Scholarly article on topic 'Universal Passivation for p++ and n++ Areas on IBC Solar Cells'

Universal Passivation for p++ and n++ Areas on IBC Solar Cells Academic research paper on "Nano-technology"

CC BY-NC-ND
0
0
Share paper
Academic journal
Energy Procedia
OECD Field of science
Keywords
{"amorphous silicon" / "universal passivation" / "IBC solar cells" / "interdigitated back contact" / "high efficiency" / "laser doping"}

Abstract of research paper on Nano-technology, author of scientific article — Kai Carstens, Morris Dahlinger, Erik Hoffmann, Jürgen R. Köhler, Renate Zapf-Gottwick, et al.

Abstract We present a universal passivation for both, phosphorus and boron doped surfaces of interdigitated back contact solar cells. The amorphous silicon layer shows excellent passivation on boron emitter and phosphorus back surface field as long as the deposition temperature is below T dep = 200°C. The amorphous silicon enables saturation current densities J o,em = 46 fA/cm2 for laser doped p++ boron emitters with sheet resistance R sh,em = 90Ω/sq as well as J o,bsf = 73 fA/cm2 for laser doped n++ phosphorus back surface fields with R sh,bsf°=°36°Ω/sq. Integration of the amorphous silicon low temperature passivation into our laser processed IBC solar cells yields a mean efficiency η = 22.8%. The open circuit voltage gain ΔV oc amounts to ΔV oc = 5mV compared to cells with thermal silicon dioxide rear passivation. The best solar cells achieve a confirmed efficiency η = 23.24%.

Academic research paper on topic "Universal Passivation for p++ and n++ Areas on IBC Solar Cells"

CrossMark

Available online at www.sciencedirect.com

ScienceDirect

Energy Procedia 77 (2015) 779 - 785

5th International Conference on Silicon Photovoltaics, SiliconPV 2015

Universal passivation for p++ and n++ areas on IBC solar cells

Kai Carstens*, Morris Dahlinger, Erik Hoffmann, Jürgen R. Köhler, Renate Zapf-Gottwick and Jürgen H. Werner

Institute for Photovoltaics, University of Stuttgart, Pfaffenwaldring 47, 70569 Stuttgart, Germany

Abstract

We present a universal passivation for both, phosphorus and boron doped surfaces of interdigitated back contact solar cells. The amorphous silicon layer shows excellent passivation on boron emitter and phosphorus back surface field as long as the deposition temperature is below Tdep = 200°C. The amorphous silicon enables saturation current densities Jo,em = 46 fA/cm2 for laser doped p++ boron emitters with sheet resistance Rsh,em = 90 Q/sq as well as Jo,bsf = 73 fA/cm2 for laser doped n++ phosphorus back surface fields with Rsh,bsf = 36 Q/sq. Integration of the amorphous silicon low temperature passivation into our laser processed IBC solar cells yields a mean efficiency ^ = 22.8%. The open circuit voltage gain AVoc amounts to AVoc = 5 mV compared to cells with thermal silicon dioxide rear passivation. The best solar cells achieve a confirmed efficiency ^ = 23.24%.

© 2015The Authors.PublishedbyElsevierLtd. Thisis an open access article under the CC BY-NC-ND license (http://creativecommons.Org/licenses/by-nc-nd/4.0/).

