Metal-organic frameworks (MOFs) are microporous materials possessing exciting applications, in gas storage and separation, catalysis, artificial photosynthesis, energetic materials1,2. The diversity of applications stems from the modular node-and-linker composition, where different combinations of building blocks yield materials with a range of properties. Currently, MOF design is principally based on empirical experimental approaches dependent on chemical intuition. In contrast, the development of ab initio methods for predicting structures and properties of MOFs could provide new, unexpected opportunities for theory-driven. Recently3, our group has reported the first-principles crystal structure prediction (CSP) method for MOFs, based on the ab initio random structure searching (AIRSS)4 and Wyckoff Alignment of Molecules (WAM) algorithms. Here, we will demonstrate the first examples for the prediction of hexafluosilicate-based MOFs, combined with mechanochemical synthesis and crystal structure determination. The theoretical CSP guided mechanochemical synthesis provides a more effective way for MOFs design. Moreover, to further improve the efficiency of our MOF CSP method, where accurate but computationally-demanding periodic DFT is utilized, machine learning (ML) structure prediction approaches are studied. Atom-wise interatomic potentials were trained based on periodic DFT simulation data5. Subsequently, the potentials which showed excellent agreement between DFT and ML, were used to predict the energies and optimize the geometries of randomly-generated structures. The far lower cost of ML-based calculations allows us to explore the largest MOF structures, containing more formula units per crystallographic unit cell. Overall, we expect this ML approach to provide us with a better CSP method in terms of accuracy, efficiency and the range of potential MOF structures that can be predicted.

Metal-organic framework design: theoretical crystal structure prediction and properties study

Fall 2022 abstracts, posters, and presentations will launch in mid-September.

ACS Fall 2022

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.