INGOLD LECTURE. Reactive Intermediates: Carboxylic Acid Enols and Other Unstable Species A. J. Kresge Department of Chemistry, University of Toronto, Toronto, Ontario, Canada M5S 3H6 1 Introduction Enols are the tautomeric isomers of carbonyl compounds through which many important chemical and biological reactions occur, and, if we wish to understand these reactions, and through understanding to control them, we must understand the chemistry of enols. Most enols, however, are also quite unstable, both thermodynamically and kinetically, and their lifetimes in solution are usually very short. This has impeded their study. Enols have been known for more than a century,' but very little hard quantitative information about the rates and equilibrium constants of their reactions was available for very nearly all of that time. The situation changed dramatically about 15 years ago when we and others developed methods for generating enols in greater than equilibrium amount under conditions where they could be observed directly and their reactions studied in detail.This has produced a wealth of reliable new information about en01 isomers of simple aldheydes and ketones.* That work is now being extended to the more difficult task of examining the much more labile enols of carboxylic acids, esters, and amides. This review will begin with a short description of some of our work on enols of aldehydes and ketones in order to illustrate the methods that we have used and the kinds of information that can be obtained.That will then be followed by an application of these methods to enols of carboxylic acids and their derivatives. 2 Methods We have found flash photolysis to be an especially useful technique for studying the chemistry of short-lived enols. An example of this method is provided by the Norrish type I1 photoelimination of 5-hydroxypentan-2-one1,eqn. (1).3 Irradiation of this substance leads to hydrogen abstraction from the y-position by the oxygen atom of the photoexcited carbonyl group to give the diradical2, which then Jerry Kresge received his undergraduate education at Cornell University and obtained a PhD in synthetic organic chemistry at the University of Illinois working with Nelson Leonard. He then spent the next four years learning physical organic chemistry, jirst as a Fulbright Scholar with Hughes and Ingold at University College London and then as a Postdoctoral Fellow,jrst with H.C. Brown at Purdue and then with C. G. Swain at M. I. T. Kresge 'sjrst independent research position was as Associate Chemist at Brookhaven National Laboratory. From there, he moved to an academic post at the Illinois Institute of Technology, and, some twenty years ago, he moved again, to Toronto, where he is now Emeritus Professor of Chemistry. Kresge has worked in a number of areas of physical organic chemistry, using kinet- ics, acid-base catalysis, and isotope efSects a5 principal investigative tools. His accom-plishments have been recog-nized by a number of honours: he has held Guggenheim, Killam, and Yamada Fellow- ships, and has received the Syntex Award and the Morley Medal; he has also been visiting professor at many universities.1 2 3 4 fragments to the enol of acetone 3 and the enol of acetaldehyde 4. These reactions occur very rapidly, and, if the process is initiated by a sufficiently short and intense burst of light in a flash photolysis apparatus, the enols can be generated much faster than they keton- ize to their carbonyl isomers. The ketonization reactions can then be monitored by following changes in UV absorbance of the enols. Ketonization of acetone enol is about two orders of magnitude faster than ketonization of acetaldehyde enol, and accurate rate con- stants for both enols can easily be obtained.We have also made enols by the flash photolytic oxidation of alcohols and reduction of carbonyl compounds: as is illustrated in eqn. (2) for acetone. Irradiation of acetone in the presence of iso-propyl alcohol leads to an intermolecular analogue of the Norrish type I1 reaction, giving a pair of ketyl radicals 5;these then dispro- portionate, producing the enol of acetone and regenerating the alcohol. Photohydration of acetylenes, eqn. (3), is another useful method OH 0 of generating en01.s.