4 Reaction Mechanisms Part (ii) Polar Reactions By T. W. BENTLEY Department of Chemistry University Colfege of Swansea Swansea SA2 8PP 1 Introduction Electronic theories have dominated the interpretations of organic reactions during the past forty years and within the past ten years the concept of orbital symmetry has received much attention. Whilst orbitals and curly arrows the pictorial accompaniments of these qualitative theories are both useful and appealing other important factors including the effects of solvent and ion pairing are often neglected.* As shown by recent work on addition and elimination (see later) the stereochemistry of products may be controlled by formation and solvation of ion pairs rather than by stereoelectronic shifts of electron pairs or orbitals.There are other indications that these electronic theories are not panaceas. Both curly arrows and orbital-symmetry arguments suggest that the Diels-Alder reaction between butadiene and ethene would proceed via the symmetrical transition state shown in Scheme 1. However it is a fundamental postulate of r 1 i+ Scheme 1 theoretical chemistry that electrons ‘move’ faster than nuclei (e.g. Born-Oppen-heirner approximation Franck-Condon principle). Thus the representation in Scheme 1 requires that the nuclei move in a remarkably concerted manner especially remarkable considering that the collision complex of butadiene and ethene must be highly vibrationally excited. Recently using a quantitative MO treatment (MIND0/3) without symmetry constraints it has been calculated that one of the new bonds (e.g.C-1 to C-6)is almost completely formed while the other has hardly begun to form at all.’ Similar arguments as well as additional * It is particularly unfortunate that effects of solvent and ion pairing are often disregarded whereas it appears to be ‘well known’ that the iobes of a p-orbital are blue and green! ‘ M.J. S. Dewar A. C. Griffin and S. Kirschner J. Amer. Chem. Soc. 1974 96 6225; see also Chapter 4 part (i). 111 T. W. BmtieJ experimental evidence have led Bordwell to question the prevalence of concerted reactions involving formation and cleavage of four or more bonds (e.g.E2 S,2’),2 It is often difficult to rule out the formation of reactive ion-pair intermediates.Consequently there is little room to sit back complacently thinking that a particular mechanism ‘is’ as shown by Sneen and co-workers’ serious challenge of the classical S,2 mechanism -however some major deficiencies of Sneen’s interpretation are presented in the following section. Studies of reaction mechanisms are important aids to calculations of reaction rates and help to ensure that reactions the rates of which are to be calculated or correlated are proceeding by the same mechanism. This chapter includes a section on empirical force-field calculations of bridgehead reactivity which now spans a range of relative rates of over twenty powers of ten. Also a timely review has been published encouraging the inductive approach to chemistry and pointing out some of the approximations in transition-state the~ry.~ Isotopes are being used increasingly for studies of reaction mechanisms.Many of these investigations provide indirect mechanistic evidence because the structure of the product (i.e.the arrangement of atoms) is characterized more completely. Further mechanistic information is provided by kinetic isotope effects which are subtle probes of mechanism and cause minimum disturbance to the original system unlike for example substitution of a methyl group for a hydrogen. Primary deuterium isotope effects are particularly relevant to the sections on elimination and oxidation which include a discussion of a remarkable deuterium isotope effect of 36.5 for CrV* oxidation of glycolic acid. The inter- pretation of secondary deuterium isotope effects provides the basis for Shiner and co-workers’ mechanisms of aliphatic nucleophilic substitution but these are not generally accepted.A more comprehensive annual coverage of the literature of organic reaction mechanisms is presented el~ewhere.~ 2 Aliphatic Nucleophilic Substitution An advanced undergraduate text updating Bunton’s monograph (1 963) has been published,’ and there is also an account of solvolytic reactions of simple alkyl systems including detailed discussion of ion pairs.6 Since substitution and elimination usually occur concurrently research discussed in this section is relevant to both types of reaction.* Study of concurrent F. G. Bordwell Accounts Chem. Res. 1972,5 374; 1970 3 281.D. L. Bunker Accounts Chem. Res. 1974 7 195; see also L. P. Hammett ‘Physical Organic Chemistry’ McGraw-Hill New York 1970 pp. 1-3. ‘Organic Reaction Mechanisms 1974’ ed. A. R. Butler and M.J. Perkins Wiley London 1975. .S. R. Hartshorn ‘Aliphatic Nucleophilic Substitution’ Cambridge University Press London 1973. J. M. Harris Pragr. Phys. Org. Chem. 1974 11 pp. 89-173;(a) p. 140; (b) p. 155; see also D. J. Raber J. M. Harris and P. von R. Schleyer in ‘Ions and Ion Pairs in Organic Reactions’ ed. M. Szwarc Vol. 2 Wiley New York 1974 pp. 247-374. * If one accepts the principle of microscopic reversibility mechanistic conclusions from elimination reactions are also relevant to addition reactions. However some critical comments on applications of this principle to solution chemistry are presented later.Reaction Mechanisms-Part (ii) Polar Reactions elimination and substitution during solvolyses of t-butyl halides showed that the products depended on the leaving group (see Ann. Reports 1963,60 262). For solvolyses of 2-adamantyl arenesulphonates in 70 % v/v ethanol-water (Scheme 2) only substitution products were formed but the proportions of alcohol and 70"LEtOH-H, D O H + GOEt aoso20x ROH ROEt Scheme 2 ethyl ether depended markedly on the substituent X in the arylsulphonate; see Table l.7 Similar results have been obtained for 1-adamantyl solvolyses.8 Table 1 Dependence ojsubstitution product on leaving group (Scheme 2) X OMe Me H Br NO [ROH]/[ROEtJ 1.94 2.56 3.25 3.05 5.20 Therefore it appears that all the above reactions proceed through ion-pair intermediates rather than through free (i.e.symmetrically solvated) cations. The Mechanism of Sneen and Co-workers.