Advanced Materials Research Vol. 321 (2011) pp 240-245 Online available since 2011/Aug/16 at www.scientific.net © (2011) Trans Tech Publications, Switzerland doi:10.4028/www.scientific.net/AMR.321.240
Laser Textured Black Silicon Solar Cells with Improved Efficiencies Xxx Sedao1, a, Rémi Torres1, b, Thierry Sarnet1, c, Philippe Delaporte1, d and Marc Sentis1, e 1
LP3-CNRS, 163 Avenue de luminy - C. 917, 13288 Marseille cedex 9, France a
[email protected],
[email protected],
[email protected], d e
[email protected],
[email protected]
Keywords: Ultra-short pulse laser, black silicon, texturisation, absorption, solar cells.
Abstract. Femtosecond laser irradiation of silicon has been used for improving light absorption at its surface. In this work we demonstrate the successful implementation of femtosecond laser texturisation to enhance light absorption at Si solar cell surface. In order to adapt this technology into solar industry, the texturisation process is carried out in air ambient. The microstructure similar to what has been produced in vacuum can be made in air by using appropriate laser conditions. The texturised surface shows excellent optical properties with a reflectivity down to 7% without crystalline orientation dependence. Junction formation and metallisation proceeded after texturisation. Suns-Voc measurements are performed to evaluate the cell performance and decent electrical characteristics have been achieved. Introduction The formation of laser-induced micro-sized conical structures on silicon substrate has been observed firstly in reactive gas ambients, notably halogen gases such as SF6 or Cl2, but also possible in other gas species such as HS2 or O2 [1-7]. The improved properties of laser micro-structured silicon as compared to untreated silicon surface make it attractive to various applications in electronics and photonics [3-5]. Special interest has been shown in the unique optical properties micro-structured silicon exhibits, i.e., increased absorption throughout a wide spectral range, for possible applications in photo-detector and solar cell technology [5, 8-11]. Recent studies have shown that similar microstructures, “penguin-like” black silicon [8, 9], can be produced in vacuum using femtosecond laser. It is manifested that the laser texturisation is independent to the crystalline orientation of the substrate being irradiated - uniformly texturised surface is obtained on multi-crystalline silicon (mc-Si) substrate and identical absorption enhancement is achieved on different crystal grains. This makes laser texturisation a very attractive alternative to conventional texturisation method such as chemical etching to be used in mc-Si solar cell production, since the latter is problematic due to the anisotropic etch rates at different crystalline grains [11, 12]. However, laser texturisation has not reached the status of mass production technology for standard screen printed solar cells, most likely, because it is still a process that requires vacuum system which raises the production cost and compromises the throughput. In this study, further work is carried out to investigate femtosecond laser irradiation at silicon surface with the same manner but in an atmospheric ambient. The morphology and the optical properties of surface texturised silicon are studied. The laser process parameters which yield the best optical properties are used to texturise larger surface area on which solar cells is fabricated. Experiments Laser Texturisation. The micromachining experiments were performed using a Ti: sapphire laser (Hurricane model, Spectra-Physics) that was operated at 800 nm, with maximum pulse energy of 500 µJ, a repetition rate of 1 kHz and laser pulse duration of 100 fs. The laser beam was perpendicularly incident onto the sample surface. In order to get a reasonable uniform laser energy distribution, only All rights reserved. No part of contents of this paper may be reproduced or transmitted in any form or by any means without the written permission of TTP, www.ttp.net. (ID: 89.92.241.235, LP3-CNRS, MARSEILLE, France-15/02/14,11:24:01)
Advanced Materials Research Vol. 321
241
the centre part of the Gaussian laser beam was selected using a 2×2 mm2 square-shaped mask. A spot about 35×35 µm2 area was obtained by projecting the mask image onto the sample surface with an objective lens (f’ = 50 mm). The laser energy delivered to the sample surface was attenuated by coupling an analyser and a polariser and a set of neutral density filters. A computer-controlled XY-stage (for the sample) and Z-stage (for the objective lens) allowed precise positioning of the laser spot on the sample surface and enabled scanning for large area texturisation at a scan speed up to 3 mm/s. The number of laser shots (Nb) is the scan control parameter defined as the product of spot area and laser repetition rate divided by the product of transverse displacement and laser scan speed. The analyser rotation, the mechanic shutter placed in front of the polariser and the XYZ motion stages were controlled by a computer. The texturing results are in-situ monitored using a CCD camera. In order to control the ambient pressure, a vacuum chamber is equipped to house the sample, the objective and the translation stages. The vacuum level can be altered from atmospheric pressure down to ~10-5 mbar. The schematic of the system setup can be sought elsewhere [8, 10]. Cell Fabrication. Laser texturisation test was made on 400 µm thick silicon substrates, both multi-crystalline silicon (mc-Si) and n-doped (5-20 Ω.cm) monocrystalline silicon (c-Si). The solar cells were only produced on the c-Si. The thickness of