possibly because the electrostatic stabilization shown in (22) was very small.'* However it should be emphasized that the structures (22)-(24) represent transition states and the possibility that the five-co-ordinate structure (25) could be 11 D. J. McLennan Tetrahedron,1975,31,2999. 12 J. R. Pritt and M. C. Whiting J.C.S. Perkin 11 1975 1458. Part (ii) Polar Reactions an intermediate [see Ann. Reports (B) 1974,71 1231 may have been dismissed too readily -the evidence cited against this proposal appears to be only capable of ruling out a common symmetrical intermediate for E2C and SN2processes.Also from studies of secondary solvolyses it appears that there may be an unusually long-range nucleophilic interaction between solvent molecule(s) and the electron-deficient carbon and that although this interaction is moderate in size it is less susceptible to extinction by steric hindrance.'* Thus it is not certain how an E2C would be affected by steric hindrance and earlier criticisms [see Ann.Reports (B) 1973 70 1441 assuming that steric effects in E2C reactions would be of comparable magnitude to those in S,2 reactions may have been premature. At the other end of the spectrum of bimolecular eliminations are the reactions involving the formation of carbanion intermediates the conjugate bases (cB) of the starting materials (Scheme 7).If the expulsion of the leaving group (X) from the BH+ carbanion (27) is rate determining i.e. k-,[BH+] >> k2 the carbanion is formed reversibly and the mechanism is designated (EIcB),. Experimental evidence characteristic of this mechanism is obtained by deuterium exchange with the reaction medium and/or from the absence of a primary kinetic isotope effect for removal of the p -proton. For the (ElcB)R mechanism the observed rate constant kobs is given by kob= kl[B]k2/k-l[BH+]. Assuming that the equilibrium constant (kI/kl) for carbanion formation is independent of the leaving group (X) then observed rate constants for a series of substrates (26) with the same activating group (PhS02) should be propor- tional to k2 which should reflect the leaving group abilities of various X substituents in 1,2-eliminations.Results for an unusually wide range of leaving groups (e.g. CMe2NO2 CN N(Me)Ac OMe S02Ph and 'SMePh) varying in relative rates by at least 10l6show a remarkable lack of correlation with the pK of X-H or the C-X bond strength or the nucleophility of X. Since the sceptics might state that there are now enough empirical scales to enable us to correlate anything it should be emphasized that the results appeared in preliminary form!13 Experimental evidence for 'irreversible' E1cB reactions (for which k2>> k-,[BH'] -Scheme 7) is necessarily less direct as it is difficult to distinguish 13 D. R. Marshall P.J. Thomas and C.J. M.Stirling J.C.S. Chem Comm. 1975 940. 78 T. W.Bentley Ar'CH,SO,ArZ + Et,N & Ar'CHS0,Ar' k-i + Et&H (6) Ar'CHS03ArZ 5Ar'CH=SO + OAr2 H,O + Et,N + Ar'CH=SO Ar'CH,SO; + Et,&H (assuming a steady-state k 1kz[Et3Nl carbanion concentration) kobs = + k-I[Et,NH) + kz Scheme 8 from the classical E2 mechanism. Recent interest in this mechanism can be illustrated by the formation of sulphenes from aryl arylmethanesulphonates (Scheme 8). It was established that the (ElcB) mechanism occurred when Ar2=p-nitrophenyl that the observed reaction rates for a range of electron-withdrawing substituents in Ar' gave a Hammett p-value of 0.54 and that the reaction showed specific base promotion. In contrast when Ar2=2,4-dinitrophenyl p-was 2.7 the reaction showed only general base promotion and deuterium exchange did not occur.For the substrates already established to be reacting by an (ElcB) mechan- ism another Hammett correlation of values of k was obtained from the rates of deuterium exchange. Extrapolation of the correlation line gave predicted values of k which were in good agreement with the observed rate constants for the substrates which did not undergo deuterium exchange. Thiz is in agreement with the 'irrever- sible' ElcB mechanism