- Email: [email protected]
Fundamental understanding, impact, and removal of boron-rich layer on n-type silicon solar cells
Solar Energy Materials & Solar Cells 146 (2016) 58–62
Contents lists available at ScienceDirect
Solar Energy Materials & Solar Cells journal homepage: www.elsevier.com/locate/solmat
Fundamental understanding, impact, and removal of boron-rich layer on n-type silicon solar cells Kyungsun Ryu a, Chel-Jong Choi b, Hyomin Park c, Donghwan Kim c, Ajeet Rohatgi a, Young-Woo Ok a,n a
School of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA 30332-0250, USA School of Semiconductor and Chemical Engineering, Chonbuk National University, Jeonju 561-756, Republic of Korea c Department of Materials Science and Engineering, Korea University, Seoul 136-713, Republic of Korea b
art ic l e i nf o
a b s t r a c t
Article history: Received 21 August 2015 Received in revised form 16 November 2015 Accepted 18 November 2015
Most boron diffusion technologies result in the formation of an undesirable boron-rich layer (BRL) on the emitter surface. This paper reports on a study of the impact of gradual etching of the BRL on n-type silicon solar cell performance. It is found that gradual removal of the BRL improves surface passivation and bulk lifetime in the finished cell, while over-etching of the BRL results in a sharp decrease in fill factor due to the increased n-factor and series resistance. It is shown that the optimum chemical etching of the BRL formed as a byproduct of the screen-printed boron emitter diffusion used in this study raised the cell efficiency by 0.5%, resulting in 20.0% efficient large area (239 cm2) n-type solar cells. The change in BRL thickness and morphology as a function of chemical etching time was investigated by TEM and AES measurements to explain the quantitative impact of BRL removal on cell performance. & 2015 Elsevier B.V. All rights reserved.
1. Introduction Current market share of n-type mono-crystalline silicon (Si) for solar cells is only 5%, however, it has become an active area of investigation in photovoltaics (PV) because of several key advantages over its counterpart p-type Si cells for terrestrial applications. For example, n-type Si has (a) higher tolerance to detrimental metal impurities because of lower capture cross section for holes [1], (b) higher minority carrier lifetime (4 1 ms), and (c) absence of light-induced degradation (LID) [2–4] which is known to degrade solar cell efficiency of boron (B) doped p-type Si solar cells by 0.5–1% (absolute) [5]. However, formation of lowcost high performance B emitter has been a barrier for the growth of n-type solar cells partly because of the challenge in achieving effective passivation of B emitter and the formation of boron-rich layer (BRL) during the diffusion. Most current B diffusion processes used to fabricate n-type solar cell result in the formation of undesirable BRL which is located between the borosilicate glass (BSG) and p þ region [6–9]. The BRL is known to act as high recombination site and may interfere with surface passivation [9–11]. It has been reported that the BRL can also cause degradation of bulk lifetime due to crystal defects resulting from the different thermal expansion coefficients n
Corresponding author. E-mail address: [email protected] (Y.-W. Ok).
http://dx.doi.org/10.1016/j.solmat.2015.11.031 0927-0248/& 2015 Elsevier B.V. All rights reserved.
between BRL and Si, which occurs during the cooling down [10,11]. Since both surface passivation and bulk lifetime are the key to high efficiency solar cells, BRL needs to be removed to attain high efficiency n-type solar cells. Thermal oxidation is often used to remove the BRL, but it causes degradation in bulk lifetime due to the injection of impurities from the BRL into bulk [9,12]. Kessler et. al. reported that carrier lifetime degradation can be avoided either by limiting the process temperature to 850 °C and thus preventing BRL formation or through reconverting the BRL by a drive-in step in oxidizing atmosphere [11]. Recently, we have shown that chemical etching treatment can effectively remove BRL without bulk lifetime degradation in the case of B emitter formed by screen printing paste and it can also improve passivation quality of implanted B emitter [9,13–15]. However, above reports have explored only the two extreme cases: (1) cells with full BRL intact and (2) cells with BRL completely removed. The impact of gradual etching of the BRL on cell performance has not been investigated systematically to provide a better understanding of the morphology, degradation mechanism, and impact of BRL thickness on cell efficiency. Thus, this paper talks about fundamental understanding, impact, and removal of BRL to attain well passivated high efficiency n-type Si solar cells. In this paper, n-type PERT (passivated emitter, rear totally-diffused) solar cells were fabricated with controlled wet chemical etching of BRL to quantify the impact of partial and complete removal of the BRL on cell performance. In order to investigate the impact of impurities left in the partially etched BRL, a cell process
K. Ryu et al. / Solar Energy Materials & Solar Cells 146 (2016) 58–62
was developed to partially etch the BRL first after the B diffusion followed by consumption of the remaining BRL by thermal oxidation to inject the remaining impurities in the BRL into the bulk. Detailed electrical and optical measurements, transmission electron microscopy (TEM), and auger electron spectroscopy (AES) measurement were performed to examine the morphology, thickness, and composition of the partially removed BRL and its impact on solar cell parameters.
