Research
Our research group uses computational chemistry methods and machine learning to solve the challenges in designing new soft materials. To this end, we actively work on developing new computational methods that can be generalized to other research areas as well. Our expertise lies in characterizing the properties of organic molecular crystals as well as polymeric materials by utilizing methods such as quantum chemistry, semiempirical, classical, and enhanced sampling to bridge the gap in length/time scales of the typical simulations.
Force-Field Development from First Principles
Molecular dynamics simulation is an important computational tool for understanding and predicting the properties of materials. To propose new materials with the desired properties using molecular dynamics, the ability to derive force-field parameters from scratch is paramount. Our group develops data-driven approaches for building force-field from quantum chemistry calculations such as density functional theory, aimed towards material discovery through high-throughput simulations.
High-Throughput Screening Through Machine Learning
High-throughput computations provide opportunity for us to utilize the data through machine learning algorithms that can aid us in the areas where heuristics based on chemical intuitions are currently employed. Our approach to material discovery involves constructing machine learning models for efficient exploration of the chemical space with various applications such as machine learning-based potential, collective variables design, and generative model.
Mechanistic Modeling of Organic Molecular Crystals
Crystalline organic solids are ubiquitous in industries including pharmaceutical, defense, food, electronics, and energy. The quality of their products depends on the properties such as polymorph, particle size distribution, morphology, and purity, which are determined during the crystallization processes. Therefore, controlling the properties of the crystal is a crucial part of process optimization. We develop crystal growth models to predict growth morphology and employ methods like enhanced sampling and semiempirical quantum chemistry to facilitate the rare event problems.
Characterization of Lithium-Ion Battery Electrolytes
In the global trend of low-carbon pursuit, lithium-ion batteries are in the spotlight for portable devices, renewable energies, and electric vehicles due to their high energy density and long lifespan. The advent of post-lithium-ion batteries with improvements in terms of performance, production cost, safety, and lifespan, requires understanding and further optimizing of the specific components such as the cathode, anode, and electrolyte. Using computations, we seek to provide insights into the structures and transport mechanisms in electrolyte materials.
Â