430712 Application of a Genetic Algorithm to Screen Metal-Organic Frameworks: Pre-Combustion CO2/H2 Separation

Wednesday, November 11, 2015: 4:45 PM
255A (Salt Palace Convention Center)
Yongchul G. Chung1, Diego Gomez-Gualdron1, Peng Li2, Nicolaas Vermeulen2, Pravas Deria2, Fraser Stoddart2, Joseph T. Hupp2, Omar K. Farha2 and Randall Q. Snurr1, (1)Chemical and Biological Engineering, Northwestern University, Evanston, IL, (2)Chemistry, Northwestern University, Evanston, IL

Metal-organic frameworks (MOFs) are novel porous materials consisting of organic and inorganic building blocks, whose combinatorial possibilities give rise to a practically unlimited number of possible structures. Due to the large number of possible MOFs, high-throughput computational screening has been used in the literature to identify high-performing MOFs for particular applications of interest, such as methane storage and delivery. Computational screening has also provided useful structure-property relationships that can guide experimental discovery efforts. However, current approaches are “brute-force”, where full molecular simulations are carried out to calculate the property of interest for a very large number of structures that have been generated or mined from the literature. This approach may not be suitable, however, for complex systems or for very time-consuming simulations because of the prohibitively large computational cost associated with such simulations.

In this work, we have developed a Genetic Algorithm (GA) to identify high-performing MOFs for pre-combustion CO2/H2 separation from a database of ~137,000 hypothetical MOFs [1]. Firstly, grand canonical Monte Carlo (GCMC) simulations were carried out to calculate CO2/H2 uptakes under selected operation conditions for a “genetically diverse” initial generation of 100 hypothetical MOFs. Secondly, genetic operations were carried out on the preceding generation of structures to create a new one, for which GCMC simulations were carried out as well. This procedure was continued until we reached 10 generations.

Our results demonstrate that the developed GA is a robust method that can find high-performing MOFs based on different performance metrics such as CO2/H2 selectivity and CO2 working capacity in a much more computationally efficient fashion than the brute force method. One of the top MOFs identified by the GA was synthesized and activated, with the experimental CO2 and H2 adsorption values in good agreement with the molecular simulation results. Moreover, the predicted MOF has the highest CO2 working capacity (~ 6 mol/kg) among reported MOFs and zeolites in the literature under the same operating conditions [2].

#: Authors contributing equally


[1] C. E. Wilmer, M. Leaf, C.Y. Lee, O.K. Farha, B.G. Hauser, J.T. Hupp, and R.Q. Snurr, “Large-scale screening of hypothetical metal-organic frameworks”, Nature Chemistry, 2012, 4, (2), 83 – 89.
[2] Z. R. Herm, J. A. Swisher, B. Smit, R. Krishna, and J. R. Long, “Metal-Organic Frameworks as Adsorbents for Hydrogen Purification and Precombustion Carbon Dioxide Capture”, J. Am. Chem. Soc., 2011, 133, (15), 5664 – 5667.

Extended Abstract: File Not Uploaded