2. Experiment Large area (239 cm2) n-type PERT solar cells were fabricated on 170 μm-thick and 5 Ω cm resistivity n-type Si wafers with the process sequence described in Fig. 1. Both surfaces of the wafers were randomly textured with upright pyramids followed by RCA cleaning. A liquid B paste was screen-printed on the entire front side followed by a drying step at a temperature of 200 °C. The samples were then annealed at a temperature of 1000 °C for 10 min to form a B emitter. After the B annealing, rear surface planarization was performed by protecting the front surface by PECVD silicon nitride (SiNx) film followed by potassium hydroxide (KOH) etching for back planarization and subsequent removal of SiNx protective layer by hydrofluoric acid (HF) dip. Then the wafers received the chemical etching treatment [13] for different times ranging from 0 to 8 min. The mixture of HF, acetic acid (CH3COOH), and nitric acid (HNO3) (1:100:100) was used for the etching. Phosphorus (P) dopants were implanted on the entire rear side followed by 840 °C anneal/oxidation for 60 min to form n þ back surface field. During the implant anneal, a thin thermal Si oxide layer ( 10 nm on front and 30 nm on back) was grown. This many not be enough to consume full BRL layer without etching. This oxidation step injects impurities contained in the remaining BRL into the bulk. Next, an appropriate thickness of PECVD SiNx layer was deposited on the front and back SiO2 to serve as antireflective coating and effective passivation. The front grid pattern and point contacts on the back were screen-printed using Ag/Al and Ag paste, respectively, followed by a rapid co-firing step at 700 °C. Finally, a Ag reflector was printed on top of the rear screen-printed point contacts followed by a short low temperature ( 200 °C) drying step.
59
3. Results and discussions Light I–V characteristics of the n-type PERT Si solar cells with different BRL etching time are summarized in Table 1. The opencircuit voltage (Voc), short-circuit current density (Jsc), fill factor (FF), and efficiency (η) are plotted in Fig. 2 as a function of BRL etching time to gain a better understanding of the trend and impact of BRL thickness on cell parameter. Note that as the etching time increases, Jsc increases linearly while Voc increases slowly up to an etching time of 4 min and after that the cell performance starts to decrease. The FF remains essentially unchanged until 4 min etching, and after that it decreases sharply. This behavior resulted in maximum cell efficiency at 4 min BRL etching time for the screen-printed B emitter used in this study. Sharp drop in cell efficiency and FF beyond 4 min etching time can be explained by detailed series resistance (Rs) analysis using the methodology outline by D. L. Meier [16]. The TLM pattern analysis showed that contact resistance is very low until 6 min BRL etching time, but increased rapidly after 8 min etching time. It clearly shows that beyond 6 min etching, both sheet resistance and contact resistance start to increase Rs because over-etching starts to reduce emitter surface concentration. The over-etching also showed an increase in the ideality factor n from 1.06 to 1.14 which can be caused by junction leakage due to metal spiking when firing the screen-printed contacts through thinner emitter [17,18]. Finally, calculated FF values including the effect of Rs, Rshunt and n-factor, are very consistent with the measured FF. (Table 2). In addition, emitter saturation current density underneath the metal contact region (J0e,met) increases due to higher sheet resistance emitter, resulting in the decrease in Voc beyond 4 min Table 1 Summary of light I–V results of n-type PERT solar cells with different etching times Etching time (min)
Voc[mV]
Jsc[mA/cm2]
FF [%]
η [%]
0 2 4 6 8
644 646 649 646 641
38.18 38.34 38.53 38.66 38.86
79.5 79.5 79.8 78.9 77.7
19.5 19.6 20.0 19.7 19.4
Fig.1. Schematic structure of n-type PERT (passivated emitter, rear totally-diffused) cell and the process flow.
60
K. Ryu et al. / Solar Energy Materials & Solar Cells 146 (2016) 58–62
Fig.2. Light I–V characteristics of n-type PERT solar cells with different etching times.
Table 2 Summary of Rs analysis from n-type PERT solar cells with different etching times. Etching time
Sheet resistance (Ω/sq.)
