pv magazine spoke with Fan Bin, founder of GCL Optoelectronics, about how high-throughput experimentation and AI-driven optimization are accelerating perovskite development, and why stability, not efficiency, remains the key challenge on the path to mass production
Over the past decade, perovskite solar technology has made remarkable progress, both in terms of stability and efficiency, but challenges remain on the path to commercialization. GCL Optoelectronics is now ramping up its first gigawatt-scale production line using the same technology, with an initial capacity of 500 MW.
According to Fan Bin, founder of China’s GCL Optoelectronics, efficiency is no longer the main obstacle. “Our best tandem modules have reached 29.5% in the lab, and we are targeting around 27% on the mass production line. That is already very competitive compared to silicon,” he said pv magazine. “The main remaining challenge is stability.”
Even with rigorous manufacturing, achieving consistent stability at scale remains difficult. “Our modules can already pass the IEC 61215 standard even at three times the test time. But as we move to full-scale production, where one module is produced every ten seconds, it is still a challenge to ensure that each module performs equally well,” Fan explains.
To address this, GCL Optoelectronics has focused on high-throughput experimentation and AI-based optimization. In their laboratory in Singapore, the team uses a system that can manufacture about 1,800 small 1-inch cells per day and measure their performance. This generates thousands of data points every day, which are then fed into AI models in a process known as active learning. Fan Bin describes the goal: “We hope that the AI can become predictive and eventually propose and test its own material changes. This is possible today because automation tools are much more advanced than they were 20 years ago,” said Fan.
The company works with secret partners from France, Germany and China and integrates AI models such as ChatGPT and Gemini into their workflow. Although such tools are largely redundant in the silicon industry, they are proving invaluable for perovskite solar development, especially in optimizing absorber materials, buffer layers, and passivation techniques.
GCL Optoelectronics also stands out for its four-terminal tandem approach (4T). Unlike most solar manufacturing companies, which reportedly fabricate perovskite directly on top of silicon in a two-terminal (2T) configuration, Fan Bin’s team first produces large-area perovskite modules on glass and then combines them with silicon. “In the early stages, many people believed that producing perovskite modules on a square meter scale would be extremely difficult. Now we have proven that it is possible. I think it is more complicated to fabricate perovskite directly on the texture of a silicon wafer,” he said. “CATL, for example, is pursuing a similar strategy, as are some Chinese display panel manufacturers entering the perovskite field.”
In terms of manufacturing methods, small cells are produced using spin coating, but this method cannot be scaled for larger modules. Instead, GCL Optoelectronics relies on slot-die coating based entirely on solution processing. “Some groups are trying vacuum or sublimation methods, but perovskite formulations are very complex. Different components evaporate at different rates, making precise control extremely difficult. Therefore, we continue to rely on solution processing,” Fan concluded.
This content is copyrighted and may not be reused. If you would like to collaborate with us and reuse some of our content, please contact: editors@pv-magazine.com.
