Carnegie Mellon University uses artificial intelligence to develop electrolytes to accelerate battery innovation

According to foreign media reports, in order to meet unique design requirements, researchers are increasingly using machine learning to develop new materials and compounds. This novel approach helps reduce the time to develop and test materials and make new discoveries faster. At Carnegie Mellon University, Adarsh ​​Dave, a doctoral student in the Department of Mechanical Engineering and Tata Consulting Services, applied it to battery design and made amazing discoveries.

(Image source: techxplore)

Dave wants to reduce greenhouse gas emissions, and battery innovation is a relatively easy way to reduce emissions. However, because the chemical reaction process is quite complicated, it often takes a long time to achieve innovation, and the research team began looking for ways to speed it up. This study focuses on aqueous electrolytes. Dave said that this electrolyte is very suitable for storing renewable energy. "Designing high-performance water-based batteries is an important process to solve this problem. However, the number of formulas to choose from is astonishing, which is where our design process comes in."

Dave and his team built a robotic platform called Otto that measures the properties of electrolytes to determine their effectiveness in batteries. Machine learning is combined with Otto to optimize the battery's electrolytes. The computer tells Otto which electrolytes to test, and then Otto tells the computer the properties of those electrolytes. This back-and-forth relationship helps machine learning optimize and find the best electrolyte. Otto can mix and test electrolytes as quickly as a human, but unlike a human, Otto can run 24/7.

Dave and his team used machine learning to discover a "non-intuitive, novel electrolyte." Without this research, this electrolyte might still be unknown to designers. This shows the broad application prospects of machine learning in the future design process. In addition, Otto can automatically operate, speed up the testing and experimental process, and enable scientists to focus on macro research.

Professor Jay Whituck, director of the Scott Institute for Energy Innovation, said: "While robots or algorithms can't replace the intuition of a trained chemist when it comes to innovation, our system can certainly automate and accelerate routine scientific and design tasks. I hope to see colleagues in other labs take the boring stuff out of automation and really accelerate the pace of battery innovation."