Rsheet (Ω cm2)
Rgrid (Ω cm2)
Rcontact (Ω cm2)
Rsub. þRback (Ω cm2)
Total Rs (Ω cm2)
Measured Rshunt (Ω cm2)
Measured n factor
Calculated FF
0 2 4 6 8
39 47 58 73 112
0.08 0.1 0.12 0.15 0.23
0.35 0.35 0.35 0.35 0.35
0.01 0.02 0.02 0.05 0.12
0.15 0.15 0.15 0.15 0.15
0.59 0.62 0.64 0.7 0.85
4127 5755 8014 6518 7103
1.06 1.06 1.06 1.13 1.14
0.797 0.797 0.797 0.785 0.775
etching [19,20].This is the reason why cell Voc decreased after the over-etching ( 44 min). Internal quantum efficiency (IQE) measurements in Fig. 3 reveal that the initial rise in Jsc with the increased etching time is due to the increase in both short and long wavelength response of the cells. The increased IQE response in the short wavelength (o600 nm) is attributed to improved passivation or reduced front surface recombination velocity (FSRV) in field region. The increase in long wavelength response (4600 nm) is attributed to higher bulk lifetime due to reduced impurity injection from the thinner BRL. To understand the trend of IQE, we measured J0e using symmetric test structures (SiNx/SiO2/p þ np þ /SiO2/SiNx) after simulated firing. J0e value decreased from 128 to 95 fA/cm2 and also τbulk increased gradually with the increase BRL etching time. The reduced FSRV supports that BRL interferes with the B emitter passivation, therefore, gradual removal of BRL gradually improves surface passivation. The gradual increase in bulk lifetime with the reduced BRL thickness can be explained by reduced impurity content in the BRL, resulting in reduced injection of impurities into Si bulk during the subsequent oxidation. Based on our experimental results, it is clear that BRL is a source of recombination and impurities which can be injected into Si bulk if the BRL is consumed by oxidation [9,21]. After 4 min etching or complete BRL removal, there is little change in the IQE response. Note that the Jsc
Fig.3. IQE response of n-type PERT solar cells with different etching times.
continues to increase even after the complete BRL removal. This is because further etching starts to shave the B emitter. This results in a shallow emitter with lower B surface concentration (Table 2). This reduces auger recombination and heavy doping effect in the emitter, resulting in an improved blue response and Jsc. To investigate the thickness, composition, and morphological change in the BRL with the increased etching time, TEM measurements were performed. Fig. 4 shows the TEM cross section of
K. Ryu et al. / Solar Energy Materials & Solar Cells 146 (2016) 58–62
61
Fig.5. AES depth profiling analysis of the samples (a) without etching and (b) with 4 min etching.
4. Conclusions Fig.4. Cross-sectional TEM images of screen-printed B emitter with different etching times for (a) 0 min, (b) 2 min, and (c) 4 min.
screen-printed B emitter prior to BRL etching and after 2 and 4 min BRL etching. Fig. 4(a) shows that BRL is present on top of the p þ emitter in the form of dark and dense agglomerations of different size. After 2 min partial etching, the smaller agglomerated regions and particles are removed (Fig. 4(b)), but some agglomerations still remain on the p þ emitter. Finally, BRL including all the remaining agglomerations are completely removed after 4 min etching (Fig. 4(c)). These TEM images support that 4 min etching time is sufficient and optimum to remove the BRL completely for our screen-printed B emitter. After that it starts to etch the B emitter, resulting in higher sheet resistance and degradation in FF and efficiency. The chemical composition of the BRL/p þ Si interface in the above samples, Fig. 5(a) and (c), was measured by AES depth profiling. To delineate the interface of the samples, about 10 nm thick SiNx film was deposited on the samples prior to the measurement. The sample without etching showed the excess B peak at the Si interface (Fig. 5(a)), indicative of the presence of BRL. However, no B peak was detected at the interface in the sample after 4 min etching (Fig. 5(b)). TEM and AES measurements demonstrate that the BRL region is agglomerated and non-uniform with high B concentration and this can be removed by controlled chemical etching. However, these techniques were unable to detect harmful impurities in the bulk or emitter region before and after annealing.
In summary, n-type PERT solar cells with different etching times for BRL removal were fabricated and analyzed to investigate the impact of gradual removal of BRL on the device performances. When the BRL is gradually removed by the chemical etching, cell efficiency increases initially due to the reduced FSRV and increased bulk lifetime. When the BRL is completely removed after 4 min etching, without hurting p þ emitter region, a maximum cell efficiency of 20.0% was achieved for this process sequence and PERT cell structure. However, over-etching of the BRL starts to increase emitter sheet resistance, Rs, and n-factor, resulting in the significant reduction in FF and efficiency. Optimum removal of BRL raised the cell efficiency by 0.5% in this study. The morphological and compositional analysis of the BRL by TEM and AES measurements showed presence of dark agglomerations in the BRL which could be indicative of impurities and defects. Impurity content or agglomerations are reduced when the BRL is gradually thinned. This explains the reason for inferior surface passivation and bulk lifetime degradation when the BRL is present and removed or consumed by thermal oxidation.
Acknowledgment This research was supported by Basic Science Research Program (NRF-2015R1A6A1A04020421) through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Republic of Korea, and by the Converging Research Center Program (2014M3C1A8048834) through the Ministry of Science, ICT & Future Planning, Republic of Korea.