Peer review by the scientific conference committee of SiliconPV 2015 under responsibility of PSE AG

Keywords: amorphous silicon; universal passivation; IBC solar cells; interdigitated back contact; high efficiency; laser doping

1. Introduction

Aluminum oxide (AlOx) is the preferred choice for surface passivation of p-type doped silicon, due to a low interface state density and a high amount of negative fixed charges. In contrast to AlOx, silicon nitride (SiNx) passivates n-type surfaces well, due to positive fixed charges. In advanced silicon solar cell structures, like interdigitated back contact solar cells (IBC), the passivation of both n-type and p-type doped rear surface is necessary to achieve high efficiencies. The most common passivation for IBC-solar cells is a high temperature thermal silicon dioxide (SiO2), e.g. [1]. In contrast to SiO2 the full area application of a charged passivation layer,

* Corresponding author. Tel.: +49 (0)711 685-67161; fax: +49 (0)711 685-67143. E-mail address: kai.carstens@ipv.uni-stuttgart.de

1876-6102 © 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Peer review by the scientific conference committee of SiliconPV 2015 under responsibility of PSE AG doi:10.1016/j.egypro.2015.07.110

like AlOx or SiNx could lead to parasitic shunting [2], or at least passivate only one type of surface doping well. Consequently, either charged passivation layers need a structuring process or a universal layer passivating both n-type and p-type surfaces at the same time is necessary. Alternatives for such non-charged passivation layers are stacks of silicon oxide or hafnium oxide plus aluminum oxide to suppress the charges [3]. Furthermore, Dauwe and Seiffe showed that amorphous silicon based passivation is suitable for both low n- and p-type doping due to the missing fixed charges and low interface state densities [4,5]. In this work, we investigate the application of such a non-charged low temperature amorphous silicon passivation layer to IBC solar cells.

2. Low temperature passivation for p++ and n++ doped silicon.

2.1. Furnace diffused boron emitter

For investigation of the passivation ability under varying deposition parameters, especially the deposition temperature Tdep, we start with well-known phosphorus and boron furnace diffusions with sheet resistances Rsheet = 60 Q/sq. on n-type FZ wafer with resistivity p = 2 Qcm and thickness t = 280 ^m. After wet chemical cleaning, we deposit 30 nm amorphous silicon on both sides at varying deposition temperature Tdep = 150°C to Tdep = 350°C by 13.56 MHz plasma enhanced chemical vapour deposition (PECVD). Afterwards, we measure the effective minority carrier lifetimes xeff and evaluate the saturation current densities J0 with a WCT-120 lifetime tester from Sinton Instruments, using the latest published software [6].

Figure 1a and 1b show the saturation current densities for the passivated phosphorus and boron diffusion, respectively, for varying deposition temperature Tdep. The diffusion of hydrogen from amorphous silicon to the interface between amorphous and crystalline silicon explains the decreasing saturation current density J0,phos and J0,boron for increasing deposition temperature Tdep. This behavior occurs for phosphorus as well as for boron diffusions. An additional annealing step for 5 minutes at 350°C after the deposition decreases the saturation current density of the phosphorus diffusion further to J0phos ~ 60 fA/cm2 independent of deposition temperature Tdep. We assume that additional hydrogen diffusion to the interface plays the major role. In contrast to the phosphorus diffusion, the saturation current density of the boron diffusion decreases during annealing for low deposition temperatures, whereas it increases for higher deposition temperatures Tdep > 250°C. The reason for this behavior is not clear so far and needs further investigation. In the following all aSi:H layers are deposited at Tdep = 175°C.

Fig. 1. Saturation current density of (a) phosphorus furnace diffusion J0,phos and (b) boron furnace diffusion Jo,boron with sheet resistance Rsh = 60 Q/sq. for different amorphous silicon deposition temperatures Tdep. The saturation current density J0,phos and J0,boron directly after deposition of the aSi:H decreases with the deposition temperature Tdep for phosphorus and boron diffusions. After annealing at Tann = 350°C, the J0,phos of the phosphorus diffusion decreases to constant J0,phos ~ 60 fA/cm2 independent of deposition temperature Tdep. The saturation current density J0,boron of the boron diffusion increases after annealing for deposition temperature Tdep > 250°C. For lower Tdep < 250°C we measure J0,boron = 30 fA/cm2.