-Many different types of ion pairs may be present in solution but recent attention has been focused on the simple ion-pair mechanism shown in Scheme 3. In the Straightforward? S,1 reaction ki h [Nuc] RX * R'X-LR-Nuc+X-k-i contact ion-pair when ~~[NUC] >> k-1 kobs = k straightforward S,1 but when k2[Nuc] -K k- kobs= k,k,"ucl . ion-pair SN2 or SN ip ~ k-1 and when k2[Nuc] x k-borderline mechanism. Scheme 3 ' J. M. Harris A. Becker J. F. Fagan and F. A. Walden J. Amer. Chem. SOC.,1974 96 4484. D. J. Watson (University College Swansea) unpublished results; P.Luton and M. C. Whiting to be published; P. Luton Ph.D. Thesis Bristol 1972. t Use of 'straightforward' is particularly appropriate since there is much current interest in internal return from contact ion-pairs (k- > k, Scheme 3) in some S,1 reactions -see also discussion of ref. 15. 114 7‘.W. Bentley k is rate-limiting and products are produced by rapid destruction of the contact ion-pair -for simplicity elimination is excluded from Scheme 3 but the mech- anism can readily be extended to include it [see Ann. Reports (B) 1973’70 1201. Although in the classical interpretation of Ingold and co-workers it was assumed that a free cation was formed in sN1 reactions there is nothing in Ingold’s defini- tion of the term SN1 to rule out rate-limiting formation of the contact ion-pair shown in Scheme 3.As there is a long history to the debate about whether a free cation shielded on one side by the leaving group is chemically or just semantically different froma contact ion-pair suffice it to state here that some form of sequential ionization via ion pairs is now widely accepted ;6 see also Scheme 6. However Sneen’s reinterpretation of the classical sN2 mechanism has not been received favourably. His ion-pair SN2 mechanism (Scheme 3) requires rapid formation of an ion-pair followed by its rate-limiting destruction. Possibly this may occur for systems capable of stabilizing the developing positive charge (e.g. allyl) but the mechanism was also intended to account for the behaviour of simple primary and secondary substrates.Perhaps the best single piece of evidence against this spectacular generalization is the calculation of Abraham who showed that heterolysis of an alkyl chloride without assistance from rearside nucleophilic attack would require about 40kcal mol- more than the experi- mentally observed activation energy [Ann. Reports (A) 1973’70 1501. As might be expected the energetics for heterolysis of secondary substrates were calculated to be borderline but there are other good reasons to doubt the interpretation of borderline behaviour shown in Scheme 3. The experimental basis of Sneen’s generalization is a study of the effect of azide ion on the solvolysis of 2-octyl methanesulphonate in aqueous dioxan.Unfortunately it is difficult to separate the effect of azide on the ionic strength of the medium from the effect of azide as nucleophile. One approach to ‘buffer’ the solution with a large excess of the non-nucleophilic salt NaClO, led to the conclusion that the correlation of rates and products (with varying concentrations of azide) was consistent with the classical sN2 mechani~rn.~ It has also been pointed out that azide concentrations may be poor approximations for activities and that salt effects may be specific in nature so that corrections for ‘normal’ salt effects on the medium are highly questionable.” OBs kHlkD klZclk14c c6H13teH3 ROH 1.097 & 0.007 1.063 & 0.007 N -+D OBs aq. acetone I IH RN,C~H~*,CCHJ 1.106 f0.007 1.105 f0.007 ROBS Scheme 4 D.J. McLennan J.C.S. Perkin II 1974 481. lo G. A. Gregoriou Tetrahedron Letters 1974 233; P. J. Dais and G. A. Gregoriou,ibid. p. 3827. Reaction Mechanisms-Part (ill Polar Reactions There are also completely independent criticisms based on secondary kinetic isotope effects. The results given in Scheme 4 show the I4C-and a-deuterium- isotope effects on the formation of both azide and carbinol from 2-octyl bromo- benzene-p-sulphonate in aqueous acetone-NaN, using carbon- 14 as a tracer for deuterium.'' It was argued that the relatively large carbon isotope effects indi- cated that the transition state must be nearly symmetrical which is consistent with the classical S,2 mechanism not the ion-pair SN2 mechanism. Also as the a-deuterium isotope effects of 1.10 for formation of both azide and carbinol are the same within experimental error similar mechanisms must be operating but detailed interpretation is a matter ofdisagreement and speculation.The authors'" surprising suggestion that 1.10 represented competing classical S,2 (kH/kD = 1.0) and SNl (k,/k x 1.2) mechanisms is inconceivable because as weH as ignoring all the evidence for sequential ionization it requires that reactions with both azide and water as nucleophiles proceed by the same proportions of S,1 and SN2 mechanisms (ca. 50 50). An alternative explanation of the value of 1.10 is that varying amounts of nucleophilic assistance to ionization are possible in a spectrum of sN2 mechanisms,'* [with SN1as the limit where nucleophilic assis- tance is zero (cf.Ann. Reports 1951,4'8 121)]. Since a-deuterium isotope effects appear to reflect hybridization changes e.g. addition of MnO -to cinnamic acids involves a hybridization change from sp2 to sp3 and gives inverse a-PhCH=CHCO,H & PhHF-qHCO,H 5 PhCHO + HCOCO,H(?) + + Mn0,-Mn02-for k fork [a-2H]cinnamic acid k JkD = 0.77 kdkD = 1.09 [B-'H]cinnamic acid kJkD = 0.75 k JkD = 1.09 Scbeme 5 deuterium isotope effects (Scheme 5)," it might be expected that the ion-pair SN2mechanism should also give inverse isotope effects. Thus depending on the kinetic-isotope effects on k and k- (Scheme 3) the value of 1.10 may not be consistent with an ion-pair SN2 mechanism. The Mechanism of shiner and Co-workers.