Nuclear quantum effects on zeolite proton hopping kinetics explored with machine learning potentials and path...

Area of Science:

• Catalysis
• Computational Chemistry
• Materials Science

Background:

• Proton hopping is crucial in zeolite catalysis, but its kinetics are challenging to determine.
• Nuclear quantum effects (NQEs) influence proton dynamics, yet their inclusion in zeolite catalysis is computationally prohibitive.
• Existing methods struggle with the high computational cost of Density Functional Theory (DFT) for simulating these processes.

Purpose of the Study:

• To develop a computationally efficient method for accurately simulating proton hopping kinetics in zeolites.
• To rigorously include nuclear quantum effects (NQEs) in the study of proton transfer reactions within zeolites.
• To enable accurate calculations of kinetic isotope effects for proton and deuterium hopping.

Main Methods:

• Trained a reactive machine learning potential (MLP) to accurately represent the DFT potential energy surface for proton hopping.
• Employed path integral molecular dynamics (PIMD) simulations utilizing the trained MLP for extensive calculations.
• Achieved over 0.6 μs of simulation time, far exceeding standard DFT capabilities.

Main Results:

• The developed MLP accurately reproduces DFT-level accuracy for proton hopping in H-CHA zeolite.
• NQEs were found to significantly impact proton hopping kinetics up to approximately 473 K.
• PIMD simulations with deuterium were performed without retraining, enabling kinetic isotope effect calculations.

Conclusions:

• The combination of MLPs and PIMD overcomes previous computational limitations for studying NQEs in zeolite catalysis.
• This approach provides accurate reaction kinetics and enables reliable kinetic isotope effect predictions.
• The method is applicable to a broad range of temperatures and isotopic substitutions.