TSF-33887; No of Pages 4 Thin Solid Films xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Thin Solid Films journal homepage: www.elsevier.com/locate/tsf
Cu2S as ohmic back contact for CdTe solar cells Johannes Türck, Sebastian Siol, Thomas Mayer, Andreas Klein, Wolfram Jaegermann ⁎ Technische Universität Darmstadt, Department of Materials and Earth Sciences, Surface Science Division, Jovanka-Bontschits-Str. 2, D-64287 Darmstadt, Germany
a r t i c l e
i n f o
Available online xxxx Keywords: Cadmium telluride Solar cell Back contact Copper sulfide Interface X-ray photoelectron spectroscopy Valence band offset
a b s t r a c t We prepared a back contact for CdTe solar cells with Cu2S as primary contact. Cu2S was evaporated on CdCl2 treated CdTe solar cells in superstrate configuration. The CdTe and CdS layers were deposited by Closed Space Sublimation. Direct interface studies with X-ray photoelectron spectroscopy have revealed a strongly reactive interface between CdTe and Cu2S. A valence band offset of 0.4–0.6 eV has been determined. The performance of solar cells with Cu2S back contacts was studied in comparison to cells with an Au contact that deposited onto a CdCl2-treated CdTe surface that was chemically etched using a nitric-phosphoric etch. The solar cells were analyzed by current–voltage curves and external quantum efficiency measurements. After several post deposition annealing steps, 13% efficiency was reached with the Cu2S back contact, which was significantly higher than the ones obtained for the NP-etched back contacts. © 2014 Elsevier B.V. All rights reserved.
1. Introduction One of the big challenges in the fabrication of a CdTe solar cell is the back contact to p-CdTe. A direct CdTe/metal contact leads to large Schottky barrier heights in the valence band [1]. The formation of a tunnel junction is also impeded, due to the difficulties in doping CdTe p-type. The common back contacts applied to CdTe solar cells are a thin Te layer, formed by etching as a primary contact, which has a lower valence band offset [2], followed by a secondary metal contact. Cu containing back contacts are also often used. Cu diffuses into the CdTe and leads to p-doping, which improves the formation of the back contact [3–5]. To improve the back contact of a CdTe solar cell a contact interface layer with high ionization energy should lead to a small valence band offset. With an ionization energy of 5.4 eV Cu2S [6] is a promising candidate for such an application.
2. Experimental details The interface properties have been studied by stepwise deposition of Cu 2 S onto a CdCl 2 treated and cleaned CdTe layer. After every deposition step the sample was studied by X-ray photoelectron spectroscopy (XPS) to investigate the band alignment. The subsequent analysis and deposition were carried out at the DArmstadt Integrated SYstem for MATerial research (DAISY-MAT), which combines a Physical Electronics PHI 5700 multi-technique surface analysis system with several deposition chambers. Monochromatic Al Kα radiation (1486.6 eV) has been used for the XPS measurements. ⁎ Corresponding author. E-mail address:
[email protected] (W. Jaegermann).
Binding energies are given with respect to the Fermi level, calibrated by Ag and Cu references. The solar cells have been prepared on commercial Fluorine doped Tin Oxide (Pilkington TEC C15M) substrates. The CdS and CdTe layers were deposited by Closed Space Sublimation at a substrate temperature of 520 °C and a film thickness of 100 nm and 5 μm, respectively. The cells were wet chemically activated with CdCl2 soluted in methanol and annealed for 25 min at 400 °C in a tube furnace in an air atmosphere. After activation the cells were cleaned in deionized water for 5 min in an ultrasonic bath to remove the surface oxidation on the CdTe layer [7]. A CdTe solar cell with a primary Cu2S back contact of 60 nm layer thickness has been prepared. The Cu2S layer was thermally evaporated from Cu2S powder (abcr GmbH and Co. KG 99.5% purity). The substrate was not additionally heated during the deposition of the Cu2S layer. A more detailed description of the deposition parameters is given in [8]. As a secondary back contact 100 nm gold was deposited via sputtering. As a reference a CdTe solar cell with nitric-phosphoric (NP) etched back contact with a secondary contact of 100 nm Au has been manufactured. The composite of the NP etch was 70% H 3PO4, 0.8% HNO3, and 29.2% H2O by weight and the etch process is continued until bubbles cover the surface. The gold contacts were sputtered at room temperature on the solar cell covering the whole area. By mechanical scraping single solar cells with an area of 0.16 cm 2 have been prepared. The current–voltage (IV) behavior was measured with a Keithley 2400 under AM 1.5 spectra (solar simulator LOT-Oriel Typ 81150). Also the external quantum efficiency (EQE) for the solar cells has been determined. The solar cells have been annealed several times on a hotplate at temperatures between 150 °C and 225 °C, followed by repeated light IV and EQE measurements.
http://dx.doi.org/10.1016/j.tsf.2014.11.017 0040-6090/© 2014 Elsevier B.V. All rights reserved.
