Tech Insights

Automated Solar Cells: Are Robots the Future of Solar Power Innovation?

November 28, 2023 by John Nieman

Solar energy technologies are widely used to capture one of the planet's most abundant natural power sources, but solar cells can still be improved through material experimentation. Such research can be costly and time-consuming, but robots and machine learning can expedite the process, improving solar cell technology. 

Developing solar is one of the most important methods for defraying dependence on the power grid. Common solar panels can power greenhouses, appliances, and chip technology for countless electronics. However, engineers are still searching for innovative ways to advance the material technology of silicon solar cells. 


Conventional solar cells made from silicon.

Conventional solar cells made from silicon. Image used courtesy of the National Renewable Energy Laboratory 


Such research can be slow, labor intensive, and costly, however. Hence, engineers are pioneering robotics and machine learning to search for new material components for the next generation of solar cell technology more efficiently. 


Limitations of Silicon in Solar Cells

The demand for silicon solar cells has increased steadily as the collective pivot toward renewable energy continues, and there is no sign of a slowdown. 


Projected adoption of solar energy.

Projected adoption of solar energy. Image used courtesy of the Department of Energy


The reliance on silicon for solar cell production can be explained simply. Silicon is an efficient semiconductor, has a long lifetime, and is relatively low cost primarily because it is one of the most common materials on the planet. Silicon is second only to oxygen when it comes to sheer abundance of natural availability, making it an obvious choice for a myriad of engineering applications. Currently, it is used to construct 95% of the solar modules sold, making it the dominant material used and practically ubiquitous in solar technology. 

However, some drawbacks warrant further investigation into other potential materials that might help advance solar technology and push through the limits of silicon. For example, conventional silicon solar cells are not flexible enough to be installed on car roofs, given the aerodynamic curvature of virtually every vehicle. This limitation has prevented the implementation of solar power in one of the most obvious places for electric cars as they commute. 

In addition, silicon has significant costs associated with its mining and extraction, leading researchers to explore perovskites as an alternative. Even though the performance of perovskite solar cells (PSCs) has been impressive, they are, in fact, toxic because of their lead levels, which harm the environment and damage human health.

PSCs also degrade quickly, so what initially seemed like the ideal material to move solar technology forward has proven to have drawbacks. 


Machine Learning, AI, and Robotics 

Given this slow, plodding process of researching, testing, and experimenting with novel solar cell development materials, Osaka University researchers are pioneering an automated system that uses robots and machine learning algorithms to evaluate materials for solar cells. 

The research process for solar cell material is predicated on consistent criteria. Researchers need to test for toxicity, evaluate performance capabilities, and ensure the material is sufficiently abundant to supply the world’s demand for solar power. The lead researcher,  Chisato Nishikawa, has helped design a novel robotics approach to execute optical microscopy, time-resolved microwave conductivity analysis, and photoabsorption spectroscopy without human intervention. 

The team analyzed 576 thin-film semiconductor samples and was able to deploy artificial intelligence (AI) to help amass and analyze data associated with all these samples. 


Semiconductor sample films and associated analysis.

Semiconductor sample films and associated analysis. Image used courtesy of ACS Publications


The robot can evaluate prospective materials and perform these time-consuming tasks accurately and precisely, removing costly labor and increasing research efficiency. In fact, with the aid of this automated system, the team can assess materials in one-sixth the time when people handled the research exclusively.

Even though this study did not find the supposedly ideal material for the next generation of solar cells, it demonstrated the potential of AI and robotics to increase innovation efficiency dramatically. By advancing research techniques, engineers are hastening the development of solar cell technology that can better capture one of the most abundant sources of natural power.