Numerical Analysis of CsSnGeI3 Perovskite Solar Cells Using SCAPS-1D

Recently, organic-inorganic perovskite-based solar cells have become promising devices due to their unique properties in the photovoltaic field. However, the factor of toxicity, stability, high production cost and complicated fabrication processes of these devices is a challenge to their progress in commercial production. Here a numerical modelling of Caesium Tin–Germanium Tri-Iodide (CsSnGeI3) as an efficient perovskite light absorber material is carried out. In this paper, different inorganic Hole Transport Materials (HTMs) such as Cu2O, CuI, CuSbS2, CuSCN and NiO have been analyzed with C60 as the Electron Transport Material (ETM). We intend to replace the conventional hole and electron transport materials such as TiO2 and Spiro-OMeTAD which have been known to be susceptible to light induced degradation. Moreover, the influence of the Electron Transport Layer (ETL) and the perovskite layer properties, bandgap, doping concentration and working temperature for various Hole Transport Layers (HTL) on the overall cell performance have been rigorously investigated. The design of the proposed PSC is performed utilizing SCAPS1D simulator and for optimum device an efficiency greater than 30% was obtained. The results indicate that CsSnGeI3 and C60 are viable candidates for use as an absorber layer and electron transport layer in high-efficiency perovskite solar cells, with none of the drawbacks that other PSCs have.


Introduction
Photovoltaic (PV) devices which convert solar energy into electrical energy are the alternatives to reduce carbon emissions while meeting the increased demand in energy consumption. With a power conversion efficiency (PCEs) exceeding 26% [1] and long-term durability, High-purity crystalline silicon is an ideal product for the market. However, as a result of the rigorous research in this field various other PV technologies, especially on thin films (<1 mm) based on simple deposition methods have emerged. This promises a reduction in production cost and production of high-quality semiconductors. Recently, perovskite solar cells (PSCs) have emerged as one such candidate [2][3][4]. In just a few years, the efficiency of perovskite solar cells has crossed the 22% mark of CdTe solar cells. [5] Perovskite, named after the Russian scientist L.A. Perovski [6], has a crystal structure ABX 3 (X = oxygen, halogen). With the larger cation A occupying a cubo-octahedral site shared with twelve X anions and the smaller B cation in an octahedral site sharing six X anions [7]. Basically perovskite solar cells are solid-state sensitized solar cells based on dyesensitized Gratzel solar cells [8]. Although the perovskites exhibit unique optoelectronic properties like high absorption coefficient [9] in the UV visible spectrum, tuneable bandgap, large charge carrier mobility, long charge carrier diffusion length [10], defect tolerance property and so on they have some major drawbacks. One of the major disadvantage is their thermal instability caused by the photobleaching effect, moisture and oxygen induced degradation etc. this reduces their compound shelf life. Also other problems like instability due to the presence of organic cations, anomalous hysteresis, presence of toxic chemicals like lead and problems with large area flexible PV systems are also yet to be pondered [11].
Not only have the organic-inorganic hybrid absorber materials, organic charge transport materials such as Spiro-OMeTAD [12], P3HT [13] become unstable in the presence of oxygen, light, and moisture [14]. Besides that, several organic charge transport materials react with perovskite material, show hostile behavior due to hygroscopic materials, low mobility [15] and have high preparation costs. For these disadvantages, researchers are looking forward to inorganic charge transport materials as they provide good transparency in ultraviolet, visible and infrared spectrum, wide bandgap, high charge carrier mobility [16], superior thermal [17] and chemical stability, and simple synthesizing process [16].
Scientists have been trying to work in this hostile behavior of perovskite solar cells on various aspects and to develop superior perovskite materials with superior thermal [18] and chemical stability, simple synthesizing process [19] and most importantly less toxic. Two approaches are usually followed in this regard. Firstly, mixing lead with other metals which have lower toxicity such as tin-lead alloyed perovskite (CH 3 NH 3 SnxPb1-x) [20] and secondly by completely replacing lead with analogous metals [20]. Several lowtoxicity cations have been proposed for replacing Pb(II) in halide perovskites [21], including Ag(I) [22], Bi(III) [23], Sb(III) [24], Ti(IV) [25], Ge(II) [26], and Sn(II) [27]. Among these candidates, Tin is considered more suitable material because both are in the same group in periodic table and halide perovskites based on Sn(II) have shown the highest PCE, and, thus, have attracted the most attention in the PSC field. Typical Sn-based halide perovskites that have been studied include CH 3 NH 3 SnI 3 (MASnI 3 ), HC(NH 2 ) 2 SnI 3 (FASnI 3 ) [28], and CsSnI 3 [29]. Among other candidates Min Chen et al [30] proposed a nontoxic perovskite absorber material CsGeSnI 3 . While PSCs based on MASnI 3 and FASnI 3 perovskites have been shown to deliver high PCE, up to 9% [31], these materials have intrinsically low stability [32]. This is primarily attributed to the presence of the organic cation, which is prone to facile volatilization.
In this work, for the first time we investigate the simulation of CsSnGeI 3 based perovskite solar cells [33,34] with different materials as their Electron Transport Layer (ETL) and Hole Transport Layer (HTL) layers using the SCAPS 1-D simulation software. We study the impact of the change in various parameters of the HTL and absorber layers in the overall performance of the device. Also, the effect on properties of the solar cell with the change in operating temperature is studied as well.

