Reactions of substituted benzylidene Meldrum’s acids and methylthiobenzylidene Meldrum’s acids with OH – , CF 3 CH 2 O – and HOCH 2 CH 2 S – in 50% DMSO–50% water. π –Donor effects, soft acid–base interactions and transition state imbalances

A kinetic study of the reactions of phenyl substituted benzylidene Meldrum’s acids ( 5-H-Z with Z = 4-MeO, 4-Me, H, 4-Br, 3-Cl and 4-NO 2 ) and methylthiobenzylidene Meldrum’s acids ( 5-SMe-Z with Z = 4-MeO, 4-Me, H, 4-Br, 4-CF 3 and 3,5-(CF 3 ) 2 ) with OH – , CF 3 CH 2 O – and HOCH 2 CH 2 S – in 50% DMSO–50% water (v/v) at 20 °C is reported. The reactions of 5-H-Z lead to reversible attachment of the nucleophile to the substrate; with CF 3 CH 2 O – and HOCH 2 CH 2 S – rate ( k 1 Nu and k 1 − Nu ) and equilibrium constants ( K 1 Nu ) for this process could be determined while for the reactions with OH – only k 1 Nu was obtained. The reactions of 5-SMe-Z lead to substitution of the MeS group by the nucleophile via a two–


Introduction
In a series of recent papers we have reported kinetic studies of nucleophilic vinylic substitution (S N V) reactions that proceed by the attachment-detachment mechanism (eq 1) typical for substrates activated by electron withdrawing groups (X, Y). [1][2][3][4][5][6][7][8][9][10][11][12][13] We have focused mainly on reactions where the intermediate (2) accumulates to detectable levels 1,2,[5][6][7][8][9][11][12][13]  A major reason why 2 accumulates to detectable levels in these reactions is that the πacceptor groups (X, Y) provide the necessary stabilization of the intermediate by delocalizing the negative charge. This mplies that the formation of 2 should show some of the characteristic features of reactions that lead to resonance stabilized anions in general.One such feature is the relatively high intrinsic barrier or low intrinsic rate constant, 14 especially for substrates with strong π-acceptors.[15][16][17] Another is imbalanced transition states in the sense that charge delocalization into X,Y lags behind bond formation between carbon and the nucleophiles.[15][16][17] What complicates the reactivity patterns is that other factors such as the π-donor ability of the leaving group, steric effects and anomeric effects 18 play an important role as well.Nevertheless, there is an accumulating database that confirms the expected correlation between intrinsic barriers and π-acceptor strength for these reactions but there is only a relatively small number of investigations that address the question of transition state imbalance.Information about the latter comes from the study of substituent effects that provide Brønsted-type structurereactivity coefficients such as α Apart from addressing the issue of transition state imbalances, our results also allow an assessment under what conditions the π-donor effect of a para methoxy group is effective or ineffective.Furthermore, some of our data suggest the operation of the hard-soft acid-base principle.24

Results
General Procedures.All kinetic experiments were performed in 50% DMSO-50% water (v/v) at 20 °C and conducted under pseudo-first-order conditions with the vinylic substrates as the minor component.The basic features of the kinetic experiments were quite similar to those for the reactions of the parent substrates (Z = H) with OH -, CF 3 CH 2 O -and HOCH 2 CH 2 S -, respectively, reported earlier 11,25 and hence only an abbreviated account is given here.For the reactions of 5-H-Z there is no k 2 Nu -step (eq 1) because H is not a real leaving group and hence only data on the nucleophilic attachment step were obtained.
For the reactions of 5-SMe-Z with HOCH 2 CH 2 S -the intermediate accumulates to detectable levels and all rate constants in eq 1 could be determined.For the reactions of 5-SMe-Z with CF 3 CH 2 O -the kinetic results are similar to those for the reactions with HOCH 2 CH 2 S -but some ambiguities as to the nature of the second phase of the reaction render an interpretation of the data difficult. 11For the reactions of 5-SMe-Z with OH -the breakdown of the intermediate to products is faster than its formation and hence only k 1 Nu could be measured. 11ll the rate and equilibrium constants determined in this study are summarized in Tables 1 (5-H-Z) and 2 (5-SMe-Z).In the following sections the individual reactions are briefly described.

