A Comprehensive Study of CsSnI3-Based Perovskite Solar Cells with Different Hole Transporting Layers and Back Contacts

By an abrupt rise in the power conservation efficiency (PCE) of perovskite solar cells (PSCs) within a short span of time, the instability and toxicity of lead were raised as major hurdles in the path toward their commercialization. The usage of an inorganic lead-free CsSnI3-based halide perovskite offers the advantages of enhancing the stability and degradation resistance of devices, reducing the cost of devices, and minimizing the recombination of generated carriers. The simulated standard device using a 1D simulator like solar cell capacitance simulator (SCAPS) with Spiro-OMeTAD hole transporting layer (HTL) at perovskite thickness of 330 nm is in good agreement with the previous experimental result (12.96%). By changing the perovskite thickness and work operating temperature, the maximum efficiency of 18.15% is calculated for standard devices at a perovskite thickness of 800 nm. Then, the effects of replacement of Spiro-OMeTAD with other HTLs including Cu2O, CuI, CuSCN, CuSbS2, Cu2ZnSnSe4, CBTS, CuO, MoS2, MoOx, MoO3, PTAA, P3HT, and PEDOT:PSS on photovoltaic characteristics were calculated. The device with Cu2ZnSnSe4 hole transport in the same condition shows the highest efficiency of 21.63%. The back contact also changed by considering different metals such as Ag, Cu, Fe, C, Au, W, Ni, Pd, Pt, and Se. The outcomes provide valuable insights into the efficiency improvement of CsSnI3-based PSCs by Spiro-OMeTAD substitution with other HTLs, and back-contact modification upon the comprehensive analysis of 120 devices with different configurations.

In 2012, the first reported CsSnI 3 -based PSC measured the efficiency of 0.9% [22].In 2014, an improved PCE of 2.02% was reported experimentally [23].The facile oxidation of Sn 2+ to Sn 4+ , high self-doping effect, and Sn vacancy result in fast degradation of the CsSnI 3 absorber and instability of CsSnI 3 -based PSCs.Inspiringly, several effective approaches have been suggested to enhance the stability, and quality of CsSnI 3 thin films, through additive engineering, composition engineering, fabricating the perovskite quantum dots, developing a two-step sequential deposition method [24][25][26][27][28].An inverted CsSnI 3 -based PSC with efficiency >10%, was experimentally reported in previous literature [29].These reported PCEs are much lower than the best reported results for lead-based PSCs and the theoretical ideal Shockley-Queisser limit.
Recently, an increasing number of theoretical studies have been carried out to design more efficient inorganic lead-free PSC devices with different electron transporting layer (ETL) and hole transporting layer (HTL) in both normal and inverted structures [30][31][32][33][34][35].The P-type semiconductors usually regarded as HTLs for extracting the photogenerated holes from the perovskite absorber layer and blocking the carrier recombination.Conventionally, inorganic/organic HTLs have been categorized depending on their chemical structure and contents.Due to great benefits such as the easy and low cost preparation process, high chemical stability, and good hole mobility, inorganic HTLs attracted attention as potential candidates for usage in stable PSCs.Although, organic HTMs usually exhibited higher PCE than inorganic HTMs, their low conductivity and/or hole mobility induce instability of organic HTM-based PSC devices.Therefore, adding dopants is necessary for most of the effective organic HTMs to improve their conductivity and/or hole mobility.Herein, several typical inorganic HTLs such as CuI, CuSCN, Cu 2 O, CuO, MoS 2 and Cu 2 ZnSnSe 4 (CZTSe) studied in comparison to organic HTLs such as Spiro-MeOTAD, PEDOT:PSS, P 3 HT, and PTAA to provide some enlightenment for design a more efficient nontoxic lead-free CsSnI 3 -based PSC that can use for further promote PSCs performance, scalable manufacturing, and commercial application using low cost and easy preparation HTMs.Among them, Cu 2 ZnSnSe 4 (CZTSe) showed the highest efficiency.Cu 2 ZnSnSe 4 is a ptype semiconductor with high hole mobility (15.1 cm 2 v −1 s −1 ), low resistivity (0.33 Ω-cm), earth-abundant elemental constituents, and non-toxic properties, which was regarded as a promising inorganic HTL in solar cells [36,37].The simulated standard device in the present study confirmed the photovoltaic characteristic parameters of the experimentally reported results of the CsSnI 3 -based device with the highest PCE of 12.96%.By optimizing the perovskite absorber thickness, the resultant data proposed a greater thickness for improving the PCE of CsSnI 3 -based device in experiments.According to best of our search, the previous theoretical and experimental studies did not mention this issue.The HTL modification study in this paper is accomplished by considering an optimized thickness for a perovskite layer that was extracted from the agreement between experiment and simulation results.Herein, we focused on the optimization of a CsSnI 3 -based perovskite solar cell with Spiro-OMeTAD HTL as the standard cell.The final generation-recombination rate, current density-voltage and quantum efficiency of optimized simulated standard device with and without Spiro-OMeTAD HTL compared with each other.Then, the effect of Spiro-OMeTAD replacement with other HTLs is investigated.

