Author ORCID Identifier
0000-0002-2434-5201
Document Type
Dissertation
Date of Award
5-31-2026
Degree Name
Doctor of Philosophy in Mechanical Engineering - (Ph.D.)
Department
Mechanical and Industrial Engineering
First Advisor
Dibakar Datta
Second Advisor
Eon Soo Lee
Third Advisor
Bo Shen
Fourth Advisor
Samaneh Farokhirad
Fifth Advisor
Mengqiang Zhao
Abstract
The growing demand for sustainable, high-energy-density electrochemical storage has motivated the exploration of multivalent-ion batteries based on earth-abundant elements such as aluminum, calcium, magnesium, and zinc. While multivalent charge carriers offer higher theoretical energy density than lithium, their practical deployment is hindered by sluggish ion transport, strong ion-host interactions, and structural degradation of electrode materials. Identifying host materials that can reversibly accommodate multivalent ions while maintaining structural integrity remains a fundamental challenge. The dissertation develops a scalable, end-to-end computational framework that integrates density functional theory (DFT), machine learning (ML), and generative artificial intelligence (GenAI) to accelerate the discovery of next-generation multivalent-ion battery materials. The work focuses on open-tunnel transition metal oxides, whose interconnected nanoporous channels provide natural diffusion pathways for large, highly charged ions. First-principles DFT calculations are used to systematically investigate multivalent-ion insertion in representative tunnel oxides, revealing the critical role of tunnel geometry and topology in governing ion accommodation and stability. Guided by these physical insights, a generative materials discovery framework is developed by combining a Crystal Diffusion Variational Autoencoder (CDVAE) with a fine-tuned Large Language Model (LLM) trained on over 44,000 known inorganic crystal structures. This framework generates more than 20,000 candidate transition metal oxide structures, which are subsequently screened using graph based machine learning models for thermodynamic stability and electronic properties. Promising candidates are rigorously validated through full DFT structural relaxations and phonon calculations, leading to the identification of several AI-generated materials that can be implemented for industry-scale multivalent-ion-based battery applications. Collectively, this work demonstrates a generalizable and scalable paradigm for AI-driven materials discovery that bridges physics-based modeling, data-driven generation, and first-principles validation. To further connect computational discovery with experimental realization, the dissertation also introduces a retrieval-augmented synthesis planning framework, enabling data-driven guidance for the synthesis of AI-designed materials. The proposed methodology provides a foundation for industry-relevant, large-scale discovery of multivalent-ion battery materials and can be extended to other functional material classes.
Recommended Citation
Datta, Joy, "A generative ai-driven computational framework for industry-scale discovery of novel battery materials" (2026). Dissertations. 1878.
https://digitalcommons.njit.edu/dissertations/1878
Included in
Artificial Intelligence and Robotics Commons, Data Science Commons, Materials Science and Engineering Commons, Mechanical Engineering Commons
