Ionic conductivity of molten alkali-metal carbonates A(2)CO(3) (A = Li, Na, K, Rb, and Cs) and binary mixtures (Li1-xCsx)(2)CO3 and (Li1-xKx)(2)CO3: A molecular dynamics simulation

T Kiyobayashi and T Kojima and H Ozaki and K Kiyohara, JOURNAL OF CHEMICAL PHYSICS, 151, 074503 (2019).

DOI: 10.1063/1.5109912

Based on experimental data, we optimized the potential parameters for the classical molecular dynamics simulation to reproduce the volume and ionic conductivity of the molten alkali-metal carbonates A(2)CO(3) where A = Li, Na, K, Rb, and Cs at T/K = 1223 and ambient pressure. The force field was then applied to the binary mixtures (Li1-xCsx)(2)CO3 and (Li1-xKx)(2)CO3. In (Li1-xCsx)(2)CO3, the diffusion coefficient D-Cs exceeds D-Li at x > 0.6, testifying to the Chemla effect. The net ionic conductivity was broken down into the contributions from the velocity auto- and cross-correlations of each ionic species. The significant negative deviation of the real conductivity of (Li1-xCsx)(2)CO3 from the one estimated by the Nernst-Einstein (NE) relation is clearly explained by the contribution from the cross correlations; specifically, the cross term between Li(+)and CO32-, which is negative at x = 0, significantly shifts to the positive side when x increases, which is dominantly responsible for dampening the conductivity from the NE conductivity. A similar behavior was observed in (Li1-xKx)(2)CO3 with a less pronounced manner than in (Li1-xCsx)(2)CO3. These observations corroborate the precedent studies pointing to the trapping of Li+ by the anion when a lithium salt is mixed with another salt of which the cation size is greater than that of Li+.

Return to Publications page