Lead-Free Perovskite Homojunction-Based HTM-Free Perovskite Solar Cells: Theoretical and Experimental Viewpoints

Simplifying the design of lead-free perovskite solar cells (PSCs) has drawn a lot of interest due to their low manufacturing cost and relative non-toxic nature. Focus has been placed mostly on reducing the toxic lead element and eliminating the requirement for expensive hole transport materials (HTMs). However, in terms of power conversion efficiency (PCE), the PSCs using all charge transport materials surpass the environmentally beneficial HTM-free PSCs. The low PCEs of the lead-free HTM-free PSCs could be linked to poorer hole transport and extraction as well as lower light harvesting. In this context, a lead-free perovskite homojunction-based HTM-free PSC was investigated, and the performance was then assessed using a Solar Cell Capacitance Simulator (SCAPS). A two-step method was employed to fabricate lead-free perovskite homojunction-based HTM-free PSCs in order to validate the simulation results. The simulation results show that high hole mobility and a narrow band gap of cesium tin iodide (CsSnI3) boosted the hole collection and absorption spectrum, respectively. Additionally, the homojunction’s built-in electric field, which was identified using SCAPS simulations, promoted the directed transport of the photo-induced charges, lowering carrier recombination losses. Homojunction-based HTM-free PSCs having a CsSnI3 layer with a thickness of 100 nm, defect density of 1015 cm−3, and interface defect density of 1018 cm−3 were found to be capable of delivering high PCEs under a working temperature of 300 K. When compared to formamidinium tin iodide (FASnI3)-based devices, the open-circuit voltage (Voc), short-circuit density (Jsc), fill factor (FF), and PCE of FASnI3/CsSnI3 homojunction-based HTM-free PSCs were all improved from 0.66 to 0.78 V, 26.07 to 27.65 mA cm−2, 76.37 to 79.74%, and 14.62 to 19.03%, respectively. In comparison to a FASnI3-based device (PCE = 8.94%), an experimentally fabricated device using homojunction of FASnI3/CsSnI3 performs better with Voc of 0.84 V, Jsc of 22.06 mA cm−2, FF of 63.50%, and PCE of 11.77%. Moreover, FASnI3/CsSnI3-based PSC is more stable over time than its FASnI3-based counterpart, preserving 89% of its initial PCE. These findings provide promising guidelines for developing highly efficient and environmentally friendly HTM-free PSCs based on perovskite homojunction.


Introduction
Perovskite solar cells (PSCs) have attracted a lot of attention in the photovoltaic community in recent years owing to their high power conversion efficiencies (PCEs) and low-cost can outperform lead-based ones [32]. Though great progress has been made in FASnI 3or MASnI 3 -based devices, the presence of organic ions in FASnI 3 poses some concerns about the durability of PSCs based on organic-inorganic hybrid Sn-based perovskites due to the volatile nature of organic constituents. Despite the poor morphology of their films, all-inorganic Sn-based perovskite may possibly be a better choice in this scenario. Since CsSnX 3 (here X represent halides: I, Br, or Cl) perovskites have tolerance factors that are closer to 1, the geometric arrangement of CsSnX 3 is more durable [33]. The CsSnX 3 perovskites have low energies for the phase transition from the orthorhombic to the tetragonal phase and, as a result, low phase transition temperatures [34]. Furthermore, the narrow band gap (1.3 eV), high absorption coefficient (10 4 cm −1 ), and high hole mobility (42 cm 2 V −1 s −1 ) of CsSnI 3 make it an excellent perovskite for high optical absorption and charge-carrier transport [35][36][37][38]. In comparison to the theoretical limit, the PCE of reported PSCs based on CsSnI 3 is only about 5.03% [39]. However, it is important to consider the oxidation of Sn 2+ to Sn 4+ because this will induce perovskite instability and decrease the efficiency of Sn-based PSCs [40]. In this context, metallic Sn was used by Lin et al. to reduce the Sn 2+ from oxidation [41]. Other ways have been described, such as adding chemicals to the perovskite precursor, graphene-Sn quantum dots [42], tin halide (SnF 2 /SnCl 2 ) [43][44][45], hypophosphorous [46], hydrazine [47], ethylenediammonium [48], and 1,4-bis(trimethylsilyl)-2 [49]. Therefore, such approaches might increase the Sn-based PSCs' stability, which would then increase their performance. However, the HTM-free PSCs seldom ever use the environmentally favorable homojunction-based perovskites. The perovskite and electron/hole transport materials form a heterojunction, but because of its excellent self-doping ability [50][51][52][53], the perovskite can also form a homojunction, which would lessen the presence of impurities that could act as carrier recombination centers [54]. The guided transport of the photo-induced electrons and holes may be aided by an internal electric field that arises in the perovskite homojunction, which would reduce additional charge-carrier recombination in the perovskite layer [55]. It is therefore a promising route to further improve PSC performances. If homojunction Sn-based perovskites might be theoretically and experimentally used to guide the fabrication of efficient HTM-free PSCs, it is currently unclear and scarcely reported.