62
K. Ryu et al. / Solar Energy Materials & Solar Cells 146 (2016) 58–62
References [1] D. Macdonald, L. Geerligs, Recombination activity of interstitial iron and other transition metal point defects in p-and n-type crystalline silicon, Appl. Phys. Lett. 85 (2004) 4061–4063. [2] A. Cuevas, M.J. Kerr, C. Samundsett, F. Ferrazza, G. Coletti, Millisecond minority carrier lifetimes in n-type multicrystalline silicon, Appl. Phys. Lett. 81 (2002) 4952–4954. [3] K. Bothe, R. Sinton, J. Schmidt, Fundamental boron–oxygen-related carrier lifetime limit in mono- and multicrystalline silicon, Prog. Photovolt. Res. Appl. 13 (2005) 287–296. [4] J.E. Cotter, J.H. Guo, P.J. Cousins, M.D. Abbott, F.W. Chen, K.C. Fisher, P-type versus n-type silicon wafers: prospects for high-efficiency commercial silicon solar cells, IEEE Trans. Electron Devices 53 (2006) 1893–1901. [5] A. Das, A. Rohatgi, The impact of cell design on light induced degradation in ptype silicon solar cells, in: Proceedings of Photovoltaic Specialists Conference (PVSC), 2011 37th IEEE, 2011, pp. 000158–000164. [6] P. Negrini, A. Ravaglia, S. Solmi, Boron predeposition in silicon using BBr3, J. Electrochem. Soc. 125 (1978) 609–613. [7] S.M. Myers, M. Seibt, W. Schroter, Mechanisms of transition-metal gettering in silicon, J. Appl. Phys. 88 (2000) 3795–3819. [8] H. Talvitie, V. Vähänissi, A. Haarahiltunen, M. Yli-Koski, H. Savin, Phosphorus and boron diffusion gettering of iron in monocrystalline silicon, J. Appl. Phys. 109 (2011) 093505–093505-5. [9] K. Ryu, A. Upadhyaya, V. Upadhyaya, A. Rohatgi, Y.-W. Ok, High efficiency large area n-type front junction silicon solar cells with boron emitter formed by screen printing technology, Prog. Photovolt. Res. Appl. 23 (2015) 119–123. [10] P.J. Cousins, J.E. Cotter, The influence of diffusion-induced dislocations on high efficiency silicon solar cells, Electron Devices IEEE Trans. 53 (2006) 457–464. [11] M.A. Kessler, T. Ohrdes, B. Wolpensinger, N.-P. Harder, Charge carrier lifetime degradation in Cz silicon through the formation of a boron-rich layer during BBr3 diffusion processes, Semicond. Sci. Technol. 25 (2010) 055001. [12] P. Sieu Pheng, L. Wensheng, B. Wolpensinger, M.A. Kessler, D. Macdonald, Tradeoffs between impurity gettering, bulk degradation, and surface
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
passivation of boron-rich layers on silicon solar cells, photovoltaics, IEEE J. 3 (2013) 261–266. K. Ryu, A. Upadhyaya, H.-J. Song, C.-J. Choi, A. Rohatgi, Y.-W. Ok, Chemical etching of boron-rich layer and its impact on high efficiency n-type silicon solar cells, Appl. Phys. Lett. 101 (2012). Y.-W. Ok, A.D. Upadhyaya, Y. Tao, F. Zimbardi, K. Ryu, M.-H. Kang, A. Rohatgi, Ion-implanted and screen-printed large area 20% efficient n-type front junction Si solar cells, Sol. Energy Mater. Sol. Cells 123 (2014) 92–96. Y. Tao, A. Payne, V.D. Upadhyaya, A. Rohatgi, 20.7% efficient ion-implanted large area n-type front junction silicon solar cells with rear point contacts formed by laser opening and physical vapor deposition, Prog. Photovolt. Res. Appl. 22 (2014) 1030–1039. D.L. Meier, D.K. Schroder, Contact resistance: its measurement and relative importance to power loss in a solar cell, Electron Devices IEEE Trans. 31 (1984) 647–653. Y. Tao, Y.-W. Ok, F. Zimbardi, A. Upadhyaya, J. Lai, S. Ning, V. Upadhyaya, A. Rohatgi, Fully Ion-implanted and screen-printed 20.2% efficient front junction silicon cells on 239 cm2 n-type CZ substrate, photovoltaics, IEEE J. 4 (2014) 58–63. M.M. Hilali, A. Rohatgi, B. To, A review and understanding of screen-printed contacts and selective-emitter formation, in: Proceedings of the 14th workshop on crystalline silicon solar cells and modules,2004, pp. 80401–83393. U. Jä ger, S. Mack, A. Kimmerle, A. Wolf, R. Preu, Influence of doping profile of highly doped regions for selective emitter solar cells, in: Proceedings of the Photovoltaic Specialists Conference (PVSC), 2010 35th IEEE, 2010, pp. 003185– 003189. A. Cuevas, P.A. Basore, G. Giroult‐Matlakowski, C. Dubois, Surface recombination velocity of highly doped n‐type silicon, J. Appl. Phys. 80 (1996) 3370–3375. S.P. Phang, D. Macdonald, Direct comparison of boron, phosphorus, and aluminum gettering of iron in crystalline silicon, J. Appl. Phys. 109 (2011) 073521.