Theoretical Methods and Device Structures
Theoretical simulation is a valuable tool for designing and characterizing devices under different conditions.Moreover, simulation reduces the cost and time associated with their experimental investigation by optimizing the necessary parameters and material properties before doing experiments.

Computational Simulation Details
In the present study, the solar cell capacitance simulator structures (SCAPS-1D) software version 3.3.08 is utilized to find the photovoltaic characteristics of short-circuit current density (J sc ), open-circuit voltage (V oc ), fill factor (FF), and power conversion efficiency (PCE) by resolving coupled one-dimensional Poisson's, and continuity semiconductor equations.SCAPS-1D is most popular software among photovoltaic researchers as its simu-lation results demonstrated a close match with the experimental results [38].A solar cell device in SCAPS is defined as number of semiconductor layers, in which the transportation of electrons and holes consider by solving the carrier continuity equations and Poisson's equations iteratively for both electrons and holes.The Poisson's equation (Equation ( 1)) relates the charges to the electrostatic potential by following equation: Here, Ψ represents the electrostatic potential, q is the electron charge, ε 0 represents the permittivity of free space, ε r is the relative permittivity of the semiconductor material, p(n) is the hole (electron) concentration, N A − (N D + ) is the density of the ionized acceptors (donors), n t (p t ) is the trapped electron (hole), and x is the position coordinate.Because of the simultaneous investigation of recombination, generation, drift, and diffusion, the continuity equations are used.The continuity equations for the change in the concentration of electrons (Equation ( 2)) and holes (Equation ( 3)) are given as [39]: where G n (G p ) is the electron (hole) generation rate, n p (p n ) is the electron (hole) concentration in the p-region (n-region), n p0 (p n0 ) is the equilibrium electron (hole) concentration in the p-region (n-region), τ n and τ p denote electron and hole lifetime, µ p (µ n ) is the hole (electron) mobility, E is the electric field, and D n (D p ) is the electron (hole) diffuse on coefficient.
The photovoltaic parameters such as PCE, V oc , J sc , and FF are related to each other by the following equations [38]: where P max is the maximum power achievable from the point on the current-voltage (I-V) curve of a solar module under illumination that the product of current and voltage is maximum.

Device Structure
In most of the perovskite device configurations, the perovskite absorber layer is sandwiched between an electron transport layer (ETL), and a hole transport layer (HTL).According to the semiconductor physics, ETL and HTL are responsible for preventing photogenerated charge carriers from recombination and boosting charge transportation by creating the carrier separation paths to deliver electrons and holes from perovskite absorber to the electrodes and go into the external circuit through the metal (Au) contacts.
The simulated CsSnI 3 -based perovskite solar cell in the present study has a normal (n-i-p) structure with TiO 2 ETL.The exact values of the necessary input parameters for simulation using SCAPS-1D, are difficult to obtain by simulation.The input parameters extracted from literatures are listed in Table 1 [34,[40][41][42].The schematic illustration of device configuration and corresponding energy levels of different layers in simulated device and the band alignment of different HTLs with a perovskite absorber layer are shown in Figure 1.
Table 1.Input parameters of different layers of FTO, TiO 2 , CsSnI 3 , HTLs for simulated devices in the present study by SCAPS-1D software.

