Estimation of critical dimension and line edge roughness using a neural network

While electron-beam (e-beam) lithography is widely employed in the pattern transfer, the proximity effect makes features blurred, and the stochastic nature of the exposure and development processes causes the roughness in the feature boundaries. In an effort to reduce the proximity effect and line edge roughness (LER), it is often necessary to estimate the critical dimension (CD) and LER. In our previous study, the e-beam lithographic process was modeled using the information extracted from SEM images for the estimation of CD and LER. This modeling involves several parameters to be determined and tends to require a long computation time. In this study, the possibility of improving the accuracy of the CD and LER estimation using a neural network (NN) is investigated. In the NN-based estimation, the explicit modeling of the e-beam lithographic process can be avoided. This paper describes the method of estimating the CD and LER using a NN, including the issues of training, tuning, and sample reduction and presents results obtained through an extensive simulation.