Fidelity of rectangular patterns printed with 0.3-NA MET optics

Abstract

Arrays of rectangular patterns of various sizes were printed with the EUV micro-exposure tool (MET) at the Lawrence Berkeley National Laboratory (LBNL) using the chemically-amplified resist MET-1K; and their fidelity to the mask patterns was evaluated. The experimental results showed that the shortening of resist patterns in the lengthwise direction was greater for smaller patterns. For example, the line-end shortening of half-pitch (hp) 45-nm patterns was about 20-25 nm on one side, while that of hp-90-nm patterns was less than 10 nm. However, simulated aerial images exhibited little shortening, even for hp-45-nm patterns. On the other hand, considerable shortening appeared in hp-45-nm patterns after post-exposure-baking (PEB) process. When the acid diffusion length in the PEB process was assumed to be 20 nm, the calculated shapes of resist patterns agreed well with the experimental results for various sizes. Printing experiments showed that lowering the PEB temperature improved fidelity, probably due to the shorter acid diffusion length. Thus, we concluded that acid diffusion is the main cause of shortening in rectangular patterns printed with the MET. For better pattern fidelity, the acid diffusion length must be reduced in accordance with the reduction in pattern size.

Electron and Hole Transfer in Anion-Bound Chemically Amplified Resists Used in Extreme Ultraviolet Lithography

“…It has been speculated that the main cause of resist blur in a chemically amplified resist is acid diffusion. [3][4][5] The increase in acid generator concentration and the suppression of the subsequent acid diffusion is one way to simultaneously improve resolution and LER. [6][7][8][9][10][11] On the other hand, the feature size is shrinking so markedly that the uniformity of acid generator distribution in resist matrices is becoming a serious problem.…”