EUV mask roughness can recover litho-tool aberrations (Conference Presentation)

Speckle from an EUV mask adds to the line edge roughness of the final image in resist, so is typically minimized for better critical dimension control. However, the roughness of the mask can also be utilized constructively, for probing the pupil function of an aerial imaging system or a lithography scanner. The spectrum of the speckle image generated from an EUV mask blank encodes the system aberrations under a weak scattering approximation. We show that the properties of EUV masks are suitable for achieving a good balance between weak scattering and speckle contrast. Using this concept, we demonstrate in-situ experimental recovery of field-of-view dependent aberrations from blank areas of an EUV mask.     EUV masks are naturally rough at the scales seen by 13.5 nm light, creating weak diffused light that fills the entire pupil of the imaging system. Additionally, since most materials are only weakly scattering at soft X-ray wavelengths including EUV, the scattered light acts as a perturbation on the background illumination, recombining with it interferometrically to encode the pupil phase in the final speckle. We present an algorithm based on the phase contrast transfer function to use illumination angle diversity for extracting the pupil phase from the measured speckle spectrum at the camera plane. However, since the contrast transfer function is a linearization of the image intensity in terms of the object phase, it relies on the mask being a weak phase object. The exact properties of the EUV mask roughness needed for the linearization to apply are described. Measurements on the SHARP EUV microscope at the Lawrence Berkeley National Lab on mask blanks shows them to be weakly scattering, while still providing sufficient speckle contrast for aberration estimation. Additionally, the method can be used to probe aberrations across the field-of-view, using the speckle in any blank area of the mask for single-shot in-situ recovery of imaging system aberrations. While the method is shown to work on speckle from an aerial imaging tool, an extension to resist images of speckle from lithography scanner tools is being evaluated. This uses surface profile measurements of the speckle captured as the latent image on the exposed resist (before develop) to quantify aberrations in the lithography tool, under actual operating conditions. The recovered aberrations allow for high resolution reconstruction of the mask image in aerial imaging tools, or for compensating scanner aberrations using source-mask or pupil optimization techniques.