Lowering the Exponential Wall: Accelerating High-Entropy Alloy Catalysts Screening using Local Surface Energy Descriptors from Neural Network Potentials
Abstract
Computational screening is indispensable for the efficient design of high-entropy alloys (HEAs), which hold considerable potential for catalytic applications. However, the chemical space of HEAs is exponentially vast with respect to the number of constituent elements, making even machine learning-based screening calculations time-intensive. To address this challenge, we propose a rapid method for predicting HEA properties using data from monometallic systems (or few-component alloys). Central to our approach is the newly introduced local surface energy (LSE) descriptor, which captures local surface reactivity at atomic resolution. By applying linear regression, we successfully screened adsorption energies of molecules on HEAs based on LSEs derived from monometallic systems. Furthermore, we developed high-precision models by employing both classical and quantum machine learning. Our method enabled CO adsorption-energy calculations for 1000 quinary nanoparticles, comprising 201 atoms each, within a few days, considerably faster than density functional theory, which would require hundreds of years or neural network potentials, which would have taken hundreds of days. The proposed approach accelerates the exploration of the vast HEA chemical space, facilitating the design of novel catalysts.
AI-Generated Overview
-
Research Focus: The study focuses on developing a computational methodology to predict the molecular adsorption energies of high-entropy alloys (HEAs) using local surface energy (LSE) as a descriptor derived from neural network potentials (NNPs), addressing the complex chemical space of HEAs.
-
Methodology: The authors introduce a new LSE descriptor that captures the local surface reactivity at atomic resolution. They utilize a regression model based on monometallic adsorption data to predict CO adsorption energies on HEA nanoparticles, employing classical and quantum machine learning methods to enhance predictive accuracy.
-
Results: The methodology achieved a strong correlation between the LSE and CO adsorption energies, enabling rapid calculations for 1000 HEA nanoparticles, drastically reducing computation time from years with DFT methods to just a few days.
-
Key Contribution(s): The study provides a novel LSE descriptor that effectively links atomic-level reactivity to adsorption energy predictions, allowing for efficient screening of HEA compositions to identify promising catalytic materials.
-
Significance: This approach substantially accelerates the computational exploration of HEAs, addressing the limitations of conventional methods and facilitating the discovery of novel catalysts with improved performance.
-
Broader Applications: The developed methodology can be applied to optimize catalyst design for various chemical reactions, extend to other material systems, and incorporate environmental factors for enhanced predictive capabilities in materials science research.