-Shiner and co-workers have pro- posed a sequential mechanism for ionization in which under different conditions as many as four steps may be rate-limiting.I4 A simplified form of their mechan- ism is shown in Scheme 6 which has been drawn to show the similarity with Scheme 3.Additional steps are the classical SN2 process and k the dissociation V. F. Raaen T. Juhlke F. J. Brown and C. J. Collins J Amer. Chem. Soc. 1974 % 5928. '* J. M. Harris R. E. Hall and P. von R. Schleyer J. Amer. Chem. Soc. 1971 93 2551. l3 D. G. Lee and J. R. Brownridge J. Amer. Chem. Soc. 1974 96 5517; see also E. A. Halevi Progr. Phys. Org. Chem. 1963 1 168. l4 V. J. Shiner jun. in 'Isotope Effects in Chemical Reactions' ed. C. J. Collins and N. S. Bowman Van Nostrand Reinhold New York 1970.116 T. W.Bentley R-Nuc R + /IX-solvent-separated ion-pair Scheme 6 of a contact ion-pair to a solvent-separated ion-pair. It is claimed that the four possible rate-limiting steps can be identified by their characteristic a-deuterium isotope effects. As these are somewhat dependent on the leaving group typical values for solvolyses of secondary sulphonates are given below (i) k,/k z 1.0-classical S,2 mechanism ; (ii) k,/k = 1.15-1.16 -k is rate-limiting e.g. solvolysis of 3,3-dimethyl-2- butyl sulphonates (1) in protic solvents (Scheme 7); (iii) k,/k = 1.22-1.25-k is rate-limiting e.g. trifluoroacetolysis of 2-propyl toluene-p-sulphonate solvolyses of 2-adamantyl sulphonates ;l2 (iv) k,/k zz 1.05-1.15 (depending on the solvent) -k is rate-limiting; this interpretation corresponds to Sneen’s.The above interpretation is supplemented by studies of additions to double bonds thought to give rise to the contact ion-pair (R’X-). Addition of bromo- benzene-p-sulphonic acid to propene in trifluoroacetic acid leads to 2-propyl- bromobenzene-p-sulphonate whereas a similar reaction of t-butylethylene leads to the rearranged trifluoroacetate (2). Consequently it appeared that for propene collapse of the contact ion-pair was more rapid than attack by nucleophile (k,) or further dissociation (k3,Scheme 6) but that for t-butyl- ethylene collapse of the contact ion-pair was prevented by rapid rearrangement CH H I ICH3C-CCH CH H k I IS CH,C-$CH CH3 H HOBs 1 1 -CH,C-C=CH &H3 OBs k-’ AH3 OBs bH3 (1) Ik3 I I OBs CH CF3COO CH3 (2) Scheme 7 (k3,Scheme 7).” Thus it was argued that the relative rates of secondary solvo- lyses were influenced by ‘hidden return’ i.e.k2 < k- (Scheme 3) or k and k3 < k-(Scheme 6). As an independent test of these far-reaching proposals the analogous 1-adamantylmethyl carbinyl system (3) was st~died.’~ Because of the extra ring (a) V. J. Shiner jun. and W. Dowd J. Amer. Chem. Soc. 1969 91,6528; (b) V. J. Shiner jun. R. D. Fisher and W. Dowd ibid. p. 7748. I6 T. W. Bentley S. H. Liggero M. A. Imhoff and P. von R. Schleyer J. Arner. Chem. Sot. 1974 96 1970. Reaction Mechanisms-Part (ii) Polar Reactions Figure 1 Logarithms of rate constants at 25 "C for trifluoroacetolysis of secondary tosylutes (X = OTs) plotted against Co*(data from refs.16 and 17) strain introduced by rearrangement acetolysis of (3) leads mainly to the un- rearranged acetate (4) and the alkene (6),along with small amounts of the re- arranged acetate (5). Accordingto Shiner's proposals this lack of rearrangement Scheme 8 118 T. W.Bentley indicates that hidden return in (3) should increase and therefore the rate of solvolysis of (3) should be slower than that of (1). The experimental results showed that after allowing for the well-established OBs/OTs rate factor of three (3) solvolysed slightly faster (not slower) than (l) in agreement with classical carbo-cation theory. Also the linear correlation (Figure 1) of the logarithms of trifluoroacetolysis rate constants of simple secondary toluene-p-sulphonates (tosylates) including 2-propyl and (l) with CF* suggests that all the compounds have the same rate-limiting step presumably formation of contact ion-pair (k,).16 The point for 3-methyl-2-butyl” is slightly off the correlation line (Figure l) possibly owing to anchimeric assistance (rate x ca.3) by the neighbouring hydrogen. These criticisms cast considerable doubt on the basis of Shiner’s proposals i.e. both the interpretation of a-deuterium isotope effects and certain applications of the principle of microscopic reversibility are brought into question. It appears that ion pairs produced from additions to double bonds behave differently from those produced by heterolysis of a covalent bond in that the former appear to be more likely to collapse to covalent substrate.This may be due to different solvation of short-lived intermediates and suggests that the principle of micro- scopic reversibility should be used cautiously to interpret reaction mechanisms in solution. ?Ihe Mechanism of Schleyer and Co-workers.-Recent studies of 2-adamantyl tosylate (7) a hindered secondary substrate which solvolyses without detectable nucleophilic assistance (i.e. &l) have caused reconsideration of the role of solvent as nucleophile in the solvolyses of other secondary substrates [cJ Ann. Reports (B) 1972 69 1621. Until recently it could have been argued that the relative rates of secondary solvolyses depended so markedly on solvent that it was not possible to make a meaningful comparison of solvolysis rates.However because the relative rates of secondary solvolyses in hexafluoroisopropyl alcohol and trifluoroacetic acid are very similar (Table 2) it may be argued that these solvolyses are close to limiting (i.e. S,1) and therefore reflect true carbenium ion reactivity.Is This accounts for the linear correlation of the trifluoroacetolysis data in Figure 1,16 which is not observed in more nucleophilic solvents.6b The relative rate data (Table 2) suggests that solvolysis of secondary substrates (except 2-adamantyl) in more nucleophilic solvents is assisted by the solvent behaving as nucleophile. Both acetic and formic acids appear to be relatively A. Pross and R. Koren Tetrahedron Letters 1974 1949.Is F. L. Schadt P. von R. Schleyer and T. W. Bentley Tetrahedron Letters 1974 2335. Reaction Mechanisms-Part (ii) Polar Reactions Table 2 Relative rates of secondary solvolyses" Solt'ent EtOH CH3C0,H HC0,H Tosyiate 2-adamantyl{7) 1.P exo-2-norbornyl(8) 10 400 endo-Znorbornyl(9) 34 cyclopentyl(l0) 6 250 3-pentyl 1 560 I.@ 4 140 14 280 40 1 .ob 1940 1.14 27 5.3 cyclohexyl 108 8.3 1.5 CF,CO,H 1.P 52W 0.46' 3.w 0.85 0.30 97 wt. % (CF,),CHOH 1.P ----0.19d 2-propyl 910 13 0.9 6' 0.01 0.024' a Titrimetric or conductimetric rate constants at 25 "C. Defined as 1 .O for each solvent. 