Reactions of 5-H-Z with OH -.
The reaction refers to eq 2 and the observed pseudo-first-order rate constants are given by eq 3. Plots of k obsd vs. [OH -] were obtained in the [OH -] range from 0.015 to 0.10 M (6 points) from which k 1 OH was determined; the intercepts were too small for a determination of k -1 .Due to the high reactivity of HOCH 2 CH 2 S -, its reactions with 5-H-Z had to be run at pH 7.34, a pH well below the pK a (10.57) of HOCH 2 CH 2 SH, in order to keep the free HOCH 2 CH 2 S - concentrations low.This pH was maintained with an N-methylmorpholine buffer.Even at this pH, the intercepts of the plots according to eq 5 were too small to yield reliable k 1 − RS values in the [HOCH 2 CH 2 S -] range from 1.1 × 10 -6 to 1.6 × 10 -5 M (8 points).In this case, k 1 RS being the equilibrium constant; the K 1 RS values were obtained in chloroacetate buffers or HCl solution by applying classical spectrophotometric methodology.

Reactions of 5-SMe-Z with OH -.
As shown previously for 5-SMe, 11 the reaction with OH - leads to the corresponding substitution product which, under most conditions, is ionized and is best represented as the enolate 8a ↔ 8b. 9 The rate limiting step is nucleophilic attachment of OH -to 5-SMe, eq 7 and hence k obsd is given by eq 8. Data were obtained in [OH -] range from 0.04 to 0.24 M (6 points).
Reactions of 5-SMe-Z with CF 3 CH 2 O -.As described earlier, 11 the reaction of 5-SMe with CF 3 CH 2 O -buffers in characterized by two kinetic processes 26 ; the same is true for the substituted derivatives.This process could be observed at 335 nm (loss of 5-SMe) as well as at 260 nm (formation of the intermediate).The data at 335 nm were of better quality and hence the results obtained at this wavelength were used in our kinetic analysis.Plots of k obsd vs. [CF 3 CH 2 O -] obtained at pH 14.6 (ca.1.3 × 10 -2 to 1.9 × 10 -1 M, 8 points) were linear either without intercept or with a small one.The slopes correspond to k 1 TFEO .The interpretation of the intercepts is ambiguous because they represent a complex combination of contributions from k 1 − TFEO and a k 1

OH
[OH -] term from the reaction with OH -. 26 Hence no kinetic information was extracted from these intercepts.
The second process was only observed at 260 nm (loss of the intermediate); it is much slower than the first.As discussed earlier, 11 it is not clear what reaction this process is referring to; even though the dependence of k obsd on [CF 3 CH 2 O -] is consistent with rapid equilibrium formation of the intermediate shown in eq 9 followed by rate limiting conversion of the intermediate to products, 28 the spectral changes associated with this process imply a more complex and poorly understood reaction.We therefore refrain from analyzing the kinetic data.
Reactions of 5-SMe-Z with HOCH 2 CH 2 S -.The kinetic behavior of these reactions is similar to that for the reactions of 5-SMe-Z with CF 3 CH 2 O -, i.e. there is a fast process leading to the formation of the corresponding intermediate (eq 10) followed by a slow process, where now intermediate formation represents a fast preequilibrium and product formation is rate limiting (eqs 10a and 10b); in this case the spectral evidence is fully consistent with this interpretation. 11e fast process was studied at pH 10.  Z         The large values imply that the negative charge of the intermediate is relatively close to the substituent and not highly delocalized into the two ester groups, in agreement with independent evidence that charge delocalization plays only a secondary role in the strong stabilization of the Meldrum's acid anion.[31][32][33][34] By way of contrast, for morpholine addition to substituted βnitrostyrenes, eq 14, 29 ρ(K 1

