Numerical Simulations and Machine Learning for Flexible Materials

Shuzhou Li

School of Materials Science and Engineering, Nanyang Technological University, Singapore (lisz@ntu.edu.sg)

Abstract

Polymer semiconductors as a key component of electronic skin need to maintain the coexistence of stretchability and electrical functionalities. However, repeated stretching-releasing cycles inevitably lead to the charge mobilities decreasing and poor working performance of polymer semiconductors. We developed a method combining molecular dynamics simulations and charge transport theory to obtain the morphology-mobility relationship of amorphous Poly(3-hexylthiophene) (P3HT). The simulation results show that the hole mobility decreases by 6% along the strain direction after three stretching-releasing cycles with 80% strain. These results are due to the chain alignment change caused by the mechanical operations. The stretched P3HT material presents higher charge mobility due to its better chain alignment while the compressed P3HT shows lower charge mobility because of the poor chain alignment. Repeated stretching-releasing cycles lead to the chain alignment parameters decreasing along the deformation direction with accumulation and saturation effects. The repeated cycles also result in the primitive path length decreasing, which indicates polymer chain spatial distribution more localized after repeated deformations.

A large database is desired for machine learning (ML) technology to make accurate predictions of materials physicochemical properties based on their molecular structure. When a large database is not available, such as in elastomer materials, development of proper featurization method based on physicochemical nature of target proprieties could also improve predictive power of ML models even if with smaller database. We introduce structure-based multilevel (SM) descriptors of elastomers, derived solely from molecular structure that is universally available. With our SM descriptors, a machine learning model accurately predicts elastomer toughness with a remarkable accuracy of 0.91. Furthermore, an high-throughput screening pipeline is established to swiftly screen elastomers with targeted toughness. The user-friendly and low computational cost SM descriptors make them a promising tool to significantly enhance high-throughput screening in the search for novel materials.