- Email: [email protected]
Simulation of high efficiency silicon solar cells with a hetero-junction microcrystalline intrinsic thin layer
Energy Conversion and Management 72 (2013) 141–146
Contents lists available at SciVerse ScienceDirect
Energy Conversion and Management journal homepage: www.elsevier.com/locate/enconman
Simulation of high efficiency silicon solar cells with a hetero-junction microcrystalline intrinsic thin layer Mehdi Hamid Vishkasougheh ⇑, Bahadir Tunaboylu Industrial Engineering, Istanbul Sßehir University, Altunizade Mah. Kusßbakısßı sok, No: 27, Üsküdar, Istanbul, Turkey
a r t i c l e
i n f o
Article history: Available online 5 April 2013 Keywords: Silicon solar cell Numerical simulation Thin films Hetero-junction
a b s t r a c t The solar cells using silicon technology have been modeled and fabricated reaching 19% cell efficiency in the past. In an effort to maximize efficiency and reduce cost to reach the grid parity, thin films of silicon are being investigated. In this study, a solar cell hetero-junction with an intrinsic thin layer (HIT) was simulated on a p-type substrate, which can be manufactured with standard silicon manufacturing processes. The influence of different parameters such as the temperature, the back surface field, different layer thicknesses, different doping concentrations for p and n type layers, ZnO and ITO as transparent conductive oxides with plane and texturized surface shapes and densities of interface defects (Dit) on the efficiency was investigated. For simulation of hetero-structures, AFORS-HET software was used in the study. Our results indicate that by optimizing different parameters of hetero-structure thin films, a high performance can be obtained using nanostructured surfaces up to an efficiency of 25% for HIT silicon solar cells. Optimized design parameters for HIT silicon solar cell for fabrication are proposed. Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction Hetero-junction with intrinsic thin-layer (HIT) solar cells, developed by Sanyo Ltd. in 1994, offer low cost fabrication for a high-efficiency solar cells compared to crystalline silicon solar cells with diffused p–n junctions [1–4]. Fabrication of HIT solar cells is comparatively simple and more importantly, does not require high temperature steps. Thinner and lower grade silicon wafers typically experience wafer bowing in the high temperature, the back-surface field (BSF) process [1,3] can be used in this lowtemperature (<200 °C) process [5]. This feature could also result in additional cost savings. HIT solar cells consist of ultra-thin amorphous silicon on crystalline silicon absorber layers. The a-Si:H/c-Si hetero-interface is of functional importance because the junction properties determine solar cell efficiency. In this investigation, a thin micro-crystalline intrinsic silicon layer is considered between the a-Si:H/c-Si hetero-interface to suppress the carrier recombination at this junction. Also the hydrogen dilution is reported to be a key deposition parameter that controls the film quality and phase formation [6,7]. Though SANYO’s original design used a n-type substrate as the absorber for the HIT solar cell, the current research concentrates on developing the HIT solar cell on a p-type substrate because of its popularity in the photovoltaic industry [8].
⇑ Corresponding author. Tel.: +90 5319246864. E-mail address: [email protected] (M.H. Vishkasougheh). 0196-8904/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.enconman.2012.10.025
Al-BSF is usually applied because of its easy fabrication process [9,3]. However, processing at a high-temperature (850 °C) could lead to the degradation of lower grade Si wafers. An alternative is to use a-Si:H(p+) to create the BSF for p-type c-Si that could be deposited at low processing temperatures (<200 °C). Efficient crystalline silicon heterojunction solar cells were fabricated on p-type wafers using amorphous silicon back contact layers [10]. In comparison, Al-BSF, an alternative material with larger conduction band offsets, could provide a much more effective mirror for the minority carrier electrons and a low back surface recombination. This leads to higher solar cell performance. However, the disadvantage of this material is that the larger band offset in the valence band edges would present a large barrier for majority carrier holes to flow through to the back contact [11]. It is well-known that the performance of a hetero-junction cell critically depends on densities of interface defects (Dit). Experimentally, the Dit can be modified by a hydrogenated termination of the Si wafer surface [12] or by passivating the defects by deposition of a thin lc-Si:H(i) layer [2]. Even if optimum parameters are utilized, solar cell efficiency is greatly limited by recombination at the transparent conductive oxide TCO/a-Si:H(n) interface. With the high work function of transparent conductive oxide (UTCO), an electron injection barrier develops at its interface with a-Si:H(n). This barrier limits the open-circuit voltage Voc by upwards band bending that develops at the TCO/a-Si:H(n) interface [13]. Hence, the value of UTCO should be carefully tuned to obtain the highest quality solar cells. We used numerical computer simulation to address this issue. We investigated the influence of (a) lc-Si:H(i) layer as buffer, (b)
142
M.H. Vishkasougheh, B. Tunaboylu / Energy Conversion and Management 72 (2013) 141–146
back surface field, such as a-Si:H(p+), (c) Dit, (d) two different transparent conductive oxide layers such as ZnO and ITO with texturized and planar surfaces and (e) temperature on solar cell performance. Finally, the optimal parameters were determined. The AFORS-HET (automat for simulation of hetero-structures) software was used as the numerical simulation tool. Its reliability was proven by many references [8,14].
