Ring-closure reactions. Part 15. Solvent effects on cyclic aralkyl ether formation by intramolecular Williamson synthesis

摘要

J.C.S. Perkin I1 Ring-closure Reactions. Part 15.j Solvent Effects on Cyclic Aralkyl Ether Formation by Intramolecular Williamson Synthesis By Antonella Dalla Cort, Gabriello Illuminati,' Luigi Mandolini,' and Bernard0 Masci," Centro di Studio sui Meccanismi di Reazione del Consiglio Nazionale delle Ricerche, c/o lstituto di Chimica Organica, UniversitA di Roma, 00185 Roma, Italy Product analysis, rate data, and effective mdarity (e.m.) values have been obtained for the formation of catechol polymethylene ethers by the intramdecular alkylation of o-o-bromoalkoxyphenoxides in Me,SO-water (99 : 1, v/v). Twelve ring sizes were investigated in the range 6-32. Comparison with similar data for the reaction in EtOH-water (75 : 25, v/v) (except for n 32) showed that the e.m.values are largely independent of solvent despite the large solvent effect observed in both the cyclization reaction and the related intermolecular model reaction. Small effects only were observed in the medium ring region. The main possible factors playing a role in solvent effects on intramolecular cyclization have been critically analysed in the light of the data. A THOROUGH understanding of the reactivity of intra- molecular reactions requires structural and solvation effects to be assessed on a firm experimental basis. The large rate accelerations of intra- over inter-molecular reactions (the so-called intramolecular effect) have stimulated a great deal of experimental as well as theore-tical study, which are of interest not only pev se, but also in connection with the elucidation of the mechanism of enzyme a~tion.~-~ Whereas structure effects have been the subject of several investigations, less attention has been devoted to solvent effects in spite of the fact that solvation has been suggested to have a role in the intra- molecular effe~t.39~ A suitable approach to fill this gap is to carry out a solvent-sensitive reaction in two sol- vents of basically different nature and to test ring-closure tendencies over a wide spectrum of ring sizes.A simple reaction is desirable in order to avoid com-plications arising from complex reaction mechanisms and/or possible changes in the rate-limiting step due to solvent change. The formation of catechol polymethyl- ene ethers (2) by intramolecular amp;type alkylations of (1) (2) o-o-bromoalkoxyphenoxides (1) seems to meet these conditions.In particular, the reaction of the 2-methoxy- phenoxide ion with butyl bromide, which is an inter- molecular counterpart of reaction (1) has been found to proceed in Me,SO-water (99 : 1v/v, referred to as 99 Me,SO) lo4 faster than in EtOH-water (75 : 25 v/v, referred to as 75 EtOH). The formation of the six-teen-membered ring (2; n = 16) is similarly affected by these solvent^.^. Since we have obtained rate data for cyclization re- action (1)in 75 EtOH 'for a number of ring sizes in the range 6-24, we now report similar data for 99 Me,SO solution which permit a comparison of the effect of the solvent on the effective molarity of reaction (1) over several ring sizes including 6-14, 16, and 24. The present study also includes the formation of the 32- membered ring, which is the largest polprnethylene ring compound considered so far in rate studies. RESULTS Rate measurements were carried out at 25.0 "C in S9yo Me,SO solution by monitoring the disappearance of the phenoxide absorption at 314 nm.The anions (1) were generated in situ by adding to the parent phenols a calcul-ated amount of KOH solution.6 Initial concentrations were in the range 6 x 10-5-3 x ~O-*M, i.e. low enough to ensure that the cyclization reactions are free from any appreciable contribution of second-order polymerization. First-order plots showed good linearity to at least 80 con-version and, whenever tested, the first-order rate constants (kobs)were independent of initial concentrations in the given range.The formation of the six- and seven-membered rings was too fast to be followed by conventional spectrophoto- metry. A stopped-flow technique was required in these cases. Cyclization to the six-membered ring, which is the fastest reaction studied, is also very rapid on an absolute scale, the half-life (0.7 ms) being of the same order of magnitude as the mixing time of