Kinetic Isotope Effects in Some Reactions of Tri-(4-nitrophenyl)- methane with Alkoxide Bases in Various Alcoholic Media B Y EDWARD F. CALDIN,* ELEANOR DAWSON, RICHARD M. HYDE? AND ALAN QUEEN: University Chemical Laboratory, University of Kent at Canterbury, Canterbury, Kent Received 30th May, 1974 Rate constants and Arrhenius parameters have been determined for proton-transfer and deuteron- transfer reactions of tri-(4-nitrophenylmethane) with alkoxide ions in alcohol + toluene mixtures. The kinetic isotope effects (kH/kD) at 25°C are considerable (8.4 in ethanol+ toluene, 8.9 in propan-2-01 + toluene, 23 in t-butanol + toluene) and indicate that the rate-limiting step is a proton transfer. For the forward reactions, the results suggest that the importance of tunnelling increases with the size of the alkoxide group.For the backward reactions, the rates for proton- and deuteron-transfer are different, although they would be expected to be equal by reason of rapid exchange with the solvent ; this suggests that in the product the proton remains firmly bonded to the carbanion. The reaction (1) between tri-(4-nitrophenylmethane) (TNPM) and sodium ethoxide in ethanol has been shown in previous work to be a proton-transfer : ki kb (NOZC&4)3CH + OEt-+[(NOzC6H4)3C]- + EtOH. (1) The carbanion is deeply coloured and the kinetics of the reaction were determined spectrophotometrically. The Arrhenius plot was linear over the range +20 to -8O"C, showing no sign of the curvature which might be expected if there were an appreciable tunnel effect. It was later suggested by Lewis that the effects of tunnel- ling will be increased by steric hindrance, because the contribution of the short-range repulsive forces to the energy-barrier will increase its steepness, or more specifically its curvature at the top, which controls the tunnelling factor.The TNPM molecule is a bulky propeller-like structure and the approach to it of the large alkoxide ion might well be sterically hindered. It therefore seemed desirable to investigate the kinetics of similar reactions using more bulky alkoxides as bases, in the corresponding alcohols as solvents. Since TNPM dissolves with difficulty in the pure alcohols, the solvents also contain toluene (as in the earlier work I). t-Butoxide anion was the most obvious choice as a sterically hindered base, but we encountered difficulties in obtaining reproducible results with this system over a range of temperatures.For this reason the present studies have used ethanol, ethanC2H]ol and isopropyl alcohol as the solvents, with the lyate ions acting as the base. Less reliable data are reported for the t-butoxide reaction at 25°C. -f present address : The Wellcome Research Laboratories, Langley Court, Beckenham, Kent. $..permanent address : Department of Chemistry, University of Manitoba, Winnipeg, Canada. Visiting Research Fellow 1972. 528E. F . CALDIN, E. DAWSON, R . M. HYDE AND A. QUEEN 529 EXPERIMENTAL MATERIALS Tri-(4-nitrophenyl)methane (TNPM) was prepared by the method of Montagne (m.p. 214"C, lit.3 214-215°C). Dioxan (G.P.R.) was purified by shaking with stannous chloride to remove peroxides and refluxing with aqueous HCl for 9 h under nitrogen to remove acetaldehyde and ethylene aced, dried with KOH, refluxed with sodium, distilled, fraction- ally frozen, and kept under dry nitrogen (m.p.1 1.70-1 1.75"C, lit. 1 1.80"C). Chloroform was washed with water, dried and distilled immediately before use. Ethanol was obtained from commercial absolute alcohol by the method of Smith ; the water content (Karl Fischer) was less than 0.01 % (w/v). Ethan[2H]ol (Merck, Sharpe and Dohme, 99.5 % anhydrous) was similarly purified. t-Butyl alcohol was refluxed with sodium under nitrogen, distilled and fractionally frozen (m.p. 24.4"C, lit. 24.4"C). After decanting the liquid remaining when about 75 % of the total had crystallized, the remainder was distilled from calcium hydride under nitrogen and redistilled from sodium t-butoxide and dibutyl phthalate (b.p.82.2"C, lit. 82.2"C). It was stored over molecular sieves (Linde 3A) in a nitrogen atmosphere and used within 24 h. Isopropyl alcohol was similarly purified and stored but without a frac- tional crystallization. The final distillation was carried out after refluxing with phthalic anhydride and adding a solution of sodium isopropoxide in the dry alcohol (b.p. 82.2"C7 lit. 82.4"C). The water content (Karl Fischer) was less than 0.01 % (wlv). Toluene (B.D.H. " sulphur-free ") was refluxed with sodium under dry nitrogea and distilled (b.p. 110.2"C, lit. 110.6"C). It was stored over molecular sieves. DEUTERATION OF TNPM This was carried out as described by Caldin and T~rnalin.~ The expected isotopic purity of the product was calculated as 99.7 %; the actual isotopic purity was at least 99 % as shown by the fact that the n.m.r.spectrum in [2H] chloroform showed no trace of the peak at 4.14 p.p.m. characteristic of the acidic H in TNPM. PREPARATION OF SOLUTIONS All solvents were degassed on a vacuum line immediately before use. Solutions were prepared under dry nitrogen by a syringe technique, and protected from the atmosphere by soda-lime guard tubes. Absorption of atmospheric carbon dioxide was thus minimized ; the maximum uncertainty in the lowest base concentrations is no more than 1 % and in general it is much less. ABSORPTION SPECTRA A solution of TNPM+sodium ethoxide in toluene+ethanol (15 : 85, vlv) was mauve and showed a broad absorbance peak with a maximum at 570nm.The corresponding solution containing isopropoxide and isopropyl alcohol showed a similar peak at 565 nm, and the t-butoxide system one at 560nm. The mauve solutions faded gradually at room temperature, but the changes were shown to be negligible over the few seconds required for a kinetic run. In all cases, the " infinity " traces were reproducible after at least 50 half-lives. NATURE OF THE PRODUCT Evidence that the reaction in ethoxide+ ethanol + toluene is a proton transfer, producing a coloured tri-(4-nitrophenyl)methide carbanion, was given in the earlier paper and the large isotope effects found in the present workconfkm this view. That the coloured products in the isopropoxide and t-butoxide systems are the same as in the ethoxide system is indicated by the observation that the values of Amax and the extinction coefficients are nearly equal in all cases.530 KINETIC