2. Modeling AFORS-HET has been proven as a convenient and effective means to study the role of various parameters on the performance of HIT solar cells [5,15]. The simulated basic structure of the solar cell was TCO/a-Si:H(n)/lc-Si:H(i)/c-Si(p)/a-Si:H(p+) as shown in Fig. 1. Oxygen defects in c-Si were chosen to be located at 0.56 eV above the edge of the valence band. For the density of localized states in the band gap of amorphous silicon it has been assumed that there are both acceptor-like states and donor-like states mod-
eled by exponential band tails (Urbach tails) and Gaussian mid-gap states (associated to silicon dangling bonds) [16]. The defect-state distributions for different layers such as the a-Si:H, lc-Si:H , and the a-Si(p+) base were set with the distributions depicted in Fig. 2 [17]. Some of the simulation parameters, such as a-Si:H layers, and the c-Si(p) base are the default values in the AFORS-HET software that can be found in Table 1. Fig. 3 shows texturized front surface of TCO in a schematic view. In terms of illuminated current voltage characteristics, the global solar spectrum of 1 Sun of AM1.5 was studied with a power density of 100 mW/cm2. The device temperature was considered as 300 K. The flat band was applied to the front and back contacts to prevent a contact potential. The light reflection on the front and back contact was set to be 0.1 and 1, respectively. 3. Results and discussion 3.1. Optimization of emitter layer It was observed that as the thickness is reduced, the current becomes higher. There is difficulty in practice manufacturing a repeatable thickness value less than 3 nm, therefore the most realistic thickness is 3 nm. Using this thickness, an excellent efficiency can be obtained by the solar cell. According to Ref. [18], the thickness d of a-Si:H layer shows an optimal value in the range 5 nm < d 6 10 nm for an application as an emitter. This proves that the simulation results correlate with experimental findings. As it can be seen from Fig. 4a, it is possible to optimize a solar cell structure for achieving the highest efficiency. We have obtained efficiency levels greater than 20%. 3.2. Optimization of the intrinsic layer
Fig. 1. A schematic for the basic structure of simulated solar cell TCO/a-Si:H(n)/lcSi:H(i)/c-Si(p)/a-Si:H(p+) is shown.
While SANYO developed HIT solar cells with a very thin intrinsic a-Si:H(i) layer inserted between different type a-Si:H and c-Si, it should be noted that there is a controversy about the need of such a layer [10]. Some authors claim that it is beneficial, while others get good results without it and do not see significant improvements if they introduce it. In this investigation lc-Si:H(i) was used, one reasonable explanation for the benefit of this undoped buffer layer is that the density of states in undoped lc-Si:H is weaker than in doped a-Si:H. Therefore, we can expect to have less interface defects when the heterointerface is formed with undoped lc-Si:H rather than doped a-Si:H. AFORS-HET was used for finding the optimum thickness of this lc-Si:H intrinsic layer. Fig. 4b exhibits the effect of thickness of this layer on the solar cell external current, for a range of 3–10 nm. Solar cell external current reduces and by considering this fact that the current amount has direct impact on the final efficiency, therefore the efficiency will decrease. Our simulation results show good correlation to the experimental results from Ref. [19]. Voc depends little upon the defect density or the thickness of the intrinsic layer. Furthermore, the value of Voc is mostly controlled by the fairly simple physics of the splitting of quasi-Fermi-levels in
Fig. 2. Representations of defect state distributions of lc-Si(i) (left), a-Si(p+) (middle) and a-Si(n) (right) layers in the simulations are provided.
143
M.H. Vishkasougheh, B. Tunaboylu / Energy Conversion and Management 72 (2013) 141–146 Table 1 Some initial parameter values adopted for the HIT solar cells in the investigation. Parameters Layer thickness (nm) Dielectric constant Electron affinity (eV) Band gap (eV) Optical band gap (eV) Eff. conduction band density (cm3) Effective valance density (cm3) Electron mobility (cm2 V1 S1) Hole mobility (cm2 V1 S1) Doping concentration of accp. (cm3) Doping concentration of donors (cm3) Thermal velocity of electrons (cm s1) Thermal velocity of holes (cm s1) Layer density (g cm3) Auger recombination coefficient for electron (cm6 s1) Auger recombination coefficient for hole (cm6 s1) Direct band to band recombination coefficient (cm3 s1)
a-Si:H(n) 3–10 11.9 3.9 1.72 1.72 1 1020 1 1020 5 1 0 6 1018 1 1007 1 1007 2.328 0 0 0
Fig. 3. Schematic view of the textured TCO and the concept of light trapping due to scattering from this layer are depicted above.
the intrinsic layer. As it can be seen in Fig. 5a, the voltage is decreased by increasing the thickness of the intrinsic layer. As it can be seen in Ref. [19] this simulation result is in good agreement with the experimental data. 3.3. The influence of TCO (transparent conductive oxide) layer After optimization of the emitter and intrinsic layers, some simulation studies have been performed to clarify the impact of the front contact. Due to the low conductivity and low thickness of the amorphous emitter layer in this investigation, the application of highly conductive transparent layer was necessary to transfer the carriers to the metal contact. Radiation for texturized standard Si h1 1 1i surface was simulated by AFORS-HET software. Fig. 5b
lc-Si:H(i) 3–10 11.9 4 1.2 1.4 3 1019 2 1019 50 5 0 0 1 1007 1 1007 2.328 0 0 0
c-Si(p) 05
4 10 11.9 4.05 1.124 1.124 2.8 1019 1.04 1019 1041 412.9 1.5 1017 0 1 1007 1 1007 2.328 2.2 1031 9.9 1032 0
a-Si:H(p) 5–30 11.9 3.9 1.72 1.72 1 1020 1 1020 5 1 1 1020 0 1 1007 1 1007 2.328 0 0 0
Fig. 4. The dependence of external current of the structure as a function of the thickness of emitter (a) and the thickness of intrinsic layer (b) are shown.
exhibits the final efficiency of evaluated cell for ZnO and ITO with texturized and planar surfaces. As shown in Fig. 5b, the highest efficiency is for ZnO with a texturized surface solar cell with 24.5% and a planar structure with 23.33% efficiency and for ITO texturized and planar surface structured layers, the efficiency is 22% and 20.99%, respectively. Table 2 is giving the necessary information about the fill factor (FF), open-circuit voltage (Voc), short-circuit current (Jsc) and efficiency (g) for each of the mentioned cells in the Fig. 5b. The efficiency of the cell is the power density delivered at operating point as a function of the incident light power density, Ps, as shown in the following equation:
g ¼ Jsc V oc FF=Ps
ð1Þ
144
M.H. Vishkasougheh, B. Tunaboylu / Energy Conversion and Management 72 (2013) 141–146
Fig. 5. The dependence of voltage on the thickness of different transparent conductive oxide layers (a) and the efficiency (b) are provided.
