Predicting low-temperature free energy landscapes with flat-histogram Monte Carlo methods

Area of Science:

• Thermodynamics
• Computational Chemistry
• Fluid Dynamics

Background:

• Predicting fluid thermodynamic properties at low temperatures is computationally intensive.
• Flat-histogram grand canonical Monte Carlo simulations are powerful but can be inefficient over wide temperature ranges.

Purpose of the Study:

• To develop a computationally efficient method for predicting fluid free energy landscapes at low temperatures.
• To demonstrate the method's applicability to pure and multicomponent systems undergoing various phase transitions.

Main Methods:

• Extrapolating results from high-temperature flat-histogram grand canonical Monte Carlo simulations to lower temperatures.
• Utilizing two distinct sampling methods for demonstration purposes.

Main Results:

• The extrapolation method quantitatively predicts thermodynamic properties for systems within a specific temperature difference range.
• Beyond this range, extrapolation provides a well-informed estimate, reducing subsequent simulation refinement effort.
• A binary fluid phase diagram was quantitatively predicted from a single supercritical state simulation.

Conclusions:

• The proposed method significantly enhances the computational efficiency of flat-histogram simulations for investigating fluid thermodynamic properties.
• This approach enables accurate predictions of phase transitions and thermodynamic behavior across wide temperature ranges with reduced computational cost.