Designed to accelerate innovation in electronics, this technology could fast-track the development of advanced solar panels, sensors, and other high-performance electronic devices.
A team of MIT researchers has unveiled a fully autonomous robotic system that could supercharge the discovery of new semiconductor materials—particularly for applications like solar panels and advanced electronics. This innovation dramatically reduces the time it takes to measure a key property of semiconductors: photoconductance, or how materials respond electrically to light.
From Manual to Machine: Speeding Science
Traditionally, researchers manually test materials—a slow, tedious process. The solution uses a robotic probe guided by a machine-learning model trained with insights from material science experts. The system identifies the most data-rich contact points on a material sample, rapidly maneuvers between them, and collects measurements at a rate of over 125 per hour—with far greater precision than previous AI approaches.
Photoconductance can’t be measured without physical contact. So, the team developed a computer vision system that maps printed samples—like perovskites—and feeds them into a neural network that selects optimal probe locations. A sophisticated path planner, enhanced with controlled randomness, ensures the robot moves efficiently across uniquely shaped samples, which resemble “snowflakes” in their diversity.
Self-Supervised AI, Precision Measurements
Unlike many AI systems, this one doesn’t need labeled training data. It’s self-supervised, directly analyzing sample images to select probe points. The robotic system then executes a streamlined measurement path with the help of motors and custom software, achieving thousands of reliable measurements in just one day.
In rigorous testing, the neural network outperformed seven other AI models in speed and contact accuracy. Their path-planning algorithm also proved superior in optimizing movement. The goal? A fully autonomous lab capable of discovering novel, high-efficiency semiconductor materials to power the future of electronics and solar energy. This research is backed by major players, including First Solar, MathWorks, Eni, the U.S. Department of Energy, and the National Science Foundation—highlighting the broad interest in accelerating materials innovation for sustainable tech.
