Automated Identification of Ordered Phases for Simulation Studies of Block Copolymers

In unit cell simulations, identification of ordered phases in block copolymers (BCPs) is a tedious and time-consuming task, impeding the advancement of more streamlined and potentially automated research workflows. In this study, we propose a scattering-based automated identification strategy (SAIS) for characterization and identification of ordered phases of BCPs based on their computed scattering patterns. Our approach leverages the scattering theory of perfect crystals to efficiently compute the scattering patterns of periodic morphologies in a unit cell. In the first stage of the SAIS, phases are identified by comparing reflection conditions at a sequence of Miller indices. To confirm or refine the identification results of the first stage, the second stage of the SAIS introduces a tailored residual between the test phase and each of the known candidate phases. Furthermore, our strategy incorporates a variance-like criterion to distinguish background species, enabling its extension to multi-species BCP systems. It has been demonstrated that our strategy achieves exceptional accuracy and robustness while requiring minimal computational resources. Additionally, the approach allows for real-time expansion and improvement to the candidate phase library, facilitating the development of automated research workflows for designing specific ordered structures and discovering new ordered phases in BCPs.