Testing of pseudopotentials used in classical Car - Parrinello simulations

Abstract
Recently a classical scheme of the Car - Parrinello method was proposed to simulate charged colloidal suspensions described by the highly asymmetric `primitive model' of electrolytes. To make the simulations feasible one has to rely on a pseudopotential picture which smooths out the counterionic density profile near the macroionic surfaces. Using `exact' Monte Carlo simulations of the same model with two macroions only, we critically test the validity of the pseudopotential construction. We find that the effective forces between the macroions and the counterion density fields are satisfactorily described by pseudopotential theory for strongly interacting macroions. However, for a configuration of nearly touching macroions, there are substantial deviations. Variants of linear screening theory lead to less satisfactory comparison with the Monte Carlo data than the ` ab initio' calculations supplemented with the pseudopotential picture. Among all linear screening models, the Poisson - Boltzmann-cell model supplemented with a size correction is found to be the best description.