Enhancing displacement Talbot lithography through inverse-designed curvilinear masks and multibeam mask writing

Displacement Talbot Lithography (DTL) is a photolithography technique utilized for printing periodic patterns, finding use especially in various photonic applications. Unlike conventional proximity or projection lithography, DTL's unique image formation process brings challenges in mask design, particularly for complex wafer patterns requiring non-Manhattan (curvilinear) features on the mask. Writing such curvilinear masks poses challenges for traditional variable shaped beam writers. This paper presents a collaborative effort to create high-resolution DTL masks with curvilinear geometries, leveraging Inverse Mask Design (IMD) algorithms and the capabilities of state-of-the-art multi-beam mask writers (MBMW). Our method addresses the challenge of making suitable masks for DTL, especially for intricate target patterns. By combining the benefits of DTL, machine learning algorithms, and cutting-edge mask writing techniques, this approach expands DTL's potential applications and enhances the utility of MBMWs in advanced lithography processes.