The full text of this item is not available at this time because the student has placed this item under an embargo for a period of time. The Libraries are not authorized to provide a copy of this work during the embargo period, even for Texas A&M users with NetID.
Computational Frameworks Towards the Discovery of Zeolites and Similar Nanoporous Crystalline Materials
Abstract
Porous crystalline materials are highly valued for their exceptional capabilities in separation, catalysis, and ion exchange. In particular, zeolites are nanoporous tetrahedral aluminosilicate crystal frameworks with different Si/Al ratios, which have large internal surface areas with high gas adsorption and selectivity for many gas separation applications including carbon capture, natural gas purification, and hydrogen separation. While only about 260 zeolites have been naturally found or chemically synthesized so far, there are millions of hypothetical zeolite frameworks yet to be explored. For each zeolite framework, additional substitutions of Si atoms with Al are possi-ble. For example, there are almost 2 million possible ways of Al-substitutions in chabazite (CHA) framework alone for a Si/Al ratio of 5.0. The combinatorically large space of both pure silica and Al-substituted zeolite frameworks poses a significant challenge for discovery of new zeolites with desired properties. In this doctoral work, we propose a computationally efficient screening strategy for identifying new Al-substituted zeolites with high CO2 adsorption capabilities. The central to this strategy is a novel graph-theoretic representation for crystalline frameworks, which we call the single repeating unit (SRU) representation. In the case of zeolites, SRU representation significantly reduces the description space with fewest possible T-nodes. Furthermore, it allows the systematic analysis of groups of similar structures and reduces the need for high-fidelity molecular simulations for property estimation, thereby significantly increasing the computational efficiency of zeolite discovery for specific applications.
We propose multiple SRU identification approaches and test for over 180 known and 10,000 hy-pothetical zeolite frameworks, showcasing the benefits of this representation. In the optimization-based approach, SRU identification is formulated as a special instance of the traveling salesman problem. Despite the potential of machine learning approaches for property prediction, limitations exist even with the proposed reduced representation. This necessitates complex modeling requirements for accuracy and poses significant challenges for the inverse design of materials. Based on previous insights, Al-substitution has been known to increase CO2 gas adsorption in the frame- work, however, a thorough understanding of the mechanisms at play is lacking. To overcome the enumeration challenge posed by Al-substitution, we propose a framework for the systematic enumeration of Al-substituted frameworks, based on the SRU representation. We demonstrate that an optimal Si/Al ratio exists for the highest adsorption, and different configurations of Al substitutions lead to varying adsorption capacities, showing up to a 12% variation. Investigation utilizing the radial distribution function indicates that a uniform distribution of Al within the lattice ex-hibits superior properties compared to dense pockets of Al. The proposed framework is compared to enumerating all 2 million structures of Si/Al = 5.0 in the CHA unit cell. Multiple candidates are selected based on insights learned and metrics, showcasing high adsorption capacities, thus demonstrating the capability of the proposed framework. Compared to the CO2 adsorption in pure-silica CHA framework, the newly identified top Al-substituted CHA frameworks have 119%of increased CO2 adsorption. Importantly, we required few molecular simulations to find these top candidates. The proposed framework is amenable to the discovery of similar porous materials, including metal-organic frameworks (MOFs), covalent organic frameworks (COF), porous polymer networks (PPN), etc.
Citation
Gandhi, Akhilesh Kamlesh (2023). Computational Frameworks Towards the Discovery of Zeolites and Similar Nanoporous Crystalline Materials. Doctoral dissertation, Texas A&M University. Available electronically from https : / /hdl .handle .net /1969 .1 /199823.