Identifying misbonded atoms in the 2019 CoRE metal-organic framework database

Area of Science:

• Materials Science
• Computational Chemistry
• Crystallography

Background:

• Experimentally-derived metal-organic framework (MOF) crystal structures are crucial for computational screening.
• Existing databases like the 2019 CoRE MOF database require further cleaning to remove artifacts and ensure accuracy.
• Partially cleaned databases can lead to inaccuracies in identifying suitable MOFs for specific applications.

Purpose of the Study:

• To perform further cleaning of the 2019 CoRE MOF database.
• To identify and correct various structural artifacts, including misbonded or isolated atoms.
• To generate forcefield precursors for a large number of MOFs to facilitate computational studies.

Main Methods:

• Systematic identification of structures with isolated atoms.
• Detection of structures with atoms that are too close together (overlapping atoms).
• Identification of structures with misplaced hydrogen atoms, under-bonded, or over-bonded carbon atoms.
• Atom typing for accepted structures to enable forcefield parameterization.

Main Results:

• Several types of structural artifacts in the 2019 CoRE MOF database were identified and addressed.
• Atom typing was performed on accepted structures.
• Forcefield precursors, including net atomic charges and polarizabilities, were reported for over five thousand MOFs.

Conclusions:

• This study represents significant progress towards a completely cleaned database of experimentally-derived MOF crystal structures.
• The generated forcefield precursors enhance the utility of the CoRE MOF database for computational screening.
• Further cleaning and refinement of MOF databases are essential for advancing materials discovery.