- Email: [email protected]
Fabrication of selective-emitter silicon heterojunction solar cells using hot-wire chemical vapor deposition and laser doping
Thin Solid Films 517 (2009) 4749–4752
Contents lists available at ScienceDirect
Thin Solid Films j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / t s f
Fabrication of selective-emitter silicon heterojunction solar cells using hot-wire chemical vapor deposition and laser doping B.R. Wu a, D.S. Wuu a,⁎, M.S. Wan b, H.Y. Mao a, R.H. Horng b a b
Department of Materials Science and Engineering, National Chung Hsing University, Taichung 40227, Taiwan, ROC Institute of Precision Engineering, National Chung Hsing University, National Chung Hsing University, Taichung 40227, Taiwan, ROC
a r t i c l e
i n f o
Available online 24 March 2009 Keywords: Selective emitter Silicon Heterojunction Solar cells Hot-wire chemical vapor deposition Laser doping
a b s t r a c t The Si heterojunction (HJ) solar cells were fabricated on the textured p-type mono-crystalline Si (c-Si) substrates using hot-wire chemical vapor deposition (HWCVD). In view of the potential for the bottom cell in a hybrid junction structure, the microcrystalline Si (μc-Si) film was used as the emitter with various PH3 dilution ratios. Prior to the n-μc-Si emitter deposition, a 5 nm-thick intrinsic amorphous Si layer (i-a-Si) was grown to passivate the c-Si surface. In order to improve the indium-tin oxide (ITO)/emitter front contact without using the higher PH3 doping concentration, a laser doping technique was employed to improve the ITO/n-μc-Si contact via the formation of the selective emitter structure. For a cell structure of Ag grid/ ITO/n-μc-Si emitter/i-a-Si/textured p-c-Si/Al-electrode, the conversion efficiency (AM1.5) can be improved from 13.25% to 14.31% (cell area: 2 cm × 2 cm) via a suitable selective laser doping process. © 2009 Elsevier B.V. All rights reserved.
1. Introduction Recently, thin-film silicon/crystalline silicon (c-Si) heterojunction (HJ) solar cells have been studied extensively due to their promising performance and low-temperature fabrication process. In this type of cell, a Si thin-film emitter layer is typically deposited by plasmaenhanced chemical vapor deposition to form a p–n junction at low temperatures (b400 °C) [1,2]. The emitter layer in the HJ cell is usually highly doped with a very thin thickness which helps to reduce the series resistance and to minimize the recombination of photogenerated carriers in the layer. It is well known that the performance of the HJ cell is strongly dependent on the emitter properties and the quality of the thin-film Si/c-Si interface [3]. Efforts to improve the material quality of the Si thin-film emitter itself have focused on the development of various deposition techniques, such as hot-wire chemical vapor deposition (HWCVD) [4], very highfrequency plasma CVD, and electron cyclotron resonance CVD. Among them, the HWCVD has attracted considerable attention because both amorphous silicon (a-Si) or microcrystalline silicon (μc-Si) films can be achieved with higher deposition rate [5,6] and lower deposition temperature [7,8]. In this paper, the n-μc-Si/p-c-Si HJ solar cell was investigated, where this type of cell could be further used for the bottom cell in a hybrid junction structure [9,10]. All cells in this work were fabricated on textured c-Si substrates by HWCVD. A systematic study of various PH3 dilution ratios in the HWCVD process was carried out to optimize the n-μc-Si emitter properties for the Si HJ cell applications. The influence of the contact property between indium-tin oxide (ITO) ⁎ Corresponding author. Tel.: +886 4 22840500x714; fax: +886 4 22855046. E-mail address: [email protected] (D.S. Wuu). 0040-6090/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.tsf.2009.03.119
and n-μc-Si on the cell performance is also investigated. Especially, a laser doping technique was employed to improve the ITO/n-μc-Si contact via the formation of the selective emitter structure. Finally, the photovoltaic properties like open-circuit voltage (Voc), short-circuit current density (Jsc), fill factor (FF), and conversional efficiency (η) of the μc-Si/p-c-Si HJ cells with and without the selective emitter will be characterized. 2. Experimental All Si thin films used in this study were deposited using a HWCVD apparatus as described previously [11], where the filament temperature was monitored by an optical pyrometer and the substrate temperature was measured using a thermocouple embedded in the substrate holder. A textured Czochralski c-Si wafer (250 μm thick, boron-doped, 1–5 Ω cm) was used as the substrate. The 1-μm-thick Al back contact was deposited by electron-beam evaporation and then annealed at 600 °C in order to achieve both effective back surface field and ohmic contact [4]. Before loading the sample, a chamber cleaning process was performed using the atomic H generated on the hot tungsten wires [12] and the substrate was dipped in 1% diluted hydrofluoric acid for H-termination. The c-Si substrate was then subjected to the H atom pre-treatment process to passivate the c-Si surface [13], which was carried out by decomposing H2 in the same HWCVD chamber. A 5 nm-thick intrinsic aSi layer (i-a-Si) has been inserted between the c-Si substrate and n-μc-Si emitter layer [14]. Details of the i-a-Si layer effect on the characterization of Si HJ solar cells have been reported elsewhere [15]. Following, a 20nm-thick n-μc-Si emitter layer was deposited under various PH3 dilution ratios [RPH3 = PH3 / (SiH4 + H2 + PH3)]. Table 1 summarizes the HWCVD pretreatment and deposition parameters for the i-a-Si and n-μc-Si films.