Table 2 The fill factor, open circuit voltage, short circuit current and the efficiency of each cell. Parameter/cell
Voc (mV)
Jsc (mA/cm2)
FF%
Eff%
Texturized ZnO Plane ZnO Texturized ITO Plane ITO
626 624.9 623.1 622
47.4 45.28 42.84 40.98
82.49 82.46 82.4 82.35
24.48 23.33 22 20.99
3.4. The effect of defects in c-Si(p) layer on the efficiency As the interface defect density increases, the carrier recombination probability at the interface increases, leading to the increase in reverse saturation current and reducing the open circuit voltage and the fill factor. Therefore, the efficiency decreases as well. In order to obtain high-efficiency HIT solar cells, a surface passivation method is essential, such as plasma-assisted H passivation, in order to control interface state defect density down to 1010 cm2 V1. Fig. 6a exhibits the influence of interface state defect density on the efficiency.
Fig. 6. The impact of interface defect density in c-Si(p) on solar cell efficiency (a) and the dependency of efficiency on the BSF doping concentration (b) are illustrated above.
3.6. The effect of the BSF thickness Fig. 7a exhibits the effect of the BSF thickness on the efficiency characteristics of HIT solar cell. We can see that when the thickness increases, the open circuit voltage and short circuit current are nearly unchanged and therefore the efficiency does not change with increasing thickness. As a result, the efficiency can be considered as independent of the thickness, which is very beneficial. When the manufacturing technology is considered, one difficulty in fabricating microcrystalline silicon thin films is that the film growth rate is very low. If the required thickness is very thin, the production time can be dramatically saved. Therefore, the BSF thickness can be set to 5 nm.
3.5. The effect of a-Si(p+) BSF doping concentration on the device efficiency Fig. 6b shows the influence of BSF doping concentration on the photovoltaic characteristics of solar cell. In this graph, we can see that the doping concentration must reach a certain value, preferably more than 1020 cm3 to observe good conversion efficiency. When the doping concentration increases, the open circuit voltage, short circuit current, fill factor all increase. This can be attributed to the BSF band structure. There are two reasons that explain why high doping concentration is a guarantee of good BSF. First, when the doping concentration is low, the reflection role of BSF is not clear. Second, the barrier on the carrier transport can be reduced by increasing the doping concentration.
Fig. 7. The effect of thickness on the efficiency is shown.
M.H. Vishkasougheh, B. Tunaboylu / Energy Conversion and Management 72 (2013) 141–146
3.7. The effect of temperature on the efficiency
145
gap linearly extrapolated to absolute zero and B is a constant which is essentially independent of temperature. Substituting these equations back into the expression for I0, and assuming that the temperature dependencies of the other parameters can be neglected, gives the following equation:
One of the important factors of solar cell efficiency is the temperature. In this investigation, the impact of temperature was considered on the mentioned HIT solar cell structure. As it can be seen in Fig. 8, the efficiency will decrease gradually as the temperature is raised from 300 K to 343 K. The parameter most affected by an increase in temperature is the open-circuit voltage. As the temperature is increased, the equilibrium population of electrons ni increases exponentially, increasing the dark saturation current density (I0). Note that this effect can be stronger for the diffusion component of the dark current than that for the recombination–generation, because of stronger dependence on ni (Eqs. (2)–(4)). The increased dark current reduces Voc according to the Eq. (5). At the same time, increased temperature reduces the band gap and increases the photocurrent, since lower energy photons can now be absorbed. The net effect is a reduction in the efficiency because the loss in Voc outweighs the gain in Jsc [20]. This is illustrated in Fig. 9. The equation for I0 from one side of p–n junction is defined by the the following equation:
where Isc is the short-circuit current. The impact of increasing temperature is shown in the Fig. 9. Also Our simulation results show good correlation to the experimental results from Ref. [21].
I0 ¼ qADn2i =LN 0
4. Conclusions
ð2Þ
where q is the electronic charge, D is the diffusivity of majority carrier, L is the diffusion length of the minority carrier, N0 is doping and ni is the intrinsic carrier concentration. 2
N2i ¼ 4ð2pKT=h Þ3 ðme mþh Þ3=2 BT 3 expðEG0 =KTÞ
ð3Þ
where T is the temperature, h and k are constants, me and mh are the effective masses of electrons and holes respectively. EG0 is the band
I0 ¼ qAD=LN 0 B0 T g expðEG0 =KTÞ
ð4Þ
where B0 is a temperature independent constant. A constant, g is used instead of the number 3 to incorporate the possible temperature dependencies of the other material parameters. For silicon solar cells near room temperature, I0 approximately doubles for every 10 °C increase in temperature. Also it is known Voc has dependency to I0 according to the following equation:
V oc ¼
KT lnðIsc =IU Þ Q
ð5Þ
In this study, the impact of various parameters such as the thickness of layers, the doping concentration, the surface condition of TCO, temperature and the interface defect density on the efficiency were investigated. This simulation study targeted a set of conditions using a structure consisting an a-Si(n) emitter and lc-Si:H intrinsic layers with 3 nm thickness, with a max 1010 cm3 V1 defect density in c-Si(p) layer, having doping concentration around 1020 cm3 with 5 nm thickness in the BSF layer and a texturized ZnO transparent conductive oxide. Also the temperature dependency of proposed structure has been discussed. It has been shown that by increasing the temperature the efficiency will decrease as well. In comparison with the literature provided earlier, the simulation results show good correlation with experimental results. Our proposed multilayer structure with different optimized n and p type layers provides a promising path which enables a silicon solar cell device with 25% efficiency using a hetero-junction intrinsic thin layer. Acknowledgements The authors wish to thank Professor Vural Aksakallı for his helpful discussions and Istanbul S ß ehir University for the support to this study. References
Fig. 8. The effect of temperature on the efficiency is shown.