'Ref. 21. Ref. 18. nucleophilic compared with trifluoroacetic acid. Also less-hindered substrates appear to be more susceptible to nucleophilic attack and there appears to be a regular gradation of reactivity within the series.A quantitative estimate of the magnitude of nucleophilic solvent assistance (kJk ratio) can be obtained from Table 2 by dividing the relative rates in any solvent by the relative rates in tri- fluoroacetic acid e.g. in acetic acid 2-propyl tosylate appears to be nucleo- philically assisted by a factor of 13/0.024 = 540 whereas cyclopentyl is assisted by 280/3.0 = 93. Consequently it is clear that the common tendency to compare directly the acetolysis rates of secondary substrates is meaningless. Correction should be made for nucleophilic solvent assistance or comparisons should be made between data for solvolyses in hexafluoroisopropyl alcohol or in trifluoro- acetic acid.This leads to a previously unpublished analysis of the problem of the rates of norbornyl solvolyses discussed below. Solvolyses of Norbornyl Systems a Compromise Solution.t-Although inter-pretation of the relative rate ratio of about lo00 for solvolyses of exo-and endo- 2-norbornyl systems (Table 2) is a controversial topic it is generally agreed that the problem is a subtle one. Attempts to 'probe' the system by introducing extra substituents (e.g. by substitution at the l-position,lg or by comparison of the endo-2-norbornyl system with the corresponding tertiary compound20) into an environment which is already sterically crowded must be suspect. Ideally exo-(8) should be compared with endo-(9) but this does not help to determine l9 D.Lenoir Tetrahedron Letters 1974 1563. 2o J. M. Harris and S. P. McManus J. Amer. Chem. SOC.,1974 96 4693. 7 Helpful discussions with H. C. Brown B. Capon and P. von R. Schleyer are gratefully acknowledged. 120 T. W.Bentrey whether solvolysis of exo-(8) is unusually or solvolysis of endo-(9)is unusually slow.22 In the past there has been considerable disagreement about how to interpret the kinetic data and what constituted normal solvolytic be- haviour. If the conclusions discussed in the previous section are accepted a revised analysis of the kinetic data is necessary. Also ‘normal’ behaviour can be defined by the good linear correlation of trifluoroacetolysis rate constants with G* for acyclic tosylates (Figure l),which suggests that inductive/hyperconjugative effects dominate the effect of structure on reactivity.I6 Even in trifluoroacetolysis of cyclopentyl tosylate (lo) where relief of torsional strain may occur the rate constant is less than four times greater than that of the acyclic C5system 3-pentyl tosylate; however (10)reacts over 100 times faster than 2-propyl tosylate (Table 2).An estimate of the ‘normal’ trifluoroacetolysis rates of exo-(8)and endo-(9) can be made by correction of cyclopentyl tosylate (10)for the normal inductive/ hyperconjugative effects of the two extra carbon atoms and a rate correction factor of 3-8 is indicated by the following evidence based on a wide variety of model systems (i) Solvolysis of the tertiary norbornyl derivative (11) is about 5 timesfaster than that of the corresponding cyclopentyl derivative (12),e.g.when X = C1 k,,/k12 = 5.4,23and when X = OPNB k,,/k12 = 4.4.24 This is of the expected magnitude and direction for inductiuelhyperconjugativeeflects.2 I (10) CH -3 (1 1) (12) P Following Sargent’s argument,26‘ if it is assumed that cyclopentyl is accelerated over norbornyl by conformational effects such as relief of extra torsional strain the observed rate ratios would have to be explained by another rate-accelerating effect in the tertiary norbornyl system (I 1).Anchimeric assistance is ruled out since it has now generally been agreed that tertiary 2-norbornyt cations are essentially clas~icai.~~~~~’.~~ Also as similar results are obtained for chlorides and p-nitrobenzoates * .24 steric acceleration by relief of leaving-group strain is not appreciably different in (I 1) and (12).29Therefore it appears that only inductive/hyperconjugativeeffects lead to a consistent explanation.21 J. E. Nordlander R. R. Gruetzmacher W. J. Kelly and S. P. Jindal J.Amer. Chem. Soc. 1974 96 181. 22 (a) H. C. Brown M. Ravindranathan and E. N. Peters J. Amer. Chem. SOC. 1974 96 7351 ; (b)H. C. Brown Accounfs Chem. Res. 1973,6 377. 23 H. C. Brown and F. J. Chloupek J. Amer. Chem. Sor. 1963,85 2322. 24 H. C. Brown and M.-H. Rei J. Amer. Chem. SOC.,1964,86 5004. 2s A. Streitwieser jun. J. Amer. Chem. soc. 1956 78 4935. 26 (a)C. D. Sargent Quart. Rev. 1966 20 301 ; (b) G. D. Sargent in ‘Carbonium Ions’ Vol.111 ed. G. A. Olah and P. von R. Schleyer Wiley New York N.Y. 1972 pp. 1099-1 200. Zf H. L. Goering and J. V.Clevenger J. Amer. Chem. Soc. 1972 94 1010; H. L. Goering and K. Humski ibid. 1969 91 4594. 28 G. A. Oiah G. Liang G. D. Mateescu and J. L. Riemenschneider J. Amer. Chem. Soc. 1973 95 8698. 29 J. Slutsky R. C. Bingham P. von R. Schleyer W. C. Dickason and H. C. Brown J. Amer. Ghem. SOC.,1974 96 1969. Reaction Mechanisms-Part (ii)Polar Reactions 121 (ii) In trifluoroacetic acid trans-2-methylcyclopentyl tosylate solvolyses 2.6 times faster than cyclopentyl tosylate,2 and allowing for the extra y-carbon this would lead to a rate factor of 3.4. (iii) A maximum factor of 7-8 would be predicted from the r~* correlation for the more flexible acyclic secondary tosylates ;Figure 1.Comparison of the trifluoroacetolysis rates relative to cyclopentyl with the predicted relative rate of 3-8 leads to the residual rate factors (Table 3). It Table 3 Relative rates oj'secondary solzlolyses Tosylare CF3C0,H (Predicted)" Residuul rate factors cyclopentyl (10) 1 .ob (1 .oy __ endo-(9) 0.15 (3-8)c 20-60 too Slow em-(8) 172 (3-8)' 20-60 too fast See text. Defined as 1.0. 'Based on an estimate of normal inductive/hyperconjugative effects of two extra carbon atoms -see text. A higher rate factor would probably be obtained using polarimetric rate constants. appears that by comparison with models (lo)-( 12) exo-(8) solvolyses more rapidly than expected whereas endo-(9) solvolyses more slowly than expected.Thus at least two additional kinetic effects must be postulated. Published interpretations for the behaviour of exo-(8) include anchimeric assistance with formation of a bridged non-classical ion during heterolysis of the sulphonate bond,26a or stabilization of the incipient cation by C-C hyper-c~njugation,~' ' The behaviour of endo-(9)may or by a vertical ~tabilization.