Contents lists available at ScienceDirect
Solar Energy Materials & Solar Cells journal homepage: www.elsevier.com/locate/solmat
Fundamental understanding, impact, and removal of boron-rich layer on n-type silicon solar cells Kyungsun Ryu a, Chel-Jong Choi b, Hyomin Park c, Donghwan Kim c, Ajeet Rohatgi a, Young-Woo Ok a,n a
School of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA 30332-0250, USA School of Semiconductor and Chemical Engineering, Chonbuk National University, Jeonju 561-756, Republic of Korea c Department of Materials Science and Engineering, Korea University, Seoul 136-713, Republic of Korea b
art ic l e i nf o
a b s t r a c t
Article history: Received 21 August 2015 Received in revised form 16 November 2015 Accepted 18 November 2015
Most boron diffusion technologies result in the formation of an undesirable boron-rich layer (BRL) on the emitter surface. This paper reports on a study of the impact of gradual etching of the BRL on n-type silicon solar cell performance. It is found that gradual removal of the BRL improves surface passivation and bulk lifetime in the finished cell, while over-etching of the BRL results in a sharp decrease in fill factor due to the increased n-factor and series resistance. It is shown that the optimum chemical etching of the BRL formed as a byproduct of the screen-printed boron emitter diffusion used in this study raised the cell efficiency by 0.5%, resulting in 20.0% efficient large area (239 cm2) n-type solar cells. The change in BRL thickness and morphology as a function of chemical etching time was investigated by TEM and AES measurements to explain the quantitative impact of BRL removal on cell performance. & 2015 Elsevier B.V. All rights reserved.
1. Introduction Current market share of n-type mono-crystalline silicon (Si) for solar cells is only 5%, however, it has become an active area of investigation in photovoltaics (PV) because of several key advantages over its counterpart p-type Si cells for terrestrial applications. For example, n-type Si has (a) higher tolerance to detrimental metal impurities because of lower capture cross section for holes [1], (b) higher minority carrier lifetime (4 1 ms), and (c) absence of light-induced degradation (LID) [2–4] which is known to degrade solar cell efficiency of boron (B) doped p-type Si solar cells by 0.5–1% (absolute) [5]. However, formation of lowcost high performance B emitter has been a barrier for the growth of n-type solar cells partly because of the challenge in achieving effective passivation of B emitter and the formation of boron-rich layer (BRL) during the diffusion. Most current B diffusion processes used to fabricate n-type solar cell result in the formation of undesirable BRL which is located between the borosilicate glass (BSG) and p þ region [6–9]. The BRL is known to act as high recombination site and may interfere with surface passivation [9–11]. It has been reported that the BRL can also cause degradation of bulk lifetime due to crystal defects resulting from the different thermal expansion coefficients n
Corresponding author. E-mail address: [email protected] (Y.-W. Ok).
http://dx.doi.org/10.1016/j.solmat.2015.11.031 0927-0248/& 2015 Elsevier B.V. All rights reserved.
between BRL and Si, which occurs during the cooling down [10,11]. Since both surface passivation and bulk lifetime are the key to high efficiency solar cells, BRL needs to be removed to attain high efficiency n-type solar cells. Thermal oxidation is often used to remove the BRL, but it causes degradation in bulk lifetime due to the injection of impurities from the BRL into bulk [9,12]. Kessler et. al. reported that carrier lifetime degradation can be avoided either by limiting the process temperature to 850 °C and thus preventing BRL formation or through reconverting the BRL by a drive-in step in oxidizing atmosphere [11]. Recently, we have shown that chemical etching treatment can effectively remove BRL without bulk lifetime degradation in the case of B emitter formed by screen printing paste and it can also improve passivation quality of implanted B emitter [9,13–15]. However, above reports have explored only the two extreme cases: (1) cells with full BRL intact and (2) cells with BRL completely removed. The impact of gradual etching of the BRL on cell performance has not been investigated systematically to provide a better understanding of the morphology, degradation mechanism, and impact of BRL thickness on cell efficiency. Thus, this paper talks about fundamental understanding, impact, and removal of BRL to attain well passivated high efficiency n-type Si solar cells. In this paper, n-type PERT (passivated emitter, rear totally-diffused) solar cells were fabricated with controlled wet chemical etching of BRL to quantify the impact of partial and complete removal of the BRL on cell performance. In order to investigate the impact of impurities left in the partially etched BRL, a cell process
K. Ryu et al. / Solar Energy Materials & Solar Cells 146 (2016) 58–62
was developed to partially etch the BRL first after the B diffusion followed by consumption of the remaining BRL by thermal oxidation to inject the remaining impurities in the BRL into the bulk. Detailed electrical and optical measurements, transmission electron microscopy (TEM), and auger electron spectroscopy (AES) measurement were performed to examine the morphology, thickness, and composition of the partially removed BRL and its impact on solar cell parameters.
