Background: The Extreme UltraViolet (EUV) lithography is known as the technology that enables the scaling of semiconductor devices for high-volume manufacturing. The emerging high Numerical Aperture (NA) EUV lithography scanners increase NA from 0.33 to 0.55, enabling continued shrink down to 8 nm half-pitch, while EUV systems with NA ≥ 0.75 (hyper NA) are being explored to support further cost-effective increase in device density. To achieve the productivity targets, EUV photoresist materials and their related processes play a significant role. The effectiveness of a photoresist for microelectronic fabrication depends on several factors, among which high resolution, high sensitivity, low pattern variability and zero defectivity are the key requirements. Aim: Due to the scarcity of EUV photons hitting the wafer, high sensitivity photoresists are strongly sought after to reduce the costs of the lithographic steps while ensuring sufficient throughput. Therefore, photoresists with high EUV absorption are required to minimize the stochastic effects on patterning, such as photon and chemical shot noise, and to enable the required solubility switch during the wet or dry development step. Thus, high EUV photon absorption in photoresist is a key parameter to have high patterning performance while continuing the dimensional scaling of features. Approach: In this work we address the challenges of the reduced Depth of Focus (DoF) at the wafer level, due to the increase of NA3, by looking at the photoresist requirements in terms of photon absorption. Through rigorous stochastics simulations of imaging performance on a relevant patterning test case, we vary the EUV absorption by introducing a gradient extinction coefficient through the photoresist film. We study the effect of this gradient on stochastic effects across focus at 0.55 NA and compare it with the reference metal oxide resist. Ta-based and low-n EUV masks in combination with metal oxide resist are studied at 0.33 NA, 0.55 NA and 0.75 NA. Furthermore, we investigate potential solutions such as focus drilling to measure the validity of this approach from a theoretical point of view to mitigate the DoF reduction. Results: Best defectivity free processes are with the 6% low-n mask reflectivity at 0.33 NA and 0.55 NA. The process at 0.75 NA is the most defective. The lowest sidelobe defectivity count is given by the gradient absorption resist although overall in defectivity such a process resulted two orders of magnitude worse than its equivalent with the metal oxide resist. The gradient absorption resist can uniformly distribute the photons along the z axis; however, this observation does not explain its large defectivity. On the other hand, the number of absorbed photons for the gradient absorption resist through focus are much lower with respect to the reference metal oxide resist. This is in line with the exposure dose that drops from 58 mJ/cm2 (metal oxide resist) to 39.7 mJ/cm2 (gradient absorption resist) at CD target. The increase in photon shot noise could explain the increase in defectivity seen for the gradient absorption resist at best exposure conditions.
Simone et al. (Thu,) studied this question.