Snake Scanning for SEM: Quantification and Correction of Its Inherent Misalignment Distortion Using an External Scan Controller

Distortions in scanning electron microscope (SEM) images compromise characterization accuracy and restrict reliable quantitative analysis. Quantifying and correcting these distortions remains challenging due to the complexity of their inherent sources, such as scanning coil hysteresis and electronic circuit response delays. To address this, we independently developed a scanning controller and software system that enables customizable scanning strategies and is crucial for capturing unprocessed raw data. We utilized the characteristic row misalignment of snake scanning to split images into sub-images, measure offsets using the ORB algorithm, and apply pixel compensation. Experimental validation shows that corrected images exhibit reduced distortion artifacts, with structural similarity comparable to raster scanning results and improved reference-free quality metrics. The distortion magnitude is independent of magnification, primarily governed by dwell time, and stabilizes at a minimum level when the dwell time reaches a critical threshold. This work clarifies the relationship between scanning parameters and distortion behavior, guiding the optimization of SEM scanning strategies. Furthermore, it offers a potential scalable framework for distortion correction in related microscopy techniques. Many of these techniques also face distortion issues from hardware hysteresis or circuit delays, similar to SEM.