~ The thermal hydration of acetylenes is known to occur through enol intermediates, but, in the strongly acidic solu- tions needed to effect this process, the enol ketonizes as fast as it is formed and consequently cannot be observed.Photoexcitation, however, greatly enhances the reactivity of acetylenes towards electrophilic addition, and now the enol is generated much more rapidly than it is consumed; its reactions may therefore be moni- tored. The photohydration of acetylenes occurs through vinyl cation intermediates such as 6, and vinyl cations can also be produced by photoionization of vinyl halides. We have consequently found that we can generate enols by the flash photolytic solvolysis of vinyl chlorides as well: as is illustrated in eqn. (4). Me rvre Me Me Paramount among the things one would like to know about enols is the magnitude of the equilibrium constants for their formation 275 from keto isomers. Such keto-enol equilibrium constants, KE, have traditionally been determined by the Kurt Meyer halogen titration method, which uses the fact that enols react with halogens whereas their keto isomers do not.This method works well when enol con- tents are not too low, as is the case, for example, with P-dicarbonyl compounds. With most monofunctional aldehydes and ketones, however, the method fails badly because enol contents here are in the parts-per-million or even parts-per-billion range, and impurities that react with halogen must be excluded at better than these levels. This difficulty may be avoided by determining keto-enol equi- librium constants as ratios of rate constants: those for enolization, k,, divided by those for ketonization k,: KE = k,/k,. Values of k, are easily obtained by long-established methods such as halogen scavenging of the enol as it forms, and values of k, may be supplied by the flash photolytic methods we have developed.This technique produces keto-enol equilibrium constants whose accuracy is limited only by that with which the component rate constants can be determined, and it can give good results no matter how small the value of K, actually is. Another important property of enols is the acid strength of their hydroxy groups. Standard methods for determining acidity con- stants of stable acids cannot, of course, be applied to short-lived enols, but accurate estimates of enol pK, values may nevertheless be obtained from the way their rates of reaction change with pH.Fig. 1 shows a typical enol ketonization rate profile. At high acidity 1o5 1o4 1o3 3 lo2 10' 1oo Figure 1 Rate profile for the ketonization of the enol of indan-I-one in aqueous solution at 25 "C there is an acid-catalysed process that involves rate-determining protonation of the enol on its @-carbon by hydronium ion; this is represented by the diagonal segment with slope = -I on the left side of Fig. 1. As the acidity drops, this reaction gives way to keton- ization by @-carbon protonation of the very much more reactive enolate ion; since this process produces a hydronium ion in an initial pre-equilibrium step and then uses it up in the rate-determining step, the overall reaction is independent of acidity and it consequently appears as the 'uncatalysed' segment of slope = 0 near the centre of Fig.1. At still lower acidities, the concentration of hydronium ions is too low to sustain this process and @-carbon protonation of enolate ion by solvent water takes over; since the hydronium ion produced in the pre-equilibrium is now not used up in the rate-deter- mining step, the rate of reaction becomes inversely proportional to hydronium ion concentration and directly proportional to hydrox- ide ion concentration, giving the region of apparent hydroxide ion catalysis represented by the segment of slope = +1 on the right side of Fig. I. Eventually, however, the position of the pre-equilibrium shifts from enol to enolate, and the advantage of converting a less reactive to a more reactive substrate is lost; hydroxide ion catalysis then becomes saturated and the rate levels off to a constant value giving the second 'uncatalysed' segment of slope = 0 at the far right of Fig.1. Straightforward analysis of rate data in the region of transition from hydroxide ion catalysis to catalysis saturation gives both the rate constant for the carbon protonation step and the equi- librium constant for the prior equilibrium step. The latter, of course, CHEMICAL SOCIETY REVIEWS, 1996 is the acid dissociation constant of the enol. We have used this method to