Application of Nosé–Hoover dynamics for coarse-graining molecular systems: An evaluation of reproducibility in Lennard-Jones systems

This study proposes an application of Nosé-Hoover (NH) dynamics as a coarse-graining (CG) method for molecular simulations, offering an alternative to traditional Langevin-based approaches. The NH dynamics, known for its deterministic temperature control without stochastic forces, is adapted here to model a monoatomic Lennard-Jones system at different coarse-grained levels. The CG particle's equation of motion is derived from atomic-level dynamics, linking NH thermostat terms with system properties obtained from molecular dynamics (MD) simulations. Key parameters, including the thermostat coefficient and thermal inertia, are calibrated using the MD data to assess their impact on the dynamic and structural reproducibility of the CG model. The calibration results suggest the potential application of NH dynamics as a coarse-graining method. The effectiveness of the proposed method is then evaluated through a set of CG simulations. The CG results show stable energy regulation and promising accuracy in reproducing system properties, particularly for mass diffusion, with opportunities for further refinement in representing structural reproducibility and momentum diffusion.