Novel EPE metrology for advanced mask manufacturing

Leading-edge semiconductor IC manufacturing is constantly pushing the boundaries of not only minimum critical dimensions and pitch but also patterning complexity and diversity. As such, there is a growing consensus in the industry to quantify the sum of variations in terms of EPE (Edge Placement Errors). There are several definitions of EPE and how it relates to variations coming from litho, resist stochastics, OPC, mask, etc. These individual component breakdowns do show that the mask contribution to total wafer EPE has been increasing over the recent technology cycles and could account for nearly one-third of total EPE budget. Hence, it's becoming crucial for mask metrology to also evolve to EPE-level characterization. This paper outlines an approach to comprehensively utilize existing mask CD-SEMs to compute accurate mask EPE metrics. By application of e-beam die-to-database modeling and correction of these images, more advanced metrics such as Local Critical Dimension Uniformity (LCDU), area, EPE, and also registration measurements can be computed. These metrics provide front-end engineering teams with a better understanding of the e-beam write and process effects on mask patterning fidelity, and the ability to monitor these in production. This paper describes the development and deployment of such a mask EPE computation system. An adaptable and accurate system offers the advantage of evaluating how the intended Mask Proximity Corrections (MPC) compares to the actual results after e-beam writing and processing. Such comprehensive characterization is found to be relevant not only for critical EUV layers but also optical masks that are now being patterned with complex Curvilinear shapes that have only a few areas where simple CD structures can be measured. The roadmap of mask EPE-driven monitoring, control and optimization methodologies is expected to continually evolve in the upcoming process technologies as total EPE budgets shrink, making it important to have an accurate and adaptable metrology computation system.