Ionic diffusion in naturally-occurring aqueous solutions: use of activity coefficients in transition-state models

Abstract

The model of Lane and Kirkaldy ( Can. J. Chem. 43, 1812–1828, 1965; 44, 477–485, 1966) for estimating diffusion coefficients in aqueous electrolyte solutions from limiting ionic conductances is rederived for a system of n salts with a common anion, by utilizing the equations of Miller ( J. Chem. Phys. 71, 616–632, 1967b; 71, 3588–3592, 1967c). Salt and water activities are used, and it is of necessity assumed that the diffusion mechanism does not change with concentration. The revised model predicts the on-diagonal D v ik of a 3 M solution in NaCl-KCl-H 2O to within 5%, compared with errors of 20–30% for the original model. Errors remain constant or decrease as concentrations increase to 3 M, so that predictions of these D v ik at even higher concentrations appear promising. Relatively large errors persist in our estimates of the small, off-diagonal coefficients in this system. Measured diffusion coefficients in MgCl 2-NaCl-H 2O extend over only a limited concentration range and are of only moderate accuracy. Nevertheless, the revised model predicts D v 11 D v 22, and D v 21, the larger of the off-diagonal coefficients, with errors of only 5–20%.