ISOTOPE EFFECTS KINETIC MEASUREMENTS All kinetic measurements were made by means of a stopped-flow apparatus with spectro- photometric detection which has been recently described in detaiL6 Temperatures were controlled to within +O.Ol"C or better and measured by means of a Hewlett-Packard quartz thermometer or a platinum resistance thermometer attached to a Leeds and Northrup Mueller bridge.Changes in transmittance were kept below 4 % so that the oscilloscope deflection could be assumed to be proportional to the change of concentration within f 3 %, at the most.This is acceptable, especially since the measurements did not cover the entire change of optical density. First-order behaviour was ensured by keeping the base in large excess over the TNPM. First-order rate constants for reactions with half-lives ranging from a few seconds to about 20milliseconds were measured with an overall reproduc- ibility of 1-2 %. The values for the ethoxide reactions were determined by the curve-fitting method of Crooks, Tregloan and Zetter from photographs of the traces obtained on a Tektronix 564 storage oscilloscope. The values for the other systems were integrated values obtained using a Biomation transient recorder (type 610-B) coupled directly to a Wang 600 programmable calculator through a specially-built interface; the rate constants were available within a few seconds of carrying out the kinetic run.The values given in table 3 are the means of 100-250 individual rate constants spread over at least three half-lives in 6-10 separate kinetic runs. The rate traces were accurately exponen- tial for all systems, except that in which deutero-TNPM reacted with ethoxide in ethanol+ toluene. For this case, although the last 80 % or so of the rate traces were good exponential curves, the initial 15-20 % portions were steeper than required for a good fit. The steep portions of some of the curves were isolated by expanding the vertical gain and the time-base of the oscilloscope, and the rates were found to be roughly those associated with the proton transfer. Since similar effects were not observed with deutero-TNPM in systems containing ethan12H]ol, isopropyl alcohol, or t-butyl alcohol, and since no proton n.m.r. signal was observed for deutero-TNPM in deuterochloroform, the problem was not associated with incomplete deuteration of the substrate.We can only assume that a slow exchange between ethanol and deutero-TNPM was occurring and that this effect was negligible for the other alcohols, possible because of greater steric hindrance. The corresponding rates for the protonated and deuterated TNPM with ethoxide ions are sufficiently different to allow good rate constants for kFbS to be obtained by using only the last 80 % of the reaction curves, and this procedure was used. RESULTS In all cases the observed rate constants (itabs) increased linearly with the base concentrations.This result is consistent with a reaction of the type shown in eqn (l), provided the base is present in large excess over the TNPM. The dependence of rate on base concentration is then given by : TABLE 1 .-RATE CONSTANTS FOR THE REACTION OF TRb(4-NITROPHENYL)METHANE AND DEUTEROTRI-(4-NITROPHENYL) METHANE WITH SODIUM ETHOXIDE IN ETHANOL+TOLUENE (85 : 15 v/v) 10*1Et0-1/ temp./"C mol dm-3 k&/s- kp/dm3 mol-' s-1 0.00 0.429 0.336 0.286 8.95 2.598 0.49, fO.011 f0.20 3.926 0.641 6.575 0.880 10.11 1.195.00 10.00 15.00 20.00 25.00 30.00 35.00 40.00 45 .00 E. F. CALDIN, E. DAWSON, R . M. HYDE AND A. QUEEN TABLE 1 .-continued 0.429 0.470 2.598 0.746 3.926 0.943 6.575 1.32 10.11 1.80 0.429 0.698 2.598 1.10 3.926 1.43 6.575 1-96 10.11 2.65 2.958 lh2 3.926 2.06 6.575 2.g3 10.11 3.89 0.429 0.998 0.429 1.37 2.958 2.31 3.926 2-96 6.575 4.12 10.11 5.64 0.429 1.95 2.598 3.36 3.926 4.18 6.575 5.88 10.11 8.09 0.429 2.68 2.598 4.54 3.926 6.575 0.429 2.598 3.926 6.575 0.429 2.598 10.1 1 10.1 1 5.95 8.42 1.8 3.69 6.43 8.39 1.8 6.0 8.74 4-86 3.926 11.5 6.575 16.1 10.11 21.9 0.429 6.51 2.598 11.9 3.926 15.5 6.575 21.5 10.11 29.6 0.40, * 0.009 0.606 f 0.021 0.864 * 0.01 1.19 & 0.03 1.70 & 0.03 2.Z7 - + 0.0s 3.22 +0.14 4.28 & 0.21 5.76 & 0.24 13.9 - +0.15 20.fO.37 29. g k0.2 44.2 k 0.4 63.4 k0.3 91.9 k 1.4 128 - +2 176 - +4 238 f 4 1.548 4.87 3.174 8.94 6.450 15.8 10.14 24.2 1.548 7.45 3.174 13.8 6.450 24.2 10.14 36.7 1.548 ll.7 3.174 21.1 6.450 36.6 10.14 57.5 1.548 N2 3.174 31.8 6.450 55.3 10.14 85.3 1.548 27.5 3.174 47.8 6.450 85.5 10.14 126 1.548 40.2 3.174 70.0 10.14 186 6.450 lZ4 1.548 59.4 3.174 lo3 6.450 177 10.14 270 1.548 89.3 3.174 146 6.450 26, 10.14 377 1.59 f 0 .5 2.61 k 0.43 3.69 * 0.90 7.13 & 0.70 10. * k 1.1 15.1 f 1.1 23.1 k 2.8 39.5 - +4.1 53 1 2.Z3 - + 0.0, 3.37 & 0.0, 5-26 kO.14 7.51 kO.11 ll.4 k0.2 16.9 f 0.2 24.3 Ifr 0.4 33.6 - +0.7 Initial concentration of substrate = (5.450.3) x kf;Ibs and krbs are the observed first-order rate constants for TPNM and ['HITNPM respectively. Standard errors in k%s and kpbs are approximately &l% or better: errors shown are standard deviations. mol dm-3.532 KINETIC ISOTOPE EFFECTS TABLE Z-RATE CONSTANTS FOR THE REACTION OF DEUTEROTRI-(4-NITROPHENYL)METHANE WITH SODIUM ETHOXIDE IN DEUTEROETHANOL~TOLUENE (85 : 15 v/v) temp./"C 4.69 9.63 14.51 19.57 24.57 30.49 34.76 39.90 44.92 102[Et0-1/ mol dm-3 2.120 5.450 8.450 10.73 2.120 5.450 8.450 10.73 2.120 5.450 8.450 10.73 2.120 5.450 8.450 10.73 2.120 5.450 8.450 10.73 2.120 5.450 8.450 10.73 2.120 5.450 8.450 10.73 2.120 5.450 8.450 10.73 2.120 5.450 8.450 10.73 k,D,,ls-' 0.136 0.246 0.359 0.43 5 0.20, 0.374 0.551 0.672 0.322 0.565 0.842 1.01 0.465 0.870 1.25 1.52 0.695 1.31 1.82 2.22 1.02 1 .go 2.6, 3.28 1.54 2.77 3.81 4.72 2.23 4.1 1 5.80 7.01 3.24 5.67 8.00 9.80 kg 15-1 0.059 Ik 0.004 0.085 f 0.008 0.142 f 0.020 0.204 +0.012 0.331 f 0.01 0.467 foe056 0.759 f 0.045 1 .07 f 0.04 1.55 - + 0.0, 5.47 f0.12 8.12 - + 0 . 