Impact of laser pulse width on laser ablation process of high performance PERC cells

Impact of laser pulse width on laser ablation process of high performance PERC cells

Available online at www.sciencedirect.com ScienceDirect Solar Energy 110 (2014) 208–213 www.elsevier.com/locate/solener Impact of laser pulse width ...

2MB Sizes 0 Downloads 28 Views

Available online at www.sciencedirect.com

ScienceDirect Solar Energy 110 (2014) 208–213 www.elsevier.com/locate/solener

Impact of laser pulse width on laser ablation process of high performance PERC cells Myungsu Kim a,⇑, Donghwan Kim b, Dongseop Kim a, Yoonmook Kang c,⇑ a

Solar Energy Development Group, Samsung SDI, San #24 Nongseo-dong, Giheung-gu, Yongin-si, Gyunggi-do 446-711, Republic of Korea b Department of Materials Science and Engineering, Korea University, 145 Anam-ro, Seongbuk-gu, Seoul 136-701, Republic of Korea c KUKIST Green School, Graduate School of Energy and Environment, Korea University, 145 Anam-ro, Seongbuk-gu, Seoul 136-701, Republic of Korea Received 13 July 2014; received in revised form 30 August 2014; accepted 3 September 2014

Communicated by: Associate Editor Nicola Romeo

Abstract This study quantitatively compares the characteristics of ablation processes using nanosecond (ns) and picosecond (ps) pulse width green (532 nm) lasers. The laser ablation results are analyzed using Electron Probe Micro Analyzer (EPMA), quasi-steady-state photoconductance (QSSPC) measurements and transmission electron microscopy (TEM). The ablated using the ns green laser is predominantly melted, due to the relatively longer pulse width, and laser damage is incurred to a depth of 2.5 lm. Meanwhile, the laser ablation using the ps green laser precisely removes the thin layers on the surface without severely melting the sample and the observed laser damage depth is almost negligible. However, since the maximum damage depth (2.5 lm) using the ns laser is much shallower than the local contact depth (10–17 lm), the passivated emitter and rear cell (PERC) efficiencies using the ns and ps pulse width lasers converge to a power conversion efficiency of 19.4%. Ó 2014 Elsevier Ltd. All rights reserved.

Keywords: Laser ablation; Picosecond pulse width; Nanosecond pulse width; PERC cell

1. Introduction The advanced technologies of fabricating high efficiency solar cells are gaining considerable attention for the further improvement of the cell efficiency (Gatz et al., 2011; Wang et al., 1990; Roder et al., 2010; Kerschaver and Beaucarne, 2006). As one of the high efficiency solar cell technologies, the PERC cell is of interest and is being actively investigated, due to the high reflection and reduced recombination at the rear surface of the cell (Blakers et al., 1989; Lai et al., 2011; Yamamoto et al., 2004; Khan et al., 2014). Compared ⇑ Corresponding authors.

E-mail addresses: [email protected] (M. Kim), ddang@ korea.ac.kr (Y. Kang). http://dx.doi.org/10.1016/j.solener.2014.09.001 0038-092X/Ó 2014 Elsevier Ltd. All rights reserved.

to the fabrication scheme of the conventional screen printed cell, additional process steps are necessary to realize a PERC cell structure such as passivation layer deposition and rear local contact opening at rear side by applying laser ablation or printable etch paste methods (Stuart et al., 1996; Gamaly et al., 2002; Mangersnes et al., 2010; Mishra et al., 2010). This article focuses on the local contact opening process using ns and ps pulse width laser ablation systems which have a huge impact on the ablation quality. More specifically, a ns pulse width laser dominantly melts materials due to the relatively longer irradiation time while ps pulse width laser transfers energy to sublimate the materials which causes less ablation induced damage (Stuart et al., 1996; Knorz et al., 2002). However, since the cost of the ps laser system is approximately 5–10 times more expensive

M. Kim et al. / Solar Energy 110 (2014) 208–213

209

Fig. 1. Optical microscopy image (a and b) and 3D laser scanning image (c and d) of ns/ps pulse width lasers.