Device Simulation Methodology
Numerical simulation has a great importance in understanding the physical properties and design of solar cells. It can interpret the measurements on complex structures, design and optimisation of advanced cell structures. There are various simulation packages like SILVACO ATLAS, AMPS, COMSOL etc in the market [35].
Here we use the Solar Cell Capacitance Simulator (SCAPS) software version 3.2.00 for the simulation. It is developed by Prof. Marc Burgelman [36] at the Department of Electronics and Information Systems (ELIS) of the University of Gent, Belgium. In SCAPS various material parameters like thickness, band gap, permittivity, mobility etc and operating conditions like temperature, voltage, frequency, illumination can be set [37]. The uniqueness of SCAPS is that it can calculate almost all the parameters like spectral response, energy bands, ac characteristics, J-V curve and defect density [38][39][40] by just solving three basic semiconductor equations: continuity equations for both electron and holes and Poisson equation [41]. The basic sequence of operation of SCAPS software [42] is illustrated in Figure 1.

Device Architecture
The simulated PSC has five layers: (i) Gold (Au) as the metal contact (ii) The p-type Hole Transport Layer (HTL) CuSCN, CuSbS 2 , NiO, Cu 2 O and CuI (iii) Absorber layer as CsGeSnI 3 , (iv) The Electron Transport Layer (ETL) as C 60 (v) Transparent Conducting Oxide. So basically here we simulate five device configurations Au/CuSCN/CsGeSnI 3 /C 60 /TCO, Au/CuSbS 2 /CsGeSnI 3 /C 60 /TCO, Au/CuI/CsGeSnI 3 /C 60 /TCO, Au/Cu 2 O/CsGeSnI 3 /C60/TCO and Au/NiO/CsGeSnI 3 /C 60 /TCO. The schematic of the device and the energy level diagram of the proposed structures is shown in Figures 2 and 3 respectively. The absorber material should have high hole and electron mobility since it needs to absorb incident radiation leading to the creation of charge carriers, while the ETL and HTL layers extract and transfer electrons and holes respectively.

Device Simulation Parameters
The parameters for different layers in the simulation are chosen on the basis of theoretical considerations, experimental data and existing literature or in some cases, reasonable estimation [43][44][45][46][47][48], [37], Table 1 summarizes the simulation parameters for the configuration. In the given parameters Eg is the energy bandgap, ϵr is the relative permittivity, χ being the electron affinity, µn and µp are the electron and hole nobilities respectively. N A and N D are the densities of acceptor and donor materials whereas N C and N V are the effective densities of conduction band and valence band. The parameter values not included in the table are set identical for all layers. Neutral Gaussian distribution defect is adopted with characteristic energy being set to 0.1 eV [48].
The electron and hole capture cross section is set to 9 × 10 −15 cm 2 with the thermal velocity of both carriers fixed at 10 7 cm/s [49][50][51].

Effect of the Bandgap of the Absorber Layer
The effect of band gap on the various solar cell characteristics has been examined since the band gap of a perovskite absorber is a significant parameter [52,53]. In this study the band gap is adjusted from 1.2eV to 2.0eV. When the band gap is 1.2eV, the V oc and J sc have maximum values of 1.38V and 36mA/cm 2 respectively, as shown in Figure 4, and then both drop. Because the absorption of photons with energy greater than the band gap (which contributes to the photocurrent) diminishes as the band gap increases, the Jsc continues to drop. PCE follows the same trend as J SC , and it drops to just 8% at higher band gaps (as illustrated in Figure  3(c)). The FF remains a constant for lower values of band gap but ultimately falls off at higher values. From the observations we conclude that, as the band gap of perovskite increases the light harvesting property of the solar cell goes on worsening.

Effect of the Acceptor Density of the Absorber Layer
As the doping of the absorber layer increases, more carriers can contribute to reverse saturation current causing its rise. According to the diode circuit model, the V OC can be expressed as [54] = + 1 Here, n is representing the ideality factor, k denotes the Boltzmann constant, T is used to denote the temperature, J 0 and q is the reverse saturation current and charge respectively. It can illustrate that due to increase in temperature, the value of J 0 rises exponentially which leads to the reduction in open circuit voltage. From another point of view, when doping concentration increases, the electric field at the perovskite interface increases and the separation process will be enhanced. However, increasing recombination according to this increasing field will badly affect the performance. In this scenario most of the devices show similar decrement type of behaviour on various parameters except CuSbS 2 . As indicated in Figure 5, the open circuit voltage (V oc ), FillFactor and power efficiency reach a maximum value of 1.33V, 90.2% and 30.1% respectively at doping concentration of 10 16 cm -3 , and then decrease. While the J sc almost remains a constant at 25mA/cm 2 . The optimum value of doping concentration (N A ) was set to 10 15 cm -3 .