2 ~ 12.3 k 0.2 17.9 k 0.2 25.9 f0.7 36.6 - +0.6 55.7 f0.5 76.2 f 1.0 Initial concentration of substrate (5.3 f0.3) x mol dm-3 : errors are standard deviations.E .F. CALDIN, E. DAWSON, R . M . HYDE AND A . QUEEN 533 A plot of kobs against b therefore gives kb as the intercept and k, as the slope. Tables 1-4 list values of kobs and the derived values of kb and kf, which were obtained by cairying out a least-squares fit of the data to eqn (2) with the help of a computer pro- gram. The values of the rate constants at 25°C for H+-transfer and Df-transfer in both the forward and backward directions are compared in table 5, where the activa- tion parameters are also shown. The present data for the reaction of TNPM with ethoxide ion in ethanol+ toluene are in good agreement with the previous results of Caldin and Trickett which covered the temperature range - 80 to - 10°C ; an Arrhenius plot showing both sets of data is linear within the limits of the experimental errors.Since, however, the experimental errors were smaller in the present studies, we have used only these results in our calculations. The results in t-butyl alcohol+ toluene are subject to some uncertainty, because the rate constants appear to be sensi- tive to the presence of water, and it is difficult to be sure that residual moisture had no effect, though the results were reproducible when a given batch of solvent was used for all experiments. TABLE 3 .-RATE CONSTANTS FOR THE REACTION OF TRI-(4-NITROPHENYL)METHANE AND DEUTEROTRI-(4-NITROPHENYL)METHANE WITH SODIUM ISOPROPOXIDE IN PROPAN-ZOL + TOLUENE (85 : 15 vlv) 1O2[Pri0-1/ temp./"C mol dm-3 kZs/s-l 0.39 1.634 2.140 3.218 4.678 5.29 1.634 2.140 3.218 4.678 10.01 1.634 2.140 0.21 8 4.678 6.135 15.16 1.634 2.140 3.21 8 4.678 6.135 20.04 1.634 2.140 3.218 4.678 5.135 kF/s-I 1.57 0.60, 1.85 f0.032 2.53 3.35 2.01 0.866 2.51 k0.057 3.27 4.31 5.44 2.79 l.17 3.30 kO.06 4.29 5.87 7.20 3.52 1.51 4.17 fO.11 5.57 7.51 9.10 45.8 f 1.5 58.9 - + 1.0 74.0 f 1.5 98.8 f 1.6 125 - +3 1 Ozj?rQ-l/ temp./"C mol dm-3 kFDB/s-' 9.93 2.036 0.234 0.091, 3.772 0.351 +0.0084 4.308 0.393 6.050 0.519 6.135 0.531 7.625 0.617 15.21 2.036 0.327 0.138 3.772 0.505 +0.013 4.308 0.552 6.050 0.719 6.135 0.744 7.625 0.859 20.04 2.036 0.458 0.189 3.772 0.685 +O.OO, 4.308 0.750 6.050 0.983 6.135 0.996 7.625 1.19 kp/dm3 mol-1 s-1 7.00 *O.lfj 9.62 f O.z4 13.1 k0.1534 KINETIC ISOTOPE EFFECTS lOZ[prlO-l/ temp.I0C mol dm-1 k&&-l kE1s-I 25.07 1.634 4.64 2.16 2.140 5.42 fO.09 3.218 7.13 4.678 9.46 6.135 11.5 30.02 1.634 5.77 2.46 2.140 6.90 k0.12 3.218 9.24 4.678 lZ3 6.135 15.1 34.98 1.634 7.35 3.41 2.140 8.90 kO.14 3.218 11.5 4.678 15.2 6.135 18.7 40.03 1.634 9.17 4.41 2.140 11.3 f0.27 3.218 14.6 4.678 19.00 6.135 23.4 45.07 1.634 ll.4 5.09 2.140 13.0 +0.17 3.218 - 4.678 23.5 6.135 29.3 TABLE 3.-continued lO*lprlO-]/ temp./"C mol dm-3 kF-ls-1 25.07 1.880 0.590 2.036 O.6z2 3.012 0.791 3.772 0.911 4.308 1.01 4.524 1.06 6.050 1.32 7.625 1.58 30.02 2.036 0.828 3.378 l.14 4.308 1.35 6.050 1.74 6.135 1.78 7.625 2.09 34.99 2.036 3.378 4.308 6.050 6.135 7.625 40.05 2.036 3.378 4.308 6.050 6.135 7.625 45.01 2.036 3.378 4.308 6.050 6.135 7.625 1 .06 2.28 1-48 1.75 2.3 1 2.73 1.41 1.95 2.29 3.02 2-98 3-56 1.82 2.48 2.91 3.74 3.80 4.50 *Is-' 0.268 +0.006 0.372 f 0.01 2 0.46 - + 0.01 3 0.635 0.01 8 0.851 + 0.01 0 mol-1 s-1 17.3 - +O., 22.7 f0.2 30.0 - + 0.2 38.6 + 0.3 47.9 f0.2 Initial concentration of substrate = (4.0f0.4) x mol dm-3: errors are standard deviations, TABLE 4.-hTE CONSTANTS FOR THE REACTION OF TRI-(4-NITROPHENYL)METHANE AND DEUTEROTRI-(h'ITROPHENn)METHANE WITH SODIUM f-BUTOXIDE IN t-BUTYL ALCOHOL+ TOLUENE (85 : 15 v/v) AT 25.0+ 0.05"C *Is-' lo*lButo-]/ mol d ~ n - ~ kFb/s-l kg Is- kf/dm3 mol-1 s-1 1.13 l.27 0.33 77.8 0.57 0.20~ 0.18 3.83 1.74 1.62 2.07 0.252 2.33 2.02 3.52 0.315 2.91 2.74 5.56 0.395 3.60 3.10 1.16 1.24 fO.09 f2.2 1.13 0.21~ f 0 .0 ~ +0.i8 Errors are standard deviations.rn TABLE 5.-ARRHENIUS PARAMETERS AND ISOTOPE EFFECTS FOR THE REACTION OF TRI-(4-NITROPHENYL)METHANE WITH ALKOXIDE IONS IN ALCOHOL + crj TOLUENE SOLUTIONS (85 : 15 v/v) system I system I1 system I11 (ethanol+ toluene+EtO-) (ethan[2H]ol+ toluene+ EtO-) (propan-2-olf toluene + PriO-) (t-butanol+ toluene+ ButO-) forward backward forward backward forward backward forward backward - ------I_- ---- - k H 63.4f 0.3 1.70f 0.03 154+ 2 2.1 6f0.09 9O.4+3 0.28+0.06 kD 7.5k0.1 I 0.071 +O.O07 18.5k0.2 0.34k0.02 17.3f0.1 0.27f0.006 3.83k0.2 O.l8+O.O4 kH/kD 8.4k0.2 24f3 8.9f0.2 8.1k0.5 23+2 1.6f 0.08 12.60+ 0.06 11.46fo.i 5 8.4of0.09 9.53 & 0.23 4% 13.93+ 0.10 16.6k0.6 13.47k0.06 13.85k0.44 9.74fO.09 11.1 6fo.15 log AZs 1 1 .04+ 0.04 8.63 f 0.11 8.3 6& 0.07 7.29f 0.1 7 log AFbs 11 .I 0 - t 0.07 11 .O2 fsO.41 11.15+O.O5 9.7f0.3 8.38k0.06 7.60+0.10 AH$ l2.O1+O.O6 10.87+0.15 7.80+0.09 8.94fO.23 -AS$ 10.0+0.2 21.0k0.5 23 -3 f 0.3 27.2f0.8 -AS$ 9.7+ 0.3 10.1 & 1.6 9.5k0.2 16.1+ 1.5 22.2+O.3 25.7 0.5 AH; 1 .14f O.& - - 1.1f0.3 - 1.4+ 0.2 AH; -2.71 k0.69 - 0.4+ 0.5 EoHbS Ezs- E$s 1 3 3 f 0.1 6 5.2fO.74 1.54+0.18 1 s6k0.38 log Azs/AFbs 0.06+0.11 2.4f 0.5 0.02f0.13 0.31 40.27 AH$ 13.34+ 0.10 16.O5+OS9 12.88+0.06 13.25+ 0.44 9.1 5+ 0.09 10.57kO.15 AS; 1 1 .O+ 0.7 - 5+ 1 As; 0.4f 1.9 7+2 4f 1 kf and Af at 25°C in dm3 mol-' s-l; k b and Ab at 25°C in s-l; E and AH in kcal mol-I; AS in cal deg-l mol-'; 1 cal = 4.184 J; errors are standard deviations.0 Z w536 KINETIC ISOTOPE EFFECTS DISCUSSION We refer to the systems as I, I1 and 111 as shown in table 5.Let us first focus attention only on the forward reactions (I&). For system I, the results show no marked abnormalities. The values of kH/kD, of - EFbs) and of kfFbs/kfrbs might be explained on the basis of classical transition-state theory, but they are larger than usual and a small tunnelling correction is not excluded. The individual values of the rate constants and activation parameters are comparable with those previously determined for similar reactions of other carbon acids. The negative entropy of activation is consistent with the bimolecular process, and its size suggests that only a small change of solvation occurs on. passage into the transition state. This conclusion is supported by the fact that the entropy of activation is very nearly the same for the corresponding reaction of deutero-TNPM in.ethanol and in ethan[2H]ol. The positive values of ASo show that there is a decrease of solvation on forming products, presumably because of the delocalisation of the charge on the anion. The solvent isotope effect on kD, i.e. kD(EtOH)/kD(EtOD), at 25°C is 0.39 for the forward reaction and 0.21 for the backward reaction. For system 11, the large negative entropy of activation suggests considerable solvation of the transition state by isopropyl alcohol, and this is supported by the smaller activation energy compared with system I. Such compensatory changes in AH * and AS* are frequently observed for processes controlled by solvation. Similar trends