2.2. Laser diffused boron emitter

For investigation of the quality of our laser doped boron emitter, we use float-zone (FZ) grown n-type silicon wafer with a thickness t = 280 ^m and a resistivity p = 2 Qcm. For doping profiles comparable to those of the solar cell process, we apply the same process sequence to fabricate the boron doped emitter: precursor sputtering and laser irradiation on one side of the wafer, POCl3-diffusion of the front surface field (FSF), cleaning and oxidation [7,8]. The sheet resistance Rsh,em is adjustable by variation of the laser pulse energy density and the thickness of the precursor layer [7]. We compare two different passivation concepts: First, the thermal SiO2 acts as passivation layer with a silicon nitride on top for additional hydrogen supply. For the second approach hydrofluoric acid removes the thermal oxide on the rear and the low temperature amorphous silicon passivation layer is deposited by PECVD on both sides.

Figure 2a and 2b show the effective minority carrier lifetime xeff and the Auger corrected inverse effective lifetime 1/xeff - 1/xAuger of a laser doped boron emitter with Rshem = 90 Q/sq. The inset schematically shows the sample structure with the passivated boron emitter on the front and a passivated shallow phosphorus diffusion (FSF) on the rear. The passivation with a thermal oxide results in a measured saturation current density J0,em = 120 fA/cm2. Adding corona charges on top of the SiO2 is similar to the application of aluminum oxide. The charges supply an additional field effect, which reduces the saturation current density to J0em = 55 fA/cm2. This finding suggests the thermal SiO2 initially still having some positive fixed charges, limiting the surface passivation and therefore the effective lifetime xef. Unfortunately, corona charges are not stable and it is not possible to integrate them into a solar cell process. After removal of the thermal oxide in diluted hydrofluoric acid, the surface is passivated with amorphous silicon on both sides. This reduces the saturation current density to J0em = 46 fA/cm2 and shows the good quality of our laser doped boron emitter and the excellent passivation properties of the aSi:H.

Fig. 2. (a) Effective minority carrier lifetime Tef and (b) Auger-corrected inverse effective lifetime 1/Tef - 1/xAuger of a laser doped boron emitter with Rsh,em = 90 Q/sq. The inset shows a sketch of the sample structure. The passivation with thermal SiO2 and SiNx capping limits the effective lifetime Tef. The saturation current density amounts J0,em = 120 fA/cm2. Corona charges supply an additional field effect passivation, similar to aluminum oxide, and reduce the saturation current density to J0,em = 55 fA/cm2. The effective lifetime xef increases. Replacing the thermal SiO2 by aSi:H reduces the saturation current density further to J0,em = 46 fA/cm2.

Figure 3 shows the saturation current densities J0,em versus the sheet resistance Rsh of our boron doped emitter. The saturation current density J0em = 30 fA/cm2 at Rsh,em = 107 Q/sq. to J0,em = 51 fA/cm2 at Rsh,em = 54 Q/sq. is excellent with the amorphous silicon low temperature passivation.

For n-type IBC solar cells, additionally to a good boron emitter passivation, a low saturation current density J0,bsf of the laser doped phosphorus back surface field is necessary. For the preparation of the laser doped back surface field (BSF), the remaining phosphosilicate glass (PSG) from the FSF-diffusion serves as precursor [8]. Figure 3 also

shows the saturation current density J0,bsf of the phosphorus BSF on Czochralski-grown (CZ) n-type wafer with saturation current density J0,bsf = 73 fA/cm2 at a sheet resistance Rsh,bsf = 36 Q/sq, to J0,bsf = 43 fA/cm2 at Rsh,bsf = 95 Q/sq. Herewith, the aSi:H passivation layer is well suited for the rear side passivation of IBC solar cells.

p emitter aSi:H

40 60 80 100 sheet resistance R' [Q/sq]

Fig. 3. Saturation current densities of laser doped boron emitter J0,em and phosphorus BSF J0,bsf decreases with sheet resistance Rsh. The amorphous silicon passivation layer supplies good passivation with low J0,em and J0,bsf for both, the highly phosphorus doped BSF and boron doped emitter.

