Convergence of Kirkwood-Buff Integrals of Ideal and Nonideal Aqueous Solutions Using Molecular Dynamics Simulations

The computation of Kirkwood-Buff integrals (KBIs) using molecular simulations of closed systems is challenging due to finite system-size effects. One of the problems involves the incorrect asymptotic behavior of the radial distribution function. Corrections to rectify such effects have been proposed in the literature. This study reports a systematic comparison of the proposed corrections (as given by Ganguly et al. J. Chem. Theory Comput. 2013, 9, 1347-1355 and Kruger et al. J. Phys. Chem. Lett. 2013, 4, 4-7) to assess the asymptotic behavior of the RDFs, the KBIs, as well as the estimation of thermodynamic quantities for ideal urea-water and nonideal modified-urea-water mixtures using molecular dynamics simulations. The results show that applying the KBI correction suggested by Kruger et al. on the RDF corrected with the Ganguly et al. correction (denoted as B-KBI) yields improved KBI convergence for the ideal and nonideal aqueous mixtures. Different averaging regions in the running KBIs (correlated or long-range) are assessed, and averaging over the correlated region for large system sizes is found to be robust toward the change in the degree of solvent nonideality and concentration, providing good estimates of thermodynamic quantities. The study provides new insights into improving the KBI convergence, the suitability of different averaging regions in KBIs to estimate thermodynamic properties, as well as the applicability of correction methods to achieve KBI convergence for nonideal aqueous binary mixtures.