README.md

metadata

license: cc-by-4.0
pretty_name: QMOF
tags:
  - chemistry
  - materials-science
  - metal-organic-frameworks

QMOF

The QMOF Database (Rosen et al., 2021 and 2022), version 18: metal-organic frameworks and coordination polymers relaxed with PBE-D3(BJ). It is the database of the MOF Explorer of the Materials Project.

Structures	20,372
Atoms	2,312,743, 17 to 500 per structure (114 on average, median 96)
Elements	79, from H to Pu
With a band gap at HLE17 / HSE06* / HSE06	10,756 / 10,811 / 10,810

Labels

Field	Source	Unit
total_energy	outputs.pbe.energy_total	eV, of the cell
band_gap	outputs.pbe.bandgap	eV

formation_energy_per_atom and energy_above_hull are NaN.

properties[...]	Source	Unit
band_gap_hle17, band_gap_hse06_10hf, band_gap_hse06	outputs.<theory>.bandgap	eV, NaN when not computed
energy_vdw	outputs.pbe.energy_vdw, the D3(BJ) part of the energy	eV
net_magmom	outputs.pbe.net_magmom	μB
pld, lcd	pore-limiting and largest cavity diameters (Zeo++)	Å
density	info.density	g/cm³
synthesized	1 if the structure comes from an experimental database	0 or 1

Things to know

• Not on the scale of the crystal datasets. total_energy is a raw PBE-D3(BJ) energy with the settings of QMOF. It cannot be compared with the energies of the Materials Project or of LeMat-Bulk, and there is no convex hull.
• Large cells. One structure in ten has more than 200 atoms. The radius graph lists the pairs of atoms of each structure: use smaller batches than for MP-20.
• Experimental and hypothetical. synthesized is 0 for the hypothetical structures (from ToBaCCo, Boyd & Woo, and others).
• Identifiers are not in the file. all_index.parquet gives for each row its qmof_id, its name in the source database, its formula, its DOI and its MOFid.
• The partial charges, bond orders and magnetic moments per atom of the source are not in the file.
• No structure with partial occupancy was found in the database.
• Licence: CC BY 4.0. Cite the two papers of the database.

Format

HDF5 files of flat concatenated arrays, the format of the other datasets of materials-toolkits:

Dataset	Shape	Type	Content
lattice	(n_struct, 3, 3)	float32	lattice vectors as rows, in Å; the identity for a molecule
num_atoms	(n_struct,)	int64	number of atoms of each structure
ptr	(n_struct,)	int64	index of the first atom of each structure in x and z
x	(n_atoms, 3)	float32	fractional coordinates; positions in Å for a molecule
z	(n_atoms,)	int64	atomic numbers
pbc	(n_struct, 3)	bool	whether the structure is periodic along each lattice vector
total_energy, formation_energy_per_atom, energy_above_hull, band_gap	(n_struct,)	float32	labels, NaN when unknown
properties/<name>	(n_struct,)	float32	the properties listed above, NaN when unknown

The position of an atom in Å is x @ lattice. The lattice of a crystal is rebuilt from the lengths and the angles of its cell, in the standard orientation (first vector along x), and its fractional coordinates are wrapped into the cell. The batch array is an artefact of the writer and is not used when reading. Row i of an index file describes structure i.

Usage

from torchms.data import HDF5Dataset

dataset = HDF5Dataset.from_hub("materials-toolkits/qmof", "all.hdf5")
structures = dataset[:32]            # batched Structures
structures.properties                # the named properties, one value per structure

Built with torchms.

Source and checks

• QMOF Database, figshare, version 18: qmof.json and qmof_structure_data.json of qmof_database.zip, checked with its MD5.
• Built by scripts/build_qmof.py. Every structure was compared with the database, read by pymatgen: number of atoms, elements, coordinates, lattice, total energy, the four band gaps and the largest cavity diameter.
• Repository: materials-toolkits/qmof, file all.hdf5.