A calibrated photoresist model for pattern prediction

Abstract

Aerial image metrics and non-rigorous resist modeling have been used in computational lithography to estimate the resist CD on the wafer, with varying degrees of success. However, the complex illumination and optical-proximity correction strategies used at the 90 nm node and below demand accurate simulation at the wafer level, allowing engineers to make informed decisions when optimizing the process. A calibrated, rigorous model simulating the physical-chemistry of exposure, reaction-diffusion and development can produce highly-accurate predictions, even when extrapolated. The advantage of the rigorous physical approach is that simulations retain predictive accuracy even when the experimental conditions are modified.In this work, the rigorous physical resist model available in PROLITH is calibrated for a commercially available ArF resist using top-down CDSEM data, including focus-exposure matrices, CD vs. mask pitch and mask-error enhancement data. All calibration data consist of 1D mask structure e.g. lines and spaces. It will be shown that the resulting resist model, while calibrated using 1D mask structures, is able to accurately predict the performance of 2D mask structures with an area fidelity metric of 98.34% under the illumination conditions used for calibration. It will also be shown that, additionally, the calibrated resist model is able to predict 2D patterns robustly under alternate illumination schemes without the need for re-calibration. The resist model accurately describes the experimental results, validating the extrapolated model predictions. Accurate, calibrated models can be used to improve the process with much less impact on time and resources than empirical methods i.e. wafer fabrication and inspection.

Impact of mask three-dimensional effects on resist-model calibration

“…Calibration of full-physical resist models [1][2][3][4] has a long record of use and is constantly being refined to offer better predictability to rigorous simulators, such as Sentaurus Li-thography™ of Synopsys or Prolith™ of KLA-Tencor. We used a large experimental critical dimension ͑CD͒ data set, which was generated for the purpose of optical proximity correction ͑OPC͒-model building, to assess the importance of mask 3-D ͑or mask-topography͒ effects in resist-model calibration.…”