Solvent effects empirical measures

Solvents exert their influence on organic reactions through a complicated mixture of all possible types of noncovalent interactions. Chemists have tried to unravel this entanglement and, ideally, want to assess the relative importance of all interactions separately. In a typical approach, a property of a reaction (e.g. its rate or selectivity) is measured in a laige number of different solvents. All these solvents have unique characteristics, quantified by their physical properties (i.e. refractive index, dielectric constant) or empirical parameters (e.g. ET(30)-value, AN). Linear correlations between a reaction property and one or more of these solvent properties (Linear Free Energy Relationships - LFER) reveal which noncovalent interactions are of major importance. The major drawback of this approach lies in the fact that the solvent parameters are often not independent. Alternatively, theoretical models and computer simulations can provide valuable information. Both methods have been applied successfully in studies of the solvent effects on Diels-Alder reactions. 

In Section 8.4 we will encounter many empirical measures of solvent polarity. These are empirical in the sense that they are model dependent that is, they are defined in terms of a particular standard reaction or process. Thus, these empirical measures play a role in the study of solvent effects exactly analogous to that of the substituent constants in Chapter 7.)... 

Some of these model-dependent quantities were formulated as measures of a particular phenomenon, such as electron-pair donor ability but many of them have been proposed as empirical measures of solvent polarity, with the goal, or hope, that they may embody a useful blend of solvent properties that quantitatively accounts for the solvent effect on reactivity. This section describes many, although not all, of these empirical measures. Reichardt has reviewed this subject. 

The rate constants in organic reaction in a solvent generally reflect the solvent effect. Various empirical measures of the solvent effect have been proposed and correlated with the reaction rate constant [5]. Of these, some measures have a linear relation to the solubility parameter of the solvent. The logarithms of kj and k2/ki were plotted against the solubility parameter of toluene, NMP and DMSO[6] in Fig. 2. As shown in Fig.2, the plots satisfied the linear relationship. The solvent polarity is increased by the increase of solubility parameter of the solvent. It may be assumed that increase of unstability and solvation of Ci due to the increase of solvent polarity make the dissociation reaction of Ci and the reaction between Ci and COisuch as SNi by solvation[7] easier, respectively, and then, k2/ki and ks increases as increasing the solubility parameter as shown in Fig. 2. 

Similar solvent effects on HFS constants have been observed for other aminyl-oxides such as diphenyl aminyloxide [207, 212], di-r-butyl aminyloxide [218, 385], r-butyl aminyloxide [213, 217], and 2,2,6,6-tetramethylpiperidyl-l-oxide [384, 387, 388]. The constants of di-t-butyl and two other aminyloxides have been proposed as an empirical solvent polarity parameter because a( " N) is easily measured in most solvents [218, 389] cf. Section 7.4. 

Solvatochromic fluorescent probe molecules have also been used to establish solvent polarity scales. The solvent-dependent fluorescence maximum of 4-amino-V-methylphthalimide was used by Zelinskii et al. to establish a universal scale for the effect of solvents on the electronic spectra of organic compounds [80, 213], More recently, a comprehensive Py scale of solvent polarity including 95 solvents has been proposed by Winnik et al. [222]. This is based on the relative band intensities of the vibronic bands I and III of the % - n emission spectrum of monomeric pyrene cf. Section 6.2.4. A significant enhancement is observed in the 0 0 vibronic band intensity h relative to the 0 2 vibronic band intensity /m with increasing solvent polarity. The ratio of emission intensities for bands I and III serves as an empirical measure of solvent polarity Py = /i/Zm [222]. However, there seems to be some difficulty in determining precise Py values, as shown by the varying Py values from different laboratories the reasons for these deviations have been investigated [223]. 

A more simplified but likewise sueeessful empirical two-parameter approach for the deseription of solvent effects has been proposed by Krygowski and Fawcett [113]. They assume that only specific solute/solvent interactions need to be eonsidered. These authors postulated that the solvent effeet on a solute property A can be represented as a linear funetion of only two independent but eomplementary parameters describing the Lewis aeidity and Lewis basicity of a given solvent. Again, for reasons already mentioned, the t(30) values were chosen as a measure of Lewis acidity. In addition, Gut-mann s donor numbers DN [26, 27] were chosen as a measure of solvent basicity cf. Table 2-3 and Eq. (7-10) in Sections 2.2.6 and 12, respectively). Thus, it is assumed that the solvent effect on A can be described in terms of Eq. (7-62) . 