~ be explained by 'steric hindrance to ionization',22b possibly caused by decreased ability of the solvent to form hydrogen bonds to the three oxygen atoms of the sulphonate anion. Since the rate-limiting step is probably formation of a contact ion-pair 'steric hindrance to solvation of the leaving group' may be a preferable modification of the interpretation.The residual rate factors (Table 3) are similar and correspond to energies of only 2 kcal mol-'. Thus if the above analysis is correct the remaining problems of the interpretation are extremely subtle. Recently calculations purporting to show that the rate constant for 1,2-shift in a classical 2-norbornyl cation would be as high as 1OI2 s-' did not adequately account for the con- current 1,Zshift of the counterion and its hydrogen-bonded solvent molecules ;32 the same paper contains an interesting but biased analysis of the norbornyl problem including criticisms of the ESCA data28 of stable ions [cf.Ann. Reports (B) 1972 69 801 and of transition-state theory. '"F. R. Jensen and B. E. Smart J. Amer. Chem. SOC.,1969 91 5688. '*T. G. Traylor W. Hsnstein H.J. Berwin N. A. Clinton and R. S. Brown J. Amer. Chem. SOC.,1971 93 5715. 32 F. K. Fong. J. Amer. Chem. SOC.,1974 96 7638. 122 T. W. Bentley 3 Elimination Since there is not general agreement on the mechanistic interpretations of nucleophilic substitution formally involving cleavage of only one bond and formation of only one bond it is not surprising that the mechanisms of elimination are a matter of great debate. As with substitution reactions it is possible to recog- nize relatively clear mechanistic categories e.g. Elcb El E2H (classical E2) but consideration of ion pairs leads to a variety of possible mechanistic interpretations. General agreement on the mechanism of substitution would appear to be a prerequisite for agreement on mechanisms of elimination particularly for solvo- lytic reactions and for eliminations by those weak bases which are good nucleo- Scheme 9 philes.This can be illustrated by the results of a detailed study of the products from cyclopentyl bromobenzene-p-sulphonate (Scheme 9 Table 4).33 Whilst Table 4 Elimination from cyclopentyl bromobenzene-p-sulphonate" Reaction conditions Alkene k,lkDb yield (%) antilsyn" 1SM-NaOEt-EtOH ca. 20 5.77 1.15-1.18 trifluoroethanol-water aqueous ethanol 50-76 12-27 0.27 1.37 1.23-1.25 1.15-1.18 See Scheme 9. a-Deuterium isotope effect including competitive substitution. reaction with sodium ethoxide gave predominantly anti-elimination (E2H-like) in trifluoroethanol-water syn-elimination was favoured by about 4 1.A possible rationalization of the solvolytic results for which no base is available to remove the P-hydrogen involves the five-co-ordinate intermediate first suggested nearly twenty years ago (Scheme 10);6a*34 this proposal is similar to the E2C mechanism which usually leads to anti-elimination except that bonding between the nucleo- phile or leaving group and the B-hydrogen has not been specified in Scheme 10. The authors'33 proposal that the reaction in aqueous ethanol was an El elimina- tion does not appear to be consistent with the a-deuterium isotope effect (Table 4); 33 K. Humski V. SendijareviC and V. J. Shiner jun, J. Amer. Chem. SOC.,1974 96 6187. 34 S. Winstein D. Darwish and N. J. Holness J. Amer. Chem. Soc. 1956 78 2915. Reaction Mechanisms-Part (ii) Polar Reactions substitution (inversion) syn-elimination OBs 6-anti-elimination Scheme 10 their proposed mechanism for trifluoroethanol involved attack on a preformed ion-pair (see also refs.2 and 39 and closely parallels the mechanism of substitu- tion (Scheme 3) criticized above. Recent studies of the E2C mechanism [Ann. Reports (B) 1968,65 138 ; 1972 69 1801 have concentrated on the isotope-effect criterion. Primary hydrogen isotope effects on the rates of bimolecular elimination of substituted cyclohexyl tosylates and bromides are small (kH/kD -2-3) for very E2C-like reactions pass through a maximum of about 6 for E2H-like reactions and are small again (2-3) for very E2H-like reactions. Movement within the E2C-E2H spectrum was achieved by increasing the acidity of the substrate or the basicity of the base or by changing the leaving group from tosylate to bromide.36 Secondary isotope effects indicated significant changes in hybridization at both C and C, con-sistent with an E2C me~hanism.~’ As there is now considerable evidence for the E2C mechanism one wonders to what extent an inability to represent the mech- anism in an aesthetically pleasing way (e.g.using curly arrows) is proving to be an ‘activation’ barrier to its general acceptance. It is known that crown ethers can affect the amount of product formed by syn-elimination. When the ether complexes with the cation more of the base is dissociated. In some cases it appears that anti-elimination is favoured by disso- ciated base whereas syn-elimination is favoured when the base is associated which can be explained by simultaneous co-ordination of the metal counterion (M) with the base and the leaving group (13).A striking example of this effect is the influence of 18-Crown-6 (1 moll-’) on elimination from exo-2-norbornyl [exo-3-’H]tosylate (14) by the sodium salt of 2-cyclohexylcyclohexanol in triglyme (Table 5).38 35 W. T. Ford Accounts Chem. Res. 1973 6,410. 36 D. Cook R. E. J. Hutchinson J. K. McLeod and A. J. Parker J. Org. Chem. 1974 39,534. ” I>. Cook R. E. J. Hutchinson and A. J. Parker J. Org. Chem. 1974 39,3029. 38 R.A. Bartsch and R. H. Kayser J. Amer. Chem. SOC.,1974 96,4346; R. A.Bartsch E. A. Mintz and R. M. Parlman ibid.p. 4249. 