than that of the ns laser system, it is not a cost effective solution. For this reason, the laser ablation qualities using ns and ps pulse width lasers are thoroughly compared. 2. Experiments The laser-ablated PERC cells were fabricated with 6-in., boron-doped, p-type Czochralski (CZ) wafers. The wafers were textured using alkaline solution to form random pyramids and remove any marks on the wafer. To form a p–n junction, POCl3 precursor was diffused on the wafer in the standard tube thermal furnace and the sheet resistance of emitter was 60 X/sq. After diffusion, a SiNx AR layer of 100 nm thickness was coated on the front of the wafer. In the following emitter removal step, the pyramids on the backside of the wafer were etched in KOH solution for 4 min to generate a flat surface. In a batch process, the deposited SiNx layer protected etching on the front side wafer. Next, the Al2O3 layer of 30 nm thickness was deposited using atomic layer deposition (ALD). On top of the Al2O3 layer, a SiNx layer of 320 nm thickness was deposited to prevent paste through-problem. In the next step, the rear passivation layer was locally removed by LA process with various combinations of diameter and pitch of local contacts. Laser systems with two different pulse widths are applied to ablate passivation layer of PERC cell, viz. ns or ps pulse widths. The pulse widths of the ns and ps green lasers are 20 ns and 9 ps, respectively, and the wavelength for both laser systems is 532 nm. The laser energy density is adjusted by changing the current and stage height. The beam shape is defined by optimizing the lens position and galvanometer movement. The power densities of the lasers are measured using a Newport power detector. The samples were then screen-printed with a Ag/Al bus bar for module assembly and fired. To prevent void formation

in the local contact, eutectic Al–Si alloy paste was applied. After firing the cells, the samples were laser-isolated. Each group comprised 10 samples. The fabricated cells were characterized under 1 sun intensity (AM1.5G). Before characterization, the reference cell measured at Fraunhofer ISE was used for calibration. 3. Results and discussion First, the visual ablation features obtained using the ns and ps pulse width lasers were carefully examined using an optical microscope and 3D laser scanning system after ablating the samples (Fig. 1). The sample ablated using an ns pulse width laser with a power of 3.2 W was characterized and its optical microscopy image in Fig. 1(a) shows molten silicon at the center of the laser-ablated hole. The thin layer of SiNx having a purple color is well removed and the silicon, underneath SiNx, was melted due to the pulse width of ns order. The other characteristic of the ns laser ablation is the protruded edge around the laserablated hole and its height is measured to be about 8 lm (Fig. 1c). On the other hand, the laser ablation using the ps pulse width more precisely removes the passivation thin layer (Fig. 1b). The square shaped pyramid vestige underneath the laser-ablated spot remains as a crystal-like morphology rather than molten silicon. The profile of the laser-ablated sample in Fig. 1(d) shows almost the same height as the non-laser ablated region. This demonstrates that the ps laser ablation removes the thin films with much less damage on the surface. In addition, the residues of the thin passivation materials after ablation were quantitatively analyzed using EPMA (Fig. 2). EPMA is a powerful tool to quantify the concentration of residues and its working principle can be found elsewhere (Ro et al., 1999; Dreer et al., 1999).

210

M. Kim et al. / Solar Energy 110 (2014) 208–213

Fig. 2. Elemental analysis of residual nitrogen and aluminum using EPMA.

Fig. 3. impVoc and effective lifetime depending on laser damage removal using ns pulse with laser (a) and ps pulse width laser (b).

M. Kim et al. / Solar Energy 110 (2014) 208–213

211

Fig. 4. TEM cross section images of laser-ablated samples. (a) Samples ablated using nanosecond pulse width laser. (b) Red circle inside (a) is magnified to show interfacial a-Si damage. (c) is taken from the sample processed by the ps pulse width laser. (d) Red circle inside of (c) is magnified. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

As can be seen in Fig. 2(a), the nitrogen (N) and aluminum (Al) residues after ablation using the ns laser can be visualized. The N and Al mappings represent SiNx and Al2O3, respectively. The majority of N is well removed at all power densities, though some small particles remain in the laserablated region. The concentration of the remaining N particles is around 3–6%. On the other hand, the concentration of residual Al2O3 varied depending on the laser irradiation density. At a laser intensity of 3.2 W, a significant amount of Al remains with concentrations ranging from 3% to 13%. When the intensity is increased over 6.4 W, most of the Al is removed.

Fig. 5. The optical microscope image of local contact depth.

Using the ps laser the quality of removal of the passivation layer is also analyzed in Fig. 2(b). In the case of the ps pulse width laser with an intensity of 2.8 W, most of the SiNx layer is removed, as intended. However, the concentrations of Al remaining at the center and edge are 7% and 42%, respectively. When the intensity is increased above 5.6 W, the Al residues almost disappear. Based on the EPMA results, the laser energies for ablation have been optimized as 3.2 W for the ns and 2.8 W for the ps pulse width lasers. The criteria to determine the proper laser energy are primarily related to the concentration of N residues representing SiNx because Al and O elements are included in the screen printing paste. The laser damage depth in the laser ablation process is also a pivotal characteristic for judging the process suitability. The wavelength and pulse width of the laser are the main variables influencing on the laser damage depth (Chichkov et al., 1996). The quasi-steady-state photoconductance (QSSPC) technique is utilized to analyze the laser damage by reading the impVoc trend as the laser damage is gradually removed. The samples have an identical structure of real cells which are processed from texturing to Al2O3 and SiNx passivations. Then, these samples are laser-ablated using the ns and ps lasers at the optimized intensities of 3.2 W and 2.8 W, respectively. The laser damaged samples are dipped in KOH etching solution and the damage is gradually removed as the time elapses from 10 to 30 and 60 s with

