Predicting shape-memory effect in thermoplastic polymers using machine learning assisted framework

Shape memory plastics are materials that can recover a pre-programmed shape, in response to a stimulus (often, temperature). Such smart materials find application in biomedical devices, among other areas. Discovery of new shape-memory plastics relies of extensive experimentation, making it slow and expensive. In this project, we will examine the possibility of using machine learning frameworks to correlate shape-memory performance to structural attributes, including, the polymer molecular weight, entanglement density, morphology, crystallinity, glass transition temperature, thermal expansion, etc. These will be correlated using machine learning models, and the predictive capabilities of these models will be evaluated. Data will be obtained from the literature and will be matched with measurements performed in the laboratory. This project is aimed at developing a robust, physics-informed model that can accurately predict the shape memory properties of a polymer, given its structural attributes. This project will be carried out in close collaboration with a company that uses data science to accelerate materials innovation (polymerize.io). The student will have the opportunity to work with scientists from the company. Students interested in this project should have familiarity with python programming. Some exposure to polymer science and engineering and to data science is preferable, but not mandatory.