124 T. W. Bentley Table 5 Eliminationfrom exo-2-norborny~[exo-3-2H]tos~~lute (14) 18-Crown-6 % of total hydrocarbon present (15) (16) Nortricyclene No 98.0 0" 2.0 Yes 70.0 27.2 2.8 a Lower limit of detection <2 % (n.m.r.). Similar research has been carried out on acyclic systems. In a detailed analysis of the stereochemistry of elimination reactions of 3-hexyltrimethylammonium iodide (17) promoted by phenoxide bases in mixtures of t-butyl alcohol and dimethyl sulphoxide it has been shown that trans-3-hexene is formed in both syn-and anti-pathways whereas at least 94% of cis-3-hexene is formed by the anti-pathway. The cation of the metal phenoxide influences the proportion of syn-elimination and in 20 % dimethyl sulphoxide the syn -+ trans pathway ranges from 34% for potassium to 74% for lithium phenoxide.Because the leaving group is positively charged attack by associated base leading to a transition state resembling (13) is unfavourable. Thus in contrast to the previous example (14) attack by dissociated base favours syn-elimination possibly oia the cyclic transition state (18).39 CH3CH2CHCH2CH2CH3 I I-+N(CH,) (17) (18) As it requires only 1.4 kcal mol-I to change a product distribution from 50 50 to 90 10 generalizations about the factors influencing products should be applied cautiously e.g.base association plays an important role in determining the trans-cis-alkene ratios in elimination reactions of 1-phenylprop-2-yl chlorides but in contrast to the behaviour of non-activated halides an associated base leads to preferential formation of trans-alkene~.~' From a study of a simple system 2-iodobutane for which there is considerable evidence for a classical E2 mechanism it appears that the positional orientation of the elimination products is mainly controlled by the base strength.Orientation was most sensitive to base strength when the ,&hydrogen was removed by an oxygen base the order oxygen > nitrogen > carbon favouring 'hard' bases. Only for very hindered bases such as 2,6-di-t-butylphenoxide did steric effects become important .41 39 J. K. Borchardt and W. H. Saunders jun. J. Amer. Chem. SOC.,1974 96 3912. 'O S. Alunni E. Baciocchi R. Ruzziconi and M. Tingoli J. Org.Chem. 1974 39 3299. 41 R. A. Bartsch K. E. Wiegers and D. M. Guritz J. Amer. Chem. SOC.,1974 96 430. Reaction Mechanisms-Part (ii) Polar Reactions 4 Electrophilic Addition In accordance with Ingold’s systematic notation the term Ad will be used for electrophilic addition reactions. Akenes.-Because the kinetic expressions for addition of hydrogen halides to alkenes are often complex mechanistic interpretations benefit from detailed studies of selected systems. Furthermore rigorous experimental techniques may be required to prevent isomerization of the starting alkenes. For addition of hydrogen bromide in acetic acid to cis- or trans-but-2-enes7it appears that the presence of hydrogen ion bromide ion and oxygen or light is necessary for alkene isomerization.Although addition of a bromine atom to the alkene to give a #I-bromoalkyl radical is implicated even a large excess of a typical radical inhibitor did not quench the reaction and reactions were carried out on degassed solutions in the dark under a helium atm0sphe1-e.~~ The stereochemistry of addition of both deuterium bromide and {O-2H]acetic acid to both cis- and trans-but-2-ene is 84 f2‘%; anti (16 & 2 ”/ syn) and is invari- ant over a 100-fold concentration range of deuterium bromide and unaffected by the presence of added lithium bromide or lithium perchlorate. Three mechanistic possibilities were considered. The first an Ad,2 (formally the reverse El) process involving formation of a carbenium ion-bromide ion contact ion-pair which collapses to syn-addition product or rearranges to produce the anti-addition product was ruled out because added bromide ion did not affect the stereo- chemistry.The second competing syn A42 and anti Ad,3 (formally the reverse E2H) processes was ruled out because when taken with the kinetic expression +d[RBr],dt = k,[alkene] [HBr] + k,[alkene] [HBrI2 (1) [equation (l)],it predicts that product stereochemistry should be dependent on the concentration of hydrogen bromide. The results appear to be consistent with competing syn and anti AdE3 mechanisms (19) and (20) in which alkyl bromide is derived only from the third-order term and alkyl acetate is derived only from the second-order term in equation (1). Kinetic isotope effects support the inter- pretation that the hydrogen in transition states (19) and (20) is supplied by co- valent hydrogen bromide rather than protonated acetic Nuc = HBr Br- or CH,C02H Allowing for pre-equilibria (e.g..n-complexes or ion pairs) there is not neces- sarily a termolecular elementary step. The products may be determined by ‘* D. J. Pasto G. R. Meyer and B. Lepeska J. Amer. Chem. SOC.,1974 96 1858; see also R. C. Fahey C. A. McPherson and R. A. Smith ibid.,p. 4534. 126 T. W.Bentley addition of either acetic acid not explicitly included in equation (l),or hydrogen bromide to a complex of alkene and another molecule of hydrogen bromide. Since this complex could be attacked by hydrogen bromide acetic acid or bromide ion the identical stereochemistry of addition of both acetate and bromide is explained as well as the dependence ofthe ratio of acetate to bromide products on the initial concentrations of hydrogen bromide.However the reaction of both cis-and trans-but-2-enes cannot proceed through a common intermediate (at least not all the reaction). Consequently although it may be fortuitous that both cis- and trans-alkenes undergo the same amount of unti- addition this aspect of the interpretation needs further attention. Using gas-phase 'H n.m.r. spectroscopy it has been found that the uncatalysed reaction between alkenes and hydrogen chloride at room temperature and ten atmospheres pressure has a half-life of about four days (k = 4.2 f1 x 1 mol-' s-at 25 "C). The reaction rate is not affected by an increase in surface area (e.g.added glass capillary tubing) but appears to be retarded by increased temperature. Similar results were obtained for 2-methylpropene and the reaction at room temperature appears to be ca. lo9 times faster than predicted from the equilibrium established at high temperatures. The low-energy path (not available at higher temperatures?) may involve reversible formation of a complex which is unstable at high temperatures. It was suggested that this pathway could not be the reverse of the decomposition pathway although some charge separation appears to be present because no 1-chloropropane was obtained from pr~pene.