SilentDan44/mof-structural-database

---

license: cc-by-4.0 pretty_name: MOF Structural Database tags: - materials-science - metal-organic-frameworks - porous-materials - crystal-structures - zeo++

MOF Structural Database

A consolidated SQLite database of metal-organic framework (MOF) structures and structural descriptors. This snapshot combines source records from ARC-MOF, the public CoRE MOF 2019 dataset, and QMOF, and adds a common schema, structure fingerprints, MOF-family/group metadata, and Zeo++ pore properties.

This release contains structural and pore-property data only. It does not contain adsorption isotherms, NIST adsorption records, CIF file contents, or the isotherm-to-structure crosswalk.

Snapshot

• 281,970 material records
• 242,517 records from ARC-MOF
• 19,081 records from CoRE MOF
• 20,372 records from QMOF
• 261,866 records with Zeo++ pore properties; 20,104 records have no Zeo++ result in this snapshot
• 281,970 records with a composition-rdf-v1 structure fingerprint
• SQLite file: mof-public.db.gz (about 190 MB compressed; about 1.2 GB uncompressed)

Counts refer to source records in the merged database; they should not be interpreted as counts of unique structures. The database includes identifiers for canonical records, duplicate groups, and MOF families where available.

Database schema

The SQLite database contains one table, materials. Each row represents a material record and includes:

• Identifiers and provenance: material_id, source_database, source_material_id, source_url, source_filename, doi, external_ids
• Basic composition and structure labels: canonical_name, formula, elements, structure_hash
• Fingerprints: structure_fingerprint, structure_fingerprint_version
• Pore properties: pore_properties, pore_properties_version, pore_properties_updated_at
• Grouping: duplicate_group_id, canonical_material_id, mof_family_id, mof_family_key, mof_family_method
• Parsing and source metadata: parse_status, metadata_json, created_at, updated_at
• cif_path: a path from the build environment. It is not a downloadable CIF and is not expected to resolve outside that environment.

The Zeo++ properties include largest included sphere, pore-limiting diameter, largest sphere along the free path, unit-cell volume, density, accessible and non-accessible surface areas, and probe-occupiable volumes and void fractions. Values are stored as JSON in pore_properties, with units indicated by the JSON key names.

The recorded Zeo++ configuration uses probe radius 1.86 Å, high-accuracy mode, 2,000 surface samples, and 20,000 volume samples. The recorded property version is zeopp-99d004b8-v1. The fingerprint version is composition-rdf-v1.

Files and use

Download mof-public.db.gz from this repository and decompress it:

gzip -dk mof-public.db.gz

Query it with Python's built-in SQLite support:

import sqlite3

con = sqlite3.connect("mof-public.db")
rows = con.execute(
    "SELECT material_id, formula, source_database, pore_properties "
    "FROM materials LIMIT 5"
).fetchall()

The database contains CIF path strings and source references, but not the CIF files themselves. Some source structures may have additional access or redistribution conditions at their original providers; consult the cited source records before retrieving or redistributing those structure files.

Data sources and attribution

Please cite the source datasets and papers when using this compilation:

1. ARC-MOF — Jake Burner et al., ARC–MOF: A Diverse Database of Metal-Organic Frameworks with DFT-Derived Partial Atomic Charges and Descriptors for Machine Learning, Chemistry of Materials 35 (2023), 900–916. Paper DOI; ARC-MOF data record, version v7.
2. CoRE MOF 2019 — Yongchul G. Chung et al., Advances, Updates, and Analytics for the Computation-Ready, Experimental Metal–Organic Framework Database: CoRE MOF 2019, Journal of Chemical & Engineering Data 64 (2019), 5985–5998. Paper DOI; public CoRE MOF data. The public 2019 dataset is identified as CC BY 4.0 by the CoRE-MOF project.
3. QMOF — Andrew S. Rosen et al., Machine Learning the Quantum-Chemical Properties of Metal–Organic Frameworks for Accelerated Materials Discovery, Matter 4 (2021), 1578–1597, and High-Throughput Predictions of Metal–Organic Framework Electronic Properties, npj Computational Materials 8 (2022), 112. 2021 paper DOI; 2022 paper DOI; QMOF project and license statement.

The source datasets state CC BY 4.0 terms for the data used here. This compilation is released under Creative Commons Attribution 4.0 International (CC BY 4.0). Credit the original sources, link to the license, and indicate changes when reusing the data. See the CC BY 4.0 deed.

Limitations

• The source databases have different curation rules and may contain related, duplicate, or variant structures. Use the record IDs and grouping fields when comparing entries.
• Zeo++ properties are absent for 20,104 records in this snapshot. Check pore_properties_version before using these fields.
• CIF files are not bundled. The cif_path field is build-environment-specific; use the source identifiers and source links to locate records at their original providers.
• No isotherms, adsorption measurements, NIST material registry entries, or matching crosswalk are included.
• The presence of a record in this database is not an independent validation of its structure or computed properties.

License

This compiled dataset is distributed under CC BY 4.0, with attribution to the source datasets and papers listed above. Third-party structure files are not included; consult their original providers' terms before obtaining or redistributing them.