Physical and numerical aspects of sodium ion solvation free energies via the cluster-continuum model. Issue 48 (5th December 2022)

• Record Type: Journal Article
• Authors: Otlyotov, Arseniy A.; Itkis, Daniil; Yashina, Lada V.; Cavallo, Luigi; Minenkov, Yury
• Is Part Of: Physical chemistry chemical physics. Volume 24:Issue 48(2022)
• Page Start: 29927
• Page End: 29939
• Publication Date: 2022-12-05
• Subjects: Chemistry, Physical and theoretical -- Periodicals; 541.3
• DOI: 10.1039/d2cp03583a
• Languages: English
• ISSNs: 1463-9076

Abstract

Abstract : Na + solvation free energies are obtained in 8 solvents through the cluster-continuum approach. The Pearson correlation coefficient between the predicted and experimental data is 0.82, and the mean signed/unsigned errors are 0.3/1.4 kcal mol −1 . Abstract : Sodium cation solvation Gibbs free energies (Δ G solv (Na + )) have been obtained in water, dimethylformamide, dimethyl sulfoxide, ethanol, acetone, acetonitrile, and methanol through the "monomer cycle" cluster-continuum approach where a solvent reference state is described by infinitely separated molecules. The following steps are vital for obtaining reliable Δ G solv (Na + ) values: (a) a meticulous conformational search involving dispersion corrected density functional theory (DFT-D) and the continuum solvation model (CSM); (b) gas-phase DFT-D geometry optimization followed by single-point (SP) domain-based local pair natural orbital coupled clusters including single, double, and partly triple excitation (DLPNO-CCSD(T)) calculations in conjunction with the complete basis set extrapolation; (c) advanced statistical thermodynamic treatment of the low harmonic frequencies (<100 cm −1 ) to obtain the robust gas-phase Gibbs free energy correction; (d) gas-phase and dielectric continuum SP with non-electrostatic contributions included in the CSM; (e) an evaluation of the relative thermodynamic stability of the Na + (S) n clusters to identify the number of explicit solvent molecules n to be considered. Our refined computational protocol is promising with a Pearson correlation coefficient between the predicted and experimental data, ρ, of 0.82, and the mean signed and mean unsigned errors of 0.3 and 1.4 kcal mol −1, respectively.