Berson, J.A., Hamlet, Z. and Mueller, W.A. (1962). The Correlation of Solvent Effects on the Stereoselectivities of Diels-Alder Reactions by Means of Linear Free Energy Relationships. A New Empirical Measure of Solvent Polarity. J.Am.Chem.Soc., 84,297-304. 

Thus, the chemical shift itself is an incredibly rich and precise source of information. Indeed, these shifts can be thought of as a unique fingerprint of the molecule under the conditions of the measurement. The obvious question that arises is why a 3D structure cannot be derived from chemical shifts. In theory, there is no reason why this should not be possible however, in practice, the ab initio models required for accurately calculating the above contributions in a dynamic and complex system such as a protein, in particular when solvent effects are considered, are beyond our current computational capabilities. All is not lost however, as the fact that proteins consist of a limited number of residue types that are connected through the same repetitive bonding structure allows for derivation of vastly simplified empirical models for approximating the abovementioned complex contributions. 

Extensive studies have been made of solvent effects on atom transfer reactions involving ions [12]. In the case of reaction (7.3.23), the rate constant decreases from 250M s in A-methylpyrrolidinone to 3 x 10 M s in methanol. This effect can be attributed to solvation of the anionic reactant Cl and the anionic transition state [12]. Since the reactant is monoatomic, its solvation is much more important. It increases significantly with solvent acidity leading to considerable stabilization of the reactants. As a result the potential energy barrier increases and the rate decreases with increase in solvent acidity. As shown in fig. 7.7, this leads to an approximate linear relationship between the logarithm of the rate constant and the solvent s acceptor number AN, an empirical measure of solvent acidity (see section 4.9). Most of the results were obtained in aprotic solvents which have lower values of AN. The three data points at higher values of AN are for protic solvents. 

Revalues are an empirical parameter of solvent polarity which are derived by measurement of the long-wave UV/ Visible absorption band of the negative solvachromatic Pyri-dinimn-N-phenoxide betaine dyes in the solvent being studied. Higher values are an indication of greater solvent polarity. The above values were taken from the text by Reichardt C (1990) Solvents and Solvent Effects in Organic Chemistry, 2nd edn. VCH Publishing, Weinheim, p 365... 

Many empirical measures of solvent polarity have been developed.One of the most useful is based on shifts in the absorption spectrum of a reference dye. The position of absorption bands is sensitive to solvent polarity because the electronic distribution, and therefore the polarity, of the excited state is different from that of the ground state. The shift in the absorption maxima reflects the effect of solvent on the energy gap between the ground state and excited state molecules. An empirical solvent polarity measure called (SO) is based on this concept. Some values for common solvents are given in Table 3.33 along with the dielectric constants for the solvents. It can be seen that a quite different order of polarity is given by these two quantities. 

The effects of solvent on charge-transfer maxima have been reviewed by Murrell and Reichardt , the latter in connection with empirical measures of solvent polarity. The rules of thumb are as follows For complexes of the non-ionic type, the charge-transfer maximum should be red shifted in polar solvents, blue shifted in non-polar solvents. For ionic complexes, the opposite behavior is anticipated. However, a recent attempt to correlate absorption maxima of complexes with solvent polarity has failed, and it appears there may be exceptions to the general rules. 

Because of the often-observed inadequacies of the dielectric approach, that is, using the dielectric constant to order reactivity changes, the problem of correlating solvent effects was next tackled by the use of empirical solvent parameters measuring some solvent-sensitive physical property of a solute chosen as the model compoimd. Of these, spectral properties such as solvatochromic and NMR shifts have made a spectacular contribution. Other important scales are based on enthalpy data, with the best-known example being the donor mun-ber (DN) measuring solvent s Lewis basicity. 

---