Trapping light in solar cells with disordered hyperspectral uniformity

Conventional solar cells use relatively thick (~160 µm) crystalline silicon (c Si) layers to absorb most incoming sunlight. Reducing the absorber thickness to 1–20 µm offers multiple benefits: more efficient charge extraction, lower material cost and usage, relaxed crystal-quality requirements, kerf loss–free and roll to roll manufacturing, and new applications requiring lightweight or curved devices (e.g., automotive).
Ultra thin absorbers, however, suffer from incomplete optical absorption, and conventional light trapping methods are often incompatible. Inspired by disordered hyperspectral uniformity (“hyperuniformity”)—a form of natural order in disordered systems—this Ph.D. thesis investigates light scattering layers that exhibit this special class of spatial correlations.
First, we develop novel numerical tools to analyze light trapping in arbitrary scattering morphologies embedded in full device stacks. These tools handle large disordered systems and provide crucial insight for the rational design of functional optoelectronic devices such as solar cells. Next, we demonstrate 16.7% efficient, 4.8 µm thick c Si solar cells grown epitaxially from SOI wafers.
Our hyperuniform light trapping and anti reflection nanopatterns outperform periodic designs thanks to a superior ability to couple the broad solar spectrum into many guided modes of the absorbing layer.
We further test the devices under real world conditions including oblique illumination, which reveals important pattern design directives applicable to many types of scattering layers. Finally, to maximize impact, we demonstrate a large area self assembly method to produce near hyperuniform nanopatterns at industrially relevant scales, compatible with scalable manufacturing for more sustainable and future-proof photovoltaics.