Photovoltaic Characterization of Standard Device
The photocurrent density-voltage (J-V) and quantum efficiency measurements of the simulated standard PSC device with and without Spiro-OMeTAD recorded using SCAPS-1D software are presented in Figure 2. The fill-factor (FF), short circuit current (Jsc), open circuit voltage (Voc), and power conversion efficiency (PCE) were deduced from the J-V characteristics.A standard simulated AM 1.5 G irradiation system (1000 W.m −2 ) and a bias potential scan from 0.0 V to +1.0 V was applied with a scan speed of 100 mV.s −1 for characterization of PSCs.According to the previous literature, the main responsibility of HTL is effecting on Voc [41].Therefore, a disparity appears between series and recombination resistances (Rs and Rrec, respectively) in simulated standard devices with and without HTL.Effect of perovskite thickness modification on photovoltaic parameters of standard optimized device presented in Figure 2. The highest efficiency of 18.06% was obtained for a CsSnI3 perovskite layer with a thickness of 800 nm.

Photovoltaic Characterization of Standard Device
The photocurrent density-voltage (J-V) and quantum efficiency measurements of the simulated standard PSC device with and without Spiro-OMeTAD recorded using SCAPS-1D software are presented in Figure 2. The fill-factor (FF), short circuit current (J sc ), open circuit voltage (V oc ), and power conversion efficiency (PCE) were deduced from the J-V characteristics.A standard simulated AM 1.5 G irradiation system (1000 W.m −2 ) and a bias potential scan from 0.0 V to +1.0 V was applied with a scan speed of 100 mV.s −1 for characterization of PSCs.According to the previous literature, the main responsibility of HTL is effecting on V oc [41].Therefore, a disparity appears between series and recombination resistances (R s and R rec , respectively) in simulated standard devices with and without HTL.Effect of perovskite thickness modification on photovoltaic parameters of standard optimized device presented in Figure 2. The highest efficiency of 18.06% was obtained for a CsSnI 3 perovskite layer with a thickness of 800 nm.

Photovoltaic Characterization of Standard Device
The photocurrent density-voltage (J-V) and quantum efficiency measurements of the simulated standard PSC device with and without Spiro-OMeTAD recorded using SCAPS-1D software are presented in Figure 2. The fill-factor (FF), short circuit current (Jsc), open circuit voltage (Voc), and power conversion efficiency (PCE) were deduced from the J-V characteristics.A standard simulated AM 1.5 G irradiation system (1000 W.m −2 ) and a bias potential scan from 0.0 V to +1.0 V was applied with a scan speed of 100 mV.s −1 for characterization of PSCs.According to the previous literature, the main responsibility of HTL is effecting on Voc [41].Therefore, a disparity appears between series and recombination resistances (Rs and Rrec, respectively) in simulated standard devices with and without HTL.Effect of perovskite thickness modification on photovoltaic parameters of standard optimized device presented in Figure 2. The highest efficiency of 18.06% was obtained for a CsSnI3 perovskite layer with a thickness of 800 nm.By increasing the absorber layer thickness, more amount of generated carriers push to move from the traps and the charge separation ability of perovskite absorber increases.Therefore, the generated current density increases because of the higher rate of producing the photogenerated carriers [33,43,44].Although, the generation per volume unit decreases (Figure 3a).As shown in Figure 3b, higher increase in the absorber layer thickness results in creating the additional traveling paths for carriers to transfer through the absorber.Therefore, the recombination rate and series resistance decrease by creating a bigger barrier between HTL and ETL.Thus, a thicker perovskite layer produces a weaker built-in electric field, which leads to a reduction of V oc and FF by decreasing the recombination and cell series resistance [45].

, x FOR PEER REVIEW 6 of 12
By increasing the absorber layer thickness, more amount of generated carriers push to move from the traps and the charge separation ability of perovskite absorber increases.Therefore, the generated current density increases because of the higher rate of producing the photogenerated carriers [33,43,44].Although, the generation per volume unit decreases (Figure 3a).As shown in Figure 3b, higher increase in the absorber layer thickness results in creating the additional traveling paths for carriers to transfer through the absorber.Therefore, the recombination rate and series resistance decrease by creating a bigger barrier between HTL and ETL.Thus, a thicker perovskite layer produces a weaker built-in electric field, which leads to a reduction of Voc and FF by decreasing the recombination and cell series resistance [45].By increasing absorber layer thickness, the current density of electron and holes increase (Figure 3c).Although, the total current density in Figure 2b shows a saturated amount for devices contain a perovskite absorber thicker than 800 nm.The results plotted in Figure 3d-f demonstrate energy levels considered fixed and the thickness is the only parameter that is modified in this section.