because when k-,[Et,NH] <k, equation (9) reduces to kohs = k,[Et,N]. Furthermore the effect of added Et,NH was as predicted by equation (9),also consistent with an irreversible ElcB mechanism rather than with a classical E2 mechanism.14 As the departing ability of (OAr2)- increases k increases until it reaches 1013s-' when the 'lifetime' of the carbanion would be less than the vibration time of the S-OAr bond.Then the carbanion would not exist as an identifiable species and the irreversible ElcB mechanism would have merged into a concerted E2 mechanism with carbanion-like character. l5 An extreme case of an ElcB reaction occurs with the cyanide (29) which reacts with trimethylamine to form such a stable carbanion that elimination does not occur. Using pyridine or collidine as bases elimination of CF3CH20-from the carbanion (30) to give the alkene (31)is first order in carbanion and first order in protonated base (BH'). The role of the BH' may be to assist electrophilically in the removal of the leaving group and this new mechanism was designated E2cB.16 14 J.F. King and R. P. Beatson Tetrahedron Letters 1975 973. 15 K. T. Douglas A. Steltner and A. Williams J.C.S. Chem. Comm. 1975 353; see also R. A. More O'Ferrall and P. J. Warren J.C.S. Chem. Comm. 1975,483. 16 M. Albeck S. Hoz,and Z. Rappoport J.C.S.Perkin II 1975 628. 79 Part (ii)Polar Reactions 5 CarbonAadity The acidities of weak carbon acids are relevant to wide areas of organic dhemistry from synthesis to electronic theory. Interpretation of substituent effects in the gas phase is important in evaluating electronic theories of organic chemistry and comparisons between gas phase and solution give information about solvent effects. There has been considerable activity recently in studies of gas-phase effects and dimethyl sulphoxide (DMSO)has been found to be a very useful solvent for studies in solution.One of the advantages of DMSO is its ability to increase the basicity of an aqueous hydroxide so1ution.” Also the pure solvent has a high dielectric constant is strongly dissociating and can be used to determine equilibrium acidities without complications from the effects of ion association.’8 In a short cautionary note Bordwell et al.19 have warned that the determination of acidities from kinetic l2 t. 10 14 18 22 26 30 34 38 42 46 50 Equilibrium Acidities in DMSO (kcal /mole) Figure Equilibrium acidities of carbon acids in DMSO solution plotted against intrinsic gas- phase acidities (Reproduced by permission from J.Amer. Chem. SOC.,1975,97,3226) 1’ W. S. Matthews J. E. Bares J. E. Bartmess F. G. Bordwell F. J. Cornforth,G. E. Brucker,2. Margolin, R. J. McCallum G. J. McCollum and N. R. Vanier J. Amer. Chem SOC.,1975,97,7006. 18 D. W. Earls J. R. Jones T. G Rumney and A. F. Cockerill J.C.S.Perkin ZI 1975 54. l9 F. G. Bordwell W. S. Matthews and N. R. Vanier J. Amer Chem. SOC.,1975,97,442. T. W.Bentley measurements (e.g. deuterium exchange) is particularly susceptible to complications from the effects of ion association and internal return -apparent carbanion stabilities based on kinetic acidities may change by several orders of magnitude depending on the nature of the cation anion and solvent. Consequently discussions of carbanion stabilities should relate to equilibrium acidities although in organic synthesis advantage can be taken of different kinetic acidities [e.g.Ann.Reports (B),1974,71 4451. As expected the degree of dissociation of an acid in solution is much greater than in the gas phase -e.g. the Hammett p value for dissociation of rnetu-and pura-substituted benzoic acids is 1.0in water (by definition) 2.5 in DMSO and 10.0in the gas phase.