“…Nevertheless, a rationalization of the p K a values of a series of related compounds can be successfully attempted in terms of factors, such as bond polarization and/or relative stability of the protonated and deprotonated species. Both factors can be conveniently expressed by molecular descriptors, attainable at different levels of theory, which, in turn, can be used in a linear free energy relationship (LFER) analysis to estimate p K a values. ,,− These descriptors have been previously calculated at both low level of theory (empirical, AM1, HF/STO-3G, HF/3-21G, HF/3-21G(d)) ,,,− ,,− and Hartree–Fock (HF) or DFT methods with larger basis sets. Although, as a general rule, the highest possible computational methods have to be adopted, there are several examples where the use of HF or DFT methods with larger basis sets does not necessarily lead to significant improvements in correlations with the experimental data. − ,, It has also been demonstrated that the atomic charge densities on the α-hydrogens, δ(H α ), can be used to monitor even small variations of acidity as a function of conformational and/or configurational changes, when these changes influence the degree of polarization of the bonds involved in the dissociation process. ,,, Thus δ(H α ) values have been calculated and used to predict the p K a of the investigated ketones as previously reported. , As a first step, we have determined the most representative conformers of the keto form of the experimentally investigated 2-nitrocycloalkanones both at the semiempirical AM1 and B3LYP/6-31G(d,p) levels of theory.…”