Photovoltaic Characterization of Standard Device vs. Experimental Results and HTL Free Device
The simulated CsSnI3-based standard device with/without HTL compared in Figure 4a.The contact resistance between the metal electrode and perovskite layer causes increasing the series resistance.Therefore, the fill factor and current density of the HTL-free device decrease in comparison to the standard device with spiro-OMeTAD HTL.The HTL acts as a separator layer between perovskite absorber and metal back electrode to enhance the stability and efficiency of perovskite device (Figure 4a).The existence of HTL leads to increasing the amount of collected holes from photo-generated carriers by incident pho- By increasing absorber layer thickness, the current density of electron and holes increase (Figure 3c).Although, the total current density in Figure 2b shows a saturated amount for devices contain a perovskite absorber thicker than 800 nm.The results plotted in Figure 3d-f demonstrate energy levels considered fixed and the thickness is the only parameter that is modified in this section.

Photovoltaic Characterization of Standard Device vs. Experimental Results and HTL Free Device
The simulated CsSnI 3 -based standard device with/without HTL compared in Figure 4a.The contact resistance between the metal electrode and perovskite layer causes increasing the series resistance.Therefore, the fill factor and current density of the HTL-free device decrease in comparison to the standard device with spiro-OMeTAD HTL.The HTL acts as a separator layer between perovskite absorber and metal back electrode to enhance the stability and efficiency of perovskite device (Figure 4a).The existence of HTL leads to increasing the amount of collected holes from photo-generated carriers by incident photons at a given energy (QE), and photon absorption at longer wavelengths (Figure 4b).The theoretical characterization results of standard PSC device simulated by SCAPS software at perovskite thickness of 330 nm are in good agreement with the previous experimental report of 12.96% (Figure 4c) [46].The simulation results in Figure 4c and data listed in Table 2 demonstrate the considered thickness in experimentally fabricated devices is important issue for more efficient CsSnI3-based PSC devices.According to the current calculation results, increasing the perovskite thickness from 330 nm to 800 nm improves the efficiency of fabricated devices from 12.96% to 18.11% at room temperature.The thickness of fabricated HTL in devices that can assess HTLs between 300 nm to 800 nm is important for improving efficiency.The simulation results in Figure 4c and data listed in Table 2 demonstrate the considered thickness in experimentally fabricated devices is important issue for more efficient CsSnI 3 -based PSC devices.According to the current calculation results, increasing the perovskite thickness from 330 nm to 800 nm improves the efficiency of fabricated devices from 12.96% to 18.11% at room temperature.The thickness of fabricated HTL in devices that can assess HTLs between 300 nm to 800 nm is important for improving efficiency.The operating temperature for all of devices was considered fixed at 300 K in the present study.The effect of work operating temperature modification on the optimized standard device parameters is investigated.According to the data shown in Figure 5, the highest efficiency of 18.15% is calculated at 320 K for the standard device.By increasing the temperature, the efficiency will decrease rapidly.Because of the increase in the number of charge carriers reaching to the electrodes, a continuous increase in the saturation current is observable.Increasing the reverse saturation current, due to higher surface recombination, causes the V oc to decline with an increase in temperature, while FF increases up initially, then abruptly starts to decline.The main reason for the fill factor falling is due to the increased resistance device.

Photovoltaic Characterization of Simulated Device with HTL Modification
In the simulated CsSnI3-based perovskite devices with normal structure, the existence of a hole transport layer helps the photogenerated holes to transfer easily from the absorber layer to metal back contact and also prevent holes recombination with electrons by rapid reflecting them to the circuit through metal back contact.It is noticeable that the input parameters of Spiro-OMeTAD in SCAPS software extract from experiments data.In experiment, the Spiro-OMeTAD is doped with Li-TFSI and tBP dopants for increasing the favorite properties of HTL.Therefore, Spiro-OMeTAD is very expensive.Herein, some proposed p-type semiconductors as HTL in the literature include CuI, CuSbS2, CBTS,