Nu
) is much smaller (≈ 1.09), 35 consistent with the strong delocalization of the negative charge into the nitro group.
The fact that ρ(K 1

Nu
) for HOCH 2 CH 2 S -attachment to 5-SMe-Z (2.21) is significantly smaller than for addition of the same nucleophile to 5-H-Z (2.92) calls for comment.A possible interpretation of this result is that the soft-soft 24 acid-base interactions between the thiolate ion and the MeS group in 5-SMe-Z provide some extra stabilization to the intermediate, thus attenuating the substituent effect on the stability of the intermediate.This is reminiscent of findings in the reaction of thiolate ions with Fischer carbene complexes which were interpreted in a similar way. 36he ρ(k 1

Nu
) values for nucleophilic attack are very similar (0.98 to 1.18) for all reactions, i.e.Hammett plots (Figure 5).The most plausible explanation for this reduced π-donor effect is that π-donation by the MeS group (11a ↔ 11b) is more effective and greatly diminishes the importance of 11c.The fact that λ max (334 nm) for 5-SMe-MeO is the same as for the other 5-SMe-Z derivatives (4-Me, H, 4-Br: 334 nm; 4-CF 3 and 3,5-(CF 3 ) 2 : 335 nm) is consistent with our explanation.
Transition state imbalances.One of the major motivations for this study was to learn more about potential transition state imbalances in the nucleophilic attachment steps.Such imbalances where charge delocalization into the π-acceptor group(s) in carbanion forming reactions lags behind bond formation (see, e.g.12) appears to be the norm [15][16][17] and would be expected to occur in the reactions reported in this paper.
They would manifest themselves in α nuc n values that exceed β nuc n ; this is because, due to the closer proximity of the negative charge to the Z-substituent at the transition state than in the adduct, the stabilization of the transition state by electron withdrawing substituents is disproportionately strong and hence the substituent effect on k 1 Nu is exalted.complex dependence of the imbalance on the π-acceptor groups, the nature of the nucleophile and on whether or not the substrate has a leaving group.Comparison of entries 4, 6 and 7 suggests that for the same LG the imbalance is relatively large for substrates with strong πacceptors (NO 2 ) but small when the π-acceptors are weaker (COOR and CN groups).This is the same pattern observed in the deprotonation of carbon acids activated by the same πacceptors 15,17 and reflects the fact that the differences between the charge distribution at the transition state and that of the carbanion increases with the increasing strength of the π-acceptor.
Comparison of entries 3 and 5 shows a large increase in the imbalances when the hydrogen is replaced by a MeS group.This increase is the result of a larger α nuc n value.A possible (speculative) explanation is that a destabilization of the substrate by electron withdrawing substituents could enhance α nuc n .Such a destabilization would occur if the electrostatic interaction of Z with the partial positive charge on the sulfur atom (see 11b) is stronger than its interaction with the partial negative charge because the latter is partially delocalized.With respect to the dependence of the imbalance on the nucleophile, too many factors such as charge, central atom and size come into play to allow any meaningful conclusions to be drawn.

Nu
) values are relatively large, consistent with the notion that the negative charge on the respective intermediates or adducts is not particularly strongly delocalized.

Nu
) value for HOCH 2 CH 2 S -attachment to 5-SMe-Z compared to 5-H-Z may be attributed to a stabilizing soft-soft acid-base interaction in the intermediates derived from 5-SMe-Z.
(3) For the reactions of 5-H-Z the logk 1 (4) The α nuc values increase with decreasing reactivity, suggesting a trend from a somewhat reactant-like transition state to a more central one, as predicted by the Hammond-Leffler postulate.
(5) For all the reactions where α nuc could be determined, α nuc exceeds β nuc , indicating the presence of a transition state imbalance.The size of the imbalance depends on several factors most of which are poorly understood except for a definite trend towards larger imbalances with increasing π-acceptor strength of the activating groups.