3. Laser doped IBC solar cells

3.1. Process flow

Figure 4a shows the process flow for the production of our IBC solar cells [7,8]. We fabricate the solar cells on one-sided textured CZ-grown n-type silicon wafer with a thickness t = 165 ^m and a resistivity p = 2 Qcm. The process starts with sputtering of the boron containing precursor. During laser doping with a green, line shaped, nanosecond pulsed Nd:YAG laser, the silicon melts and the boron diffuses into the liquid silicon [7]. After defect free recrystallization [9], the crystalline silicon is locally boron doped at the irradiated areas. The sheet resistance is easily adaptable by variation of the pulse energy density and the precursor layer thickness [7]. For our solar cells, we use a boron doped emitter with sheet resistance Rsh,em = 90 Q/sq. and a phosphorus BSF with sheet resistance Rsh,bsf = 36 Q/sq. Then, wet chemical etching removes the remaining precursor at the not irradiated areas and cleans the wafer surface. The following shallow POCl3 diffusion serves as front surface field (FSF) and the remaining PSG as precursor for the following laser doping of the back surface field (BSF). Phosphorus laser doping is done with the same laser system used for boron doping. After wet chemical removal of the remaining PSG and cleaning of the wafer surface, we apply two different passivation concepts on the rear side: First, a 14 nm thin thermal oxide acts as passivation layer. The oxidation also serves as a drive-in step for all doping profiles. A PECVD dielectric layer stack supplies additional hydrogen and improves the light trapping on the rear side. In a second approach, hydrofluoric acid removes the thermal oxide on the rear and the above described low temperature amorphous silicon passivation layer is deposited, including the same dielectric layer stack from the first approach. The removal of the thermal oxide and re-passivation with amorphous silicon ensures the same doping profiles for SiO2 and aSi:H passivated solar cells and therefore good comparability. After the passivation, the process continues with local laser ablation for contact openings and an evaporated aluminum metallization. The aluminum metallization is wet-chemically structured in hot phosphoric acid using a laser structured etching mask.

Figure 4b shows a schematic sketch of the finished solar cell including the interdigitated boron and phosphorus doping, the passivation layer stack and aluminum contacts on the polished rear side. In summary, we have fabricated IBC solar cells without any lithography steps applying two different passivation schemes. All doped areas are laser doped, the dielectric stack on the rear side is locally laser ablated for contact openings and the metallization is structured by a laser structured etching mask.

Fig. 4. (a) Process flow for the fabrication of laser doped back contact solar cells. The rear side is passivated with two different passivation schemes: either a thermal SiO2 or an amorphous silicon low temperature passivation. (b) Sketch of the finished solar cell with laser doped emitter and BSF, passivation layer stack and aluminum contacts.

3.2. Comparison of solar cells with thermal SiO2 and aSi:H passivation

Table 1 shows the saturation current densities of the used passivated and not passivated contacted emitter, with sheet resistance Rsh,em = 90 Q/sq. and BSF with sheet resistance Rsh,bsf = 36 Q/sq., as well as the gap between emitter and BSF on the rear side and the FSF on the front side of the solar cell with sheet resistance Rsh,FSF > 250 Q/sq. It distinguishes between the passivation with a thermal oxide and amorphous silicon. The saturation current densities J0 for the contacted areas are taken from King et al. [10,11]. The recombination in the bulk is not included to the calculation so far. The area weighted saturation current density stems from the area fraction f times the respective saturation current density J0. The passivated emitter shows the largest improvement when replacing the thermal oxide with amorphous silicon. The saturation current density reduces from J0,em,Si02 = 120 fA/cm2 to J0,em,aSi = 46 fA/cm2 for the used Rshem = 90 Q/sq. boron emitter.

