Innovative approach for roughness characterization of arbitrary polygons

In the era of extreme ultraviolet lithography, reducing pattern roughness is crucial for improving device efficiency and reliability. Line edge roughness (LER) and line width roughness (LWR) are common metrics used to characterize roughness on horizontal or vertical line structures in SEM images. Obtaining accurate LER and LWR information requires to follow specific guidelines as described in the so called IMEC protocol.¹ One of the most important rules of the protocol is not to use any filtering operation during the detection of the lines contour in the SEM image. Besides, the protocol is limited to the case of lines only. Generalizing this protocol to more complex patterns has not been achieved so far without using filtering which inherently makes the roughness results unexploitable.² In this paper we propose a novel methodology to extend the validity of the protocol for extracting roughness from arbitrary patterns in SEM image.