The solar absorbent is a material for converting such energy into thermal energy or electric energy. A research group of the Argonne National Laboratory of the US Department of energy (DOE) has developed a machine learning method for screening thousands of compounds as solar absorbers.
A recent study by the US Department of energy shows that by 2035, solar energy can power 40% of the country's electricity. According to the researchers, machine learning will play an important role in achieving this noble goal. As a form of artificial intelligence (AI), machine learning uses a combination of large data sets and algorithms to imitate human learning methods. It learns from the training of sample data and past experience to make better predictions.
At present, the main absorber of solar cells is silicon or cadmium telluride. But their manufacturing costs are still quite high and energy intensive. The researchers used their machine learning method to evaluate the solar energy characteristics of a class of materials called halide peroxides.
In the past decade, many researchers have been studying peroxides because of their remarkable efficiency in converting sunlight into electrical energy. They also offer the prospect of lower cost and energy input for material preparation and battery manufacturing.
Argonne researchers trained their method with the data of hundreds of halide peroxide components, and then applied it to more than 18000 components as test cases. This method evaluates the key characteristics of these components, such as stability, ability to absorb sunlight, structure that is not easy to break due to defects, and others. They found that the calculated results were in good agreement with the relevant data in the scientific literature. Moreover, these findings reduced the number of components worthy of further study to about 400.
Maria Chan, who led the study, said: "our candidate list includes compounds that have been studied, compounds that have not been studied, and even compounds that are not included in the original 18000 compounds." So we are very excited about this. "
The team next plans to use experiments to test the predicted results. The ideal scenario is to use an independent discovery laboratory, such as polybot of Argonne's nanomaterials Center (CNM), a user facility of the US Department of energy's office of science. Polybot combines the power of robotics and artificial intelligence to drive scientific discovery with little or no human intervention.
By using independent experiments to synthesize, characterize and test the best of their hundreds of main candidates, Chen and her team expect that they can also improve the current machine learning methods.
"We are indeed in a new era of applying artificial intelligence and high-performance computing to material discovery." "In addition to solar cells, our design method can be applied to LED and infrared sensors," Chan said