212

M. Kim et al. / Solar Energy 110 (2014) 208–213

Table 1 PERC cell parameters depending on laser pulse widths. Pulse width

Laser power (Watt)

Jsc (mA/sq)

Voc (mV)

Eff. (%) (STDEV)

FF (%)

Rs (ohm  sq)

Rsh (ohm  sq)

n Factor

20 ns

3.2

38.1

638.0

79.7

0.6

40888.6

1.1

9 ps

2.8

38.2

640.1

19.4 (0.05) 19.4 (0.07)

79.4

0.6

44016.1

1.1

the etch rate of 5 lm/min. The etched samples are re-passivated with Al2O3 using atomic layer deposition and the QSSPC data is generated from each sample. Interestingly, the implied Voc trends of the ns and ps pulse width lasers are similar. After ablation the initial impVoc value significantly decreases over 10 mV, mainly due to the laser damage. (Fig. 3) As the damage is removed for 10 s, impVoc recovers back about 2–5 mV. It is generally accepted that the laser damage depth from an ns pulse width laser is deeper than that from a pico second pulse width laser. However, the both obtained impVoc trends indicate that the damage depth scales from the ns and ps lasers are not significantly different. This minor difference can be compensated for by the local contact formation process during firing. This indicates that a ns pulse width laser system can be a cost effective choice which does not cause severe damage during laser ablation. To confirm the results of the laser damage depth analysis using QSSPC, the cross section of the samples laser-ablated using the ns and ps lasers was observed using TEM. The image (Fig. 4a) of the sample ablated using the ns pulse width laser clearly shows a recrystallized morphology with a depth of 100 nm at the interface between the sample and tungsten (W) coating. Tungsten is used to fix the sample for TEM. When the interface is magnified (Fig. 4b), the ˚ . The thickness of the a-Si layer is observed to be 60 A TEM image of the sample processed using the pico second laser system is visualized in Fig. 4(c). There is no observable recrystallized damage near the interface. The laser ablation using the pico second laser system does not create any recrystallized laser damage underneath the processed region and such damage is limited to the surface. The image of the same sample was also magnified to examine the a-Si resulting from the ps pulse width laser (Fig. 4d). The thickness of the amorphous silicon layer resulting from the pico ˚ , which is shallower than second laser system is about 32 A ˚. that obtained the from nano second laser system, viz. 60 A The results of the laser damage analysis using TEM confirm that the ps laser system creates less damage than the ns laser system. However, since the damage depth after ablation from the ns and ps pulse width lasers is not significantly different, both ns and ps pulse width laser systems can be used for ablation, considering the local contact depth after completing cell fabrication ranges from 10 to 17 lm (Fig. 5) (Stuart et al., 1996; Gamaly et al., 2002; Knorz et al., 2002; Liu et al., 1979; Kim et al., 2013). The efficiencies and parameters of the PERC cells ablated using the ns and ps pulse width lasers are compared in Table 1. For local contact ablation, the ablated diameter

and pitch are consistently maintained to be 90 lm and 600 lm in the case of both the ns and ps pulse width laser systems. Each group has 10 samples and the resulting efficiencies obtained using the ns and ps pulse width laser systems converge to 19.4%, which indicates that the difference in the degree of laser damage resulting from the ablations using the ns and ps pulse width lasers is negligible. When each cell parameter is compared, the fill factor of the ps pulse width group is 0.3% smaller, while Voc is 2.1 mV higher, than those of the ns pulse width laser. These deviations may originate from the slight variation of the local contact diameter size. 4. Conclusions The PERC solar cells with a high efficiency of 19.4% were successfully fabricated using ns and ps pulse width lasers. The laser ablation using the ns pulse width green laser removed the SiNx and Al2O3 layers well at the optimized laser energy density. The processed local contact showed a molten morphology, which was attributed to the ns pulse width, but the measured laser damage depth of less than 1 lm did not degrade the cell performance, since the rear local contact depth is 10–17 lm. The ablation process using the ps pulse width laser demonstrated outstanding ablation quality, mainly sublimating the thin layers with less damage than ns pulse width laser ablation. After ablation, a pyramid shaped vestige having a crystalline morphology still remained, due to the intrinsic feature of the pico second order pulse width laser. Acknowledgements This work was supported by the National Research Foundation of Korea Grant funded by the Korean Government (MSIP)” (2014, University-Institute cooperation program); by the Human Resources Development program (No. 20124030200120) of the Korea Institute of Energy Technology Evaluation and Planning (KETEP) Grant funded by the Korea government Ministry of Trade, Industry and Energy. References Blakers, Andrew W., Wang, Aihua, Milne, Adele M., Zhao, Jianhua, 1989. 22.8% Efficient silicon solar cell. Appl. Phys. Lett. 55 (13), 1363–1365. Chichkov, B.N., Momma, C., Nolte, S., von Alvensleben, F., Tunnermann, A., 1996. Femtosecond, picosecond and nanosecond laser ablation of solids. Appl. Phys. A 63 (2), 109–115. Dreer, Sabine, Wilhartitz, Peter, Mersdorf, Edgar, Piplits, Kurt, Friedbacher, Gernot, 1999. Quantitative analysis of thin aluminium-oxynitride films by EPMA. Microchim. Acta 131 (3–4), 211–218.

