Reactivity-selectivity relationships

The transition states of free radical reactions generally show evidence of polar character wherein electron transfer to or from the radical has occurred (20). Thus, the electron affinity or ionization potential of the radical involved should affect the reaction. The much higher electron affinity (16) of ROo than CH3 radicals no doubt alters the transition state so that the reactivities toward it show less selectivity. The results of Szwarc and Binks (22) center around the fact that only carbon radicals were used for the correlation, and thus the electron affinity does not vary sufficiently to show in the correlation any deviation from the expected reactivity-selectivity relationship. 

An extensive review with many examples125 shows that the reactivity-selectivity principle cannot be used to predict the selectivity of a reaction except in unique systems where one reaction is close to or diffusion controlled. The relative importance of the Hammond effect and the frontier-orbital effects determines the reactivity-selectivity relationship that will be found in a particular system. The review also concludes that the Hammond-Leffler a-value cannot be used as an indicator of transition-state structure. 

The chemical reactivity most associated with dioxiranes is the electrophilic transfer of oxygen to electron-rich substrates (e.g., epoxidation, N-oxidation) as well as oxygen insertion reactions into unactivated C-H bonds. The reactivity-selectivity relationships among these types of reactions has been examined in depth by Curci. The reaction kinetics are dependent upon a variety of factors, including electron-donor power of the substrate, electrophilicity of the dioxirane, and steric influences (95PAC811]. 

This review, therefore, attempts to survey the current status of reactivity-selectivity relationships by exploring the basis and variety of such relationships. The part they have played in a number of selected areas of mechanistic interest is also discussed. No attempt is made to cover every facet of mechanistic study. Rather, the aim has been to point out different reactivity-selectivity relationships that have been noted in such studies and to focus attention on the use of such relationships as a mechanistic tool. In addition, an attempt has been made to outline the limitations of the reactivity-selectivity principle. This serves not only to increase the utility of the principle but also strengthens the basis and justification for its use by rationalizing apparent failures that have been reported. 

It is important to note that the above presentation, justifying the reactivity-selectivity principle, is based on a number of fundamental assumptions. First, it is assumed that the Leffler-Hammond postulate is valid, which in turn implies that the reaction under consideration obeys a rate-equilibrium relationship [eqn (2)]. This assumption often cannot be verified since for reactions of highly active species such as carbenes, free radicals, carbonium ions, etc., equilibrium constants are generally not measurable. However it follows that for reactions which do not conform to a rate-equilibrium relationship, no reactivity-selectivity relationship is expected. Also, in Fig. 4, the difference in the free energy of the... 

This section shows the part reactivity-selectivity relationships have played in a number of areas of mechanistic interest. The discussion considers a number of different aspects simultaneously. 

Thus the observation of reactivity-selectivity relationships is reported and the utilization of the principle pointed out. In addition, an attempt is made to indicate the validity of the various measures of selectivity as well as the possible rationalizations of the observed breakdown in the principle which have been reported for these specific areas. 

Bell (1959) has observed that the experimental data for the halogenation of a series of carbonyl compounds also yield a reactivity-selectivity relationship. Thus reactive compounds show a low measure of selectivity while unreactive compounds show a high measure of selectivity. The data are listed in Table 2. Bell et al. 

Kresge et al. (1971a) have observed a reactivity-selectivity relationship in proton transfer reactions, though of poor correlation. 

The preceding discussion has suggested that a (or 0) may be considered a measure of transition state structure even if the expected reactivity-selectivity relationship is not observed. There is, however, strong evidence to suggest that the Br nsted coefficient does not always reflect the degree of proton transfer in the transition state. 

The use of reactivity-selectivity relationships in the study of solvolysis reactions has been somewhat limited, although during recent years the potential of such relationships as a mechanistic tool has become apparent. 

Raber et al. (1971b) noted that, in addition to 2-octyl mesylate, a number of primary and secondary substrates which also underwent solvolysis with substantial nucleophilic solvent assistance all showed considerably higher selectivities than expected from the reactivity-selectivity relationship illustrated in Fig. 8. They concluded that, while the failure of these points to correlate with the carbocations did point to a mechanistic difference between the two groups, the conclusion... 

The reactivity-selectivity relationship observed by Harris et al. (1973, 1974a) and illustrated in Fig. 10 provides additional support for this view. Thus 1-adamantyl, 2-adamantyl, and exo-2-norbonyl chlorides, which are hindered to back-side nucleophilic attack, show negative selectivity values. The reason for the negative selectivity value exhibited by 4,4 -dichlorodiphenylchloromethane and for its failure to conform to the reactivity-selectivity correlation is obscure. However, the result itself suggests that product formation is predominantly via the solvent-separated ion pair. 

A reactivity-selectivity relationship based on the use of an ambident nucleophile has been observed by Okamoto and Kinoshita... 

