Photovoltaic power generation, a key renewable energy resource, commonly uses wafer-based crystalline silicon solar cells. An alternative is thin-film silicon solar cells (TFSSCs), which would particularly benefit systems with generation capacity of more than one gigawatt because of the abundance and non-toxicity of their source materials. In TFSSCs, trapping the incident light within thin silicon films is crucial to improve the photon absorption and conversion efficiency. To scatter the incident light and elongate the optical path length inside the cell, TFSSCs use textured substrates, which normally have randomized textures with sizes ranging from sub-microns to several microns. Potentially, a TFSSC's optical path length can be enhanced by a factor of 4n2, where n is the refractive index of thin-film silicon.1
In recent years, developers have sought a more sophisticated platform for light trapping, studying periodically textured substrates or surface gratings.2, 3 For example, if we apply an optimized periodic texture to amorphous silicon (a-Si:H) solar cells, it improves the conversion efficiency as well as the short-circuit current density (JSC) at the same level as state-of-the-art random textures.3 However, we have yet to demonstrate the full potential of using periodic textures in microcrystalline silicon (μc-Si:H) solar cells, which need more efficient light confinement because of their very small absorption coefficient in the IR region. Excessively steep textures, such as binary surface-relief gratings or pyramidal textures with V-shaped valleys, can induce defects in μc-Si:H films and impair photovoltaic performance,4 so it is important to find textures that are suitable for high-quality film growth as well as good light confinement.
Figure 1. Scanning electron microscope images of four periodic honeycomb textures for thin microcrystalline solar cells with periods of (a) 1μm, (b) 1.5μm, (c) 2μm, and (d) 3μm.
Photolithography produces a wide variety of periodic textures that are uniform and easily reproduced. Although expensive, the technology provides the best way to make a systematic survey of periodic textures for μc-Si:H solar cells. We can, alternatively, achieve a smooth profile of textures for high-quality μc-Si:H films using gray scale masks or isotropic etching.
Figure 2. External quantum efficiency (EQE) spectra of the 1μm-thick μc-Si:H solar cells fabricated on a flat substrate and a honeycomb-textured substrate with P = 1.5μm.
We developed periodically textured substrates with hexagonal dimple arrays, or 'honeycomb texture' (see Figure 1), and applied them to single-junction μc-Si:H solar cells.5 We chose the substrate-configuration in μc-Si:H solar cells because of the non-transparency of our substrates.
We fabricated the honeycomb textures by forming a hexagonal photo-resist pattern on a silicon wafer with a thermally-grown silicon dioxide (SiO2) film. We then soaked the wafer in a diluted buffered hydrofluoric acid solution to transfer the hexagonal pattern to the SiO2 film. We removed the remaining resist and deposited a stacking of silver and gallium-doped zinc oxide (Ag/ZnO:Ga) films on the honeycomb textured SiO2 layer by sputtering in order to obtain a highly reflective and conductive surface. Using plasma-enhanced chemical vapor deposition, we deposited n-, i- and p-μc-Si:H layers, where n- and p- type layers are made by doping phosphorous and boron into the silicon during deposition, and the i-layer is the intrinsic (or non-doped) microcrystalline silicon. We further deposited a film of indium tin oxide (In2O3:Sn) and a silver finger grid by sputtering to make μc-Si:H cells. In this work, we fabricated honeycomb textures with periods of P = 1.0–4.0μm and aspect ratios of H/P=0.1–0.25, where H denotes the peak height of the texture.
From the survey, we obtained a high JSC of 26.2mA/cm2 in a 1-μm-thick μc-Si:H cell with an active area of 1cm2 by using the honeycomb texture with P = 1.5μm and H/P= 0.23 (see Figure 2). This JSC is almost 50% higher than that of the reference flat cell (17.5mA/cm2), showing a significant improvement in the cell's light confinement. Furthermore, we achieved an efficiency of 10.1% using the honeycomb texture, which improved by 30% compared with the flat cell. We found the optimum period to be longer than that predicted by optical simulations, suggesting that the flattening of the surface texture during film growth plays an important role in the operation of TFSSCs.
So far our results have shown that periodic textures could improve the performance of μc-Si:H solar cells. In future work we would apply the results described here to the superstrate-configuration, which is the basis for construction of most current TFSSCs. However, we would need to develop a low-cost fabrication process to make such periodic textures commercially viable.
This work was supported by the New Energy and Industrial Technology Development Organization, Japan.
Hitoshi Sai, Michio Kondo
Research Center for Photovoltaic Technologies (RCPVT)
National Institute of Advanced Industrial Science and Technology
Hitoshi Sai received his PhD from Tohoku University in 2004 and is currently a researcher at RCPVT. His interests include light management in thin-film solar cells.
Michio Kondo received his PhD from Osaka University in 1987 and is director of RCPVT.
Photovoltaic Power Generation Technology Research Association (PVTEC)
Kimihiko Saito received his PhD from Tokyo Institute of Technology in 2010. He manages PVTEC's Tsukuba Laboratory and is a professor at Fukushima University.
1. E. Yablonovitch, J. Opt. Soc. Am. A 72, p. 899, 1982.
2. Z. Yu, A. Raman, S. Fan, Opt. Express 18, p. A366, 2010.
3. C. Battaglia, C. M. Hsu, K. Soderstrom, J. Escarre, F. J. Haug, M. Charriere, M. Boccard, ACS Nano 6, p. 2790, 2012.
4. M. Python, O. Madani, D. Domine, F. Meillaud, E. Vallat-Sauvain, C. Ballif, Solar Energy Mater. Solar Cells 93, p. 1714, 2009.
5. H. Sai, K. Saito, M. Kondo, Appl. Phys. Lett. 101, p. 173901, 2012.