determine pK, values of very short-lived enols, some with lifetimes of only a microsecond. As shown in eqn. (3,keto-enol isomerization and acid ioniza- Ph,C& + H* tion of the enol form two legs of a thermodynamic cycle whose third member is ionization of the keto form as a carbon acid.Once the equilibrium constants K, and KF for the first two legs have been determined, therefore, that for the third member, KF, can be calcu- lated as the product K,K:. In some cases the keto isomer is suffi- ciently acidic to allow Kf to be determined directly, and then the cycle serves as a test of the internal consistency of the data, for the pK values should sum to zero around the cycle. The diphenyl- acetaldehyde system, shown in eqn. (5), provides an especially good example of such a situation, for here we were able to measure each of the three constants by two independent methods.' The results obtained, pK, = 0.98 5 0.04, pK: = 9.40 ? 0.01 and pK5 = 10.42 ? 0.02, do sum up to a value that is less than the rather small combined experimental uncertainty of the individual measurements: ZpK = 0.04 ? 0.05.A sample of results we have obtained for representative simple aldehydes and ketones is presented in Table 1. It may be seen that Table 1 Equilibrium constants for some simple aldehyde and ketone systems in aqueous solution at 25 "CU Substrate QK, PKE QK,KP 6.23 10.50 16.73 8.33 10.94 19.27 7.96 10.34 18.31 Ph 3.86 11.63 15.49 0.98 9.40 10.42phFoPh 6.39 11.70 18.09d "Ionic strength = 0.10 mol I-'; acidity constants are concentration quotients at this ionic strength. K, for acetaldehdyde is very small: less than 1 ppm of en01 is in equilibrium with the keto form.The enol content of acetone is even lower, by some two orders of magnitude. This difference between acetone and acetaldehyde may be attributed to stabilization of the keto isomer of acetone by its additional methyl group, and the next entry in Table 1 shows that phenyl is just about as good as methyl in this respect. Methyl substitution in the @-position, as in iso-butyraldehyde, on the other hand, raises KE, which may be under- stood in terms of alkyl group stabilization of the enol double bond. Phenyl groups in this position are even better than methyls, as shown by the relatively large enol content of diphenylacetaldehyde. A conformational effect appears to operate in the case of cyclo- hexanone, for its enol content is considerably greater than that of corresponding acyclic ketones and also greater than that of cyclo- pentanone and cycloheptanone.4 The acid strength of these enols is similar to that of phenols, as might be expected from the vinyl alcohol structures they have in REACTIVE INTERMEDIATES CARBOXYLIC ACID ENOLS AND OTHER UNSTABLE SPECIES-A J KRESGE common There is not much variation in pK: with structure for the group of substances shown in Table 1, much less than that in pK, or pKt Structural effects on pKt, moreover, appear to parallel those on pK,, and there is in fact a fairly good linear relationship between these two quantities for a substantial number of aldehydes and ketones 3 Mandelic Acid and Derivatives As is illustrated by the examples provided above, the enol isomers of simple aldehydes and ketones are usually highly labile and thermodynamically unstable substances Carboxylic acid enols are even more labile and unstable This is apparent, for example, from a comparison of the keto-enol transformation of acetaldehyde, eqn (6),with that of acetic acid, eqn (7) For acetaldehyde, pK, = 6 23 q04==(a 0) OH OH and the hydronium ion catalytic coefficient for ketonization of the enol is k; = 33 mol I 1 s Neither of these quantities has as yet been measured directly for acetic acid, but reliable estimates can be made Guthrie has presented several different arguments that lead to the consistent result pK, = ca 20p and k; = 10s-lo9 mol I I s-I, may be estimated on the basis of the experimentally determined value k; = 7 X lo5mol 1 s I for the dimethyl ether of acetic acid enol, CH2=C(OMe),,Io and the fact that methyl vinyl ethers are one to two orders of magnitude less reactive than the corresponding enols This comparison shows that the enol content of simple car- boxylic acids with no enol-stabilizing substituents can be expected to be many orders of magnitude smaller than that of simple alde- hydes and ketones and that the rates of ketonization of these