Fig. 9. Effect of temperature on p–n junction J(V) characteristic. The arrow indicates the direction of increasing intensity or temperature [20].
[1] Tucci M, Noce MD, Bobeico E, Roca F, Cesare GD, Palma F. Comparison of amorphous/crystalline heterojunction solar cells based on n- and p-type crystalline silicon. Thin Solid Films 2004;355:451–2. [2] Voz C, Munoz D, Fonrodona M, Martin I, Puigdollers J, Alcubilla R, et al. Bifacial heterojunction silicon solar cells by hot-wire CVD with open-circuit voltages exceeding 600 mV. Thin Solid Films 2006;415:511–2. [3] Xu Y, Hu ZH, Diao HW, Cai Y, Zhang SB, Zeng XB, et al. Heterojunction solar cells with n-type nanocrystalline silicon emitters on p-type c-Si wafers. J NonCryst Solids 2006;352:1972. [4] Kleider JP, Chouffot R, Gudovskikh AS, Cabarrocas PR, Labrune M. Electronic and structural properties of the amorphous crystalline silicon interface. Thin Solid Films 2009;517:6386–91. [5] MiKio T, Hitoshi S, Yukihiro Y, Eiji M, Akira T, Makoto T. An approach for the higher efficiency in the HIT cells[C]. In: Thirty first IEEE photovoltaic specialists conference; 2005. p. 866–71. [6] Froitzheim A, Angermann H, Brendel K, Elstner L, Füssel W, Kliefoth K, et al. Simulation and study of the influence of the buffer intrinsic layer. In: Proceedings of the seventeenth European photovoltaic solar energy conference. München, Germany; 2000. p. 1238. [7] Froitzheim A, Brendel K, Elstner L, Fuhs W, Kliefoth K, Schmidt M. Interface recombination in heterojunctions of amorphous and crystalline silicon. J NonCryst Solids 2002;299–302:663–7.
146
M.H. Vishkasougheh, B. Tunaboylu / Energy Conversion and Management 72 (2013) 141–146
[8] Zhao L, Zhou CL, Li HL, Diao HW, Wang WJ. Sol Energy Mater Sol Cells 2008;92:673–81. [9] Tucci M, Cesare GD. 17% efficiency heterostructure solar cell based on p-type crystalline silicon. J Non-Cryst Solids 2004;663:338–40. [10] Wang Qi, Page MR, Iwaniczko E, Xu Yueqin, Roybal L, Bauer R, et al. Efficient heterojunction solar cells on p-type crystal silicon wafers. Appl Phys Lett 2010;96:013507. [11] Wang TH, Page MR, Iwaniczko E, Levi DH, Yan Y, Branz HM, Wang Q. Toward better understanding and Improved Performance of silicon heterojunction solar cells. In: Fourteenth workshop on crystalline silicon solar cells and modules winter park. Colorado, USA; 2004. p. 1. [12] Ferreira GM, Ferlauto AS, Chen C, Koval RJ, Pearce JM, Ross C, et al. Kinetics of silicon film growth and the deposition phase diagram. J Non-Cryst Solids 2004;13:338–40. [13] Centurioni E, Iencinella D. Role of front contact work function on amorphous silicon/crystalline silicon heterojunction solar cell performance. IEEE Electron Dev Lett 2003;24:177–9. [14] Ren BY, Zhang Y, Guo B, Zhang B, Li HY, Xu Y, et al. Acta Energy Solaris Sinica 2008;29(9):1112–6. [15] Stangl R, Froitzheim A, Schmidt M, Fuhs W. Design criteria for amorphous/ crystalline silicon heterojunction solar cells – a simulation study. In: Kurokawa K, [u.a.] editor. Proceedings of the third world conference on
[16] [17]
[18]
[19]
[20] [21]
photovoltaic energy conversion: joint conference of thirteenth PV science & engineering conference. Thirtieth IEEE PV specialists conference. Eighteenth European PV solar energy conference. Osaka International Congress Center ‘‘Grand Cube’’, Osaka, Japan, 11–18 May 2003/WCPEC-3. vol. 2. Osaka; 2003. p. 1005–8. Norberto HC, Arturo MA. Sol Energy Mater Sol Cells 2010;94:62–7. Dao VA, Heo J, Choi H, Kim Y, Park S, Jung S, et al. Simulation and study of the influence of the buffer intrinsic layer, back-surface field, densities of interface defects, resistivity of p-type silicon substrate and transparent conductive oxide on heterojunction with intrinsic thin-layer (HIT) solar cell. Sol Energy 2010;84:777–83. Schmidt M, Korte L, Laades A, Stangl R, Schubert Ch, Angermann H, et al. Physical aspects of a-Si:H/C-Si hetro-junction solar cells. Thin Solid Films 2007;515:7475–80. Sawada T, Terada N, Tsuge S, Baba T, Takahama T, Wakisaka K, et al. Highefficiency a-Si:H/c-Si heterojunction solar cells. In: Proceedings of the first WCPEC conference. Hawaii, USA; 1994. p. 1219–26. Nelson Jenny. The physics of solar cells. UK: Imperial college press; 2010. Hamadeh H. Temperature dependence of pin solar cell parameters with intrinsic layers made of pm-Si:H and low crystalline volume fraction mc-Si:H. Renewable Energy 2010;35:1419–23.