2. Experiment Large area (239 cm2) n-type PERT solar cells were fabricated on 170 μm-thick and 5 Ω cm resistivity n-type Si wafers with the process sequence described in Fig. 1. Both surfaces of the wafers were randomly textured with upright pyramids followed by RCA cleaning. A liquid B paste was screen-printed on the entire front side followed by a drying step at a temperature of 200 °C. The samples were then annealed at a temperature of 1000 °C for 10 min to form a B emitter. After the B annealing, rear surface planarization was performed by protecting the front surface by PECVD silicon nitride (SiNx) film followed by potassium hydroxide (KOH) etching for back planarization and subsequent removal of SiNx protective layer by hydrofluoric acid (HF) dip. Then the wafers received the chemical etching treatment [13] for different times ranging from 0 to 8 min. The mixture of HF, acetic acid (CH3COOH), and nitric acid (HNO3) (1:100:100) was used for the etching. Phosphorus (P) dopants were implanted on the entire rear side followed by 840 °C anneal/oxidation for 60 min to form n þ back surface field. During the implant anneal, a thin thermal Si oxide layer ( 10 nm on front and 30 nm on back) was grown. This many not be enough to consume full BRL layer without etching. This oxidation step injects impurities contained in the remaining BRL into the bulk. Next, an appropriate thickness of PECVD SiNx layer was deposited on the front and back SiO2 to serve as antireflective coating and effective passivation. The front grid pattern and point contacts on the back were screen-printed using Ag/Al and Ag paste, respectively, followed by a rapid co-firing step at 700 °C. Finally, a Ag reflector was printed on top of the rear screen-printed point contacts followed by a short low temperature ( 200 °C) drying step.
59
3. Results and discussions Light I–V characteristics of the n-type PERT Si solar cells with different BRL etching time are summarized in Table 1. The opencircuit voltage (Voc), short-circuit current density (Jsc), fill factor (FF), and efficiency (η) are plotted in Fig. 2 as a function of BRL etching time to gain a better understanding of the trend and impact of BRL thickness on cell parameter. Note that as the etching time increases, Jsc increases linearly while Voc increases slowly up to an etching time of 4 min and after that the cell performance starts to decrease. The FF remains essentially unchanged until 4 min etching, and after that it decreases sharply. This behavior resulted in maximum cell efficiency at 4 min BRL etching time for the screen-printed B emitter used in this study. Sharp drop in cell efficiency and FF beyond 4 min etching time can be explained by detailed series resistance (Rs) analysis using the methodology outline by D. L. Meier [16]. The TLM pattern analysis showed that contact resistance is very low until 6 min BRL etching time, but increased rapidly after 8 min etching time. It clearly shows that beyond 6 min etching, both sheet resistance and contact resistance start to increase Rs because over-etching starts to reduce emitter surface concentration. The over-etching also showed an increase in the ideality factor n from 1.06 to 1.14 which can be caused by junction leakage due to metal spiking when firing the screen-printed contacts through thinner emitter [17,18]. Finally, calculated FF values including the effect of Rs, Rshunt and n-factor, are very consistent with the measured FF. (Table 2). In addition, emitter saturation current density underneath the metal contact region (J0e,met) increases due to higher sheet resistance emitter, resulting in the decrease in Voc beyond 4 min Table 1 Summary of light I–V results of n-type PERT solar cells with different etching times Etching time (min)
Voc[mV]
Jsc[mA/cm2]
FF [%]
η [%]
0 2 4 6 8
644 646 649 646 641
38.18 38.34 38.53 38.66 38.86
79.5 79.5 79.8 78.9 77.7
19.5 19.6 20.0 19.7 19.4
Fig.1. Schematic structure of n-type PERT (passivated emitter, rear totally-diffused) cell and the process flow.
60
K. Ryu et al. / Solar Energy Materials & Solar Cells 146 (2016) 58–62
Fig.2. Light I–V characteristics of n-type PERT solar cells with different etching times.
Table 2 Summary of Rs analysis from n-type PERT solar cells with different etching times. Etching time
Sheet resistance (Ω/sq.)
