Petr Cizmar, Andras E Vladar, and Michael T Postek

Nanometer Scale Metrology Group, Precision Engineering Division, MEL, NIST


Development of novel accurate and quick imaging and metrology techniques is inevitable for future improvements in semiconductor industry and biological research. Current imaging techniques in the scanning electron microscopy (SEM) suffer from severe distortions, especially, where high magnification is needed. The same issue appears in the scanning Helium-ion microscopy as well. Image distortions cause significant errors in the nanometer-scale metrology, since its methods mostly work with SEM images or line-scans. Main origin of these distortions is the sample drift with respect to the charged-particle beam. However, there are methods, currently developed at NIST, that allow for efficient correction of these drift-related distortions, which include the employment of a dual-axis laser interferometer, or drift-corrected image composition.


For assessment of the new techniques, the artificial SEM (and Helium-ion microscope) image modeling library (Artimagen) has been developed. This library, implemented in C++) is capable of modeling artificial images, which very closely resemble real ones. All important effects present in the SEM images are deterministically and repeatably modeled, while the ones in real images are neither repeatable, nor deterministic. This is the fundamental advantage of the modeled images over the real ones. The Artimagen library has been released as a public-domain, open-source, and multi-platform software.