The sum of all area weighted saturation current densities f x J0i according to table 1

J0,sum =Z f X J0,i (1)

results in the total saturation current density J0,sum,SiO2 = 118 fA/cm2 and J0 ,sum,aSi = 82 fA/cm2 for the passivation with a thermal oxide and amorphous silicon, respectively. According to the one-diode equation, the expectable open circuit voltage

Voc = vth • ln

V J0,ges )

is calculated from the thermal voltage Vth = 25.7 mV at a temperature T = 298.15 K and an assumed short circuit current density Jsc = 41 mA/cm2. We obtain Voc,Sio2 = 683 mV and Voc,aSi = 692 mV for the passivation with a thermal oxide and amorphous silicon, respectively.

Table 1. Comparison of area weighted saturation current densities for thermal silicon dioxide and amorphous silicon as rear side passivation. The area weighted saturation current densities sum to Jo,sum = 2 Jo = 118 fAcm-2 and 82 fAcm-2 for thermal SiO2 and amorphous silicon, respectively.

Rear side passivation Jo,siO2 [fA/cm2] Jo,asi [fA/cm2]

Area fraction f [%] f x J0Si02 [fA/cm2]

f x Jo,asi [fA/cm2]

Emitter passivated Emitter contacted BSF passivated BSF contacted Gap passivated FSF passivated

120 1500 70 1000 7 12

1500 73 1000 7 12

29 1 20 100

60 15 20 10 1 12

23 15 21 10 1 12

Table 2 shows the mean and standard deviation of the current/voltage data of solar cells fabricated according to our process flow on CZ-wafers. The short circuit current density Jsc is almost the same for amorphous silicon and thermal oxide. The cells with an active area A = 4 cm2 are measured in-house with a calibrated reference cell. The best solar cell, measured by ISE CalLab, has a calibrated efficiency ^max = 23.24%. As a consequence of the lower saturation current densities, the open circuit voltage improves with aSi:H passivation from VocSiO2 = 677 mV to Voc,aSi = 682 mV. The increase by AVoc = 5 mV is slightly lower than expected from our calculation. We assume an additional recombination path not considered in the one-diode model to be the reason for lower open circuit voltage. Since the solar cells are made of CZ-wafer, additional bulk recombination, which is not included to the calculation so far, might be a possible reason for the lower Voc and AVoc. In contrast to the voltage, the fill factor FF decreases from FFSiO2 = 82.2% to FFaSi = 81.8%. As the pseudo fill factor pFF =84.2% is not affected by the change of the passivation layer, we assume that the not yet optimized laser ablation process for the aSi:H passivation layer stack leads to an increased electrical contact resistance. We achieve a mean efficiency ^ = 22.8% for both passivation schemes due to the slightly lower fill factor for the aSi:H passivated cells.

Table 2. Mean IV-Data and standard deviation of solar cells with an area A = 4 cm2. The open circuit voltage improves by AVoc = 5 mV when replacing the thermal SiO2 with an amorphous silicon passivation layer.

Rear side passivation Jsc [mA/cm2] Voc [mV] FF [%] pFF [%] r) [%] Amount

Thermal SiO2 41.0±0.2 677±2 82.2±0.5 84.2±0.8 22.8±0.3 11

Amorphous silicon 40.9±0.1 682±2 81.8±0.4 84.2±0.3 22.8±0.1 6

4. Conclusion

Our amorphous silicon low temperature passivation shows excellent passivation on furnace diffused boron emitters, as long as the deposition temperature is below Tdep < 200°C. The reason for inferior passivation at higher deposition temperatures is not clear so far. Additionally to furnace diffused boron emitter, low temperature amorphous silicon passivates both, laser doped p++ boron emitters and n++ phosphorus back-surface fields well. The saturation current densities are as low as J0em = 46 fA/cm2 for the boron emitter with sheet resistance Rshem = 90 Q/sq and J0bsf = 73 fA/cm2 for the phosphorus BSF with sheet resistance Rsh,bsf = 36 Q/sq. The integration of the amorphous silicon low temperature passivation into our fully laser doped solar cell process results in IBC solar cells with an open circuit voltage increased by AVoc = 5 mV compared to cells with thermal SiO2. The best solar cell

achieves a confirmed efficiency ц = 23.24%. Adjusting the parameters for laser ablation and further process optimization will result in even higher cell efficiencies in the future.

