Researchers at Northwestern University have designed a new technique that uses a combination of mathematics and machine learning to produce an optimal material for light management in solar cells while simultaneously producing the nanostructured surfaces that work to speed up the fabrication process. The approach could offer an alternative to standard trial-and-error nanomanufacturing and design methods that are time-consuming and require huge resources, increasing the cost-effectiveness of prototype nanophotonic devices.

With much study into nanophotonic materials for light absorption in ultra-thin and flexible solar cells, as reported in Proceedings of the National Academy of Sciences [Lee et al. PNAS (2017) DOI: 10.1073/pnas.1704711114], this approach, which has the advantage of being fast, highly scalable and streamlined, bridges the gap between design and nanomanufacturing, helping broadband light absorption in solar cells. As leader of the design component, Wei Chen, points out, “Instead of designing a structure element by element, we are now designing and optimizing it with a simple mathematic function and fabricating it at the same time”.

“We designed for the whole spectrum of sunlight frequencies, so the solar cell can absorb light over broadband wavelengths and over a wide collection of angles”Wei Chen

The team hope a similar principle could be used to implement color into clothing without the need for dyes and also to develop anti-wet surfaces. For solar cells, the best nanostructure surface comprise quasi-random structures, although such patterns can be problematic and time-consuming to design as there are so many geometric variables to optimize simultaneously to find the optimal surface pattern to absorb the most light.

Using nano-lithography is impractical as it takes far too much time to print. To get round this, they fabricated the quasi-random structures with wrinkle lithography, a nanomanufacturing technique whose stages can be integrated with concurrent design of nanostructures and function, and which quickly transfers wrinkle patterns into different materials to allow practically an unlimited number of quasi-random nanostructures. Wrinkling is a straightforward technique for scalably manufacturing nanoscale surface structures through the application of strain to a substrate.

In demonstration, 3D photonic nanostructures for light trapping on a silicon wafer that could potentially by used as a solar cell were quickly optimized. The material was found to absorb 160% more light in the 800 to 1,200 nanometer wavelength range than alternative designs.

As Chen revealed “We designed for the whole spectrum of sunlight frequencies, so the solar cell can absorb light over broadband wavelengths and over a wide collection of angles”. The researchers now hope to apply their method to a range of other materials, including polymers, metals and oxides, for other photonics applications.