Silicon films with light-trapping structures are fabricated based on Bi2O3 nano-islands, which are obtained by annealing Bi nano-islands in the air at 400°C. The topography exhibits the maximum altitude of over 600nm and the root-mean-square roughness of 150nm, with the lateral size of single island of about 1μm. Highly crystallized sputtered silicon, realized by Cu-induced crystallization, is used to be a light-absorbing layer. Reflectivity of the samples with different thickness of silicon has been studied to reveal the light-trapping efficiency. The average reflectivity under AM1.5 illumination spectrum is 12% when silicon is 480nm thick and the reflectivity for the long wavelength region between 800nm and 1100nm is less than 10% when the silicon is 1.2μm thick. This is a promising low-cost structure for crystallized silicon thin-film solar cells with high efficiency.
© 2010 OSA
Thin-film solar cells (TFSC) have attracted much attention in recent years for the advantage of material usage. Generally, the TFSC can be divided into several kinds, such as silicon [1,2], compound , organic  and dye-sensitized TFSC . Only some products of silicon and compound TFSC can be found in the photovoltaic market, like hydrogenated amorphous silicon (a-Si:H), hydrogenated microcrystalline silicon (mc-Si:H) , CdTe , CIGS  and GaAs  TFSC, and other kinds are still under research. Si TFSC is believed to be one of the most promising candidates to dominate the photovoltaic market in future, since the material reserve is abundant and the manufacturing is relatively mature. Due to the light-induced creation of metastable defects, which is known as the Staebler-Wrronski effect, a-Si solar cells are always degraded to some extent soaked by light . Recrystallization from a-Si to poly-Si is an effective way to solve the problem, and the absorption spectrum edge can be extended from 800nm to 1100nm. Various Si recrystallization technologies have been reported, like excimer laser annealing (ELA) , solid-phase crystallization (SPC)  and metal induced crystallization (MIC) . We chose MIC to complete the recrystallization, because it can be operated at a temperature lower than 500 °C and soda-lime glass can be used as a substrate . Al, Cu and Ni are the most common promising metals to realize MIC [15–17]. Al inducing process introduces a heavy p-type doping. Cu and Ni are two possible low-cost metals to obtain highly crystallized poly-Si which is suitable for fabricating solar cells. We chose Cu in this paper.
Nevertheless, a-Si has higher absorption strength compared to poly-Si due to direct absorption character. To maintain the light absorption, thicker active layer is needed for poly-Si, which leads to a higher cost. Light-trapping can increase the light absorption ratio which allows decreasing the volume of the active layer . Light-trapping structure usually can be introduced by fabricating textured topography of Si films, such as pyramids structured surface on monocrystalline silicon produced by anisotropic etching , textured ZnO surface produced by wet-etching and textured Al produced by anodic oxidation [20,21]. Light trapping can be realized by two factors simultaneously, multiple reflections by surface textures and enhancement of the internal reflections by scattering the incident light at oblique angles beyond the total reflection angle . However, in most cases, complicated treatment is needed to obtain textured topography, like soaking in an acid solution, lithography, etc. which increases the fabrication cost and increase the risk of introducing unexpected impurities. A simple process to introduce an efficiency light-trapping structure without complicated treatment is obviously attractive.
In this paper, we develop a simple way to fabricate a light-trapping structure with textured highly-crystallized sputtered silicon film based on Bi2O3 nano-islands. Only two traditional deposition technologies, sputtering and evaporation, are used here and both are available on large areas. So in principle the preparation method can be applied to large substrate.
A schematic sketch of the cross section of the light-trapping structure is shown in Fig. 1 . Soda-lime glass is used here as a substrate. All the films are deposited by DC magnetron sputtering except Bi. A 20nm thick Cr layer is first deposited, which is supposed to enhance the adhesion of Bi nano-islands and glass. Next Bi is evaporated to fabricate the nano-islands at a higher substrate temperature of 180 °C, as shown in Fig. 2(a) . The average thickness of the Bi film is 100nm. Then the Bi islands are annealed in the air at 400 °C for 10 minutes. During this process, the Bi nano-islands will be oxidized to be Bi2O3 nano-islands. However, Bi2O3 presents α-phase at room temperature and hence it is not well conductive . So a 100nm thick Ag layer is deposited to cover the islands, behaving as a back electrode and a light reflector simultaneously. Subsequently a 20nm thick Cu layer is deposited as the inducing metal. At last, a 1.2μm thick Si is deposited from a p-type silicon target (99.999%). The whole structure is annealed in a furnace, filled with nitrogen to protect against oxidation, at 530 °C for 10 hours.
