High-throughput Zeolite Structure Prediction

Zeolites are an important class of inorganic crystalline materials that have been widely used in the fields of petroleum refining,petrochemical industry,and fine chemical industry,as catalysts,adsorbents,and ion-exchangers.To date,over 200 types of zeolite frameworks have been discovered,but they cannot meet the rapidly growing demand for new zeolites with superior physical and chemical properties.Because of few understanding of zeolite formation mechanism,exploring not-yet-discovered hypothetical zeolites become more challenging.Predicting not-yet-discovered hypothetical zeolites provides a large pool of candidate structures for synthetic chemists,but it remains challenging because of their intrinsic structural complexity and diversity.Here,we developed a zeolite prediction method by enumerating Wyckoff position combination according to the size of unite cell and the number of T atoms.Using framework-building algorithms that assemble T atoms in fxed Wyckoff position,all the enumerated Wyckoff position combinations can be explored equally,leading to a more comprehensive search in the configuration space.On the basis of these studies,we are able to build,optimize,evaluate,rank,and screen unknown zeolite structure models day and night seamlessly.By enumerating all possible combinations of Wyckoff positions for framework T atoms,this approach can build zeolite framework models much more efficiently and equally than previous approachesThe main results in the thesis are summarized as follows:1.In space group R-3m,we obtained 93005 types of Wyckoff position combinations.Considering the restriction in T atom distribution densities,the number of Wyckoff position combinations is reduced by 50 times to 1785.From these Wyckoff position combinations,we build 63811 tetrahedral network models.After adding bridging oxygen atoms,we obtain 25701 successfully optimized zeolite models using SLC potentials.After evaluation according to the energy and the LID criteria,we obtain 111 feasible zeolite structures.2.In space group Ibca,we obtained 107095 types of Wyckoff position combinations.Considering the restriction in T atom distribution densities,the number of Wyckoff position combinations is reduced by 50 times to 1135.From these Wyckoff position combinations,we build 156724 tetrahedral network models.After adding bridging oxygen atoms,we obtain 25701 successfully optimized zeolite models using SLC potentials.After evaluation according to the energy and the LID criteria,we obtain 511 feasible zeolite structures.3.In contrast,we define the same conditions to filter out Deem_PCOD database,The Atlas of Prospective Zeolite Structures and our result.In R-3m space group,there were 24 feasible low-energy zeolite structures in Deem_PCOD database and 39 feasible low-energy zeolite structures in siliver database.Under the same conditions,we obtained 85 feasible low-energy zeolite structures.In Ibca space group,there were 37 feasible low-energy zeolite structures in Deem_PCOD database and we obtained 163 feasible low-energy zeolite structures.Beyond any dispute,providing all possible atom position combinations is counted as essential factors contributing to a complete and efficient configuration space search.