Chad Mirkin's team at Northwestern University has figured out how to design new materials in hours instead of years—and they're doing it by abandoning the painstaking trial-and-error methods that have defined materials science for generations.
The breakthrough matters because the world faces urgent material challenges: better batteries for renewable energy, more efficient semiconductors, lighter stronger alloys. Traditionally, discovering a single promising new material has meant years of laboratory work, thousands of failed experiments, and a large measure of luck. But Mirkin's megalibrary platform—a tool that synthesizes millions of tiny material candidates simultaneously on a single chip—collapses that timeline dramatically. In a study published in Science Advances, the Northwestern team demonstrated they can now screen more than a million different material samples in less than 30 minutes using a technique called second harmonic generation microscopy.
This time isn't just faster—it's transformative. The researchers challenged their megalibrary to search through thousands of chemical combinations to find piezoelectric materials, substances that generate electricity when pressed or bent. These materials are essential for ultrasound imaging, sensors, motion detectors, and energy-harvesting devices. The platform not only discovered a promising candidate but went further: the team deliberately engineered a piezoelectric material designed to operate at a specific temperature, completing the entire design process within hours. This represents a fundamental shift from passive discovery to active engineering.
The megalibrary's speed advantage over emerging "self-driving labs"—automated robotic systems that propose, develop, and test materials iteratively—is stark. Self-driving labs work step-by-step, refining experiments one after another. The megalibrary operates in massive parallel, evaluating enormous numbers of candidates at once. As graduate student Jarod Beights, a co-first author of the study, put it: "Compared to the megalibrary, which moves at a sprint, self-driving labs are basically crawling."
What makes this especially significant is not just the speed but the data. The megalibrary generates high-quality datasets at an unprecedented scale—exactly what artificial intelligence systems need to learn and improve. Where older approaches produced scattered data points, the megalibrary produces millions of carefully characterized samples. This abundance of training data could unlock the next generation of AI-assisted materials discovery, creating a virtuous cycle: faster discovery generates better data, which trains better AI, which enables faster discovery.
Mirkin, the George B. Rathmann Professor of Chemistry at Northwestern and founding executive director of the International Institute for Nanotechnology, first introduced the megalibrary platform in 2016. It was already revolutionary—compressing years of searching into a single day. But this new research shows the platform can do something more: deliberately design materials with predetermined properties, something that would have been "extraordinarily difficult to find through conventional experimentation," as the team notes.
The implications extend far beyond the laboratory. As materials discovery accelerates, so does the pace of technological innovation. Better materials mean better devices, more efficient energy systems, and solutions to problems we haven't yet solved. "We're about to witness the meteoric rise of materials discovery," Mirkin said, "and this is just the start."