基于机器学习的光刻坏点检测研究进展  

Research Progress of the Lithography Hotspot Detection Based on Machine Learning

作　　者：盖天洋 粟雅娟[1] 陈颖[1,2] 韦亚一 Gai Tianyang;Su Yajuan;Chen Ying;Wei Yayi(Key Laboratory of Microelectronics Devices and Integrated Technology , Institute of Microelectronics ,Chinese Academy of Science , Beijing 100029, China;University of Chinese Academy of Sciences ,Beijing 100049, China)

机构地区：[1]中国科学院微电子研究所微电子器件与集成技术重点研究室,北京100029 [2]中国科学院大学,北京100049

出　　处：《微纳电子技术》2019年第6期421-428,434,共9页Micronanoelectronic Technology

基　　金：国家科技重大专项资助项目(2017ZX02315001;2017ZX02101004)

摘　　要：基于机器学习的坏点检测技术已经成为光刻坏点检测的重要研究方向,在新技术节点开发与物理设计验证中具有重要意义。按照基于机器学习的光刻坏点检测技术的流程,依次介绍了特征提取、机器学习模型建模和待测样本提取等步骤中面临的问题,综述了近年研究中针对以上问题提出的关键技术及其优劣。对基于机器学习的坏点检测技术的发展方向和面临的挑战进行了展望。目前,完全基于机器学习技术的坏点检测技术中数据生成成本巨大,精度尚不满足集成电路行业应用的要求,因此与光刻仿真模型、图形匹配等传统方法的结合是基于机器学习的坏点检测技术最容易应用于实际生产的技术途径。The hotspot detection technology based on machine learning has become an important research direction of the lithography hotspot detection,and is of great significance in the development of new technology nodes and physical design verification.According to the flow of the lithography hotspot detection technology based on machine learning,the existing problems in the feature extraction,machine learning model building and sample capture are introduced in turn.The key technologies for the above problems proposed in recent years are reviewed,and their advantages and disadvantages are compared.Then,the development trend and challenges of the hotspot detection technology based on machine learning are prospected.In present,the cost of the data generation in the hotspot detection technology entirely based on machine learning is huge,and the accuracy does not meet the requirements of the application of the integrated circuit industry.Therefore,the combination of traditional methods,such as the lithography simulation model and pattern matching may be the best way to apply the hotspot detection technology based on machine learning to the actual production.

关 键 词：机器学习 坏点检测 集成电路物理设计 计算光刻 设计工艺联合优化 

分 类 号：TN305.7[电子电信—物理电子学]