Of course, there are many ways to optimize solar energy collection. One might add concentration optics, or possibly a mechanical solar tracker. Silicon surface texturing is another simple and effective technique to boost the energy conversion efficiency of most module designs. Instead of forming the nice, smooth gray surface that some of us remember from the cells in old calculators, it is instead helpful to design the top and bottom of the silicon sheet to include unique patterns. Example topologies include small pyramids, gratings to enhance back reflection, or quasi-random structures to increase inter-material scattering. These textures both prevent loss of photons from surface reflection and also enhance absorption of photons within the silicon. Furthermore, recent strides in promising fabrication techniques, such as nanoimprint lithography, may enable profile design at nanometer precision and over large scales.
However, determining the ideal pattern for a given panel material and module geometry is not so simple. Simulation is an invaluable tool early on in the design stage, but typical patterning structures fall within a range of sizes that are challenging to accurately and efficiently simulate together. For example, pyramidal textures are large enough to require analysis via ray tracing, whereas the sub-micron features of reflection gratings must be modeled via wave propagation (e.g., using rigorous coupled wave analysis). When different textures are combined within the same silicon wafer solar cell, designers have a difficult choice: either rely on inexact ray tracing, or perform a computationally demanding simulation of the entire structure using a wave-based approach.
This work, by Eisenlohr et al., proposes a very simple yet powerful toolset to efficiently simulate both ray and wave-based phenomena together within a single multi-layered structure. Instead of relying upon a unified simulation pipeline, their strategy splits the trajectory of light through each cell into a set of discrete reflections, including absorption or loss in between. This enables a three-step modeling approach. First, the authors treat each optical surface separately to determine their optical reflection and transmission properties, using any desired model (e.g., with a wave-based simulation, or analytically). Second, the authors discretize the input-output relationship of optical power at each surface into a power redistribution matrix. Third, each optical surface of interest is connected together, along with effects within the silicon substrate and any other possible thin films or coatings, through a series of matrix equations. The end result is simple and efficient modeling platform for both ray and wave-based effects, which accurately predicts the response of a wide variety of surface texture combinations.
The benefit of Eisenlohr et al.’s formalism does not just lie within its improved efficiency and potential ability to scale to large structures. In addition, it provides a helpful conceptual shift away from the common strategy of treating the effects of optical rays and waves as two distinct categories. Through their merger, one can simultaneously take advantage of the computational efficiency of ray tracing and the accuracy of diffractive modeling. This strategy mirrors recent procedures applied within the field of computer graphics, which can quickly render images containing complicated optical reflections, subsurface scattering and diffraction. Hopefully, this new modeling pipeline can eventually help designers find ideal surface texture combinations to produce cheap and efficient solar cells for the masses.
You must log in to add comments.