4750
B.R. Wu et al. / Thin Solid Films 517 (2009) 4749–4752
Table 1 Deposition parameters of Si thin films by HWCVD used in this study for Si heterojunction solar cells. Deposition parameters
H-treatment
i-a-Si
n-μc-Si
Wire temperature (°C) Substrate temperature (°C) Pressure (mTorr) Process gas flow (sccm)
1650 150 200 H2 = 80
1650 250 100 SiH4 = 5
1750 200 100 SiH4 = 2 H2 = 17.82–17.98 PH3 = 0.02–0.18
After the n-μc-Si emitter deposition, a laser doping technique was used to improve the ITO/n-μc-Si contact performance. A phosphorusdoped SiO2 solution was prepared as a dopant source using a sol-gel process. It was spun on the n-μc-Si emitter film, prebaked at 80 °C for 20 min and then postbaked at 200°C for 1 h under atmosphere. An NdYAG laser (wavelength: 355 nm; repetition rate: 10 Hz) with various power densities (64 to 318 mJ/cm2, beam size = 7.85 × 10− 3 cm2) was scanned through a metallic mask with periodic circular hole to make the heavily doped selective emitter. The hole diameter is kept at 0.5 mm with different spacings from 0.625 to 5 mm. After the laser irradiation process, the residual dopant source was removed using a wet etching process. Then the ITO front contact (80 nm thick) and Ag grid contact (1 μm thick) were successively deposited on the emitter film using an electron beam evaporator. Finally, the cell structure (Ag grid/ITO/n-μc-Si selective emitter/i-a-Si/textured p-c-Si/Alelectrode) was fabricated and shown in Fig. 1. The cell characteristics (area: 2 cm × 2 cm) were obtained through the current–voltage (I–V) measurements at100 mW/cm2 with an AM 1.5-like spectrum. 3. Results and discussion The control of the dopant concentration in the emitter layer is the key point in the photovoltaic properties of a HJ cell since it determines the equilibrium between defect density and position of Fermi level at the interface [16]. The first step in the present work was the deposition of singly doped films in order to determine the doping efficiency of phosphine in our system. Here the energy level at absorption coefficient of 104 cm− 1 (E04) of the n-μc-Si film was obtained from the transmission spectrum [17]. The carrier concentration (Nc) of the doped emitter film was determined using the Van der Pauw-Hall measurement. Fig. 2 shows the dependence of E04 and Nc of the n-μc-Si films on the PH3 dilution ratio (RPH3). It is found that both the E04 and Nc are strongly dependent on the RPH3. When the RPH3 increases from 0.1 to 0.9%, the Nc increases rapidly from 5 × 1015
Fig. 1. Schematic cross-section (a) and fabrication process (b) of Si heterojunction solar cell used in this study.
Fig. 2. Variation of optical band gap (E04) and carrier concentration (Nc) of n-type μc-Si emitter layer as functions of PH3 dilution ratio [RPH3 = PH3 / (SiH4 + H2 + PH3)].
