AI-driven, autonomous lab at Argonne transforms materials discovery
Artistic rendering depicting Polybot, the AI-driven automated material laboratory.
Image by Argonne National Laboratory
Argonne, UChicago researchers harness artificial intelligence and automation for electronic polymer processing
Plastic that conducts electricity might sound impossible. But there is a special class of materials known as “electronic polymers” that combines the flexibility of plastic with the functionality of metal. This type of material opens the door for breakthroughs in wearable devices, printable electronics and advanced energy storage systems.
Yet making thin films from electronic polymers has always been a difficult task. It takes a lot of fine-tuning to achieve the right balance of physical and electronic properties. Researchers at Argonne National Laboratory and the University of Chicago have created an innovative solution to this challenge with artificial intelligence.
They used a tool called Polybot — an AI-driven, automated materials laboratory — to explore processing methods and produce high-quality films. The laboratory is located at the Center for Nanoscale Materials, a user facility at Argonne.
Polybot is the latest method in autonomous discovery, a general approach that combines robotics with the power of AI to accelerate discovery and innovation.
Polybot performing automated workflow. The video is accelerated by four times. (Video by Argonne National Laboratory.)
“Polybot operates on its own, with a robot running the experiments based on AI-driven decisions,” said Jie Xu, a scientist at Argonne. “We are creating a method that highlights how AI and automation can transform chemical engineering and materials science.”
The researchers used Polybot to solve key challenges in electronic polymer processing. For example, the final properties of these materials are influenced by a complex production history. There are nearly a million possible combinations in the fabrication process that can affect the final properties of the films — far too many possibilities for humans to test.
“We faced limited resources and had little knowledge about the vast processing options,” said Henry Chan, a computational materials scientist at Argonne. “Using AI-guided exploration and statistical methods, Polybot efficiently gathered reliable data, helping us find thin film processing conditions that met several material goals.”
Polybot helped the researchers simultaneously optimize two key properties: conductivity and coating defects. Improving conductivity and reducing defects makes devices more reliable and boosts electrical performance.
This fully automated platform streamlines the formulation, coating and post-processing steps, allowing for quick experimentation and data collection. As a result, the team was able to create thin films with average conductivity comparable to the highest standards currently achievable. They also developed “recipes” for large-scale production of these films.
Further, one of the key achievements of this project, according to Argonne research scientist Aikaterini Vriza, is the use of advanced computer programs that can process and analyze images.
“We are creating a method that highlights how AI and automation can transform chemical engineering and materials science.”
—Jie Xu, Argonne scientist
“These programs not only helped us perform experiments and create films, but they also allowed us to capture images and evaluate the quality of the films,” Vriza said. “This information was crucial to our efforts to produce high-quality, highly conductive films.”
In addition to making films, the researchers collected valuable data, which they plan to share with the scientific community through a database. This adds significant value to their work.
“Data is critical,” Vriza said. “We support open-source research, and by sharing this data, we hope to motivate the community to contribute to, test and improve our methodology.”
The impact of this work goes beyond just making electronic polymers in the laboratory. It also establishes important guidelines for large-scale production. The recipes and instructions from this research provide practical advice for scientists and manufacturers who want to explore the potential of electronic polymers in various applications.
This research made use of the Materials Engineering Research Facility at Argonne for electronic printing support and the National Synchrotron Light Source II at DOE’s Brookhaven National Laboratory for wide-angle X-ray scattering characterization.
Other contributors to the work include Chengshi Wang, Yeon-Ju Kim, Rohit Batra, Arun Baskaran, Pierre Darancet, Logan Ward, Yuzi Liu, Maria K.Y. Chan, H. Christopher Fry and Christina S. Miller from Argonne; Naisong Shan and Nan Li from The University of Chicago; and Subramanian K.R.S. Sankaranarayanan from both Argonne and the University of Illinois, Chicago.
“This project is just the beginning,” Xu said. “We’ve shown that our approach works. Next, we want to dive deeper into using AI and automated processes to tackle more real-world challenges and help discover new materials.”
Citation “Autonomous platform for solution processing of electronic polymers.” Chengshi Wang, Yeon-Ju Kim, Aikaterini Vriza, Rohit Batra, Arun Baskaran, Naisong Shan, Nan Li, Pierre Darancet, Logan Ward, Yuzi Liu, Maria K. Y. Chan, Subramanian K.R.S. Sankaranarayanan, H. Christopher Fry, C. Suzanne Miller, Henry Chan & Jie Xu, Nature Communications, Feb. 17, 2025.
Funding: U.S. Department of Energy Office of Basic Energy Sciences, the Laboratory Directed Research and Development program at Argonne and the University of Chicago.