Rsheet (Ω cm2)
Rgrid (Ω cm2)
Rcontact (Ω cm2)
Rsub. þRback (Ω cm2)
Total Rs (Ω cm2)
Measured Rshunt (Ω cm2)
Measured n factor
Calculated FF
0 2 4 6 8
39 47 58 73 112
0.08 0.1 0.12 0.15 0.23
0.35 0.35 0.35 0.35 0.35
0.01 0.02 0.02 0.05 0.12
0.15 0.15 0.15 0.15 0.15
0.59 0.62 0.64 0.7 0.85
4127 5755 8014 6518 7103
1.06 1.06 1.06 1.13 1.14
0.797 0.797 0.797 0.785 0.775
etching [19,20].This is the reason why cell Voc decreased after the over-etching ( 44 min). Internal quantum efficiency (IQE) measurements in Fig. 3 reveal that the initial rise in Jsc with the increased etching time is due to the increase in both short and long wavelength response of the cells. The increased IQE response in the short wavelength (o600 nm) is attributed to improved passivation or reduced front surface recombination velocity (FSRV) in field region. The increase in long wavelength response (4600 nm) is attributed to higher bulk lifetime due to reduced impurity injection from the thinner BRL. To understand the trend of IQE, we measured J0e using symmetric test structures (SiNx/SiO2/p þ np þ /SiO2/SiNx) after simulated firing. J0e value decreased from 128 to 95 fA/cm2 and also τbulk increased gradually with the increase BRL etching time. The reduced FSRV supports that BRL interferes with the B emitter passivation, therefore, gradual removal of BRL gradually improves surface passivation. The gradual increase in bulk lifetime with the reduced BRL thickness can be explained by reduced impurity content in the BRL, resulting in reduced injection of impurities into Si bulk during the subsequent oxidation. Based on our experimental results, it is clear that BRL is a source of recombination and impurities which can be injected into Si bulk if the BRL is consumed by oxidation [9,21]. After 4 min etching or complete BRL removal, there is little change in the IQE response. Note that the Jsc
Fig.3. IQE response of n-type PERT solar cells with different etching times.
continues to increase even after the complete BRL removal. This is because further etching starts to shave the B emitter. This results in a shallow emitter with lower B surface concentration (Table 2). This reduces auger recombination and heavy doping effect in the emitter, resulting in an improved blue response and Jsc. To investigate the thickness, composition, and morphological change in the BRL with the increased etching time, TEM measurements were performed. Fig. 4 shows the TEM cross section of
K. Ryu et al. / Solar Energy Materials & Solar Cells 146 (2016) 58–62
61
Fig.5. AES depth profiling analysis of the samples (a) without etching and (b) with 4 min etching.
4. Conclusions Fig.4. Cross-sectional TEM images of screen-printed B emitter with different etching times for (a) 0 min, (b) 2 min, and (c) 4 min.
screen-printed B emitter prior to BRL etching and after 2 and 4 min BRL etching. Fig. 4(a) shows that BRL is present on top of the p þ emitter in the form of dark and dense agglomerations of different size. After 2 min partial etching, the smaller agglomerated regions and particles are removed (Fig. 4(b)), but some agglomerations still remain on the p þ emitter. Finally, BRL including all the remaining agglomerations are completely removed after 4 min etching (Fig. 4(c)). These TEM images support that 4 min etching time is sufficient and optimum to remove the BRL completely for our screen-printed B emitter. After that it starts to etch the B emitter, resulting in higher sheet resistance and degradation in FF and efficiency. The chemical composition of the BRL/p þ Si interface in the above samples, Fig. 5(a) and (c), was measured by AES depth profiling. To delineate the interface of the samples, about 10 nm thick SiNx film was deposited on the samples prior to the measurement. The sample without etching showed the excess B peak at the Si interface (Fig. 5(a)), indicative of the presence of BRL. However, no B peak was detected at the interface in the sample after 4 min etching (Fig. 5(b)). TEM and AES measurements demonstrate that the BRL region is agglomerated and non-uniform with high B concentration and this can be removed by controlled chemical etching. However, these techniques were unable to detect harmful impurities in the bulk or emitter region before and after annealing.
In summary, n-type PERT solar cells with different etching times for BRL removal were fabricated and analyzed to investigate the impact of gradual removal of BRL on the device performances. When the BRL is gradually removed by the chemical etching, cell efficiency increases initially due to the reduced FSRV and increased bulk lifetime. When the BRL is completely removed after 4 min etching, without hurting p þ emitter region, a maximum cell efficiency of 20.0% was achieved for this process sequence and PERT cell structure. However, over-etching of the BRL starts to increase emitter sheet resistance, Rs, and n-factor, resulting in the significant reduction in FF and efficiency. Optimum removal of BRL raised the cell efficiency by 0.5% in this study. The morphological and compositional analysis of the BRL by TEM and AES measurements showed presence of dark agglomerations in the BRL which could be indicative of impurities and defects. Impurity content or agglomerations are reduced when the BRL is gradually thinned. This explains the reason for inferior surface passivation and bulk lifetime degradation when the BRL is present and removed or consumed by thermal oxidation.
Acknowledgment This research was supported by Basic Science Research Program (NRF-2015R1A6A1A04020421) through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Republic of Korea, and by the Converging Research Center Program (2014M3C1A8048834) through the Ministry of Science, ICT & Future Planning, Republic of Korea.
