Simulating an Applied Voltage in Molecular Dynamics Using Charge Optimized Many Body (COMB3) Potentials

Technological development of energy sources such as fuel cells and batteries has been enhanced by the progress of computational models and simulations used to study these electrochemical systems, more specifically, the electrode/electrolyte interface.  Here, we present a methodology for applying a voltage during a dynamic charge molecular dynamics (MD) simulation.  Our method is developed within the third generation charge optimized many body (COMB3) potential [1] framework and implemented in the Large-scale Atomic/Molecular Massively Parallel Simulator (LAMMPS) software package [2].  The COMB3 potential incorporates the electronegativity-equalization (EE) principle in its dynamic charge equilibration scheme.  The EE principle requires the electronegativity at all atomic sites be equal, assuming fixed atomic positions.  The partial charge on each atom is dynamically equilibrated by solving Euler-Lagrange equations of motion, a method developed by Rick et al. [3].  Only changes in the charge equations of motion are shown for brevity, but a complete description of the charge equilibration steps in COMB3 can be found in ref. 1.  Electrode COMB3 (ECOMB3) development includes a simplistic view of varying the atomic charge equations of motion during the charge equilibration steps to simulate an applied voltage for various electrochemical systems.  In ECOMB3, the mean electronegativities at the cathode and anode differ by an applied voltage, V.  This electronegativity imbalance generates a new set of charge equations of motion for the anode and cathode while the electrolyte (non-electrode) equation of motion remains unchanged.  Using the ECOMB3 method, we simulate an applied voltage for a system of liquid water between two Cu electrodes.  The aim of these simulations is to study the differing dynamical and structural properties of water molecules at the Cu/H₂O interface as opposed to molecules in the bulk.  More specifically, we analyze the charge density distribution across the interface to demonstrate the practicality of using a charge transfer scheme in such a simulation.  We also report density distribution profiles for atomic oxygen and hydrogen in water, indicating the ordering of molecules near the interface.

References

[1]         T. Liang et al., Materials Science & Engineering R-Reports 74(2013).

[2]         S. Plimpton, Journal of Computational Physics 117(1995).

[3]         S. W. Rick, S. J. Stuart, and B. J. Berne, Journal of Chemical Physics 101 (1994).