Photovoltaic Characterization of Simulated Device with HTL Modification
In the simulated CsSnI 3 -based perovskite devices with normal structure, the existence of a hole transport layer helps the photogenerated holes to transfer easily from the absorber layer to metal back contact and also prevent holes recombination with electrons by rapid reflecting them to the circuit through metal back contact.It is noticeable that the input parameters of Spiro-OMeTAD in SCAPS software extract from experiments data.In experiment, the Spiro-OMeTAD is doped with Li-TFSI and tBP dopants for increasing the favorite properties of HTL.Therefore, Spiro-OMeTAD is very expensive.Herein, some proposed p-type semiconductors as HTL in the literature include CuI, CuSbS 2 , CBTS, MoS 2 , MnO 2 , P 3 HT, Cu 2 O, CuSCN, CZTSe, CuO, MoO x , PTTA, and PEDOT:PSS investigated to find an appropriate replacement for Spiro-OMeTAD.The resultant J-V curves and external quantum efficiency (QE) for simulated devices with different HTLs and configurations of FTO/TiO 2 /CsSnI 3 /HTL/Au plotted in Figure 6.
Among all of devices with different HTLs, simulated devices Cu 2 O, CuSCN, Cu 2 ZnSnSe 4 (CZTSe), CBTS, CuO, MoS 2 , and MoO x showed better efficiency than Spiro-OMeTAD and the device with Cu 2 ZnSnSe 4 (CZTSe) as an HTL layer showed the highest efficiency of 21.63% (Table 2) while the minimum PCE achieved for a device with PE-DOT:PSS HTL (Figure 6a).The favorable band alignment of Cu 2 O offers a higher current density while lower current density obtained from a device with MoS 2 HTL owing to unsatisfactory band alignment with perovskite absorber and lower hole transportation to metal electrodes.Therefore, the HTL's band alignment influences the flow of photogenerated holes in CsSnI 3 -based PSCs.The corresponding QE as a function of wavelength (λ) is studied in the range of 400 nm-1000 nm as shown in Figure 6b.In this study, the ETL, light absorption, and incident photons to the perovskite layer were fixed.The QE started to increase from 400 nm and dropped after 820 nm, corresponding to the band edge of each active material.The maximum QE was obtained for the device with a CZTSe HTL as expected from J-V characteristics and the minimum for the device with a MoS 2 HTL.

Photovoltaic Characterization of Simulated Device with HTL Modification
In the simulated CsSnI3-based perovskite devices with normal structure, th ence of a hole transport layer helps the photogenerated holes to transfer easily fr absorber layer to metal back contact and also prevent holes recombination with el by rapid reflecting them to the circuit through metal back contact.It is noticeable t input parameters of Spiro-OMeTAD in SCAPS software extract from experiments d experiment, the Spiro-OMeTAD is doped with Li-TFSI and tBP dopants for increas favorite properties of HTL.Therefore, Spiro-OMeTAD is very expensive.Herein proposed p-type semiconductors as HTL in the literature include CuI, CuSbS2, MoS2, MnO2,P3HT, Cu2O, CuSCN, CZTSe, CuO, MoOx, PTTA, and PEDOT:PSS gated to find an appropriate replacement for Spiro-OMeTAD.The resultant J-V and external quantum efficiency (QE) for simulated devices with different HTLs an figurations of FTO/TiO2/CsSnI3/HTL/Au plotted in Figure 6.

Conclusions
The present work reported a comparison between different CsSnI3-based PSC devices with modified HTLs and considering different metal back contacts in a normal structure.simulated devices were characterized by SCAPS software.The standard device with Spiro-OMeTAD HTL showed good overlap with experiment results in 330 nm thickness.The effect of perovskite thickness and working temperature on photovoltaic parameters of devices investigated.Spiro-OMeTAD replaced with ten different HTLs.The device with Cu2ZnSnSe4 (CZTSe) as an HTL layer showed the highest efficiency of 21.63%.The metal back contact also changed by modifying the work function of the back contact among 120 configurations.Thus, several promising and competitive configurations for high efficient CsSnI3-based PSCs are proposed.

Figure 1 .
Figure 1.Schematic presentation of (a) device configuration, and (b) corresponding energy band levels of different layers of simulated inverted CsSnI3-based PSCs.