As FASnI 3 and CsSnI 3 have similar perovskite crystal structures, fabricating CsSnI 3 on FASnI 3 may produce a homojunction layer that is high quality and has fewer defects. In light of this, conducting thorough theoretical and experimental examinations of the performance of FASnI 3 /CsSnI 3 homojunction-based HTM-free PSCs could be a promising strategy to offer guidelines for environmentally friendly devices. In this regard, using SCAPS software, we first investigated the parameters of p-type CsSnI 3 that influence the performance of lead-free homojunction-based HTM-free PSCs. The HTM-free PSCs in this study use a lead-free homojunction-based perovskite (FASnI 3 /CsSnI 3 ) as the photon harvesting layer. In the light absorption spectrum, the homojunction-based FASnI 3 /CsSnI 3 perovskite demonstrated better photon harvesting. Additionally, the homojunction's builtin electric field, which was identified using SCAPS simulations, promotes the directed movement of the photo-induced charges, lowering carrier recombination losses. With optimal conditions, the lead-free homojunction-based HTM-free PSCs display a PCE of 19.02%. Moreover, a homojunction-based HTM-free PSC was fabricated to verify the validity of the simulations. According to the experimental approach, the FASnI 3 /CsSnI 3 perovskite-based homojunction was fabricated by a two-step process [55]. As a result, the FASnI 3 /CsSnI 3 homojunction-based HTM-free device demonstrated better performance with a PCE of 11.77% compared to the FASnI 3 -based HTM-free PSC (PCE = 8.94%).

Simulation Procedure
For experimental scholars investigating solar cells, finding the ideal structure and materials have always been a major difficulty. This procedure can be expedited by the simulation tool. The unique capabilities of the SCAPS include, but are not restricted to, modeling up to seven layers and computing numerous parameters, such as spectral response, energy bands, J-V curve, and defect density, by only resolving three fundamental semiconductor equations. It is simple to operate and can be utilized in both light and dark conditions. Key processes in photovoltaic systems are defined by SCAPS modeling techniques, allowing for the intuitive and systematic classification of optimal operating conditions for each parameter. In this research, a lead-free homojunction-based perovskite with a device layout of FTO/TiO 2 /FASnI 3 /CsSnI 3 /Au was proposed in HTM-free PSCs, as illustrated in Figure 1a,b. The illumination source in the simulation was the AM l.5G solar radiation spectrum with an intensity of 1000 W·m −2 . Important material parameters were gathered from prior experimental reports and are summarized in Table 1 to validate the viability of the simulations. Two interface layers were introduced: one between the FASnI 3 and CsSnI 3 and the other one between the CsSnI 3 and gold (Au) electrode, accounting for the interfacial defect density (denoted by IDL1 and IDL2 in Table 1). The defects in the perovskite layer were configured to have a characteristic energy of 0.1 eV and to be in a neutral Gaussian distribution with energy 0.6 eV above the valence band. Interfacial defects were regarded as neutral single defects with distribution energy of 0.6 eV above the valence band. ials 2023, 13, x FOR PEER REVIEW simulation tool. The unique capabilities of the SCAPS include, but are not restric modeling up to seven layers and computing numerous parameters, such as spec sponse, energy bands, J-V curve, and defect density, by only resolving three funda semiconductor equations. It is simple to operate and can be utilized in both light an conditions. Key processes in photovoltaic systems are defined by SCAPS modelin niques, allowing for the intuitive and systematic classification of optimal operatin ditions for each parameter. In this research, a lead-free homojunction-based pero with a device layout of