**4. Discussion**

Solvolytic dediazoniations have long been studied, and there is now a substantial body of knowledge on the reactions. For instance, the sensitivity of dediazoniations towards solvent polarity has been analyzed by employing the Winstein–Grunwald equation [26,27], yielding slopes for 2MBD, 3MBD, and 4MBD of *m* = −0.026, −0.018, and −0.058, respectively. These values are quite low and reflect the astonishing insensitivity of dediazoniations to the nature of the solvent [1]. The low values are interpreted on the grounds of similarities in the structure and the charge allocation between the parent aryldiazonium ion and the corresponding aryl cation [14,50,51].

The selectivity values for a number of aryldiazonium ions have also been obtained, assuming that the nucleophilic attack on the carbocation is under kinetic control. In aqueous solution, aryldiazonium ions have a small selectivity against reacting nucleophiles, compared to water, typically ranging from 0.4 to 6 [46], following the order Br− > Cl− > SO2− 4 > H2O. Such values are orders of magnitude lower than those observed for nucleophiles competing with water in reactions with stabilized carbocations [14,52,53], and are essentially constant with solvent composition. Table 1 shows the selectivity values for 2-, 3-, and 4-MBD. Such low values are consistent the formation of highly reactive aryl cations and with the preassociation of the stepwise mechanism, as shown in Scheme 1A.

**Table 1.** Selectivity values for 2-, 3-, and 4MBD. Data from references [14,22,54].


Figure 5 shows that, at about 80% MeOH/H2O, the yields of ArOH and ArOMe are equal to each other. That is, equal yields of dediazoniation products are obtained when the concentrations of nucleophiles in solution are approximately 8.3 M for [H2O] and approximately 14.6 M for [MeOH], indicating that ArN<sup>+</sup> 2 ions show favored solvation with H2O molecules. This preferential solvation is fully consistent with the moderate variation in *k*obs on changing the MeOH content of the solvent, Figure 2. Moreover, barriers for the capture of carbocations by the solvent [55] are sufficiently small, so that the product distribution reflects the solvent distribution in the first solvation shell of the reactant, suggesting, therefore, that the solvation of the ground state ArN<sup>+</sup> 2 is essentially the same as that in the proximity of the transition state. Thus, at low MeOH content, the presumed aryl cations are mostly solvated by water molecules, and a relatively modest increase in the concentrations of NaCl, LiCl, or LiClO4 (in comparison to that of water) does not have significant effects on both the dediazoniation rate constants (Figure 3), and on product distributions (Figure 6).

The sensitivity of dediazoniations to changes in the solvent polarity can be illustrated by means of a Winstein–Gruwald plot, as shown in Figure 8, where a linear relationship between log(*k*/*k*w) with the *Y* parameter is found, with slopes of *m* = −0.032 ± 0.001 (2MBD), *m* = −0.037 ± 0.001 (3MBD), and *m* = −0.072 ± 0.003 (4MBD), consistent with the low selectivity of aryldiazonium ions to solvent effects [52,56]. This low sensitivity can be attributed to the similarity in the structure and charge distribution of the parent aryldiazonium ion, compared to that of the corresponding aryl cation.

**Figure 8.** Winstein–Grunwald plots for 2MBD, 3MBD, and 4MBD. Y values were collected from Leffer et al. [57].

Surprisingly, when dediazoniations are carried out in solutions with high concentrations of MeOH, a different behavior is observed. The addition of NaCl or LiCl does not modify *k*obs values, as seen in Figure 3D,E, and the analysis of the product distribution confirms the expected formation of the chlorobenzene derivatives. However, upon the addition of LiClO4, both *k*obs (Figure 3F) and the yields of the corresponding methyl phenyl ethers (ArOMe, Figure 7B,D) decrease. We note that chromatograms do not show the formation of the reduction products (toluenes, Ar-H) in a large extent and, thus, we can safely presume that the mechanism of the reaction does not change in the presence of LiClO4. Thus, the variations in the yields of Ar-OMe can be attributed to the replacement of MeOH molecules in the solvation shell by ClO<sup>−</sup>4 ions, presumably leading to the formation of the potentially explosive (in the solid state) aryldiazonium perchlorates.

Our hypothesis regarding perchlorate ions modifying the solvation shell of aryldiazonium ions seems to be supported by the results of Cruz et al. [58], who carried out a molecular dynamic study of dediazoniations and calculated the local concentrations of solvents around diazonium ions, concluding that, in the absence of salts, the local solvent composition varied linearly on the molar fraction of water in methanol/water mixtures, although they found non-linear variations for other solvent mixtures. The same authors also found that the calculated number of solvent molecules around the aryldiazonium cation can predict the product distribution when preferential solvation is considered. However, results must be taken with some caution because of the failure to identify the new product formed in the presence of LiClO4.

The rates of dediazoniations with uncharged nucleophiles (such as MeOH or H2O) are hardly affected by the nature of the solvent, as shown in Figure 2, because the electric charges are neither destroyed nor created in the rate-determining step. In fact, the rates of heterolytic dediazoniations have been reported to vary only by a factor of 9 when carried out in 19 solvents [1]. Thus, the low variation in *k*obs with [MeOH] suggests that the solvation of the ground-state is similar to the proximity of the transition state; otherwise, unimolecular reactions, where the nucleophilic attack of the solvent is rate-determining, would lead to a strong dependence of *k*obs on nucleophile concentration, which is not observed. The rates of the nucleophilic attack on aryl cations have been reported to be close

to the diffusion control limit (~10<sup>9</sup> <sup>M</sup>−1·s<sup>−</sup>1) and the preassociation of the nucleophile with the aryl cation does not account for much of the trapping [1,2,46].

A different situation is found when charged nucleophiles are present in the solvation shell, as they may distort the charge distribution, compared to that of neutral nucleophiles. The Mayr's reactivity scale is certainly useful to predict whether an electrophile/nucleophile reaction can be expected to take place at room temperature, or to predict the selectivity of competing nucleophiles. The rate constants of electrophile– nucleophile combination reactions are mostly dependent on four factors: nucleophilicity, electrophilicity, temperature, and solvent. Strictly, it is not possible to separate these four factors, mainly because the values of the electrophilicity and nucleophilicity parameters are usually linked to the type of solvent, whose physical properties may change with changing temperatures. Nucleophilicity scales may also be affected by changes in temperature [33,59]. However, to our knowledge, an all-embracing theory of electrophile–nucleophile reactions is still not in sight, and further work on the effects of electrolytes on ionic reactions is needed.

**Supplementary Materials:** The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/compounds2010005/s1, Figure S1: The chromatogram in Figure 1.

**Author Contributions:** Conceptualization C.B.-D.; investigation, S.L.-B. and C.B.-D.; writing— original draft preparation, C.B.-D.; writing—review and editing, S.L.-B., C.B.-D.; visualization. S.L.-B. and C.B.-D. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** The authors thank University of Vigo (Dep. C11) for financial support during the sabbatical leave of C.B.-D.

**Conflicts of Interest:** The authors declare no conflict of interest.
