Simulation-based analysis of influence of error on super-resolution optical inspection

Microfabricated structures such as semiconductors and MEMS continue shrinking as nanotechnology expands, demand that measures microfabricated structures has risen. Optics and electron beam have been mainly used for that purpose, but the resolving power of optics is limited by the Rayleigh limit and it is generally low for subwavelength-geometry defects, while scanning electron microscopy requires a vacuum and induces contamination in measurement. To handle these considerations, we propose optical microfabrication inspection using a standing-wave shift. This is based on a super-resolution algorithm in which the inspection resolution exceeds the Rayleigh limit by shifting standing waves with a piezoelectric actuator. While resolution beyond the Rayleigh limit by proposed method has been studied theoretically and realized experimentally, we must understand the influence of experimental error factors and reflect this influence in the calibration when actual application is constructed. The standing-wave pitch, initial phase, and noise were studied as experimental error factors. As a result, it was confirmed that super-resolution beyond the Rayleigh limit is achievable if (i) standingwave pitch error was 5% when standing-wave pitch was 300 nm or less and (ii) if initial phase error was 30° when standing-wave pitch was 300 nm. Noise accumulation was confirmed in studies of the noise effect, and a low-pass filter proved effective against noise influence.