激光埋栅高效太阳能电池

Presented at the 21 stEuropean Photovoltaic Solar Energy Conference, September4-8, Dresden (2CV.3.17) HIGH-EFFICIENCY SOLAR CELLS WTH LASER-GROOVED BURIED CONTACT FRONT AND LASER-FIRED REAR FOR INDUSTRIAL PRODUCTION O. Schultz1, S.W. Glunz 1, W. Warta 1, R. Preu1, A. Grohe 1, M. K?ber 1, G.P. Willeke 1R.Russel2, J. Fernandez2, C. Morilla 2, R. Bueno2, I. Vincuer ía2 1Fraunhofer ISE, Heidenhofstrasse 2, 79110 Freiburg, Germany 2BP Solar Espa?a, Poligono Industrial Tres Cantos, s/n Zona Oeste, 28760 Tres Cantos , (Madrid) Spain Phone ++761-4588-5355; Fax ++761-4588-9250, email: Oliver.Schultz@ise.fraunhofer.de ABSTRACT: The laser-grooved buried contact (LGBC) cell produced by BP Solar combines the advantages of a selective emitter and a fine line grid structure on the front. A solar cell with such a front structure is well capable of 20% efficiency on monocrystalline silicon. The current rear metallization in the production process is a full area aluminium back surface field. Although of high quality, the internal reflectance and rear surface passivation can be improved to increase cell efficiency. An industrially viable process for the rear surface of high efficiency cells is the laser-fired rear contact process (LFC) in which aluminium is fired through a dielectric passivation layer. The two laser technologies LGBC and LFC are combined into a single cell structure and applied to monocrystalline silicon. Solar cells with 19.8% efficiency were achieved on 108 cm 2 large FZ silicon of 0.5 ? cm base resistivity. Keywords: High-Efficiency, c-Si, Laser Processing 1 INTRODUCTION To enable a further and sustainable growth of photovoltaic industry a further cost reduction is an important issue. One very effective way to achieve this goal is a higher electrical energy conversion efficiency of the produced solar cells at competitive costs. Therefore many concepts for solar cells of high efficiencies have been developed on lab scale in the last decade. Three products have made it into commercial production, Sunpower′ s Back Junction Cell [1], Sanyo ′ s HIT-Cell [2] and BP Solar ’ s Laser-Grooved Buried Contact cell [3]. The Laser-Grooved Buried Contact (LGBC) cell, which is in production at BP Solar at the multi-Megawatt scale, consists of laser grooves on the front which are highly phosphorus-doped and filled with metal in a plating process. Due to the incorporation of a selective emitter, which reduces front contact resistance, and the fine line grid structure, which reduces shading losses to a minimum, this front has all features of a high-efficiency solar cell structure. The rear surface currently consists of a full-area aluminium back surface field. Especially when thin wafers are considered, the optical properties of the rear (internal reflection for light-trapping) and the rear surface passivation become increasingly important. Although of high quality, these parameters of the aluminium BSF limit the performance of the standard cell and have to be improved in order to increase the efficiency of the cells. One solution to achieve an excellent passivation simultaneously with a high internal reflection is the application of a dielectric layer on the rear, e.g. silicon oxide or silicon nitride, in combination with evaporated aluminium. The contacts are locally fired through the aluminium and the passivation layer with the LFC process [4]. In this paper we report on the results of the combination of the LGBC front and the LFC rear structure and first steps to move the developed cell structure to BP Solar manufacturing equipment. 2 EXPERIMENTAL 2.1 Cell structure For a proof of principle 330 μ m thick p-type FZ material is used in a first step to evaluate the potential of the cell structure. The front surface is covered with a thick thermal oxide in order to realize optimum surface passivation of the lowly doped phosphorus emitter. The rear surface is also passivated by a thick thermal oxide in order to achieve optimum conditions like they are standard in high-efficiency lab scale solar cells. For the industrial approach boron-doped magnetically confined Czochralsky silicon (MCz) of reduced oxygen content is used and the thick silicon oxide layer is replaced by a stack system of equal optical and nearly identical electrical properties [5]. A schematic illustration of the cell structure is shown in Figure 1. Figure 1 : Cell structure which combines the LGBC front with the LFC rear. The cell is fabricated by random pyramid texturing of the (100) crystal orientated wafer surface. After phosphorus diffusion of a lightly doped emitter a LPCVD silicon nitride antireflection-coating is deposited. Then laser grooves of 20 μ m width and 30 μ m depth are cut into the front silicon surface. Subsequent etching in hot NaOH removes residual silicon and surface crystal damage. The grooves are then doped to a sheet resistivity of less than 10 ? /sq using a POCl 3 source. The rear Presented at the 21 stEuropean Photovoltaic Solar Energy Conference, September4-8, Dresden (2CV.3.17) contact comprises a thin silicon oxide film and PECVD-SiN x layers underneath a 2 μ m thick evaporated aluminium layer. A laser is used to fire small (100 μ m diameter) areas of the aluminium layer through the oxide film to form a local back contact. The front metallization for both cell structures used in this study consisted of the standard industrial plating process and was performed at BP Solar. 