*' Thus it is surprising (at least to some chemists' expectations) that the gas-phase acidities of a series of nitriles and ketones closely parallel the equilibrium acidities in DMSO (Figure). However with the benefit of hindsight it may be argued that the failure of substituent effects to be attenuated appreciably is probably because the charge is delocalized in each case over the entire anion whereas for carboxylic acids the negative charge is largely localized on the carboxylate anion.21 Related studies of enthalpies of solvation of alkoxide anions in DMSO suggest that steric hindrance to ion solvation becomes increasingly important as the size of the alkyl group increases.22 Comparisons between DMSO and water can also be made -water is a strong hydrogen bond donor and also a strong hydrogen bond acceptor whereas DMSO acts as a strong hydrogen bond acceptor but is only a very weak donor.Consequently oxygen acids (e.g. phenols carboxylic acids) as well as carbon acids where the negative charge resides mainly on oxygen (e.g. nitroalkanes) are more acidic in water than in DMSO because the anions are solvated by strong hydrogen bonding.*' 6 Enolition of Ketones One of the first organic reactions to be investigated mechanistically the halogenation of acetone occurs by rate-determining enolization (except at very low halogen concentrations) and the rate law in acetate-buffered solution is well established [equation (lo)].The term kA[HA]is kinetically equivalent to kL[H,O'][A-] and is thought to represent a mechanism in which protonation of the ketone is followed by removal of a proton by acetate (Scheme 9).23 It was argued that the kB term represented general base catalysis of proton removal to give an enolate anion stabilized by hydrogen bonding to water because there would be no thermodynamic 20 R.Yamdagni T. B. McMahon and P.Kebarle J. Amer. Chem. Soc. 1974 % 4035. *J F. G. Bordwell J. E. Bartmess G. E. Drucker 2. Margolin and W. S. Matthews J. Amer. Chem. Soc. 1975,97,3226. 22 E. M. Amett D. E. Johnston and L. E. Small J. Amer. Chem. SOC.,1975,97,5598. 23 G. E. Lienhard and T. C. Wang J. Amer. Chem. Soc. 1969,91 1146; J. Toullec and J. E. Dubois J. Amer. Chem. SOC.,1974,96 3524. Part (ii) Polar Reactions kdA-1I + IH-C-C I / \ = 7 IH-C I -C /OH \ Scheme 9 advantage in transferring a proton from water to the enolate anion (pK =11)24325-note however that k,[B] is kinetically equivalent to kb[BH'][oH]. Recent studies have confirmed that the third-order term (k,) is not an artifact from a solvent or specific salt effect and a mechanism was proposed (similar to that for kB) in which the base-catalysed removal of the proton and the enolate anion was stabilized by hydrogen bonding to the This explanation implies that the Bronsted 0 values should be similar for kBand kAB,and from the known value of 0.88 for k a value of 0.68 was predicted for the p value for kAB.However the observed value of p for kAB was 0.15 which suggests that an increase in the effectiveness of the catalysing base is largely cancelled by a decreasing effectiveness of the catalysing acid.Consequently the dependence of k on the strength of the catalysing acid must be large and a Bronsted a! value of 0.5 was estimated. These and other results suggest that the third-order term represents true bifunctional catalysis with partial proton abstrac- tion by acetate ion and a significant movement of the proton of acetic acid towards the carbonyl oxygen atom in the transition Some authors have dismissed the possibility of termolecular processes but it may be that the involvement ofprotons as the third component of a termolecular reaction (as suggested for k,)* is not an unlikely process because an appropriate hydrogen may be in the solvent shell in close proximity to the reaction site (see also later discussion of ref.26). If the proton to be transferred is already hydrogen bonded to the appropriate atom (e.g. to the carbonyl oxygen for k,) it could be argued that this constitutes pre-equilibrium formation of a 'complex' and that the reaction is really bimolecular -e.g. for k the bimolecular reaction would be between acetate ion and the 'complex' of the ketone and acetic acid.However extension of this argument involves the reasonable though unusual proposition that a beaker full of water contains one 'complex' molecule made up of hydrogen-bonded H,O units. A