FTO/TiO2/FASnI3/CsSnI3/Au was proposed in HTM-free PS illustrated in Figure 1a,b. The illumination source in the simulation was the AM l.5G radiation spectrum with an intensity of 1000 W·m −2 . Important material parameter gathered from prior experimental reports and are summarized in Table 1 to valid viability of the simulations. Two interface layers were introduced: one between the and CsSnI3 and the other one between the CsSnI3 and gold (Au) electrode, account the interfacial defect density (denoted by IDL1 and IDL2 in Table 1). The defects perovskite layer were configured to have a characteristic energy of 0.1 eV and to neutral Gaussian distribution with energy 0.6 eV above the valence band. Interfac fects were regarded as neutral single defects with distribution energy of 0.6 eV abo valence band.     Using SCAPS-1D software developed by researchers at the University of Ghent, the photovoltaic performance of the devices under AM1.5G solar illumination was examined [68]. By resolving the following Equations (1)-(3), which are encoded in SCAPS-1D, it was possible to obtain the current-voltage (J-V) characteristic curves, external quantum efficiency (EQE), generation/recombination of charges, and electric field distribution.
where ε, Ψ, q, n, p, n t , p t , N d , N a , G, D, and E represent the permittivity, electrostatic potential, elementary charge, density of free electrons, density of free holes, density of trapped electrons, density of trapped holes, donor doping density, acceptor doping density, generation rate, diffusion coefficient, and electric field, respectively.

Fabrication of PSCs
The FTO-substrates were ultrasonically cleaned after being immersed for 20 min in acetone, deionized water, detergent solution, and isopropyl alcohol, respectively. The FTO-substrates were treated with UV-ozone for 15 min following nitrogen blow drying. An aqueous stock solution of 2 M TiCl 4 (stored in the freezer) was diluted to the desired proportion after cleaning the FTO-substrates. The as-cleaned substrates were then dipped in this solution and maintained in a closed vessel in a 70 • C oven for one hour. To create a compact layer of TiO 2 (denoted as c-TiO 2 ), the substrates were dried at 100 • C in the air for an hour after being rinsed with deionized water and ethanol. Spin-coating the TiO 2 nanoparticles paste in ethanol at 2000 rpm for 25 s resulted in the deposition of a mesoscopic TiO 2 layer (denoted as m-TiO 2 ). The substrate was subsequently annealed for 20 min at 500 • C. After annealing, a diluted TiCl 4 aqueous solution in deionized water was applied to the FTO/c-TiO 2 /m-TiO 2 substrate for 20 min at 70 • C. The substrate was then exposed to UV-ozone for 10 min. SnI 2 (0.143 g) and FAI (0.140 g) were combined with 1 mL of DMSO to form the precursor solution for the FASnI 3 . The FASnI 3 layer was then created by one-step spin-coating of the FASnI 3 precursor solution for 25 s at 4000 rpm. Then, 0.5 mL of diethyl ether was carefully injected during the spinning process as an anti-solvent to speed up the formation of perovskite film. The resultant perovskite layers were annealed for 10 s at 60 • C and then for 12 min at 100 • C. Then, the FTO/c-TiO 2 /m-TiO 2 /FASnI 3 coated substrates were used to fabricate a homojunction FASnI 3 /CsSnI 3based perovskite layer, as schematically illustrated in Figure 2. With a SnI 2 thickness of ∼65 nm (formed by vapor deposition) and a dipping duration of 50 s (equivalent to a SnI 2 /CsI ratio of 1.6), the top CsSnI 3 perovskite layer was produced. The as-prepared FTO/TiO 2 /FASnI 3 /CsSnI 3 substrate was then annealed at 70 • C for 10 min. Note that we followed the two-step method from a previous report [55]. Lastly, to fabricate 80 nm electrodes onto the perovskite films and complete the FTO/c-TiO 2 /m-TiO 2 /FASnI 3 /Au and FTO/c-TiO 2 /m-TiO 2 /FASnI 3 /CsSnI 3 /Au devices, gold was thermally evaporated.

