Mask Defect Detection by Combining Wiener Deconvolution and Illumination Optimization

Abstract

In the extreme ultraviolet (EUV) lithography process, mask defect is inevitably replicated on chips hence the yield and quality of the product are directly related to the mask quality. Mask microscopy resolution is then an essential specification. In this work, a high-efficiency method for enhancing the resolution of mask defect is proposed based on illumination optimization and Wiener deconvolution. To validate this approach, we established a verification apparatus designed to achieve a theoretical resolution of 3.0 μm with visible light. Remarkably, the empirical results demonstrated that the actual resolution attained is as low as 2.5 μm, which can be extrapolated to 53 nm resolution based on the wavelength ratio to an EUV source. The verification demonstrates a significant improvement for various periodic fringes. Moreover, the augmented capability of the apparatus facilitates the identification of mask defects.