the substrates is not optimised for solar cells but it was the only material we had to start with. The surface area of the solar cell measures 1.5×1.5 cm2. The surfaces of the silicon were first phosphorus diffused from a POCl3 source in order to create a n+ layer which helps the formation of a back ohmic contact, while the n+ front layer was chemically removed before the laser treatment. After the laser process, the sample was boron implanted by Plasma Immersion and treated by a rapid thermal annealing step. The junction depth made using this technique is about 150 nm. After p-n junction formation, the electrical contacts were formed by physical vapour deposition. Both surfaces remained un-passivated. The details about cell processing roadmap can be found in [8]. Characterisation. The surface morphology was studied using an optical microscope and a scanning electron microscope (SEM). For getting detailed information about the microstructure, the conical features developed after laser irradiation are sliced off from the silicon substrate on which they grew and studied using a high resolution transmission electron microscope (HRTEM). In order to evaluate the optical properties of texturised surface, an integrating sphere was used to measure the total reflectance, with a spectrophotometer operating between 340 and 1940 nm. The electrical performance of the solar cells was evaluated under illuminated condition using a Sinton Suns-Voc tester. Results and Discussions Surface Microstructure. Some typical surface morphological features appear after laser irradiation in air ambient can be seen from the SEM images presented in Fig. 1. The process parameter Nb was 500 shots and laser fluence was varied between 0.5 and 1.2 J/cm2. Micrometer-sized conical features are formed in the laser irradiated regions. Increase of the feature dimensions while decrease of feature density is observed with increased laser fluence. Similar process is carried out in vacuum as well. Despite the difference in quantitative characteristics, increased height with increased quantity of the number of laser shots or laser fluence and decreased density with increased laser fluence, appear as common trends, independent from the ambient pressure. In all the cases of air ambient processing, large amount of nano-scale particles can be seen across the entire micro-conical features and the amount of the particles is larger when compared to the features made under vacuum. These nano-particles are re-deposited material from the ejected plume. In vacuum, the low background pressure significantly reduces the re-deposition while in air ambient this re-deposition is inevitable due to the lack of extraction drive. This re-deposition does not cause any problems during micro-structuring process until laser fluence reaches sufficient high such that the re-deposition tends to clog the space between individual cones. Unlike laser texturisation in vacuum,
242
Advanced Research on Intelligent Materials and Mechanical Engineering
as the laser fluence goes beyond 1.2 J/cm2 the conical features start to be merged among the re-deposits. Similar tendency too holds for increasing the number of laser shots: at any given laser fluence, the accumulation of re-deposits tends to submerge the conical features with increasing Nb. Femtosecond laser irradiation introduces surface disorder [13, 14]. Our TEM investigation reveals that a thin layer in the order of 100 nm consisting of disordered crystalline, poly-crystalline and some times amorphous silicon has formed at the outermost region of the conical features. It may be concluded that the formation of this layer is due to fast re-solidification of silicon melt. The discussion on the microstructure evolution during laser texturisation is not in the scope of this article. The details of micro-structural study will be given in a separate report. (a)
(b)
(c)
(d)
Fig. 1. SEM photos of surfaces texturised in atmospheric ambient. Nb = 500, laser fluence: (a) 0.5 J/cm2, (b) 0.7 J/cm2, (c) 0.9 J/cm2, and (d) 1.2 J/cm2. Optical Properties. Fig. 2 shows the measured reflectance for an unstructured silicon surface and laser texturised surfaces. Micro-structured silicon has a drastically decreased reflectance over the entire spectrum. The reduced reflection in the visible and near IR spectral range can be interpreted as a result of irradiation trapping between the inter-cone areas, which leads to multiple reflections and thus enhanced light collection. The lowered reflectance results in a featureless, near-unity absorption in the entire wavelength region 400-800 nm and also reduced reflectivity at region 800-1200 nm compared to the reference. Over the whole spectrum, different reflectance is obtained by varying laser fluence, with lowest reflectance achieved at the lowest laser fluence and the reflectance slowly increases with laser fluence. The only observation contrary to this trend is seen between the two high fluence settings at spectral region 400-800 nm, where the reflectance of the surface irradiated at 0.9 J/cm2 is higher than that at 1.2 J/cm2. Similar reduction of surface reflectance is identified between texturisation in air ambient and in vacuum, although more careful comparison reveals that a different tendency of reflectivity change with varying laser fluence exists [8-10], which could possibly be attributed to the geometrical difference of the features made in different ambients: the micro-cones made in air are more pillar-shaped compared to those made in vacuum. This assumption is being examined using optical modeling software FDTD Lumerical, the results from which will be reported in a separate paper.
