- Email: [email protected]
Study on influences of CdZnS buffer layer on CdTe solar cells
Accepted Manuscript Study on influences of CdZnS buffer layer on CdTe solar cells
Xiaobo Xu, Xiaoyan Wang, Wenping Gu, Si Quan, Zan Zhang PII:
S0749-6036(17)30500-1
DOI:
10.1016/j.spmi.2017.05.033
Reference:
YSPMI 5014
To appear in:
Superlattices and Microstructures
Received Date:
27 February 2017
Revised Date:
14 May 2017
Accepted Date:
14 May 2017
Please cite this article as: Xiaobo Xu, Xiaoyan Wang, Wenping Gu, Si Quan, Zan Zhang, Study on influences of CdZnS buffer layer on CdTe solar cells, Superlattices and Microstructures (2017), doi: 10.1016/j.spmi.2017.05.033
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Highlights: 1. Material parameters of Cd1-xZnxS are selected and confirmed with different x. 2. The optimal x content is found to be 0.6 for the improved performance. 3. the optimal Cd1-xZnxS width is found to be 30nm. 4. The efficiency increases from 15.42% to 17.71% with the optimized CdZnS layer.
ACCEPTED MANUSCRIPT Study on influences of CdZnS buffer layer on CdTe solar cells * Xiaobo Xu,* Xiaoyan Wang, Wenping Gu, Si Quan, Zan Zhang School of Electronic and Control Engineering, Chang'an University, Xi'an 710064, PR China
ABSTRACT: This paper deals with the Cd1-xZnxS/CdTe solar cell, with the structure parameters confirmed by basic theories, reports and experiments. The influences of x composition and the buffer layer width on the open circuit voltage, the short circuit current, the fill factor, and the conversion efficiency are discussed by numerical simulation. It is shown that the optimized buffer layer should be Cd0.4Zn0.6S, and the width is 30nm based on the simulation. The Cd1-xZnxS introduction indeed increases the blue portion absorption of solar spectrum and the efficiency of CdTe solar cells is thus improved. Key words：CdTe, solar cell, Cd1-xZnxS PACS:
73.40.Lq, 88.40.jn, 88.40.H-
1.Introduction
Over the past three decades, the Ⅱ-Ⅵ semiconductor compound Cadmium Sulfide (CdS) has been extensively investigated from the research community owing to the efficient use in the fabrication of solar cells. In traditional CdTe solar cells, the CdS thin film is applied as a window material and as a heterojunction partner, due to its high transparency, direct band gap transition, high electron affinity and n-type conductivity [1, 2]. As the solar irradiance is rich in photons with energies between 1.4 eV and 3 eV, the band gap of CdS is hence relatively low (2.42 eV) [3], the photons with energy higher than 2.42 eV may be absorbed by the CdS layer, while the Ⅱ-Ⅵ semiconductor compounds show absorption efficiency as high as 104-105 cm-1, the higher energy photons are absorbed at the surface, where the depletion region is far away and the recombination is substantial, causing little contribution to the light current. In order to improve the performance of CdTe solar cells, an appropriate window layer should be introduced. In fact, the band gap of CdS is adjustable by adding Zn to substitute Cd, in the form of ZnxCd1-xS, depending on the Zn:Cd ratio. Since ZnxCd1-xS has a larger band gap than CdS, the use of the ZnxCd1-xS window layer instead of CdS is suitable to circumvent window absorption losses as well as lattice mismatch problems associated with a quaternary solar absorber layer [4]. Although some earlier reports focused on Cd1-xZnxS/CdTe solar cells [5, 6], they were constrained to be low Zn content situation, because early growth techniques show that for the Zn concentration of x>1, the resistivity of Cd1-xZnxS is large enough to counterpart absorption improvement, leading to sharp decline of the short circuit current [7-9]. But with the advances of the growth technique, recent researches show decreased resistivity with the increase of x [10]. The best performance is observed for x=0.7 from Cd1-xZnxS solar cells deposited via the AP-MOCVD method, where the Cd1-xZnxS width is 240nm [11]. In this simulation, not only the optimized Zn content but also the Cd1-xZnxS layer width are discussed to obtain the highest efficiency.
2. CdTe Cell Design
The optimized structure is designed to be TCO/Cd1-xZnxS/CdTe/ZnTe as shown in Fig. 1. ZnTe (bandgap 2.26eV) is introduced to be inserted prior to the back contact to form the back surface field (BSF), in order to inhibit the recombination of the generated carriers by reflecting to the Cd1-xZnxS/CdTe junction [12]. Cd1-xZnxS is introduced due to the blue portion absorption of the solar spectrum for the CdS window layer, which causes a decrease in the current density of solar cells.
