Solvent property-ion conductivity relationship for lithium, sodium and potassium ions in non-aqueous solvents using QSER

Abstract

Non-aqueous solvents are applicable in industrial and chemical reactions as well as some analytical methods such as electrochemistry. Application of quantitative structure-electrochemistry relationship (QSER) modeling in non-aqueous media can reduce trial-and-error experiments and more accurate estimation of target properties with lower number of runs. In the current study, for the first time, QSER was applied in conductometry as a wide field in electrochemistry. The target properties for modeling were the limiting molar conductivities of three alkali metal ions (Li+, Na+, K+) in water and various organic solvents; and empirical scales of the utilized solvents were applied as the independent set of variables. R2 in training sets of Li+, Na+ and K+ models were 0.94, 0.95 and 0.90 respectively. Models also showed good prediction ability and R2 in test sets of Li+, Na+ and K+ models were 0.95, 0.94 and 0.98 respectively. In addition, more statistical evaluations were done to show the predictability and stability of proposed QSERs. One of the advantages of suggested QSER models based on solvent empirical scales is the clear meaning of the solvent scales and their ability in description of solute-solvent interactions. Based on obtained results, the impact of some solvent features such as electron-donating power, electron-donating power, dipolarity-dipolarizability, hydrogen bonding, hydrogen-bond acceptor basicity and solvent acceptor number on the conductivity of lithium, sodium and potassium ions were discussed.