Contents lists available at SciVerse ScienceDirect
Energy Conversion and Management journal homepage: www.elsevier.com/locate/enconman
Simulation of high efficiency silicon solar cells with a hetero-junction microcrystalline intrinsic thin layer Mehdi Hamid Vishkasougheh ⇑, Bahadir Tunaboylu Industrial Engineering, Istanbul Sßehir University, Altunizade Mah. Kusßbakısßı sok, No: 27, Üsküdar, Istanbul, Turkey
a r t i c l e
i n f o
Article history: Available online 5 April 2013 Keywords: Silicon solar cell Numerical simulation Thin films Hetero-junction
a b s t r a c t The solar cells using silicon technology have been modeled and fabricated reaching 19% cell efficiency in the past. In an effort to maximize efficiency and reduce cost to reach the grid parity, thin films of silicon are being investigated. In this study, a solar cell hetero-junction with an intrinsic thin layer (HIT) was simulated on a p-type substrate, which can be manufactured with standard silicon manufacturing processes. The influence of different parameters such as the temperature, the back surface field, different layer thicknesses, different doping concentrations for p and n type layers, ZnO and ITO as transparent conductive oxides with plane and texturized surface shapes and densities of interface defects (Dit) on the efficiency was investigated. For simulation of hetero-structures, AFORS-HET software was used in the study. Our results indicate that by optimizing different parameters of hetero-structure thin films, a high performance can be obtained using nanostructured surfaces up to an efficiency of 25% for HIT silicon solar cells. Optimized design parameters for HIT silicon solar cell for fabrication are proposed. Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction Hetero-junction with intrinsic thin-layer (HIT) solar cells, developed by Sanyo Ltd. in 1994, offer low cost fabrication for a high-efficiency solar cells compared to crystalline silicon solar cells with diffused p–n junctions [1–4]. Fabrication of HIT solar cells is comparatively simple and more importantly, does not require high temperature steps. Thinner and lower grade silicon wafers typically experience wafer bowing in the high temperature, the back-surface field (BSF) process [1,3] can be used in this lowtemperature (<200 °C) process [5]. This feature could also result in additional cost savings. HIT solar cells consist of ultra-thin amorphous silicon on crystalline silicon absorber layers. The a-Si:H/c-Si hetero-interface is of functional importance because the junction properties determine solar cell efficiency. In this investigation, a thin micro-crystalline intrinsic silicon layer is considered between the a-Si:H/c-Si hetero-interface to suppress the carrier recombination at this junction. Also the hydrogen dilution is reported to be a key deposition parameter that controls the film quality and phase formation [6,7]. Though SANYO’s original design used a n-type substrate as the absorber for the HIT solar cell, the current research concentrates on developing the HIT solar cell on a p-type substrate because of its popularity in the photovoltaic industry [8].
⇑ Corresponding author. Tel.: +90 5319246864. E-mail address: [email protected] (M.H. Vishkasougheh). 0196-8904/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.enconman.2012.10.025
Al-BSF is usually applied because of its easy fabrication process [9,3]. However, processing at a high-temperature (850 °C) could lead to the degradation of lower grade Si wafers. An alternative is to use a-Si:H(p+) to create the BSF for p-type c-Si that could be deposited at low processing temperatures (<200 °C). Efficient crystalline silicon heterojunction solar cells were fabricated on p-type wafers using amorphous silicon back contact layers [10]. In comparison, Al-BSF, an alternative material with larger conduction band offsets, could provide a much more effective mirror for the minority carrier electrons and a low back surface recombination. This leads to higher solar cell performance. However, the disadvantage of this material is that the larger band offset in the valence band edges would present a large barrier for majority carrier holes to flow through to the back contact [11]. It is well-known that the performance of a hetero-junction cell critically depends on densities of interface defects (Dit). Experimentally, the Dit can be modified by a hydrogenated termination of the Si wafer surface [12] or by passivating the defects by deposition of a thin lc-Si:H(i) layer [2]. Even if optimum parameters are utilized, solar cell efficiency is greatly limited by recombination at the transparent conductive oxide TCO/a-Si:H(n) interface. With the high work function of transparent conductive oxide (UTCO), an electron injection barrier develops at its interface with a-Si:H(n). This barrier limits the open-circuit voltage Voc by upwards band bending that develops at the TCO/a-Si:H(n) interface [13]. Hence, the value of UTCO should be carefully tuned to obtain the highest quality solar cells. We used numerical computer simulation to address this issue. We investigated the influence of (a) lc-Si:H(i) layer as buffer, (b)
142
M.H. Vishkasougheh, B. Tunaboylu / Energy Conversion and Management 72 (2013) 141–146
back surface field, such as a-Si:H(p+), (c) Dit, (d) two different transparent conductive oxide layers such as ZnO and ITO with texturized and planar surfaces and (e) temperature on solar cell performance. Finally, the optimal parameters were determined. The AFORS-HET (automat for simulation of hetero-structures) software was used as the numerical simulation tool. Its reliability was proven by many references [8,14].