3. Results and discussion
The key step to introduce light-trapping is to fabricate Bi nano-islands. According to Fig. 2(a), the average lateral size of the islands is about 800nm and it can be easily tuned by varying the weight of Bi. Because the melting point of Bi is only 271 °C, the Bi nano-islands cannot maintain the topography during the following silicon recrystallization process that will be operated at 530 °C.
However, the melting point of Bi2O3 is over 800 °C, much higher than the temperature needed in the metal induced crystallization process. Bi2O3 can be easily obtained by annealing Bi in the air, which would keep the rough surface topography well as shown in Fig. 2 (b) and Fig. 2(c). After annealing, the lateral sizes of Bi2O3 islands become a little larger than Bi islands. The average lateral size increases to about 1μm, and many smaller islands exist among the bigger ones. Figure 2(d) shows the surface topography of 1.2 μm thick a-Si as a result of the transfer of that of the Bi2O3 islands. Figure 2(e) and Fig. 2(f) shows the topography of poly-Si obtained after annealing, more angular than that of a-Si. The Bi2O3 islands coating behaves quite stable during the whole annealing process, demonstrating good adhesion.
Figure 3(a) is the Atomic force microscopy (AFM) image of Bi2O3, which shows maximum altitude of over 600nm. Figure 3(b) shows the surface topography of 1.2μm thick Si deposited on Bi2O3 nano-islands after annealing. According to Fig. 2 and Fig. 3, the surface topographies of Bi2O3 and poly-Si nano-islands are slightly different. The poly-Si nano-islands are a little bigger and sharper than Bi2O3 nano-islands. To further analyze the alteration of the topography with Si thickness, root-mean-roughness (Rms) of samples with different Si thickness is characterized, and the result is shown in Fig. 4 . Apparently, the Rms changes with the Si thickness. When Si thickness increases from 720nm to 960nm, Rms decreases from 150nm to 120nm, and when it increases from 960nm to 1200nm, Rms goes back up to nearly 160nm.
The light trapping efficiency relies on light scattering into angles beyond the total reflection angle, and hence relies on the intensity scattered into large oblique angles. The total haze of reflected light in air by the Bi2O3 nano-islands coating is measured to be 87% (±5%). The light source used is a standard light source—-CIE (Commission Internationale de L'Eclairage) C illuminant. Furthermore, angle resolved scattering (ARS) dependence on light scattering in air is also measured, shown in Fig. 5 . The incident light source used is a laser beam with λ = 632.8nm and the incident angle is set nearly normal at 5° (θi). The scattering property of Bi2O3 coating is better than the Asahi-U substrate . This indicates that the coating is rather effective in scattering the light to large angles and hence the light path in Si absorber can be effectively increased. However, it should be noted that the real scattering will be different, since the coating contact with a high-index Si layer.
Samples with different volume ratios of Si and Cu are studied to optimize the Si crystallization. Figure 6 shows the Raman spectrum of the samples. The Si thickness is fixed at 1.2μm and the ratio is varied from 120:1 to 30:1. With the Cu thickness increasing, the Raman peak first drifts from 514.74cm−1 to 518.13cm−1, and then back to 513.05cm−1. This indicates the alteration of the grain size and the crystal lattice of Si. The grain size is inversely proportional to the drift of the peak from 518 cm−1 and the Full Width at Half Maximum (FWHM) .
According to Fig. 6(a), the peak of the curve corresponding to the sample with 20nm thick Cu is centered at 518.13 cm−1, so the optimal thickness of Cu is 20nm and the optimal volume ratio of Si and Cu is 60:1. Figure 6(b) shows the Gauss fitting curves of the curve. The Raman curve can only be split by two Gauss curves centered at 518cm−1 and 510cm−1, which represent the poly-Si and microcrystalline Si (mc-Si), respectively. The crystallization ratio can be defined as ,
Fixing the volume ratio of Si and Cu at 60:1, four samples based on the Bi2O3 nano-islands have been fabricated, with different thickness of Si from 480nm to 1.2μm and the increment is 240nm, to clarify the light trapping efficiency. The reflectivity is examined by a spectrometer with an integrating sphere (AvaSpec-2048x14. Avantes company), as shown in Fig. 7(b) to Fig. 7(e). For comparison, reflectivity of a 1.2μm thick poly-Si film with smooth surface topography has also been examined as shown in Fig. 7(a), which is obtained with the same parameters as the other four samples but without the Bi2O3 nano-islands layer as shown in Fig. 1. According to Fig. 7, the reflectivity of samples based on the light-trapping structure is significantly lower than that of the one with smooth topography. Although the absorption coefficient of Si in the long wavelength region is smaller than that in the short wavelength region, the reflectivity is lower in the long wavelength region and that is accordant to the lateral size of the surface nano-islands. When the thickness of Si is 480nm, the reflectivity shows obvious interference, especially when the wavelength is over 600nm. With the Si thickness increasing, the interference becomes weaker. When the thickness reaches 1.2μm, the reflectivity between 800nm and 1100nm is below 10%, where light-trapping is most necessary, because it is close to the band gap of crystalline silicon and the absorption coefficient decrease rapidly in this region.
