AI To Go For the Best Solar Perovskites

• A team of researchers in the US is using machine learning technology to optimize the composition of the material. It will help to predict the performance of perovskite solar cells and design strategies. According to the research conducted by 2000 peers, they reviewed perovskite publications which allowed them to collect over 300 data points.

The University of Central Florida has signed for the AI to the perovskite solar cell research. The objective is to develop a system that will help them identify the best materials for the cells.

According to the researchers, machine learning will help scientists develop perovskite cells or improve devices by using a tool that takes into account the cost, efficiency and flexibility issues.

The team from the US collected data from 2,000 publications that were reviewed by the peer and came up with a 300 data point. This information was subjected to a rigorous analysis. The objective was to enable the AI-based system to help in defining perovskite solar cell combinations that give optimal results.

Findings

According to Javan Thomas – the Perovskite Sollar Cells author, the results obtained show that machine learning tools could be used to craft perovskite materials. It will also help to investigate the physics used in developing efficient perovskite solar cells (PSCs). In the book, Javan predicted the best recipe that could be used to make PSCs using band-gap perovskites.

Early this year, researchers at Massachusetts Institute of Technology conducted similar research and came up with a process of screening the new perovskite compounds. They developed the process during their search for perovskite that could be used in developing highly efficient solar cells. The MIT findings showed that the process increased the rate of analyzing and synthesizing new compounds. It did this by speeding up the analyzing process. It increased the speed by a factor of 10 and exposed 2 sets of materials that needed to be studied further.

By using computer modeling, it is possible to narrow down the candidates as it was demonstrated by the work done at San Diego and the University of Sandiego. But scientists will need to go through the process of analyzing the material and synthesizing it in the lab.