Artificial Intelligence Solves 100-Year-Old Electronics and Energy Storage Mystery

A breakthrough by researchers at Columbia Engineering has resolved a long-standing scientific puzzle that has challenged scientists for over a century. By harnessing artificial intelligence, the team developed a tool capable of revealing the hidden atomic structures of nanocrystals, materials so small and disordered that conventional methods have failed to analyze them. This achievement not only opens new doors in materials science but also demonstrates how far AI has come in aiding scientific discovery.

The century-old challenge

For more than 100 years, scientists have struggled to determine the atomic structures of nanocrystals—tiny, disordered materials with wide applications in fields such as electronics, energy storage, and archaeology. Traditional X-ray diffraction techniques, passing a beam through large, pure crystals to produce readable patterns, fall short with nanocrystals. These smaller structures scatter X-rays into chaotic, unreadable patterns, making their internal structure difficult to interpret.

A custom AI solution

A significant advancement was made when a team from Columbia Engineering in New York solved this long-standing problem using artificial intelligence. By developing a PXRDnet tool, the researchers could decode the atomic structure of crystals as small as 10 angstroms—thousands of times thinner than a human hair.

“The AI solved this problem by learning everything it could from a database of many thousands of known, but unrelated, structures,” said Simon Billinge, professor of materials science and applied physics and applied mathematics at Columbia Engineering. “Just as ChatGPT learns the patterns of language, the AI model learned the patterns of atomic arrangements that nature allows.”

Scientific impact and future possibilities

This AI-driven approach marks a significant leap forward in materials science, offering new ways to identify and characterize previously inaccessible nanomaterials. It reminds us how rapidly artificial intelligence is evolving to support, rather than replace, human researchers.

“When I was in middle school, the field was struggling to build algorithms that could tell cats from dogs,” said Gabe Guo, who led the project at Columbia. “Now, studies like ours underscore the massive power of AI to augment the power of human scientists and accelerate innovation to new levels.”

The complete study was titled “Ab initio structure solutions from nanocrystalline powder diffraction data via diffusion models.” This peer-reviewed publication signals the credibility and broader importance of the research.

A sign of things to come

“What excites me is that with relatively little background knowledge in physics or geometry, AI was able to learn to solve a puzzle that has baffled human researchers for a century,” said Hod Lipson, chair of the Department of Mechanical Engineering at Columbia Engineering. “This is a sign of things to come for many other fields facing long-standing challenges.”

Acknowledgement

This article is based on original reporting by Anthony Cuthbertson, published in The Independent. The research was conducted by Columbia Engineering in New York, with key contributions from Simon Billinge, Gabe Guo, and Hod Lipson. Their findings were published in Nature Materials in a study titled “Ab initio structure solutions from nanocrystalline powder diffraction data via diffusion models.”

The team developed an AI tool called PXRDnet, which was trained on tens of thousands of known materials to determine the atomic structures of nanocrystals.

The original article can be accessed here: AI solves 100-year-old mystery to supercharge scientific discovery – The Independent.