Research

AI-Assisted Prediction of Pharmaceutical Ink Printability Using Rheology and Few-Shot Domain Adaptation

This research focuses on developing a data-driven framework for predicting the 3D printability of pharmaceutical polymer inks using rheological properties and printing-process parameters. Extrusion-based pharmaceutical 3D printing is strongly influenced by material characteristics such as viscosity, viscoelasticity, and structural recovery, together with processing conditions such as printing pressure and speed. However, substantial differences among formulations make it difficult for conventional machine-learning models to generalize to previously unseen materials. To address this limitation, the project develops a similarity-guided supervised few-shot domain adaptation framework that transfers knowledge from previously characterized formulations to a new formulation using only a small number of labeled target samples. The study uses a rheology-informed dataset containing 379 observations across 41 pharmaceutical ink formulation domains and applies formulation-wise cross-domain evaluation. Feature selection and explainable machine learning are also incorporated to identify the rheological and processing variables most relevant to printability. The overall goal is to reduce empirical trial-and-error, material consumption, and experimental effort while enabling more efficient and reliable development of printable pharmaceutical formulations.

Rational Design and Synthesis of High-Performance Solid-State Electrolytes for Next-Generation Lithium Batteries

This project focuses on the rational design and synthesis of high-performance solid-state electrolytes (SSEs) for next-generation all-solid-state lithium batteries. The study targets two technologically important ceramic electrolyte families: superionic conductors, particularly lithium aluminum titanium phosphate (LATP), and garnet-type lithium lanthanum zirconium oxide (LLZO). First-principles simulations will be used to screen different compositions and stoichiometries based on key properties, including electrochemical stability and mechanical performance. The most promising candidates will then be experimentally synthesized, with particular emphasis on wet-chemical processing followed by controlled calcination. This approach is designed to achieve improved compositional homogeneity, controlled particle morphology, and tunable porous nanostructures that can facilitate Li-ion transport compared with conventional dry-processing routes. Synthesis parameters will be systematically correlated with phase formation, microstructure, and ionic conductivity to identify optimal processing–structure–property relationships. Advanced synchrotron X-ray and neutron scattering, together with electrochemical impedance spectroscopy, will validate the predicted structural modifications and transport behavior. The project ultimately aims to establish a computation–synthesis–characterization feedback loop for developing SSEs with improved ionic conductivity, providing fundamental guidance for scalable, high-performance solid-state battery technologies.

Page updated

Google Sites

Report abuse