2.2 Simulation A one-dimensional solar cell simulation was performed with PC1D [6]. Two scenarios as internal reflector on the rear were calculated: a) alloyed aluminium BSF b) evaporated aluminium on top of a layer of SiO 2. Internal reflection was assumed to be Rback= 69% (diffuse) for the alloyed case and 96% (specular) for the evaporated rear. Wafer thickness and rear surface recombination velocity were varied according to the technological limits of alloyed aluminium BSF with respect to the optical and electrical properties (compare [7]). Alloyed aluminium BSF is expected to have recombination velocities S back in excess of 1000 cm/s, dielectrically passivated rears have values below 500 cm/s. The results for high quality boron-doped MCz or FZ silicon with minority carrier lifetimes of τ bulk=1000 μ s are shown in Figure 2. 0 1000 2000 3000 4000 500016.517.017.518.018.519.019.520.020.521.0 back reflector SiO2 Al thickness 300 μ m 200 μ m 150 μ m 100 μ m FZ/MCz-material 1.5 ? cm, high lifetimeEfficiency[%]Sback [cm/s]Figure 2 : Simulation of solar cell efficiency for high-quality MCz or FZ silicon. As internal reflectors an alloyed aluminium BSF or evaporated aluminium on top of SiO 2 are considered. Higher efficiencies are obtained on 300 μ m than on 100 μ m thick wafers for the whole parameter range. In order to reach efficiencies in excess of 20%, S back values below 500 cm/s are needed, technologically this is realized with dielectric passivation and evaporated aluminium plus the LFC process. In this regime of the high-efficiency cell structure the difference between the wafer thicknesses becomes marginal, i.e. a thickness reduction can be tolerated without large efficiency loss. The situation is different when boron-doped Cz silicon with a typical minority carrier lifetime of τ bulk=14 μ s is used (see Figure 3). This lifetime was measured for industrial Cz material with oxygen after carrier-induced boron-oxygen related degradation. For high rear surface recombination velocities the 300 μ m thick cell still yields a higher efficiency than the 100 μ m thick one. Improved rear passivation of the order of S back≈ 1000 cm/s results in comparable performance regardless of cell thickness. For the case of effective rear dielectric passivation 0 1000 2000 3000 4000 500016.517.017.518.018.519.019.520.020.521.0back reflector SiO 2 Al thickness 300 μ m 200 μ m 150 μ m 100 μ m Cz-material 1.5 ? cm, degradedEfficiency[%]Sback [cm/s]Figure 3 : Simulation of solar cell efficiency for standard boron-doped Cz silicon. As internal reflectors an alloyed aluminium BSF or evaporated aluminium on top of SiO 2are considered. (Sback≤ 500 cm/s) with good light-trapping a substantial gain of ?η ≈ 0.7% absolute is calculated. According to the simulation efficiencies in excess of 20% cannot be expected for the low-lifetime oxygen-rich Cz silicon of 1.5 ? cm. However, an increase of efficiency can be achieved by decreasing wafer thickness when the appropiate cell structure is applied. This supports the trend towards thinner wafers originally pushed from cost-saving needs and silicon shortage. 3 RESULTS AND DISCUSSION 3.1 Laboratory process With the reference process with full oxide surfaces solar cells were manufactured on 330 μ m thick boron-doped FZ silicon of 0.5 ? cm base resistivity. Efficiencies of the two best cells for two different cell areas, respectively, are listed in Table I. Table I : I-V parameters of LGBC/LFC cell on 330 μ m thick FZ cell. Cell ID Area [cm2] Voc [mV] j sc[mA/cm 2] FF [%] η [%]12-3 129 664 38.4 76.4 19.5 12-4 129 656 37.9 78.1 19.4 12-5 108 658 38.3 76.2 19.2 12-6 108 662 38.4 77.7 19.8 Excellent open-circuit voltages and high j sc values were achieved. The fill factor has not yet reached its full potential of FF ≥ 80% as it is proven for the LGBC process with aluminum BSF [8] as well as for the LFC rear with screen-printed front contacts [9]. Yet, efficiencies up to 19.8% were achieved on large cell area. This is a significant improvement compared to best cell performance achieved with the production compatible Al-BSF which yields 18.3% on Cz silicon material [8]. The reason can be deduced from the quantum efficiency and reflectance measurements displayed in Figure 4. Besides the higher material quality (FZ instead of Cz) the rear surface is well passivated and carrier collection is significantly improved especially for long-wavelength photons. The gain is enhanced by the mirror-like rear which leads to an excellent light trapping. Applying this Presented at the 21 stEuropean Photovoltaic Solar Energy Conference, September4-8, Dresden (2CV.3.17) cell structure to even thinner wafers of 140 μ m with adjusted metallization technology already resulted in efficiencies of 20.1% on large wafer area [10]. 