MIT Uses AI To Accelerate Discovery Of New Materials For 3D Printing

Researchers at MIT and BASF have developed a data-driven system that speeds up the process of discovering new 3D printing materials with multiple mechanical properties. Credit: Courtesy of the researchers

A new machine learning system costs less, generates less waste, and can be more innovative than manual discovery methods.

The growing popularity of 3D printing for manufacturing all kinds of items, from personalized medical devices to affordable homes, has created increased demand for new 3D printing materials designed for very specific uses.

To reduce the time required to discover these new materials, researchers at MIT have developed a data-driven process that uses machine learning to optimize new 3D printing materials with multiple characteristics, such as toughness and compressive strength.

By streamlining material development, the system reduces costs and environmental impact by reducing the amount of chemical waste. The machine learning algorithm could also spur innovation by suggesting unique chemical formulations that human intuition might miss.

“Material development is always a manual process. A chemist walks into a lab, mixes ingredients by hand, prepares samples, tests them, and comes up with a final formulation. But rather than having a chemist who can only do a few iterations over a period of days, our system can do hundreds of iterations over the same period, ”says Mike Foshey, Mechanical Engineer and Project Manager in the Computational Design and Manufacturing department. Group (CDFG) of the Computer Science and Artificial Intelligence Laboratory (CSAIL), and lead co-author of the article.

Other authors include co-lead author Timothy Erps, a technical associate at CDFG; Mina Konakovic Lukovic, CSAIL post-doctoral fellow; Wan Shou, a former post-doctoral fellow at MIT who is now an assistant professor at the University of Arkansas; senior author Wojciech Matusik, professor of electrical engineering and computer science at MIT; and Hanns Hagen Geotzke, Herve Dietsch and Klaus Stoll from BASF. The research was published on October 15, 2021 in Scientists progress.

Optimize discovery

In the system developed by the researchers, an optimization algorithm performs much of the discovery process through trial and error.

A materials developer selects a few ingredients, enters details of their chemical compositions into the algorithm, and defines the mechanical properties that the new material should have. Then the algorithm increases and decreases the amounts of these components (like turning the knobs on an amplifier) ​​and checks how each formula affects the properties of the material, before arriving at the ideal combination.

Then the developer mixes, processes and tests this sample to determine the actual performance of the material. The developer reports the results to the algorithm, which automatically learns from experience and uses the new information to decide on another formulation to test.

“We believe that for a number of applications this would surpass the conventional method, as you can rely more on the optimization algorithm to find the optimal solution. You wouldn’t need an expert chemist on site to preselect material formulations, ”explains Foshey.

Researchers have created a free and open source materials optimization platform called AutoOED that incorporates the same optimization algorithm. AutoOED is a comprehensive software package that also allows researchers to perform their own optimization.

Make materials

The researchers tested the system using it to optimize formulations of a new 3D printing ink that cures when exposed to ultraviolet light.

They identified six chemicals for use in formulations and set the algorithm’s goal of finding the best performing material in terms of toughness, compressive modulus (stiffness) and strength.

Maximizing these three properties manually would be particularly difficult as they can be conflicting; for example, the strongest material may not be the stiffest. Using a manual process, a chemist usually tries to maximize one property at a time, which results in a lot of experiments and a lot of waste.

The algorithm proposed 12 best performing materials that exhibited optimal tradeoffs between the three different properties after testing only 120 samples.

Foshey and coworkers were surprised at the wide variety of materials the algorithm was able to generate, and said the results were much more varied than expected based on the six ingredients. The system encourages exploration, which could be particularly useful in situations where specific material properties cannot be easily discovered intuitively.

Faster in the future

The process could be further accelerated through the use of additional automation. The researchers mixed and tested each sample by hand, but the robots could operate the dispensing and mixing systems in future versions of the system, Foshey explains.

Later, researchers would also like to test this data-driven discovery process for uses beyond the development of new 3D printing inks.

“It has wide applications in materials science in general. For example, if you wanted to design new, more efficient and less expensive types of batteries, you could use a system like this to do it. Or if you wanted to optimize the paint for a performance car that is environmentally friendly, this system could do that too, ”he says.

Because it presents a systematic approach to identifying optimal materials, this work could be a major step towards achieving high-performance structures, says Keith A. Brown, assistant professor in the Department of Mechanical Engineering at Boston University.

“The emphasis on new material formulations is particularly encouraging as it is a factor often overlooked by researchers who are limited by the commercially available materials. And the combination of data-driven methods and experimental science allows the team to identify materials efficiently. Since experimental efficiency is something all experimenters can relate to, the methods here have a chance to motivate the community to adopt more data-driven practices, ”he says.

Reference: “Accelerated Discovery of 3D Printing Materials Using Data-Driven Multiobjective Optimization” by Timothy Erps, Michael Foshey, Mina Konaković Luković, Wan Shou, Hanns Hagen Goetzke, Herve Dietsch, Klaus Stoll, Bernhard von Vacano and Wojciech Matusik, October 15, 2021, Scientists progress.
DOI: 10.1126 / sciadv.abf7435

The research was supported by BASF.

About Edward Fries

Check Also

Organist Peter Conte will perform in concert | Herald Community Newspapers

Renowned organist Peter Richard Conte will give a recital on May 13 at St Agnes …

Leave a Reply

Your email address will not be published.