Corresponding author. E-mail address: [email protected] (X. Xu)
*
ACCEPTED MANUSCRIPT Back contact ZnTe
P-CdTe
N-Cd1-xZnxS TCO(SnO2) glass
Fig. 1
Optimized CdTe solar cell structure
WxAMPS [13] is used as the simulation tool. To start a simulation, the basic material parameters of SnO2 [14], CdS [14] and CdTe [12, 15] are assumed by reports and experiments, including the doping concentration, the dielectric constant, the bandgap, the mobility, the affinity, and so on, as shown in Table 1. It should be noted that although the absorption coefficient of CdTe is as high as 10-5 cm-1, and thin film CdTe layers have been successfully fabricated by different growth techniques [16, 17], the CdTe layer is normally prepared with higher film thickness in order to avoid shunting effects due to the formation of pinholes [18]. In this paper, the width is set to be 4 μm.
Table 1 Material parameters
Para.
SnO2
CdTe
ZnTe
W (μm)
0.5
4
0.2
ε/ε0
9
9.4
9.67
μn (cm2/V/s)
100
320
330
μp (cm2/V/s)
25
40
80
n,p (cm-3)
1×1017
2×1014
1×1018
Eg (eV)
3.6
1.5
2.26
NC (cm-3)
2.2×1018
8×1017
7×1016
NV (cm-3)
1.8×1019
1.8×1019
2×1019
χ
4
4.28
3.5
For the Cd1-xZnxS layer, the variable x causes the variations of material parameters. As some values are insensitive to the CdTe cell performance, such as the dielectric constant ε/ε0, the conduction band density, and the valence band density NV. It is reasonable to set them invariable for simplicity: ε/ε0=9.3, NC =2.1×1018cm-3, NV =1.7×1019cm-3. The other parameters are listed in table 2 [19].
Table 2 Material parameters of Cd1-xZnxS
Para.
x=0.05
x=0.08
x=0.1
x=0.2
x=0.3
x=0.5
x=0.6
x=0.8
x=0.9
μn (cm /V/s)
100
100
95
85
75
70
65
60
50
μp (cm /V/s)
40
40
35
30
25
20
15
10
7
Eg (eV)
2.48
2.50
2.55
2.58
2.64
2.70
3.07
3.33
3.43
χ
4.47
4.46
4.44
4.38
4.32
4.26
4.14
4.02
3.94
2 2
ACCEPTED MANUSCRIPT It should be noted that these parameters are quite different from techniques, we take the most significant bandgap as example: Eg=2.62 eV at x=0.2 for vacuum elaborated thin film [20], while Eg=2.66 eV at x=0.2 for spraying pyrolysis method [9], and the bandgap also depends on the annealing temperature [7, 21, 22], such as Eg=2.48 eV at x=0.05 as deposited but Eg=2.44 eV at x=0.05 as annealed at 400°C for 30 minutes [7]. Therefore the above values are just set to be typical, to show the influences on the cell performances. All simulations are modeled under the standard test condition (1000W/m2, 25°C ), for simplicity, the front and rear recombination rates are the same as 1×107cm/s, the front and rear reflectivities are 0.1 and 0.8, respectively. Mid-gap donor-like defect is assumed [14], and the density is set to be 1×1015cm-3, the doping concentration of Cd1-xZnxS is assumed to be 1.1×1018cm-3, the Cd1-xZnxS width is first set to be 50nm to obtain the optimal x composition. 3. Discussion Fig. 2 and Fig.3 show variations of critical parameters for different Zn contents, including the short circuit current density Jsc, the open circuit voltage Voc, the fill factor FF and the conversion efficiency η.
Fig. 2 Open circuit voltage VOC and short circuit current JSC for different ratio x
Fig. 3 Fill factor FF and conversion efficiency η for different ratio x
Fig. 4 Electron lifetime of CdTe layer
As the energy band gap of Cd1-xZnxS thin film increases with the increase of Zn concentration, the absorption edge shifts toward lower wavelength, more photons are absorbed by CdTe layer and the short circuit current density is hence improved from 23.118mA/cm2 at x=0.05 to 23.757 mA/cm2 at x=0.6 (a relative 2.76% increase). The slow variation between x=0.3 and x=0.6 is because most absorbable photons have been transmitted to the CdTe layer and the subsequent improvement is minor. It should be noted that the density drops slightly to 23.651 mA/cm2 at x=0.8. This decreased density may be due to the lattice mismatch induced recombination. As the open circuit voltage changes approximately logarithmically with the short circuit current, VOC
ACCEPTED MANUSCRIPT increases with x. The variations is just 1.4%, and the VOC can be regarded as constant, which is verified by Fig. 4 where the minority lifetime of the CdTe layer can be regarded as invariable with x. The efficiency is shown to increase from 15.42% to 16.48% during the range of x=0.05~0.6 (a relative 6.87% increase), because of the increases of the short circuit current, the open circuit voltage and the fill factor based on the equation η=VOCJSCFF /G, where G is the solar irradiation density. Values plateau beyond x=0.3 before dropping off at x>0.6. The decline is apparently because of the decreased short circuit current. As the fill factor (a relative 2.47% increase) plays a significant role on the efficiency, the fill factor should be further analyzed. Based on the empirical equation of Green [23], the fill factor is dominated by the serial resistance RS and the shunt resistance RSH. If we define the normalized open circuit voltage as vOC=VOC/(nkT/q), the ideal fill factor FF0 is expressed as
FF0
vOC ln vOC 0.72 vOC 1
(1)
Further, the specific fill factors FFs and FFsh are correlated to the serial resistance and the shunt resistance, respectively.
