返回
Template Design for Complex Block Copolymer Patterns Using a Machine Learning Method
DOI:10.1021/acsami.3c05018.png)
摘要
En 中文
This study represents the first attempt to address theinversedesign problem of the guiding template for directed self-assembly(DSA) patterns using solely machine learning methods. By formulatingthe problem as a multi-label classification task, the study showsthat it is possible to predict templates without requiring any forwardsimulations. A series of neural network (NN) models, ranging fromthe basic two-layer convolutional neural network (CNN) to the largeNN models (32-layer CNN with 8 residual blocks), have been trainedusing simulated pattern samples generated by thousands of self-consistentfield theory (SCFT) calculations; a number of augmentation techniques,especially suitable for predicting morphologies, have been also proposedto enhance the performance of the NN model. The exact match accuracyof the model in predicting the template of simulated patterns wassignificantly improved from 59.8% for the baseline model to 97.1%for the best model of this study. The best model also demonstratesan excellent generalization ability in predicting the template forhuman-designed DSA patterns, while the simplest baseline model isineffective in this task.
Keyword:
block copolymer
directed self-assembly
machinelearning
inverse design
lithography
期刊
IF:
8.2
论文数:
6.1W
被引数:
38.7W
机构
引用论文
Self-Directed Self-Assembly of 3D Tailored Block Copolymer Nanostructures三维定制嵌段共聚物纳米结构的自定向自组装
ACS NANO
IF16
Pattern Placement Accuracy in Block Copolymer Directed Self-Assembly Based on Chemical Epitaxy
ACS NANO
IF16
Reversible Morphology Control in Block Copolymer Films via Solvent Vapor Processing: An in Situ GISAXS Study
MACROMOLECULES
IF5.2
Directed Assembly of a Cylinder-Forming Diblock Copolymer: Topographic and Chemical Patterns
MACROMOLECULES
IF5.2

