薄脆饼
干扰(通信)
材料科学
基质(水族馆)
光电子学
炸薯条
晶圆规模集成
理论(学习稳定性)
准确度和精密度
光学
电子工程
计算机科学
物理
电信
工程类
频道(广播)
海洋学
量子力学
机器学习
地质学
作者
Jiaxing Sun,Zhisong Li,Haojie Zhang,Jinlong Song,T. Zhai
标识
DOI:10.1016/j.rinp.2023.107146
摘要
The precise measurement of the thickness of substrate and epitaxial layers in epi-wafers is essential in the semiconductor chip fabrication process. This study presents a new method for measuring the thickness of epi-wafers using Fourier Transform Infrared (FTIR) technology, which allows for precise, simultaneous, and fast measurement of both substrate and epitaxial layers. To address the low accuracy of traditional FTIR techniques, a novel VMD-LSP algorithm is introduced, capable of precisely extracting frequency components from the original spectral signal to calculate the precise thickness of the substrate and epitaxial layers. This study also introduces an adaptive spectral excision algorithm and a high-precision thickness extraction algorithm to handle variations in the infrared light absorption of different epi-wafer materials, leading to weakened interference signals, and the challenges that stem from the unique double-layer structure of the wafer. These algorithms efficiently reduce the impact of poor spectral quality and multiple interferences on measurement results, resulting in improved measurement accuracy and stability. During the experimental phase, we compared our results obtained by ellipsometry as truth to evaluate the accuracy and stability of our proposed method. The results show an 85% reduction in measurement error and a 65% improvement in stability compared to traditional methods. As a result, our method meets the need for fast and efficient epi-wafer thickness measurement in semiconductor chip manufacturing, providing a feasible and effective solution to the challenge of epi-wafer thickness measurement.
科研通智能强力驱动
Strongly Powered by AbleSci AI