Figure 2 .
Figure 2. The effect of perovskite thickness modification on photovoltaic parameters of (a) current density, (b) open circuit voltage, (c) fill factor, and (d) efficiency of simulated standard devices with Spiro-OMeTAD HTL.

Figure 1 .
Figure 1.Schematic presentation of (a) device configuration, and (b) corresponding energy band levels of different layers of simulated inverted CsSnI 3 -based PSCs.

Micromachines 2023 , 12 Figure 1 .
Figure 1.Schematic presentation of (a) device configuration, and (b) corresponding energy band levels of different layers of simulated inverted CsSnI3-based PSCs.

Figure 2 .
Figure 2. The effect of perovskite thickness modification on photovoltaic parameters of (a) current density, (b) open circuit voltage, (c) fill factor, and (d) efficiency of simulated standard devices with Spiro-OMeTAD HTL.

Figure 2 .
Figure 2. The effect of perovskite thickness modification on photovoltaic parameters of (a) current density, (b) open circuit voltage, (c) fill factor, and (d) efficiency of simulated standard devices with Spiro-OMeTAD HTL.

Figure 3 .
Figure 3.The effect of perovskite thickness modification on (a) generation rate, (b) recombination rate of photo-generated carriers per volume unit in different layers of simulated devices with different perovskite absorber thickness, (c) current density of electrons (Jn) and holes (Jp), and conduction band position (Ec), valence band position (Ev), Fermi energy levels for device with perovskite thickness of (d) 100 nm, (e) 330 nm, and (f) 800 nm.

Figure 3 .
Figure 3.The effect of perovskite thickness modification on (a) generation rate, (b) recombination rate of photo-generated carriers per volume unit in different layers of simulated devices with different perovskite absorber thickness, (c) current density of electrons (J n ) and holes (J p ), and conduction band position (E c ), valence band position (E v ), Fermi energy levels for device with perovskite thickness of (d) 100 nm, (e) 330 nm, and (f) 800 nm.

Figure 4 .
Figure 4. (a) The current density-voltage (J-V) curves, and (b) quantum efficiency curves of devices with/without Spiro-OMeTAD as the HTL.(c) Comparison between J-V curves of simulated device and experimental results.

Figure 4 .
Figure 4. (a) The current density-voltage (J-V) curves, and (b) quantum efficiency curves of devices with/without Spiro-OMeTAD as the HTL.(c) Comparison between J-V curves of simulated device and experimental results.

Figure 5 .
Figure 5.The effect of work temperature on the photovoltaic parameters (a) open circuit voltage, (b) short circuit current density, (c) fill factor, and (d) efficiency of simulated standard PSC.

Figure 5 .
Figure 5.The effect of work temperature on the photovoltaic parameters (a) open circuit voltage, (b) short circuit current density, (c) fill factor, and (d) efficiency of simulated standard PSC.

Figure 5 .
Figure 5.The effect of work temperature on the photovoltaic parameters (a) open circuit (b) short circuit current density, (c) fill factor, and (d) efficiency of simulated standard PSC.

Figure 6 .
Figure 6.(a) The current density-voltage (J-V), and (b) quantum efficiency curves of simulated devices with different HTLs.

3. 4 .
Photovoltaic Characterization of Simulated Device with HTL by Modification of Metal Back ContactAs discussed in the previous section, in the normal structures the photogenerated holes extracted through the metal back contact to the external circuit.Herein, ten different metals of Ag, Cu, Fe, C, Au, W, Ni, Pd, Pt, and Se are used as back contact in simulated devices to investigate the best combination of each HTL for having more efficient devices values among 120 different configurations.The resultant photovoltaic characteristic data are listed in Figure7.It is noticeable to experimentalists that the fabricated devices with Se, Pt, and Pd back contacts revealed higher efficiency than commonly used Au electrodes for most of the devices with different HTLs.The most characteristic parameter essentially affected by the metal back contact modification in simulated PSCs is the fill factor, due to modification in surface recombination and series resistance created by connection between metal back contact and HTLs.

Figure 7 .
Figure 7. Optimization of photovoltaic parameters of simulated PSCs with different metal back contacts for different HTLs.

Table 2 .
The photovoltaic parameters of a simulated device with different HTLs.

Table 2 .
The photovoltaic parameters of a simulated device with different HTLs.