would also be observed if tunnelling is more important in system I1 than in system I.The isotope effect is hardly changed from the value found for ethoxide ion as base, but the value of (Erbs -EF& 1.54k0.2 kcal mol-l, is large enough to suggest that tunnelling is becoming a significant factor in the kinetics. The maximum value of this quantity that can be simply explained by the classical transition state theory * is 1.15 kcal mol-I. The view that tunnelling increases with increasing bulk of the base is supported by the very large isotope effect we have found when t-butoxide is the base (system 111). Unfortunately, as we have already indicated, there is some doubt as to the accuracy of this result. However, even if the derived values of kH and kD for the forward reaction are in error by as much as 25 %, which seems unlikely, the isotope effect is still large and indicative of tunnelling.When the backward reactions are considered, the present results lead to a number of problems in analysing the data in terms of eqn (1) for proton transfer and eqn (3) and (4) for deuteron transfer : 3 4 SD + B-+S-+ DB HB 11 fast 1 SH + B-+S- + HB. (3) (4) The first is that the observed values of k: and k t are different. Equilibration of deuterium with the solvent would be rapid, so that step 4 should be negligible; provided that step 1 is rapid compared to steps 2 and 4, the reverse reaction should correspond to step 2, so that kF and k? should be identical. If this condition does not apply, it would be expected that, with the deutero-compound, exponential rate curves would not be observed and that the rate would increase during the course of the reaction as the concentration of SH increases.However, although the rate curves for the transfer of deuterium to ethoxide in ethanol +toluene are exponential over onlyE. F . CALDIN, E . DAWSON, R . M . HYDE AND A . QUEEN 537 the last 80-85 % of the reaction, the initial parts of the curve are steep, so that an apparent rate decrease is observed over the first 15-20 %. The large values of kH/kD for the reverse reactions are, therefore, not in accord with eqn (I), (3) and (4). We suggest that the product which is observed in the stopped-flow experiments is the species (S-----H-B) in which the proton is strongly hydrogen-bonded to the anion. As a consequence, equilibration of this proton with the hydroxyl protons of the solvent is slow.This suggestion is in line with recent proposals by Streitwieser concerning the nature of the ionic species formed in reactions between triarylmethanes and caesium in cyclohexylamine. Kollmeyer and Cram lo have reached simiIar con- clusions from their studies of alkoxide-catalysed proton-deuteron exchange reactions of diphenyl- and triphenyl-methanes. The reaction scheme should therefore be written : k f kb SH + B-+(S-----.H-B). Acceptance of this scheme still leaves one problem to be solved. The large values of k?/k?, (E?bs-E:bs) and &bs/A:I,s for the reverse reaction with ethoxide as base all point to a large tunnelling factor. It is difficult to explain these values in other term.Although it is true that the rate constants kb are small, and probably subject to greater errors than the standard deviations indicate, nevertheless any reasonable allowance for this, which must take into account the self-consistency of the data and the agree- ment with the earlier low-temperature values of Caldin and Trickett, would still leave large differences between the proton and deuteron transfer reactions. It is difficult to rationalize a large tunnelling factor in one direction in a reversible process with a small effect in the other. A comparison of the data for ethanol + toluene as solvent with those for etha~~[~H]ol+ toluene as solvent does not help in supplying an answer to this problem; besides solvent effects on the equilibrium, it would be necessary to allow for the fact that ethoxide is a stronger base in deuteroethanol than ethanol while the deuteroalcohol is the stronger acid.Moreover, the reversal of the situation as the solvent and base are changed in systems I, I1 and I11 suggests that solvent reorganization must play an important part in determining the observed rates and activation parameters. Another instance of a solvent-dependent isotope effect has recently been reported.ll " Solvent lag " has been previously suggested l 2 as a neglected factor in fast reactions and would necessarily alter isotope effects. For example in a reaction where solvent molecules are more strongly held in the transition state than in the initial state, k,/k, would be greater than if equilibrium were main- tained.For the present reaction, -AS$ > -AS%, indicating (in the absence of tunnelling) that solvent reorganization is greater for the reverse reaction than for the forward one, and any solvent lag would lead to a larger isotope effect for this step, as observed. However, solvent lag should also lead to an increase in AS* (or A ) and the effect should be larger for H than for D. Consequently, &bSb,l&bs should be less than unity if tunnelling is excluded, which is not the case. Clearly, more data are requiIed before the problem can be resolved. The authors gratefully acknowledge grants in support of this work by the S.R.C. (to R. M. H.), the National Research Council of Canada and the Research Board of the University of Manitoba and helpful discussions with Professors R.P. Bell and J. R. Reeffe. E. F. Caldin and J. C. Trickett, Trans. Faraday SOC., 1953, 49, 772. E. S. Lewis and L. H. Funderburk, J. Amer. Clzem. SOC., 1967, 89, 2322. Montagne, Rec. Trav. chim., 1905, 24, 105.538 KINETIC ISOTOPE EFFECTS E. L. Smith, J. Chenz. SOC., 1927, 1284. E. F. Caldin and G. Tomalin, Trans. Faraduy Soc., 1968, 64, 2814. E. F. Caldin, J. E. Crooks and A. Queen, J. P/i.vs. E, 1973, 6, 930. ’ J. E. Crooks, M. S. Zetter and P. A. Tregloan, J. Phys. E, 1970, 3, 73. R. P. Bell, The Proton in Chemistry (Chapman and Hall, London, 2nd edn., 1973), chap. 12 ; M. E. Schneider and M. J. Stern, J. Amer. Chem. SOC., 1972, 94, 1517. A. Streitwieser, J. R. Murdoch, G. Hafelinger and C. J. Chang, J. Amer. Chem. SOC., 1973,95, 4248.lo D. J. Cram and W. D. Kollmeyer, J. Amer. Chem. SOC., 1968,90, 1791. l1 E. F. Caldin and S. Mateo, Chem. Comm., 1973, 854. l2 R. P. Bell, Disc. Faraduy Soc., 1965, 39, 16 ; M. M. Kreevoy and R. A. Kretchmer, J. Amer. Chem. SOC., 1964,86,2435. Kinetic