Transport coefficients of electrolyte solutions from Smart Brownian dynamics simulations

We present results of Brownian dynamics simulations of aqueous 1-1 electrol yte solutions in the 1-molar concentration range. The electrical conductivi ty and the self-diffusion coefficients obtained from the simulations are co mpared to experimental data. The interaction potential between the ions is modeled by pairwise repulsive 1/r(n) soft-core interactions (n = 9 or n = 1 2) and Coulomb forces. We take into account hydrodynamic interactions and i ntegrate the stochastic equations of motion with large time steps of about 100 femtoseconds, combined with an acceptance criterion known from the Smar t Monte Carlo method. In this way, details of the dynamics of particles in close contact are not considered and the short-ranged repulsive forces act effectively as constraint forces preventing overlap configurations. The len gths of the performed simulations (about 10 nanoseconds) and the number of ions (216) allow to obtain single particle as well as collective transport coefficients with sufficient precision. For this purpose we use Kubo expres sions which can be applied on the mesoscopic time scale of Brownian dynamic s simulations. It is shown that hydrodynamic interactions must be taken int o account to obtain agreement with the experimental data. They lower the el ectrical conductivity, as expected, but increase the self-diffusion coeffic ients, confirming a recent finding for colloids. (C) 1999 American Institut e of Physics. [S0021-9606(99)51415-3].