FFs FF0 1 rs
(2)
v 0.7 FF0 FFsh FF0 1 OC vOC rsh
(3)
with rs=RS/(VOC/ISC), rsh=RSH/(VOC/ISC). Apparently FFs decreases with the increase of Rs, while FFsh increases inversely with the increase of RSH. The serial resistance and shunt resistance dependent fill factor FF is finally expressed as [24]
v 0.7 FF0 1 rs FF FF0 1 rs 1 OC vOC rsh
(4)
Fig. 5 RS and RSH for different ratio x
The values of Rs and RSH obtained under different Zn concentrations are shown in Fig. 5. RS decreases and RSH increases dramatically with the increased x between x=0.05~0.3. According to the above models, FF shows a sharp rise, which is consistent with the curve in Fig. 3. As x increases further, both the RS and the RSH deviate little, and FF is hence stable. According to the experimental data [11], as x=0.9, RSH drops to lower than 500 Ω/cm2, and Rs grows to higher than 16 Ω/cm2, which mean a sharp deterioration of the cell performance. In our simulation by setting Rsh=410 Ω/cm2 and Rs =16.9Ω/cm2 as the input values, the corresponding results are VOC=0.9144V, JSC=23.63 mA/cm2, FF=43.9%, η=9.53%, which show similar decline trends with the experiment [11]. As x=0.6 corresponds to the highest efficiency, the Cd0.4Zn0.6S buffer layer width is next discussed for further optimization of the device. The width is changed from 10nm to 100nm to show the variations of the cell performance, in order to obtain the optimal width.
ACCEPTED MANUSCRIPT
Fig. 6 Open circuit voltage VOC and short circuit current JSC for different buffer layer widths
Fig. 7 Fill factor FF and conversion efficiency η for different buffer layer widths
As the Cd0.4Zn0.6S buffer layer width declines from 100nm to 30nm, the short circuit current increases quickly due to shorter transmission path for the absorbable photons, and the open circuit voltage and the fill factor are nearly constant, therefore the efficiency is dominated by the short circuit current and rises significantly. But with a continuous decrease of the buffer layer width, the short circuit current, the open circuit voltage and the fill factor all deteriorate for the reason of pinhole existence and the corresponding recombination induced loss, and the efficiency is hence decreased. The optimal buffer layer width of Cd0.4Zn0.6S is thus 30nm, where JSC, VOC, FF, and η are 25.468 mA/cm2, 922 mV, 75.4%, and 17.71%, respectively.
4. Conclusion
The optimized Cd1-xZnxS layer is confirmed in the CdTe solar cell design. Numerical simulations were carried out by the variations of the Zn content and the Cd1-xZnxS width. It is shown that the efficiency increases sharply with x=0.05~0.3 and keeps almost unchanged with higher x compositions, the optimal Zn concentration is 0.6; The buffer layer width also influences the cell performance, with longer photon transmission path for larger width, and pinhole formation for smaller width, the optimal Cd1ZnxS width is found to be 30nm. The research of the Cd1-xZnxS layer optimization provides constructive guidelines for the future
x
design and fabrication of CdTe solar cells. Acknowledgment The work was supported by National natural science foundation (NO. 61504011), China postdoctoral science foundation (No. 2013M540732), and the fundamental research funds for the central universities(No.310832161002, 310832163402, 310832151090).
参考文献
[1] G. Perna, V. Capozzi, M. Ambrico, V. Augelli, T. Ligonzo, A. Minafra, L. Schiavulli, M. Pallara, Structural and optical characterization of undoped and indium-doped CdS films grown by pulsed laser deposition, Thin Solid Films, 453–454 (2004) 187-194.
ACCEPTED MANUSCRIPT [2] A.S.Opanasyuk, D.I. Kurbatov, M.M. Ivashchenko, I.Y. Protsenko, Properties of the Window Layers for the CZTSe and CZTS Based Solar Cells, Journal of Nano- and Electronic Physics, (2012). [3] J. Han, C. Liao, T. Jiang, C. Spanheimer, G. Haindl, G. Fu, V. Krishnakumar, K. Zhao, A. Klein, W. Jaegermann, An optimized multilayer structure of CdS layer for CdTe solar cells application, Journal of Alloys and Compounds, 509 (2011) 5285-5289. [4] T. Yamaguchi, Y. Yamamoto, T. Tanaka, Y. Demizu, A. Yoshida, (Cd,Zn)S thin films prepared by chemical bath deposition for photovoltaic devices, Thin Solid Films, 281–282 (1996) 375-378. [5] M.S. Hossain, M.M. Aliyu, M.A. Matin, M.A. Islam, K. Sopian, M.R. Karim, N. Amin, Prospects of ZnxCd1-xS window layer in CdTe thin film solar cells from numerical analysis, in:
Developments in Renewable Energy Technology (ICDRET), 2012 2nd International Conference on the, 2012, pp. 1-4.
