Inside the microchips powering your devices, atoms aren’t just randomly scattered. They follow a hidden order that can change how semiconductors behave.
A team of researchers from the Lawrence Berkeley National Laboratory (Berkeley Lab) and George Washington University has, for the first time, observed these tiny patterns, called short-range order (SRO), directly in semiconductors.
This discovery is a game-changer, as understanding how atoms naturally arrange themselves could let researchers design materials with desirable electronic properties. Such control could revolutionize quantum computing, neuromorphic devices that mimic the brain, and advanced optical detectors.
“This is the first time the individual structure of these SRO domains has been shown experimentally,” Andrew Minor, one of the researchers and a professor at UC Berkeley, said.
Decoding the secret atomic arrangement
Until now, the arrangement of rare trace atoms mixed into semiconductors remained a mystery. These small amounts of tin, silicon, or other elements are not enough to form large repeating patterns, so scientists couldn’t tell if they were random or ordered.
Plus, traditional microscopy simply couldn’t zoom in close enough with clarity. The hidden patterns mattered “because the property that’s being changed by this local ordering is the most important property for microelectronics, the band gap, which is what controls the electronic properties,” Minor said.
The researchers attempted to solve this problem by combining advanced microscopy with machine learning. First, they studied a germanium sample containing small amounts of tin and silicon using a powerful electron microscopy technique called 4D-STEM.
Initial images were messy because the faint signals from tin and silicon were overwhelmed by strong germanium signals. To fix this, the researchers added an energy-filtering device that improved contrast, making subtle repeating atomic patterns visible for the first time.
Then, to identify these patterns, they used a pre-trained neural network, which detected six recurring motifs—distinct atomic arrangements, but the exact structures were unclear. That’s where researchers from George Washington University stepped in.
They built a machine-learning model capable of simulating millions of atoms. By performing simulated 4D-STEM, the team tested different arrangements until the motifs in the simulation matched the experimental data. This seamless combination of high-resolution imaging, energy filtering, and AI modeling finally revealed the hidden atomic order in semiconductors.
“It’s remarkable that modeling and experiment can work seamlessly to unravel SRO structural motifs for the first time,” Tianshu Li, co-lead researcher and a professor at George Washington University, said.
A discovery that promises big changes
This finding could transform how semiconductors are designed. By controlling short-range order at the atomic level, researchers could tailor the band gap and other key electronic properties, enabling faster quantum computers, brain-inspired devices, and advanced optical sensors.
It also represents a major step forward in understanding materials that were previously too small or complex to study directly. “We are opening the door to a new era of information technology at the atomic scale,” Lilian Vogl, first author of the study and a postdoc researcher at UC Berkeley, said.
However, there are some limitations as well. For instance, signals from SRO can be masked by defects or atomic motion at room temperature, and researchers are still mapping how these motifs influence material behavior.
Further research will focus on exploring such effects, aiming to manipulate atomic arrangements for new device designs.
The study is published in the journal Science.