the instrument. As pre-viously noted 6** the infinity spectra showed a residual phenoxide absorption which disappeared on adding a drop of concentrated hydrochloric acid. This phenomenon is partly due to a competing intramolecularly assisted elimin- ation of the E2 type.6** Since the side-reaction uses up another equivalent of base, the amount of KOH stock solution added at zero time was adjusted accordingly. The yields of cyclic products were calculated as 100 (1 -OD/OD,), where OD is the amount of absorption which dis- appeared upon acidification at infinity time, and OD, refers to zero time.Table 1 reports the results of the rate measure- ments and shows that, apart from the eight- and nine- membered rings, the extent of the side-reactions does not exceed 10 of the total reaction, so that only a minor cor- rection is required in most cases to convert the hobs values into the corresponding hintrrtvalues. In the case of the 24-and 32-membered rings the correction was somewhat more uncertain, because of a minor disturbance probably due to some precipitation of the products, which was still appreci- able at the lowest possible concentration, ca.5 x 10Pnr. Analysis of the kinetic solution at the end of the reaction showed that the eight-membered ring was fornied in 78 amp; 2 yield, in fairly good agreement with the value (83) estimated spectrophotonietrically. Furthermore, cyclis-ations leading to ring compounds (2; n = 11-14 and 16) have been carried out on a preparative scale under condi- tions close to those of the kinetic experiments,V with isolated yields in the range of 58-737;. TABLE1 Kinetic data for the ring-closure reactions of in !MyoMe,SO at 25 "C of Table 1 and of the effective molarity (e.m.)* data re- ported in Table 2.A complete set of e.m. values in 75:/0 EtOH at 50 "C has been reported,' except for ring size 32 whose study was made impracticable in this medium by very low solubility. Whenever possible, extrapolation to 25 "C from data at different temper- atures has been carried out on the basis of the known temperature coefficients of the rate constants.' It is worth noting that in spite of the fact that cyclisation rates are significantly faster in 99 Me,SO than in 75'j/, EtOH, rate ratios and e.m. values exhibit relatively small or, in some cases, no changes at all. In the case of the six-membered ring, the rate of formation is affected by the solvent change in exactly the same way as in the intermolecular model reaction.In some compounds with larger than six-membered rings, appreciable solvent effects are observed beyond the experimental error, allowance being made for the different temperature used. To discuss solvent effects on intramolecular reactions we may wonder to what extent the inherent reactivity of the end-groups of cto-bifunctional chain molecules is affected by possible changes in the solvation sliells around the end-groups as a function of the length of tlie chain. The simplest conceivable model is to consider the structure of the solvation shells as independent of chain length for a given cyclisation series. Examination of the entropies of activation for the present reaction in 75 EtOH,6 as well as for other cyclisation reactions,1deg; led to the suggestion that the above model is reasonably good, with the possible exclusion of the very short chains.Thus, the free energy change due to solvent reorganization when the two solvation shells surrounding the functional groups are replaced by a comtnon solv- o--OC,H,OCH,,-,Br no kobs/S-l 6 1.02 * 0.01 x lo3 7 4.07 f0.07 8 2.66 f.0.18 x 10-l 9 3.91 amp; 0.02 x 10 9.20 0.20 x 10-3 11 9.34 amp; 0.01 x 10-2 12 1.11 amp; 0.01 x 10-2 is 9.94 + 0.03 x 10-3 14 8.43 om x 10-3 16 8.79 0x3 x 10-3~ kintrs (99 Me,SO) Yield kintra (I kintrals-l (75 EtOH) 92 9.4 x 102 9 700 90 3.7 4 no0 89 2.2 x 10-1 7 300 67 2.6 x 4 600 91 8.4 x 10-3 2 400 92 8.6 x a 900 93 1.0 x 10-2 92 9.1 x 10-3 90 7.6 x 10-3 15 200 91 8.0 x 10-3 13 100 24 1.22 f0.08 x lov3 ca.95 1.2 x 10-2 5 100 32 1.25 * 0.05 x 10-2 ca. 95 1.2 x 10-2 Inter 2.95 x 10-l 9 700 a Size ol the ring to be formed. Average from 2-4 independent runs. c As based on spectrophotometric analysis. Calculated as koba (yield yo/lOO). At 25 "C. Rate con-stants in 75 EtOH were evaluated from data at different temperatures.' f The present value compares well with that (9.04 f0.1 1 x s-l) previously r~ported.