top of page
Research and Work
The traditional approach to discovering new materials has relied on a time-consuming and resource-intensive trial-and-error process. Fortunately, the rapid advancements in computer hardware and artificial intelligence have paved the way for a more efficient approach to material discovery. By harnessing simulation and data-driven science, we can significantly expedite the search for promising new material candidates.
1. Prediction of Surfactant Adsorption Effectiveness of a Novel Surfactant at Air-Water Interfaces (2023 - Present)
2. The Discovery of Surfactant Molecules Used for Fire Suppression (2022 - Present)
To accurately predict the properties of molecules, it is crucial to establish a robust molecular representation. In our research, we leverage molecular dynamics (MD) simulations in conjunction with various machine learning algorithms to develop surfactant-specific descriptors. These descriptors, computed through MD simulations, are specifically tailored to accurately characterize surfactant properties.
Our primary focus is on predicting the surfactant adsorption effectiveness, as it plays a pivotal role in determining essential surfactant attributes such as foaming, wetting, and emulsification. However, our methodology can be extended to forecast additional surfactant properties, including surface tension, critical micelle concentration (CMC), and wettability.
It's worth noting that this study is an integral part of the SERDP project, where we aim to advance our understanding of surfactants and their various applications.
Using Molecular Dynamics (MD) simulation and Machine/Deep learning (ML/DL), I have developed a framework for molecule design to invent new materials from the vast space of hypothetical materials. Currently, my work has been focusing on the discovery of surfactant molecules used for fire suppression, which SERDP funds. This study can be generalized to other types of material discovery.
3. Prediction of a Permeability of Small Molecules across Lipid Membranes Using Machine Learning (2022)
In this project, we aim to predict the permeability of small molecules across lipid membranes using several machine-learning algorithms. We plan to use the dataset of a published paper which includes more than 90,000 samples containing molecular-structure-related information and permeability. Supervised learning algorithms will be used. This project has been conducted through the Computer Science class in the Fall of 2022 (CS 5824).
bottom of page