to 3.7 × 1019 cm− 3, indicating a good PH3 doping efficiency in the emitter layer. On the other hand, the E04 first increases intensively from 1.81 to 2.01 eV when the RPH3 increases from 0.1 to 0.3% and then gradually increases to 2.08 eV when the RPH3 increases to 0.9%. The photovoltaic parameters (Jsc, Voc, FF, and η) of the HJ cells as functions of RPH3 in the n-μc-Si emitter are shown in Fig. 3. Apparently, the cell performance can be improved when the RPH3 increases from 0.1 to 0.5%. An optimum η value of 13.25% was achieved under a RPH3 of 0.5%. It was found that both the Jsc and FF values dropped when the RPH3 increased from 0.5 to 0.9%. These results indicate that a decrease of the doping concentration in the n-μc-Si emitter allows a shift of the Fermi level toward midgap and thus a reduction of the electric field at the interface. Under higher phosphorus doping levels, the increase of defect density in the emitter could cause recombination and a dramatic reduction of photovoltaic properties of the HJ cell [16]. In order to improve the ITO/emitter front contact without using the higher PH3 doping concentration, a selectively doped emitter via a metal mask was performed using the laser doping technique. The effect of laser irradiation density on the photovoltaic parameters of the HJ cell is illustrated in Fig. 4. As shown in this figure, all cells were
Fig. 3. Photovoltaic characteristics of Si heterojunction solar cells (Ag grid/ITO/n-μc-Si emitter/i-a-Si/p-c-Si/Al electrode) fabricated under various PH3 dilution ratio (RPH3) using HWCVD.
B.R. Wu et al. / Thin Solid Films 517 (2009) 4749–4752
4751
Fig. 5. Current–voltage characteristics of Si heterojunction solar cells with and without selective emitter measured at 100 mW/cm2 with an AM 1.5-like spectrum. Fig. 4. Photovoltaic characteristics of Si heterojunction solar cells (Ag grid/ITO/n-μc-Si selective emitter/i-a-Si/p-c-Si/Al electrode) fabricated under various laser power densities (Plaser).
fabricated under the RPH3 of 0.5% and treated by the laser scan with the hole spacing of 5 mm. It was found that the Jsc, FF and η data increased as the laser power density (Plaser) increased from 64 to 255 mJ/cm2. Further increasing the laser power density from 255 to 318 mJ/cm2 yielded the degraded Jsc, FF and η values. However, the Voc value decreased monotonously when the Plaser increased from 64 to 318 mJ/ cm2, especially for Plaser ≥ 255 mJ/cm2. It could be due to the enhanced cross doping problem at the film/substrate interface. Another possible reason for this could be attributed to the laser-induced damage in the i-a-Si and n-μc-Si films. The reason is still not clear. Nevertheless, a maximum conversion efficiency of 14.02% can be achieved for the Si HJ solar cell under a Plaser of 255 mJ/cm2. The result reveals that the selectively doped emitter structure can improve the ITO/n-μc-Si contact property, thus enhancing the μc-Si/c-Si HJ cell performance. Further improvement in the ITO/n-μc-Si contact can be performed by optimizing the hole spacing (D) of the metallic mask during the laser doping process. The D effect on the performance of the Si HJ solar cells is summarized in Table 2 under a fixed Plaser of 255 mJ/cm2. The data reported here were taken from the average measured value of five cell samples. It was found that the Jsc, FF, and η values increased as the D decreased from 5 to 1.25 mm. This could be due to the sufficient doping area, lowering the ITO/emitter contact resistance. Further reducing the D value (e.g. D = 0.625 mm) resulted in an inferior Voc value and cell performance. It could be attributed to the emitter damage caused by the heavily doped concentration. However, the Si HJ cells with selective emitters (D = 5 to 1.25 mm) show the equivalent or better efficiencies than that of the original cell. A best conversion efficiency of 14.31% is achieved for the selective-emitter μc-Si/c-Si HJ cell under the D value of 1.25 mm and Plaser of 255 mJ/cm2. Since the area of the selective emitter is changing, it is reasonable to suspect that the grain sizes will also have an effect on the formation of the emitters. However, from our
Table 2 Photovoltaic properties of the Ag grid/ITO/n-μc-Si selective emitter/i-a-Si/ p-c-Si/Al solar cell (area: 2 cm × 2 cm). Spacing (mm)
η (%)
Voc (V)
Jsc (mA/cm2)
FF (%)
Original cell 5 2.5 1.25 0.625
13.25 14.02 14.11 14.31 11.57
0.583 0.566 0.562 0.550 0.532
30.61 31.88 32.19 33.25 30.32
74.2 77.7 78.0 78.3 71.7
An Nd-YAG laser was scanned through a metallic mask with periodic circular hole to make the heavily doped selective emitter where the hole diameter is kept at 0.5 mm with different spacings (D) from 0.625 to 5 mm.
