|
An Extreme Ultra-Violet (EUV) photo mask is a complex engineering marvel. It is integral to semiconductor manufacturing, as it holds the design layout information intended for the device. The mask complexity allows many potential points of variation during the mask manufacturing process. Some of these deviations from the ideal mask remain benign to wafer patterning but others impact the wafer pattern. To date, reported studies have characterized the impact of local mask defects in the absorber pattern and within the multilayer mirror of an EUV mask [1], as well as various non-local mask deficiencies such as multilayer mirror surface roughness and contamination growth [2]. The results revealed relationships between mask quality limitations and stochastic failures on the wafer.
The studies to date were based on line-and-space patterns to reduce optical complexity and were completed using NA 0.33. The study proposed will extend the research effort into two-dimensional (2D) structures such as contacts, via, and tip-to-tip containing line/space structures, to further understand the impact of mask defects in complex imaging configurations. The study will also investigate these 2D structures patterned with high-NA EUV lithography. As a start, one-dimensional structures will be imaged with high-NA EUV using conditions from previous studies to provide a continuous baseline.
The introduction of 2D structures brings optical and computational complexities into the study. The optics contain more complex information than found in the line/space patterns used in previous studies. This may lead to additional impacts not observed in previous work. In addition, due to the complexities of these patterns, the computational effort becomes larger, which may necessitate improved algorithms or learning methodologies. Additionally, in the anamorphic high-NA system the choice of pattern orientation in the 4x or 8x direction is thought to be based on critical feature type such as tip-to-tip or minimum CD, but the introduction of stochastic mask analysis may influence this preference. This work is expected to contribute to a better understanding of which mask quality aspects must be tightened for high-NA EUV.
|
