Fast event-driven simulations for soft spheres: from dynamics to Laves phase nucleation

Area of Science:

• Computational physics
• Materials science
• Statistical mechanics

Background:

• Conventional molecular dynamics (MD) simulations face limitations with steeply varying interaction potentials due to short time step requirements.
• Simulating systems with potentials like the Weeks-Chandler-Andersen (WCA) potential necessitates careful time step selection in MD to balance accuracy and speed.

Purpose of the Study:

• To evaluate the event-driven Monte Carlo (EDMC) approach as a superior alternative to MD for simulating systems with steep interaction potentials.
• To demonstrate the capability of EDMC in accurately reproducing static thermodynamic properties and system dynamics.

Main Methods:

• Utilized an event-driven Monte Carlo (EDMC) simulation method, building upon the work of Peters and de With.
• Applied EDMC to systems interacting via the Weeks-Chandler-Andersen (WCA) potential in the canonical ensemble.
• Compared EDMC performance against time-driven MD simulations, particularly at low temperatures.

Main Results:

• EDMC accurately reproduces static thermodynamic properties and closely mimics system dynamics for the WCA potential.
• EDMC simulations run over an order of magnitude faster than MD at low temperatures.
• The absence of a finite time step in EDMC eliminates the accuracy-speed trade-off inherent in MD.

Conclusions:

• EDMC is an efficient and accurate method for simulating systems with steep potentials, outperforming traditional MD.
• The EDMC approach enables exploration of phase behavior at extremely low temperatures.
• Spontaneous nucleation and growth of Laves phases were observed at significantly lower temperatures than previously reported using EDMC.