much clearer case of electrophilic catalysis by complex formation of the general base-catalysed enolization of ketones has been observed in the iodination of acetylpyridines. Metal ions catalysed the reaction of 2-acetylpyridine (33) but not that of 4-acetylpyridine (32). It was calculated that the Zn2' complex reacted over * In contrast however the other terms (kAand kB) appear to represent stepwise processes (Scheme 9). E. S. Hand and W. P. Jencks J. Amer. Chem. SOC., 24 1975,97,6221.25 A. F. Hegarty and W. P. Jencks J. Amer. Chem. SOC.,1975 97 7188; see also R. Breslow 'Organic Reaction Mechanisms' Benjamin New York 2nd. Edn. 1969 pp. 52-56. 26 s. Rosenberg S. M. Silver J. M. Sayer and W. P. Jencks J. Amer. Chem. SOC.,1974,96,7986. T. W.Bentley Q lo3times faster and the Cu2+ complex over lo5times faster than the uncomplexed substrate presumably via a transition state like (34) whereas 4-acetylpyridine (32) could not form a similar complex.27 7 Tetrahedral Intermediates Carbinolrrmines.-The formation of imines hydrazones and related derivatives from carbonyl compounds is often drawn as a two-step process [equation (1l)] \ I / RNH,+ CEO * RNH-C-OH -* RN=C +H2O (11) / I \ As the dehydration step is acid catalysed at pH values below neutrality most reactions of this type undergo a change in rate-determining step from dehydration to formation of the carbinolamine.For the addition of methoxylamine (NH,OMe) to p-substituted benzaldehydes this change in rate-limiting step occurs at pH “4 inferred from the ‘break’ in the pH-rate profile (i.e.graph of kobsus. pH). At pH = 1 these reactions also show a second break in the pH-rate profile,26 whereas reactions between methoxylamine and acetone and between the aromatic aldehydes and semicarbazide do not.28 The mechanism shown in Scheme 10is consistent with this and other experimental evidence.26 + A-+I HA 2 RNH,-C-OH / RN-\CG k-1 I T+ qk- IRNH-C-OHI /% RN=C \ g To RN>\C& 5 R&H2-C-O-I Note Values for the 11 rate / k-2 I constants are given in ref.27 T* Scheme 10 27 B.G. Cox J. Amer. Chem. Soc. 1974,%,6823;see also P.Woolley J.C.S. Chem. Comm 1975,579. 28 J. M.Sayer B. Pinsky A. Schonbrunn and W. Washtien J. Amer Chem. SOC.,1974,96,7998;see also G. P. Tuszynskiand R. G. Kallen J. Amer. Chem Soc. 1975,97,2860. Part (ii) Polar Reactions According to this mechanism the reaction occurs as follows. Below pH =1 the dominant mechanism is the acid-catalysed addition of the amine to the aldehyde (k,) to give T'. Concurrent with this process around pH 1 and dominant towards pH 4 is the hydronium-ion catalysed 'trapping' (k3) of the highly unstable zwitterionic intermediate T. However this process has two possible rate-determining steps -in the lower pH region the uncatalysed addition step (k2)appears to be rate-limiting whereas towards pH 4 the concentration of the hydronium ion catalyst decreases and the amount of unprotonated methoxyamine available to react via k2 increases so that diff usion-controlled process (k3) becomes rate-determining.* Around pH 4 acid-catalysed trapping of T(k3)becomes so slow that it ceases to be a significant process and reaction occurs by rate-determining trapping of Tby proton transfer through the surrounding solvent (k,=lo6-lo7 s-') [see also Ann.Reports (B) 1974 71 1311. At higher pH the dehydration step (k5)is rate-determining. Consequently the second break in the pH-rate profile occurs when the acid- catalysed direct route to T' (k,)takes over from the uncatalysed route via T.For the reaction represented by k the following argument (as well as others) suggests that the hydronium ion acts as a general acid and that protonation is 'concerted' with nucleophilic attack by amine (see also discussion of reference 25).26 The rate constant k',for nucleophilic attack on the protonated aldehyde is given by k;=k,K (at pH 0). Assuming that pK for protonated p-nitrobenzaldehyde was -8.1 and knowing kl,this gives avalue of 2.4 x lo-" mol-' 1 s-' for k',,which would rule out this stepwise mechanism because