Laser Edge Isolation for High-efficiency Crystalline Silicon Solar Cells

Transcription

1 Journal of the Korean Physical Society, Vol. 55, No. 1, July 2009, pp Laser Edge Isolation for High-efficiency Crystalline Silicon Solar Cells Dohyeon Kyeong, Muniappan Gunasekaran, Kyunghae Kim, Heejae Kim, Taeyoung Kwon, Inyong Moon, Youngkuk Kim and Kyumin Han School of Information and Communication Engineering, Sungkyunkwan University, Suwon Junsin Yi School of Information and Communication Engineering & Department of Energy Science, Sungkyunkwan University, Suwon (Received 26 August 2008) Edge isolation is an important step in industrial type solar cell processes. The POCl 3 emitter diffusion technique makes thin n-doped layers on all of the surfaces of the device, including the edges and the rear surface. The front and the back contacts are connected to this n-layer around the edge, which results in shunts. In this research, we used laser edge isolation to remove these shunts, thus obtaining results on the solar cell parameters. Usually, edge isolation is carried out at the end of the solar cell process; however, we altered the process steps so that the edge isolation was carried out before the SiN x deposition. This passivates the laser-induced damage. Various laser edge isolation conditions were studied by comparing the final solar cell efficiencies. From our results, we confirmed that laser edge isolation prior to SiN x deposition is good. PACS numbers: J, D, Cf, Rv, W Keywords: Edge isolation, Solar cell, Passivation, SiN x I. INTRODUCTION Currently, lasers are used to directly machine microstructures and surface patterns onto a wide range of materials, including plastics, metals, and semiconductors. Laser material processing provides many advantages over traditional methods, and the applications are innumerable. In addition, there are many cases in which no other tool can compete with the speed and the precision of the laser [1, 2]. Lasers play a very important role in the manufacture of both thin-film and thicker silicon-based photovoltaic cells. During solar-cell production, p-doped wafers are coated with an outer layer of n-doped silicon to form a large area p-n junction, which generates electrical power [3, 4]. Due to the inter connection between the points of contact in the front and the back contacts through the n-layer, an edge shunt occurs. This edge shunts accounts for 80% of the total loss mechanism, and more specifically the shunt resistance (R sh ) represents a high-conductivity path running parallel across either the p-n junction or the cell edges [5]. The parasitic emitter diffusion that wraps around the wafer edges simply resolved through edge isolation. Edge isolation is nothing more than a continuous groove yi@yurim.skku.ac.kr; Tel: ; Fax: scribed completely through the n-type layer to achieve electrical isolation [6, 7]. This groove should be narrow and as close to the edge as possible in order to maximize the cell s active area. There are different methods for performing edge isolation on silicon solar cells, including plasma barrel etching, dry etching, and dispensing of etching paste, among others. In industry, plasma etching of wafer stacks is very common, but this step is not inline capable and requires the use of undesired chemicals [8, 9]. In-line wet etching is another method that also uses chemicals and too many cleaning processes. Therefore, there is a demand for other options that can be performed inline and that are easy to implement and cheap to realize. Among the aforementioned edge isolation methods, the most successful one is a laser edge isolation process. The laser edge isolation employed in industry requires a significant proportion of very high value capital and equipment in order to produce these photovoltaic cells in a cost effective manner. In this study, we employed a 20-watt fiber laser with a wavelength of 1064 nm to perform edge isolation on an industrial silicon solar cell process. The laser parameters, such as power, frequency, and scanning speed, were adjusted to optimize the laser edge isolation. The edge isolation was carried out on both the front and the back sides. In addition, the edge isolation was carried out before and after SiN x deposition so as to identify the better passivation effect in order to improve the solar-cell parameters.