car- boxylic acid enols will be so fast as to approach the diffusion controlled limit Daunting realizations such as these have directed studies of car- boxylic acid enols to systems in which special structural features stabilize the enol One such approach uses bulky aromatic substitu- ents to block access to the enol double bond, through whose proto- nation ketonization must take place This is a technique that has produced remarkably stable enols of aldehydes and ketones, as shown in work pioneered by Fuson some 50 years ago and lately elaborated by Rappoport Its application to carboxylic acid systems has produced enols with lifetimes as long as several hours, l2 but these long lifetimes have also produced complications in the form of radical-producing oxidation reactions of the very elec- tron-rich enols Our own approach has been to increase the lifetime of the enol somewhat less by introducing subctituents that we know, from our studies of aldehyde and ketone systems, stabilize enols through their electronic effects When we began our work, there were reports in the literature that irradiation of esters of benzoylformic acid 7 leads to a Norrish type I1 reaction, eqn (8), which produces 7 8 hydroxy(pheny1)ketene8 l4 We reasoned that in aqueous solution this ketene would be hydrated to an enol 9, eqn (9),and that the p-OH and P-Ph groups of this enol might slow its ketonization to 9 10 mandelic acid 10 sufficiently so that the enol would decay more slowly than it was formed and thus be observable We found that flash photolysis of benzoylformic acid esters in aqueous solution did indeed produce transient species which, through solvent isotope effects and the form of acid-base catalysis, we could identify as a rapidly reacting ketene and a somewhat more stable but still rapidly reacting enol The rate profile for ketonization of mandelic acid enol is shown in Fig 2 It is similar to rate profiles for the ketonization of simple log I 1”o3 w 102 tuu L“”’lo1 lo3 105 10’ log 10” 10j3 toi5 WyM Figure 2 Rate profile for the ketonization of mandelic acid enol in aqueous solution at 25 “C aldehyde and ketone enols (cf Fig I), with the exception that the bend representing acid ionization of the enol comes at a lower acidity because this enol is a stronger acid (pKt = 6 62) This is typical of carboxylic acid enols and may be attributed to the pres- ence of the second, geminal hydroxy group, a similar acid-strength- ening effect may be seen in going from ethanol 11,with pK, = 15 9, to acetaldehyde hydrate 12,with pK, = 13 6 4-OH 11 12 In order to obtain a keto-enol equilibrium constant for the man- delic acid system by the rate constant ratio method (KE= k,/k,), we measured rates of acid-catalysed enolization of the acid This reac- tion proved to be very slow, and we had to resort to high tempera tures (140-155“C) in order to get conveniently measurable rates Extrapolation of these data to 25 “C then gave a rate constant, which, when combined with the acid-catalysed rate of ketonization, provided the result pK, = 15 4 Combination of that value with pKk = 6 6 then produced pKE = 22 0 for mandelic acid ionizing as a carbon acid These results show the enol content of mandelic acid to be very low but still significantly greater than the estimate pK, = 20 for acetic acid The difference may be attributed to the p-OH and p-Ph groups of mandelic acid enol, for such substituents are known to stabilize enol isomers in simple aldeheyde and ketone systems 2 Our characterization of the mandelic acid keto-enol system, in addition to being of general chemical interest, has proved to be of specific value to biological chemists in connection with the enzy- matic racemization of mandelic acid by mandelate racemase, and the current controversy over whether the enzyme achieves its effi- ciency by an electrostatic effect or by formation of a strong hydro- gen bond We hdve written the acid ionization of mandelic acid enol, eqn (lo), as involving one of the geminal hydroxy groups, specifically that trans to phenyl, because we found the trans-enol of phenyl-acetaldehyde, 13,to be a stronger acid than the cwisomer 14 l7 In principle, however, the ionization of mandelic acid enol might also 278 A Ph PhoH u 14 involve the hydroxy group a to phenyl.In order to investigate this matter, we wished to examine an enol for which such ionization is blocked by conversion of this hydroxy to methoxy, and we hoped to be able to make that substance, 15, by hydration of the ketene generated through the photo-Wolff reaction shown in eqn. (11). 