phase behavor of architecture-controlled bottlebrush copolymer system using graph convolutional network

In this work, we present studied the phase behavior of architecture-controlled bottlebrush copolymer in solution system using Graph Convolutional Network (GCN) based on Dissipative Particle Dynamics (DPD) simulation dataset. study with Graph Convolutional Neural Network (GCN) for predicting phase behavior from single chain parameters. Bottlebrush copolymer architecture was encoded by graph including connectivity, side chain length, bead types, and repulsion parameter of DPD simulation. First, single bottlebrush copolymer chain properties such as radius of gyration, volume of chain, asphericity, etc., was predicted using GCN with over Initially, the result of GCN show over 95% accuracy about single bottlebrush chain properties, such as radius of gyration, volume of chain, asphericity, etc., with the graph. Second, we put thesusede predictedthese single chain properties to predict into multi-chain self-assembly behavior in solution using multinomial classification model to learn morphologies of bottle brush copolymer in selective solution states. With this model, we made generated the phase diagram of self-assembled morphologies of bottlebrush copolymers having various architectureswe did not simulate. This work could significantly reduce the searching space for designing target morphology of bottle brush copolymers and provide information which single chain properties influence the phase behavior in selective solventsolution state.