~ The inter-molecular model reaction is the alkylation of 2-methoxy-phenoxide ion with butyl bromide. The rate constant (hinter) is given in 1 mol-1 s-1 (from ref. 5). DISCUSSION The rate data reported in Table 1 show that in 99 Me,SO the dependence of the reactivity on the chain length displays essentially the same features as those TABLE2 Comparison of e.m.data in 75 EtOH and 99 Me,SO for the ring-closure reactions of o--OC,H,OCH,,,-,Br E.m./M (I E.m./M E.ITl./M b Em. (75 HtOH; 50 "C) E.m. (750/, EtOFI; 25 "C) n (99 Me,SO, 25") (75 EtOII, 50") (75 EtOH, 25") Em. (99 Me2SO; 25 "C) E.m. (99 Me,SO; 25 "C) 6 3 200 3 900 3 800 7 12.5 36 25 8 0.75 1.8 0.99 9 0.088 0.31 0.19 10 0.028 0.15 0.11 11 0* 029 0.046 0.032 12 0.034 0.057 13 0.031 0.038 14 0.026 0.022 0.016 16 0.027 0.023 0.020 24 0.041 0.064 0.070 32 0.041 a This work. previously recorded in 75 EtOH,' namely, (i) a steep drop on increasing the chain length for the lower members of Gie series, (ii) the lack of a reactivity minimum in the medium ring region, and (iii) a remarkable insensitivity of the reaction rate to chain length in the large ring re- gion.The last point is further stressed by the 32-membered ring, whose ease of formation is comparable to those of other large rings. A 'loser Of reactivity data in the two vents is rendered possible by inspection of the last column 1.2 1 .o 2.9 2.0 2.3 1 .:I 3.5 2.2 5.4 3.9 1.6 1.1 1.7 1.2 0.85 0.62 0.85 0.74 1.6 1.7 From ref. 7. ation shell in the transition state assumed to be essentially constant ation series and, except for special cases,' iiimilar to the corresponding quantity for the intermolecular model reaction.The observed insensitivity of the e.m. values to solvent change indicates that in both solvents the quantities ( AG:solbsol;..)illtra.-( AG:solv.)illtrr are vanishingly * Calculated as kintra/klnter,where hinterrefers to an intermolec- ular model reaction (see footnote g in Table 1). A full definition of e.m. may be found in ref. 7. small in most cases. This is a striking result since the AGTsolv. term in an amp;2 reaction involving an anionic nucleophile is known to be quite large and should differ appreciably in the two solvents investigated. For in- stance, we have estimated that ASmlV. for the present reaction in 75 EtOH is a large positive quantity, i.e.+ 17 cal mol-l K-l, which is believed to arise mainly from extensive desolvation of the nucleophile in the transition state.The above evidence definitely demonstrates that solvation does not provide an explanation for the in- J.C.S. Perkin I1 solvent contributions to the Kintrafkjn+,er ratio occurs in two solvents which belong to neatly different categories.16 Thus, the most important factor upon transfer from the protic 75 EtOH to the essentially aprotic 99 Me,SO is the change of the inherent reactivities of the end- groups. Also, this work further emphasizes the import- ant role of the e.m. parameter in a general discussion of reactivity in intramolecular reactions. The finding that e.m. is, to a good approximation, independent of solvent, provides another illustration of its definition as a genuine, absolute measure of the ring-closure tendency of bi-functional chain molecules.tramolecular effect and that the major factor determining e.m. values is structure. Similar conclusions, although in a less straightforward manner, were drawn by Dafforn and Koshland l1 in a study of lactone and thiolactone formation in water and sulpholan, as possible models of solvation effects in enzymatic reactions, as well as by Bruice and Turner1, in an analogous comparison of anhydride formation in water and in Me,SO plus 1~-water. The most significant solvent effects, though small, are observed with the medium rings and presumably are related to the strain of the ring to be formed. It is worth noting that the e.m. values are generally higher in 75 EtOH than in 99 Me,SO.A tentative explanation may be found in terms of changing bond-breaking and bond-making contributions at the transition state in the two solvents by applying Thornton's reaction bond rules l3 to the reaction at hand. Transfer from 7576 EtOH to 99yoMe,SO will increase the nucleophilicity of the nucleophile and decrease the leaving-group ability of the leaving group. The two factors oppose each other, with the net result that motion along the re-action co-ordinate (parallel