ARI and Beckman Researchers Receive NSF Grant for Automated Polymer Research System

Researchers at the Applied Research Institute are part of a team at the University of Illinois Urbana-Champaign that has received a $3.6 million Major Research Instrumentation grant from the National Science Foundation which was announced on September 13.

The grant will facilitate the acquisition of a custom-built, automated, high-throughput (HT) system for polymer formulation and characterization to accelerate the design and development of new polymer systems. This instrument will enable research to be performed around-the-clock with very high accuracy. Having one machine perform all the experiments, instead of multiple students and scientists, greatly increases the repeatability of the data. The system will also include a fully-integrated software and database system that will enable all the collected data and experimental conditions to be automatically recorded in databases. This improved speed, accuracy, and digital organization will enable researchers to leverage recent advances in machine learning to further catapult their output. This acceleration means that new materials can be developed and put into commercial production much quicker.

Dr. Daniel Krogstad, a senior research scientist at ARI in the Materials and Manufacturing group and co-principal investigator (PI) on the project, said this machine fills an important gap in materials research and development. Current research and development approaches in polymer science can result in development cycles that last decades before the materials reach the market. This is a critical problem when key societal challenges, such as plastic waste management, global climate change, and food scarcity, require solutions now. Materials research is slow because experiments take time and resources, and the materials can have dozens of components and hundreds of potential variables. Many organizations haven’t been able to invest in HT-systems like this before. But this new machine will help decrease the timeline of new materials development.

“It’s a gamechanger in every way. It fundamentally changes how you do research and the speed at which you do research,” he said. “It will allow us to efficiently explore large variable spaces to understand how the polymer formulation and processing conditions will influence the key properties of the polymer systems.”

The machine, which will take almost two years to be custom built, will feature automated thermal (how materials behave at different temperatures) and rheological (how materials behave under different stress and flow conditions) characterization, which are broadly applicable to nearly every polymer system.

Using machine learning algorithms to analyze the data that is generated will help identify correlations and insights in the data that the researchers may not otherwise observe and get closer to optimized combinations of variables more quickly and efficiently. Polymer scientists have been slow to adopt machine learning approaches in their research because there is often not enough data. This high-throughput instrument will solve that problem.

Eliezer Colina Morles, a senior research engineer at ARI in the Algorithms and Software group and one of the senior personnel on the project, said, “We suggested creating and using machine learning algorithms as a powerful tool for modeling complex processes involving many numerical and categorical variables on multiple time scales.”

Dr. Michal Ondrejcek, a senior software engineer at ARI in the Algorithms and Software group and another of the senior personnel on the project, explained that the active learning frameworks the new machine will use will be suitable to the type of experimental datasets that will be generated.

The machine learning component, he said, “will enable insights into the key experimental features that can be used to tune the rheological properties.”

This new machine will let researchers increase the complexity of problems they can work on and accelerate development cycles of materials, which will be important considerations for future partnerships involving this new system.

The machine will be housed at the Beckman Institute at Illinois and will open up new possibilities for ongoing research and collaboration with future partners.

The established collaboration of the team at ARI, comprised of researchers in the Materials and Manufacturing and Algorithms and Software groups, will accelerate the impact they can have at the early stages of using this new instrument, Krogstad said.

Ondrejcek, echoing this sentiment, looks forward to the continued collaboration this new machine will facilitate for the research groups at ARI. “It will enhance polymer research at ARI and solidify collaboration between the Materials and Manufacturing and Algorithms and Software groups,” he said.

“The facility will catalyze a greater transition from basic research to practical applications, forging robust partnerships between ARI and industry, and addressing real-world manufacturing challenges,” said Ngoc Nguyen, a staff research scientist and one of the senior personnel on the project.

The full list of PIs and senior personnel on this project include: PI: Charles Schroeder; Co-PIs: Ying Diao, Daniel Krogstad, Nancy Sottos, Sameh Tawfick; Senior Personnel: Jeff Baur, Paul Braun, Martin Burke, Eliezer Colina Morles, Randy Ewoldt, Christopher Evans, Damien Guironnet, Nick Jackson, Katie Matlack, Jeffrey Moore, Ngoc Nguyen, Michal Ondrejcek, Simon Rogers, Charles Sing, and Antonia Statt.