experimental data, the laser-treated emitter (Plaser = 255 mJ/cm2, D = 5 mm, no phosphorus spin coating) showed nearly the same in the conversion efficiency as compared with that of the original cell (without laser treatment). This exhibits that the laser treatment on the HWCVD μc-Si film has less effect on the contact resistance of the ITO/emitter in the Si heterojunction cell. Fig. 5 compares the I–V characteristics for the HJ cells with and without the selective emitter structure under AM1.5 illumination. It is worthy to note that the sample with selective-emitter structure shows the better Jsc, and FF but a lower Voc (550 mV as compared to 583 mV). The conversion efficiency can be improved from 13.25% to 14.31% via a suitable selective laser doping process. Moreover, the series and shunt resistances of the samples were determined using the dark I–V data. The series resistances were 0.23 Ω and 0.68 Ω and the shunt resistances were 1.78 × 105 Ω and 1.81 × 105 Ω for the best cell with and without the selective emitter process, respectively. These results show the improvement of the cells performance with the change of contact characteristics. It suggests that the proposed selective-emitter structure is very promising for future fabrication of high-efficiency HJ solar cells. 4. Conclusion A cell structure of Ag grid/ITO/n-μc-Si selective emitter/i-a-Si/ textured p-c-Si/Al-electrode was fabricated using a combination of HWCVD and laser doping techniques. The effects of selectively doped emitter on the characteristics of μc-Si/c-Si HJ solar cells have been studied. The initial efficiency up to 13.25% can be obtained for the Si HJ cells under a PH3 dilution rate of 0.5%. By using the selective-emitter structure, the cell properties were improved. The optimization of the selective-emitter layer was conducted in terms of Nc, Plaser, and D. Our best selective-emitter n-μc-Si/p-c-Si HJ cell (area: 2 cm × 2 cm) shows an efficiency of 14.31% with a Voc of 550 mV, a Jsc of 33.25 mA/cm2, and an FF of 78.3%. This is a promising result for the fabrication of lowcost and high-efficiency Si HJ solar cells. Acknowledgment This work was supported by the National Science Council of the Republic of China under contract no. NSC-96-2628-E-005-090-MY3. References [1] M. Tucci, G.D. Cesare, J. Non-Cryst. Solids 338–340 (2004) 663. [2] E. Centurioni, D. Iencinella, R. Rizzoli, F. Zignani, IEEE Trans. Electron Devices 51 (2004) 1818. [3] B. Jagannathan, W.A. Anderson, Sol. Energy Mater. Sol. Cells 44 (1996) 165.
4752
B.R. Wu et al. / Thin Solid Films 517 (2009) 4749–4752
[4] H.M. Branz, C.W. Teplin, D.L. Young, M.R. Page, E. Iwaniczko, L. Roybal, R. Bauer, A.H. Mahan, Y. Xu, P. Stradins, T. Wang, Q. Wang, Thin Solid Films 516 (2008) 743. [5] H.S. Povolny, X. Deng, Thin Solid Films 430 (2003) 125. [6] E. Iwaniczko, Y. Xu, R.E.I. Schropp, A.H. Mahan, Thin Solid Films 430 (2003) 212. [7] T.H. Wang, E. Iwaniczko, M.R. Page, D.H. Levi, Y. Yan, H.M. Branz, Q. Wang, Thin Solid Films 501 (2006) 284. [8] S.S. Lee, M.S. Ko, C.S. Kim, N.M. Hwang, J. Cryst. Growth 310 (2008) 3659. [9] J. Damon-Lacoste, P. Roca i Cabarrocas, P. Chatterjee, Y. Veschetti, A.S. Gudovskikh, J.P. Kleider, P.J. Ribeyrone, J. Non-Cryst. Solids 352 (2006) 1928. [10] Y. Xu, Z. Hu, H. Diao, Y. Cai, S. Zhang, X. Zeng, H. Hao, X. Liao, E. Fortunato, R. Martins, J. Non-Cryst. Solids 352 (2006) 1972.
[11] D.S. Wuu, S.Y. Lien, H.Y. Mao, B.R. Wu, I.C. Hsieh, P.C. Yao, J.H. Wang, W.C. Chen, Thin Solid Films 498 (2006) 9. [12] A. Masuda, H. Matsumura, Thin Solid Films 395 (2001) 112. [13] M.W.M. van Cleef, J.K. Rath, F.A. Rubinelli, C.H.M. van dert Werf, R.E.I. Schropp, W.F. van der Weg, J. Appl. Phys. 82 (1997) 6089. [14] S.K. Kim, J.C. Lee, S.J. Park, Y.J. Kim, K.H. Yoon, Sol. Energy Mater. Sol. Cells 92 (2008) 298. [15] S.Y. Lien, B.R. Wu, J.C. Liu, D.S. Wuu, Thin Solid Films 516 (2008) 747. [16] M. Tucci, Sol. Energy Mater. Sol. Cells 57 (1999) 249. [17] K.D. Mackenzie, J.R. Eggert, D.J. Leoplod, Y.M. Li, S. Lin, W. Paul, Phys. Rev. B31 (1985) 2198.