2 Laser Edge Isolation for High-efficiency Crystalline Silicon Dohyeon Kyeong et al Fig. 1. Processing steps for solar cell fabrication: (a) edge isolation as a final step, and (b) edge isolation before SiN x deposition. The shunted areas were detected with the help of liquidcrystal thermal foils. The effects of laser edge isolation were identified by analyzing the final solar efficiency. Fig. 2. Schematic of the various edge isolation processes: (a) solar cell without edge isolation, (b) front edge isolation, (c) back edge isolation, and (d) front edge isolation with SiN x layer. II. EXPERIMENT 1. Solar Cell Fabrication We followed conventional process steps for the fabrication of single-crystalline Si solar cells. These steps included POCl 3 diffusion, plasma-enhanced chemical vapor deposition (PECVD) for silicon nitride (SiN x ) deposition as an anti-reflection coating (ARC), and screenprinted metallization. All of the textured p-type silicon wafers (125 mm 125 mm) were diffused by a pentavalent impurity (phosphorus) in an open-tube furnace by using a conventional POCl 3 diffusion source at 850 C for 7 min for the pre-deposition, followed by 16 min of drive-in. The sheet resistance of the n + emitter layer was Ω/. An 80-nm SiN x layer was deposited on the front side by using PECVD at 450 C as the AR Coating. The refractive index of the SiN x film was maintained at 2.0. The front and back metallizations were carried out by using a screen printing technique with standard silver and aluminum pastes. This was followed by baking (150 C) and co-firing (920 C) the material in a conveyer belt furnace (Seirratherm). Finally, the laser edge isolation was performed with a fiber laser. The process sequence for the solar cell fabrication is illustrated in Fig Laser Edge Isolation A fiber laser with a wavelength of 1064 nm was used to scribe a groove of 0.2 mm inside the perimeter all around the wafer on either the front or the back. Our

3 -126- Journal of the Korean Physical Society, Vol. 55, No. 1, July 2009 Fig. 4. Photograph of an edge isolated solar cell. Fig. 5. Thermal image of an edge shunt (5inch CZ wafer): (a) without edge isolation and (b) with laser edge isolation. Fig. 3. SEM images (500X) of scribe lines with different laser powers: (a) 20%, (b) 40%, (c) 60%, and (d) 80%. goal was to investigate whether it were feasible to electrically isolate the p/n junction by using laser grooving and to improve the edge isolation process for better efficiency. The wafer was mounted on a vacuum chuck attached to a smooth, moveable stage. The laser beam was focused onto the wafer through the use of a fused silica lens. During the focus mode, the beam remained stationary while the stages moved the wafer in a direction perpendicular to the laser beam. The wafers were scribed with different laser powers at a stage speed of 100 mm/s; then, the wafers were tested using a shunt meter for the shunts. A schematic of the present edge isolation is shown in Fig. 2. We studied the effects of front and back edge isolation on the cell parameters. Because we used a laser for the edge isolation, there would be a heat-induced defect at the scribe line; therefore, we passivated the defects using SiN x with a small change in the process sequence (Fig. 1(b)). Edge isolation lines with different laser parameters were characterized using a scanning electron microscope (SEM). The edge shunts were identified by using a liquid crystal thermal foil (Shunt meter). The fabricated cells with laser edge isolation were characterized using illuminated I-V measurements under AM 1.5 conditions. III. RESULT AND DISCUSSION 1. Laser Isolation Laser scribing was carried out using a fiber laser with optimized laser parameters for isolating the n-layer. The edge isolation were shown to isolate the active area of the device from the enhanced recombination or shunting that occurs at the wafer edge. Laser power percentages, such as 20%, 40%, 60% and 80%, were used from the 100% total power of the laser (100% maximum laser output power were 20 W), with a repetition frequency of 10 khz at a fixed scribing rate of 100 mm/s. We used a focused laser beam for the entire edge isolation process to minimize the scribing