24.04.2026
In February this year, a GAČR-funded project focused on high-speed scanning probe microscopy was launched. Over the next three years, researchers from CEITEC at Brno University of Technology (BUT), in collaboration with the Czech Metrology Institute (CMI), will work on accelerating measurement processes using this technique and improving data analysis through custom-developed artificial intelligence algorithms. The results are expected to be applied in both scientific and industrial practice, for example in monitoring living cells or semiconductor chip manufacturing.
High-speed atomic force microscopy allows scientists to study the surface properties of various materials. It uses microscopes with a tiny probe tip only a few nanometers in size (millionths of a millimeter), which repeatedly scans the sample surface along regular paths over areas measured in micrometers (smaller than a human red blood cell). The probe maps surface shapes with accuracy comparable to the distance between individual atoms, producing extremely high-resolution images. Although such measurements usually take only a few minutes, this can still be too slow in some cases, especially for samples that change over time, such as living biological materials or dynamic non-living systems. Capturing these changes continuously is a key motivation of the project, officially titled High-Speed Atomic Force Microscopy with Sparse Sampling.
“Our joint work aims to speed up scanning by optimizing the path of the microscope tip so that only truly useful data is measured. At the same time, we want to create an algorithm capable of evaluating images without needing to reconstruct missing surface data as an intermediate step,” says principal investigator Petr Klapetek from the Czech Metrology Institute. He explains that in conventional high-speed scanning, hundreds of images are collected across the surface, but for rapidly changing samples the probe may miss critical moments of change. Events occurring within milliseconds may not be captured continuously, causing important details to be lost. Worldwide, researchers are therefore seeking ways to speed up scanning, typically by modifying microscope hardware to move the probe faster. However, mechanical acceleration increases errors due to vibrations. “Since we understand the inherent limitations of scanning microscopy, we chose a different approach. We aim to bypass the problem through smarter measurement rather than major hardware changes,” Klapetek adds.
Software-based acceleration overcomes hardware limits
A key goal of the project is to design new scanning trajectories that remain regular but are arranged in optimized patterns to maximize information gained in minimal time. Combined with more efficient data processing, this could increase scanning speed up to fivefold.
“Data processing is absolutely crucial for us. We need to develop AI algorithms capable of analyzing and interpreting images from such sparse sampling grids directly, without needing information from the missing areas. This will be our biggest challenge,” adds David Nečas from CEITEC BUT, noting that CEITEC’s role includes scan pattern design and mathematical modelling, while all experimental measurements and testing will be conducted at the Czech Metrology Institute.
The result of this three-year collaboration should be a unified theoretical framework supported by experimental validation. It could be used in biological laboratories studying time-dependent changes in living cells, as well as in semiconductor manufacturing, where high-speed scanning microscopy is used for quality control and defect detection. The method can measure not only surface shape but also electrical and other properties.
“In the future, it could even become part of industrial production lines, where high-speed scanning microscopes would perform real-time imaging and evaluation directly in quality control systems,” Nečas says.
CEITEC BUT is also currently collaborating with the Czech Metrology Institute and other academic and industrial partners on the European DINAMO project, which focuses on developing data processing methods for calibrating all types of scanning probe microscopes. The CEITEC team, led by Nečas, is working on standardizing these methods so that even non-specialists can easily configure the instruments and measure samples with high precision.
