Found in knee replacements and bone plates, aircraft components, and catalytic converters, the exceptionally strong metals known as multiple principal element alloys (MPEA) are about to get even stronger through artificial intelligence.
Sanket Deshmukh, associate professor in chemical engineering, and his team have designed a new MPEA with superior mechanical properties using a data-driven framework that leverages the supercomputing power of explainable artificial intelligence (AI).
Their findings are published in npj Computational Materials.
“This work demonstrates how data-driven frameworks and explainable AI can unlock new possibilities in materials design,” said Deshmukh.
“By integrating machine learning, evolutionary algorithms, and experimental validation, we are not only accelerating the discovery of advanced metallic alloys, but also creating tools that can be extended to complex material systems such as glycomaterials—polymeric materials containing carbohydrates.”
Elemental synergy, extraordinary properties
MPEAs are valuable because of their exceptional mechanical properties and versatility. Composed of three or more metallic elements, these alloys are designed to offer excellent thermal stability, strength, toughness, and resistance to corrosion and wear. Because they can withstand extreme conditions for longer periods than traditional alloys, they’re ideal for applications in aerospace, medical devices, and renewable energy technologies.
The team’s primary objective was to develop a new alloy with superior mechanical strength compared to the current model.
Traditionally, designing MPEAs has involved trial and error, which is slow and costly. But Deshmukh and his team are exploring the vast possibilities of designing MPEAs using explainable AI.
One major difference between standard AI and explainable AI is that traditional AI models often behave like “black boxes”—they generate predictions, but we don’t always understand how or why those predictions are made. Explainable AI addresses this limitation by providing insight into the model’s decision-making process.
In its work, the team used a technique called SHAP (SHapley Additive exPlanations) analysis to interpret the predictions made by its AI model. This enabled team members to understand how different elements and their local environments influence the properties of the MPEAs. As a result, they gained not only accurate predictions, but also valuable scientific insight.
AI can quickly predict the properties of new MPEAs based on their composition and optimize the combination of elements for specific applications. Using large data sets from experiments and simulations, AI can help explain the mechanical behaviors of MPEAs, guiding the design of new advanced alloys.
“Leveraging explainable AI accelerates our understanding of MPEAs’ mechanical behaviors. It could transform the traditional expensive trial-and-error materials design into a more predictive and insightful process,” said Fangxi “Toby” Wang, postdoctoral associate in chemical engineering and researcher on the project.
“Our design workflow, combining advanced machine learning and evolutionary algorithms, provides interpretable insights into materials’ structure-property relationships, offering a robust approach for the discovery of diverse advanced materials.”
Collaboration drives breakthroughs
Deshmukh teamed up with partners across disciplines and institutions on the research: Tyrel McQueen, professor of materials science and engineering at Johns Hopkins University, and Maren Roman, professor of sustainable biomaterials at Virginia Tech and director of GlycoMIP, a National Science Foundation Materials Innovation Platform.
“Working on a project this interdisciplinary is a treat,” said Allana Iwanicki, a graduate student in materials science and engineering at Johns Hopkins, who synthesized and tested the alloys. “This work bridges two fields: computational biomaterials and synthetic inorganic materials. It is exciting to achieve results meaningful to both groups.”
After initially focusing on these solvent-free systems, Deshmukh and his team have already extended this computational framework to design more complex materials, such as new glycomaterials, with potential applications in a wide range of products, including food additives, personal care items, health products, and packaging materials.
These advancements not only highlight the translational nature of this research, but also pave the way for future breakthroughs in material science and biotechnology.
“Our interdisciplinary collaboration across two National Science Foundation Materials Innovation Platforms not only allows us to develop transferable tools and platforms, but also highlights how partnerships at the intersection of computation, synthesis, and characterization can drive transformative breakthroughs in both fundamental science and real-world applications,” said Deshmukh.
More information:
Fangxi Wang et al, Experimentally validated inverse design of FeNiCrCoCu MPEAs and unlocking key insights with explainable AI, npj Computational Materials (2025). DOI: 10.1038/s41524-025-01600-x.
Virginia Tech
Citation:
Explainable AI framework reveals how element combinations boost alloy strength and durability (2025, May 15)
retrieved 15 May 2025
from https://techxplore.com/news/2025-05-ai-framework-reveals-element-combinations.html
This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no
part may be reproduced without the written permission. The content is provided for information purposes only.