Experimental Section
Materials.The substrates in the 5-H-Z series were prepared as described by Schuster et al. 41 The substrates in the 5-SMe-Z series were available from a previous study. 42l other materials were from the sources described before. 25thodology.All procedures, including the spectrophotometric equilibrium determinations, were the same as described in references 9 and 25.

1 Nu , k 1 −
because this allows the determination of all individual rate constants (k Nu and k 2 Nu ).Some representative systems studied to date are the reactions of 4-LG, 5-LG, 6-LG and 7-LG with alkoxide and thiolate ions as well as amines in 50% DMSO-50% water at 20 °C.

Reactions of 5 - 1 −TFEO . However, k 1 − 1 −
H-Z with CF 3 CH 2 O -and HOCH 2 CH 2 S -.The reactions are shown in eq 4 with Nu -= CF 3 CH 2 O -or HOCH 2 CH 2 S -; k obsd is given by eq 5.With CF 3 CH 2 O -, the reactions were run in trifluoroethanol buffers at pH values between 13.15 and 13.39.In calculating the free CF 3 CH 2 O -(TFEO) concentration, the homo-association equilibrium (eq 6, K assoc = 1.8 M -1 6 ) was taken into account.Plots of k obsd vs. [CF 3 CH 2 O -] with concentrations ranging from about 2.5 × 10 -3 to about 2.2 × 10 -3 M (6 points) were linear and yielded k 1 TFEO ; the intercepts were too small for an accurate determination of k TFEO could be determined by following the breakdown of the intermediategenerated at high pH-in a triethylamine buffer at 10.5.At this pH eq 4 favors the reactants and eq 5 simplifies to k obsd ≈ k 1 − TFEO We also examined the possibility of general acid catalysis of CF 3 CH 2 O -expulsion by Et 3 NH + by running the reaction at several [Et 3 NH + ] from 0.006 to 0.03 M. Within this range k obsd increased slightly with increasing [Et 3 NH + ] for some substrates and decreased slightly for some other substrates, but no consistent pattern emerged and hence the reported k TFEO values represent averages of k obsd .

1 − 1 RS / k 1 − 1 TFEO , k 1 TFEO , k 1 − 1 RS , k 1 RS , k 1 − 1 RS , k 1 RS , k 1 −
55 and 10.60.Plots of k obsd vs. [HOCH 2 CH 2 S -], in a [HOCH 2 CH 2 S -] range from 10 -3 to 1.3 × 10 -1 M (10-12 points), yielded straight lines according to eq 11; for Z = MeO and Me the intercepts were large enough to yield a k RS value.For Z = 4-Br, 4-CF 3 and 3,5-(CF 3 ) 2-3 measurable intercepts were obtained at pH 9.0 where lower [HOCH 2 CH 2 S -] (7.6 × 10 -6 to 1.5 10 M) could be used by buffering the solutions with DABCO.The slow process was studied at pH 10.55 and 10.60 for Z = 4-MeO, 4-Me, 4-Br and at pH 9.0 for Z = 4-Br, 4-CF 3-and 3,5-(CF 3 ) 2 .It conforms to eq 12; Fig. 1 shows a representative plot of k obsd vs. [HOCH 2 CH 2 S].Least squares analysis of the data allowed a determination of k 1 RS and k 2 RS .In those cases where k 1 RS could be determined from eq 12 the agreement with k 1 RS obtained as kRS is excellent (Table2).DiscussionHammett and Brønsted correlations.Hammett plots are shown in Fig. 2 for the reactions of 5-H-Z and 5-SMe-Z with OH -(k 1 OH ) and of 5-SMe-Z with CF 3 CH 2 O -(k 1 TFEO ), in Fig. 3 for the Issue in Honor of Prof. O. reactions of 5-H-Z with CF 3 CH 2 O(K TFEO ), in Fig. 4 for the reactions of 5-H-Z with HOCH 2 CH 2 S -(K RS ), and in Fig. 5 for the reactions of 5-SMe-Z with HOCH 2 CH 2 S - (K RS , k 2 RS ).The correlations are good except that in most cases the points for Z = 4-MeO deviate from the least squares line defined by the other substituents, a feature discussed below.The ρ values summarized in Table3were obtained by omitting the 4-MeO substituents from the correlations.

