(2dx) Rational Design of Sustainable Chemical Solutions with Reaction Networks and Data Science

Traditional industrial processes have brought humanity to the brink of total ecological and climate disaster. Out-of-control greenhouse gas production drives anthropogenic climate change, while pollution by plastics, metals, and small molecules threatens the wellbeing of human and non-human populations. There is an urgent need to develop sustainable technologies that can both minimize the impact of historical emissions (e.g. removing pollutants from the environment) and eliminate future emissions (e.g. improving energy efficiency and replacing fossil fuels with renewable resources and energy storage). While the challenges to sustainability are varied, they are alike in that they can be addressed via engineered chemical reactions and interactions.

The design of sustainable technologies today is primarily guided by labor-intensive trial and error and incremental improvements on existing solutions. First-principles rational design is attractive to efficiently identify chemical solutions, but the design of reactive technologies - for instance, selectively decomposing electrolytes in metal-ion batteries and catalysts to chemically recycle polymer waste to monomers - is inherently challenging. Such systems are often far from equilibrium, involving many reactions which can occur simultaneously in multiple phases and compete with desired pathways.

In my research, I rely on theory, modeling, and data science to improve our understanding of molecular reactivity and use this understanding to accelerate chemical transitions to sustainability. I am motivated to develop tools combining simulated and experimental data to optimize reactivity, with a particular interest in domains where reaction mechanisms are poorly understood, such as electrochemistry and multi-phase catalysis. With the methods that I develop, I aim to design chemistries for high-performance batteries, energy-efficient electrochemical synthesis, chemical polymer recycling, and pollution management.

Research Experience:

During my doctoral studies, I have realized a data-driven methodology to predict the behavior of complex reactive systems and applied it to obtain unprecedented insights into the stability and long-term performance of metal-ion batteries. To achieve high energy densities, metal-ion batteries operate at extreme potentials outside of the thermodynamic stability window of most electrolytes. As a result, electrolytes degrade at battery electrodes in complex reaction cascades. This reactivity has a key role in determining long-term battery health. Unmitigated electrolyte decomposition leads to rapid capacity loss, while well-designed electrolytes form interphases which passivate electrode surfaces and limit ongoing parasitic reactions. Despite the importance of interphase formation in metal-ion batteries, the mechanisms of electrolyte reactivity remain mysterious, preventing the rational design of advanced energy storage technologies.

My research into predicting and understanding reactivity in metal-ion batteries has spanned three main areas:
1. Using high-throughput density functional theory (DFT), I have constructed datasets of molecular and reaction properties. This data, which includes some of the largest collections of charged, open-shell, and metal-coordinated molecules yet published, is open to the scientific community, aiding in studies of battery electrochemistry and various chemical machine learning (ML) tasks.
2. I have developed techniques to construct and analyze chemical reaction networks (CRNs), enabling the efficient exploration of (electro)chemical reactivity in systems involving up to thousands of species and millions of reactions. With this methodology, I can automatically identify plausible reaction products and their formation mechanisms with minimal prior knowledge of the systems of interest.
3. Combining automated CRN analysis with microkinetic modeling and experimental characterization, I have improved our understanding of electrolyte degradation and interphase formation. My simulations have for the first time revealed the origins of the observed bilayer interphase structure entirely from first principles. Moreover, I have recently employed CRN analysis to interpret mass spectroscopy data, elucidating the mechanisms of gas evolution and solvent decomposition in Mg-ion electrolytes. The insights obtained by these studies have revealed key descriptors to control molecular reactivity in batteries, improving the design of next-generation electrolyte components.

In this poster, I will present an overview of this work, focusing on the development of a database of computed molecular properties integrated into the Materials Project, as well as projects related to predicting interphase composition and structure in Li-ion batteries.

