Optical proximity correction (OPC) using neural network

Abstract

a popular example is ML-OPC based on regression methods,which however has shown limited prediction accuracy. We propose new ML-OPC, in which a neural network classifier serves as a mask bias model. A few techniques are applied to enhance basic ML-OPC: parameterization of layout segments using polar Fourier transform basis functions, dimensionality reduction through weighted principal component analysis, and sampling of training layout segments. We also propose fast network size optimization by rudundant node pruning and incremental training to increase coverage of ML-OPC. Training segments are typically imbalanced over the range of mask biases, which causes large prediction error for segments that appear less frequently. This is resolved by synthetic data generation, class reorganization, and an adaptive learning rate. Experiments with new ML-OPC indicate that prediction error of mask bias and training time are reduced by 29% and 90%,respectively, compared to state-of-the-art ML-OPC.

Machine learning-guided optical proximity correction (ML-OPC) has recently been proposed to replace extremely time-consuming model-based OPC