(599d) Beyond the Conventional High-Throughput Computational Screening of MOFs

The emerging metal-organic frameworks (MOFs) have received significant attentions for many potential applications, because MOFs offer huge flexibility in design by varying metal nodes and organic linkers, as well as framework topologies. Numerous experimental studies have been directed to identify record-breaking MOFs for gas separation particularly CO₂ capture. However, navigating through the large design space of MOFs to discover promising candidates is technically challenging and time consuming. High-throughput computational screening (HTCS) can facilitate this process by fast evaluation of the adsorption performance in a large collection of MOFs in silico, thus alleviating experimental synthesis burden. HTCS approaches can be used to explore not only existing experimental MOF databases, but also hypothetical databases, and recommend top-performing MOFs for experimentalists to attempt their synthesis. However, conventional HTCS neglects considering the synthesizability of recommended MOFs, and whether they are stable at the manufacturing and process conditions. To bridge this gap, in the present study, we have incorporated the synthesizability and crucial stability metrics with the conventional HTCS scheme, taking post-combustion CO₂ capture as a case study. Starting from nearly 15,200 hypothetical MOFs, top-performing candidates were firstly identified based on their gas uptake and selectivity. Then, the synthesizability and stability were evaluated for the top 151 MOFs. Free energy approach based on thermodynamic integration was used to assess the synthesis likelihood (thermodynamic stability) of the MOFs at ambient temperature. Mechanical stability was evaluated using the elastic constants obtained through molecular dynamics (MD) simulations. Specifically, the Born’s stability criteria was used to discard the unstable MOFs. Furthermore, machine learning models were employed to predict if the MOFs would survive activation process during solvent removal and whether they are thermally stable. Finally, all the stability measures were combined to identify the top-performing stable MOFs. The approach developed in this study emphasizes on incorporating crucial stability metrics, along with separation performance, to shortlist promising MOFs for CO₂ capture, and it can also be further used for other important industrial applications.