[6] A. Belghachi, N. Limam, Analysis of Cd1-xZnxS window effect on CdTe solar cells, in:
Renewable and Sustainable Energy Conference (IRSEC), 2013
International, 2013, pp. 5-9. [7] S.D. Chavhan, S. Senthilarasu, S.-H. Lee, Annealing effect on the structural and optical properties of a Cd1−xZnxS thin film for photovoltaic applications, Applied Surface Science, 254 (2008) 4539-4545. [8] L.C. Burton, T.L. Hench, ZnxCd1−xS films for use in heterojunction solar cells, Applied Physics Letters, 29 (1976) 612-614. [9] M. Bedir, R. Kayali, M. Oztas, Effect of the Zn concentration on the characteristic parameters of ZnxCd1-xS thin films developed by spraying pyrolysis method under the nitrogen, Turkish Journal of Physics, 26 (2002) 121-126. [10] H. Yao, H. Shen, X. Zhu, J. Jiao, J. Li, W. Wang, Influence of Cd source concentration on photo-current response property of CdxZn1−xS film prepared by chemical bath deposition, Ceramics International, 42 (2016) 2466-2471. [11] G. Kartopu, A.J. Clayton, W.S.M. Brooks, S.D. Hodgson, V. Barrioz, A. Maertens, A.L. Dan, S.J.C. Irvine, Effect of window layer composition in Cd 1− x Zn x S/CdTe solar cells, Progress in Photovoltaics Research & Applications, 22 (2013) 18-23. [12] A. Nowshad, Y. Akira, K. Makoto, Effect of ZnTe and CdZnTe Alloys at the Back Contact of 1-µm-Thick CdTe Thin Film Solar Cells, Japanese Journal of Applied Physics, 41 (2002) 2834. [13] Y. Liu, Y. Sun, A. Rockett, A new simulation software of solar cells - WxAMPS, Solar Energy Materials and Solar Cells, 98 (2012) 124-128. [14] M. Gloeckler, A.L. Fahrenbruch, J.R. Sites, Numerical modeling of CIGS and CdTe solar cells: setting the baseline, in:
Photovoltaic Energy
Conversion, 2003. Proceedings of 3rd World Conference on, 2003, pp. 491-494 Vol.491. [15] N. Amin, K. Sopian, M. Konagai, Numerical modeling of CdS/CdTe and CdS/CdTe/ZnTe solar cells as a function of CdTe thickness, Solar Energy Materials and Solar Cells, 91 (2007) 1202-1208. [16] N. Amin, T. Isaka, A. Yamada, M. Konagai, Highly efficient 1 μm thick CdTe solar cells with textured TCOs, Solar Energy Materials and Solar Cells, 67 (2001) 195-201. [17] E.W. Jones, V. Barrioz, S.J.C. Irvine, D. Lamb, Towards ultra-thin CdTe solar cells using MOCVD, Thin Solid Films, 517 (2009) 2226-2230. [18] X. Wu, J.C. Keane, R.G. Dhere, C. Dehart, D.S. Albin, A. Duda, T.A. Gessert, S. Asher, D.H. Levi, P. Sheldon, 16.5%-efficient CdS/CdTe polycrystalline thin film solar cell, in:
17th European PVSEC, 2001, pp. 995-1000.
[19] M.S. Hossain, N. Amin, M.M. Aliyu, M.A. Matin, M.K. Siddiki, T. Razykov, K. Sopian, ZnxCd1-xS as prospective window layer in CdTe thin film solar cells from numerical analysis, in:
Photovoltaic Specialists Conference (PVSC), 2011 37th IEEE, 2011, pp. 001223-001228.
[20] P. Kumar, A. Misra, D. Kumar, N. Dhama, T.P. Sharma, P.N. Dixit, Structural and optical properties of vacuum evaporated CdxZn1−xS thin films, Optical Materials, 27 (2004) 261-264. [21] S. Chavhan, R.P. Sharma, Growth, structural and optical transport properties of nanocrystal Zn1−xCdS thin films deposited by solution growth technique (SGT) for photosensor applications, Journal of Physics and Chemistry of Solids, 66 (2005) 1721-1726. [22] T.P. Kumar, S. Saravanakumar, K. Sankaranarayanan, Effect of annealing on the surface and band gap alignment of CdZnS thin films, Applied Surface Science, 257 (2011) 1923-1927. [23] M.A. Green, Solar Cells - Operating Principles, Technology and System Application, University of NSW, Kensington, Australia, 1992. [24] N.H. Reich, W.G.J.H.M. van Sark, E.A. Alsema, R.W. Lof, R.E.I. Schropp, W.C. Sinke, W.C. Turkenburg, Crystalline silicon cell performance at low light intensities, Solar Energy Materials and Solar Cells, 93 (2009) 1471-1481.