300 400 500 600 700 800 900 1000 1100 12000.00.10.20.30.40.50.60.70.80.91.0LGBCfront / LFC rear (19.8%)LGBC front / aluminium BSF rear (18.3%)Reflectance,ExternalQEWavelength [nm]Figure 4 : Quantum efficiency and reflectance measurements taken on cell structures with LGBC front and LFC or aluminium BSF rear. 3.2 Production process To test the process on material closer to industrial standards, solar cells were made from MCz of 1 ? cm base resistivity and 250 μ m thickness. All front surface processing was performed in the production line of BP Solar: The front surface is covered with a LPCVD-SiN x(instead of 105 nm oxide like in the all-oxide approach). On the rear a stack system of dielectric layers [5] was used instead of the thick thermal oxide. The electrical characteristics of a cell just before plating of the front grooves can be seen in Figure 5. All other processing steps including rear surface metallization with LFC were completed. 0 100 200 300 400 500 600 700010203040Current[mA/cm2]SI-34.refV oc,measured = 656 mVJ sc,assumed = 38 mA/cm 2FF pseudo = 0.826 Voltage [mV]Figure 5 : Suns-V OC measurement of a LGBC/LFC on MCz of 250 μ m thickness. The solid line represents a 1-diode model fit. All process steps except front surface plating (last process step) were already executed. 656 mV of V OC were measured and a pseudo-fill factor 82.6 % is calculated. High V OC values of 656 mV were measured. Further analysis leads to the conclusion that the bulk minority carrier lifetime is very high after all high-temperature processes (τ bulk ≥ 500 μ s). In combination with the well passivated rear (S rear ≤ 300 cm/s, compatible with the measurement) this cell structure has the potential to achieve high efficiencies. The high pseudo-fill factor of FFpseudo≥ 82% shows that the diode quality is excellent. Similar results to the all-oxide approach are therefore expected for completed solar cells after front surface metallization. In order to demonstrate the production viability of the jointly developed process sequence in the production line of the industrial partner, the LFC contacting scheme was performed at the BP Solar facilities. The results for standard high-efficiency cells processed at Fraunhofer ISE and LFC contacted at BP Solar are given in Table 2. Equally high values for all parameters were achieved with the BP Solar LFC process as for the Fraunhofer ISE process. This holds for both, the thick thermal oxide as well as the stack system of dielectric layers. Table II : I-V parameters of best LFC cell with BP Solar LFC process. The cells were made on 250 μ m thick FZ silicon, cell area is 4 cm 2 aperture area. Cell-ID rear Voc [mV] jsc[mA/cm 2] FF [%] η [%]nrp27-22.1 thick oxide 668 39.4 80.0 21.0 nrp32-11.4 stack 673 38.7 78.9 20.6 4 SUMMARY The LGBC front and a dielectrically passivated rear with LFC are two proven processes suitable for thin wafers to produce a high-efficiency solar cell structure. Their combination resulted in 19.8% efficiency cells on FZ silicon wafers with thick thermal oxides for surface passivation. For the industrial process the front surface and metallization processes were performed in the fabrication line of BP Solar. MCz silicon was used and the rear surface was passivated with a stack system of dielectric layers. 20% efficient solar cells with this cell structure are within reach. ACKNOWLEDGEMENT The authors would like to thank all members of the Fraunhofer ISE Solar Cell Department for their contributions to this work. REFERENCES [1] W. P. Mulligan, D. H. Rose, M. J. Cudzinovic et al., Proceedings of the 19th European Photovoltaic Solar Energy Conference, Paris, France (2004) 387. [2] M. Tanaka, S. Okamoto, S. Tsuge et al., Proceedings of the 3rd World Conference on Photovoltaic Energy Conversion, Osaka, Japan (2003) 955. [3] N. B. Mason, T. M. Bruton, and M. A. Balbuena, PV in Europe - Prom PV Technology to Energy Solutions, Rome, Italy (2002) 227. [4] E. Schneiderl?chner, R. Preu, R. L ü demann et al., Progr. Photovolt. 10 (2002) 29. [5] O. Schultz, M. Hofmann, S. W. Glunz et al., Proceedings of the 31st IEEE Photovoltaic Presented at the 21 stEuropean Photovoltaic Solar Energy Conference, September4-8, Dresden (2CV.3.17) Specialists Conference, Orlando, Florida, USA (2005) 872. [6] D. A. Clugston and P. A. Basore, Proceedings of the 26th IEEE Photovoltaic Specialists Conference, Anaheim, California, USA (1997) 207. [7] M. Hermle, E. Schneiderl?chner, G. Grupp et al., Proceedings of the 20th European Photovoltaic Solar Energy Conference, Barcelona, Spain (2005) 810. [8] N. B. Mason, T. M. Bruton, S. Gledhill et al., Proceedings of t