Analyzing the Influence of Source/Drain Growth Height and Lateral Growth Depth in FinFETs Through XGBoost and SHAP

Abstract

In FinFETs, the shape of the source/drain is crucial for performance, as it supplies charge and induces stress in the channel. Due to the 3D structure and numerous components of FinFETs, experimentally analyzing the performance impact of source/drain shapes is time-consuming and costly. This study employs machine learning and the SHAP method to analyze the influence of source/drain shapes on FinFET performance, focusing on growth height and lateral growth depth. These factors’ effects on key performance indicators such as on-current and threshold voltage are confirmed. SHAP analysis further substantiates the results’ reliability and significance. Our findings contribute to understanding and improving the performance of increasingly complex and miniaturized semiconductor device structures.