BRITE Relaunch: Leveraging Advances in Artificial Intelligence to Predict Mechanical Behavior of Two-Dimensional Atomic Sheets with Defects

This Boosting Research Ideas for Transformative and Equitable Advances in Engineering (BRITE) Relaunch project will use modeling to study the mechanical behavior of two-dimensional materials formed by laterally interfacing sheets that are one atomic layer in thickness. When the sheets are laterally interfaced, they may form defects within the interfacing region. These defects influence the mechanical behavior, such as elasticity and strength, which are key properties that must be understood for the materials to be integrated into new technology. Advances in artificial intelligence and machine learning will be used to overcome challenges in computational time and resources that are usually needed for traditional modeling of these materials. Accelerating the modeling will advance researchers’ efforts in making new two-dimensional materials. The research will be made widely available to benefit the US economy and society. This BRITE Relaunch addresses fundamental barriers to predictive mechanics of materials and structures needed for modeling of two-dimensional nanomaterials integrated into new devices, such as for flexible electronics or batteries. Specifically, the research will focus on a workflow to predict elastic properties, mechanical stress-strain behavior, and fracture strength of two-dimensional lateral heterostructures with defects at the interface between two-dimensional atomic sheets. The project integrates novel artificial intelligence and machine learning techniques to predict properties that can match the accuracy of traditional modeling with density functional theory. The same workflow will automate the search for defects at interfaces through geometry optimization and automate the data clustering for training of algorithms. This project will serve as the foundation for future research on predictive modeling and inverse design (in which the structure of the material is computed from the desired properties, a long-term goal of the community) of two-dimensional materials by developing a materials agnostic workflow. The findings will accelerate the research communities’ efforts on modeling stress-strain and fracture strength for two-dimensional lateral heterostructures needed to predict the mechanical behavior of new 2D material systems. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.