~~ No doubt the issues raised between the above interpretation and the principle of microscopic reversibility will require further arbitration.Whilst the broad outline of the mechanism of addition of bromine to alkenes is generally agreed detailed interpretations are still being debated.44 Relative rate data for addition of bromine to a series of alkenes in 1,1,2-trichlorotri- fluoroethane solution at -35 "C show a considerably smaller range than rates previously obtained for methanol at 25°C. This was interpreted as resulting from a change in the position of the transition state along the reaction co- ordinate so that it was more like the reversibly formed n-complex than the ethylene bromonium ion a-comple~.~~ Alkynes.-If protonation of an alkyne by hydrogen chloride yielded a 'free' vinyl cation which would be linear subsequent attack by nucleophile should lead to both anti- and syn-addition products.However recent data for addition of hydrogen chloride to a variety of alkynes are consistent with competing Ad,2 and A43 mechanisms with the predominant mode of addition being syn and anti respectively. Contact ion-pairs are clearly implicated in the AdE2 mechanism because unlike the ,443 (see above) addition of chloride ion does not increase the percentage of chloride formed. For example phenylethyne 43 F. Amar D. R. Dalton G. Eisman and M. J. Haugh Tetrahedron Letters 1974 3033 3037. 44 S. P. McManus and D. W. Ware Tetrahedron Letters 1974 4271 and references there cited. 45 G. A. Olah and T. R. Hockswender jun. J. Amer. Chem. SOC.,1974 96 3574. Reaction Mechanisms-Part (ii) Polar Reactions 127 undergoes addition of hydrogen chloride (0.8moll-') in acetic acid to give the expected chloride (90%) along with acetophenone (10 %).Added tetramethyl- ammonium chlaride (0.2 moll- ') reduces the yield of chloride slightly whilst the reaction rate increases by a factor of two probably due to a medium effect. In contrast [1-'HI hex-1-yne undergoes addition of hydrogen chloride (0.75 mol 1-I) in acetic acid to give anti-(21) and syn-(22) products in the ratio of 40 :60; addition of tetramethylammonium chloride (0.2 moll- ') changes the ratio to 90 10,and considerably decreases the yield of ketone (23).46 V ci H CI D \ D Ph H \ /D Ph \/ Ph-CEC-D-/c=c\ /c=c\ + Although the syn contact ion-pair is almost certainly formed first in the A42 mechanism both syn-and anti-addition products may result if 'leakage' occurs i.e.the syn ion-pair may isomerize to the anti ion-pair at a rate competitive with the rate of collapse of the ion-pair to covalent product. Consequently both the solvent and the presence of catalyst may influence the stereochemistry (Table 6),47 Table 6 Stereochemistry of addition of hydrogen chloride to phenyl[1-'H]ethyne (24)" Ratio Solvent Catalyst syn(25):anti(26)* Acetic Acid -60 :40' Dichloromethane -65:35 Sulpholane -75 :25 Sulpholane ZnC1 50 :50 Results from ref. 47. Estimated error 5% See also ref. 46. but the effects are not large in energy terms as both products are probably formed from the syn ion-pair by very rapid reactions. Whilst Table 6 does show one 50 :50 syn :anti ratio theauth01-s'~~ alternative interpretation that this implicated a symmetrically substituted free vinyl cation seems unlikely as other workers observed that added chloride ion reduced the amount of alkyl chloride formed in competition with ketone.46 46 R.C.Fahey M. T. Payne and D.-J. Lee J. Org. Chem.. 1974 39 1124. 47 F. Marcuzzi G. Melloni and G. Modena Tetrahedron Letters 1974 413. 128 T. W.Bentley Similarly addition of trifluoroacetic acid to both the allene (27) and the alkyne (29) (Scheme 11) gave the same ratio of 2 :E trifluoroacetates (30) and (32) which was interpreted as indicating that the vinyl cation (28) was a common intermediate.48 However an alternative explanation is that in the weakly nucleophilic solvent trifluoroacetic acid the two ion pairs (31) and (32) may interconvert more rapidly than they collapse to products (30)and (33).-H H 7 \+ H+ H,C=C=C /H++ +-H,C-C_C-CH \ /C=C-cH3 CH3 H3C (28) (27) or (29) H CH3 \ +/ / \ H3C/c=c OCOCF H H ,CH, \ I \ /OCOCF3 (31) /c=c\ H3C OCOCF 11 H3C CH E -(30) H -OCOCF3 Z-(33) H3C';'\CH3 Scheme 11 5 Bridgehead Reactivity Currently semi-empirical quantum mechanical' and empirical valence force- field (molecular mechanics)49 calculations are being applied to problems of structure conformation mechanism and reactivity. In sharp contrast to the above discussion emphasizing the roles of ion pairs and solvation effects attempts to interpret mechanism and reactivity in polar reactions by calculation usually require the approximation that the reactions are carried out in the gas phase.As might be expected if these calculations are compared directly with solution chemistry some alarming discrepancies arise. Bridgehead reactivity is one of the more successful areas of polar reactions in solution where calculations have been attempted. The change in strain (AHcalc)during the reaction is evaluated from empirical independent terms for bond stretching angle strain torsional strain and non- bonded strain. To reduce the number of empirical parameters required the hydrocarbon is used as a model for the bridgehead halide or sulphonate. Also R. H. Summerville and P. von R. Schleyer J. Amer. Chem. SOC.,1974 96 1 I 10.49 W. Parker R. L. Tranter C. 1. F. Watt L. W. K. Chang and P. von R. Schleyer J. Amer. Chem. SOC.,1974 96 7121. Reaction Mechanisms-Part (ii) Polar Reactions I 1 1 I I 1 I 1 I 1 I 1 18 ---16 -+Fey ii -14 -El-a- 12 - -10 8--&--@ I -@a-@ -+-p @-a 6-2-C1 Y,-4-$-4 2-CI 0-H&H -2 -H -4 --& -I I I I I I I I I 6 I I 0 12 3 4 5 6 7 8 9 10 -log k(70 "C) Figure 2 Calculated diflerences in strain energy between hydrocarbon and bridgehead cation (AHcalc)plotted against -log of the experimental rate constants in 80% ethanol ar 70 "C cf. R.C. Bingham and P. von R. Schleyer 3. Amer. Chem. SOC.,1971,93 3 189 with additional data for l-chZorobicyclo[2,2,2]octanecalculated from results for the corresponding bromide assuming k,,/k, = 40,and for l-chlorobicycio[3,3,3]undecane (35)jronare$ 49 the free cation is assumed to be a good approximation to the transition state.Then (AHcalc) is plotted against the experimentally observed rate constant. The results for chlorides at present the most extensive series available having a common leaving group are shown in Figure 2. Considering the wide range of relative rates (> 10") correlated the errors in extrapolations of experimental data at various temperatures to 70 "C,and the 130 T. W.Bentley approximations made in the calculations the results in Figure 2 are quite en- couraging. Predictions of reactivity within a rate factor of 10 should be possible using values of (AHcalc)for new systems by extrapolation or interpolation of a correlation line for data in Figure 2 e.g.for l-chlorobicyclo[4,4,4]tetradecane (34) and l-chlorobicyclo[3,3,3]undecane (39 Scheme 12."' Although the re1 rates -ca. lo4 1 (predicted) ca. lo6 ca. lo3 1 Scheme 12 predicted very high reactivity of (34) has yet to be tested experimentally the high reactivity of (35) indicates how much progress has been made from the time when all bridgehead positions were thought to be very unreactive in S,1 reactions. 