Reflectance spectral characteristics and imaging mechanism of silicon-based lithography chips based on coaxial broadband illumination

Abstract

The optical performance of silicon-based lithography chips is crucial for optimizing optoelectronic devices. However, as lithography processes advance to the nanoscale, traditional single-wavelength light sources face challenges in resolving the complex surface structures of silicon-based lithography chips. To address this issue, this study proposed a tunable broadband spectral illumination microscopy imaging method based on reflection characteristics for chip detection, and explored the effects of different lighting conditions on lithographic imaging quality. Through an in-depth analysis of the interaction between light and silicon-based lithography chips, the study revealed the feedback effects of illumination conditions on surface electric field distribution and imaging performance. Using finite element simulations, a silicon-based chip model was established to simulate surface electric field distribution and reflection spectral characteristics under different wavelength light sources. The results clearly indicated the significant role of short-wavelength light in exciting surface charges and enhancing the local electric field. The study identified the key influence mechanisms of multi-wavelength light sources in optimizing lithographic imaging quality. The findings demonstrated that the green wavelength band at 525 nm exhibited the best performance in exciting surface charges and enhancing the local electric field. The average reflectance reached 38.45 % and 41.37 % on simple and complex surfaces, respectively, while the Strehl ratio was the highest at 0.49 and 0.43, indicating an effective improvement in imaging quality. Additionally, the study confirmed that the primary mechanism for electric field enhancement was the localized surface plasmon resonance effect. This study provides a theoretical basis for developing characterization methods for complex surface structures.