The secret to how steel hardens and shape-memory alloys snap into place lies in rapid, atomic-scale shifts that scientists have struggled to observe in materials. Now, Cornell researchers are revealing how these transformations unfold, particle by particle, through advanced modeling techniques.
Using custom-built computer simulations, Julia Dshemuchadse, assistant professor of materials science and engineering in Cornell Engineering, and Hillary Pan, Ph.D. '25, have visualized solid-solid phase transitions in unprecedented detail, capturing the motion of every particle in a theoretical material as its crystal structure morphs into another.
Their findings, published July 23 in the Proceedings of the National Academy of Sciences, reveal not only classical transformation mechanisms but also entirely new ones, reshaping how scientists understand this fundamental process in materials science.
"Most prior research either reports on the before and after stage of the transformations or it discusses them from a theoretical perspective," Dshemuchadse said. "Our computational study is the first to fill the gap between these two more traditional approaches. We simulate the transformations directly, and we can track particle by particle how one structure forms from the other."
The researchers focused on transformations between two of the most common crystal structures: face-centered cubic and body-centered cubic sphere packings. These structures are found across a wide range of materials - from soft-matter systems like plastics to hard metals like iron and steel, where such transformations play a key role in industrial processes like metallic hardening.
"There's no camera fast enough to capture the resolution you need in order to know what exactly is happening in between," Pan said, "and X-ray diffraction techniques provide limited information about how the transformation is actually proceeding."
Starting with small simulations of about 4,000 particles and then scaling up to more than 100,000 particles, the researchers designed the models to explore general transformation behavior using abstract, tunable particles. This allowed them to characterize multiple transformation pathways, including three well-known mechanisms that have been proposed for atomic systems: the Bain, Kurdjumov-Sachs and Nishiyama-Wassermann orientation relationships.
The simulations found systems in which the material's microstructure and temperature dictate the transformation pathway taken, and they revealed a stable intermediate phase on the path from body-centered cubic to face-centered cubic.
But one of the study's most surprising discoveries was a completely new way the transformation could happen: Particles in the material shifted together in a coordinated, multiunit shearing motion that had not been predicted or seen before.
"Importantly, the study shows that the pathways taken are not clearly determined by comparing before and after configurations of the material," Dshemuchadse said, "which suggests that researchers have rightfully struggled with classifying these transformations when unable to observe them in action."
Instead, the pathways are linked to the shape of the underlying particle interactions. The insights could help experimentalists interpret data from material systems by providing simulated templates for transformations that remain invisible in real time.
"It's possible lab experiments could be designed to tune particle interactions in order to replicate the different transition pathways we're seeing," said Pan, adding that previous studies have suggested hydrodynamics can play a role in pathway selection for soft materials.
The research was based upon work supported by the National Science Foundation and the Camille and Henry Dreyfus Foundation.
Syl Kacapyr is associate director of marketing and communications for Cornell Engineering.
