Two-dimensional materials, which consist of a single layer of atoms, exhibit unusual properties that could be harnessed for a wide range of quantum and microelectronics systems. But what makes them truly special are their flaws. “That’s where their true magic lies,” said Alexander Weber-Bargioni at the Department of Energy’s Lawrence Berkeley National Laboratory (Berkeley Lab). Defects down to the atomic level can influence the material’s macroscopic function and lead to novel quantum behaviors, and there are so many kinds of defects that researchers have barely begun to understand the possibilities. One of the biggest challenges in the field is systematically studying these defects at relevant scales, or with atomic resolution.
Artificial intelligence suggests a way forward. Researchers at Berkeley Lab recently unveiled a new, fast, and readily reproducible way to map and identify defects in two-dimensional materials. It uses convolutional neural networks, which are an application of artificial intelligence, to quickly analyze data from autonomous experiments, which in recent years have become a powerful tool for imaging these exotic materials.
“Defects can be used advantageously, or they can cause issues with the macroscopic function of the material,“ said John Thomas, a postdoctoral research fellow in the Weber-Bargioni Group at the Molecular Foundry, a DOE Office of Science user facility at Berkeley Lab where this research was conducted. Thomas devised the approach that couples AI with autonomous discovery. “This combination gives us a nice way to screen for defects and measure them,” he said. The method could dramatically reduce the time required to characterize two-dimensional materials and use them in next-generation quantum and electronic devices. The scientists reported their research in a paper published in npj Computational Materials.