Please cite this article as: J. Türck, et al., Cu2S as ohmic back contact for CdTe solar cells, Thin Solid Films (2014), http://dx.doi.org/10.1016/ j.tsf.2014.11.017
2
J. Türck et al. / Thin Solid Films xxx (2014) xxx–xxx
a
3. Results
3000
b
1000 0
Cu 2p3/2
0.0 0.2
Cd 3d5/2 Te 3d5/2 Cu 2p3/2 S 2p
0.4 0.6 0.8 1.0 1.2 CdTe
Cu2S
Cd 3d5/2
0.1 substrate intensity [arb. unit]
1 10 100 adsorbate thickness [nm]
-0.2
1.4 S2p
Cd 3d5/2 Te 3d5/2 Cu 2p3/2 S 2p
2000
0.1 substrate
EVBM-EF [eV]
To determine the band alignment of the interface between Cu2S and CdTe XPS measurements were carried out while successively depositing Cu2S on a CdTe layer which was CdCl2 treated and cleaned. XP spectra of the valence band region and the main core level emissions of the substrate and film have been measured, and are shown in Fig. 1. The XP spectra of the substrate, which were cleaned with deionized water, show almost no oxidation of the Te 3d emission. TeO2 shows a chemical shift of 3.5 eV compared to CdTe (TeO2 3d5/2 577.3 eV [9]). The position of Cd 3d5/2 core level for CdO is 404.8 eV [10], which is only 0.6 eV below the binding energy in CdTe, so no conclusion about the oxidation of Cadmium could be made. The O1s emission is almost not visible. Therefore, it can be argued, that the surface of the CdTe substrate has almost no oxidation. In Fig. 2a the evolution of the substrate and adsorbate peak intensities and in Fig. 2b the development of core level binding energies are shown versus the logarithmic adsorbate layer thickness. For the first three deposition steps the Cd 3d and Te 3d emission intensities are attenuated and the Cu 2p and S 2p emissions are increased as expected. The Cu to S ratio is 2:1. During the following deposition steps the Te emission intensity further decreases and S increases, but after deposition of 15 nm i.e. step 5 in Fig. 2 unexpectedly the intensity of Cu drops and the intensity of Cd increases. This is a clear indication of a strong exchange reaction between Cd and Cu. The formation of an interface layer of CuxTe [11] and CdS is suggested. After the two last deposition steps the Cd 3d emission
intensity with ASF [arb. unit]
4000
3.1. Interface experiment CdTe/Cu2S
1 10 100 adsorbate thickness [nm]
300 nm Fig. 2. a) Evolution of the intensity during the interface experiment. The intensities are divided by the atomic sensitivity factors (AFS) to scale the intensities to comparable values. After the deposition of 15 nm Cu2S a reaction between the substrate and the adsorbate is visible. b) Evolution of the core level binding energies of the CdTe/Cu2S interface experiment. All the binding energies are drawn in respect to the corresponding VBM. The CdTe emissions are shifted to lower binding energies. After the reaction between CdTe and Cu2S the Cu2S emission is shifted to higher binding energies.
90 nm 15 nm
6 nm 3 nm 1.5 nm
0.4 nm 0 nm 166 164 162 160 binding energy [eV] Te 3d5/2
936 934 932 930 binding energy [eV] Cu LMM
407 406 405 404 binding energy [eV]
VB
intensity [arb. unit]
300 nm 90 nm
0.26 eV
15 nm 6 nm 3 nm 1.5 nm
0.4 nm 0 nm 575 570 binding energy [eV]
0.78 eV
6
5
4 3 2 1 binding energy [eV]
0
-1
Fig. 1. Photoemission spectra of the core level and valence band emissions from the interface experiment measured after each Cu2S deposition cycle on a CdTe substrate. The bottom lines show the CdTe substrate. With etch deposition step of Cu2S the core level emission of Cu and S is increased and the core level emission of Cd and Te is damped. From the valence band (VB) spectra the position of the valence band maximum for the CdTe substrate has been determined. The valence band maximum of the Cu2S was determined from the last deposition step of the Cu2S.
