The Sun constantly transmits trillions of watts of energy to the Earth. It will continue to do so for billions of years. However, we have only just begun to tap into this abundant and affordable renewable energy source.
Solar absorbers are the material used to convert this energy into heat or electricity. Maria Chan, a scientist at the US Department of Energy’s (DOE) Argonne National Laboratory, has developed a machine learning method to test many thousands of compounds as solar absorbers. Her co-author on the project was Arun Mannodi-Kanakitodi, a former Argona postdoctoral fellow who is now an assistant professor at Purdue University.
“We are truly in a new era of applying artificial intelligence and high-performance computing to material discovery.” — Maria Chan, a scientist at the Center for Nanomaterials
“According to a recent study by the Ministry of Energy, solar energy could provide 40% of the country’s electricity by 2035,” Chan said. “And it can help decarbonise the grid and provide a lot of new jobs.”
Chan and Mannoji-Kanakitadi are betting that machine learning will play an important role in realizing this lofty goal. A form of artificial intelligence (AI), machine learning uses a combination of large data sets and algorithms to mimic the way humans learn. It learns by training with sample data and past experience to make better and better predictions.
In Thomas Edison’s day, scientists discovered new materials through a laborious process of trial and error with many different candidates until one worked. For the past several decades, they have also relied on time-consuming calculations that require up to a thousand hours to predict material properties. Now they can reduce both detection processes by invoking machine learning.
Currently, the main absorber of solar cells is silicon or cadmium telluride. Such cells are now commonplace. But they remain quite expensive and energy-intensive to produce.
The team used their machine learning method to evaluate the solar energy properties of a class of materials called halide perovskites. Over the past decade, many researchers have studied perovskites because of their remarkable efficiency in converting sunlight into electricity. They also offer the prospect of significantly lower cost and energy consumption for materials preparation and cell generation.
“Unlike silicon or cadmium telluride, the possible variations of halides combined with perovskites are essentially limitless,” Chan said. “Thus, there is an urgent need to develop a method that can narrow down the number of promising candidates to a manageable number. Machine learning is the perfect tool for this.”
The team trained their method on data from several hundred halide perovskite compositions and then applied it to more than 18,000 compositions as a test case. The method evaluated these compositions on key properties such as stability, ability to absorb sunlight, a structure that does not break easily due to defects, and more. The calculations were in good agreement with the corresponding data in the scientific literature. In addition, the findings reduced the number of compositions worthy of further study to about 400.
“Our list of candidates includes compounds that have already been studied, compounds that no one has ever studied, and even compounds that were not in the original 18,000,” Chan said. “So we’re very excited about that.”
The next step will be to test the predictions with experiments. An ideal scenario would be to use a stand-alone discovery lab, e.g Palibot in Argon Center for Nanomaterials (CNM), US Department of Energy Office of Science. Polybot combines the power of robotics with artificial intelligence to advance scientific discovery with virtually no human intervention.
By using stand-alone experiments to synthesize, characterize and test the best of several hundred top candidates, Chan and her team believe they can also improve the current machine learning method.
“We are truly in a new era of applying artificial intelligence and high-performance computing to material discovery,” Chan said. “In addition to solar cells, our design methodology can be applied to LEDs and infrared sensors.”
This research is reported in the article in Energy and environmental sciencecalled “Data-driven design of a new halide perovskite alloy.»
The research was supported by the Department of Energy’s Office of Science. The researchers used computing resources from the National Energy Research Computing Center, a user agency of the Department of Energy’s Office of Science, and Bebop, which is operated by the Argonne Laboratory Computing Resource Center.
About the Argonne Nanomaterials Center: The Center for Nanoscale Materials is one of five Department of Energy Nanoscale Science Centers, the nation’s premier institutions for interdisciplinary nanoscale research supported by the Department of Energy’s Office of Science. Together, the NSRCs comprise a set of complementary facilities that provide researchers with state-of-the-art capabilities to fabricate, process, characterize, and model nanomaterials, and are the largest infrastructure investment of the National Nanotechnology Initiative. NSRCs are located at DOE’s Argonne, Brookhaven, Lawrence Berkeley, Oak Ridge, Sandia, and Los Alamos National Laboratories. For more information on DOE’s NSRC, please visit https://science.osti.gov/User-Facilities/U ser-Facilities-at-a-Glance.
Argonne National Laboratory seeks solutions to urgent national problems in the field of science and technology. The first Argonne National Laboratory in the country conducts advanced fundamental and applied scientific research in almost all scientific disciplines. Argonne researchers work closely with researchers from hundreds of companies, universities, and federal, state, and municipal agencies to help them solve their specific challenges, advance America’s scientific leadership, and prepare the nation for a better future. Argonne is run by employees from more than 60 countries UChicago Argonne, Ltd for US Department of Energy Office of Science.
US Department of Energy Office of Science is the largest supporter of basic research in the physical sciences in the United States and works to solve some of the most pressing problems of our time. For more information visit https://energy.gov/science.
Granted Argonne National Laboratory. Joseph E. Harmon
Appreciate CleanTechnica’s originality and clean tech news coverage? Consider becoming a CleanTechnica member, supporter, technician or ambassador — or patron on patreon.
Have a tip for CleanTechnica, want to advertise, or suggest a guest for our CleanTech Talk podcast? Contact us here.