Isotope Effects in Some Reactions of Tri-(4-nitrophenyl)- methane with Alkoxide Bases in Various Alcoholic Media B Y EDWARD F. CALDIN,* ELEANOR DAWSON, RICHARD M. HYDE? AND ALAN QUEEN: University Chemical Laboratory, University of Kent at Canterbury, Canterbury, Kent Received 30th May, 1974 Rate constants and Arrhenius parameters have been determined for proton-transfer and deuteron- transfer reactions of tri-(4-nitrophenylmethane) with alkoxide ions in alcohol + toluene mixtures. The kinetic isotope effects (kH/kD) at 25°C are considerable (8.4 in ethanol+ toluene, 8.9 in propan-2-01 + toluene, 23 in t-butanol + toluene) and indicate that the rate-limiting step is a proton transfer.For the forward reactions, the results suggest that the importance of tunnelling increases with the size of the alkoxide group. For the backward reactions, the rates for proton- and deuteron-transfer are different, although they would be expected to be equal by reason of rapid exchange with the solvent ; this suggests that in the product the proton remains firmly bonded to the carbanion. The reaction (1) between tri-(4-nitrophenylmethane) (TNPM) and sodium ethoxide in ethanol has been shown in previous work to be a proton-transfer : ki kb (NOZC&4)3CH + OEt-+[(NOzC6H4)3C]- + EtOH. (1) The carbanion is deeply coloured and the kinetics of the reaction were determined spectrophotometrically.The Arrhenius plot was linear over the range +20 to -8O"C, showing no sign of the curvature which might be expected if there were an appreciable tunnel effect. It was later suggested by Lewis that the effects of tunnel- ling will be increased by steric hindrance, because the contribution of the short-range repulsive forces to the energy-barrier will increase its steepness, or more specifically its curvature at the top, which controls the tunnelling factor. The TNPM molecule is a bulky propeller-like structure and the approach to it of the large alkoxide ion might well be sterically hindered. It therefore seemed desirable to investigate the kinetics of similar reactions using more bulky alkoxides as bases, in the corresponding alcohols as solvents.Since TNPM dissolves with difficulty in the pure alcohols, the solvents also contain toluene (as in the earlier work I). t-Butoxide anion was the most obvious choice as a sterically hindered base, but we encountered difficulties in obtaining reproducible results with this system over a range of temperatures. For this reason the present studies have used ethanol, ethanC2H]ol and isopropyl alcohol as the solvents, with the lyate ions acting as the base. Less reliable data are reported for the t-butoxide reaction at 25°C. -f present address : The Wellcome Research Laboratories, Langley Court, Beckenham, Kent. $..permanent address : Department of Chemistry, University of Manitoba, Winnipeg, Canada.Visiting Research Fellow 1972. 528E. F . CALDIN, E. DAWSON, R . M. HYDE AND A. QUEEN 529 EXPERIMENTAL MATERIALS Tri-(4-nitrophenyl)methane (TNPM) was prepared by the method of Montagne (m.p. 214"C, lit.3 214-215°C). Dioxan (G.P.R.) was purified by shaking with stannous chloride to remove peroxides and refluxing with aqueous HCl for 9 h under nitrogen to remove acetaldehyde and ethylene aced, dried with KOH, refluxed with sodium, distilled, fraction- ally frozen, and kept under dry nitrogen (m.p. 1 1.70-1 1.75"C, lit. 1 1.80"C). Chloroform was washed with water, dried and distilled immediately before use. Ethanol was obtained from commercial absolute alcohol by the method of Smith ; the water content (Karl Fischer) was less than 0.01 % (w/v).Ethan[2H]ol (Merck, Sharpe and Dohme, 99.5 % anhydrous) was similarly purified. t-Butyl alcohol was refluxed with sodium under nitrogen, distilled and fractionally frozen (m.p. 24.4"C, lit. 24.4"C). After decanting the liquid remaining when about 75 % of the total had crystallized, the remainder was distilled from calcium hydride under nitrogen and redistilled from sodium t-butoxide and dibutyl phthalate (b.p. 82.2"C, lit. 82.2"C). It was stored over molecular sieves (Linde 3A) in a nitrogen atmosphere and used within 24 h. Isopropyl alcohol was similarly purified and stored but without a frac- tional crystallization. The final distillation was carried out after refluxing with phthalic anhydride and adding a solution of sodium isopropoxide in the dry alcohol (b.p.82.2"C7 lit. 82.4"C). The water content (Karl Fischer) was less than 0.01 % (wlv). Toluene (B.D.H. " sulphur-free ") was refluxed with sodium under dry nitrogea and distilled (b.p. 110.2"C, lit. 110.6"C). It was stored over molecular sieves. DEUTERATION OF TNPM This was carried out as described by Caldin and T~rnalin.~ The expected isotopic purity of the product was calculated as 99.7 %; the actual isotopic purity was at least 99 % as shown by the fact that the n.m.r. spectrum in [2H] chloroform showed no trace of the peak at 4.14 p.p.m. characteristic of the acidic H in TNPM. PREPARATION OF SOLUTIONS All solvents were degassed on a vacuum line immediately before use. Solutions were prepared under dry nitrogen by a syringe technique, and protected from the atmosphere by soda-lime guard tubes.Absorption of atmospheric carbon dioxide was thus minimized ; the maximum uncertainty in the lowest base concentrations is no more than 1 % and in general it is much less. ABSORPTION SPECTRA A solution of TNPM+sodium ethoxide in toluene+ethanol (15 : 85, vlv) was mauve and showed a broad absorbance peak with a maximum at 570nm. The corresponding solution containing isopropoxide and isopropyl alcohol showed a similar peak at 565 nm, and the t-butoxide system one at 560nm. The mauve solutions faded gradually at room temperature, but the changes were shown to be negligible over the few seconds required for a kinetic run. In all cases, the " infinity " traces were reproducible after at least 50 half-lives. NATURE OF THE PRODUCT Evidence that the reaction in ethoxide+ ethanol + toluene is a proton transfer, producing a coloured tri-(4-nitrophenyl)methide carbanion, was given in the earlier paper and the large isotope effects found in the present workconfkm this view.That the coloured products in the isopropoxide and t-butoxide systems are the same as in the ethoxide system is indicated by the observation that the values of Amax and the extinction coefficients are nearly equal in all cases.530 KINETIC