Phase field mapping for accurate, ultrafast simulations of directed self-assembly

Abstract

Block copolymer self-assembly is a powerful tool for nanoscale patterning which benefits from predictive simulations. Two classes of simulations are self-consistent field theory (SCFT), which is accurate but computationally expensive, and phase field models, which are faster but historically less accurate. We refine a mapping procedure that uses results from SCFT to optimize parameters in a phase field model for diblock copolymers. We validate the performance of this optimized phase field model with regards to accuracy and computational speed in perfect and defective configurations. The optimized phase field model is significantly faster than SCFT and more accurate than previous phase field models, making it a viable design tool for directed self-assembly processes.