极紫外光刻
平版印刷术
模具(集成电路)
材料科学
光电子学
计算机科学
纳米技术
作者
Toshiyuki Todoroki,Takashi Hanamoto,Takashi Kamochi,K.I. Gondaira,Arosha Goonesekera,Hiroki Miyai
摘要
The actinic patterned mask inspection tool "ACTIS" has become essential for EUV photomask qualification in mask shops and semiconductor fabs. It provides high-resolution, high-throughput inspection that can detect all types of mask defects with lithographic implications because it uses the 13.5nm EUV light. ACTIS is equipped with a high-brightness EUV light source with an optimally sized etendue. It minimizes the heat load on the pellicles while increasing the number of photons on the detector to improve inspection sensitivity and throughput. High NA EUV lithography is scheduled to be introduced for the angstrom-generation technology node. Lasertec has developed a new model of ACTIS, the ACTIS A300, to meet the requirements of high-NA mask inspection. In high-NA EUV imaging, the effect of pattern edge roughness on inspection becomes more significant because the image resolution is higher than that of the current NA imaging. This paper discusses how Die-to-Database inspection using a machine learning-based reference image generation model minimizes the effects of edge roughness and further improves the sensitivity of A300.
科研通智能强力驱动
Strongly Powered by AbleSci AI