has almost disappeared and a stoichiometric Cu2S layer has formed as indicated by the Cu to S ratio of 2:1. The development of the substrate core level binding energies in Fig. 2b is plotted after subtracting the initial binding energy difference between the core level and the CdTe valence band maximum (VBM). For the adsorbate core levels the binding energies are plotted after subtracting the binding energy difference between the respective adsorbate core level and the Cu2S VBM of the final deposition step. The position of the valence band maximum was identified for the substrate and the last deposition step of the Cu2S layer by a linear extrapolation of the low binding energy edge of the XP spectra. For the intermediate steps the valence bands of the two materials superpose each other and cannot be determined. The development of the CdTe and Cu2S valence band maxima is concluded from the evolution of the core level binding energies. With the first 3 deposition steps the binding energy of substrate and adsorbate core level emission lines increases in parallel corroborating the absence of a chemical reaction. Instead downward band bending in the CdTe substrate by 0.2 eV is indicated. Then the Te binding energy changes more strongly than the binding energy of the Cd emission and the binding energy of Cu decreases while the binding energy of S increases, corroborating the exchange reaction between Cd and Cu. In the following deposition steps the decrease of Cu and S binding energies indicates band bending in the growing Cu2S layer. Due to the intermediate reaction phase the exact value is hard to determine but is in the range between 0.15 eV and 0.35 eV. This leads to a valence band discontinuity between 0.4 eV and
Please cite this article as: J. Türck, et al., Cu2S as ohmic back contact for CdTe solar cells, Thin Solid Films (2014), http://dx.doi.org/10.1016/ j.tsf.2014.11.017
FF [%]
70 65 60 55 50 45 0.85 0.80 0.75 0.70 0.65 0.60 24 23 22 21 20
annealing temperature 150 °C
200 °C
225 °C
annealing time [min] Fig. 4. Evolution of the efficiency (η), fill factor (FF), open circuit voltage (UOC) and short circuit current (JSC) for the solar cells during the annealing on the hotplate at several temperatures. No improvement of the performance of the NP-etched back contact is visible. For the solar cell with a Cu2s back contact an improvement of the FF and UOC is visible, which also leads to an improved efficiency.
40 30 2
J [mA/cm ]
The CdTe Solar cell with the Cu2S back contact does not show a good performance as deposited, showing an efficiency of only 6.7%. The reasons for this are the low open circuit voltage and fill factor. In Fig. 4 the characteristic values of the solar cells with Cu2S and NP-etched back contact are compared with each other during the annealing on the hotplate. All light IV cell parameters of the device with the NP-etched back contact are relatively stable over the heat treatment with an efficiency of around 11% compared to the sample with the Cu2S containing back contact. The fill factor, short circuit current and the open circuit voltage are also relatively stable after the first annealing step. For the solar cell with the Cu2S back contact the behavior is totally different. At annealing temperature of 150 °C an increase of the fill factor to 65% is visible. With further annealing steps at 200 °C it was also possible to improve the open circuit voltage to almost 800 mV. The fill factor improved to 70% by a annealing at 225 °C. This way an efficiency of 13% could be reached with the Cu2S back contact. This significant improvement can be attributed to healing of interface states [13], which could lead to an improved band alignment at the back contact. The doping of the CdTe by Cu diffusion into the absorber could also explain the improvement of solar cells with Cu2S back contact. For longer annealing times at 225 °C a degradation of the short circuit current and the fill factor was observed. In Fig. 5 some selected light IV-curves and external quantum efficiencies are shown. The solar cell with the NP-etched back contact shows a roll over in the light IV curves, this still remains during the annealing of the cells. The Cu2S back contact also has a roll over directly after the deposition. With the annealing step on the hotplate the roll over gets less prominent and the fill factor of the cells with Cu2S back contact improves. The short circuit current is significantly higher for the Cu2S back contact. From the EQE measurement an improvement in the longer wavelength range is recognized, an indication of either less recombination at the back contact or improved carrier life times due to doping of the CdTe with Cu. Improved collection in the space charge region could also be an explanation for the improved EQE, because the electric field in the space charge region is increasing with a stronger p doped CdTe due to Cu diffusion.
Cu2S back contact NP-etched back contact
0 5 10 15 20 25 30 35 40 45 50 55
3.2. Solar cells with Cu2S back contact
3
as deposited
η [%]
14 12 10 8 6
UOC [V]
0.6 eV between CdTe and Cu2S. Using literature vales for the band gaps of CdTe (1.45 eV [12]) and Cu2S (1.2 eV [6]) the respective conduction band discontinuity is given between 0.15 eV and 0.35 eV. The described photoemission data are summarized in the interface band diagram (Fig. 3). The band diagram indicates a barrier of 0.2 eV for holes generated in CdTe and a driving force of 0.4–0.6 eV for the transfer to Cu2S, which on the other hand is also a loss in photo voltage. In addition the disordered CuxTe– CdS interphase is likely to have a high number of recombination states.