Effect of the Donor Density of the ETL Layer
Donor density or charge carrier density of ETL layers of each solar cell had been varied between 10 16 cm −3 and 10 20 cm −3 as shown in Figure 6 to find variation in Efficiency, V oc , J sc and FF of the solar cells. For all of the solar cells, when we vary the acceptor density between 10 15 cm −3 and 10 20 cm −3 can see that J sc decreased but V oc , FF, and Efficiency increased with the increase of acceptor density. For increasing donor density of the absorber layer, J sc decreased from 25.5mA/cm 2 . From the PN-junction model, we can find the equation Where, A is the quality factor of the diode, J 0 is the reverse saturation current density, n i is the intrinsic concentration, q is the electronic charge, L is the diffusion length, D is the diffusion coefficient, N a and N d are the donor and acceptor doping concentrations respectively and the subscripts p and n refer to holes and electrons respectively. When N d increases, the saturation current J 0 will decrease according to the above equation. And when J 0 decreases, V oc increases. Increased carrier concentration, on the other hand, promotes the recombination process by introducing recombination centres or traps into the layer, which reduces the ability of photogenerated carriers to be collected at the front, lowering Jsc. Again, as V oc increases, PCE and FF also increase for all of the solar cells Cu 2 O, CuI and NiO based solar cells all had almost the same maximum PCEs of 33%.

Effect of Temperature on the Device
Here we studied the change in the temperature of a solar cell and its impact on the overall performance of the cell. At first we kept the temperature at 300 K in the simulation. Then the temperature was changed from 300 to 400 degrees Celsius to take into account the influence of the working temperature on the PCE, Voc, Jsc and FF of all solar cells, for the ideal absorbent layer and the thickness of the buffer layer. Figure 7 shows that the current density remains relatively constant with increasing temperatures, although the Voc, PCE and FillFactor of the PSC decrease. This might be due to the thermal dissociation of the absorbent layer [58]. Generally, increased temperature affects parameters such as carrier mobility, electron and hole concentration and band gap of the materials that changes the resistance and thus, PCE, V oc and FF get declined. When the temperature was increased to 400 K, the efficiency of the device decreased to 24.87% for most of the devices except CuSbS 2 . V oc decreased from 1.32 V to 1.14 V. This may be due to the rise in reverse saturation current, which decreases the voltage. So, 300 K temperature was used as working temperature for optimal performance. At this temperature the maximum achievable efficiency of the model is 30.01%, FillFactor is 88.87%, J sc =25.58 mA/cm 2 , and V oc =1.32V. Figure 8 depicts the quantum efficiency (QE) curves of the various PSCs (a). It clearly shows that the studied PSCs have the same QE curve with minor variations. Almost the full visible range of the solar spectrum is covered by the QE curve. The QE is nearly constant in the region from 300 nm to 500 nm and then it decreases till 800 nm except for the device with NiO as the HTL. It should be pointed out here that the QE in SCAPS is the external QE. Figure 9 shows simulated J-V curves for the proposed PCE with optimised C 60 as well as the absorber layer using various inorganic materials as HTL. Most of the combinations C 60 /CuI, CuSCN, NiO, Cu 2 O cell has a high open circuit voltage of 1.32V and short circuit current density (J sc ) of 25.99 mA/cm 2 respectively except CuSbS 2 which has a slightly lower open circuit voltage of 1.16V.

Conclusion
The solid-state perovskite solar cell models were controllably structured in SCAPS simulator and were successfully verified by comparing with performance parameters found in different literatures. The results show C 60 as an alternate ETL having the potential to be used with CsGeSnI 3 absorber layer and can effectively replace organic ETLs which are expensive and easily degradable. We studied the various PV parameters with different HTM layers CuSCN, Cu 2 O, NiO, CuSbS 2 and CuI with the proposed ETL (C 60 ) and absorber (CsGeSnI 3 ). After simulation we found that (i) C 60 as ETL gives a considerable efficiency of 15.54% with CsGeSnI 3 (ii) the optimum thickness of C 60 (ETL) and Absorber layer was found to 100nm and 500 nm respectively. (iii) 600 nm-700 nm of absorber was found to be appropriate to get better PV parameters; and (iv) the defect density in absorber should be maintained up to order of 10 15 cm -3 (v) studied the effect of band gap variation and finally (vi) conclude that CsGeSnI 3 based PSC's are very sensitive to temperature. The increase in PSC temperature affects material conductivity which leads to the degradation of PV parameters. The device simulation being employed here not only substantiates our understanding obtained from these simulation results but also gives us a wider view in selecting the ETL materials. Our simulation will be useful for more deep understanding of the mechanism, material characterization, stability enhancing, and efficiency elevation of perovskite solar cells.