(2dv) Bridging Physics-Informed and Data-Driven Materials Designs for Deep Decarbonization

Electrifying and decarbonizing the chemical industry is a pressing chemical engineering mission of our time. The chemical industry is the world’s biggest energy consumer and the third-largest source of emissions, where the overproduction of CO₂ has led to severe problems such as global warming. To tackle these issues, governments and industrial players worldwide have set ambitious targets to reduce carbon emissions. For example, the current U.S. administration aims to half the emissions of 2005 by 2030 to reach net zero by 2050. To facilitate such deep decarbonization, it is crucial to electrify the chemical industry—transforming earth-abundant molecules into green chemicals and fuels using electricity converted from solar and wind energy. A core element in achieving this goal lies in developing novel clean energy technologies enabled by unprecedented materials (e.g., catalysts). Unfortunately, there is no time left for materials design as usual, as inventing new materials is a painstakingly slow process. On average, previous innovations have taken 20 years to discover a material and bring it to market. To accelerate materials design for deep decarbonization, it is imperative to bypass traditional research paradigms that rely too heavily on tedious trial and error, unsystematic chemical intuition, and pure serendipity.

To this end, I aim to build a joint computational–experimental research program to develop physics-informed and data-driven blueprints to boost materials design for decarbonization by combining physical chemistry, material descriptors, atomistic simulations, operando characterizations, and machine learning. I aim to bypass traditional research paradigms that largely rely on chemical intuition and serendipity, which can hardly grapple with balancing the reactivity and durability of functional interfaces and elucidating their intricate, dynamic nature. I will couple physics-informed approaches and bleeding-edge data-driven tools to boost materials design out of an immense configurational space of compositions and structures and accelerate their global optimization. My research group will be highly cross-disciplinary—pushing the cutting edges of physical chemistry, materials science, and chemical engineering and spanning the boundaries of theory, experiments, simulations, and machine learning. Moreover, my group aspires to build a diverse team that empowers all to collectively construct non-conformist solutions to the most recalcitrant societal challenges—climate change, pollution, energy poverty, and food insecurity.

Doctoral Research:

I obtained my Ph.D. in Materials Science and Engineering from the Massachusetts Institute of Technology (MIT) in 2022, advised by Prof. Yang Shao-Horn, and collaborated with Prof. Yuriy Román-Leshkov. My Ph.D. research focuses on building quantitative physics-informed design principles of catalyst materials to optimize their reactivity and durability for catalyzing electrochemical water oxidation. Notably, I have combined electrochemistry and reaction kinetics with first-principles atomistic simulations and synchrotron X-ray spectroscopies to unveil how to mechanistically understand and quantitatively engineer these materials to realize optimal activity and stability. I have elucidated how rationally controlling the electronic structure of transition metal compounds can effectively tune their chemical bondings, modulate key reaction barriers, and thus optimize their reactivity and durability, offering quantitative predictive power to boost materials design for decarbonization.

Postdoctoral Research:

My postdoctoral research at MIT, working with Prof. Rafael Gómez-Bombarelli, highlights how materials design can be further accelerated by coupling physics-based design principles with machine learning. Specifically, my work focuses on multicomponent oxides—an uncharted class of materials with great promise for decarbonization but with too-high structural complexity (e.g., cation orderings) for an exhaustive investigation. To conquer this challenge, I have established data-driven ordering descriptors that universally rationalize and accurately predict experimental cation orderings in these complex oxides. Furthermore, I have developed equivariant graph neural networks to accurately infer key cation ordering–dependent properties by learning intrinsic symmetries, enabling efficient discoveries of new multicomponent oxides over a design space of up to billions of materials.

Selected Honors and Awards:

• 2023, Finalist, Distinguished Young Scholars Seminar, Department of Chemical Engineering, University of Washington
• 2023, ENFL Future Investigator Spotlight, Energy and Fuels Division, American Chemical Society
• 2021, Graduate Student Award, Materials Research Society

Teaching Interests

Throughout my academic journey, I have continuously sought out opportunities to deepen my engagement with teaching. For instance, at MIT, I worked as a teaching assistant for a course, “Electrochemical Energy Conversion and Storage: Fundamentals, Materials, and Applications,” which is cross-listed in three engineering departments, including the Department of Chemical Engineering. I also had the opportunity to participate as a trainee in MIT’s semester-long Kaufman Teaching Certificate Program to receive systematic training on evidence-based teaching techniques grounded in the scholarship of teaching and learning. These experiences have prepared me to teach chemical engineering courses at an undergraduate or graduate level, such as thermodynamics and kinetics. I am also excited about opportunities to teach or develop graduate-level courses related to my research interests and expertise, such as (1) fundamentals of electrochemical energy conversion and storage and (2) machine learning and data science for chemical engineering. As a future faculty member, I am committed to building a supportive, diverse, and inclusive learning and mentoring environment that uplifts all.