Here’s what you’ll learn when reading this story:

• Although 3D-printed steel comes with a lot of advantages—including increased flexibility and low lead times—few metals are specifically designed for the process of printing.
• Scientists from Purdue University and the University of South China have developed a new alloy by using a machine algorithm to track 81 physicochemical features of elements, as well as how those elements interact with the 3D printing process.
• The resulting metal was rust-proof, had a strength increase of 30 percent, and its ductility doubled—a major leap forward for additive metal manufacturing.

3D-printing metal comes with some obvious benefits. Unbounded by typical tooling and traditional manufacturing, structures built with these methods can be highly complex, reduce material waste, and significantly cut down on lead times. But a lot of the metals used for these purposes today weren’t developed as 3D-printing material—at least, not at first.

According to 3Dprint.com, many of today’s additive metal materials were originally developed for forging or casting, and later adopted for 3D-printing purposes. But as with anything jerry-rigged to fit some other purpose, this can lead to strength issues, defects, or other inefficiencies—especially during the heating and cooling process required for laser powder bed fusion (LPBF) printers.

Now, researchers behind a new international study have created an “interpretable machine learning” model that can churn through 81 fundamental physicochemical features of elements—drilling down to their very atomic radii and electron behaviors—to develop an ultra-strong 3D-printable rust-proof alloy. The algorithm also accounted for how the material would react to the 3D printing process itself, meaning that the material it helped create was developed with the application specifically in mind. The results of the study were published in the International Journal of Extreme Manufacturing.

“This strategy has dramatically accelerated the discovery process and enabled the introduction of a low-cost, short-process strategy for additively manufacturing UHSDS [ultra-high strength and ductility steels] with exceptional corrosion resistance, thereby overcoming critical limitations in current additively manufactured steels,” the authors wrote.

The metal produced by the algorithm is quite a mouthful (Fe-15Cr-3.2Ni-0.8Mn-0.6Cu-0.56Si-0.4Al-0.16C), but the results were immediately promising. According to the AI model, the material should be able to withstand roughly 1,713 Megapascals (Mpa) and stretch more than 15 percent before breaking. When the researchers tested this new alloy using LPBF printers, they found that these predictions matched perfectly with physical experimentation.

According to a press release, this performance represents about a 30 percent increase in strength (compared to a metal’s raw printed state) and a doubling of its ductility. They found that the short, six-hour heat treatment of the metal created nanoscale particles of copper and nickel-aluminum that effectively blocked structural defects from spreading. Additionally, the fact that the material also resists corrosion means it has a much wider potential for application—especially in the aerospace and marine sectors, where materials often interact directly with moisture. According to the study, this new alloy degrades only 0.105 millimeters per year, which is better than some leading commercial stainless steels.

The authors argue that the physicochemical feature-machine learning (PF-ML) design strategy is a cost-effective way to advance additive metal manufacturing, though the features will need to be retooled with each new material class. But it could represent the breakthrough that the industry needs to finally create strong, rust-resistant metals with the speed and flexibility that made 3D printers famous in the first place.