k' is about lo2greater than the expected upper limit for diffusion-controlled processes. The authors26 note that the hydro- nium-ion catalysed reaction is the first example of a reaction involving rapid proton transfer to or from an electronegative atom that has been observed to proceed by concurrent kinetically distinct 'concerted' (k,)and stepwise (k2,k3) pathways for proton transfer with a single catalyst.Thus concerted catalysis may occur even when the intermediate (T) in the stepwise pathway is stable enough to have a finite lifetime (cf.reference 29). Studies of other amines supported the prediction that the concerted pathway would be favoured as the basicity of the amine decreased.28 From Esters.-An unusually large negative activation energy (ca.-10 kcal mol-') has been determined for the third-order reaction between n-butylamine and p-nitrophenyl trifluoroacetate in either chlorobenzene or 1,2-dichloroethane -also AS+is -70 cal K-' mol-'! When the reaction is catalysed by general bases it is first order in ester amine and catalyst but the rate constants do not follow the Bronsted catalysis law.The role of the catalysing base (or the second molecule of butylamine) appears to be to assist in the removal of a proton from the amine (Scheme 1l).30 The first reaction [equation (12)] is thought to be rapid and it may proceed by two sequential bimolecular processes or by one termolecular step. Unimolecular col- lapse of the intermediate [equation (13)] was thought to involve essentially simul- taneous breaking of the N-H and C-0 bonds so that formation of completely free ions was avoided. As equation (13) represents conversion of a zwitterion (35)into * Note that k refers to rate constant whereas reaction rate devends on k x appropriate concentrations.z9 W. P. Jencks Chem. Rev. 1972,72,705. 30 T. D.Sin@ and R. W. Taft J. Amer. Chem. SOC.,1975,97,3867; see also C.-W. Su and J. W. Watson J. Amer. Chem. SOC.,1974,96,' 1854. T. W.Bentley 0- CF,C//O + RNH + B CF,(!-OAr I \ OAr R-N+-H....B I H (35) Scheme 11 an ion pair (36)and a good leaving group is departing from an electron-rich centre the activation energy may be small but positive -the negative activation energy arises because reaction (12)is strongly exothermic [see also equation (5)J. However if the 1,2-dichloroethane solvent contains 0.487 mol I-’ methanol AH* for aminolysis is close to zero (a sharp increase from -10kcal mol-’!) whereas in the same solvent AH*for methanolysis is +4.4 kcal mol-’.Methanol may hydrogen- bond to the zwitterion intermediate (35) and it was suggesed that the marked change in activation parameters between ‘hydrophobic’ and ‘hydrophilic’ environments may provide a model for temperature regulatory action in certain enzymatic processes.3o 8 Micellar Catalysis Micelles are the high molecular weight aggregates of surface-active agents (e.g. detergents) containing a long hydrocarbon chain (hydrophobic) and a polar or ionic group (hydrophilic). Specific interactions between a substrate and either the hyd- rophobic or the hydrophilic part of the micelle in dilute solution can give rise to large rate effects either enhancement or inhibition and there is considerable interest in the kinetics and mechanism of these processes as catalysts and models for various biochemical processes.Consequently there have been many investigations of the reactions of carboxylic esters,31 but the following discussion will summarize some other recently published studies. The acid-catalysed rearrangement of hydrazobenzene (37)to benzidine (38) is accelerated by a factor of 1.5X lo3 by anionic micelles of sodium lauryl (37) (38) + o,p-isomer (39) sulphate larger than expected by analogy with rate accelerations observed in other specific hydrogen-ion catalysed reactions. The rearrangement is slightly inhibited by neutral micelles (possibly a solvent effect) whereas the reaction is sharply inhibited by the cationic micelles of cetyltrimethylammonium bromide presumably because 31 J.H. Fendler and E. J. Fendler ‘Catalysis in Mioellar and Macromolecular Systems’ Academic Press New York,1975; E. J. Fendler and J. H. Fendler Adu. Phys. Org. Chem 1970,8,271; see also Ann. Reports (A) 1973,70 167. Part (ii) Polar Reactions the hydrazobenzene is taken into the micellar phase from which hydrogen ions are excluded. The rate laws for both the normal and anionic micelle-catalysed reaction appear to be the same (first order in hydrazobenzene and second order in H+) although there is inevitably some uncertainty in the distribution of hydrogen ions between the micelles and the bulk solvent. Also addition of more of the detergent than is necessary to take up all of the hydrazobenzene into the micelles decreases the rate constant because it reduces the possibility of havingsubstrate plus two hydrogen ions in the same micelle.The anionic micelles did not change the product composition [80% of (38) 20% of (39)] and it was suggested that the micelles catalysed the reaction by bringing the three particles together in the micellar phase prior to formation of the transition state. Thus the effective concentrations of reactants increase and the unfavourable loss of entropy in forming the transition state is reduced. This argument suggests that such effects should be more important for third-order processes than for processes of lower order and is supported by the observation that the second-order rearrange- ment of 1,2-di-o-tolylhydrazine is only accelerated sixty-fold (cf.1500above) at the optimum micelle concentration.Also from more detailed studies of rearrangement of (37)at various acidities it was suggested that the second protonation step (either on carbon or nitrogen) may be rate-determining whereas previously it has been generally assumed that both proton transfers are pre-eq~ilibria.~~ Formation of micelles may also influence stereochemistry -e.g. hydrolysis of 2-octyl trifluoromethanesulphonate by alkyl-oxygen fission proceeds with predo- minant retention of configuration and is retarded about 300-fold by both cetyl- trimethylammonium bromide and sodium lauryl sulphate whereas predominant inversion occurs during normal hydroly~is.~~ As the stereochemistry of secondary solvolyses is known to be influenced by double inversion processes [e.g.Ann. Reports 1965,62 2391 it may be fruitful to check whether 2-octyl bromide or the corresponding sulphate are reaction intermediates. Deamination of amines (40) in aqueous perchloric acid has been shown to depend on micelle formation by the amines themselves. For (40a) the critical micelle concentration (the concentration at which the micelles first become detectable) is NH OH aq. HCIO Me pH 3.5 Me (40) (41) (42) (a) R = Et (b) R = CH2CH,CHMe2 0.35 moll-' whereas it is 0.1 mol I-' for (40b) which has a larger alkyl group. At the critical micelle concentrations direct substitution [(40a) to (41a)l occurs with excess inversion of stereochemistry whereas the 1,2-hydride shift [(40a) to (42a) and of (40b) to (42b)l occurs with excess retention of configuration.However at high micelle concentrations the opposite results are obtained -substitution occurs with excess retention of stereochemistry whereas the 1,2-hydride shifts occur with excess 32 C. A. Bunton and R. J. Rubin TetrahedronLetters 1975 55 59. 33 K. Okamoto T. Kinoshita and H. Yoneda J.C.S. Chem. Comm. 1975,922; see also C. N. Sukenik B.-A. Weissmann and R. G. Bergman J. Amer. Chem. Soc. 1975,M 445. 86 T. W. Bentley inversion. This correlation between direct substitution and hydride shift may be due to asymmetric ~olvation.