A research group led by Professor Akinori Saeki and Graduate Student Chisato Nishikawa (at the time of the study) of the Graduate School of Engineering at Osaka University combined manually operated measuring equipment with robotics to measure optical properties, microwave conductivity, and microscope images, and developed a unique system that performs automatic measurements. Using this system, they succeeded in improving the performance of perovskite solar cell materials that do not contain toxic elements. The study was published in JACS Au on October 23, 2023.
Lead perovskite solar cells, which can be fabricated using a solution coating process, are being developed to replace conventional inorganic solar cells due to expectations for them to be cheaper and lighter. Over the past 14 years, the conversion efficiency of these cells has improved to 26%, almost equivalent to that of silicon solar cells. However, as the toxicity of lead is a major concern, research is also underway to develop solar cells made of relatively low-toxicity elements.
Bismuth and antimony are two candidate semiconductor materials to replace lead. However, the conversion efficiency of these materials is frequently as low as 1%. Furthermore, investigations into the composition of raw materials in solution, as well as the solution application process (solvent, additives, heat treatment temperature, etc.), have not been undertaken. In addition to the several process parameters involved, the fabrication and evaluation of coated solar cells require considerable time and resources.
To date, the research group has developed a microwave conductivity method that quickly and easily measures signals that correlate with solar cell performance. Until now, all measurements were made manually, which limited the materials and conditions that could be examined, as well as the effectiveness of the method in the presence of other performance-affecting factors.
To address these issues, they have independently developed an automatic evaluation device using a cooperative robot and succeeded in significantly reducing the measurement time (by approximately sixfold), as well as increasing accuracy (by approximately fivefold). This not only dramatically increases the number of conditions and materials that can be examined, but also allows researchers to focus on more creative tasks.
In addition to microwave conductivity measurements, the automated evaluation system incorporates measurements of optical absorption and emission spectra, which are important properties of solar cell thin films, as well as optical microscopy, which can observe thin film surface morphology. The measured optical micrographs can be automatically subjected to histogram analyses of shading, fast Fourier transform analyses, and particle analyses. Consequently, a large quantity of highly accurate and uniform experimental data can be obtained from a single thin film sample. They have also developed their own control and analysis software for the equipment, making it a one-of-a-kind evaluation method.
Using this automated evaluation system, they investigated the composition, additives, and heat treatment temperatures of lead-free solar cells composed of cesium, bismuth, antimony, and iodine. Specifically, thin film samples were prepared under 576 conditions using combinations of 12 composition ratios, four additives, three additive concentrations, and four heat treatment temperatures.
Based on the automatic measurement results, they fabricated solar cells under 40 conditions and evaluated their conversion efficiency. They found that the conversion efficiency of an element fabricated without additives at a low heat treatment temperature was 0.35%. The newly developed material process successfully improved this figure to 2.36%. Furthermore, the obtained solar cell conversion efficiency and automated measurement data were examined by machine learning and statistical analysis. Consequently, the standard deviation of microwave conductivity signals and the shading histogram obtained by optical microscopy were evaluated as effective search guidelines to explore high-efficiency material processes. The automated evaluation of high-efficiency lead perovskite solar cell thin films was found to be consistent with this search guideline, demonstrating the validity of the proposed model.
Saeki stated, "In recent years, there has been a lot of research into material exploration and phenomenon clarification using machine learning. The automation of experiments such as this study is very effective for obtaining large quantities of uniform experimental data that are optimal for machine learning, as well as exploring areas where existing data is scarce. Although the measurements themselves were automated in this study, the 576 thin film samples were prepared manually by Ms. Nishikawa. In the future, we plan to automate the production of thin films, and these results represent the first step toward that goal."
Journal Information
Publication: JACS Au
Title: Exploration of Solution-Processed Bi/Sb Solar Cells by Automated Robotic Experiments Equipped with Microwave Conductivity
DOI: 10.1021/jacsau.3c00519
This article has been translated by JST with permission from The Science News Ltd. (https://sci-news.co.jp/). Unauthorized reproduction of the article and photographs is prohibited.