Multi-technique reference measurement of critical dimension and shape for model validation in scanning electron microscopy

In semiconductor device manufacturing, 3D structure measurements of newly complex devices would be valuable alongside conventional scanning electron microscopy (SEM) and critical dimension SEM measurements. Since these measurements are based on the SEM’s grayscale secondary electron (SE) signal intensity, accurate measurement of structure needs the correct physics to link SE intensity to structure. However, different models use different approximations of the physics. For model validation, we measured the same topography using different techniques. With e-beam lithography, we patterned line/space arrays on a thin, flat silicon membrane. We acquired images of the same parts of one such array using secondary electrons in SEM and transmitted electrons in scanning transmission electron microscopy (STEM). We compared them with a subsequent image section by focused ion-beam at the same location. The determined discrepancy of line heights was 6.3 nm ± 6.1 nm. The 6.1 nm is an expanded uncertainty equal to three times the combined standard uncertainty, derived from several components that will be described. Our findings contribute to a wide range of measurements in SEM and STEM, providing insights that have the potential to be useful for combined metrology in the future.