Prediction and Analysis of Film Stress and Warpage based on Machine Learning

Abstract

Stoney equation is widely used to estimate the film stress in semiconductor industry however it is accurate only under particular situation The purpose of this thesis is to predict film stress and wafer warpages while the Stoney equation could not be used such as multilayers structure finite deformation and discontinue films To achieve the goal several machine learning models are constructed based on the finite element database The final machine learning model is chosen with the best accuracy The final models constructed in this work could predict the film stress or the radius of curvature To get the warpage of the deformed wafer a topology-reconstruction algorithm is proposed and be demonstrated in two and three-dimensional with good performance Finally the machine learning model and the algorithm are integrated in graphical user interface

Date of Award	2020
Original language	English
Supervisor	Kuo-Shen Chen (Supervisor)