16 15 Flash photolysis of methyl phenyldiazoacetate 16 did produce a transient species which we identified as an enol, but the product of the reaction was methyl mandelate 17 rather than the expected methyl ether of mandelic acid 18, eqn.( 12).Is We postulate that this MeOH: - p h M G (12) Ph OH 18 substance was formed by insertion of the methyoxycarbonylcar- bene 19, formed by loss of nitrogen from methyl phenyldiazo- acetate, into an O-H bond of solvent water. Such carbene insertion reactions are well known, and a number of different mechanisms have been proposed for them, all of which involve only the carbenic carbon. Our detection of an enol intermediate in the process indi- cates that the carbonyl group is involved as well and that the reac- tion occurs as shown in eqn. (1 3); the process might therefore be 17 more accurately described as conjugate addition of water across the entire carbonylcarbene functional group.We have since found other examples of such conjugate additions, in the cyclic system shown in eqn. (14)and also in the amide system N2 OH OH CHEMICAL SOCIETY REVIEWS, 1996 shown in eqn. (15). This has taken us into the chemistry of enols of carboxylic acid esters and enols of carboxylic acid amides. 4 Phenylcyanoacetic Acid We have found that the enol20 of phenylcyanoacetic acid 21 can be generated from a diazo compound precursor, 22, eqn. ( 16).19 In this K CN 22 20 21 case flash photolysis gives the Wolff rearrangement rather than con- jugate addition of water to a carbonylcarbene, and the ketene so pro- duced is hydrated to the enol.The rate profile for ketonization of this enol, shown in Fig. 3, is \b pK':=8.70 1o2 ioo tw100 -104 L.,.. ,, 4L*'*L'L 12 [H*W Figure 3 Rate profile for the ketonization of phenylcyanoacetic acid enol in aqueous solution at 25 "C somewhat different from those of Figs. 1 and 2 in that there is an additional bend at the high acidity end. Through solvent isotope effects and the form of acid-base catalysis, we have been able to assign this bend to ionization of the first O-H group of the enol ,and the bend at [H+] = ca. mol 1-I to ionization of the second O-H group. The plateau before the first bend then represents reac- tion through pre-equilibrium ionization of the enol followed by rate- determining carbon protonation of the enolate ion, and the diagonal segment of slope = -I represents the same reaction with enolate ion as the initial state.Near [H+] = lop6mol I-', carbon protona- tion of the more reactive dianion by water takes over, with reaction first starting from the monoanion as the initial state, giving a region of apparent hydroxide-ion catalysis, and then reaction starting from the dianion as the initial state, leading to catalysis saturation. With the aid of rates of enolization measured by bromine scav- enging, we were able to characterize the phenylcyanoacetic acid keto-enol system completely. The results are presented in Scheme 1. It may be seen that the enol is quite strongly acidic with pK: = 0.99 CN CN PKE = 7.22-l'h&O$-I OH 0.99 -+ Ht Scheme 1 REACTIVE INTERMEDIATES CARBOXYLIC ACID ENOLS AND OTHER UNSTABLE SPECIES-A J KRESGE for its first ionization, this makes the enol more acidic than the car- because their conjugate bases are stabilized by delocalization of boxylic acid group of its keto form’ Even in its second ionization, negative charge into the five-membered ring, which produces with pK’F = 8 70, this enol is more strongly acidic than the enols of an aromatic cyclopentadienyl anion The enol contents are high simple aldehydes and ketones (cf Table l), which attests to the strong acidifying effect of the cyano group The difference between the first and second pK,, ApK, = 7 7, is not unlike that for the first and second ionization of carbonic acid, ApK, = 6 4 The cyano group also stabilizes the neutral enol strongly, raising the enol content of phenylcyanoacetic acid over that of mandelic acid by eight orders of magnitude 5 Cyclopen tadienecarboxy lic Acid and Derivatives The cyclopentadiene moiety also stabilizes enols strongly, and this effect has been used in studies of aldehyde and ketonez0 as well as carboxylic21-23 acid systems The enol 23 of cyclopentadiene-carboxylic