News

A team of researchers from the Massachusetts Institute of Technology (MIT) have used U.S. National Science Foundation (NSF) ACCESS allocations on the Expanse system at the San Diego Supercomputer Center (SDSC) at the University of California San Diego School of Computing, Information and Data Sciences to uncover a hidden atomic process that challenges long-held assumptions about disorder in metals. The finding shows that the tiny building blocks of metals — atoms — can arrange themselves in surprising ways when the metal is melted, cooled or shaped.

“Using detailed computer simulations on SDSC’s Expanse, we tracked how these atomic patterns, known as short-range order or SRO, change during manufacturing,” said Rodrigo Freitas, a materials science and engineering professor at MIT. “We found that instead of settling into known, stable chemical arrangements, the atoms can fall into unusual configurations that only appear under fast or forceful conditions.”

Freitas said the discovery reveals that even familiar manufacturing processes — from slow casting to extreme deformation — can push metals into previously unrecognized atomic states. Under these conditions, atoms retain hidden order instead of becoming fully random, exposing physical mechanisms that challenge classical ideas of disorder.

The study, which was published in Nature Communications, shows that atomic motion and order persist even in everyday metal processing, expanding our understanding of how physical processes shape materials. The way atoms arrange themselves in a metallic alloy depends on how the metal is processed during manufacturing — whether it is slowly cast, rapidly solidified or heavily deformed. For decades, researchers assumed that these manufacturing routes, especially the most extreme ones, would erase any chemical order, leaving the alloy fully mixed.

“Our study shows instead that metals are never fully random at the atomic scale,” Freitas explained. “No matter the conditions, atoms always retain subtle local chemical patterns, known as short-range order or SRO.”

This finding introduces a new piece of physics in how metals respond when driven far from equilibrium. Atoms don’t simply lose order. Instead, motion and heat drive subtle, purposeful shifts among the atoms, allowing the metal to form entirely new, surprisingly stable structures that exist only under these extreme conditions.

It’s like trying to shuffle a deck of cards. One might expect that shuffling long enough would eventually produce a completely random deck. Freitas said that the team found that metallic alloys don’t work that way: their ‘deck’ always keeps a trace of order — for example, a slightly higher than random chance of finding a jack of spades near a four of hearts — no matter how long or hard you shuffle.

Growing evidence shows that SRO can affect a wide range of behaviors — from heat capacity and thermal expansion to electrical resistivity, magnetism, mechanical strength, corrosion resistance and even radiation tolerance. Despite this broad impact, SRO has often been overlooked in alloy design. Only recently have researchers begun to ask whether it could serve as a deliberate ‘design knob’ to tune performance. Those early studies, however, were mostly carried out under laboratory conditions, without carefully accounting for how alloys are actually made in practice. As a result, while the materials science and engineering community recognized SRO as a possible design mechanism, its evolution during real manufacturing remained unclear.

Freitas’ work began with a practical question in metal manufacturing: how fast do chemical elements mix during processing? The goal was to develop a simple relationship that could be used in the lab to predict and design alloys with controlled amounts of SRO, alongside other design parameters.

“But when we began measuring mixing rates, we found that the alloy never reached a fully random state. This was unexpected, because no known physical mechanism could explain such behavior,” Freitas said. “It pointed to a new piece of physics in metals. It was one of those cases where applied research led to a fundamental discovery: what began as an attempt to quantify mixing during manufacturing revealed new physics instead.”

To probe this unexpected behavior, the team turned to large-scale simulations on the Expanse supercomputer at UC San Diego. The team modeled millions of atoms to capture fleeting rearrangements that occur during deformation, events that experiments cannot resolve. They developed advanced computational tools, including machine-learned interatomic potentials and large-scale molecular dynamics simulations, to quantify subtle chemical order. These simulations revealed two distinct regimes of atomic organization: quasi-equilibrium states resembling conventional alloys, and far-from-equilibrium states never observed before.

Expanse-generated simulation of a metallic alloy under deformation. Colored spheres represent different atomic species, while green lines trace defects that shuffle atoms but still leave behind subtle chemical patterns. Credit: Rodrigo Freitas

“The results were striking as we identified two distinct kinds of chemical patterns,” Freitas explained. “One set mimics the familiar patterns of equilibrium alloys, but at an effectively higher temperature — what we call quasi-equilibrium states. Even more surprising were far-from-equilibrium states with no analog under any known conditions outside manufacturing.”

He said that this was the first time such patterns were observed.

Observing them was only half the challenge as the scientists also had to explain how they arise. That meant revisiting fundamental physics: how complex systems behave when driven far from equilibrium. By reducing the problem to its essentials, Expanse was used for a model that reproduced the key features of our simulations.

The mechanism turned out to be both unexpected and intuitive. Dislocations — well-known line defects that move as metals deform — emerged as the key players. As they move through the crystal, dislocations shuffle nearby atoms, much like reshuffling cards in a deck.

“Conventional wisdom held that this process simply erased order, but we found that dislocations also carry consistent ‘preferences’ — favoring some atomic swaps over others,” Freitas said. “The result is not perfect randomness but a steady remnant pattern, shaped by the competition between disorder and bias. This hidden bias in atomic shuffling gives rise to the new nonequilibrium states we uncovered.”

In a follow-up study, the team used this approach to study how SRO develops across a wide range of manufacturing conditions. By systematically varying temperature and strain rate, the scientists isolated the competing processes that create and erase these patterns and translated that behavior into a simple predictive framework. The result: a map linking standard processing parameters to the chemical patterns metals retain. With this map, engineers can begin to think of SRO as a practical design knob — tunable by controlling the same parameters already central to casting, metalworking and additive manufacturing.

This work was supported by the MathWorks Ignition Fund, MathWorks Engineering Fellowship Fund, and the Portuguese Foundation for International Cooperation in Science, Technology, and Higher Education in the MIT–Portugal Program. This material is based upon work supported by the Air Force Office of Scientific Research (award no. FA9550-25-1-0199), through the Young Investigator Program. The use of Expanse was supported via NSF ACCESS (allocation no. MAT210005).