Research Interests:

As a professor of chemical engineering, I will focus my research efforts on designing reactive processes for sustainable technologies. Research activities will be focused in three thrusts:
1. Understanding reactivity with CRNs: I will expand on the methods that I have developed to aid in the construction and analysis of CRNs. I aim to build tools to automatically construct and analyze multi-phase CRNs, with reactions occurring simultaneously at solid interfaces and in fluids. This will enable in-depth analysis of reactivity in application areas that have thus far been inaccessible, such as coupled heterogeneous-homogeneous catalysis. I will further leverage active learning to predict the properties of reactions in CRNs (e.g. thermodynamics and kinetics), minimizing the need for expensive electronic-structure calculations and unlocking the possibility of high-throughput CRN studies.
2. Theoretical-experimental reaction optimization: Rational design of reactivity requires effective application of synergistic simulated and experimental results. I am interested in using ML-guided automated experimentation to discover new reaction mechanisms and improve homogeneous and heterogeneous catalytic chemistries. To efficiently identify optimal solutions, I will train ML models combining descriptors from electronic structure calculations, CRN analysis, and prior experiments; these models will then automatically select experiments that are most likely to improve on metrics of interest.
3. Extreme multi-scale modeling of energy and sustainability technologies: Successful interventions to mitigate ecological and climate impacts must operate effectively on multiple scales: on the fundamental level of individual molecules; on the device level; and on the system level, where devices exist within environmental, industrial, and economic contexts. I intend to bridge these various scales, employing atomistic-scale predictions to inform device-scale models and further utilizing those device-scale models to inform predictions on systems-scale impacts. In this way, high-level performance targets can be effectively tied to fundamental chemistry and provide feedback to guide reaction engineering efforts.

In the initial stages of my research, I intend to develop these capabilities by focusing on a small number of application areas, namely electrolyte degradation and interphase formation in Na-ion batteries and electrochemical synthesis of ammonia from nitrogen and nitrate, with the aim of expanding to other domains (e.g. chemical recycling of polyolefins and catalytic removal of pollutants from air) in the future.

My research portfolio will be necessarily interdisciplinary and highly collaborative. In addition to utilizing core competencies in chemistry and chemical engineering, I will rely on advances in computer science, data science, and systems engineering. I expect to collaborate closely with experimentalists to perform characterization studies (especially of device performance), as well as domain experts in ML, mechanical engineering, thermal engineering, and technoeconomic modeling.

Teaching Interests:

My approach to teaching is grounded in the question: how can the classroom, in addition to providing information, be a site of wellness and excitement? This begins with the relationship between professor and student. I work to understand my students as people and as learners and collaboratively develop the courses I teach with students to give them ownership over their education. I also use concepts from universal design for learning, developing curricula to ensure that all students have multiple routes to engage with the course, develop understanding, and demonstrate their mastery of the course material.

With significant classroom instruction experiences leading workshops, laboratory sessions, and small-group discussions, I am well equipped to teach any core undergraduate or graduate chemical engineering course. My expertise in modeling molecular reactivity prepares me particularly well to teach courses in thermodynamics and kinetics. To give students the necessary skills to develop sustainable electrochemical technologies and meaningfully address climate change, I would be interested in developing or teaching an undergraduate elective course on electrochemistry, covering the theory of electrochemical cells as well as its application to common devices. Finally, I strongly desire to teach a course in modern scientific ethics. This course for undergraduate or graduate students would introduce ethics broadly, discuss questions of scientific morals and ethics, and challenge students to reason through case studies from current events and their own communities, critiquing existing power structures while developing their own personal and collective sense of right and wrong.

Regarding mentoring, my style can best be described as individual and proactive. As a graduate student, I have had the great fortune to mentor a diverse group of undergraduates with varied interests (algorithms, science policy) and goals (entering industry or graduate school, gaining confidence in scientific communication). I begin each of these relationships by getting to know the person who I am working with and developing projects that would help them to build the skills necessary to meet their goals. This ensures that students are performing work that is meaningful for them and that the mentoring relationship is mutually beneficial. I aim to be consistently available for mentees – not just when they need help – serving as a sounding board and a source of encouragement and accountability. Knowing well the key role that mentors can have in opening doors, especially for students from disadvantaged backgrounds who lack social capital, I seek to connect mentees with opportunities in research, networking, and professional development.