1 − 1 − 1 − 2 Nu and k 1 − 1 Nu) = ρ(k 1 Nu)/ρ(K 1 Nu) and ρ n (k 1 −Nu ) = ρ(k 1 − 1 Nuvs. log K 1 Nu and log k 1 − 1 Nu/dlog K 1 Nu = α nuc n = ρ n (k 1 Nu) and dlog k 1 −Nu /dlog K 1 Nu= −α lg n = ρ n (k 1 − 1 Nu= 3 . 32 ×
essentially independent of nucleophile and substrate.The ρ(k Nu ) values for the reverse reaction are similar for the reactions of 5-H-Z with CF 3 CH 2 O -and of 5-SMe-Z with HOCH 2 CH 2 S -(-1.19 and -1.05) but for the reaction of 5-H-Z with HOCH 2 CH 2 S -, ρ(k Nu ) = -1.95 is substantially more negative, as required by the higher ρ(K 1 Nu ).The ρ(k 2 Nu ) value for the reaction of 5-SMe-Z with HOCH 2 CH 2 S -(-0.90) is about the same as ρ(k Nu ) (-1.05).This is reasonable since both k Nu refer to the expulsion of a thiolate ion from the intermediate.Information about the transition state of the first step can be obtained from the normalized ρ values, i.e. ρ n (k Nu )/ρ(K 1 Nu ); they can also be obtained directly from Brønsted-type plots of log k Nu vs. log K 1 Nu (see Fig. 6 for a representative example), with dlog k Nu ).The α nuc n values increase from 0.33 for the reactions of 5-H-Z with HOCH 2 CH 2 S -to 0.45 for the reactions of the same substrate with CF 3 CH 2 O -to 0.53 for the reactions of 5-SMe-Z with HOCH 2 CH 2 S -, suggesting a trend from a somewhat reactant-like to a more central transition state.This trend parallels the decrease in reactivity from high for the reactions of 5-H-Z with HOCH 2 CH 2 S -(for Z = H: K 1 Nu = 5.38 × 10 10 M -1 , k 1 Nu = 1.44 × 10 7 M -1 s -1 ) to intermediate for the reactions of 5-H-Z with CF 3 CH 2 O -(for Z = H: K 1 Nu = 6.34 × 10 6 M -1 , k 1 Nu = 2.06 × 10 4 M -1 s -1 ) to low for the reaction of 5-SMe-Z with HOCH 2 CH 2 S -(for Z = H: K 10 2 M -1 , k 1 Nu = 9.22 × 10 2 M -1 s -1 ); it is consistent with the Hammond-Leffler postulate.37,38A similar trend has been observed for β nuc for the reactions of secondary alicyclic amines with a series of electrophilic alkenes.16

Nu and log K 1 Nu values for the 4 - 1 − 1 Nu and log K 1 Nu values for the 4 - 1 −
MeO derivative deviate negatively from the Hammett plots while log k Nu deviates positively.This is the result of the πdonor effect of the MeO group (10b).In the reactions of 5-SMe-Z this effect is greatly diminished because of the π-donor effect of the MeS group (11b) and the log k MeO derivatives deviate positively, and the log k Nu values negatively, from the Hammett plots.

Table 1 .
Rate and Equilibrium Constants for the Reactions of 5

Table 3 .
Hammett ρ Values a for the Rate and Equilibrium Constants of the Various Steps of the Reactions of 5-H-Z and 5-SMe-Z with OH -, CF 3 CH 2 O -and HOCH 2 CH 2 S -in 50% DMSO-50% water (v/v) at 20 °C.