2. Modeling AFORS-HET has been proven as a convenient and effective means to study the role of various parameters on the performance of HIT solar cells [5,15]. The simulated basic structure of the solar cell was TCO/a-Si:H(n)/lc-Si:H(i)/c-Si(p)/a-Si:H(p+) as shown in Fig. 1. Oxygen defects in c-Si were chosen to be located at 0.56 eV above the edge of the valence band. For the density of localized states in the band gap of amorphous silicon it has been assumed that there are both acceptor-like states and donor-like states mod-
eled by exponential band tails (Urbach tails) and Gaussian mid-gap states (associated to silicon dangling bonds) [16]. The defect-state distributions for different layers such as the a-Si:H, lc-Si:H , and the a-Si(p+) base were set with the distributions depicted in Fig. 2 [17]. Some of the simulation parameters, such as a-Si:H layers, and the c-Si(p) base are the default values in the AFORS-HET software that can be found in Table 1. Fig. 3 shows texturized front surface of TCO in a schematic view. In terms of illuminated current voltage characteristics, the global solar spectrum of 1 Sun of AM1.5 was studied with a power density of 100 mW/cm2. The device temperature was considered as 300 K. The flat band was applied to the front and back contacts to prevent a contact potential. The light reflection on the front and back contact was set to be 0.1 and 1, respectively. 3. Results and discussion 3.1. Optimization of emitter layer It was observed that as the thickness is reduced, the current becomes higher. There is difficulty in practice manufacturing a repeatable thickness value less than 3 nm, therefore the most realistic thickness is 3 nm. Using this thickness, an excellent efficiency can be obtained by the solar cell. According to Ref. [18], the thickness d of a-Si:H layer shows an optimal value in the range 5 nm < d 6 10 nm for an application as an emitter. This proves that the simulation results correlate with experimental findings. As it can be seen from Fig. 4a, it is possible to optimize a solar cell structure for achieving the highest efficiency. We have obtained efficiency levels greater than 20%. 3.2. Optimization of the intrinsic layer
Fig. 1. A schematic for the basic structure of simulated solar cell TCO/a-Si:H(n)/lcSi:H(i)/c-Si(p)/a-Si:H(p+) is shown.
While SANYO developed HIT solar cells with a very thin intrinsic a-Si:H(i) layer inserted between different type a-Si:H and c-Si, it should be noted that there is a controversy about the need of such a layer [10]. Some authors claim that it is beneficial, while others get good results without it and do not see significant improvements if they introduce it. In this investigation lc-Si:H(i) was used, one reasonable explanation for the benefit of this undoped buffer layer is that the density of states in undoped lc-Si:H is weaker than in doped a-Si:H. Therefore, we can expect to have less interface defects when the heterointerface is formed with undoped lc-Si:H rather than doped a-Si:H. AFORS-HET was used for finding the optimum thickness of this lc-Si:H intrinsic layer. Fig. 4b exhibits the effect of thickness of this layer on the solar cell external current, for a range of 3–10 nm. Solar cell external current reduces and by considering this fact that the current amount has direct impact on the final efficiency, therefore the efficiency will decrease. Our simulation results show good correlation to the experimental results from Ref. [19]. Voc depends little upon the defect density or the thickness of the intrinsic layer. Furthermore, the value of Voc is mostly controlled by the fairly simple physics of the splitting of quasi-Fermi-levels in
Fig. 2. Representations of defect state distributions of lc-Si(i) (left), a-Si(p+) (middle) and a-Si(n) (right) layers in the simulations are provided.
143
M.H. Vishkasougheh, B. Tunaboylu / Energy Conversion and Management 72 (2013) 141–146 Table 1 Some initial parameter values adopted for the HIT solar cells in the investigation. Parameters Layer thickness (nm) Dielectric constant Electron affinity (eV) Band gap (eV) Optical band gap (eV) Eff. conduction band density (cm3) Effective valance density (cm3) Electron mobility (cm2 V1 S1) Hole mobility (cm2 V1 S1) Doping concentration of accp. (cm3) Doping concentration of donors (cm3) Thermal velocity of electrons (cm s1) Thermal velocity of holes (cm s1) Layer density (g cm3) Auger recombination coefficient for electron (cm6 s1) Auger recombination coefficient for hole (cm6 s1) Direct band to band recombination coefficient (cm3 s1)
a-Si:H(n) 3–10 11.9 3.9 1.72 1.72 1 1020 1 1020 5 1 0 6 1018 1 1007 1 1007 2.328 0 0 0
Fig. 3. Schematic view of the textured TCO and the concept of light trapping due to scattering from this layer are depicted above.
the intrinsic layer. As it can be seen in Fig. 5a, the voltage is decreased by increasing the thickness of the intrinsic layer. As it can be seen in Ref. [19] this simulation result is in good agreement with the experimental data. 3.3. The influence of TCO (transparent conductive oxide) layer After optimization of the emitter and intrinsic layers, some simulation studies have been performed to clarify the impact of the front contact. Due to the low conductivity and low thickness of the amorphous emitter layer in this investigation, the application of highly conductive transparent layer was necessary to transfer the carriers to the metal contact. Radiation for texturized standard Si h1 1 1i surface was simulated by AFORS-HET software. Fig. 5b
lc-Si:H(i) 3–10 11.9 4 1.2 1.4 3 1019 2 1019 50 5 0 0 1 1007 1 1007 2.328 0 0 0
c-Si(p) 05
4 10 11.9 4.05 1.124 1.124 2.8 1019 1.04 1019 1041 412.9 1.5 1017 0 1 1007 1 1007 2.328 2.2 1031 9.9 1032 0
a-Si:H(p) 5–30 11.9 3.9 1.72 1.72 1 1020 1 1020 5 1 1 1020 0 1 1007 1 1007 2.328 0 0 0
Fig. 4. The dependence of external current of the structure as a function of the thickness of emitter (a) and the thickness of intrinsic layer (b) are shown.
exhibits the final efficiency of evaluated cell for ZnO and ITO with texturized and planar surfaces. As shown in Fig. 5b, the highest efficiency is for ZnO with a texturized surface solar cell with 24.5% and a planar structure with 23.33% efficiency and for ITO texturized and planar surface structured layers, the efficiency is 22% and 20.99%, respectively. Table 2 is giving the necessary information about the fill factor (FF), open-circuit voltage (Voc), short-circuit current (Jsc) and efficiency (g) for each of the mentioned cells in the Fig. 5b. The efficiency of the cell is the power density delivered at operating point as a function of the incident light power density, Ps, as shown in the following equation:
g ¼ Jsc V oc FF=Ps
ð1Þ
144
M.H. Vishkasougheh, B. Tunaboylu / Energy Conversion and Management 72 (2013) 141–146
Fig. 5. The dependence of voltage on the thickness of different transparent conductive oxide layers (a) and the efficiency (b) are provided.
