(582f) Towards Process-Materials Co-Optimization: Automatic Generation of Optimizable MOF Structure-Function Relationships

Metal-organic frameworks (MOFs) are promising adsorbent materials for next-generation gas adsorption applications^1,2. The traditional screening-based adsorbent selection workflow is inefficient and costly, due to the vast design space of highly modular and flexible MOF structures. To that end, there is a need to develop systematic search methods to identify optimal MOF designs³. In addition, it is necessary to integrate MOF design with adsorption process design for the optimal process-level performance⁴. Currently, many adsorption process optimization efforts simply ignore the adsorbent’s design space. Other studies that do attempt to incorporate materials design merely consider the analytical isotherm model’s parameters as the decision variables, a practice that still bypasses the actual MOF molecular design space^5,6. This work proposes a new workflow to generate optimizable MOF structure-function relationships, linking MOF molecular design space to their process-level adsorption properties (i.e., isotherm models) and thus enabling direct process-materials co-optimization.

We illustrate our proposed workflow with MOF structures and adsorption data obtained from the hMOF database^7,8. We first generate a pool of descriptors, including pore geometries (e.g., pore size distribution, pore volumes) via the Zeo++ toolkit⁹ and structural identities (e.g., topology, building blocks) via the MOFid algorithm¹⁰. We are able to reconstruct a MOF’s molecular structure from its vector of descriptors, which is critical for MOF structure optimization. We then exploit the adsorption data by fitting them into the dual-site Langmuir (DSL) equation, which is commonly utilized in process optimization to describe adsorption equilibria⁴. For this, we take advantage of parameter estimation routines available within the IDAES Integrated Platform¹¹ to automate the DSL parameter fitting process. Finally, we learn a surrogate model that satisfactorily links the MOFs descriptors (i.e., features) with the DSL parameters (i.e., targets) using the ALAMO suite^12,13. This workflow leverages state-of-the-art open-source tools and is highly automated. The resulting predictive structure-function relationships enable the inverse design of a MOF’s molecular structure from its process-level adsorption properties. It can thus be easily integrated into optimization models for process design, realizing process-materials co-optimization in a common framework.

References

(1) Zhao, D.; Yuan, D.; Zhou, H.-C. The Current Status of Hydrogen Storage in Metal–Organic Frameworks. Energy & Environmental Science 2008, 1 (2), 222–235.

(2) Simmons, J. M.; Wu, H.; Zhou, W.; Yildirim, T. Carbon Capture in Metal–Organic Frameworks—a Comparative Study. Energy & Environmental Science 2011, 4 (6), 2177–2185.

(3) Colón, Y. J.; Snurr, R. Q. High-Throughput Computational Screening of Metal–Organic Frameworks. Chemical Society Reviews 2014, 43 (16), 5735–5749.

(4) Farmahini, A. H.; Krishnamurthy, S.; Friedrich, D.; Brandani, S.; Sarkisov, L. From Crystal to Adsorption Column: Challenges in Multiscale Computational Screening of Materials for Adsorption Separation Processes. Industrial & Engineering Chemistry Research 2018, 57 (45), 15491–15511.

(5) Khurana, M.; Farooq, S. Integrated Adsorbent-Process Optimization for Carbon Capture and Concentration Using Vacuum Swing Adsorption Cycles. AIChE Journal 2017, 63 (7), 2987–2995.

(6) Khurana, M.; Farooq, S. Integrated Adsorbent Process Optimization for Minimum Cost of Electricity Including Carbon Capture Bya VSA Process. AIChE Journal 2019, 65 (1), 184–195.

(7) Wilmer, C. E.; Farha, O. K.; Bae, Y.-S.; Hupp, J. T.; Snurr, R. Q. Structure–Property Relationships of Porous Materials for Carbon Dioxide Separation and Capture. Energy & Environmental Science 2012, 5 (12), 9849–9856.

(8) Bobbitt, N. S.; Chen, J.; Snurr, R. Q. High-Throughput Screening of Metal–Organic Frameworks for Hydrogen Storage at Cryogenic Temperature. The Journal of Physical Chemistry C 2016, 120 (48), 27328–27341.

(9) Willems, T. F.; Rycroft, C. H.; Kazi, M.; Meza, J. C.; Haranczyk, M. Algorithms and Tools for High-Throughput Geometry-Based Analysis of Crystalline Porous Materials. Microporous and Mesoporous Materials 2012, 149 (1), 134–141.

(10) Bucior, B. J.; Rosen, A. S.; Haranczyk, M.; Yao, Z.; Ziebel, M. E.; Farha, O. K.; Hupp, J. T.; Siepmann, J. I.; Aspuru-Guzik, A.; Snurr, R. Q. Identification Schemes for Metal–Organic Frameworks to Enable Rapid Search and Cheminformatics Analysis. Crystal Growth & Design 2019, 19 (11), 6682–6697.

(11) Gunter, D. K.; Agarwal, D. A.; Beattie, K. S.; Boverhof, J. R.; Cholia, S.; Cheah, Y.-W.; Elgammal, H.; Sahinidis, N. V.; Miller, D.; Siirola, J.; others. Institute for the Design of Advanced Energy Systems Process Systems Engineering Framework (IDAES PSE Framework); Lawrence Berkeley National Lab.(LBNL), Berkeley, CA (United States), 2018.

(12) Cozad, A.; Sahinidis, N. V.; Miller, D. C. Learning Surrogate Models for Simulation-Based Optimization. AIChE Journal 2014, 60 (6), 2211–2227.

(13) Cozad, A.; Sahinidis, N. V. A Global MINLP Approach to Symbolic Regression. Mathematical Programming 2018, 170 (1), 97–119.