Theoretical simulation of performances in CIGS thin-film solar cells with cadmium-free buffer layer

    Corresponding author: Fang Wang,
  • 1. College of Information Science and Technology, Nanjing Forestry University, Nanjing 210037, China
  • 2. School of Science, Nanjing Forestry University, Nanjing 210037, China

Key words: solar cellshetero junction structurecadmium-free buffer layer

Abstract: Copper indium gallium selenium (CIGS) thin film solar cells have become one of the hottest topics in solar energy due to their high photoelectric transformation efficiency. To real applications, CIGS thin film is covered by the buffer layer and absorption layer. Traditionally, cadmium sulfide (CdS) is inserted into the middle of the window layer (ZnO) and absorption layer (CIGS) as a buffer layer. However, the application of the GIGS/CdS thin film solar cells has been limited because of the environmental pollution resulting from the toxic cadmium atom. Although zinc sulfide (ZnS) has been proposed to be one of the candidates, the performance of such battery cells has not been investigated. Here, in this paper, we systematically study the possibility of using zinc sulfide (ZnS) as a buffer layer. By including the effects of thickness, concentration of a buffer layer, intrinsic layer and the absorbing layer, we find that photoelectric transformation efficiency of ZnO/ZnS(n)/CIGS(i)/CIGS(p) solar cell is about 17.22%, which is qualified as a commercial solar cell. Moreover, we also find that the open-circuit voltage is~0.60 V, the short-circuit current is~36.99 mA/cm2 and the filled factor is~77.44%. Therefore, our results suggest that zinc sulfide may be the potential candidate of CdS as a buffer layer.


1.   Introduction
  • It is well known that the thin films of CuInSe$_{\mathrm{2}}$-based compounds are remarkable semiconductor materials with inherently passive grain boundaries, wide compositional tolerance. By doping an appropriate amount of gallium (Ga) into Indium (In) sites, the band gap of copper Indium gallium selenium (CIGS) can range from 1.0 to 2.5 eV[1]. Consequently, people can get the excellent match of the absorption band gap and the solar spectrum through adjusting the gallium (Ga)/indium (In) proportion[2, 3]. Moreover, to be used as solar cells, CIGS thin film has shown many significant characters, such as the highest light-electricity conversion efficiency of all thin film solar cells, ability to be manufactured on low cost glass or flexible substrates, long-term battery performance stability, and low light so it has received extensive attention from the international photovoltaic industry. Therefore, solar cells based on such materials could quite possibly be broadly used to solve the energy crisis.

    On the other hand, in the real solar-cell battery, the cell structure is constructed by substrate/back electrode/absorption layer/buffer layer/window/former electrode. At present, cadmium sulfide (CdS) is used as a buffer layer in most of the CIGS based solar cells[4]. However, a cadmium atom is toxic and unfriendly to the environment[5]. Besides, the band gap of the CdS is about 2.4 eV[6], which is not large enough to transmit the whole part of visible light and limit the further improvement of the light-electricity conversion efficiency. Thus seeking an appropriate semiconductor material to substitute the CdS and realizing the copper based solar cell without cadmium are extremely urgent in today's research. As we know, the ZnS is an important Ⅲ-Ⅵ semiconductor material with a direct band gap of about 3.7 eV at room temperature[7, 8]. They can effectively reduce the blu-ray absorption, increasing the range of spectral response, so as to improve the efficiency of solar cells[9]. Moreover, ZnS has been widely used in p-n junction diodes and solar cells due to its non-toxic character, wide band gap, the unique feature of the photoelectric effect and high refractive index. Meanwhile, ZnS as an n-type semiconductor, suits heterojunction formation with a p-type CIGS absorption layer, constituting the structure of the multilayer film solar cells[10].

    In 2004, Nakada $\textit{et al. }$made 100 nm thick ZnS as the buffer layer by means of three consecutive sedimentary layers, and reached an efficiency of 18.6% for the CIGS solar cell[11]. America National Renewable Energy Laboratory (NREL) adopts a deposit of 100 nm thick ZnS as the buffer layer, and found that the efficiency of solar cell can reach 18.5%[12]. They also predicted that precipitating 120 nm thick ZnS buffer layer can obtain higher efficiency. Platzer-Bjorkman $\textit{et al.}$ have developed a cadmium-free CIGS Solar cell, whose efficiency can reach 19.7% with an area of 0.5 cm$^2$[13]; however, the buffer layer is not explored in their research.

    Although ZnS is one of the possible candidates of CdS, the performance of the complex solar-cell battery is still not explored. In this paper, we use the zinc sulfide (ZnS) substituting for the CdS as the buffer layer in the CIGS thin film solar cells[14]. The effects of thickness and doping concentration of each layer on short-circuit current density ($J_{\mathrm{sc}})$, open-circuit voltage ($V_{\mathrm{oc}})$, filled factor (FF) and efficiency (Effi) have been systematically analyzed with the numerical simulations.

