Using neural network potential (NNP), we defined local surface energy (LSE). LSE is a metric that reflects local reactivity at an atomic resolution.
In this study, we use universal NNP M3GNet.
Using this metric LSE, you can predict CO adsorption energy on high entropy alloy surfaces.
We provide a user-friendly container environment with Docker and Jupyter Lab to avoid cumbersome dependencies and environment construction costs.
Execute the following command in the directory where docker-compose.yml is located. A docker image will be created, referring to the Dockerfile.
docker-compose build
Run the following command in the directory containing docker-compose.yml.
docker-compose up -d
Then, run the following command in some browser. you can connect to the Docker container through Jupyter Lab. The settings are in docker-compose.yml.
http://localhost:18080
You can run our tutorial notebook in tutorial
We offer a sample code to generate atomic energy.
We offer a sample code to create nanoparticle.
1000 HEA nano particles structure data are hosted on figshare.
T. Shiota, K. Ishihara, W. Mizukami, Lowering the Exponential Wall: Accelerating High-Entropy Alloy Catalysts Screening using Local Surface Energy Descriptors from Neural Network Potentials ,
arXiv:2402.18433 [quant-ph]
Chen, C., Ong, S.P. A universal graph deep learning interatomic potential for the periodic table. Nat Comput Sci 2, 718–728 (2022). https://doi.org/10.1038/s43588-022-00349-3.