Calibration of ESCAP resist simulation parameters from consideration of printed CD pitch bias, CD measurement offset and wafer thermal history

Abstract

In this work an automated optimization routine is used to modify modeling parameters for a chemically amplified photoresist, with the goal of minimizing the error observed between lithography simulation and experimental results. It is shown that a basic tuning procedure modifying, optimizing only CD measurement offset and acid generation efficiency, improves the fit significantly. Further improvements can be made by optimization of the diffusiondeprotection kinetic parameters, in combination with the two aforementioned values. It is shown further improvement is observed if the actual temperature profile experienced in the postexposure bake process is considered and the temperature dependence of both the diffusion and the deprotection processes are optimized. This parameter values that result in this improvement infer a temporal offset in the start, and finish, ofdeprotection and acid diffusion.

Processing chemically amplified resists on advanced photomasks using a thermal array

“…The variation in linewidth for this thermally activated process ranges generally between 1 and 10 nm per 1°C change in bake temperature [2]. With linewidths decreasing below 100 nm, there is a need to minimize temperature variations and maximize process robustness for thermal processing of resists, otherwise a significant portion of the process error budget will be consumed [3][4][5][6][7][8][9]. Recently, chemically amplified photoresists (CARs) have been used in the manufacturing of advanced photomasks where critical dimension (CD) specifications have required the introduction of laser pattern generators that operate in the wavelength range of 248-257 nm [10].…”