effect) will be little affected. However, the indicated solvent change would favour a simultaneous decrease of both nucleophile-carbon and carbon-leaving-group distances (perpendicular effect) and cause the transition state to be tighter and, speci- fically, the oxygen-carbon distance to be shorter in 99 Me,SO than in 75 EtOH.In consequence, the cyclic transition states leading to the medium-sized rings in 99 Me,SO should be affected to a greater extent by the steric strain typical of medium rings, i.e., transannular interactions and bond opposition forces,14 which are due to steric crowding. A final comment may be devoted briefly to the pro- blem of solvent-induced conformational changes of chain molecules and of the possible effects of these changes on the ease of cyclisation, as pointed out by Winnik.15 Although the present comparison is confined to two sol- vents only, which makes it difficult to assess subtle effects of this sortJ15 the observed behaviour, notably the virtual constancy of the e.m.values for the large rings, seems to indicate the absence of a significant influence of any pos-sible modification of the chain conformation upon trans- fer from 75 EtOH to 99y0Me,SO. As a concluding remark, we note that in spite of minor effects detected in the medium ring region, for the remaining ring sizes the essential cancellation of tlie EXPERIMENTAL The mixed solvent (99 Me,SO) and the KOH stock solution (2 x 10-2~in 93 Me,SO) were prepared, stored, and handled as before.G The apparatus was largely as pre- viously reported.G Stopped-flow kinetics were carried out on a Durrurn-Gibson stopped-flow spectrophotoineter model D-110, matched with a Hewlett-Pacltard storage oscillo- scope model 1207 R.MateviaZs.-l, 28-Dibronio-octacosane was prepared in 33 yield by a modification l7 of the Woolford procedure,18 m.p. 78-79 "C (from EtOAc) (1it.,l8 78-79 "C). 28-Bromo-octacosyl o-hydroxyphenyl ether was synthesized from 1,28- dibromo-octacosane and catechol in a way similar to that reported for the preparation of 20-bromoicosyl o-hydroxy- phenyl ether. Elution on silica gel of the crude reaction product with CC1, gave a 29 recovery of unchanged di- bromoalltane. Further elution with CHC1, gave the pure title conipound in 37 yield, m.p. 65-67 "C. The spec- tral properties were similar to those of the lower homo- logues @ and confirnied the expected structure. Bromine content was within 0.5 of theory.Other materials were available from previous investigations.8* @ Product analysis was carried out for the cyclisation of (1; n = 8) on a scaled-up kinetic experiment. After the standard work-up, the yield of cyclic product was deter- mined by g.1.c. analysis (internal standard) on a 1m column, packed with 2 SE 30 plus 0.5 FFAP on Chromosorb W 60-80, operated at 120". 0/398 Received, 1lth March, ISSO REFERENCES Part 14, C. Galli, G. Illuminati, and L. Mandolini, J. Org.Chem., 1980, 45, 311. R. D. Gandour, 'Transition States of Biochemical Pro-cesses ', eds. R. D. Gandour and R. L. Showen, Plenum Press, New York, 1978, ch. 14, p. 529. T. H. Fife, Adv. Phys. Org. Chrm., 1975, 11, 1. R. L. Showen, ' Transition States of Biochemical Pro-cesses ', eds.R. D. Ganclour and R. L. Showen, Plenum Press, New York, 1978, ch. 2, p. 77. L. Mandolini, B. Masci, and S. Roelens, J.Ovg. Chew, 1577,42, 3733. G. llluininati, L. Mandolini, and R.Masci, J. Amer. Chew. SOC.,1975, 97, 4960. G. Illuminati, L. Mantlolini, and B. Masci, J. Amer. Chem. SOC.,1977, 99, 6308.* G. Illuminati, L. Mandolini, and euro;3. Masci, J. Anzer. Chew SOC.,1974, 96, 1422. L. Mandolini and euro;3. Masci, J. Ovg. Chem., 1977, 42, 2840. L. Mandolini, J. Amer. Chem. SOC., 1978, 100, 5.50. ti. A. Lkfforn and D. E. Ko.~hland,jun., J. Amer. Chmz. SOC., 1977, 99, 7246. la T. C,. Rriiice and A. Tiirner, J. Amr. Chem. SOC.,1970, 92, 3422. j3 E. I. Thornton and E. R. Thornton, ' Transition States of Biochemical Processes ', eds. I. D. Gaticlour and I. L. Showen, Plenum Press, New York, 1978, ch. 1, p. 1. la J. Sicher, Progv. Stcrcochcm., 1962, 8, 202. Ib M. A. Winnik, Accounts Cltem. Res., 1977, 10, 173. l8 G. Tlluniinati, * 'Techniques of Cheniistry ', ed. M. R. J. Dack, Wiley, New York, 1976, vol. 8/2, ch. 12, p. 159. l7 C. Galli, G. Giovannelli, G. Illuminati, and L. Mandolini, J. Org. Chcm., 1979, 44, 1258. la R. G. Woolford, Canad. J. Chem., 1962, 40, 1846.