ISOTOPE EFFECTS KINETIC MEASUREMENTS All kinetic measurements were made by means of a stopped-flow apparatus with spectro- photometric detection which has been recently described in detaiL6 Temperatures were controlled to within +O.Ol"C or better and measured by means of a Hewlett-Packard quartz thermometer or a platinum resistance thermometer attached to a Leeds and Northrup Mueller bridge.Changes in transmittance were kept below 4 % so that the oscilloscope deflection could be assumed to be proportional to the change of concentration within f 3 %, at the most. This is acceptable, especially since the measurements did not cover the entire change of optical density. First-order behaviour was ensured by keeping the base in large excess over the TNPM. First-order rate constants for reactions with half-lives ranging from a few seconds to about 20milliseconds were measured with an overall reproduc- ibility of 1-2 %. The values for the ethoxide reactions were determined by the curve-fitting method of Crooks, Tregloan and Zetter from photographs of the traces obtained on a Tektronix 564 storage oscilloscope. The values for the other systems were integrated values obtained using a Biomation transient recorder (type 610-B) coupled directly to a Wang 600 programmable calculator through a specially-built interface; the rate constants were available within a few seconds of carrying out the kinetic run.The values given in table 3 are the means of 100-250 individual rate constants spread over at least three half-lives in 6-10 separate kinetic runs. The rate traces were accurately exponen- tial for all systems, except that in which deutero-TNPM reacted with ethoxide in ethanol+ toluene. For this case, although the last 80 % or so of the rate traces were good exponential curves, the initial 15-20 % portions were steeper than required for a good fit.The steep portions of some of the curves were isolated by expanding the vertical gain and the time-base of the oscilloscope, and the rates were found to be roughly those associated with the proton transfer. Since similar effects were not observed with deutero-TNPM in systems containing ethan12H]ol, isopropyl alcohol, or t-butyl alcohol, and since no proton n.m.r. signal was observed for deutero-TNPM in deuterochloroform, the problem was not associated with incomplete deuteration of the substrate. We can only assume that a slow exchange between ethanol and deutero-TNPM was occurring and that this effect was negligible for the other alcohols, possible because of greater steric hindrance. The corresponding rates for the protonated and deuterated TNPM with ethoxide ions are sufficiently different to allow good rate constants for kFbS to be obtained by using only the last 80 % of the reaction curves, and this procedure was used.RESULTS In all cases the observed rate constants (itabs) increased linearly with the base concentrations. This result is consistent with a reaction of the type shown in eqn (l), provided the base is present in large excess over the TNPM. The dependence of rate on base concentration is then given by : TABLE 1 .-RATE CONSTANTS FOR THE REACTION OF TRb(4-NITROPHENYL)METHANE AND DEUTEROTRI-(4-NITROPHENYL) METHANE WITH SODIUM ETHOXIDE IN ETHANOL+TOLUENE (85 : 15 v/v) 10*1Et0-1/ temp./"C mol dm-3 k&/s- kp/dm3 mol-' s-1 0.00 0.429 0.336 0.286 8.95 2.598 0.49, fO.011 f0.20 3.926 0.641 6.575 0.880 10.11 1.195.00 10.00 15.00 20.00 25.00 30.00 35.00 40.00 45 .00 E.F. CALDIN, E. DAWSON, R . M. HYDE AND A. QUEEN TABLE 1 .-continued 0.429 0.470 2.598 0.746 3.926 0.943 6.575 1.32 10.11 1.80 0.429 0.698 2.598 1.10 3.926 1.43 6.575 1-96 10.11 2.65 2.958 lh2 3.926 2.06 6.575 2.g3 10.11 3.89 0.429 0.998 0.429 1.37 2.958 2.31 3.926 2-96 6.575 4.12 10.11 5.64 0.429 1.95 2.598 3.36 3.926 4.18 6.575 5.88 10.11 8.09 0.429 2.68 2.598 4.54 3.926 6.575 0.429 2.598 3.926 6.575 0.429 2.598 10.1 1 10.1 1 5.95 8.42 1.8 3.69 6.43 8.39 1.8 6.0 8.74 4-86 3.926 11.5 6.575 16.1 10.11 21.9 0.429 6.51 2.598 11.9 3.926 15.5 6.575 21.5 10.11 29.6 0.40, * 0.009 0.606 f 0.021 0.864 * 0.01 1.19 & 0.03 1.70 & 0.03 2.Z7 - + 0.0s 3.22 +0.14 4.28 & 0.21 5.76 & 0.24 13.9 - +0.15 20.fO.37 29. g k0.2 44.2 k 0.4 63.4 k0.3 91.9 k 1.4 128 - +2 176 - +4 238 f 4 1.548 4.87 3.174 8.94 6.450 15.8 10.14 24.2 1.548 7.45 3.174 13.8 6.450 24.2 10.14 36.7 1.548 ll.7 3.174 21.1 6.450 36.6 10.14 57.5 1.548 N2 3.174 31.8 6.450 55.3 10.14 85.3 1.548 27.5 3.174 47.8 6.450 85.5 10.14 126 1.548 40.2 3.174 70.0 10.14 186 6.450 lZ4 1.548 59.4 3.174 lo3 6.450 177 10.14 270 1.548 89.3 3.174 146 6.450 26, 10.14 377 1.59 f 0 . 5 2.61 k 0.43 3.69 * 0.90 7.13 & 0.70 10. * k 1.1 15.1 f 1.1 23.1 k 2.8 39.5 - +4.1 53 1 2.Z3 - + 0.0, 3.37 & 0.0, 5-26 kO.14 7.51 kO.11 ll.4 k0.2 16.9 f 0.2 24.3 Ifr 0.4 33.6 - +0.7 Initial concentration of substrate = (5.450.3) x kf;Ibs and krbs are the observed first-order rate constants for TPNM and ['HITNPM respectively.Standard errors in k%s and kpbs are approximately &l% or better: errors shown are standard deviations. mol dm-3.532 KINETIC ISOTOPE EFFECTS TABLE Z-RATE CONSTANTS FOR THE REACTION OF DEUTEROTRI-(4-NITROPHENYL)METHANE WITH SODIUM ETHOXIDE IN DEUTEROETHANOL~TOLUENE (85 : 15 v/v) temp./"C 4.69 9.63 14.51 19.57 24.57 30.49 34.76 39.90 44.92 102[Et0-1/ mol dm-3 2.120 5.450 8.450 10.73 2.120 5.450 8.450 10.73 2.120 5.450 8.450 10.73 2.120 5.450 8.450 10.73 2.120 5.450 8.450 10.73 2.120 5.450 8.450 10.73 2.120 5.450 8.450 10.73 2.120 5.450 8.450 10.73 2.120 5.450 8.450 10.73 k,D,,ls-' 0.136 0.246 0.359 0.43 5 0.20, 0.374 0.551 0.672 0.322 0.565 0.842 1.01 0.465 0.870 1.25 1.52 0.695 1.31 1.82 2.22 1.02 1 .go 2.6, 3.28 1.54 2.77 3.81 4.72 2.23 4.1 1 5.80 7.01 3.24 5.67 8.00 9.80 kg 15-1 0.059 Ik 0.004 0.085 f 0.008 0.142 f 0.020 0.204 +0.012 0.331 f 0.01 0.467 foe056 0.759 f 0.045 1 .07 f 0.04 1.55 - + 0.0, 5.47 f0.12 8.12 - + 0 .2 ~ 12.3 k 0.2 17.9 k 0.2 25.9 f0.7 36.6 - +0.6 55.7 f0.5 76.2 f 1.0 Initial concentration of substrate (5.3 f0.3) x mol dm-3 : errors are standard deviations.E . F. CALDIN, E. DAWSON, R . M . HYDE AND A . QUEEN 533 A plot of kobs against b therefore gives kb as the intercept and k, as the slope. Tables 1-4 list values of kobs and the derived values of kb and kf, which were obtained by cairying out a least-squares fit of the data to eqn (2) with the help of a computer pro- gram. The values of the rate constants at 25°C for H+-transfer and Df-transfer in both the forward and backward directions are compared in table 5, where the activa- tion parameters are also shown.The present data for the reaction of TNPM with ethoxide ion in ethanol+ toluene are in good agreement with the previous results of Caldin and Trickett which covered the temperature range - 80 to - 10°C ; an Arrhenius plot showing both sets of data is linear within the limits of the experimental errors. Since, however, the experimental errors were smaller in the present studies, we have used only these results in our calculations. The results in t-butyl alcohol+ toluene are subject to some uncertainty, because the rate constants appear to be sensi- tive to the presence of water, and it is difficult to be sure that residual moisture had no effect, though the results were reproducible when a given batch of solvent was used for all experiments.TABLE 3 .