6 hide Hydrolysis The site of protonation of amides in dilute acids and hence the mechanism of acid-catalysed hydrolysis of amides has received further attention [cf Ann. Reports (B),1972,69,390] with experimental evidence favouring hydrolysis by a displacement reaction of the N-conjugate acid.Spectroscopic evidence suggests that addition of water to benzamide in 100% sulphuric acid causes a tautomeric change from 0-protonated to N-protonated amide.50 For acid-catalysed (pH 5) exchange of the primary amide (36) it appears that exchange of H is slower than that of He because restricted rotation about the C-N bond of RCONH3+ preserves the inequivalence of the hydrogens. The interpretation is con-sistent with hydrogen exchange in amidinium ions (37).51 In the alternative 0\ /*, C-N /\ R He interpretation involving protonation of oxygen (38) He and H are non-equiva- lent but it is not clear why H exchanges more slowly than He 50 M. Liler J.C.S. Perkin II 1974 71.51 C. L. Perrin J. Amer. Chem. SOC.,1974 96 5628 5631. Reaction Mechanisms-Part (ii) Polar Reactions The extent to which these conclusions can be generalized is not yet clear e.g. deamination of the N-nitroso-2-pyrrolidone appears to proceed by hydrolysis I NO H/\NO H’ \“-OH Scheme 13 of the N-protonated species (Scheme 13),’ but deamination of N-n-butyl-N-nitrosoacetamide may occur by rate-limiting attack by water on the 0-protonated amide.’ 7Ester Aminolysis Many reactions involving the carbonyl group proceed via tetrahedral inter- mediates (T),but the rate-limiting step may be markedly dependent on the pH of the solution [Ann. Reports (B) 1970 67 751. Previous mechanistic inter- pretations of ester aminolysis have recently been modified and extended to the (general) Scheme 14.54 For most esters direct attack of the amme on the ester 0 H 0-OH 0 k1 k2 II I I I I1 RNH + C-OR s RN+C-OR e RN-C-OR -+ RN-c + -OR / k- I \ k-2 I I \ H H Tk (41) To (42) (product) HO 0-I + II I RN-C -OR RN-C-OR I\ II H H (product) T-(43) Scheme 14 s2 B.C. Challis and S. P. Jones J.C.S. Chem. Comm. 1974 748. 53 C. N. Berry and B. C. Challis J.C.S. Perkin ZI 1974 1638. ’4 A. C. Satterthwait and W. P. Jencks J. Amer. Chem. SOC.,1974,% 7018,7031 ;see also P. Y. Bruice and T. C. Bruice ibid. p. 5533. 132 T. W.Bentley (k,) appears to be rapid and reversible and for uncatalysed reactions of alkyl acetates it is suggested that the rate-limiting step is proton transfer through water (k2).General-base-catalysed reactions may involve rate-limiting removal of a proton from T' (41) by a second molecule of amine so that T' is trapped before it reverts to starting material.Similarly the general-acid-catalysed reac- tion involves trapping of T' by rate-limiting proton donation to give T+ (39) which proceeds to products via To(42). Because the proton exchanges are not sufficiently rapid to equilibrate the tetrahedral intermediates the same intermediate(s) are not necessarily formed by hydrolysis of imidates (40). Consequently the ratio of products formed by imidate hydrolysis does not necessarily indicate which step is rate-limiting in the corresponding ester aminolysis. This evidence modifies previous interpreta- tions which assumed that direct attack of amine on ester (k,) was rate-limiting [Ann.Reports (B),1968,65 781.Apparently the direct attack of amine on ester (k,) is only rate-limiting for reactive phenyl esters which accounts for a small sensitivity to substituents observed for such reactions. Less reactive phenyl acetates may react by rate- limiting expulsion of phenolate ion (k3)from T* (41),which explains why this step is independent of pH and of buffer. 8 Three-electron Oxidation One of the most extensively studied mechanisms of oxidation is the CrV' oxidation of alcohols which is usually drawn as a two-electron process involving rate-limit- ing hydrogen transfer in a chromate ester to yield ketone and Cr" (Scheme 15). Although the detailed mechanisms of the process are still under in~estigation,~~ HH HH H & \/ o rate \ Cr"' + \/ / c \ Ii __+ /C=O + Cr"' C -limiting /\ CO2H OH C0,H 0-Cr-OH CO,H II 0 Scheme 15 Cr" appears to be responsible for one-electron oxidations [Cr" -+ Cr"'] giving rise to side-reactions involving carbon-carbon bond cleavage reactions.56 Good evidence for the thermodynamically favourable three-electron oxidation [CrV1-+Cr'"] has recently been obtained from studies of primary deuterium isotope effects in the CrV1 oxidation of glycolic acid (Table 7).57 The primary isotope effect of 6.15 at low concentrations is consistent with a two-electron oxidation (Scheme 15).As both the rate constants and the isotope effects increase at higher concentrations it appears that there is a competing " H.Kwart and J. H. Nickle J. Amer. Chem. SOC.,1974 96 7572. ''" K. B. Wiberg and S. K. Mukherjee J. Amer. Chem. SOC.,1974 96,6647. F. Hasan and J. RoEek J. Amer. Chem. SOC.,1974,% 6802. Reaction Mechanisms-Part (ii) Polar Reactions Table 7 Dependence of deuterium-isotope effect on concentration for CrV' oxidation of glycolic acid (44). [Glycolic acidl/mol I-' 103kJs-' 103kD/s-' k JkD 0.0145 0.0826 0.0134 6.15 3.20 4.2 0.334 12.6 6.01 13.3 0.365 36.5 reaction in which two C-H or C-D bonds are broken in a single rate-limiting step. A possible mechanism involving three-electron oxidation of the complex (45) is outlined in Scheme 16. Among the other evidence for three-electron H HOC'HCO~H + /\ + C02-+ Cr"' I 0 G' CHO 'H Scheme 16 oxidation is the effect of radical scavengers on stoicheiometry and several co-oxidations (e.g.2-propanol with glycolic acid or with oxalic acid) appear to involve three-electron oxidations.