Figure 7(f) shows the average reflectivity (Ra) corresponding to Si thickness under AM1.5 illumination spectrum, which is defined asFig. 1, both the topography alteration of surface and the thickness alteration of the Si layer are crucial to the reflectivity. It can be seen from Fig. 7(f) that Ra is limited between 11.5% and 13% when the Si thickness is tuned from 480nm to 1.2μm, much lower than that of bare crystalline silicon which is over 35%. However, Ra does not decrease monotonically with Si thickness, and it goes up to 13% at 960nm. This agrees with the Rms alteration presented in Fig. 4. The absorption ratio of the incident light increases with the Si thickness, meanwhile, the surface topography changes with the Si thickness, which may result in a weakened light-trapping efficiency. According to Fig. 7, the increase of reflectivity from about 12% to 13% when the Si thickness increases from 480nm to 960nm indicates that the topography alteration of the surface is dominant in this process and the light-trapping efficiency is weakened. In contrast, the decrease of reflectivity from 13% to 11.5% when the Si thickness increases from 960nm to 1.2μm indicates that the thickness alteration of Si is dominant in this process and more incident light is absorbed.
We develop a simple process to fabricate a light-trapping structure by annealing Bi nano-islands in the air. Bi2O3 nano-islands with average lateral size of about 1μm and maximum altitude of over 600nm are thus obtained, which transfer a textured topography to the subsequently deposited Si film and realize a light-trapping structure with the root-mean-square roughness of over 150nm. A comparison of samples with different volume ratios of Si and Cu is studied to optimize the recrystallization of the textured a-Si. It is proved to be totally crystallized when the ratio is 60:1. Also, the relationship between the reflectivity and the wavelength is studied to reveal the light-trapping efficiency and the result shows good light-trapping efficiency in the long wavelength region. Between 800nm and 1.1μm, which is an extended absorption region from a-Si to poly-Si but with low absorption coefficient, the reflectivity is below 10% when the Si film is 1.2μm thick. Furthermore, we study the relationship between the average reflectivity under AM1.5 illumination spectrum and the Si film thickness and discover that without any extra treatment, the average reflectivity is about 12% when the Si film is only 480nm thick.
We acknowledge Yang Li for fruitful discussion and Honggang Gou for providing convenience of spectrum measurement. This work was supported by the National Natural Science Foundation of China (No. 50706022), the “863” High Technology Research and Development Program of China (No. 2007AA05Z429), the National Science and Technology Support Program of China (No. 2006BAJ01A11-1), the Key-lab Program of National Lab for Information Science and Technology of Tsinghua University of China and Open Program of State Key Lab of Integrated Optoelectronics of China (No. IOSKL-KF200915).
References and links
1. P. D. Maycock, “PV review: World Solar PV market continues explosive growth,” Refocus 6(5), 18–22 (2005). [CrossRef]
2. M. J. McCann, K. R. Catchpole, K. J. Weber, and A. W. Blakers, “A review of thin-film crystalline silicon for solar cell applications. Part 2: Foreign substrates,” Sol. Energy Mater. Sol. Cells 68, 135–171 (2001). [CrossRef]
3. R. W. Birkmire, “Compound polycrystalline solar cells: Recent progress and Y2 K perspective,” Sol. Energy Mater. Sol. Cells 65(1-4), 17–28 (2001). [CrossRef]
4. T. Ameri, G. Dennler, C. Lungenschmied, and C. J. Brabec, “Organic tandem solar cells: A review,” Energy Environ. Sci. 2(4), 347 (2009). [CrossRef]