JSC [mA/cm²]
J. Türck et al. / Thin Solid Films xxx (2014) xxx–xxx
20 10 0
Cu2S back contact after deposition 150°C 10 min 200°C 15 min 225°C 05 min NP-etched back contact
-10 -20
Cu2S
CdTe
0.0
0.4 U [V]
0.8
Eg=1.45 eV
EF EVBM=0.8 eV EVBM
0.2 eV
Eg=1.2 eV
EVBM=0.26 eV
0.8
EQE
ECBM
interface reaction layer
1.0
0.6 Cu2S back contact after deposition 225°C 05 min 225°C 15 min NP-etched
0.4
0.15-0.35 eV 0.2
ΔEVB=0.4-0.6 eV
0.0 400
Fig. 3. Band diagram for the CdTe/Cu2S interface. The band edges in the interface region are schematically drawn to indicate the overall alignment of energy levels.
500
600 700 800 wavelength λ [nm]
900
Fig. 5. Selected IV curves (top) and EQE curves (bottom).
Please cite this article as: J. Türck, et al., Cu2S as ohmic back contact for CdTe solar cells, Thin Solid Films (2014), http://dx.doi.org/10.1016/ j.tsf.2014.11.017
4
J. Türck et al. / Thin Solid Films xxx (2014) xxx–xxx
4. Conclusions It has been shown that Cu2S can be used as a back contact material for CdTe solar cells. The band aliment between Cu2S and CdTe was determined by photoelectron spectroscopy with an in-situ interface experiment. Due to reaction at the interface between CdTe and Cu2S, it is only possible to give a range for the measured valence band offset, which is between 0.4 eV and 0.6 eV. CdTe solar cells with 13% efficiency have been manufactured with the Cu2S back contact, after an annealing of the solar cell. This is considerably higher than the efficiency of the NPetched cell. The long-term stability has not yet been tested and needs further investigation. References [1] S.H. Demtsu, J.R. Sites, Effect of back-contact barrier on thin-film CdTe solar cells, Thin Solid Films 510 (2006) 320. [2] D. Kraft, A. Thiβen, M. Campo, M. Beerbom, T. Mayer, A. Klein, et al., Electronic properties of chemically etched CdTe thin films: role of te for back-contact formation, in: R.W. Birkmire, D. Lincot, R. Nouti, H.W. Schocl (Eds.), Mater. Res. Soc. Symp. Proc., 688, Materials Research Society, San Francisco, 2011, p. H7.5.1. [3] R. Scheer, H.-W. Schock, Chalcogenide Photovoltaics, Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, Germany, 2011.
[4] C. Corwine, A.O. Pudov, M. Gloeckler, S.H. Demtsu, J.R. Sites, Copper inclusion and migration from the back contact in CdTe solar cells, Sol. Energy Mater. Sol. Cells. 82 (2004) 481. [5] S. Erra, C. Shivakumar, H. Zhao, K. Barri, D.L. Morel, C.S. Ferekides, An effective method of Cu incorporation in CdTe solar cells for improved stability, Thin Solid Films. 515 (2007) 5833. [6] G. Liu, T. Schulmeyer, J. Brötz, A. Klein, W. Jaegermann, Interface properties and band alignment of Cu2S/CdS thin film solar cells, Thin Solid Films. 431–432 (2003) 477. [7] D. Levi, D. Albin, D. King, Influence of surface composition on back-contact performance in CdTe/CdS PV devices, Prog. Photovoltaics Res. Appl. 8 (2000) 591. [8] S. Siol, H. Sträter, R. Brüggemann, J. Brötz, G.H. Bauer, A. Klein, et al., PVD of copper sulfide (Cu2S) for PIN-structured solar cells, J. Phys. D. Appl. Phys. 46 (2013) 495112. [9] H. Hayashi, N. Shigemoto, S. Sugiyama, N. Masaoka, K. Saitoh, X-ray photoelectron spectra for the oxidation state of TeO2–MoO3 catalyst in the vapor-phase selective oxidation of ethyl lactate to pyruvate, Catal. Lett. 19 (1993) 273. [10] F. Golestani-Fard, T. Hashemi, K.J.D. Mackenzie, C.A. Hogarth, Formation of cadmium stannates studied by electron spectroscopy, J. Mater. Sci. 18 (1983) 3679. [11] B. Späth, K. Lakus-Wollny, J. Fritsche, C.S. Ferekides, A. Klein, W. Jaegermann, Surface science studies of Cu containing back contacts for CdTe solar cells, Thin Solid Films. 515 (2007) 6172. [12] D. Bonnet, P. Meyers, Cadmium-telluride—material for thin film solar cells, J. Mater. Res. 13 (1998) 2740. [13] J. Yun, K. Kim, D. Lee, B. Ahn, Back contact formation using CuTe as a Cu-doping source and as an electrode in CdTe solar cells, Sol. Energy Mater. Sol. Cells. 75 (2003) 203.
Please cite this article as: J. Türck, et al., Cu2S as ohmic back contact for CdTe solar cells, Thin Solid Films (2014), http://dx.doi.org/10.1016/ j.tsf.2014.11.017
