(303g) Fluctuating Charge Approach for Modeling Polarizable Electrodes in Classical Molecular Simulations

Atomistic simulations of electrolytes and other soft-matter systems at charged surfaces is a challenging problem. These systems usually require accurate treatment of polarization effects (both in electrolyte and in the electrode) as well as necessity to deal with long length and time scales.   For example, electrolyte containing Li-salts require explicit inclusion of electronic polarizabilities (via induced dipoles or Drude oscillators) for reliable predictions of Li-ion solvation structure and transport properties.  The structure and dynamics of electrolyte near charged (due to externally applied voltage) electrode surface are strongly coupled to the details and fluctuations of the charge distribution on the electrode surface. Many simulation methods assign charges uniformly to all electrode atoms and keep them constant during simulations. However, such approach might not correctly capture the mutual coupling between electrode and electrolyte polarization. The particular scheme employed to assign charges on electrode atoms used in simulations have to correspond the physical nature of the electrode. For example, in order to model the electrode as classical conductors, the charges on electrode should be distributed such that the same electrostatic potential is obtained at every location in the electrode. In contrast, in order to capture the semiconducting character of graphene additional terms have to be added to the free energy to correctly capture the heterogeneity in charge distribution in the electrode and its coupling with electrolyte as a function of applied voltage.  In this presentation we will discuss new technique to model an electrode with fluctuating charge approach. A version of fluctuating electrode charge approach that does not require electrode charges to be Gaussian distributed will be presented. Such approach is easier to implement into standard simulation codes. A comparison of results and efficiency between simulations using the constant charge and constant potential methods will be discussed.