Contents lists available at ScienceDirect
Thin Solid Films j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / t s f
Fabrication of selective-emitter silicon heterojunction solar cells using hot-wire chemical vapor deposition and laser doping B.R. Wu a, D.S. Wuu a,⁎, M.S. Wan b, H.Y. Mao a, R.H. Horng b a b
Department of Materials Science and Engineering, National Chung Hsing University, Taichung 40227, Taiwan, ROC Institute of Precision Engineering, National Chung Hsing University, National Chung Hsing University, Taichung 40227, Taiwan, ROC
a r t i c l e
i n f o
Available online 24 March 2009 Keywords: Selective emitter Silicon Heterojunction Solar cells Hot-wire chemical vapor deposition Laser doping
a b s t r a c t The Si heterojunction (HJ) solar cells were fabricated on the textured p-type mono-crystalline Si (c-Si) substrates using hot-wire chemical vapor deposition (HWCVD). In view of the potential for the bottom cell in a hybrid junction structure, the microcrystalline Si (μc-Si) film was used as the emitter with various PH3 dilution ratios. Prior to the n-μc-Si emitter deposition, a 5 nm-thick intrinsic amorphous Si layer (i-a-Si) was grown to passivate the c-Si surface. In order to improve the indium-tin oxide (ITO)/emitter front contact without using the higher PH3 doping concentration, a laser doping technique was employed to improve the ITO/n-μc-Si contact via the formation of the selective emitter structure. For a cell structure of Ag grid/ ITO/n-μc-Si emitter/i-a-Si/textured p-c-Si/Al-electrode, the conversion efficiency (AM1.5) can be improved from 13.25% to 14.31% (cell area: 2 cm × 2 cm) via a suitable selective laser doping process. © 2009 Elsevier B.V. All rights reserved.
1. Introduction Recently, thin-film silicon/crystalline silicon (c-Si) heterojunction (HJ) solar cells have been studied extensively due to their promising performance and low-temperature fabrication process. In this type of cell, a Si thin-film emitter layer is typically deposited by plasmaenhanced chemical vapor deposition to form a p–n junction at low temperatures (b400 °C) [1,2]. The emitter layer in the HJ cell is usually highly doped with a very thin thickness which helps to reduce the series resistance and to minimize the recombination of photogenerated carriers in the layer. It is well known that the performance of the HJ cell is strongly dependent on the emitter properties and the quality of the thin-film Si/c-Si interface [3]. Efforts to improve the material quality of the Si thin-film emitter itself have focused on the development of various deposition techniques, such as hot-wire chemical vapor deposition (HWCVD) [4], very highfrequency plasma CVD, and electron cyclotron resonance CVD. Among them, the HWCVD has attracted considerable attention because both amorphous silicon (a-Si) or microcrystalline silicon (μc-Si) films can be achieved with higher deposition rate [5,6] and lower deposition temperature [7,8]. In this paper, the n-μc-Si/p-c-Si HJ solar cell was investigated, where this type of cell could be further used for the bottom cell in a hybrid junction structure [9,10]. All cells in this work were fabricated on textured c-Si substrates by HWCVD. A systematic study of various PH3 dilution ratios in the HWCVD process was carried out to optimize the n-μc-Si emitter properties for the Si HJ cell applications. The influence of the contact property between indium-tin oxide (ITO) ⁎ Corresponding author. Tel.: +886 4 22840500x714; fax: +886 4 22855046. E-mail address: [email protected] (D.S. Wuu). 0040-6090/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.tsf.2009.03.119
and n-μc-Si on the cell performance is also investigated. Especially, a laser doping technique was employed to improve the ITO/n-μc-Si contact via the formation of the selective emitter structure. Finally, the photovoltaic properties like open-circuit voltage (Voc), short-circuit current density (Jsc), fill factor (FF), and conversional efficiency (η) of the μc-Si/p-c-Si HJ cells with and without the selective emitter will be characterized. 2. Experimental All Si thin films used in this study were deposited using a HWCVD apparatus as described previously [11], where the filament temperature was monitored by an optical pyrometer and the substrate temperature was measured using a thermocouple embedded in the substrate holder. A textured Czochralski c-Si wafer (250 μm thick, boron-doped, 1–5 Ω cm) was used as the substrate. The 1-μm-thick Al back contact was deposited by electron-beam evaporation and then annealed at 600 °C in order to achieve both effective back surface field and ohmic contact [4]. Before loading the sample, a chamber cleaning process was performed using the atomic H generated on the hot tungsten wires [12] and the substrate was dipped in 1% diluted hydrofluoric acid for H-termination. The c-Si substrate was then subjected to the H atom pre-treatment process to passivate the c-Si surface [13], which was carried out by decomposing H2 in the same HWCVD chamber. A 5 nm-thick intrinsic aSi layer (i-a-Si) has been inserted between the c-Si substrate and n-μc-Si emitter layer [14]. Details of the i-a-Si layer effect on the characterization of Si HJ solar cells have been reported elsewhere [15]. Following, a 20nm-thick n-μc-Si emitter layer was deposited under various PH3 dilution ratios [RPH3 = PH3 / (SiH4 + H2 + PH3)]. Table 1 summarizes the HWCVD pretreatment and deposition parameters for the i-a-Si and n-μc-Si films.
