Viscosity Calculations with Hybrid Particle-Field Molecular Dynamics Simulations

We calculate fluid viscosities using hybrid particle-field (HPF) molecular dynamics (MD) simulations. Due to the absence of explicit particle collisions, HPF by itself exhibits artifacts, resulting in nonlinear velocity profiles in nonequilibrium MD simulations of shear flow. However, integrating multiparticle collision dynamics (MPCD) into HPF successfully mimics collision dynamics, producing linear velocity profiles that closely match those of reference Lennard-Jones (LJ) simulations of the same system, which use pair interactions. By tuning the MPCD collision period T_col, we can adjust the viscosity η of a given state point. We present a general equation of the viscosity η(ρ, T), as a function of the density ρ and the temperature T, and its relationship with T_col, demonstrating a weak dependence of HPF-MPCD systems on temperature and density compared to LJ fluids. Our results highlight the scalability and versatility of HPF and HPF-MPCD methods for large-scale, high-sampling simulations, offering a compelling alternative to traditional MD for simulating complex fluid systems. We demonstrate that HPF and HPF-MPCD models significantly outperform traditional MD in terms of computational efficiency, achieving up to 3 orders of magnitude faster simulations. Momentum conservation analysis shows that the 8-point central difference gradient interpolation method provides superior accuracy and stability compared to staggered lattice interpolation, making it ideal for long-term simulations.