Advanced Materials Research Vol. 321
243
It can be seen from Fig. 2 that in the atmospheric ambient, the surface texturised using laser fluence 0.5 J/cm2 has the lowest reflectance. This parameter set together with the second best set 0.7 J/cm2 was applied to texturise larger surface areas for the fabrication of solar cells.
Fig. 2. Reflectance of laser texturised silicon surfaces. The black-coloured curve represents of the reference, the reflectance at a polished silicon surface. Electrical Performance. The electrical performance is listed in Table 1. The key factors derived from Suns-Voc measurements are open circuit voltage Voc, pseudo fill factor pFF and pseudo efficiency pEff. The efficiencies between 6 and 7% from these cells are, to the authors’ knowledge, the best record ever made out of black silicon cells. The cell texturised with laser fluence 0.5 J/cm2 has a Voc of 0.27 Volts, while the one texturised using 0.7 J/cm2 has a higher Voc of 0.35 Volts. Compared to that of 0.5 J/cm2, the pFF of the cell made using 0.7 J/cm2 is mediocre. However, it should be noted that the low shunt resistance Rsh of this cell (see in Table 1) indicates the existence of a shunting problem. This shunt might have been made during p-n junction formation step: a direct connection between n+ and p+ layers introduced in the course of plasma emersion doping. Without the shunting problem, further improvement in Voc, pFF and pEff can be expected. Additional treatments to eliminate the shunting are planned in order to confirm this. Table 1. Electrical characteristics from Sinton Suns-Voc measurements Laser fluence [mJ/cm2] Specs 500 700 Voc [mV] 271 345 pFF [%] 67 58 pEff [%] 6.3 7.0 2 Rsh [Ωcm ] 150000 191 Considering we are new to device fabrication, there were several aspects involved in solar cell production that we are not very skilled in, these preliminary results are quite positive. A few issues associated with cell fabrication will be addressed in the immediate future, such as optimisation of metal contact and surface passivation, which will reduce serial resistance, increase blue response, short circuit current as well as open circuit voltage. These will considerably improve cell fill factor and cell efficiency. Aside from these engineering issues, there is room to improve in the technology roadmap of laser texturisation for solar cells. The micro-structuring process leaves the surface layer highly disordered [5]. As mentioned in Surface Microstructure section, the surface layer of the conical features is re-solidified poly-crystalline silicon or crystalline silicon with stress. This deformation of the lattice undoubtedly promotes high rates of free carrier recombination. On the other hand, since the p+ layer boron implementation was 150 nm deep, it seems that the p-n junction was formed very close to or probably inside the deformation region. This is undesirable since high density of recombination
244
Advanced Research on Intelligent Materials and Mechanical Engineering