acid itself 24, was made by hydration of the ketene 25, which was generated in two ways by photo-Wolf reaction of the corresponding diazo compound, 26, and also by photo-induced elimination of HBr from 2-bromophenol, 27, as shown in eqn 17 This enol and that of mandelic acidi5 were the first carboxylic acid enols to be Characterized in aqueous solution, unfortunately, however, there are problems with some of the originally published data on the cyclopentadienecarboxylic acid system,21b and revised values have not yet been made public The monobenzo B2Ib’*and dibenzo 2923 analogue of cyclo- pentadienecarboxylic acid enol have also been investigated The 29 30 former is of some commercial importance to the photolithographic industry, for the diazonaphthoquinone 30 precursors of its deriva- tives are constituents of photoresistsz4 and the enols are intermedi- ates in the photolithographic process 22 It may be seen from the results obtained, summarized in Table 2, that these enols are quite strongly acidic, this is Table 2 Summary of results for cyclopentadienecarboxylicacid keto-enol systems0 PK,E 13 19 22 PK ’,” 8 7h 83 96 PKF 8 46 93 95 PK E 5oh 66 83 pK ’,“ 13 7 15 2 17 9 uIonic strength = 0 10 mol I I, acidity constants are concentration quotients at this ionic strength Equilibnum constants are as defined in Scheme 1 hRevised value provided by Professor Wirz as well because the enols are stabilized by fulvenoid aromatic delocalization It IS significant that the acids become somewhat weaker and the enol contents somewhat smaller with increasing benzo substitution, for the resonance energy of the cyclo- pentadienyl ring can be expected to decrease with benzo sub stitution 6 Acetoacetic Acid Keto-enol tautomerism in P-keto esters has been investigated almost since the beginning of enol chemistry, and these systems provide some of the most extensively documented examples of such isomerism available today In striking contrast, very few studies of tautomerism in P-keto acids have been carried out, and only one of these has dealt with the prototype substance, acetoacetic acid 25 The acetoacetic acid system is especially interesting because two different enols may be formed, the carboxylic acid enol31, and the ketone enol32, eqn (1 8) The ketone enol might be expected to be 31 32 the much more stable of the two and the overwhelmingly pre- dominant enol form in solution, for keto-enol equilibrium constants for simple ketones are very much greater than those for simple acids cf pK, = 8 for acetone vs pK, = 20 for acetic acid Much of this difference between acetone and acetic acid, however, must be an initial-state effect, inasmuch as the keto form of acetic acid is stabilized by conjugation of its carbonyl group with the adjacent hydroxy substituent, eqn (19), whereas such conjugative initial state stabilization is not possible in the case of acetone Initial state effects, of course, have no bearing on the relative abundance of the two enols of acetoacetic acid, for both are in equilibrium with the same initial state keto form [eqn (18)] The relative amounts of the two enols will be determined, rather, by the intrinsic stabilities of the enols themselves, and this may well be not as disparate as SUE- gested by the acetone-acetic acid comparison It is difficult to address this matter experimentally because the enol content of acetoacetic acid is small (pK, = 2 25, vide infru) and the enol(s) are short-lived We have therefore performed ab initio calculations on the system at the MP2/6-3 11+G** level The results show an energy difference of AE = 11 kcal mol I (1 cal = 4 184 J) between the two enols, with the ketone enol as the more stable isomer z6 This result is thus qualitatively consistent with the prediction made from the acetone-acetic acid comparison It is sig nificant, however, that the calculated energy difference is less than the 16 kcal mol I that corresponds to the pK, difference for acetone-acetic acid We obtained the enol of acetoacetic acid by hydration of acetylketene33, eqn (20), which was generated by flash photolysis of 2,2,6-trimethyl-4H-1,3-dioxin-4-one34, eqn (21) 27 The rate profile for ketonization of the enol showed bends that could be attributed to ionization of its two acidic groups and from which the pK, values shown in eqn (22) were obtained The first of these, pK, = 4 05, shows the