Abstract
Substantial discrepancies are commonly observed when comparing the predicted and measured optical performance of deep-ultraviolet tungsten wire grid polarizers. Particularly, the extinction ratio is strongly impaired. Therefore, we investigate major differences between assumed and actual achieved properties regarding geometry and material of the grating structure as the origin of theses discrepancies. We find an improvement potential for the extinction ratio of one order of magnitude by improving the material and a factor of four by improving the geometry. Our results allow for a purposeful revision of fabrication processes and will therefore significantly contribute to the improvement of deep-ultraviolet wire grid polarizers.