Intelligent Photolithography Corrections Using Dimensionality Reductions

Abstract

With the shrinking of the IC technology node, optical proximity effects (OPC) and etch proximity effects (EPC) are the two major tasks in advanced photolithography patterning. Machine learning has emerged in OPC/EPC problems because conventional optical-solver-based OPC is time-consuming, and there is no physical model existing for EPC. In this work, we use dimensionality reduction (DR) algorithms to reduce the computation time of complex OPC/EPC problems while the prediction accuracy is maintained. Also, we implement a pure machine learning approach where the input masks are directly mapped to the output etched patterns. While one photolithographic mask can generate many experimental patterns at once, our pure ML-based approach can shorten the trial-and-error period in the photolithographic correction. Additionally, we demonstrate the automation in SEM images preprocessing using feature detection, and this facilitates intelligent manufacturing in semiconductor processing. The input vector dimensions are effectively reduced by two orders of magnitude while the observed mean squared error is not affected significantly. The computation runtime is reduced from 4804 s of the baseline calculation to 10 s-200 s The MSE values changed from the baseline 0.037 to 0.037 for singular value decomposition (SVD), to 0.039 for independent component analysis (ICA), and to 0.035 for factor analysis (FA).

Optical Proximity Correction Using Machine Learning Assisted Human Decision

“…To alleviate the long simulation runtime, numerous machine learning-based OPC models (MLOPC) have been proposed [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29]. Works from R. Frye [28] and P. Jedrasik [29] have implemented unsupervised neural networks for e-beam lithography and optical lithography for OPC, respectively.…”

“…Y. Shin et al use recurrent neural networks for MLOPC [24]. P. Parashar et al use dimensionality reduction techniques to reduce network sizes [21]. Yuan et al use unsupervised learning and general adversarial networks to shorten the OPC runtime.…”

Classification of lines, spaces, and edges of resist patterns in scanning electron microscopy images using unsupervised machine learning

“…To reduce the error in the printing, the earliest corrections in mask design could be tracked back to 1981. 6) Recently, deep learning (neural networks) has been used in the field of optical lithography for optical correction, [7][8][9][10][11][12] hotspot correction, [13][14][15][16][17][18][19] etching bias correction, 11,[20][21][22] and simulation model construction. 23) The neural networks were not only used for prediction and defect correction but also implemented in image classification.…”

TCAD augmented generative adversarial network for hot-spot detection and mask-layout optimization in a large area HARC etching process

“…PLEASE CITE THIS ARTICLE AS DOI: 10.1063/5.0093076 time-consuming, and there is no reliable physical model for high aspect ratio etch proximity correction (HEPC). 9 Moreover, many actual experimental data, e.g., crosssectional etch profiles for various process operating recipes (RCPs) and layout designs, are needed to achieve the high accuracy of these proximity correction techniques. However, it is difficult to get actual etching profile data at the early stage of fabrication development and optimization for the next generation (higher stack layers) 3D NAND device.…”