Equation (33) suggests that the failure to observe a reactivity-selectivity relationship for electrophile-nucleophile combination reactions is due to the counter influence of the solvent. The parameter 0 represents the inherent reactivity of the electrophile and is large for unreactive electrophiles, while for reactive electrophiles P is small. Now, it is the reactive electrophiles which are... 

What is clear, is that solvent effects may play a substantial, complex, and yet often subtle role in the solvolytic reaction mechanism, and that further study is required to increase the limited understanding which has been achieved to date. The original reactivity-selectivity relationships obtained by Sneen et al. (1966a) and Raber, Schleyer et al. (1971a) are now seen to be entirely fortuitous. The N+ correlation has demonstrated that the selectivities of cations are independent of their stability. Hence, the observed relationships are a result of the averaging of selectivities of attack at the different ion pair stages, and are not simple reactivity-selectivity relationships at all. 

Reactivity-selectivity relationships are obtainable in carbene chemistry provided an independent selectivity parameter is found. Under such circumstances the relative rate data then serve as a measure of relative reactivity. This has been done with a number of Hammett studies. Thus the addition of dichlorocarbene to compounds [ 1 ] — [3] (listed in Table 15) indicates that a reactivity-selectivity relationship is obtained using p as a measure of selectivity. A more reactant-like transition state is obtained for the more reactive substrate in accordance with the Leffler- Hammond postulate. 

However, such a reactivity-selectivity relationship is not always observed. For example, the selectivity data for reaction of substituted phenylcarbenes with isobutene and tranr-2-butene are listed in Table 16 (Moss, 1973). Interestingly, the data appear to support the reactivity- selectivity principle. An electron donating substituent, Z, is thought to stabilize the arylcarbene. Thus the data in Table 16 were interpreted (Moss, 1973 Closs and Moss, 1964) as signifying... 

Early examples of reactivity-selectivity relationships in aromatic substitutions are limited since, in the absence of absolute rate data, it is often difficult to assign relative reactivity to the different electrophiles. For certain cases where the relative reactivity order may be assumed, a reactivity-selectivity relationship was noted. For example, bromina-tion with the reactive species Br+ results in lower selectivity than with the less reactive species Br2 (de la Mare and Harvey, 1956 Brown, 1957). However, it appears that no general reactivity-selectivity relationship exists in electrophilic aromatic substitution reactions, for there exist slow, unselective reactions such as aromatic... 

More recently, Olah et al. (1970) have varied the reactivity of the electrophile in a systematic way and observed a number of convincing examples of reactivity- selectivity relationships. The Friedel-Crafts benzylation reaction on benzene and toluene was conducted for a large number of substituted benzyl chlorides. The data are in Table 18 and indicate that the selectivity of electrophilic attack decreases as the substituent becomes more electron withdrawing. 

Since such substituents are expected to decrease electrophile stability, the observed trend represents a particularly clear example of a reactivity-selectivity relationship. Isomer distribution of the benzylated toluenes, while not correlating well with the relative rate data, do show a similar trend. 

Reactivity-selectivity relationships play an important part in free radical chemistry for the same reasons as in carbene chemistry and electrophilic substitution. Absolute rate constants for free radical reactions are not generally available (and when they are known they are often associated with large systematic errors), and the use of relative rate studies is an important technique in the study of free radical reactions. A comprehensive monograph dealing with various... 

In spite of these uncertainties, however, the utility of the reactivity-selectivity principle has been illustrated for a number of diverse areas of mechanistic interest. Such applications are being extended to other areas as well. For example, Olah has recently studied the mechanism of electrophilic addition to multiple bonds using selectivity data and concluded that the transition states of the bromine addition to alkenes are of a 7r-complex nature (Olah and Hockswender, 1974). Finally the large number of reactivity-selectivity relationships which have been discovered offer considerable experimental support for the various expressions and formulations of the Hammond postulate whose profound effect on modem mechanistic chemistry is now beyond question. 

As far as the tertiary benzylic solvolyses are concerned, any structural and mechanistic perturbations are reflected only in the variation of the p parameter. The p value for a reaction series is a parameter of intermolecular selectivity and can change sensitively with the reactivity (or the stability of transition state). This behaviour is often referred to as adherence to the reactivity-selectivity relationship (RSR), where the selectivity (5) may vary inversely with the intrinsic reactivity of members of a reaction series, as formulated in equation (3),... 

In the bromination in methanol of several styrenes, X-C6H4C(R )=CH2, of differing reactivities (Ruasse et al., 1978), the reactivity increase corresponds roughly to an attenuation of both polar and resonance selectivities, p and pr, respectively. The reactivity-selectivity relationship based on the data of Table 11 is approximately linear for pr but not for p. The curvature of the Pr vs. log(ko)R plot was attributed to the change in the charge distribution due to charge delocalization in R , but the pr vs. log(ko)R plot was linear since pr should be independent of the charge distribution. This led to the conclusion that the problem of the balance between the transition-state position and the... 