area. Scanner mode edge isolation utilizes a large area for scribing; this leads to a reduction in the active cell area. The SEM images of the various scribing lines with different powers are pictured in Fig. 3. Fig. 3(a) shows the line scribed using 20% laser power; at this laser condition, the n-layer was not completely isolated. We also observed that the scribed line were not continuous. The 60% and the 80% scribe lines, shown in Figs. 3(c) and 3(d) are good enough to isolate the n-layer, but the heataffected zone is high, and the damage induced due to overheating is great. The line scribed using a 60% laser power is good enough in all aspects; minimum scribe width (30 µm), less induced damage, and optimum scribe depth. Fig. 4 shows the edge isolated scribe line made on the front side at a distance of 0.2 mm from the edge of the cell. 2. Shunt Resistance The fabricated solar cells were completely mapped for the shunt distribution analysis. The shunt resistance (R sh ) represents a parallel high-conductivity path across the p-n junction or the cell edges and decreases the efficiency of the cells by increasing the leakage current, which lowers the maximal output power (P m ), the opencircuit voltage (V oc ), and the fill factor (FF) [10,11]. R sh is crucial to PV performance, especially at reduced irradiance levels. When the intensity level falls, as on cloudy days or when the sun is lower in the sky, R sh becomes an increasing concern [12]. Fig. 5 shows a picture taken manually at the eagle view of the shunt meter, which shows the lock in the thermogram of the cells before and after edge isolation. We took special care in our solarcell process to avoid handling defects like cracks, holes, scratches, and aluminum at the surface, as these defects

4 Laser Edge Isolation for High-efficiency Crystalline Silicon Dohyeon Kyeong et al Table 1. Comparison of solar cell parameters for laser and manual edge isolation. Edge Isolation V oc I sc FF Eff R s R sh (mv) (A) (%) (%) (mω) (Ω) Laser Mechanical grinding Table 2. Solar cell parameters for front and back edge isolated cells. Laser Edge V oc I sc FF Eff R s R sh Isolation (mv) (A) (%) (%) (mω) (Ω) Front (0.5 mm) Front (0.1 mm) Back (0.5 mm) Fig. 5. Thermal image of an edge shunt (5inch CZ wafer): (a) without edge isolation and (b) with laser edge isolation. cause shunts. In most of our solar cells, we observed only edge shunts (Fig. 5(a)). After edge isolation using the optimized laser parameters, the edge shunts were absent (Fig. 5(b)), and the measured high shunt resistance confirmed proper edge isolation. 3. Solar Cell Characterization The performances of the laser edge isolated solar cells were measured under AM 1.5 condition at 25 C. To compare the laser and manual edge isolation (manual edge isolation means that the edge was manually ground using salt paper), We show the solar cell parameters obtained for both cases in Table 1. The solar cell parameters obtained for manual and laser edge isolation are almost identical. The advantages of laser edge isolation compared to mechanical grinding are innumerable. Laser edge isolation can be used as an inline process, which is compatible with industrial solar cell fabrication. Because of this desirable result of laser edge isolation, we also investigated the rear side edge isolation and edge isolation before the SiN x deposition (Figs. 2(b) and 2(c)). The solar cell parameters obtained for the different edge isolations are mentioned in three different tables. For Tables 1 and 3, we used similar grade wafers, and for Table 2, we used a different wafer. Rear side edge isolation yields better efficiency compared to front edge isolation because the front active area is designed entirely so as to maximize the final efficiency output. Table 2 showed improved short circuit current with increasing front active area. In our LIV measurement setup, the solar cell was placed on a copper base, which may cause a shunt through the outer rim of the scribe line. Because of this, we cannot expect better LIV characteristics from the rear side edge isolation at certain times. If the measurement setup is properly designed so that the edges are isolated from the bottom electrode, then we can