| Work Detail |
An international research group has attempted for the first time to build a perovskite solar cell with the help of ChatGPT. The experiment helped scientists identify a number of materials for the cell composition and the results were cells with higher power conversion efficiency compared to that of reference cells built without the material proposed by the large linguistic model. .
Researchers led by Chinas Nankai University have explored ChatGPTs ability to generate hypotheses for materials science and identify untested molecules capable of reducing surface recombination and thus boosting the efficiency of perovskite solar cells.
“We are still nowhere near the Jarvis system from the Iron Man movies, but we are getting close, and what we show in this paper is that we are close enough for the system to be really useful for generating hypotheses,” he told pv magazine corresponding author, T. Jesper Jacobsson. “Admittedly, it took a bit of luck, some domain knowledge, and asking the right kind of question, but it worked nonetheless.”
To find molecules for surface passivation in hybrid perovskite solar cells with a pin architecture, the group used ChatGPT 3.5, with a data cutoff in September 2021. Comparing their interaction with the chat model with brainstorming, the group He stated that it was a continuous dialogue and an exchange of questions and answers. Among the requests they entered into the broad language model (LLM), for example, there was information on compounds available on the market, or at least easily synthesized, reasonably priced, and not excessively toxic.
However, ChatGPT didnt do everything. The scientists manually verified the general plausibility of the suggested materials, then checking academic databases to see if they had already been explored. Through this process, they identified polyallylamine (PAA).
“PAA is a water-soluble biodegradable polymer with applications in areas such as medicine, the synthesis of nanoparticles and the chelation of heavy metal ions,” explained the research group. “ChatGPT provided us with suggestions for other molecules as well, but based on price, availability, toxicity, structural similarity to other surface passivations, and the lack of previous reports using PAA for this purpose, we decided that PAA would be an interesting molecule to use.” explore experimentally.
After a real-world experiment, scientists made 125 devices. The device structure followed a standard pin architecture composed of a soda-lime glass/indium tin oxide (SLG/ITO) substrate, a hole transport layer of MeO-2PACz, an electron transport layer of PCBM-60/ BCP and a silver metal contact. The absorber was based on a triple cation perovskite Cs0.05FA0.91MA0.04PbBr0.15I2.85 with a band gap of 1.54 eV.
Before depositing the PCBM layer, a thin layer of PAA was applied on the perovskite film by spin-coating. In some of the experiments, the PAA solution was 0.015% in isopropyl alcohol (IPA), while in others it was 0.025% or 0.05%. Control samples without PAA treatment were also manufactured.
“The average performance of the device increased by around 2% of units, with a maximum performance of 22.75%,” say the researchers. “This result provides a compelling demonstration of the value inherent in human-AI collaboration.”
The results were presented in “ The use of ChatGPT to generate experimentally testable hypotheses for improving the surface passivation of perovskite solar cells ,” published in Cell. Reports Physical Science . The team consisted of scientists from Nankai University in China and Linköping University in Sweden. |
