Stretching-Dependent Charge Mobility Prediction Using Machine Learning in Single-Crystal Flexible Electronics

High-performance molecular single crystal materials are seen as promising candidates for next-generation flexible electronics. Their electromechanical properties are crucial factors that will impact the performance of flexible electronic devices. Here, a method combining molecular dy-namics simulations, machine learning, and charge transport theory was developed to obtain the mechanical strain-charge mobility relationship for molecular single crystals. The simulation re-sults show that mechanical strains cause charge mobility anisotropy variation. Specifically, in pentacene crystal, stretching along the x-axis will enlarge the charge mobilities along all direc-tions in the xy-coordinate plane, while strain along the y-axis will reduce the charge mobilities. These results are due to the charge transport network change caused by the mechanical opera-tions. When the pentacene crystal is stretched along the x-axis, the inter-molecular distance de-creases, resulting the increase of electronic coupling which favors charge transport. When the pentacene crystal is stretched along the y-axis, however, the inter-molecular distance increases instead. This further hinders effective charge transport. It was also found that dynamic disorder plays a crucial role in determining charge transport properties of molecular single crystals since static electronic coupling value gives wrong description of charge mobility anisotropy. Our find-ings provide microscale knowledge about the dependence of molecular arrangement and charge mobility anisotropy on external stretching in molecular single crystals, which can help to broad-en the application of molecular single crystal in flexible electronics.