Funded by the National Science Foundation’s Designing Materials to Revolutionize Our Engineering Future (DMREF) program, researchers from the Department of Materials Science and Engineering at Texas A&M University used an artificial intelligence materials selection framework ( AIMS) to discover a new shape memory alloy. The shape memory alloy showed the highest operating efficiency achieved so far for nickel-titanium materials. Additionally, their data-driven framework offers proof-of-concept for future materials development.
Shape memory alloys are used in various fields where compact, lightweight, solid-state actuators are required, replacing hydraulic or pneumatic actuators as they can deform when cold and return to their original shape when heated . This unique property is essential for applications such as aircraft wings, jet engines and automotive components, which must withstand repeated and recoverable shape changes.
There have been many advances in shape memory alloys since their beginnings in the mid-1960s, but at a cost. Understanding and discovering new shape memory alloys required extensive research through experimentation and ad hoc trial and error. Although many of these have been documented to aid other applications of shape memory alloys, new alloy discoveries have occurred over the decades. Approximately every 10 years, a significant shape memory alloy composition or system has been discovered. Moreover, even with advances in shape-memory alloys, they are hampered by their low energy efficiency caused by incompatibilities in their microstructure during large shape change. Plus, they’re notoriously difficult to design from scratch.
To address these shortcomings, Texas A&M researchers combined experimental data to create an AIMS computational framework capable of determining optimal material compositions and processing those materials, leading to the discovery of a new composition of shape memory alloy.
“When designing materials, sometimes you have multiple goals or constraints that conflict, which is very difficult to work around,” said Dr. Ibrahim Karaman, Chevron I Professor and Head of the Department of Materials Science and Engineering. materials. “Using our machine learning framework, we can use experimental data to find hidden correlations between characteristics of different materials to see if we can design new materials.”
The shape memory alloy found during the study using AIMS has been predicted and proven to achieve the tightest hysteresis ever recorded. In other words, the material exhibited the lowest energy loss when converting thermal energy into mechanical work. The material exhibited high efficiency when thermally cycled due to its extremely small transformation temperature window. The material also exhibited excellent cyclic stability under repeated actuation.
A nickel-titanium-copper composition is typical of shape memory alloys. Nickel-titanium-copper alloys typically have 50% titanium and form a single-phase material. Using machine learning, the researchers predicted a different composition with titanium equaling 47% and copper equaling 21%. Although this composition is in the two-phase region and forms particles, they help improve the properties of the material, explained William Trehern, a doctoral student and graduate research assistant in the Department of Materials Science and Engineering and first author of the publication.
In particular, this high-efficiency shape-memory alloy lends itself to thermal energy harvesting, which requires materials capable of capturing and utilizing waste energy produced by machinery, and to thermal energy storage. , which is used to cool electronic devices.
Most notably, the AIMS framework provides the ability to use machine learning techniques in materials science. Researchers see the potential to discover more shape memory alloy chemistries with desired characteristics for various other applications.
“It’s a revelation to use machine learning to find connections that our brains or known physical principles may not be able to explain,” Karaman said. “We can use data science and machine learning to accelerate the pace of material discovery. I also think we can potentially uncover new physics or new mechanisms behind the behavior of materials that we didn’t know about before if we let’s pay attention to the connections that machine learning can find.”
Other contributors include Dr. Raymundo Arróyave and Dr. Kadri Can Atli, professors in the Department of Materials Science and Engineering, and materials science and engineering undergraduate student Risheil Ortiz-Ayala.
“While machine learning is now widely used in materials science, most approaches to date focus on predicting the properties of a material without necessarily explaining how to process it to achieve the target properties,” said Arroyave. “Here, the framework considered not only the chemical composition of the candidate materials, but also the processing needed to achieve the properties of interest.”