High dynamic separation performance of metal–organic frameworks for D2/H2: Independent or competitive adsorption?

Most traditional D2/H2 separation techniques are energy‐intensive with low efficiency. Metal–organic frameworks (MOFs) provide a promising solution for D2/H2 separation due to their excellent chemical and structural characteristics. Here, machine learning‐assisted high‐throughput computational screening was employed to identify the high‐performance MOFs for the dynamic D2/H2 separation. Extensive data analysis reveals that there were two adsorption behaviors in the optimal MOFs, independent adsorption and competitive adsorption, and the independent adsorption was favorable for the preferential adsorption of D2. To quantify these two adsorption behaviors, we introduced and defined overlap degree (OD) and independence degree (ID), and developed a software for the rapid assessment of OD/ID. After batch simulation of the breakthrough curves of 2000 optimal MOFs, ~80% MOFs exhibited independent occupancy, confirming its contribution to good dynamic separation capabilities. This work provides a new idea for designing MOFs with independent adsorption behavior to improve the dynamic separation performance of hydrogen isotopes.