4750
B.R. Wu et al. / Thin Solid Films 517 (2009) 4749–4752
Table 1 Deposition parameters of Si thin films by HWCVD used in this study for Si heterojunction solar cells. Deposition parameters
H-treatment
i-a-Si
n-μc-Si
Wire temperature (°C) Substrate temperature (°C) Pressure (mTorr) Process gas flow (sccm)
1650 150 200 H2 = 80
1650 250 100 SiH4 = 5
1750 200 100 SiH4 = 2 H2 = 17.82–17.98 PH3 = 0.02–0.18
After the n-μc-Si emitter deposition, a laser doping technique was used to improve the ITO/n-μc-Si contact performance. A phosphorusdoped SiO2 solution was prepared as a dopant source using a sol-gel process. It was spun on the n-μc-Si emitter film, prebaked at 80 °C for 20 min and then postbaked at 200°C for 1 h under atmosphere. An NdYAG laser (wavelength: 355 nm; repetition rate: 10 Hz) with various power densities (64 to 318 mJ/cm2, beam size = 7.85 × 10− 3 cm2) was scanned through a metallic mask with periodic circular hole to make the heavily doped selective emitter. The hole diameter is kept at 0.5 mm with different spacings from 0.625 to 5 mm. After the laser irradiation process, the residual dopant source was removed using a wet etching process. Then the ITO front contact (80 nm thick) and Ag grid contact (1 μm thick) were successively deposited on the emitter film using an electron beam evaporator. Finally, the cell structure (Ag grid/ITO/n-μc-Si selective emitter/i-a-Si/textured p-c-Si/Alelectrode) was fabricated and shown in Fig. 1. The cell characteristics (area: 2 cm × 2 cm) were obtained through the current–voltage (I–V) measurements at100 mW/cm2 with an AM 1.5-like spectrum. 3. Results and discussion The control of the dopant concentration in the emitter layer is the key point in the photovoltaic properties of a HJ cell since it determines the equilibrium between defect density and position of Fermi level at the interface [16]. The first step in the present work was the deposition of singly doped films in order to determine the doping efficiency of phosphine in our system. Here the energy level at absorption coefficient of 104 cm− 1 (E04) of the n-μc-Si film was obtained from the transmission spectrum [17]. The carrier concentration (Nc) of the doped emitter film was determined using the Van der Pauw-Hall measurement. Fig. 2 shows the dependence of E04 and Nc of the n-μc-Si films on the PH3 dilution ratio (RPH3). It is found that both the E04 and Nc are strongly dependent on the RPH3. When the RPH3 increases from 0.1 to 0.9%, the Nc increases rapidly from 5 × 1015
Fig. 1. Schematic cross-section (a) and fabrication process (b) of Si heterojunction solar cell used in this study.
Fig. 2. Variation of optical band gap (E04) and carrier concentration (Nc) of n-type μc-Si emitter layer as functions of PH3 dilution ratio [RPH3 = PH3 / (SiH4 + H2 + PH3)].
to 3.7 × 1019 cm− 3, indicating a good PH3 doping efficiency in the emitter layer. On the other hand, the E04 first increases intensively from 1.81 to 2.01 eV when the RPH3 increases from 0.1 to 0.3% and then gradually increases to 2.08 eV when the RPH3 increases to 0.9%. The photovoltaic parameters (Jsc, Voc, FF, and η) of the HJ cells as functions of RPH3 in the n-μc-Si emitter are shown in Fig. 3. Apparently, the cell performance can be improved when the RPH3 increases from 0.1 to 0.5%. An optimum η value of 13.25% was achieved under a RPH3 of 0.5%. It was found that both the Jsc and FF values dropped when the RPH3 increased from 0.5 to 0.9%. These results indicate that a decrease of the doping concentration in the n-μc-Si emitter allows a shift of the Fermi level toward midgap and thus a reduction of the electric field at the interface. Under higher phosphorus doping levels, the increase of defect density in the emitter could cause recombination and a dramatic reduction of photovoltaic properties of the HJ cell [16]. In order to improve the ITO/emitter front contact without using the higher PH3 doping concentration, a selectively doped emitter via a metal mask was performed using the laser doping technique. The effect of laser irradiation density on the photovoltaic parameters of the HJ cell is illustrated in Fig. 4. As shown in this figure, all cells were
Fig. 3. Photovoltaic characteristics of Si heterojunction solar cells (Ag grid/ITO/n-μc-Si emitter/i-a-Si/p-c-Si/Al electrode) fabricated under various PH3 dilution ratio (RPH3) using HWCVD.
