Matching the Probability Distribution On-the-Fly: OPES-Based Dual-Level Free Energy Calculation

Area of Science:

• Computational Chemistry
• Chemical Physics
• Reaction Dynamics

Background:

• Free energy profiles are crucial for understanding chemical processes.
• Accurate free energy profile calculations in complex systems are computationally demanding.
• The reference potential method (RPM) reduces costs but struggles with dissimilar Hamiltonians.

Purpose of the Study:

• To develop a computationally efficient method for accurate free energy profile calculations.
• To address the slow convergence issues of RPM when low-level and high-level Hamiltonians differ.
• To improve sampling efficiency and accuracy in complex chemical environments.

Main Methods:

• Proposed dual-level OPES, combining on-the-fly probability enhanced sampling (OPES) with RPM.
• Utilized an OPES bias potential for accelerated barrier crossing.
• Employed a probability distribution correction potential (Vcorr) to enhance Hamiltonian overlap.
• Validated on a mathematical model and glycine's intramolecular proton transfer in water.

Main Results:

• Dual-level OPES improved the overlap between sampled and high-level Hamiltonian distributions.
• Accelerated free energy profile convergence observed in the model system compared to direct RPM.
• Significantly reduced free energy profile error for glycine proton transfer.
• Achieved a speedup of tens of times over direct high-level simulations for the glycine case.

Conclusions:

• Dual-level OPES effectively enhances sampling and accuracy in free energy profile calculations.
• The method overcomes limitations of RPM when Hamiltonians are dissimilar.
• Dual-level OPES offers a significant computational advantage for complex chemical reaction studies.