Table 2 The fill factor, open circuit voltage, short circuit current and the efficiency of each cell. Parameter/cell
Voc (mV)
Jsc (mA/cm2)
FF%
Eff%
Texturized ZnO Plane ZnO Texturized ITO Plane ITO
626 624.9 623.1 622
47.4 45.28 42.84 40.98
82.49 82.46 82.4 82.35
24.48 23.33 22 20.99
3.4. The effect of defects in c-Si(p) layer on the efficiency As the interface defect density increases, the carrier recombination probability at the interface increases, leading to the increase in reverse saturation current and reducing the open circuit voltage and the fill factor. Therefore, the efficiency decreases as well. In order to obtain high-efficiency HIT solar cells, a surface passivation method is essential, such as plasma-assisted H passivation, in order to control interface state defect density down to 1010 cm2 V1. Fig. 6a exhibits the influence of interface state defect density on the efficiency.
Fig. 6. The impact of interface defect density in c-Si(p) on solar cell efficiency (a) and the dependency of efficiency on the BSF doping concentration (b) are illustrated above.
3.6. The effect of the BSF thickness Fig. 7a exhibits the effect of the BSF thickness on the efficiency characteristics of HIT solar cell. We can see that when the thickness increases, the open circuit voltage and short circuit current are nearly unchanged and therefore the efficiency does not change with increasing thickness. As a result, the efficiency can be considered as independent of the thickness, which is very beneficial. When the manufacturing technology is considered, one difficulty in fabricating microcrystalline silicon thin films is that the film growth rate is very low. If the required thickness is very thin, the production time can be dramatically saved. Therefore, the BSF thickness can be set to 5 nm.
3.5. The effect of a-Si(p+) BSF doping concentration on the device efficiency Fig. 6b shows the influence of BSF doping concentration on the photovoltaic characteristics of solar cell. In this graph, we can see that the doping concentration must reach a certain value, preferably more than 1020 cm3 to observe good conversion efficiency. When the doping concentration increases, the open circuit voltage, short circuit current, fill factor all increase. This can be attributed to the BSF band structure. There are two reasons that explain why high doping concentration is a guarantee of good BSF. First, when the doping concentration is low, the reflection role of BSF is not clear. Second, the barrier on the carrier transport can be reduced by increasing the doping concentration.
Fig. 7. The effect of thickness on the efficiency is shown.
M.H. Vishkasougheh, B. Tunaboylu / Energy Conversion and Management 72 (2013) 141–146
3.7. The effect of temperature on the efficiency
145
gap linearly extrapolated to absolute zero and B is a constant which is essentially independent of temperature. Substituting these equations back into the expression for I0, and assuming that the temperature dependencies of the other parameters can be neglected, gives the following equation:
One of the important factors of solar cell efficiency is the temperature. In this investigation, the impact of temperature was considered on the mentioned HIT solar cell structure. As it can be seen in Fig. 8, the efficiency will decrease gradually as the temperature is raised from 300 K to 343 K. The parameter most affected by an increase in temperature is the open-circuit voltage. As the temperature is increased, the equilibrium population of electrons ni increases exponentially, increasing the dark saturation current density (I0). Note that this effect can be stronger for the diffusion component of the dark current than that for the recombination–generation, because of stronger dependence on ni (Eqs. (2)–(4)). The increased dark current reduces Voc according to the Eq. (5). At the same time, increased temperature reduces the band gap and increases the photocurrent, since lower energy photons can now be absorbed. The net effect is a reduction in the efficiency because the loss in Voc outweighs the gain in Jsc [20]. This is illustrated in Fig. 9. The equation for I0 from one side of p–n junction is defined by the the following equation:
where Isc is the short-circuit current. The impact of increasing temperature is shown in the Fig. 9. Also Our simulation results show good correlation to the experimental results from Ref. [21].
I0 ¼ qADn2i =LN 0
4. Conclusions
ð2Þ
where q is the electronic charge, D is the diffusivity of majority carrier, L is the diffusion length of the minority carrier, N0 is doping and ni is the intrinsic carrier concentration. 2
N2i ¼ 4ð2pKT=h Þ3 ðme mþh Þ3=2 BT 3 expðEG0 =KTÞ
ð3Þ
where T is the temperature, h and k are constants, me and mh are the effective masses of electrons and holes respectively. EG0 is the band
I0 ¼ qAD=LN 0 B0 T g expðEG0 =KTÞ
ð4Þ
where B0 is a temperature independent constant. A constant, g is used instead of the number 3 to incorporate the possible temperature dependencies of the other material parameters. For silicon solar cells near room temperature, I0 approximately doubles for every 10 °C increase in temperature. Also it is known Voc has dependency to I0 according to the following equation:
V oc ¼
KT lnðIsc =IU Þ Q
ð5Þ
In this study, the impact of various parameters such as the thickness of layers, the doping concentration, the surface condition of TCO, temperature and the interface defect density on the efficiency were investigated. This simulation study targeted a set of conditions using a structure consisting an a-Si(n) emitter and lc-Si:H intrinsic layers with 3 nm thickness, with a max 1010 cm3 V1 defect density in c-Si(p) layer, having doping concentration around 1020 cm3 with 5 nm thickness in the BSF layer and a texturized ZnO transparent conductive oxide. Also the temperature dependency of proposed structure has been discussed. It has been shown that by increasing the temperature the efficiency will decrease as well. In comparison with the literature provided earlier, the simulation results show good correlation with experimental results. Our proposed multilayer structure with different optimized n and p type layers provides a promising path which enables a silicon solar cell device with 25% efficiency using a hetero-junction intrinsic thin layer. Acknowledgements The authors wish to thank Professor Vural Aksakallı for his helpful discussions and Istanbul S ß ehir University for the support to this study. References
Fig. 8. The effect of temperature on the efficiency is shown.