2.   Computational details
  • The results of CGIS based heterojunction solar cells discussed in the following parts are carried out by the wxAMPS software[15-18]. In wxAMPS, the technique of finite differences and the Newton-Raphson interaction method are used to solve the one-dimensional Poisson equation and the hole and electron continuity equations that are subject to appropriate boundary conditions[19]. These equations are written as:

    where $\phi$ is the electrostatic potential and $\varepsilon(x) $ the permittivity, $n(x)$ the free electron, $p(x)$ free hole, $n_{\rm t}(x)$ trapped electron, $p_{\rm t}(x)$ trapped hole, $N_{\rm D}^{+}{x}$ the ionized donor-like doping and $N_{\rm A}^{-}{x}$ the ionized acceptor-like doping concentrations, and $\xi $ the electric field. All these parameters are a function of the coordinate position $x$. $G$ is the generation rate, $D$ the diffusion coefficient, and $q$ the charge of electron. Recombination currents are calculated with the Shockley-Read-Hall (SRH) model for bulk defects and extension of the SRH model for interface defects. The SRH interface approach allows carriers from both conduction and valence bands to participate in the interface recombination process[20].

3.   Solar cell structure model and parameters
  • As shown in Fig. 1, the structure of simulated solar cell consists of ZnO/ZnS(n)/CIGS(i)/CIGS(p) multilayers, where the absorbing layer is p-type CIGS, the intrinsic layer is weak p-type CIGS, and the window layer is ZnO with high transmittance. The thicknesses of the ZnS buffer layer and CIGS absorbing layer vary from 1 to 30 nm. In addition, the thickness of the intrinsic layer is tuned from 10 to 50 nm. The doping concentrations for ZnS and CIGS vary from 1 $\times$ 10$^{\mathrm{13}}$ to 1 $\times$ 10$^{\mathrm{19\thinspace }}$cm$^{\mathrm{-3}}$. The main simulated parameters of the solar cells are all listed in Table 1. AM1.5 radiation is used as the illumination source with a power density of 0.1 W/cm$^{\mathrm{2}}$ in our computations[21-25].

4.   Results and discussions

    4.1.   The influences of doping concentration of the buffer layer ZnS and absorbing layer CIGS on the battery performance

  • The variations of $V_{\mathrm{oc}}$, $J_{\mathrm{sc}}$, FF and Effi as a function of the doping concentration of the buffer layer ZnS are shown in Figs. 2(a)-2(d). It indicates that the values of $V_{\mathrm{oc}}$, $J_{\mathrm{sc}}$, FF and Effi are almost unchanged when the doping concentration is smaller than 5 $\times$ 10$^{\mathrm{16}}$ cm$^{\mathrm{-3}}$. As the doping concentration of the buffer layer ZnS ranges from 5 $\times$ 10$^{\mathrm{16\thinspace }}$ to 5 $\times$10$^{\mathrm{18}}$ cm$^{\mathrm{-3}}$, the $J_{\mathrm{sc}}$, FF and Effi have been significantly increased while the $V_{\mathrm{oc}}$ has a slight decrease from about 0.616 to 0.602 V. The $V_{\mathrm{oc}}$, $J_{\mathrm{sc}}$, FF and Effi of the GIGS solar cell are proved to be stable in numerical values when the doping concentration is higher than 5 $\times$ 10$^{\mathrm{18}}$ cm$^{\mathrm{-3}}$. Although the built-in electric field voltage can be improved because of heavy doping concentration and so as to increase the minority carriers through the junction area, the heavy doping concentration brings a lot of defects, e.g. the increasing dark current, the decrease of the life of minority carriers and the diffusion length. Therefore, the $J_{\mathrm{sc}}$ will be kept unchanged with the increase of doping concentration as shown in Fig. 2(b).

    From the results shown in Fig. 3, we can see that the values of the $V_{\mathrm{oc}}$, $J_{\mathrm{sc}}$, FF and Effi almost keep their initial values as the doping concentration of the absorbing layer CIGS is increased to 1 $\times$ 10$^{\mathrm{13}}$ -5 $\times$ 10$^{\mathrm{18}}$ cm$^{\mathrm{-3}}$. However, when the doping concentration of the absorbing layer CIGS reaches the range of 5$\times$10$^{\mathrm{18}}$ -5 $\times$ 10$^{\mathrm{19}}$ cm$^{\mathrm{-3}}$, the $V_{\mathrm{oc}}$, $J_{\mathrm{sc}}$, FF and Effi are significantly increased. The four typical indexes of the CIGS solar cell are almost kept on a stable level when the doping concentration of the absorbing layer CIGS is higher than 5$\times$10$^{\mathrm{19}}$ cm$^{\mathrm{-3}}$. It indicates that the optimal doping concentration of the absorption layer is about 5$\times$10$^{\mathrm{19}}$ cm$^{\mathrm{-3}}$.

