Autonomous Construction of Phase Diagrams of Block Copolymers by Theory-Assisted Active Machine Learning

Abstract

Equilibrium phase diagrams serve as blueprints for rational design of nanostructured materials of block copolymers, but their construction is time-consuming and requires profound expertise. Herein, by virtue of the knowledge of self-consistent field theory (SCFT), the active-learning method is developed to autonomously construct the phase diagrams of block copolymers. Without human intervention, the SCFT-assisted active-learning method can rapidly search the undetected phases and efficiently reproduce the complicated phase diagrams of diblock copolymers and multiblock terpolymers via decreasing the number of sampling points to about 20%. It is clearly demonstrated that the combined uncertainty sampling/random selection scheme in the active-learning method shows the outperformance in spite of a small amount of initial data set. This work highlights the promising integration of theoretical modeling with machine learning and represents a crucial step toward rational design of nanostructured materials.

Dimensional control over metal halide perovskite crystallization guided by active learning

“…All calculations were performed using Crystals v. 14.23c. 32 Table S4. Experiment variables and their constraints in primary screening.…”

Dimensional Control over Metal Halide Perovskite Crystallization Guided by Active Learning

“…This efficiency is beneficial for constructing material phase diagrams, [31][32][33] -AL has been used to accelerate the acquisition of phase and composition diagrams of multicomponent materials including ferroelectric ceramics, 34 piezoelectric materials, 35 phase-change materials, 36 catalysts, 37 and metal halide perovskite thin films. 38 In these workflows, samples have been typically labeled using data acquired from simulations, [31][32][33] existing datasets, 37 or high-throughput characterization of existing material libraries. 36,37,38 Controlling the dimensionality of MHP derivatives requires the analogous task of mapping the reaction conditions that produce specific phases in different regions of synthetic composition space.…”

“…32 Complete crystallographic data, in CIF format, has been deposited with the Cambridge Crystallographic Data Centre. CCDC 210021 and 2110022 contain the supplementary crystallographic data for this paper.…”

Selection of Advances in Theory and Simulation during the First Decade of ACS Macro Letters

“…This work also highlights the importance of model interpretability and domain knowledge in assessing the robustness and reliability of models built on practical chemical data sets, which typically possess strong internal structure and chemical bias. Li, Lin, and co-workers utilized an active-learning method in conjunction with SCFT calculations of the standard Gaussian model of copolymers at randomly and uncertainty-selected parameter points to efficiently train a high-quality surrogate model for the construction of the complex phase diagrams of multiblock copolymers …”

“…Li, Lin, and co-workers utilized an active-learning method in conjunction with SCFT calculations of the standard Gaussian model of copolymers at randomly and uncertainty-selected parameter points to efficiently train a high-quality surrogate model for the construction of the complex phase diagrams of multiblock copolymers. 50 In the context of biopolymers, the relation between chemical sequences of short amphiphilic biopolymers and the freeenergy landscape of translocation across a lipid membrane has been explored via autoencoder neural networks. 51 This paper by Werner also discussed the physical interpretation of the model's low-dimensional latent space.…”