Fig. 9. Effect of temperature on p–n junction J(V) characteristic. The arrow indicates the direction of increasing intensity or temperature [20].
[1] Tucci M, Noce MD, Bobeico E, Roca F, Cesare GD, Palma F. Comparison of amorphous/crystalline heterojunction solar cells based on n- and p-type crystalline silicon. Thin Solid Films 2004;355:451–2. [2] Voz C, Munoz D, Fonrodona M, Martin I, Puigdollers J, Alcubilla R, et al. Bifacial heterojunction silicon solar cells by hot-wire CVD with open-circuit voltages exceeding 600 mV. Thin Solid Films 2006;415:511–2. [3] Xu Y, Hu ZH, Diao HW, Cai Y, Zhang SB, Zeng XB, et al. Heterojunction solar cells with n-type nanocrystalline silicon emitters on p-type c-Si wafers. J NonCryst Solids 2006;352:1972. [4] Kleider JP, Chouffot R, Gudovskikh AS, Cabarrocas PR, Labrune M. Electronic and structural properties of the amorphous crystalline silicon interface. Thin Solid Films 2009;517:6386–91. [5] MiKio T, Hitoshi S, Yukihiro Y, Eiji M, Akira T, Makoto T. An approach for the higher efficiency in the HIT cells[C]. In: Thirty first IEEE photovoltaic specialists conference; 2005. p. 866–71. [6] Froitzheim A, Angermann H, Brendel K, Elstner L, Füssel W, Kliefoth K, et al. Simulation and study of the influence of the buffer intrinsic layer. In: Proceedings of the seventeenth European photovoltaic solar energy conference. München, Germany; 2000. p. 1238. [7] Froitzheim A, Brendel K, Elstner L, Fuhs W, Kliefoth K, Schmidt M. Interface recombination in heterojunctions of amorphous and crystalline silicon. J NonCryst Solids 2002;299–302:663–7.
146
M.H. Vishkasougheh, B. Tunaboylu / Energy Conversion and Management 72 (2013) 141–146
[8] Zhao L, Zhou CL, Li HL, Diao HW, Wang WJ. Sol Energy Mater Sol Cells 2008;92:673–81. [9] Tucci M, Cesare GD. 17% efficiency heterostructure solar cell based on p-type crystalline silicon. J Non-Cryst Solids 2004;663:338–40. [10] Wang Qi, Page MR, Iwaniczko E, Xu Yueqin, Roybal L, Bauer R, et al. Efficient heterojunction solar cells on p-type crystal silicon wafers. Appl Phys Lett 2010;96:013507. [11] Wang TH, Page MR, Iwaniczko E, Levi DH, Yan Y, Branz HM, Wang Q. Toward better understanding and Improved Performance of silicon heterojunction solar cells. In: Fourteenth workshop on crystalline silicon solar cells and modules winter park. Colorado, USA; 2004. p. 1. [12] Ferreira GM, Ferlauto AS, Chen C, Koval RJ, Pearce JM, Ross C, et al. Kinetics of silicon film growth and the deposition phase diagram. J Non-Cryst Solids 2004;13:338–40. [13] Centurioni E, Iencinella D. Role of front contact work function on amorphous silicon/crystalline silicon heterojunction solar cell performance. IEEE Electron Dev Lett 2003;24:177–9. [14] Ren BY, Zhang Y, Guo B, Zhang B, Li HY, Xu Y, et al. Acta Energy Solaris Sinica 2008;29(9):1112–6. [15] Stangl R, Froitzheim A, Schmidt M, Fuhs W. Design criteria for amorphous/ crystalline silicon heterojunction solar cells – a simulation study. In: Kurokawa K, [u.a.] editor. Proceedings of the third world conference on
[16] [17]
[18]
[19]
[20] [21]
photovoltaic energy conversion: joint conference of thirteenth PV science & engineering conference. Thirtieth IEEE PV specialists conference. Eighteenth European PV solar energy conference. Osaka International Congress Center ‘‘Grand Cube’’, Osaka, Japan, 11–18 May 2003/WCPEC-3. vol. 2. Osaka; 2003. p. 1005–8. Norberto HC, Arturo MA. Sol Energy Mater Sol Cells 2010;94:62–7. Dao VA, Heo J, Choi H, Kim Y, Park S, Jung S, et al. Simulation and study of the influence of the buffer intrinsic layer, back-surface field, densities of interface defects, resistivity of p-type silicon substrate and transparent conductive oxide on heterojunction with intrinsic thin-layer (HIT) solar cell. Sol Energy 2010;84:777–83. Schmidt M, Korte L, Laades A, Stangl R, Schubert Ch, Angermann H, et al. Physical aspects of a-Si:H/C-Si hetro-junction solar cells. Thin Solid Films 2007;515:7475–80. Sawada T, Terada N, Tsuge S, Baba T, Takahama T, Wakisaka K, et al. Highefficiency a-Si:H/c-Si heterojunction solar cells. In: Proceedings of the first WCPEC conference. Hawaii, USA; 1994. p. 1219–26. Nelson Jenny. The physics of solar cells. UK: Imperial college press; 2010. Hamadeh H. Temperature dependence of pin solar cell parameters with intrinsic layers made of pm-Si:H and low crystalline volume fraction mc-Si:H. Renewable Energy 2010;35:1419–23.