Spectrophotometric Analysis of Stochastic Hybrid Black Silicon Nanostructures for Crystalline Silicon Photovoltaic Cells

By J. J. Tyson

Published November 2023

Black silicon nanotextures offer exceptionally low levels of reflectance and are of interest to the field of solar photovoltaics. The nanowires that form this texture create a graded refractive index, allowing light to be absorbed with high efficiency. Explored here is the application of these nanotextures on top of conventional microscale pyra- mids, combining the advantages forwarded by the latter, predominantly being a second chance for absorption, and the former, being light steering. These structures are known as hybrid black silicon. Variations on the wet chemical etch parameters are explored and related to topological features, which can, in–turn, be related to front surface reflectance. The hybrid black silicon textures created are shown to exhibit reflectance as low as 0.7%.

An advanced hemispherical reflectometry system is reported, designed for measuring the optical characteristics of a variety of samples resolved against wavelength, angle of incidence, and polarisation. Variable angle reflectance data enables a new perspective on the interaction between electromagnetic waves and nanostructures, which do not interact with light in the same way as their microscale counterparts. The versatility of this data is demonstrated for photovoltaics when combined with geographic spectral irradiance data. Solar cell optical performance, when structured with the textures measured in the reflectometer, is successfully predicted should that cell be placed in Southampton, UK. This mathematical formulation is capable, alongside the appropriate angle–resolved reflectance data, of approximating the optical performance of a given sample when situated almost anywhere in the world.

Supporting the reflectance data gathered through this work is a black silicon nanowire model, pseudorandomised using custom–made algorithms against a set of desired surface features. The model uniquely generates complex surface topologies that meet the requirements of electromagnetic wave optics simulations. The model reported showcases an accuracy within the ±2% relative error against measured reflectance, and offers a new, fast, and accurate way of simulating nanostructures without the need to manufacture them in bulk.

This work is licensed under CC BY-SA 4.0

---

Contents

TEXT

---

Introduction

Motivation

Amid an energy crisis, innovative and sustainable sources present themselves in various forms. Fossil fuels are not sustainable, and come paired with significant, often irreversible environmental changes to our planet. Using either disadvantage as a driving force, research into ‘renewable’ energy sources has ramped-up considerably over the past couple of decades. The impact potential of these renewables on the science, technology, and engineering sectors is high; alongside their rapid adoption within residential, commercial, and industrial settings, the inevitable transition between the traditional fossil and progressive renewable fuels is already under way.

Techniques have been developed to harvest wind, tidal, wave, solar, vibrational, and thermal energies with varied eciencies and points of application. Nuclear reactions are also habitually referred to as being within the renewables family despite having an inherently limited, albeit extensive, supply of reactants. One could argue that the greatest source of energy, delicately utilised by all forms of life on the planet, comes directly from the sun. Solar radiation provides an average surface power density of 1.361 kW/m^{2 1} at the outer limit of Earth’s atmosphere. Adjusted accordingly for atmospheric attenuation, this average power density yields a figure of approximately 1 kW/m² at sea level across the world, all of which is available for direct solar energy harvesting using photovoltaic (PV) devices. Over three generations, these technologies have seen many advances in their structure, formulation, and manufacture, yielding many different forms of solar PV. Despite this, each is designed to operate based on the same principles outlined by the quantum theory of light and semiconductor physics.

Table 1: An overview of the three types of solar PV. Shorthand expansions: amorphous Si (a-Si),² cadmium telluride (CdTe),³ copper indium gallium selenide (CIGS),⁴ organic PV (OPV) ⁵.

Technologies	Comments
a-Si	Inexpensive, poor efficiency
(m)c-Si	Tried and tested, long service life, high optical losses
CdTe, CIGS	Inexpensive, elemental scarcity (In, Te), high toxicity (Cd)
Perovskites, OPV	Efficiency to rival Si, simple to manufacture, poor stability

The objective of this research is to improve the anti-reflectivity of silicon-based solar cells using novel nanoscale surface texturing techniques. Over time the improvement of both crystalline (c-Si) and multi-crystalline silicon (mc-Si) based PV has slowed dramatically due to the Shockley-Queisser limit, though this has been increasing again in more recent years. This limit describes the maximum efficiency that can be obtained with a single junction semiconductor substrate based on its band gap energy. For silicon, with a band gap of 1.17 eV, [^6] this limit is widely accepted to be between 32% and 33%. [^7] The current leading research-grade c-Si cell holds an efficiency of 26.8% [^8] and can be seen on the National Renewable Energy Laboratory’s best research cell efficiencies chart in figure 1.1.

Figure 1.1: The best research cell eciencies; a time-lapse visualisation on the progression of single junction c-Si and mc-Si PV technologies. This plot is courtesy of the National Renewable Energy Laboratory, Golden, CO. (2022).

Given that the quality of solar grade silicon (SG-Si) for manufacturing solar cell technologies is already high, with a purity metric of 6 N, other methods of increasing the efficiency must be explored separate from substrate preparation in developments where only one junction is being employed. The use of higher purity substrates such as electronic grade silicon (EG-Si) is not warranted for solar applications ⁶. This is due to the fact that there is very little to gain in overall efficiency, and that the cost of producing the large substrates required for solar PV would be extortionate at this purity level. There has been the development of multi-junction devices that are capable of exceeding the single junction Shockley-Queisser limit ⁷. One of the first examples of such a device arose in 1982 in the form of a two-junction aluminium gallium arsenide to gallium arsenide cell with an efficiency surpassing 21% ⁸. However, despite its potential offerings, it is arguably one of the most expensive cell structures to manufacture due to its complexity and layer-by-layer fabrication method ^9,10,11. Surface texturing techniques have, as an alternative, become the next step in increasing this efficiency further by enabling as much light to be drawn into the device as possible.

References

[^6] W. Bludau, A. Onton, and W. Heinke, “Temperature dependence of the band gap of silicon,” Journal of Applied Physics, vol. 45, pp. 1846–1848, Apr. 1974.

[^7] S. R¨uhle, “Tabulated values of the Shockley–Queisser limit for single junction solar cells,” Solar Energy, vol. 130, pp. 139–147, June 2016.

[^8] M. A. Green, E. D. Dunlop, G. Siefer, M. Yoshita, N. Kopidakis, K. Bothe, and X. Hao, “Solar cell eciency tables (version 61),” Progress in Photovoltaics: Research and Applications, vol. 31, pp. 3–16, nov 2022.

Crystal Growth_, vol. 360, pp. 61–67, dec 2012.