Research

Overview 

The overarching goal of our group is to leverage the advances in multiscale molecular simulations and data science tools to accelerate the computational design and experimental realization of functional materials to address current and future societal challenges related to energy, environment, and human health. This vision is enabled by training a wonderful group of graduate and undergraduate researchers, and is reinforced by a wide network of friendly experimental and theory collaborators cross the department, campus, and other top-ranked institutions. Our work is supported by a diverse array of grants from internal UC Davis funds, federal funding agencies (both DOE and NSF), California-based national laboratories, and industry. Some of our recent successes are summarized below. 

  1. Leveraging 3-dimensional active site environments in metal-organic frameworks (MOFs) 
  2. Addressing the complexities of Al distributions in zeolites 
  3. Theory-guided EXAFS analyses for unprecedented characterization of precisely synthesized atomically dispersed catalysts 
  4. Machine learning assisted automation of surface-enhanced Raman spectroscopy (SERS) data 

Scientific Vision and Priorities 

Severe droughts in California. Wildfires in Oregon. 100 °F temperatures in the Arctic. Flooding in Germany. If the last few years have taught us anything, it is that time has now come for society to reckon with the detrimental impact of climate change on the environment and on human health. While some of the direct (e.g., rising CO2 levels, water eutrophication etc.) and indirect (e.g., increased prevalence diabetes, cancers etc.) impacts are clear, the list of unforeseen problems will keep growing. Scientists and policymakers across the globe are working to anticipate and mitigate these challenges.

The development of new functional materials and novel processes lies at the heart of these efforts. Although the scientific domains may differ (e.g., cost-effective fuel cell catalysts, adsorbents for CO2 capture, real-time sensors for cancer diagnostics), the core concepts of materials discovery remain similar. It is logistically impossible to synthesize, characterize, and test thousands of candidates for all possible applications. To alleviate the bottlenecks associated with traditional trial-and-error approaches, the central theme in our work is to develop robust and versatile materials development pipelines using atomistic simulations, data science tools, and experimental collaborations (both within academia and industry). Our work is typically organized according to the following Thrusts: 

  1. How does the molecular-level structure of the active site affect its catalytic performance? How can these structure-function relations be used to design more cost effective and better performing materials and processes?
  2. How can theory help interpret complex characterization experiments (e.g., synchrotron X-ray absorption, surface-enhanced Raman), and provide unprecedented insights into the molecular structure of the active site?
  3. How can we make our predictions and analyses more robust and reliable, while reducing computational costs?
  4. How can advances in data science be used streamline and automate data acquisition workflows and eventually, enable high-throughput experiments?

While the above Thrusts are presented in the context of catalysis research, these questions are materials-agnostic, and are broadly relevant to other applications such as chemical separations and sensing. The following three core characteristics of our group make us uniquely suited to address the above challenges:

  1. Versatility of our simulation tools across different material classes and application domains.

  2. Robustness of our predictions, which arise from our expertise with (single-reference) wave function theory, density functional theory, force field methods, and emphasis on careful benchmarking studies.
  3. Use of modern data science approaches including open-source software development and FAIR (Findable, Accessible, Interoperable, and Reusable) data practices.