Parametrizing linear generalized Langevin dynamics from explicit molecular dynamics simulations.

Area of Science:

• Computational physics
• Statistical mechanics
• Atomistic simulations

Background:

• Understanding complex dynamics in many-particle systems is crucial.
• Macroscopic systems often involve key degrees of freedom interacting with a thermal bath.
• The linear generalized Langevin equation simplifies bath interactions via a memory kernel.

Purpose of the Study:

• To develop a more accurate method for parametrizing the memory kernel in generalized Langevin equations.
• To compare frequency-domain and time-domain approaches for this task.
• To evaluate the suitability of the rigid bond method.

Main Methods:

• Derivation of the linear generalized Langevin equation using linear projection.
• Development of a Fourier-based (frequency-domain) parametrization method.
• Comparison with traditional time-domain methods and the rigid bond approach.

Main Results:

• The frequency-domain method outperforms time-domain analogues.
• The widely used rigid bond method is generally inappropriate.
• The rigid bond approach systematically overestimates relaxation times except in specific harmonic bath scenarios.

Conclusions:

• A Fourier-based frequency-domain approach offers superior parametrization of memory kernels.
• The rigid bond method should be used with caution due to potential overestimation of relaxation times.
• Accurate modeling of many-particle dynamics requires careful selection of parametrization methods.