Interlaboratory Round–Robin Study of Errors and Best Practices for Scanning Electron Microscope Digital Image Correlation
摘要
Scanning electron microscope digital image correlation (SEM-DIC), also referred to as high-resolution digital image correlation (HR-DIC), is a powerful method to measure displacements and strains at the micro to nanoscale. While SEM-DIC is gaining popularity, the nature of various errors in SEM-DIC is not clear.
ObjectiveTo advance the practice, this work seeks a more extensive and cohesive understanding of the types and causes of errors in SEM-DIC. Previous works have identified SEM imaging errors that are problematic for DIC, including spatial distortions, drift distortions, noise, stabilization errors, and line jumps.
MethodsThe work presented here outlines a unique, interlaboratory, round-robin format to investigate how beam and hardware settings can optimize SEM-DIC image quality. Characterization using multiple SEMs of different vendors, models, and types (e.g., electrostatic and electromagnetic columns) helps identify the types, magnitudes, and causes of SEM-DIC errors, as well as the optimum conditions to reduce such errors.
ResultsAs part of the iDICs round-robin challenge, over 2,000 total SEM images, which are available as an open-access dataset, were analyzed in this study, comprising images from five SEMs, two sample patterning types, and two magnifications. To quantify SEM-DIC errors, this work developed new error metrics to quantify the severity of line jumps, left-edge smear, horizontal skewing, vertical skewing, and translational drift.
ConclusionsThis round-robin study found that most SEM-DIC experiments should utilize high beam voltages (e.g., 30 kV), fast dwell times (e.g., 1 µs), low to moderate beam currents (e.g., less than 1 nA), no line averaging/integration, and cropping of the left-most portion of images.