Characterization of Thin-Films and PSCs
The surface and cross-sectional morphologies were examined using a s tron microscope (Hitachi S-4800). The absorption spectrum was determined Vis spectrophotometer (UV-2600). Using Edinburg PLS 980, the steady-state the obtained samples were examined. The TRPL decay of the perovskite layer using a transient state spectrophotometer (Edinburgh Institute F900) and a 4 For the J-V characteristic curves, the devices were measured using a s (Keithley 2400) using forward (−0.1 to 1.2 V) or reverse (1.2 to −0.1 V) scans simulator (XES-301S+EL-100) under AM 1.5G illumination with a power inte W cm −2 . The step voltage was set to 12 mV, and the delay duration was 10 m system (Enli Tech.) was used to determine the EQEs of the as-prepared devi

Results and Discussion
We first explore significant parameters (such as thickness, doping densi face defect density at IDL1 and IDL2, and defect density of the CsSnI3) that a formance of the homojunction-based HTM-free PSCs in order to acquire opt and provide guidelines for the experimental approaches. These parameters

Characterization of Thin-Films and PSCs
The surface and cross-sectional morphologies were examined using a scanning electron microscope (Hitachi S-4800). The absorption spectrum was determined using a UV-Vis spectrophotometer (UV-2600). Using Edinburg PLS 980, the steady-state PL spectra of the obtained samples were examined. The TRPL decay of the perovskite layers was shown using a transient state spectrophotometer (Edinburgh Institute F900) and a 485 nm laser. For the J-V characteristic curves, the devices were measured using a source meter (Keithley 2400) using forward (−0.1 to 1.2 V) or reverse (1.2 to −0.1 V) scans from a solar simulator (XES-301S+EL-100) under AM 1.5G illumination with a power intensity of 1000 W cm −2 . The step voltage was set to 12 mV, and the delay duration was 10 ms. The QE-R system (Enli Tech.) was used to determine the EQEs of the as-prepared devices.

Results and Discussion
We first explore significant parameters (such as thickness, doping density, and interface defect density at IDL1 and IDL2, and defect density of the CsSnI 3 ) that affect the performance of the homojunction-based HTM-free PSCs in order to acquire optimum values and provide guidelines for the experimental approaches. These parameters were left the same for the remaining layers of the as-simulated HTM-free PSCs. The performance variances in the cells as a function of the aforementioned parameters are elaborated in the following discussion.
Given that the thickness of the CsSnI 3 has a significant role in the incident light absorption and charge-carrier generation, the PSC's performance will be affected by varying this thickness. The perovskite layer in a PSC must be thick enough to maximize light absorption yet thick enough to boost the collection of photo-generated charge-carriers. The collection efficiency is determined by the competition between recombination and charge transfer to the corresponding contacts. Therefore, choosing the right CsSnI 3 thickness will be crucial for developing efficient homojunction-based HTM-free PSCs. Figure 3 illustrates how the photovoltaic performance and energy levels of the as-simulated HTM-free PSCs vary with CsSnI 3 thickness. As noted in Figure 3a, the J sc was increased significantly from 27.08 to 30.09 mA cm −2 , which could be attributed to a higher generation of chargecarriers due to an improved capability of the homojunction-based perovskite layer for light harvesting. According to the simulated quantum efficiency, as shown in Figure 3b, the absorption spectrum broadens with increasing CsSnI 3 thickness. It should be noted that the J sc grows slowly and tends to saturate when the CsSnI 3 thickness exceeds 300 nm because too thick perovskite could result in a significant charge-carrier recombination [64]. Figure 3c shows how the perovskite layer tends to saturate the PCE at larger thicknesses. This is because thick layers make it hard for charge-carriers to be collected in a timely manner, and they recombine before they reach the relevant electrodes. Even though the simulation results showed that PSCs with thicker CsSnI 3 layers (more than 100 nm and less than 400 nm) performed better, we still believed that it would be difficult to fabricate CsSnI 3 layers on FASnI 3 that were thicker than 100 nm because it is known that the precursor solvent used to make CsSnI 3 will dissolve the FASnI 3 bottom layer. Figure 3d shows the energy levels of the HTM-free PSCs as a function of CsSnI 3 thickness, demonstrating that the built-in electric potential (V bi ) is weakly influenced by the thickness of CsSnI 3 . This is because the doping density, as shown by Equation (4), is the primary factor controlling the V bi [69,70].
