AI now makes it possible to read the infinitesimal and explore the unsuspecting world

Artificial intelligence is revolutionizing microscopy: a giant step towards nanometric precision.

Atomic force microscopy (AFM) has long represented an essential tool for researchers in the analysis of material surfaces. However, the size of the microscope probe always limits the accuracy of the images produced. Today, the University of Illinois at Urbana-Champaign announced a significant breakthrough: a new artificial intelligence (AI) technology promises to overcome this fundamental limitation, providing unprecedented insight at the nanoscale.

Also Read:

Historical limit exceeded

For decades, AFM has allowed scientists to accurately map the surfaces of materials in three dimensions. But when surface features approach the size of the probe tip, around 10 nanometers, their resolution becomes problematic. A new AI technique developed by researchers at the University of Illinois manages to “decode” the microscope images, removing the unwanted effects of the probe width and thus allowing a resolution lower than this.

The Deep Learning Revolution

A research team led by Professor Yingjie Zhang and doctoral student Lalith Bonagiri developed an innovative deep learning algorithm. This algorithm uses an encoder-decoder framework to transform raw AFM images by identifying and removing distortions introduced by the probe. This process creates a three-dimensional surface profile with unparalleled precision.

Algorithm training and validation

To train this algorithm, the researchers created synthetic images of three-dimensional structures and simulated their AFM readings. The major challenge was to maintain the perfect brightness and contrast of the images, essential elements to preserve the significance of the data. Tests performed on gold and palladium nanoparticles proved the effectiveness of the algorithm in suppressing probe tip effects and accurately revealing the three-dimensional characteristics of the nanoparticles.

Implications and future perspectives

This advancement not only represents a significant improvement to AFM imaging, but also opens the door to new possibilities in the development of nanotechnology and the study of biological systems and materials. The researchers point out that, while promising, the technique can still be improved by training on more and better data.

Impact on research and industry

Improving the resolution of AFM images using AI could revolutionize many areas of research and industrial applications. This could lead to significant discoveries in advanced materials, biology and nanotechnology, making it possible to explore previously inaccessible regions.

This article explores a recent breakthrough in microscopy, where an artificial intelligence technique overcomes a fundamental limitation of AFM, providing unprecedented precision at the nanoscale. This progress promises to open new avenues in scientific research and technological development, enabling more detailed and precise analysis of material surfaces.

Source: https://pubs.acs.org/doi/10.1021/acs.nanolett.3c04712