carboxylic acid group of the enol to be consider- ably more acidic than the value, pK, = 5 67, that can be predicted from a correlation of acidity constants of substituted acrylic acids z8 34 + 2H++H+ The difference may be attributed to the hydrogen bond formed between the hydroxy group of the enol and the carbonyl group of the acid, which is not taken into account by the correlation, this hydrogen bond will become stronger as the acid ionizes, thus stabi lizing the ionized state This hydrogen bond also stabilizes the unionized state of the second ionization, making the enol a weaker acid than it would otherwise be This enol, with pK: = 13 18, is in fact unusually weakly acidic cf pKf = 9 48 for ionization of the enol of the corresponding methyl ester, eqn (23) 29 Another factor contributing to the weak acidity of the enol of acetoacetic acid is electrostatic repulsion between the two negative charges in the dian- ionic product In the only other investigation of the acetoacetic acid keto-enol system,25 rates of enolization were determined by bromine scav- enging, and combination of those with our results for ketonization leads to pK, = 2 25 This is a rather low value for a P-dicarbonyl system, for example, pK, = 1 10 for tautomerism of the corre- sponding methyl ester, eqn (24) 29 The difference may again be attributed to hydrogen bonding, this time in the initial state of the acetoacetic acid reaction, eqn (18), an effect that is absent from the methyl ester system Acknowledgements I am grateful to the Natural Sciences and Engineering Research Council of Canada, the United States National Institutes of Health, and the Petroleum Research Fund administered by the American Chemical Society for financial support of our work CHEMICAL SOCIETY REVIEWS, 1996 7 References 1 E Erlenmeyer, Chem Ber ,1881,14,320 2 See, e g Z Rappoport, The Chemistry of Enols, Wiley-Interscience, New York, 1990 3 Y Chiang, M Hojatti, J R Keeffe, A J Kresge, N P Schepp and J Wirz, J Am Chem SOC ,1987,109,4000 4 J R Keeffe, A J Kresge and N P Schepp, J Am Chem SOC , 1990, 112,4862 5 Y Chiang, A J Kresge, M Capponi and J Wirz, Helv Chim Acta, 1986,69,1331 6 A J Kresge, and N P Schepp, J Chem Soc ,Chem Commun ,1989, 1548 7 Y Chiang, A J Kresge and E T Krogh, J Am Chem Soc ,1988,110, 2600 8 J R Keeffe and A J Kresge, in The Chemistry of Enols, ed Z Rappoport, Wiley-Interscience, New York, 1990,ch 7 9 J P Guthrie, Can J Chem , 1993,71,2123, J P Guthrie and Z Liu, Can J Chem ,1995,73,1395 10 A J Kresge and M Leibovitch, J Am Chem SOC ,1992,114,3099 11 This work IS reviewed by H Hart, Z Rappoport and S E Biali, in The Chemistry of Enols, ed Z Rappoport, Wiley-Interscience, New York, 1990,ch 8 12 P O’Neill and A F Hegarty,J Chem SOC ,Chem Commun ,1987,744, B M Allen, A F Hegarty, P O’Neill and M T Nguyen, J Chem Soc , Perkin Trans 2, 1992,927 13 J FreyandZ Rappoport,J Am Chem Soc ,1995,117,1161,1996,118, 5182 14 M V Encinas,E A Lissi,A Zanocco,L C StewartandJ C Scaiano, Can J Chem ,1984,62,386 15 Y Chiang, A J Kresge, P Pruszynski, N P Schepp and J Wirz, Angew Chem ,lnt Ed Engl ,1990,29,792 16 J P Guthrie and R L Kluger, J Am Chem SOC , 1993, 115, 11 569, G L Kenyon, J A Gerlt, G A Petsko and J W Kozarich, Acc Chem Res ,1995,28,178 17 Y Chiang A J Kresge, P A Walsh and Y Yin, J Chem SOC ,Chem Commun ,1989,869 18 Y Chiang,A J Kresge,P Pruszynski,N P ScheppandJ Wirz,Angew Chem ,/nt Ed Engl ,1991,30,1366 19 J Andraos, Y Chiang, A J Kresge, I G Pojarlieff, N P Schepp and J Wirz, J Am Chem Soc , 1994,116,73 20 M P Harcourt and R A More O’Ferrall, J Chem Soc , Chem Commun ,1987,822,J Chem SOC ,Perkin Trans 2,1995,1415 21 (a)B Urwylerand J Wirz,Angew Chem ,Inr Ed Engf ,1990,29,790, (6)J I Kim,B Urwylerand J Wirz,J Am Chem SOC ,1994,116,954 22 J Andraos, Y Chiang, C G Huang, A J Kresge and J C Scaiano, J Am Chem SOC , 1993,115,10605 23 J Andraos, PhD Thesis, University of Toronto, 1992 24 See, eg E Reichmanis, in Polymers for Electronic and Photonic Applications, ed C P Wong, Academic Press, New York, 1993, pp 67-117 25 K J Pedersen,J Phys Chem, 1934,38,999 26 S Hoz and A J Kresge, unpublished work 27 Y Chiang, H X Guo, A J Kresge and 0 S Tee, J Am Chem SOC , 1996,118,OOOO 28 D D Perrin, B Dempsey and E P Sergeant, pKa Predictions for Organic Acids and Bases, Chapman and Hall, New York, 1981,p 127 29 J W Bunting and J P Kanter, J Am Chem SOC ,1993,115,11705