  • 4.2.   The influences of layer thickness of the intrinsic layer CIGS(i) and absorbing layer CIGS on the battery performance

  • Now, we discuss the influence of the intrinsic layer thickness CIGS(i) on the performance of the solar cell. The relationships between the solar cell performances ($V_{\mathrm{oc}}$, $J_{\mathrm{sc}}$, FF and Effi) and the thickness of the intrinsic layer are described in Fig. 4. As we can see, the short circuit current and the photoelectric conversion efficiency are increased rapidly at first and then the values reach a maximum at 0.42 $\mu $m. However, the $J_{\mathrm{sc}}$ and Effi are decreased slightly when the thickness of the CIGS (i) layer added up to 0.42 $\mu $m. On the other hand, the graphs for the variations of the $V_{\mathrm{oc}}$ and FF as functions of thickness present declining curves but the declines are relatively gentle as shown in Figs. 4(a) and 4(c). The values of the $V_{\mathrm{oc}}$ and FF have small reductions, respectively, from 0.615 to 0.595 V and 79% to 76%. The background impurity concentration limits the thickness of the intrinsic layer, and the thinner intrinsic layer can prevent the decline of the carrier collection rate caused by the background impurity concentration[26]. A too thick intrinsic layer will bring on a fall of the built-in electric field strength and then affect the carriers in the junction area. Consequently, it will reduce the short-circuit current, the fill factor and the photoelectric conversion efficiency. Taking into consideration the solar cells' performances, the thickness of the intrinsic layer should be controlled with 0.42 $\mu $m.

    As shown in Fig. 5, the curve of the $V_{\mathrm{oc}}$ is nearly flat with the increasing of the thickness for the absorbing layer (CIGS). Compared with the result of the $V_{\mathrm{oc}}$, it is quietly different in the case of the $J_{\mathrm{sc}}$ and Effi. Both the $J_{\mathrm{sc}}$ and Effi fall into a clear downward trend in the condition of the rising of the absorbing layer thickness. It indicates that the short circuit current drops rapidly when the thickness is smaller than 0.07 $\mu $m while the short circuit current declines relatively slowly in the value range larger than 0.07 $\mu $m. As we can see in Fig.5(c), the fill factor in the image has some fluctuations but it is overall kept unchanged with the adding of the absorbing layer thickness. The improvement of the conversion efficiency will be hampered due to the limit of the minority carriers' diffusion length. Consequently, the thickness of the absorbing layer CIGS should be at $\sim$10 nm for the sake of conversion efficiency. With the optimized parameters, we get an efficiency (Effi) of 17.22%, the open-circuit voltage ($V_{\mathrm{oc}})$ of 0.60 V, the short-circuit current ($J_{\rm sc}$) of 36.99 mA/cm$^{\mathrm{2}}$ and FF of 77.44% for the ZnO/ZnS(n)/CIGS(i)/CIGS(p) solar cell. In contrast to the parameters of the typical copper based solar cell CdS/CIGS ($V_{\mathrm{oc}}$: 0.689 V, $J_{\rm sc}$: 35.7 mA/cm$^{\mathrm{2}}$, FF: 78%, Effi: 19.2%)[27], the cadmium-free solar cell ZnO/ZnS(n)/CIGS(i)/CIGS(p) can not only protect the environment from cadmium pollution but also guarantee the cell performances.

5.   Conclusion
  • In summary, the effects of thickness and concentration of a buffer layer, intrinsic layer and the absorbing layer on the performances of the ZnO/ZnS(n)/CIGS(i)/CIGS(p) solar cell with a cadmium-free buffer layer are systematically studied by using the wxAMPS simulation software. Based on careful analysis, we concluded that the optimum thickness of the intrinsic layer is about 0.42 $\mu $m. For the case of the buffer layer, the best thickness and doping concentration are about 0.02 $\mu $m and 5$\times$10$^{\mathrm{18}}$ cm$^{\mathrm{-3}}$ respectively. The most appropriate thickness and doping concentration are $\sim$0.01 $\mu $m and 5 $\times$10$^{\mathrm{19}}$ cm$^{\mathrm{-3\thinspace }}$for the absorption layer. With the optimized parameters, an efficiency (Effi) of 17.22%, the open-circuit voltage ($V_{\mathrm{oc}})$ of 0.60 V, the short-circuit current ($J_{\mathrm{sc}})$ of 36.99 mA/cm$^{\mathrm{2}}$ and FF of 77.44% are obtained for the ZnO/ZnS(n)/CIGS(I)/CIGS(p) solar cell. It reveals that the ZnO/ZnS(n)/CIGS(I)/CIGS(p) solar cell may have a great potential application in the future.

Figure (5)  Table (1) Reference (27) Relative (20)

Journal of Semiconductors © 2017 All Rights Reserved