A General Stiffness-Scaling Framework for Accelerating Graph-Theoretical Kinetic Monte Carlo Simulations

Journal article

Kinetic Monte Carlo (KMC) simulations are a powerful tool for investigating catalytic reaction mechanisms, yet they often become intractably slow when certain fast, quasi-equilibrated reaction channels fire much more frequently than other processes, a problem known as stiffness. To overcome this issue, we introduce a new reaction channel-based scaling algorithm that dynamically upscales or downscales the rate constants of quasi-equilibrated channels, ensuring they remain within a user-defined time scale window. In contrast to previous methods that either fully restored original rates after nonequilibrated events or applied one-way downscaling (without the option to increase rates toward their original values), our algorithm adaptively regulates each channel throughout the simulation, and can be applied to both simple and highly complex lattice-based KMC models of catalytic systems. We demonstrate the performance of this method on three representative catalytic systems with adsorbate–adsorbate lateral interactions. First, a reverse water–gas shift (RWGS) model on Ni(111) serves as a benchmark where unscaled simulations are feasible and provide a reference for error analysis. Second, a complex and highly stiff model of dry reforming of methane (DRM) on Pt/HfC–containing 119 reversible channels, multiple site types, and 175 energetic clusters–showcases the algorithm’s robustness across a wide range of time scales and operating conditions (e.g., varying p CH4 and p CO2 ). Third, transient simulations of temperature-programmed desorption (TPD) of formate, which entails dissociation, on NiCu single-atom alloys (SAAs) illustrate the method’s ability to adapt to rapid kinetic changes. In all cases, the algorithm is able to significantly accelerate the simulations without introducing substantial error, offering a practical solution to stiffness in KMC studies of catalytic systems. The method is fully integrated into the Zacros code (release of the pertinent version pending), making it broadly accessible.

Authors: Prats, H; Li, W; Stamatakis, M

• Publication status: Published
• Peer review status: Peer reviewed
• Publisher: American Chemical Society
• Journal: Journal of Chemical Theory and Computation
• Volume: 21
• Issue: 23
• Pages: 12262-12277
• Publication date: 2025-11-18
• Acceptance date: 2025-10-28
• DOI: 10.1021/acs.jctc.5c01394
• EISSN: 1549-9626
• ISSN: 1549-9618
• Language: English