Acknowledgements

The authors acknowledge the work of Birgitt Winter, Hendrik Moldenhauer, Lydia Beisel, Leonard Bauer, Sergej Vollmer and Brigitte Lutz in wafer and solar cell processing. We thank ISC Konstanz, especially Alexander Edler and Corrado Comparotto, for providing the furnace diffused boron emitters. This work is supported by German Federal Ministry of Economics and Technology (BMWi) under project number 0325714.

References

[1] Franklin E, Fong K, Mcintosh K, Fell A, Blakers A, Kho T, Walter D, Wang D, Zin N, Stocks M, Wang EC, Grant N, Wan Y, Yang Y,

Zhang X, Feng Z and Verlinden PJ. Design, fabrication and characterisation of a 24.4% efficient interdigitated back contact solar cell. Prog Photovolt: Res Appl 2014; doi: 10.1002/pip.2556.

[2] Dauwe S, Mittelstädt L, Metz A and Hezel R. Experimental Evidence of Parasitic Shunting in Silicon Nitride Rear Surface Passivated Solar

Cells. Prog Photovolt: Res Appl 2002; 10 271-278.

[3] Simon DK, Jordan PM, Dirnstorfer I, Benner F, Richter C, Mikolajick T. Symmetrical Al2O3-based passivation layers for p- and n-type

silicon. Sol Energ Mat Sol Cells 2014; 131 72-76.

[4] Dauwe S, Schmidt J, Hezel R. Very low surface recombination velocities on p- and n-type silicon wafers passivated with hydrogenated

amorphous silicon films. Proceedings of the 29th IEEE Photovoltaic Specialists Conference; 2002. p.1246-1249.

[5] Seiffe J, Gautero L, Hofmann M, Rentsch J, Preu R, Weber S, and Eichel RA. Surface passivation of crystalline silicon by plasma-enhanced

chemical vapour deposition double layers of silicon-rich silicon oxynitride and silicon nitride. J Appl Phys 2011; 109 034105.

[6] Blum AL, Swirhun JS, Sinton RA and Kimmerle A. An Updated Analysis to the WCT-120 QSSPC measurement system using advanced

device physics. Proceedings of the 28th European Photovoltaic Solar Energy Conference; 2013. p. 1521-1523.

[7] Dahlinger M, Eisele SJ, Köhler JR, Werner JH. Laser Doped Boron Emitters with Sputtered Precursor. Proceedings of the 26th European

Photovoltaic Solar Energy Conference; 2011. p. 1152-1154.

[8] Dahlinger M, Carstens K, Köhler JR, Zapf-Gottwick R, Werner JH. Laser Doped Screen-printed Back Contact Solar Cells Exceeding 21%

Efficiency. Energy Procedia 2014; 55 410-415.

[9] Ohmer K, Weng Y, Köhler JR, Strunk HP, and Werner JH. Defect Formation in Silicon During Laser Doping. IEEE J Photovoltaics 2011; 1

183-186.

[10] King RR and Swanson MR. Studies of Diffused Boron Emitters: Saturation Current, Bandgap Narrowing, and Surface Recombination Velocity. IEEE Trans Electron Devices 1991; 38:6 1399-1409.

[11] King RR, Sinton RA, Swanson MR. Studies of Diffused Phosphorus Emitters: Saturation Current, Surface Recombination Velocity, and Quantum Efficiency. IEEE Trans Electron Devices 1990; 37:2 365-371.