Descriptors and similarity matrix for article "Density-of-states similarity descriptor for unsupervised learning from materials data" by Martin Kuban, Santiago Rigamonti, Markus Scheidgen, and Claudia Draxl

This repository contains the data used in the publication "Density-of-states similarity descriptor for unsupervised learning from materials data" by Martin Kuban, Santiago Rigamonti, Markus Scheidgen, and Claudia Draxl, preprint: https://arxiv.org/abs/2201.02187.

Description of files

• dos_data_and_fingerprints.json

  This json file contains the following data

  • Total electronic densities of states (TDOS) of materials in the C2DB [1-3]. Data IDs are correspond with the C2DB.
  • Periodic table of elements (PTE) descriptors
  • Space group (SG) descriptors

• DOS_similarity_matrix.csv

  • Similarity matrix. Data IDs correspond with the C2DB.
  • Rows and columns contain similarity values between materials
  • To load the data to a Python object, you can use, e.g., pandas.read_csv()

Both files are compressed and need to be expanded first, e.g. using unzip.

• calculate_descriptors.py

  Python script to calculate the PTE and SG desciptor values. See Notes below

• calculate_dos_fingerprints.py

  Python script to compute DOS fingerprints. See Notes below

Notes

Creation of ASE Atoms database for running script calculate_descriptors.py

To run the script calculate_descriptors.py, an ASE database with atomic structure is required. The data required for this can be downloaded from the C2DB website [3]. To create this required c2db.db file, visit the C2DB website [3] and click on Browse data. In the following search interface, modify the search parameters to include all materials. For each of the found materials, click on the button Download below the crystal structure viewer. The atomic simulation environment (ASE) [4] can be used to transform the downloaded structure files to the database file: Import each file using the ase.io module and write it to a ase.db.core.Database object.

How to obtain the DOS fingerprints

The DOS fingerprints can be obtained from file, by loading dos_data_and_fingerprints.json and using the nomad_dos_fingerprint.DOSFingerprint.from_dict() method. Alternatively, they can be generated using the script calculate_dos_fingerprints.py. The execution requires a Python package, which can be downloaded from [4].

References

[1] "The Computational 2D Materials Database: High-Throughput Modeling and Discovery of Atomically Thin Crystals" Sten Haastrup, Mikkel Strange, Mohnish Pandey, Thorsten Deilmann, Per S. Schmidt, Nicki F. Hinsche, Morten N. Gjerding, Daniele Torelli, Peter M. Larsen, Anders C. Riis-Jensen, Jakob Gath, Karsten W. Jacobsen, Jens Jørgen Mortensen, Thomas Olsen, Kristian S. Thygesen 2D Materials 5, 042002 (2018)

[2] "Recent Progress of the Computational 2D Materials Database (C2DB)" M. N. Gjerding, A. Taghizadeh, A. Rasmussen, S. Ali, F. Bertoldo, T. Deilmann, U. P. Holguin, N. R. Knøsgaard, M. Kruse, A. H. Larsen, S. Manti, T. G. Pedersen, T. Skovhus, M. K. Svendsen, J. J. Mortensen, T. Olsen, K. S. Thygesen 2D Materials 8, 044002 (2021)

[3] https://cmr.fysik.dtu.dk/c2db/c2db.html

[4] https://github.com/kubanmar/dos-fingerprints