Blank defect coverage budget for 16nm half-pitch single EUV exposure

Multilayer defects (ML-defects) are the most specific type of defects on a mask for extreme ultraviolet (EUV) lithography. The intent of this paper has been to study the practical limits of the pattern shift technique to cover such defects by the absorber pattern. We have targeted to apply pattern shift to a 16nm half-pitch EUV single exposure, interconnect-like layer, in which absorber features are predominantly as small as 64nm at mask level. Three main contributors to successful defect coverage are the lateral size of the defect, the alignment of the mask pattern to the fiducial marks and the location accuracy of the blank defects relative to these fiducial marks. For our experimental analysis, we have used a specific approach in which, rather than explicitly targeting to cover the defects to render them non-printing, we kept the possibility to study their printability, together with the possibility to assess the achieved alignment of the mask pattern to the defectivity map of the blank. This was achieved by superimposing a dedicated marker frame, with the expected defect position at its center, onto a lines-and-spaces (l/s) pattern with 16nm half-pitch (at wafer scale). The marker frame allows to determine the deviation of the defect position from the expected one, and the printing impact of the defect on the l/s pattern can be compared to its expected behavior based on its relative position within the 32nm period. It is shown that mitigation feasibility is strongly dependent on the accuracy of the defect position information. Our results suggest targeting to improve that.