Quantum lithography

Abstract

The edge definition and the interior filling of pattern features are commonly performed using the same exposing beam regardless of the feature size. Separating the two processes, as first proposed by Fulton et al. [Appl. Phys. Lett. 42, 752 (1983)], improves the efficiency of the mask making process and adds ‘‘smartness’’ to the substrate, i.e., the pattern generating tool need only provide the minimum information; the substrate and process are configured to convert the minimum information into the necessary pattern. We describe an improved level of smartness on substrates: we view the mask as a bitmapped lattice of independent pixel elements (pels), each having a given shape and carrying the relevant edge information; the desired pattern would be generated by selectively addressing a subset of the lattice and modifying the optical properties of its pels. The feature edges are effectively ‘‘quantized’’ in the sense that they can exist only in some specific places. In the edge definition step, a very fine grid whose edges are within the specifications on edge positional accuracy and precision, is inscribed on the mask blank by its manufacturer using a high resolution and high precision lithographic technique. The pitch is such that the size of the resulting tiles equals the minimum feature size of the pattern to be delineated. The limitation on the grid width is that it is not imaged by the optics upon the exposure of the mask. Once the grid is defined, the pattern can be customized by the user simply by tagging those tiles that constitute the pattern. We describe here the general principles, advantages, fabrication, and exposure of ‘‘quantum lithography’’ masks.