rely on the rear side edge isolation. Due to this, our subsequent experiments focused on front side edge isolation. The solar cell parameters increased when the front edge isolation was made at a distance of 0.1 mm, instead of 0.5 mm, from the edge. However, due to the non-uniformity in the wafer edges, the edge isolation at 0.1 mm at times went beyond the boundary of the wafer. Thus, it was difficult to control laser isolation within the wafer unless the wafer had perfect dimensions. Therefore, we suggest that the best edge isolation line should be located at a distance of 0.2 mm from the edge of the wafer. Laser scribing will always produce heat-induced defects at the scribe line, which may cause a small reduction in final solar cell efficiency due to a decrease in the carrier s lifetime. We tried to passivate the defects by altering the edge isolation step prior to the SiN x deposition. The schematic of the scribed line covered with the SiN x layer is shown in Fig. 2(d). The laser edge isolation was carried out after the emitter layer doping; then, the SiN x layer was coated. The solar cell that was fabricated with this small modification in the process steps (Fig. 1(b)) showed improved final efficiency, which is detailed in Table 3. There was a small increase in the open circuit voltage and the fill factor compared to a cell that underwent edge isolation at the end of the

5 -128- Journal of the Korean Physical Society, Vol. 55, No. 1, July 2009 Table 3. Comparison of solar cell parameters with and without SiN x passivation. Laser Edge V oc I sc FF Eff R s R sh Isolation (mv) (A) (%) (%) (mω) (Ω) Edge isolation at the final step Edge isolation before SiN x deposition solar-cell process. This results from the passivation due to the deposition of the SiN x layer on the scribe line. In addition, this passivation effect can be clearly seen though increased Rsh value up to 52.3 Ω. Despite this, the solar-cell parameters can be increased by properly engineering the scribe line and the ARC coating. Grant No. NRL-ROA REFERENCES IV. CONCLUSION Edge isolation is an important step in the solar cell process. In industries, edge isolation is carried out by either plasma etching or chemical edge isolation. Laser edge isolation is fast and compatible with the inline process for industrial solar-cell fabrication. Front and rear side laser edge isolation was successfully carried out using a 1064-nm fiber laser. Edge isolation at a distance of 0.2 mm from the edge of the wafer with optimized laser parameters yielded the best result. Laser edge isolation prior to the deposition of SiN x yielded better solar-cell parameters. With our modified solar cell process, the defects induced by laser scribing were passivated without any additional steps in the solar-cell fabrication process. Our laser edge isolation made at a scribing velocity of 100 mm/s is compatible with inline industrial cell fabrication. ACKNOWLEDGMENTS We gratefully acknowledge financial support for this work from the National Research Laboratory under [1] A. D. Compaan, I. Matulionis and S. Nakade, Op. Las. Engin. 34, 15 (2000). [2] R. Hendel, Laser Technik J. 5, 32 (2008). [3] O. Breitenstein, J. P. Rakotoniaina, M. H. Al Rifai and M. Werner, Prog. Photovolt: Res. Appl. 12, 529 (2004). [4] M. Schulze and S. Venkat, Laser Technik J. 5, 38 (2008). [5] I. M. Hamammu and K. Ibrahim, in Proceedings of International Conference on Semiconductor Electronics (Penang, Malaysia, 2002), p [6] R. H. Micheels and P. E. Valdivia, IEEE Trans. Electron Dev. 37, 353 (1990). [7] M. D. Abbott, T. Trupke, H. P. Hartmann, R. Gupta and O. Breitenstein, Prog. Photovolt: Res. Appl. 15, 613 (2007). [8] J. Arumughan, T. Pernau, A. Hauser and I. Melnyk, Solar Energy Mater. Solar Cells 87, 705 (2005). [9] A. Kress, P. Fath, G. Willeke and E. Bucher, in Proceedings of Second World Conference and Exhibition on PVSEC (Vienna, 1998), p [10] M. Wolf and H. Rauschenbach, Adv. Energy Conv. 3, 455 (1963). [11] O. Breitenstein, M. Langenkamp, O. Lang and A. Schirrmacher, Solar Energy Mater. Solar Cells 65, 55 (2001). [12] T. J. McMahon, T. S. Basso and S. R. Rummel, in Proceedings of the 25th Photovoltaic Specialists Conference (Washington, DC, 1996), p