Perovskites are a group of materials that are vying to replace silicon-based solar photovoltaics, which are now in widespread use. They offer panels that are significantly lighter and thinner, that can be manufactured in vast quantities with ultra-high throughput at room temperature rather than hundreds of degrees, and that are easier and less expensive to transport and install. However, it has taken a long time to turn these materials from minor laboratory studies into a product that can be mass-produced cheaply.
Perovskites are a class of materials fighting to replace silicon-based solar photovoltaics, which are currently in use. They provide panels that are substantially lighter and thinner, can be created in large quantities at room temperature rather than hundreds of degrees, and are easier and less expensive to transport and install. However, it has taken a long time to transform these materials from small-scale laboratory experiments into a mass-marketable commodity.
Even within one manufacturing process among numerous options, producing perovskite-based solar cells requires optimising at least a dozen or so variables at once. A new system based on a revolutionary approach to machine learning, on the other hand, might hasten the development of optimum production procedures and aid in the realisation of the next generation of solar power.
The technique, which was created over the last few years by academics at MIT and Stanford University, allows machine learning to include data from previous experiments as well as information based on personal observations by experienced personnel. This improves the accuracy of the results and has already led to the production of perovskite cells with an 18.5 percent energy conversion efficiency, which is competitive in today's market.
Tonio Buonassisi, MIT professor of mechanical engineering, Stanford professor of materials science and engineering Reinhold Dauskardt, recent MIT research associate Zhe Liu, Stanford doctoral graduate Nicholas Rolston, and three others recently published a study in the journal Joule.
Perovskites are a class of layered crystalline compounds determined by the atoms' crystal lattice structure. There are thousands of such compounds that can be made in a variety of methods. While spin-coating is used in most lab-scale perovskite material development, it is not viable for larger-scale manufacturing, thus corporations and labs throughout the world have been looking for ways to translate these lab materials into a practical, manufacturable product.
"Taking a lab-scale technique and transferring it to something like a startup or a production line is always a major issue," says Rolston, who is now an assistant professor at Arizona State University. The team looked at a method termed rapid spray plasma processing, or RSPP, that they thought had the most potential.
A moving roll-to-roll surface, or sequence of sheets, would be used in the manufacturing process, with the precursor solutions for the perovskite compound sprayed or ink-jetted as the sheet rolled by. The material would then go through a curing stage, producing a continuous and quick output "with throughputs that are higher than any existing photovoltaic technology," according to Rolston.
"The key breakthrough with this platform is that it will allow us to scale in ways that no other material has," he continues. "Due to the processing that is done, even materials like silicon require a substantially longer timescale." Whereas [this method] is more akin to spray painting."
At least a dozen variables can influence the result of the process, some of which are more controlled than others. The composition of the beginning materials, the temperature, the humidity, the processing path speed, the distance of the nozzle used to spray the material onto a substrate, and the curing procedures are all factors to consider. Many of these variables might interact, and if the procedure takes place outside, humidity, for example, may be uncontrollable. Because it is impossible to evaluate all conceivable combinations of these factors through testing, machine learning was used to guide the experimental procedure.
However, while most machine-learning systems use raw data such as electrical and other properties of test samples, they rarely include human experience such as qualitative observations of the visual and other properties of the test samples made by the experimenters, or information from other experiments reported by other researchers. So, using a probability factor based on a mathematical technique called Bayesian Optimization, the researchers devised a way to include such outside information into the machine learning model.
Reference: “Machine learning with knowledge constraints for process optimization of open-air perovskite solar cell manufacturing” by Zhe Liu, Nicholas Rolston, Austin C. Flick, Thomas W. Colburn, Zekun Ren, Reinhold H. Dauskardt and Tonio Buonassisi.
You must be logged in to post a comment.