For the last 5years a joint venture has pursued a research program studying and enhancing the ability of optical inspection tools to meet the inspection needs of extreme ultraviolet (EUV) and other next generation lithographies (NGLs). In this article we present a survey of results we have obtained for patterned inspection of NGL masks. The NGL technologies that we have studied include two electron projection lithographies, EUV, and step and flash imprint lithography (SFIL). We discuss the sensitivity of the inspection tools and mask design factors that affect tool sensitivity. In contrast to conventional photomask inspection, which primarily utilizes transmitted light for inspection, almost all NGL mask inspections are performed in reflected light. Much of the work has been directed towards EUV mask inspection and how to optimize the mask to facilitate inspection. Early EUV masks had an optical contrast of 40% or lower. Our partners have succeeded in making high contrast EUV masks ranging in contrast from 70% to 98%. Die to die and die to database inspections, at a wavelength of 257nm, of EUV masks have been achieved with a sensitivity that is comparable to what can be achieved with conventional photomasks, with a minimum detected defect size of 80nm square defects. We have inspected scattering with angular limitation projection electron-beam masks successfully. Electron-beam stencil masks, such as that used in projection reduction exposure with variable axis immersion lenses, pose a problem in that their high aspect ratio of mask thickness to minimum feature width results in low resolution transmission images. Reflected light images provide high-resolution images suitable for inspection, but will not be sensitive to defects below the inspection surface. We have run inspections on SFIL masks in die to die, reflected light in an effort to provide information concerning possible cumulative damage due to imprinting and to provide feedback on defect densities, types and sizes to improve the masks. Our defect sensitivity on SFIL masks is approximately 100nm, though we cannot run at the highest sensitivity due to large numbers of nonprogrammed defects. We have also used an inspection system, at a wavelength of 364nm, to inspect both unpatterned EUV substrates (no coatings) and blanks (with EUV multilayer coatings), and demonstrated a sensitivity of approximately 100nm to polystyrene latex spheres. This information has helped drive down the defect densities on EUV blanks and substrates by some three orders of magnitude. Extensions of conventional optical lithography have pushed out the introduction of NGL technology to the 33nm or possibly the 22nm node. Mask inspection technology will need substantial improvements in resolution to meet the NGL inspection requirements below the 45nm node.