B.R. Wu et al. / Thin Solid Films 517 (2009) 4749–4752
4751
Fig. 5. Current–voltage characteristics of Si heterojunction solar cells with and without selective emitter measured at 100 mW/cm2 with an AM 1.5-like spectrum. Fig. 4. Photovoltaic characteristics of Si heterojunction solar cells (Ag grid/ITO/n-μc-Si selective emitter/i-a-Si/p-c-Si/Al electrode) fabricated under various laser power densities (Plaser).
fabricated under the RPH3 of 0.5% and treated by the laser scan with the hole spacing of 5 mm. It was found that the Jsc, FF and η data increased as the laser power density (Plaser) increased from 64 to 255 mJ/cm2. Further increasing the laser power density from 255 to 318 mJ/cm2 yielded the degraded Jsc, FF and η values. However, the Voc value decreased monotonously when the Plaser increased from 64 to 318 mJ/ cm2, especially for Plaser ≥ 255 mJ/cm2. It could be due to the enhanced cross doping problem at the film/substrate interface. Another possible reason for this could be attributed to the laser-induced damage in the i-a-Si and n-μc-Si films. The reason is still not clear. Nevertheless, a maximum conversion efficiency of 14.02% can be achieved for the Si HJ solar cell under a Plaser of 255 mJ/cm2. The result reveals that the selectively doped emitter structure can improve the ITO/n-μc-Si contact property, thus enhancing the μc-Si/c-Si HJ cell performance. Further improvement in the ITO/n-μc-Si contact can be performed by optimizing the hole spacing (D) of the metallic mask during the laser doping process. The D effect on the performance of the Si HJ solar cells is summarized in Table 2 under a fixed Plaser of 255 mJ/cm2. The data reported here were taken from the average measured value of five cell samples. It was found that the Jsc, FF, and η values increased as the D decreased from 5 to 1.25 mm. This could be due to the sufficient doping area, lowering the ITO/emitter contact resistance. Further reducing the D value (e.g. D = 0.625 mm) resulted in an inferior Voc value and cell performance. It could be attributed to the emitter damage caused by the heavily doped concentration. However, the Si HJ cells with selective emitters (D = 5 to 1.25 mm) show the equivalent or better efficiencies than that of the original cell. A best conversion efficiency of 14.31% is achieved for the selective-emitter μc-Si/c-Si HJ cell under the D value of 1.25 mm and Plaser of 255 mJ/cm2. Since the area of the selective emitter is changing, it is reasonable to suspect that the grain sizes will also have an effect on the formation of the emitters. However, from our
Table 2 Photovoltaic properties of the Ag grid/ITO/n-μc-Si selective emitter/i-a-Si/ p-c-Si/Al solar cell (area: 2 cm × 2 cm). Spacing (mm)
η (%)
Voc (V)
Jsc (mA/cm2)
FF (%)
Original cell 5 2.5 1.25 0.625
13.25 14.02 14.11 14.31 11.57
0.583 0.566 0.562 0.550 0.532
30.61 31.88 32.19 33.25 30.32
74.2 77.7 78.0 78.3 71.7
An Nd-YAG laser was scanned through a metallic mask with periodic circular hole to make the heavily doped selective emitter where the hole diameter is kept at 0.5 mm with different spacings (D) from 0.625 to 5 mm.