5. B. Li, L. Wang, B. Kang, P. Wang, and Y. Qiu, “Review of recent progress in solid-state dye-sensitized solar cells,” Sol. Energy Mater. Sol. Cells 90(5), 549–573 (2006). [CrossRef]
6. A. V. Shah, H. Schade, M. Vanecek, J. Meier, E. Vallat-Sauvain, N. Wyrsch, U. Kroll, C. Droz, and J. Bailat, “Thin-film silicon solar cell technology,” Prog. Photovoltaics 12(23), 113–142 (2004). [CrossRef]
7. D. Bonnet, “Manufacturing of CSS CdTe solar cells,” Thin Solid Films 361–362, 547 (2000). [CrossRef]
8. F. Kessler and D. Rudmann, “Technological aspects of flexible CIGS solar cells and modules,” Sol. Energy 77(6), 685–695 (2004). [CrossRef]
9. F. Svelto, C. Flores, A. Caon, R. Contini, and E. Rossi, “The Italian activities on GaAs solar cells for space applications: Achieved results and future programmes,” Sol. Energy Mater. Sol. Cells 35, 99–104 (1994). [CrossRef]
10. D. L. Staebler and C. R. Wronski, “Reversible conductivity changes in discharge-produced amorphous Si,” Appl. Phys. Lett. 31(4), 292 (1977). [CrossRef]
11. J. S. Im, R. S. Sposili, and M. A. Crowder, “Single-crystal Si films for thin-film transistor devices,” Appl. Phys. Lett. 70(25), 3434 (1997). [CrossRef]
12. M. Miyao, I. Tsunoda, Y. Sadoh, and A. Kenjo, “Ion-beam stimulated solid-phase crystallization of amorphous Si on SiO2,” Thin Solid Films 383(1-2), 104–106 (2001). [CrossRef]
13. Y. Kawazu, H. Kudo, S. Onari, and T. Arai, “Low-Temperature Crystallization of Hydrogenated Amorphous Silicon Induced by Nickel Suicide Formation,” Jpn. J. Appl. Phys. 29(Part 1, No. 12), 2698–2704 (1990). [CrossRef]
14. Wenyan Zhang, Dejie Li, and Jian Wang. Phys, “Cu-induced poly-Si/SiO2/a-Si multilayer film with high reflectivity across the whole visible band,” Stat. Solidi RRL 3,82 (2009). [CrossRef]
15. G. Radnoczi, A. Robertsson, H. T. G. Hentzell, S. F. Gong, and M. A. Hasan, “Al induced crystallization of a-Si,” J. Appl. Phys. 69(9), 6394 (1991). [CrossRef]
16. S. W. Russell, J. Li, and J. W. Mayer, “In situ observation of fractal growth during a-Si crystallization in a Cu3Si matrix,” J. Appl. Phys. 70(9), 5153 (1991). [CrossRef]
17. Y. Kawazu, H. Kudo, S. Onari, and T. Arai, “Low-Temperature Crystallization of Hydrogenated Amorphous Silicon Induced by Nickel Suicide Formation,” Jpn. J. Appl. Phys. 29(Part 1 No. 4), 729–738 (1990). [CrossRef]
18. S.-B. Rim, S. Zhao, S. R. Scully, M. D. McGehee, and P. Peumans, “An effective light trapping configuration for thin-film solar cells,” Appl. Phys. Lett. 91(24), 243501 (2007). [CrossRef]
19. J. Xiao, L. Wang, X. Li, X. Pi, and D. Yang, “Reflectivity of porous-pyramids structured silicon surface,” Appl. Surf. Sci. 257(2), 472–475 (2010). [CrossRef]
20. J. Müller, B. Rech, J. Springer, and M. Vanecek, “TCO and light trapping in silicon thin film solar cells,” Sol. Energy 77(6), 917–930 (2004). [CrossRef]
21. H. Sai, H. Fujiwara, M. Kondo, and Y. Kanamori, “Enhancement of light trapping in thin-film hydrogenated microcrystalline Si solar cells using back reflectors with self-ordered dimple pattern,” Appl. Phys. Lett. 93(14), 143501 (2008). [CrossRef]
22. Y.-C. Lee, C.-F. Huang, J.-Y. Chang, and M.-L. Wu, “Enhanced light trapping based on guided mode resonance effect for thin-film silicon solar cells with two filling-factor gratings,” Opt. Express 16(11), 7969–7975 (2008). [CrossRef] [PubMed]
23. N. M. Sammes, G. A. Tompsett, H. Nafe, and F. Aldingera, “Bismuth Based Oxide Electrolytes-Structure and Ionic Conductivity,” J. Eur. Ceram. Soc. 19(10), 1801–1826 (1999). [CrossRef]
24. Ch. Ossadnik, S. Veprek, and I. Gregora, “Applicability of Raman scattering for the characterization of nanocrystalline silicon,” Thin Solid Films 337(1-2), 148–151 (1999). [CrossRef]
25. D. Das and M. Jana, “Hydrogen plasma induced microcrystallization in layer-by-layer growth scheme,” Sol. Energy Mater. Sol. Cells 81(2), 169–181 (2004). [CrossRef]