sites close to the p-n junction will prevent photo-generated electron-hole pairs to be collected and extracted. In the present study, it is highly likely that the local disorder of the crystalline structure compromises or even overweighs the advantages gained by enhanced surface absorption. Nonetheless, to overcome the problem, a few approaches have been proposed and will be tested: a wet etch to remove the defect layer is the most common means, either acidic or alkaline solution etch can be applied [11]. Contrary to the concept of removing the damage layer, laser annealing might be applied to heal the defects [15]. It is a more interesting option compared to the etch method since it is a dry process and it is relatively easy to be integrated with laser texturisation process. The third option is to define the junction away from the damage layer, which can be realised by the combination of minimising the thickness of laser defect layer and formation of a deep p-n junction. Summary In this report we explore the possibility of femtosecond laser texturisation of silicon surface in atmosphere ambient. The influence of laser process parameters on surface morphology and reflectance is also studied. The results show that although the process window is smaller compared to laser texturing in vacuum, the texturisation can be realised in atmospheric ambient. Laser texturisation at lower fluence results in higher density and smaller size of the conical features, which leads to a lower reflectance over the entire spectrum 350-1200 nm. The fabrication of solar cell on the texturised surface is demonstrated. The promising electrical performance with up to 7% cell efficiency encourages further efforts to be made towards developing black silicon solar cells. Acknowledgement The work was supported by the European Union within the project “Solasys - Next Generation Solar Cell and Module Laser Processing Systems”, EU Grant No. 219050. The authors acknowledge Dr. M. Schulz-Ruhtenberg for his help with Suns-Voc measurements at Fraunhofer ILT and Dr. M. Halbwax for his help on the reflectivity measurements at IEMN Laboratoire Central. References [1] A.J. Pedraza, J.D. Fowlkes, and D.H. Lowndes: Appli. Phys. Lett., Vol. 74, 1999, p. 2322. [2] C.H. Crouch, J.E. Carey, M.Y. Shen, E. Mazur, and F.Y. Genin: Appl. Phys. A., Vol. 79, 2004, p. 1635. [3] V. Zorba, P. Tzanetakis, C. Fotakis, E. Spanakis, E. Stratakis, D.G. Papazoglou, and I. Zergioti: Appl. Phys. Lett., Vol. 88, 2006, p. 081103. [4] V. Zorba, I. Alexandrou, I. Zergioti, A. Manousaki, C. Ducati, A. Neumeister, C. Fotakis, and G.A.J. Amaratunga: Thin Solid Films, Vol. 453-454, 2004, p. 492. [5] V. Zorba, N. Boukos, I. Zergioti, and C. Fotakis: Appl. Optics, Vol. 47, 2008, p. 1846. [6] C. Wu, C.H. Crouch, L. Zhao, J.E. Carey, R. Younkin, J.A. Levinson, E Mazur, R.M. Farreli, P. Gothoskar, and A. Karger: Appl. Phys. Lett., Vol. 78, 2001, p. 1850. [7] C.H. Crouch, J.E. Carey, J.M. Warrender, M.J. Aziz, E. Mazur, and F.Y. Genin: Appl. Phys. Lett., Vol. 84, 2004, p. 1850. [8] M. Halbwax, T. Sarnet, Ph. Delaporte, M. Sentis, H. Etienne, V. Vervisch, I. Perichaud, and S. Martinuzzi: Thin Solid Films, Vol. 516, 2008, p. 6791. [9] T. Sarnet, R. Torres, V. Vervisch, Ph. Delaporte, M. Sentis, M. Halbwax, J. Ferreira, D. Barakel, M. Pasquinelli, S. Martinuzzi, L. Escoubas, F. Torregrosa, H. Etienne, and L. Roux: 2008, Proceedings of International Congress of Applications of Laser & Electro-Optics, M.401. [10] R. Torres, V. Vervisch, M. Halbwax, T. Sarnet, Ph. Delaporte, M. Sentis, J. Ferreira, D. Barakel, S. Bastide, F. Torregrosa, H. Etienne, and L. Roux: J. Optoelectronics and Advanced Materials, Vol. 12, 2010, p. 621.