Xiaobo Xu, Xiaoyan Wang, Wenping Gu, Si Quan, Zan Zhang PII:
S0749-6036(17)30500-1
DOI:
10.1016/j.spmi.2017.05.033
Reference:
YSPMI 5014
To appear in:
Superlattices and Microstructures
Received Date:
27 February 2017
Revised Date:
14 May 2017
Accepted Date:
14 May 2017
Please cite this article as: Xiaobo Xu, Xiaoyan Wang, Wenping Gu, Si Quan, Zan Zhang, Study on influences of CdZnS buffer layer on CdTe solar cells, Superlattices and Microstructures (2017), doi: 10.1016/j.spmi.2017.05.033
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Highlights: 1. Material parameters of Cd1-xZnxS are selected and confirmed with different x. 2. The optimal x content is found to be 0.6 for the improved performance. 3. the optimal Cd1-xZnxS width is found to be 30nm. 4. The efficiency increases from 15.42% to 17.71% with the optimized CdZnS layer.
ACCEPTED MANUSCRIPT Study on influences of CdZnS buffer layer on CdTe solar cells * Xiaobo Xu,* Xiaoyan Wang, Wenping Gu, Si Quan, Zan Zhang School of Electronic and Control Engineering, Chang'an University, Xi'an 710064, PR China
ABSTRACT: This paper deals with the Cd1-xZnxS/CdTe solar cell, with the structure parameters confirmed by basic theories, reports and experiments. The influences of x composition and the buffer layer width on the open circuit voltage, the short circuit current, the fill factor, and the conversion efficiency are discussed by numerical simulation. It is shown that the optimized buffer layer should be Cd0.4Zn0.6S, and the width is 30nm based on the simulation. The Cd1-xZnxS introduction indeed increases the blue portion absorption of solar spectrum and the efficiency of CdTe solar cells is thus improved. Key words：CdTe, solar cell, Cd1-xZnxS PACS:
73.40.Lq, 88.40.jn, 88.40.H-
1.Introduction
Over the past three decades, the Ⅱ-Ⅵ semiconductor compound Cadmium Sulfide (CdS) has been extensively investigated from the research community owing to the efficient use in the fabrication of solar cells. In traditional CdTe solar cells, the CdS thin film is applied as a window material and as a heterojunction partner, due to its high transparency, direct band gap transition, high electron affinity and n-type conductivity [1, 2]. As the solar irradiance is rich in photons with energies between 1.4 eV and 3 eV, the band gap of CdS is hence relatively low (2.42 eV) [3], the photons with energy higher than 2.42 eV may be absorbed by the CdS layer, while the Ⅱ-Ⅵ semiconductor compounds show absorption efficiency as high as 104-105 cm-1, the higher energy photons are absorbed at the surface, where the depletion region is far away and the recombination is substantial, causing little contribution to the light current. In order to improve the performance of CdTe solar cells, an appropriate window layer should be introduced. In fact, the band gap of CdS is adjustable by adding Zn to substitute Cd, in the form of ZnxCd1-xS, depending on the Zn:Cd ratio. Since ZnxCd1-xS has a larger band gap than CdS, the use of the ZnxCd1-xS window layer instead of CdS is suitable to circumvent window absorption losses as well as lattice mismatch problems associated with a quaternary solar absorber layer [4]. Although some earlier reports focused on Cd1-xZnxS/CdTe solar cells [5, 6], they were constrained to be low Zn content situation, because early growth techniques show that for the Zn concentration of x>1, the resistivity of Cd1-xZnxS is large enough to counterpart absorption improvement, leading to sharp decline of the short circuit current [7-9]. But with the advances of the growth technique, recent researches show decreased resistivity with the increase of x [10]. The best performance is observed for x=0.7 from Cd1-xZnxS solar cells deposited via the AP-MOCVD method, where the Cd1-xZnxS width is 240nm [11]. In this simulation, not only the optimized Zn content but also the Cd1-xZnxS layer width are discussed to obtain the highest efficiency.
2. CdTe Cell Design
The optimized structure is designed to be TCO/Cd1-xZnxS/CdTe/ZnTe as shown in Fig. 1. ZnTe (bandgap 2.26eV) is introduced to be inserted prior to the back contact to form the back surface field (BSF), in order to inhibit the recombination of the generated carriers by reflecting to the Cd1-xZnxS/CdTe junction [12]. Cd1-xZnxS is introduced due to the blue portion absorption of the solar spectrum for the CdS window layer, which causes a decrease in the current density of solar cells.