62
K. Ryu et al. / Solar Energy Materials & Solar Cells 146 (2016) 58–62
References [1] D. Macdonald, L. Geerligs, Recombination activity of interstitial iron and other transition metal point defects in p-and n-type crystalline silicon, Appl. Phys. Lett. 85 (2004) 4061–4063. [2] A. Cuevas, M.J. Kerr, C. Samundsett, F. Ferrazza, G. Coletti, Millisecond minority carrier lifetimes in n-type multicrystalline silicon, Appl. Phys. Lett. 81 (2002) 4952–4954. [3] K. Bothe, R. Sinton, J. Schmidt, Fundamental boron–oxygen-related carrier lifetime limit in mono- and multicrystalline silicon, Prog. Photovolt. Res. Appl. 13 (2005) 287–296. [4] J.E. Cotter, J.H. Guo, P.J. Cousins, M.D. Abbott, F.W. Chen, K.C. Fisher, P-type versus n-type silicon wafers: prospects for high-efficiency commercial silicon solar cells, IEEE Trans. Electron Devices 53 (2006) 1893–1901. [5] A. Das, A. Rohatgi, The impact of cell design on light induced degradation in ptype silicon solar cells, in: Proceedings of Photovoltaic Specialists Conference (PVSC), 2011 37th IEEE, 2011, pp. 000158–000164. [6] P. Negrini, A. Ravaglia, S. Solmi, Boron predeposition in silicon using BBr3, J. Electrochem. Soc. 125 (1978) 609–613. [7] S.M. Myers, M. Seibt, W. Schroter, Mechanisms of transition-metal gettering in silicon, J. Appl. Phys. 88 (2000) 3795–3819. [8] H. Talvitie, V. Vähänissi, A. Haarahiltunen, M. Yli-Koski, H. Savin, Phosphorus and boron diffusion gettering of iron in monocrystalline silicon, J. Appl. Phys. 109 (2011) 093505–093505-5. [9] K. Ryu, A. Upadhyaya, V. Upadhyaya, A. Rohatgi, Y.-W. Ok, High efficiency large area n-type front junction silicon solar cells with boron emitter formed by screen printing technology, Prog. Photovolt. Res. Appl. 23 (2015) 119–123. [10] P.J. Cousins, J.E. Cotter, The influence of diffusion-induced dislocations on high efficiency silicon solar cells, Electron Devices IEEE Trans. 53 (2006) 457–464. [11] M.A. Kessler, T. Ohrdes, B. Wolpensinger, N.-P. Harder, Charge carrier lifetime degradation in Cz silicon through the formation of a boron-rich layer during BBr3 diffusion processes, Semicond. Sci. Technol. 25 (2010) 055001. [12] P. Sieu Pheng, L. Wensheng, B. Wolpensinger, M.A. Kessler, D. Macdonald, Tradeoffs between impurity gettering, bulk degradation, and surface
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
passivation of boron-rich layers on silicon solar cells, photovoltaics, IEEE J. 3 (2013) 261–266. K. Ryu, A. Upadhyaya, H.-J. Song, C.-J. Choi, A. Rohatgi, Y.-W. Ok, Chemical etching of boron-rich layer and its impact on high efficiency n-type silicon solar cells, Appl. Phys. Lett. 101 (2012). Y.-W. Ok, A.D. Upadhyaya, Y. Tao, F. Zimbardi, K. Ryu, M.-H. Kang, A. Rohatgi, Ion-implanted and screen-printed large area 20% efficient n-type front junction Si solar cells, Sol. Energy Mater. Sol. Cells 123 (2014) 92–96. Y. Tao, A. Payne, V.D. Upadhyaya, A. Rohatgi, 20.7% efficient ion-implanted large area n-type front junction silicon solar cells with rear point contacts formed by laser opening and physical vapor deposition, Prog. Photovolt. Res. Appl. 22 (2014) 1030–1039. D.L. Meier, D.K. Schroder, Contact resistance: its measurement and relative importance to power loss in a solar cell, Electron Devices IEEE Trans. 31 (1984) 647–653. Y. Tao, Y.-W. Ok, F. Zimbardi, A. Upadhyaya, J. Lai, S. Ning, V. Upadhyaya, A. Rohatgi, Fully Ion-implanted and screen-printed 20.2% efficient front junction silicon cells on 239 cm2 n-type CZ substrate, photovoltaics, IEEE J. 4 (2014) 58–63. M.M. Hilali, A. Rohatgi, B. To, A review and understanding of screen-printed contacts and selective-emitter formation, in: Proceedings of the 14th workshop on crystalline silicon solar cells and modules,2004, pp. 80401–83393. U. Jä ger, S. Mack, A. Kimmerle, A. Wolf, R. Preu, Influence of doping profile of highly doped regions for selective emitter solar cells, in: Proceedings of the Photovoltaic Specialists Conference (PVSC), 2010 35th IEEE, 2010, pp. 003185– 003189. A. Cuevas, P.A. Basore, G. Giroult‐Matlakowski, C. Dubois, Surface recombination velocity of highly doped n‐type silicon, J. Appl. Phys. 80 (1996) 3370–3375. S.P. Phang, D. Macdonald, Direct comparison of boron, phosphorus, and aluminum gettering of iron in crystalline silicon, J. Appl. Phys. 109 (2011) 073521.