Advanced Materials Research Vol. 321
245
[11] A. Poroykov, G. Untila, T. Kost, A. Chebotareva, and M. Timofeyev: 2010, Proceedings of the 25th EU PVSEC. [12] S.D. Iuliis, G. Agarwal, M. Izzi, E. Salza, L. Serenelli, and M. Tucci: 2010, Proceedings of the 25th EU PVSEC. [13] T. Tomita, R. Kumai, S. Matsuo, S. Hashimoto, and M. Yamaguchi: Appl. Phys. A., Vol. 97, 2009, p. 271. [14] M. Schade, O. Varlamova, J. Reif, H. Blumtritt, W. Erfurth, and H.S. Leipner: Anal. Bioanal. Chem., Vol. 396, 2010, p. 1905. [15] V. Vervisch, Y. Larmande, Ph. Delaporte, T. Sarnet, M. Sentis, H. Etienne, F. Torregrosa, F. Cristiano and P.F. Fazzini: Appl. Surf. Sci., Vol. 255, 2009, p. 5647.
Advanced Research on Intelligent Materials and Mechanical Engineering 10.4028/www.scientific.net/AMR.321
Laser Textured Black Silicon Solar Cells with Improved Efficiencies 10.4028/www.scientific.net/AMR.321.240 DOI References [1] A.J. Pedraza, J.D. Fowlkes, and D.H. Lowndes: Appli. Phys. Lett., Vol. 74, 1999, p.2322. doi:10.1007/s003390051517 [2] C.H. Crouch, J.E. Carey, M.Y. Shen, E. Mazur, and F.Y. Genin: Appl. Phys. A., Vol. 79, 2004, p.1635. doi:10.1007/s00339-004-2676-0 [3] V. Zorba, P. Tzanetakis, C. Fotakis, E. Spanakis, E. Stratakis, D.G. Papazoglou, and I. Zergioti: Appl. Phys. Lett., Vol. 88, 2006, p.081103. doi:10.1063/1.2177653 [4] V. Zorba, I. Alexandrou, I. Zergioti, A. Manousaki, C. Ducati, A. Neumeister, C. Fotakis, and G.A.J. Amaratunga: Thin Solid Films, Vol. 453-454, 2004, p.492. doi:10.1016/j.tsf.2003.11.144 [5] V. Zorba, N. Boukos, I. Zergioti, and C. Fotakis: Appl. Optics, Vol. 47, 2008, p.1846. doi:10.1364/AO.47.001846 [6] C. Wu, C.H. Crouch, L. Zhao, J.E. Carey, R. Younkin, J.A. Levinson, E Mazur, R.M. Farreli, P. Gothoskar, and A. Karger: Appl. Phys. Lett., Vol. 78, 2001, p.1850. doi:10.1063/1.1358846 [7] C.H. Crouch, J.E. Carey, J.M. Warrender, M.J. Aziz, E. Mazur, and F.Y. Genin: Appl. Phys. Lett., Vol. 84, 2004, p.1850. doi:10.1007/s00339-004-2676-0 [8] M. Halbwax, T. Sarnet, Ph. Delaporte, M. Sentis, H. Etienne, V. Vervisch, I. Perichaud, and S. Martinuzzi: Thin Solid Films, Vol. 516, 2008, p.6791. doi:10.1016/j.tsf.2007.12.117 [9] T. Sarnet, R. Torres, V. Vervisch, Ph. Delaporte, M. Sentis, M. Halbwax, J. Ferreira, D. Barakel, M. Pasquinelli, S. Martinuzzi, L. Escoubas, F. Torregrosa, H. Etienne, and L. Roux: 2008, Proceedings of International Congress of Applications of Laser & Electro-Optics, M. 401. doi:10.1117/12.768516 [13] T. Tomita, R. Kumai, S. Matsuo, S. Hashimoto, and M. Yamaguchi: Appl. Phys. A., Vol. 97, 2009, p.271. doi:10.1007/s00339-009-5364-2 [14] M. Schade, O. Varlamova, J. Reif, H. Blumtritt, W. Erfurth, and H.S. Leipner: Anal. Bioanal. Chem., Vol. 396, 2010, p. (1905). doi:10.1007/s00216-009-3342-3 [15] V. Vervisch, Y. Larmande, Ph. Delaporte, T. Sarnet, M. Sentis, H. Etienne, F. Torregrosa, F. Cristiano and P.F. Fazzini: Appl. Surf. Sci., Vol. 255, 2009, p.5647. doi:10.1016/j.apsusc.2008.11.010