The correlation results for the bromination of diarylethylenes [31(X,Y)] summarized in Table 14 also involve the same serious problem. The p value increases significantly as the fixed substituent Y becomes more EW. This behaviour is indeed what is expected for the quantitative reactivity-selectivity relationship. However, in Table 14, the range of variable substituents X involved in the correlation of the respective Y sets is evidently different from set to set. The correlation for the Y = p-MeO set giving p = -2.3 should be referred to as the correlation for the T-conformation where X is more EW than Y, correlations for Y = p-Me, H and p-Br sets giving p = -3.6 may be referred to the E-conformation, and those for Y = m-Hal, especially P-NO2, refer without doubts to the P-conformation. The variation of p value cited in Table 14 demonstrates nothing other than the dependence of the selectivity p upon the propeller conformation of the diaryl carbocations. While there is no doubt regarding the importance of RSR in the mechanistic studies, these results lead to the conclusion that the RSR, or most of the non-additivity behaviour of a,a-diarylcarbocation systems which have been cited as best examples of quantitative RSR, may indeed be only an artifact. 

The rate constants for reactions of highly stable triphenylmethyl and diphenylmethyl cations with various ionic and neutral nucleophiles have been measured (Gandler, 1985 McClelland etal., 1986) in aqueous acetonitrile and discussed from the view point of a reactivity-selectivity relationship. 

As already mentioned for rhodium carbene complexes, proof of the existence of electrophilic metal carbenoids relies on indirect evidence, and insight into the nature of intermediates is obtained mostly through reactivity-selectivity relationships and/or comparison with stable Fischer-type metal carbene complexes. A particularly puzzling point is the relevance of metallacyclobutanes as intermediates in cyclopropane formation. The subject is still a matter of debate in the literature. Even if some metallacyclobutanes have been shown to yield cyclopropanes by reductive elimination [15], the intermediacy of metallacyclobutanes in carbene transfer reactions is in most cases borne out neither by direct observation nor by clear-cut mechanistic studies and such a reaction pathway is probably not a general one. Formation of a metallacyclobu-tane requires coordination both of the olefin and of the carbene to the metal center. In many cases, all available evidence points to direct reaction of the metal carbenes with alkenes without prior olefin coordination. Further, it has been proposed that, at least in the context of rhodium carbenoid insertions into C-H bonds, partial release of free carbenes from metal carbene complexes occurs [16]. Of course this does not exclude the possibility that metallacyclobutanes play a pivotal role in some catalyst systems, especially in copper-and palladium-catalyzed reactions. 

The Hammond postulate is often accepted as a general principle an increase in reactivity is accompanied by a decrease in selectivity because the transition structure becomes closer to that of the reactant state as the energy barrier decreases. This idea has some truth for a hypothetical A to B reaction model where it is implicit that only a single bond change occurs moreover the Hammond postulate is predicted by the Marcus theory (above). The postulate often breaks down for reactions where more than one major bond change results in product formation. It should be emphasised that any discussion of the reactivity-selectivity relationship has to be confined to those reactions where there is no change in rate-limiting step or mechanism. 

The nucleophilicities are found to be dependent on electronic, steric, and symbiotic effects, and limited series obeyed a constant selectivity , a reactivity-selectivity or a dual-parameter linear free-energy relationship. The conclusion made was that because of different blends of the effects, the construction of a substrate-independent nucleophilicity scale was impossible at present, but an approximate scale was presented. In nucleophilic reactions on relatively long lived vinyl cations, the steric effects predominate, but at constant steric effects, reactivity-selectivity relationships were found for very short series of substrates. Additional data are required for constructing more reliable nucleophilicity scales toward neutral and positively charged vinylic carbons. 

Table IV compares the reactivity ratios of a soft (PhS-) to a hard (MeO-) nucleophile in vinylic substitution. PhS is always more reactive, and ratios lower than unity, as for 4, X = Br (4), are certainly due to elimination-addition with MeO . The ratios change by >2000-fold and are sensitive to the geometry of the substrate. An important feature is that for (3-halo-p-nitrostyrenes the ratio decreases strongly with the increased hardness of the (3-halogen (38). The lowest ratios are for the (3-fluoro derivative, whereas the differences between the chloro and bromo compounds are not so large. This behavior is similar to that in SNAr reactions. This behavior can be rationalized by symbiotic effects, which favor the soft-soft PhS--Br interaction and the hard-hard MeO-F interaction. A reactivity-selectivity relationship for vinyl bromides of different electrophilicities does not exist.

Four different probes gave short reactivity orders toward vinyl cations (1) common ion rate depression in solvolysis (2) competitive capture of solvolytically generated ions (3) direct reaction of a vinyl cation with nucleophiles and (4) competition between intra- and intermolecular nucleophilic capture. A short reactivity order is obtained in each case, but because of the different solvents and conditions the orders cannot be combined to a single series. However, a selectivity rule that governs the relative reactivities toward different vinyl cations in terms of a constant selectivity or a reactivity-selectivity relationship can be determined. 

---