Equilibrium Equality for Free Energy Difference

TL;DR

This paper introduces a new equilibrium equality that allows calculating free energy differences between any two equilibrium states using only one canonical ensemble, significantly reducing computational resources needed.

Contribution

It derives a novel equilibrium equality for free energy differences between states with different temperatures and volumes, improving efficiency over previous methods.

Findings

Efficient calculation of FED between states with different temperatures and volumes.

Validated the equality using Toda-Lattice models of various dimensions.

Reduced the number of samples needed for accurate FED estimation.

Abstract

Jarzynski Equality (JE) and the thermodynamic integration method are conventional methods to calculate free energy difference (FED) between two equilibrium states with constant temperature of a system. However, a number of ensemble samples should be generated to reach high accuracy for a system with large size, which consumes a lot computational resource. Previous work had tried to replace the non-equilibrium quantities with equilibrium quantities in JE by introducing a virtual integrable system and it had promoted the efficiency in calculating FED between different equilibrium states with constant temperature. To overcome the downside that the FED for two equilibrium states with different temperature can't be calculated efficiently in previous work, this article derives out the Equilibrium Equality for FED between any two different equilibrium states by deriving out the equality for FED between states with different temperatures and then combining the equality for FED between states with different volumes. The equality presented in this article expresses FED between any two equilibrium states as an ensemble average in one equilibrium state, which enable the FED between any two equilibrium states can be determined by generating only one canonical ensemble and thus the samples needed are dramatically less and the efficiency is promoted a lot. Plus, the effectiveness and efficiency of the equality are examined in Toda-Lattice model with different dimensions.