The New Jersey Institute of Technology's
Electronic Theses & Dissertations Project

Title:	Data-driven 2d materials discovery for next-generation electronics
Author:	Zhang, Zeyu
View Online:	njit-etd2023-048 (xiv, 110 pages ~ 23.8 MB pdf)
Department:	Department of Chemistry and Environmental Science
Degree:	Doctor of Philosophy
Program:	Chemistry
Document Type:	Dissertation
Advisory Committee:	Shakib, Farnaz A. (Committee chair) Chen, Hao (Committee member) Li, Mengyan (Committee member) Zhang, Yuanwei (Committee member) Young, Joshua (Committee member)
Date:	2023-08
Keywords:	Computational chemistry Machine learning Material database Material design Metal organic frameworks
Availability:	Unrestricted
Abstract:	The development of material discovery and design has lasted centuries in human history. After the concept of modern chemistry and material science was established, the strategy of material discovery relies on the experiments. Such a strategy becomes expensive and time-consuming with the increasing number of materials nowadays. Therefore, a novel strategy that is faster and more comprehensive is urgently needed. In this dissertation, an experiment-guided material discovery strategy is developed and explained using metal-organic frameworks (MOFs) as instances. The advent of 7r-stacked layered MOFs, which offer electrical conductivity on top of permanent porosity and high surface area, opened up new horizons for designing compact MOF-based devices such as battery electrodes, supercapacitors, and spintronics. Structural building blocks, including metal nodes and organic linkers in these electrically conductive (EC) MOFs, are recognized and taking permutations among the building blocks results in new systems with unprecedented and unexplored physical and chemical properties. With the ultimate goal of providing a platform for accelerated material design and discovery, here the foundation is laid for the creation of the first comprehensive database of EC MOFs with an experimentally guided approach. The first phase of this database, coined EC-MOF/Phase-I, is composed of 1,057 bulk and monolayer structures built by all possible combinations of experimentally reported organic linkers, functional groups, and metal nodes. A high-throughput (HT) workflow is constructed to implement density functional theory calculations with periodic boundary conditions to optimize the structures and calculate some of their most relevant properties. Because research and development in the area of EC MOFs has long been suffering from the lack of appropriate initial crystal structures, all of the geometries and property data have been made available for the use of the community through an online platform that was developed during the course of this work. This database provides comprehensive physical and chemical data of EC-MOFs as well as the convenience of selecting appropriate materials for specific applications, thus accelerating the design and discovery of EC MOF-based compact devices. Machine learning (ML), a technique of learning patterns of numerical data and making predictions, can be utilized in material discovery. Taking advantages of the EC-MOF Database, ML is adopted to predict property data that needs expensive calculations according to the crystal structures only. The implementation of ML is much faster than the HT workflow when the number of structures increases constantly.