Corresponding author. E-mail address: [email protected] (X. Xu)
*
ACCEPTED MANUSCRIPT Back contact ZnTe
P-CdTe
N-Cd1-xZnxS TCO(SnO2) glass
Fig. 1
Optimized CdTe solar cell structure
WxAMPS [13] is used as the simulation tool. To start a simulation, the basic material parameters of SnO2 [14], CdS [14] and CdTe [12, 15] are assumed by reports and experiments, including the doping concentration, the dielectric constant, the bandgap, the mobility, the affinity, and so on, as shown in Table 1. It should be noted that although the absorption coefficient of CdTe is as high as 10-5 cm-1, and thin film CdTe layers have been successfully fabricated by different growth techniques [16, 17], the CdTe layer is normally prepared with higher film thickness in order to avoid shunting effects due to the formation of pinholes [18]. In this paper, the width is set to be 4 μm.
Table 1 Material parameters
Para.
SnO2
CdTe
ZnTe
W (μm)
0.5
4
0.2
ε/ε0
9
9.4
9.67
μn (cm2/V/s)
100
320
330
μp (cm2/V/s)
25
40
80
n,p (cm-3)
1×1017
2×1014
1×1018
Eg (eV)
3.6
1.5
2.26
NC (cm-3)
2.2×1018
8×1017
7×1016
NV (cm-3)
1.8×1019
1.8×1019
2×1019
χ
4
4.28
3.5
For the Cd1-xZnxS layer, the variable x causes the variations of material parameters. As some values are insensitive to the CdTe cell performance, such as the dielectric constant ε/ε0, the conduction band density, and the valence band density NV. It is reasonable to set them invariable for simplicity: ε/ε0=9.3, NC =2.1×1018cm-3, NV =1.7×1019cm-3. The other parameters are listed in table 2 [19].
Table 2 Material parameters of Cd1-xZnxS
Para.
x=0.05
x=0.08
x=0.1
x=0.2
x=0.3
x=0.5
x=0.6
x=0.8
x=0.9
μn (cm /V/s)
100
100
95
85
75
70
65
60
50
μp (cm /V/s)
40
40
35
30
25
20
15
10
7
Eg (eV)
2.48
2.50
2.55
2.58
2.64
2.70
3.07
3.33
3.43
χ
4.47
4.46
4.44
4.38
4.32
4.26
4.14
4.02
3.94
2 2
ACCEPTED MANUSCRIPT It should be noted that these parameters are quite different from techniques, we take the most significant bandgap as example: Eg=2.62 eV at x=0.2 for vacuum elaborated thin film [20], while Eg=2.66 eV at x=0.2 for spraying pyrolysis method [9], and the bandgap also depends on the annealing temperature [7, 21, 22], such as Eg=2.48 eV at x=0.05 as deposited but Eg=2.44 eV at x=0.05 as annealed at 400°C for 30 minutes [7]. Therefore the above values are just set to be typical, to show the influences on the cell performances. All simulations are modeled under the standard test condition (1000W/m2, 25°C ), for simplicity, the front and rear recombination rates are the same as 1×107cm/s, the front and rear reflectivities are 0.1 and 0.8, respectively. Mid-gap donor-like defect is assumed [14], and the density is set to be 1×1015cm-3, the doping concentration of Cd1-xZnxS is assumed to be 1.1×1018cm-3, the Cd1-xZnxS width is first set to be 50nm to obtain the optimal x composition. 3. Discussion Fig. 2 and Fig.3 show variations of critical parameters for different Zn contents, including the short circuit current density Jsc, the open circuit voltage Voc, the fill factor FF and the conversion efficiency η.
Fig. 2 Open circuit voltage VOC and short circuit current JSC for different ratio x
Fig. 3 Fill factor FF and conversion efficiency η for different ratio x
Fig. 4 Electron lifetime of CdTe layer
As the energy band gap of Cd1-xZnxS thin film increases with the increase of Zn concentration, the absorption edge shifts toward lower wavelength, more photons are absorbed by CdTe layer and the short circuit current density is hence improved from 23.118mA/cm2 at x=0.05 to 23.757 mA/cm2 at x=0.6 (a relative 2.76% increase). The slow variation between x=0.3 and x=0.6 is because most absorbable photons have been transmitted to the CdTe layer and the subsequent improvement is minor. It should be noted that the density drops slightly to 23.651 mA/cm2 at x=0.8. This decreased density may be due to the lattice mismatch induced recombination. As the open circuit voltage changes approximately logarithmically with the short circuit current, VOC
ACCEPTED MANUSCRIPT increases with x. The variations is just 1.4%, and the VOC can be regarded as constant, which is verified by Fig. 4 where the minority lifetime of the CdTe layer can be regarded as invariable with x. The efficiency is shown to increase from 15.42% to 16.48% during the range of x=0.05~0.6 (a relative 6.87% increase), because of the increases of the short circuit current, the open circuit voltage and the fill factor based on the equation η=VOCJSCFF /G, where G is the solar irradiation density. Values plateau beyond x=0.3 before dropping off at x>0.6. The decline is apparently because of the decreased short circuit current. As the fill factor (a relative 2.47% increase) plays a significant role on the efficiency, the fill factor should be further analyzed. Based on the empirical equation of Green [23], the fill factor is dominated by the serial resistance RS and the shunt resistance RSH. If we define the normalized open circuit voltage as vOC=VOC/(nkT/q), the ideal fill factor FF0 is expressed as
FF0
vOC ln vOC 0.72 vOC 1
(1)
Further, the specific fill factors FFs and FFsh are correlated to the serial resistance and the shunt resistance, respectively.
