Innovation of Gas Separation Processes: Integrating Computational MOF Design and Adsorption Process Optimization

Considering the multiscale nature of adsorption processes, an integrated metal–organic framework (MOF) and pressure swing adsorption (PSA) process design approach is presented for gas separation. It consists of two steps: adsorbent descriptor optimization and metal–organic framework (MOF) matching. In the first step, MOFs are represented as a set of selected chemical and geometric descriptors which are treated as design variables. Based on collected adsorption isotherms of 471 different MOFs for propene/propane separation, artificial neural network (ANN)-based isotherm models are developed to serve as material–property relationships. In addition, the valid design space of the selected descriptors is confined using a classifier model and logic constraints. Combining the design space, isotherms, and four-step PSA process models, an integrated MOF descriptor PSA process design problem is formulated and solved to generate the optimal descriptors and attainable isotherms. In the second step, based on the optimal isotherms, the design of computer experiments is performed to derive model-based property–performance relationships. Then, MOF building blocks are extracted from 471 MOFs contained in the computation-ready experimental (CoRE) MOF database. With these building blocks, 45 472 hypothetical MOFs are created. After model-based and molecular simulation-based screening, six candidates are left and sent to PSA process optimization. Finally, three candidates are found to meet the predefined separation specifications and one candidate shows a better process performance than the best out of the 471 CoRE MOFs.