M. Kim et al. / Solar Energy 110 (2014) 208–213 Gamaly, E.G., Rode, A.V., Luther-Davies, B., Tikhonchuk, V.T., 2002. Precision ablation of dental enamel using a subpicosecond pulsed laser. Phys. Plasmas 9, 949–957. Gatz, Sebastian, Hannebauer, Helge, Hesse, Rene, Werner, Florian, Schmidt, Arne, Dullweber, Thorsten, Schmidt, Jan, Bothe, Karsten, Brendel, Rolf, 2011. 19.4%-Efficient large-area fully screen-printed silicon solar cells. Phys. Status Solidi – Rapid Res. Lett. 5 (4), 147–149. Kerschaver, Emmanuel Van, Beaucarne, Guy, 2006. Back-contact solar cells: a review. Prog. Photovoltaics 14 (2), 107–123. Khan, Firoz, Baek, Seong-Ho, Mobin, Abdul, Kim, Jae Hyun, 2014. Enhanced performance of silicon solar cells by application of low-cost sol–gel-derived Al-rich ZnO film. Sol. Energy 101, 265–271. Kim, Myungsu, Park, Sungchan, Kim, Dongseop, 2013. Highly efficient PERC cells fabricated using the low cost laser ablation process. Sol. Energy Mater. Sol. Cells 117, 126–131. Knorz, Annerose, Peters, Marius, Grohe, Andreas, Harmel, Christian, Preu, Ralf, 2002. Selective laser ablation of SiNx layers on textured surfaces for low temperature front side metallizations. Prog. Photovoltaics 17 (2), 127–136. Lai, Jiun-Hong, Upadhyaya, A., Ramanathan, S., Das, A., 2011. Highefficiency large-area rear passivated silicon solar cells with local AlBSF and screen-printed contacts. IEEE J. Photovoltaics 1 (1), 16–21.

213

Liu, P.L., Yen, R., Bloembergen, N., 1979. Picosecond laser-induced melting and resolidification morphology on Si. Appl. Phys. Lett. 34 (12), 864. Mangersnes, Krister, Foss, S.E., Thøgerse, Annett, 2010. Damage free laser ablation of SiO2 for local contact opening on silicon solar cells using an a-Si:H buffer layer. J. Appl. Phys. 107, 043518–043524. Mishra, R., Zhao, L., Zhang, Z., Guo, H.-W., 2010. Development of norinse screen printable etch paste for contact via in dielectric films. 35th IEEE PVSC, pp. 003536–003539. Ro, Chul-Un, Osan, Janos, Grieken, Rene Van, 1999. Determination of chemical species in individual aerosol particles using ultrathin window EPMA. Anal. Chem. 71, 1521–1528. Roder, T.C., Eisele, S.J., Grabitz, P., Wagner, C., Kulushich, G., Kohler, J.R., Werner, J.H., 2010. Add-on laser tailored selective emitter solar cells. Prog. Photovoltaics 18 (7), 505–510. Stuart, B.C., Feit, M.D., Rubenchik, A.M., Shore, B.W., Perry, M.D., 1996. Optical ablation by high-power short-pulse lasers. Phys. Rev. Lett. 74, 2248–2251. Wang, A., Zhao, J., Green, M.A., 1990. 24% Efficient silicon solar cells. Appl. Phys. Lett. 57 (6), 602–604. Yamamoto, Kenji, Nakajima, Akihiko, Yoshimi, Masashi, Sawada, Toru, Fukuda, Susumu, Suezaki, Takashi, Ichikawa, Mitsuru, Koi, Yohei, 2004. A high efficiency thin film silicon solar cell and module. Sol. Energy 77 (6), 939–949.