FFs FF0 1 rs
(2)
v 0.7 FF0 FFsh FF0 1 OC vOC rsh
(3)
with rs=RS/(VOC/ISC), rsh=RSH/(VOC/ISC). Apparently FFs decreases with the increase of Rs, while FFsh increases inversely with the increase of RSH. The serial resistance and shunt resistance dependent fill factor FF is finally expressed as [24]
v 0.7 FF0 1 rs FF FF0 1 rs 1 OC vOC rsh
(4)
Fig. 5 RS and RSH for different ratio x
The values of Rs and RSH obtained under different Zn concentrations are shown in Fig. 5. RS decreases and RSH increases dramatically with the increased x between x=0.05~0.3. According to the above models, FF shows a sharp rise, which is consistent with the curve in Fig. 3. As x increases further, both the RS and the RSH deviate little, and FF is hence stable. According to the experimental data [11], as x=0.9, RSH drops to lower than 500 Ω/cm2, and Rs grows to higher than 16 Ω/cm2, which mean a sharp deterioration of the cell performance. In our simulation by setting Rsh=410 Ω/cm2 and Rs =16.9Ω/cm2 as the input values, the corresponding results are VOC=0.9144V, JSC=23.63 mA/cm2, FF=43.9%, η=9.53%, which show similar decline trends with the experiment [11]. As x=0.6 corresponds to the highest efficiency, the Cd0.4Zn0.6S buffer layer width is next discussed for further optimization of the device. The width is changed from 10nm to 100nm to show the variations of the cell performance, in order to obtain the optimal width.
ACCEPTED MANUSCRIPT
Fig. 6 Open circuit voltage VOC and short circuit current JSC for different buffer layer widths
Fig. 7 Fill factor FF and conversion efficiency η for different buffer layer widths
As the Cd0.4Zn0.6S buffer layer width declines from 100nm to 30nm, the short circuit current increases quickly due to shorter transmission path for the absorbable photons, and the open circuit voltage and the fill factor are nearly constant, therefore the efficiency is dominated by the short circuit current and rises significantly. But with a continuous decrease of the buffer layer width, the short circuit current, the open circuit voltage and the fill factor all deteriorate for the reason of pinhole existence and the corresponding recombination induced loss, and the efficiency is hence decreased. The optimal buffer layer width of Cd0.4Zn0.6S is thus 30nm, where JSC, VOC, FF, and η are 25.468 mA/cm2, 922 mV, 75.4%, and 17.71%, respectively.
4. Conclusion
The optimized Cd1-xZnxS layer is confirmed in the CdTe solar cell design. Numerical simulations were carried out by the variations of the Zn content and the Cd1-xZnxS width. It is shown that the efficiency increases sharply with x=0.05~0.3 and keeps almost unchanged with higher x compositions, the optimal Zn concentration is 0.6; The buffer layer width also influences the cell performance, with longer photon transmission path for larger width, and pinhole formation for smaller width, the optimal Cd1ZnxS width is found to be 30nm. The research of the Cd1-xZnxS layer optimization provides constructive guidelines for the future
x
design and fabrication of CdTe solar cells. Acknowledgment The work was supported by National natural science foundation (NO. 61504011), China postdoctoral science foundation (No. 2013M540732), and the fundamental research funds for the central universities(No.310832161002, 310832163402, 310832151090).
参考文献
[1] G. Perna, V. Capozzi, M. Ambrico, V. Augelli, T. Ligonzo, A. Minafra, L. Schiavulli, M. Pallara, Structural and optical characterization of undoped and indium-doped CdS films grown by pulsed laser deposition, Thin Solid Films, 453–454 (2004) 187-194.
ACCEPTED MANUSCRIPT [2] A.S.Opanasyuk, D.I. Kurbatov, M.M. Ivashchenko, I.Y. Protsenko, Properties of the Window Layers for the CZTSe and CZTS Based Solar Cells, Journal of Nano- and Electronic Physics, (2012). [3] J. Han, C. Liao, T. Jiang, C. Spanheimer, G. Haindl, G. Fu, V. Krishnakumar, K. Zhao, A. Klein, W. Jaegermann, An optimized multilayer structure of CdS layer for CdTe solar cells application, Journal of Alloys and Compounds, 509 (2011) 5285-5289. [4] T. Yamaguchi, Y. Yamamoto, T. Tanaka, Y. Demizu, A. Yoshida, (Cd,Zn)S thin films prepared by chemical bath deposition for photovoltaic devices, Thin Solid Films, 281–282 (1996) 375-378. [5] M.S. Hossain, M.M. Aliyu, M.A. Matin, M.A. Islam, K. Sopian, M.R. Karim, N. Amin, Prospects of ZnxCd1-xS window layer in CdTe thin film solar cells from numerical analysis, in:
Developments in Renewable Energy Technology (ICDRET), 2012 2nd International Conference on the, 2012, pp. 1-4.
