反射(计算机编程)
半导体
材料科学
光谱学
光学
薄膜
光电子学
计算机科学
物理
纳米技术
量子力学
程序设计语言
作者
Jiaxing Sun,Zhisong Li,Haojie Zhang,Jinlong Song
标识
DOI:10.1016/j.optlaseng.2024.108065
摘要
The thickness of semiconductor thin films stands as a pivotal parameter dictating their operational efficacy, directly shaping their applications in the realms of electronics and optoelectronics. Nevertheless, the quandary of measuring the thickness of widely utilized hundred-micron-scale semiconductor wafers continues to vex researchers. Existing measurement techniques grapple with issues of inadequate precision, poor stability, and sluggish measurement speeds. To surmount this challenge, this study innovatively devises a method employing reflective spectral fitting technology for film thickness measurement coupled with adaptive correction. This approach integrates an enhanced frequency fitting technique, endowing it with the capability for swift and stable measurements of hundred-micron-scale semiconductor thin films. Simultaneously, addressing measurement errors stemming from complex environmental factors in industrial production, the paper proposes a thickness correction method based on carrier interference principles. This correction method ensures precise thickness calibration even when the sample isn't perfectly perpendicular to incident light, significantly augmenting measurement accuracy and stability. Experimental data demonstrates that employing the proposed methodology enhances measurement accuracy by approximately 50 % and elevates measurement stability by about 80 % compared to existing methods. Particularly noteworthy is its capability to maintain high measurement precision and stability even when the sample isn't entirely perpendicular to incident light. The calibrated measurements exhibit around a 55 % reduction in error compared to uncorrected results. This study presents an efficient and accurate solution for the challenging task of measuring hundred-micron-scale semiconductor thin film thickness in industrial settings.
科研通智能强力驱动
Strongly Powered by AbleSci AI