Dataset Card for Perovs-Dopants

Dataset Summary

The Perovs-Dopants dataset is a computational dataset designed for benchmarking machine learning interatomic potentials in defect perovskite materials. It includes density functional theory (DFT) geometry optimization calculations of over 400 perovskites with neutral substitutional defects.

Dataset Details

Creators: Intel Labs
Version: 1.0
License: CDLA Permissive 2.0 License
Number of Training Samples: 22034 DFT data points
Number of Validation Samples: 2744 DFT data points
Number of Test Samples: 2744 DFT data points
Format: xyz files

Intended Use

Primary Uses: The Perovs-Dopants dataset is intended for benchmarking machine learning interatomic potentials for defect perovskite materials, and assessing model transferability across chemical spaces. It supports high-throughput virtual screening efforts to identify promising defect perovskite candidates for functional applications.
Out-of-Scope Uses: Out-of-scope use include materials that are not included in the dataset.

Data Collection Process

The Perovs-dopants dataset was constructed by first selecting base perovskite materials from Materials Project [1], a oxide perovskite database [2], and a halide perovskite database [3], and substituting a single atom at either A-site or B-site of the perovskite structure. A-site and B-site vacancies were also included. For each defect structure, atomic configurations were optimized using CP2K with the PBE functional, and the relaxation process continued until forces on all atoms were below 0.02 eV/A.

Perovs-Dopants Dataset Description — (a) Distribution of elements in the defect dataset. A and B site elements, and X site elements are shown in the left figure, and the substituent elements are shown in the right figure. (b) t-SNE plot comparing the chemical space covered by the MPtrj and Perovs-Dopants datasets. Reproduced from "Evaluating Machine Learning Potentials on Bulk Structures with Neutral Substitutional Defects." [4]

Example: For inspecting the data, you can use common computational chemistry like ASE. Using ase==3.2.5.0, for example:

from ase.io import read

read("dopant_train.xyz", index=":")

[1] Anubhav Jain, Shyue Ping Ong, Geoffroy Hautier, Wei Chen, William Davidson Richards, Stephen Dacek, Shreyas Cholia, Dan Gunter, David Skinner, Gerbrand Ceder, and Kristin A. Persson. Commentary: The materials project: A materials genome approach to accelerating materials innovation, 2013. ISSN 2166532X.

[2] Ivano E. Castelli, David D. Landis, Kristian S. Thygesen, Soren Dahl, Ib Chorkendorff, Thomas F. Jaramillo, and Karsten W. Jacobsen. New cubic perovskites for one- and two-photon water splitting using the computational materials repository. Energy and Environmental Science, 5:9034–9043, 10 2012. ISSN 17545692. doi: 10.1039/c2ee22341d.

[3] Christopher P. Muzzillo, Cristian V. Ciobanu, and David T. Moore. High-entropy alloy screening for halide perovskites. Materials Horizons, 5 2024. ISSN 20516355. doi: 10.1039/d4mh00464g.

[4] Wang X, Park S, Lee K L K, Kurchin R C, Miret S. Evaluating Machine Learning Potentials on Bulk Structures with Neutral Substitutional Defects. In AI for Accelerated Materials Design-ICLR 2025 2025.

Citations

Dataset: If you use Perovs-Dopants in your technical work or publication, please cite relevant papers. For the dataset itself:

Wang X, Park S, Miret S. Perovs-Dopants: Machine Learning Potentials for Doped Bulk Structures. In AI for Accelerated Materials Design-NeurIPS 2024 2024.

@inproceedings{
wang2024perovsdopants,
title={Perovs-Dopants: Machine Learning Potentials for Doped Bulk Structures},
author={Xiaoxiao Wang and Suehyun Park and Santiago Miret},
booktitle={AI for Accelerated Materials Design - NeurIPS 2024},
year={2024},
url={https://openreview.net/forum?id=sEpHuS8CWQ}
}

Analysis: If you utilize or find our deeper analysis of machine learning models for doped materials useful, please cite:

Wang X, Park S, Lee K L K, Kurchin R C, Miret S. Evaluating Machine Learning Potentials on Bulk Structures with Neutral Substitutional Defects. In AI for Accelerated Materials Design-ICLR 2025 2025.

@inproceedings{
wang2025evaluating,
title={Evaluating Machine Learning Potentials on Bulk Structures with Neutral Substitutional Defects},
author={Xiaoxiao Wang and Suehyun Park and Kin Long Kelvin Lee and Rachel C. Kurchin and Santiago Miret},
booktitle={AI for Accelerated Materials Design - ICLR 2025},
year={2025},
url={https://openreview.net/forum?id=WJnJHp741K}
}

Ethical Considerations

Intel is committed to respecting human rights and avoiding causing or contributing to adverse impacts on human rights. See Intel’s Global Human Rights Principles. Intel’s products and software are intended only to be used in applications that do not cause or contribute to adverse impacts on human rights.

Contact Information

Issues: For any issues or questions regarding the dataset, please contact the maintainers or open an issue in the dataset repository.