[6] A. Belghachi, N. Limam, Analysis of Cd1-xZnxS window effect on CdTe solar cells, in:
Renewable and Sustainable Energy Conference (IRSEC), 2013
International, 2013, pp. 5-9. [7] S.D. Chavhan, S. Senthilarasu, S.-H. Lee, Annealing effect on the structural and optical properties of a Cd1−xZnxS thin film for photovoltaic applications, Applied Surface Science, 254 (2008) 4539-4545. [8] L.C. Burton, T.L. Hench, ZnxCd1−xS films for use in heterojunction solar cells, Applied Physics Letters, 29 (1976) 612-614. [9] M. Bedir, R. Kayali, M. Oztas, Effect of the Zn concentration on the characteristic parameters of ZnxCd1-xS thin films developed by spraying pyrolysis method under the nitrogen, Turkish Journal of Physics, 26 (2002) 121-126. [10] H. Yao, H. Shen, X. Zhu, J. Jiao, J. Li, W. Wang, Influence of Cd source concentration on photo-current response property of CdxZn1−xS film prepared by chemical bath deposition, Ceramics International, 42 (2016) 2466-2471. [11] G. Kartopu, A.J. Clayton, W.S.M. Brooks, S.D. Hodgson, V. Barrioz, A. Maertens, A.L. Dan, S.J.C. Irvine, Effect of window layer composition in Cd 1− x Zn x S/CdTe solar cells, Progress in Photovoltaics Research & Applications, 22 (2013) 18-23. [12] A. Nowshad, Y. Akira, K. Makoto, Effect of ZnTe and CdZnTe Alloys at the Back Contact of 1-µm-Thick CdTe Thin Film Solar Cells, Japanese Journal of Applied Physics, 41 (2002) 2834. [13] Y. Liu, Y. Sun, A. Rockett, A new simulation software of solar cells - WxAMPS, Solar Energy Materials and Solar Cells, 98 (2012) 124-128. [14] M. Gloeckler, A.L. Fahrenbruch, J.R. Sites, Numerical modeling of CIGS and CdTe solar cells: setting the baseline, in:
Photovoltaic Energy
Conversion, 2003. Proceedings of 3rd World Conference on, 2003, pp. 491-494 Vol.491. [15] N. Amin, K. Sopian, M. Konagai, Numerical modeling of CdS/CdTe and CdS/CdTe/ZnTe solar cells as a function of CdTe thickness, Solar Energy Materials and Solar Cells, 91 (2007) 1202-1208. [16] N. Amin, T. Isaka, A. Yamada, M. Konagai, Highly efficient 1 μm thick CdTe solar cells with textured TCOs, Solar Energy Materials and Solar Cells, 67 (2001) 195-201. [17] E.W. Jones, V. Barrioz, S.J.C. Irvine, D. Lamb, Towards ultra-thin CdTe solar cells using MOCVD, Thin Solid Films, 517 (2009) 2226-2230. [18] X. Wu, J.C. Keane, R.G. Dhere, C. Dehart, D.S. Albin, A. Duda, T.A. Gessert, S. Asher, D.H. Levi, P. Sheldon, 16.5%-efficient CdS/CdTe polycrystalline thin film solar cell, in:
17th European PVSEC, 2001, pp. 995-1000.
[19] M.S. Hossain, N. Amin, M.M. Aliyu, M.A. Matin, M.K. Siddiki, T. Razykov, K. Sopian, ZnxCd1-xS as prospective window layer in CdTe thin film solar cells from numerical analysis, in:
Photovoltaic Specialists Conference (PVSC), 2011 37th IEEE, 2011, pp. 001223-001228.
[20] P. Kumar, A. Misra, D. Kumar, N. Dhama, T.P. Sharma, P.N. Dixit, Structural and optical properties of vacuum evaporated CdxZn1−xS thin films, Optical Materials, 27 (2004) 261-264. [21] S. Chavhan, R.P. Sharma, Growth, structural and optical transport properties of nanocrystal Zn1−xCdS thin films deposited by solution growth technique (SGT) for photosensor applications, Journal of Physics and Chemistry of Solids, 66 (2005) 1721-1726. [22] T.P. Kumar, S. Saravanakumar, K. Sankaranarayanan, Effect of annealing on the surface and band gap alignment of CdZnS thin films, Applied Surface Science, 257 (2011) 1923-1927. [23] M.A. Green, Solar Cells - Operating Principles, Technology and System Application, University of NSW, Kensington, Australia, 1992. [24] N.H. Reich, W.G.J.H.M. van Sark, E.A. Alsema, R.W. Lof, R.E.I. Schropp, W.C. Sinke, W.C. Turkenburg, Crystalline silicon cell performance at low light intensities, Solar Energy Materials and Solar Cells, 93 (2009) 1471-1481.