~~ 9 Phase-transfer Catalysis Tetra-alkylammonium and tetra-alkylphosphonium salts catalyse reactions which are inhibited because of the inability of the reagents physically to come together -e.g.reaction between aqueous sodium cyanide and alkyl chlorides (Scheme 12). The RCI + QCN + RCN + QCl (organic phase) t 1 NaCl + QCN $ NaCN + QCl (aqueous phase) Scheme 12 quaternary salt (QX) brings the cyanide into the organic phase (e.g.benzene) as the soluble ion pair QCN and returns the displaced chloride ion to the aqueous phase where QCN can be regenerated. It appears that the rate-determining step takes place in ihe organic phase rather than in the aqueous phase at the interface or in micelles. Also the anions are less strongly solvated in the organic phase (e.g. no hydrogen bonding) and so displacement reactions occur more readily. This techni- que ‘phase-transfer is useful for many other types of reaction including generation of carbenes [Ann.Reports (B) 1974 71,1631 but only displacement reactions are reported here. The second-order reaction between thiophenoxide ion (PhS-) and 1-bromo- octane was independent of stirring speeds (200-2200 r.p.m.) and the rate of reaction was linearly dependent on catalyst concentration in contrast to micellar catalysis (Section 8). At fixed catalyst concentration the reaction rate was found to increase with the polarity of the solvent forming the organic phase (0-dichlorobenzene >benzene >heptane) and with the size of the alkyl groups in the catalyst. A correlation between rate constants and partition coefficients was noted which suggests that the major function of the catalyst is simply the solubilization of the nucleophile in the organic phase.This interpretation is supported by the increase in reaction rate with increasing ionic strength of the aqueous phase because the large organic ions are ‘salted out’ to the organic phase.36 Also from studies of Williamson ether syntheses requiring phase transfer of alkoxide anions (‘hard’) the use of bisulphates as the catalyst’s counter-ion was recommended because under the reaction conditions the unextractable ion SO:-was produced. In contrast if halide anions (‘soft’) are present either from the original catalyst (QX) or from the leaving group they are selectively transported and the reaction rates are reduced.37 Clear evidence that the phase-transfer reaction occurs in the organic phase has been obtained for displacement of n-octyl methanesulphonate with halide ions (X= C1 Br I) in water-chlorobenzene catalysed by C,H,,P’Bu,X- [equation (14)].It was shown that the activation parameters and the relative rates of the phase- n-C8H170S02CH3+KX -* n-CsH17X +CH,SO,K (14) 34 W. Kirmse G. Rauleder and H.-J. Ratajczak J. Amer. Chem. SOC.,1975 97,4141. 35 C. M. Starks and R. M. Owens J. Amer. Chem. SOC.,1973,95 3613. 36 A. W. Herriott and D. Picker J. Amer. Chem. SOC.,1975,97,2345. 37 H. H. Freedman and R. A. Dubois Tetrahedron Letters 1975 3251. Part (ii) Polar Reactions transfer reactions could be simulated under homogeneous conditions provided that small predetermined amounts of water were added to the chlorobenzene-it is known that the catalyst also transfers water to the organic phase and it was shown that this depended on the halide ion; in addition 0.15 mole of water per mole of substrate was calculated to be present in the organic phase.38 Quaternary ammonium groups anchored to a polystyrene resin have also been shown to catalyse displacement reactions and this technique may be convenient in syntheses because the catalyst can be removed by filtrati~n.~’ Insoluble polymers have also been used in a ‘three-phase test’ for reactive intermediates which involves the generation of a reaction intermediate from an insoluble polymeric precursor and its detection by trapping on a second solid phase suspended in the same solvent.Since only a fraction of the active sites of the polymer are at the surface it is assumed that any observed reaction between the two solid phases requires the existence of free intermediates.This ‘non-classical’ method has been applied to the detection of acylmidazoles from the imidazole-catalysed hydrolysis of o-nitrophenyl esters4’ and to the detection of free cy~lobutadiene.~~ 38 D. Landini A. M. Maia F. Montanari and F. M. Pirisi J.C.S.Chem. Comm 1975,950. 39 S.L. Regen J. Amer. Chem Soc. 1975,97 5956. 40 J. Rebek D. Brown and S. Zimmerman J.’Amer.Chern. SOC.,1975,97,454. 41 J. Rebek and F. Gavina J. Amer. Chem. SOC.,1975,W. 3453.