NEWS
How to Predict Diffusion in Complex Alloys
Slowing diffusion through chemistry can widen the workability window of aluminum alloys, delivering ease of use in production lines.
Written by Poornima Apte
IT’S AN ENGINEERING PROBLEM that has frustrated the automotive industry for years: How to manufacture vehicle parts using high-strength lightweight alloys if the material becomes hard and loses its formability shortly after heat treatment and quenching.
The answer is rooted in the science of diffusion—specifically in understanding how both chemistry and time affect the rate of diffusion of atoms and vacancies in complex alloys.
Researchers from the University of Michigan, led by Liang Qi, associate professor in the Department of Materials Science, have developed a computational framework that explains how both chemistry and time affect the diffusion process. The team worked in collaboration with General Motors Research & Development. The benefits of this research extend beyond automotive manufacturing to other areas, including forecasting the performance of nuclear plant building materials over time.
Alloys in Automotive Manufacturing
Lightweight aluminum alloys are favorites in vehicle production because they improve aerodynamic performance and give better fuel mileage.
Among these alloys is the 7000 series (Al, zinc, and magnesium), which has been successful in aerospace applications, and is being evaluated for automotives.
Production of the alloy involves heating the materials to a high temperature (500 ℃) so the zinc and magnesium dissolve in the aluminum matrix. Unfortunately, after the heat treatment, the final alloy hardens within an hour, and is not pliable for the long duration needed on the manufacturing floor. This happens because magnesium, and especially zinc, diffuse rapidly during the quenching process and form clusters in the aluminum matrix, kickstarting the hardening process.
The aerospace industry works around this problem by either working the material at high temperatures to preserve ductility or by slowing down the hardening process using dry ice to tamp down diffusion.
Both these methods are expensive, and prohibitively so, for the automotive industry. For any solution to be of practical use in the manufacturing of automobiles, the material must be able to be stored and worked with for days—at a low cost.
Photo: Getty
Slowing Down Diffusion
Diffusion in alloys needs vacancies, voids in the lattice structure that enable atoms to hop around.
While the number of vacancies depends on the temperature—higher temperatures lead to more vacancies and vice versa—for the automotive industry solution, the focus was on chemistry. For example, dopants might strongly pin a vacancy down and prevent it from moving, in effect slowing diffusion.
The framework that Qi and team have developed helps evaluate the effects of time and temperature as well as material chemistry on diffusion. It’s a significant step toward answering which dopant at what temperature and time frame will deliver needed diffusion rates.
“The final goal is to find some doping element that can effectively pin a mobile vacancy or help generate some cluster to effectively pin the vacancy,” Qi said. “Then we can make the diffusion much slower during quenching and during room-temperature storage.”
The advantage to manufacturing will be significant: “You won’t need to develop any special manufacturing process, you can simply use whatever production methods you have right now,” Qi said.
Computational Framework Solution
Qi and team developed a computational framework to address two goals: explain the rapid hardening phenomenon after quenching, and to evaluate if chemistry, potentially through the addition of dopants, can slow down the diffusion process. The results of the research were published in npj Computational Materials.
Clues for a framework that might explain the diffusion phenomenon are buried in quantum mechanics, Qi said, but the field of study has stalled for years because the scale of quantum mechanics is limited to single atoms and nanoseconds. By contrast, “the whole diffusion process involves billions of atoms and hours. How can we start from quantum mechanics calculations and predict behavior on a much larger scale,” Qi said.
To get there, the researchers used the kinetic Monte Carlo simulation to carry out true-to-life calculations on a small scale and plug the results into a Markov chain cluster dynamics model, which extrapolates to the wider scale needed. “If the gap between what we have and what we need is too large and it’s difficult to build one bridge, then we divide the gap into two and build two bridges,” Qi explained.
Photo: Getty
Other Applications
The computational framework developed can apply to other alloys and the study of diffusion in other settings. Qi pointed to the struggle nuclear power plants face as more of them are being developed as carbon-free energy sources.
“One problem they’re facing is predicting the long-term performance of materials and how they will be affected by temperatures and radiation,” Qi said. This too finds answers in diffusion and in the framework he and his team have developed.
Poornima Apte is a technology writer based in Walpole, Mass.
© 2026 The American Society of Mechanical Engineers. All rights reserved.