-RATE CONSTANTS FOR THE REACTION OF TRI-(4-NITROPHENYL)METHANE AND DEUTEROTRI-(4-NITROPHENYL)METHANE WITH SODIUM ISOPROPOXIDE IN PROPAN-ZOL + TOLUENE (85 : 15 vlv) 1O2[Pri0-1/ temp./"C mol dm-3 kZs/s-l 0.39 1.634 2.140 3.218 4.678 5.29 1.634 2.140 3.218 4.678 10.01 1.634 2.140 0.21 8 4.678 6.135 15.16 1.634 2.140 3.21 8 4.678 6.135 20.04 1.634 2.140 3.218 4.678 5.135 kF/s-I 1.57 0.60, 1.85 f0.032 2.53 3.35 2.01 0.866 2.51 k0.057 3.27 4.31 5.44 2.79 l.17 3.30 kO.06 4.29 5.87 7.20 3.52 1.51 4.17 fO.11 5.57 7.51 9.10 45.8 f 1.5 58.9 - + 1.0 74.0 f 1.5 98.8 f 1.6 125 - +3 1 Ozj?rQ-l/ temp./"C mol dm-3 kFDB/s-' 9.93 2.036 0.234 0.091, 3.772 0.351 +0.0084 4.308 0.393 6.050 0.519 6.135 0.531 7.625 0.617 15.21 2.036 0.327 0.138 3.772 0.505 +0.013 4.308 0.552 6.050 0.719 6.135 0.744 7.625 0.859 20.04 2.036 0.458 0.189 3.772 0.685 +O.OO, 4.308 0.750 6.050 0.983 6.135 0.996 7.625 1.19 kp/dm3 mol-1 s-1 7.00 *O.lfj 9.62 f O.z4 13.1 k0.1534 KINETIC ISOTOPE EFFECTS lOZ[prlO-l/ temp.I0C mol dm-1 k&&-l kE1s-I 25.07 1.634 4.64 2.16 2.140 5.42 fO.09 3.218 7.13 4.678 9.46 6.135 11.5 30.02 1.634 5.77 2.46 2.140 6.90 k0.12 3.218 9.24 4.678 lZ3 6.135 15.1 34.98 1.634 7.35 3.41 2.140 8.90 kO.14 3.218 11.5 4.678 15.2 6.135 18.7 40.03 1.634 9.17 4.41 2.140 11.3 f0.27 3.218 14.6 4.678 19.00 6.135 23.4 45.07 1.634 ll.4 5.09 2.140 13.0 +0.17 3.218 - 4.678 23.5 6.135 29.3 TABLE 3.-continued lO*lprlO-]/ temp./"C mol dm-3 kF-ls-1 25.07 1.880 0.590 2.036 O.6z2 3.012 0.791 3.772 0.911 4.308 1.01 4.524 1.06 6.050 1.32 7.625 1.58 30.02 2.036 0.828 3.378 l.14 4.308 1.35 6.050 1.74 6.135 1.78 7.625 2.09 34.99 2.036 3.378 4.308 6.050 6.135 7.625 40.05 2.036 3.378 4.308 6.050 6.135 7.625 45.01 2.036 3.378 4.308 6.050 6.135 7.625 1 .06 2.28 1-48 1.75 2.3 1 2.73 1.41 1.95 2.29 3.02 2-98 3-56 1.82 2.48 2.91 3.74 3.80 4.50 *Is-' 0.268 +0.006 0.372 f 0.01 2 0.46 - + 0.01 3 0.635 0.01 8 0.851 + 0.01 0 mol-1 s-1 17.3 - +O., 22.7 f0.2 30.0 - + 0.2 38.6 + 0.3 47.9 f0.2 Initial concentration of substrate = (4.0f0.4) x mol dm-3: errors are standard deviations, TABLE 4.-hTE CONSTANTS FOR THE REACTION OF TRI-(4-NITROPHENYL)METHANE AND DEUTEROTRI-(h'ITROPHENn)METHANE WITH SODIUM f-BUTOXIDE IN t-BUTYL ALCOHOL+ TOLUENE (85 : 15 v/v) AT 25.0+ 0.05"C *Is-' lo*lButo-]/ mol d ~ n - ~ kFb/s-l kg Is- kf/dm3 mol-1 s-1 1.13 l.27 0.33 77.8 0.57 0.20~ 0.18 3.83 1.74 1.62 2.07 0.252 2.33 2.02 3.52 0.315 2.91 2.74 5.56 0.395 3.60 3.10 1.16 1.24 fO.09 f2.2 1.13 0.21~ f 0 .0 ~ +0.i8 Errors are standard deviations.rn TABLE 5.-ARRHENIUS PARAMETERS AND ISOTOPE EFFECTS FOR THE REACTION OF TRI-(4-NITROPHENYL)METHANE WITH ALKOXIDE IONS IN ALCOHOL + crj TOLUENE SOLUTIONS (85 : 15 v/v) system I system I1 system I11 (ethanol+ toluene+EtO-) (ethan[2H]ol+ toluene+ EtO-) (propan-2-olf toluene + PriO-) (t-butanol+ toluene+ ButO-) forward backward forward backward forward backward forward backward - ------I_- ---- - k H 63.4f 0.3 1.70f 0.03 154+ 2 2.1 6f0.09 9O.4+3 0.28+0.06 kD 7.5k0.1 I 0.071 +O.O07 18.5k0.2 0.34k0.02 17.3f0.1 0.27f0.006 3.83k0.2 O.l8+O.O4 kH/kD 8.4k0.2 24f3 8.9f0.2 8.1k0.5 23+2 1.6f 0.08 12.60+ 0.06 11.46fo.i 5 8.4of0.09 9.53 & 0.23 4% 13.93+ 0.10 16.6k0.6 13.47k0.06 13.85k0.44 9.74fO.09 11.1 6fo.15 log AZs 1 1 .04+ 0.04 8.63 f 0.11 8.3 6& 0.07 7.29f 0.1 7 log AFbs 11 .I 0 - t 0.07 11 .O2 fsO.41 11.15+O.O5 9.7f0.3 8.38k0.06 7.60+0.10 AH$ l2.O1+O.O6 10.87+0.15 7.80+0.09 8.94fO.23 -AS$ 10.0+0.2 21.0k0.5 23 -3 f 0.3 27.2f0.8 -AS$ 9.7+ 0.3 10.1 & 1.6 9.5k0.2 16.1+ 1.5 22.2+O.3 25.7 0.5 AH; 1 .14f O.& - - 1.1f0.3 - 1.4+ 0.2 AH; -2.71 k0.69 - 0.4+ 0.5 EoHbS Ezs- E$s 1 3 3 f 0.1 6 5.2fO.74 1.54+0.18 1 s6k0.38 log Azs/AFbs 0.06+0.11 2.4f 0.5 0.02f0.13 0.31 40.27 AH$ 13.34+ 0.10 16.O5+OS9 12.88+0.06 13.25+ 0.44 9.1 5+ 0.09 10.57kO.15 AS; 1 1 .O+ 0.7 - 5+ 1 As; 0.4f 1.9 7+2 4f 1 kf and Af at 25°C in dm3 mol-' s-l; k b and Ab at 25°C in s-l; E and AH in kcal mol-I; AS in cal deg-l mol-'; 1 cal = 4.184 J; errors are standard deviations.0 Z w536 KINETIC ISOTOPE EFFECTS DISCUSSION We refer to the systems as I, I1 and 111 as shown in table 5. Let us first focus attention only on the forward reactions (I&). For system I, the results show no marked abnormalities. The values of kH/kD, of - EFbs) and of kfFbs/kfrbs might be explained on the basis of classical transition-state theory, but they are larger than usual and a small tunnelling correction is not excluded. The individual values of the rate constants and activation parameters are comparable with those previously determined for similar reactions of other carbon acids.The negative entropy of activation is consistent with the bimolecular process, and its size suggests that only a small change of solvation occurs on. passage into the transition state. This conclusion is supported by the fact that the entropy of activation is very nearly the same for the corresponding reaction of deutero-TNPM in. ethanol and in ethan[2H]ol. The positive values of ASo show that there is a decrease of solvation on forming products, presumably because of the delocalisation of the charge on the anion. The solvent isotope effect on kD, i.e. kD(EtOH)/kD(EtOD), at 25°C is 0.39 for the forward reaction and 0.21 for the backward reaction. For system 11, the large negative entropy of activation suggests considerable solvation of the transition state by isopropyl alcohol, and this is supported by the smaller activation energy compared with system I.Such compensatory changes in AH * and AS* are frequently observed for processes controlled by solvation. Similar trends