Automatic Prediction of Metal-Oxide-Semiconductor Field-Effect Transistor Threshold Voltage Using Machine Learning Algorithm

Abstract

A fast and precise threshold voltage (V-th) extraction method is required for the process design of electronic systems using metal-oxide-semiconductor field-effect transistors (MOSFETs) and its immediate on-site analysis during fabrication. The selection of a suitable V-th extraction method is a complicated task because it involves a trade-off between accuracy and simplicity according to the device scheme. Herein, an automatic-prediction method of the MOSFET V-th using machine learning (ML) is proposed. The ML model is trained with V-th, extracted using different methods (2nd derivative, constant current, and Y-function) and from various kinds of FETs (finFET, 2D FET, and metal-oxide thin-film transistors). The concept of threshold ratio (R-th) for universal V-th prediction, which considers the normalized V-th within certain V-G ranges, is suggested. The precision and accuracy of ML models are statistically verified by calculating the root mean square error (RMSE), mean absolute error, and mean coefficients of determination (R-2) values. The universal ML model (k-nearest neighbor (kNN)) achieves 1.35% of RMSE and 0.98 of R-2 for the best score. The ML model eliminates the ambiguity in V-th extraction and provides objective V-th prediction for most FET schemes used in the semiconductor industry and research field.