experimental data, the laser-treated emitter (Plaser = 255 mJ/cm2, D = 5 mm, no phosphorus spin coating) showed nearly the same in the conversion efficiency as compared with that of the original cell (without laser treatment). This exhibits that the laser treatment on the HWCVD μc-Si film has less effect on the contact resistance of the ITO/emitter in the Si heterojunction cell. Fig. 5 compares the I–V characteristics for the HJ cells with and without the selective emitter structure under AM1.5 illumination. It is worthy to note that the sample with selective-emitter structure shows the better Jsc, and FF but a lower Voc (550 mV as compared to 583 mV). The conversion efficiency can be improved from 13.25% to 14.31% via a suitable selective laser doping process. Moreover, the series and shunt resistances of the samples were determined using the dark I–V data. The series resistances were 0.23 Ω and 0.68 Ω and the shunt resistances were 1.78 × 105 Ω and 1.81 × 105 Ω for the best cell with and without the selective emitter process, respectively. These results show the improvement of the cells performance with the change of contact characteristics. It suggests that the proposed selective-emitter structure is very promising for future fabrication of high-efficiency HJ solar cells. 4. Conclusion A cell structure of Ag grid/ITO/n-μc-Si selective emitter/i-a-Si/ textured p-c-Si/Al-electrode was fabricated using a combination of HWCVD and laser doping techniques. The effects of selectively doped emitter on the characteristics of μc-Si/c-Si HJ solar cells have been studied. The initial efficiency up to 13.25% can be obtained for the Si HJ cells under a PH3 dilution rate of 0.5%. By using the selective-emitter structure, the cell properties were improved. The optimization of the selective-emitter layer was conducted in terms of Nc, Plaser, and D. Our best selective-emitter n-μc-Si/p-c-Si HJ cell (area: 2 cm × 2 cm) shows an efficiency of 14.31% with a Voc of 550 mV, a Jsc of 33.25 mA/cm2, and an FF of 78.3%. This is a promising result for the fabrication of lowcost and high-efficiency Si HJ solar cells. Acknowledgment This work was supported by the National Science Council of the Republic of China under contract no. NSC-96-2628-E-005-090-MY3. References [1] M. Tucci, G.D. Cesare, J. Non-Cryst. Solids 338–340 (2004) 663. [2] E. Centurioni, D. Iencinella, R. Rizzoli, F. Zignani, IEEE Trans. Electron Devices 51 (2004) 1818. [3] B. Jagannathan, W.A. Anderson, Sol. Energy Mater. Sol. Cells 44 (1996) 165.
4752
B.R. Wu et al. / Thin Solid Films 517 (2009) 4749–4752
[4] H.M. Branz, C.W. Teplin, D.L. Young, M.R. Page, E. Iwaniczko, L. Roybal, R. Bauer, A.H. Mahan, Y. Xu, P. Stradins, T. Wang, Q. Wang, Thin Solid Films 516 (2008) 743. [5] H.S. Povolny, X. Deng, Thin Solid Films 430 (2003) 125. [6] E. Iwaniczko, Y. Xu, R.E.I. Schropp, A.H. Mahan, Thin Solid Films 430 (2003) 212. [7] T.H. Wang, E. Iwaniczko, M.R. Page, D.H. Levi, Y. Yan, H.M. Branz, Q. Wang, Thin Solid Films 501 (2006) 284. [8] S.S. Lee, M.S. Ko, C.S. Kim, N.M. Hwang, J. Cryst. Growth 310 (2008) 3659. [9] J. Damon-Lacoste, P. Roca i Cabarrocas, P. Chatterjee, Y. Veschetti, A.S. Gudovskikh, J.P. Kleider, P.J. Ribeyrone, J. Non-Cryst. Solids 352 (2006) 1928. [10] Y. Xu, Z. Hu, H. Diao, Y. Cai, S. Zhang, X. Zeng, H. Hao, X. Liao, E. Fortunato, R. Martins, J. Non-Cryst. Solids 352 (2006) 1972.
[11] D.S. Wuu, S.Y. Lien, H.Y. Mao, B.R. Wu, I.C. Hsieh, P.C. Yao, J.H. Wang, W.C. Chen, Thin Solid Films 498 (2006) 9. [12] A. Masuda, H. Matsumura, Thin Solid Films 395 (2001) 112. [13] M.W.M. van Cleef, J.K. Rath, F.A. Rubinelli, C.H.M. van dert Werf, R.E.I. Schropp, W.F. van der Weg, J. Appl. Phys. 82 (1997) 6089. [14] S.K. Kim, J.C. Lee, S.J. Park, Y.J. Kim, K.H. Yoon, Sol. Energy Mater. Sol. Cells 92 (2008) 298. [15] S.Y. Lien, B.R. Wu, J.C. Liu, D.S. Wuu, Thin Solid Films 516 (2008) 747. [16] M. Tucci, Sol. Energy Mater. Sol. Cells 57 (1999) 249. [17] K.D. Mackenzie, J.R. Eggert, D.J. Leoplod, Y.M. Li, S. Lin, W. Paul, Phys. Rev. B31 (1985) 2198.