would also be observed if tunnelling is more important in system I1 than in system I. The isotope effect is hardly changed from the value found for ethoxide ion as base, but the value of (Erbs -EF& 1.54k0.2 kcal mol-l, is large enough to suggest that tunnelling is becoming a significant factor in the kinetics. The maximum value of this quantity that can be simply explained by the classical transition state theory * is 1.15 kcal mol-I. The view that tunnelling increases with increasing bulk of the base is supported by the very large isotope effect we have found when t-butoxide is the base (system 111).Unfortunately, as we have already indicated, there is some doubt as to the accuracy of this result. However, even if the derived values of kH and kD for the forward reaction are in error by as much as 25 %, which seems unlikely, the isotope effect is still large and indicative of tunnelling. When the backward reactions are considered, the present results lead to a number of problems in analysing the data in terms of eqn (1) for proton transfer and eqn (3) and (4) for deuteron transfer : 3 4 SD + B-+S-+ DB HB 11 fast 1 SH + B-+S- + HB. (3) (4) The first is that the observed values of k: and k t are different. Equilibration of deuterium with the solvent would be rapid, so that step 4 should be negligible; provided that step 1 is rapid compared to steps 2 and 4, the reverse reaction should correspond to step 2, so that kF and k? should be identical. If this condition does not apply, it would be expected that, with the deutero-compound, exponential rate curves would not be observed and that the rate would increase during the course of the reaction as the concentration of SH increases.However, although the rate curves for the transfer of deuterium to ethoxide in ethanol +toluene are exponential over onlyE. F . CALDIN, E . DAWSON, R . M . HYDE AND A . QUEEN 537 the last 80-85 % of the reaction, the initial parts of the curve are steep, so that an apparent rate decrease is observed over the first 15-20 %. The large values of kH/kD for the reverse reactions are, therefore, not in accord with eqn (I), (3) and (4).We suggest that the product which is observed in the stopped-flow experiments is the species (S-----H-B) in which the proton is strongly hydrogen-bonded to the anion. As a consequence, equilibration of this proton with the hydroxyl protons of the solvent is slow. This suggestion is in line with recent proposals by Streitwieser concerning the nature of the ionic species formed in reactions between triarylmethanes and caesium in cyclohexylamine. Kollmeyer and Cram lo have reached simiIar con- clusions from their studies of alkoxide-catalysed proton-deuteron exchange reactions of diphenyl- and triphenyl-methanes. The reaction scheme should therefore be written : k f kb SH + B-+(S-----.H-B).Acceptance of this scheme still leaves one problem to be solved. The large values of k?/k?, (E?bs-E:bs) and &bs/A:I,s for the reverse reaction with ethoxide as base all point to a large tunnelling factor. It is difficult to explain these values in other term. Although it is true that the rate constants kb are small, and probably subject to greater errors than the standard deviations indicate, nevertheless any reasonable allowance for this, which must take into account the self-consistency of the data and the agree- ment with the earlier low-temperature values of Caldin and Trickett, would still leave large differences between the proton and deuteron transfer reactions. It is difficult to rationalize a large tunnelling factor in one direction in a reversible process with a small effect in the other.A comparison of the data for ethanol + toluene as solvent with those for etha~~[~H]ol+ toluene as solvent does not help in supplying an answer to this problem; besides solvent effects on the equilibrium, it would be necessary to allow for the fact that ethoxide is a stronger base in deuteroethanol than ethanol while the deuteroalcohol is the stronger acid. Moreover, the reversal of the situation as the solvent and base are changed in systems I, I1 and I11 suggests that solvent reorganization must play an important part in determining the observed rates and activation parameters. Another instance of a solvent-dependent isotope effect has recently been reported.ll " Solvent lag " has been previously suggested l 2 as a neglected factor in fast reactions and would necessarily alter isotope effects. For example in a reaction where solvent molecules are more strongly held in the transition state than in the initial state, k,/k, would be greater than if equilibrium were main- tained. For the present reaction, -AS$ > -AS%, indicating (in the absence of tunnelling) that solvent reorganization is greater for the reverse reaction than for the forward one, and any solvent lag would lead to a larger isotope effect for this step, as observed. However, solvent lag should also lead to an increase in AS* (or A ) and the effect should be larger for H than for D. Consequently, &bSb,l&bs should be less than unity if tunnelling is excluded, which is not the case. Clearly, more data are requiIed before the problem can be resolved. The authors gratefully acknowledge grants in support of this work by the S.R.C. (to R. M. H.), the National Research Council of Canada and the Research Board of the University of Manitoba and helpful discussions with Professors R. P. Bell and J. R. Reeffe. E. F. Caldin and J. C. Trickett, Trans. Faraday SOC., 1953, 49, 772. E. S. Lewis and L. H. Funderburk, J. Amer. Clzem. SOC., 1967, 89, 2322. Montagne, Rec. Trav. chim., 1905, 24, 105.538 KINETIC ISOTOPE EFFECTS E. L. Smith, J. Chenz. SOC., 1927, 1284. E. F. Caldin and G. Tomalin, Trans. Faraduy Soc., 1968, 64, 2814. E. F. Caldin, J. E. Crooks and A. Queen, J. P/i.vs. E, 1973, 6, 930. ’ J. E. Crooks, M. S. Zetter and P. A. Tregloan, J. Phys. E, 1970, 3, 73. R. P. Bell, The Proton in Chemistry (Chapman and Hall, London, 2nd edn., 1973), chap. 12 ; M. E. Schneider and M. J. Stern, J. Amer. Chem. SOC., 1972, 94, 1517. A. Streitwieser, J. R. Murdoch, G. Hafelinger and C. J. Chang, J. Amer. Chem. SOC., 1973,95, 4248. lo D. J. Cram and W. D. Kollmeyer, J. Amer. Chem. SOC., 1968,90, 1791. l1 E. F. Caldin and S. Mateo, Chem. Comm., 1973, 854. l2 R. P. Bell, Disc. Faraduy Soc., 1965, 39, 16 ; M. M. Kreevoy and R. A. Kretchmer, J. Amer. Chem. SOC., 1964,86,2435.