AI helps microscopes find the most informative nanoscale features in a sample

Researchers at Oak Ridge National Laboratory (ORNL) have developed an AI framework called SimuScan that assists atomic force microscopes (AFM) in identifying important nanoscale features within a sample. AFM, which can reveal structures as small as molecules, typically requires expert judgment from users regarding where to scan, how to adjust settings, and which features deserve closer study.

The new AI system, however, reduces this burden, making AFM faster, more consistent, and better suited for high-throughput research. The challenge in using AFM lies in the interpretation of its images, which reflect both the sample and the measurement process. SimuScan tackles this issue by generating synthetic AFM images, complete with automatic labels tied directly to the simulated object geometry, allowing AI models to be trained without the need for large volumes of hand-annotated experimental data.

The synthetic images recreate common imperfections encountered in real AFM scans, such as tip effects, scanner drift, electronic noise, contamination, and surface roughness, ensuring the AI models can accurately identify features in real-world samples.

Brief written by urgent.news from Phys.org's own syndicated text. Machine-written — may contain errors; check the original before relying on it.