where k, T, q, N A , N D , and n i represent Boltzmann constant, temperature, elementary charge, acceptor doping density, donor doping density, and intrinsic density, respectively. These findings reveal that when the light-harvesting perovskite layer is very thin, poor photon absorption results in low photocurrents; however, if the perovskite layer is very thick, recombination issues develop due to poor charge-carrier extraction efficiency. The simulation findings showed that a CsSnI 3 layer with an optimal thickness of 100 nm can be used to obtain efficient homojunction-based HTM-free PSCs. Another crucial element that influences the PSC's photovoltaic metrics is the doping density of the perovskite layer. We therefore look at the critical role that the CsSnI 3 doping density plays in the homojunction-based HTM-free PSCs. Figure S1 shows the alterations in the J-V characteristic curves, PCEs, and energy levels with changes in the acceptor doping density (N A ). Figure S1a,b shows that no discernible changes in the photovoltaic parameters were seen at N A from 10 13 to 10 15 cm −3 , demonstrating that the relatively modest doping density does not alter the inherent property of CsSnI 3 . The better photovoltaic performance of the as-simulated HTM-free PSCs can be due to the larger V bi (as shown in Figure S1c,d and Equation (4)) brought on by the larger N A , which is increased from 10 15 to 10 18 cm −3 . As shown in Figure S1b, the PCE improves when the N A increases from 10 15 to 10 18 cm −3 and it begins to decline as the N A of the CsSnI 3 exceeds 10 18 cm −3 . High doping density can help with charge-carrier separation, but too much doping inhibits charge-carrier mobility and causes substantial charge-carrier recombination, which lowers PCE. The simulation results reveal that the optimal value of N A is close to 10 18 cm −3 .
where k, T, q, NA, ND, and ni represent Boltzmann constant, temperature, elementary charge, acceptor doping density, donor doping density, and intrinsic density, respectively. These findings reveal that when the light-harvesting perovskite layer is very thin, poor photon absorption results in low photocurrents; however, if the perovskite layer is very thick, recombination issues develop due to poor charge-carrier extraction efficiency. The simulation findings showed that a CsSnI3 layer with an optimal thickness of 100 nm can be used to obtain efficient homojunction-based HTM-free PSCs. Another crucial element that influences the PSC's photovoltaic metrics is the doping density of the perovskite layer. We therefore look at the critical role that the CsSnI3 doping density plays in the homojunction-based HTM-free PSCs. Figure S1 shows the alterations in the J-V characteristic curves, PCEs, and energy levels with changes in the acceptor doping density (NA). Figure S1a,b shows that no discernible changes in the photovoltaic parameters were seen at NA from 10 13 to 10 15 cm −3 , demonstrating that the relatively modest The perovskite layer is the main part for the generation of charge-carriers. As a result, the PSC's performance is greatly influenced by the properties of the perovskite layer. Defect density is a crucial indicator of how well the perovskite layer performs. Therefore, fabricating high-quality perovskite active layers with low defect densities and less nonradiative recombination is essential for producing highly efficient PSCs. The nature and density of defect states in perovskites are very sensitive to the film deposition conditions, according to prior calculations and experimental data [71]. Due to the lack of stoichiometric compositions at the surfaces of grains and the possibility that perovskite constituents could disintegrate during the thermal annealing process and leave defects, it is believed that the majority of the defects in perovskite films are found at the grain boundaries or the interface [54,72]. In this context, we have investigated the impact of defect densities on the performance of the homojunction-based HTM-free PSCs, specifically the effect of bulk defects in the perovskite layer and defects at the interfaces. Figure 4a displays the J-V characteristic curves as function of defect density in the CsSnI 3 . It demonstrates that the PCE begins to decline when the perovskite layer has a high defect density. Higher defect densities in low-quality perovskite layers raise carrier recombination rates because defects trap charge-carriers, as seen in Figure 4b. It is currently challenging to significantly reduce the defect density in the CsSnI 3 layer due to the fact that the CsSnI 3 perovskite is unstable in the air and requires sophisticated fabrication techniques [36,45,73]. According to the simulation results, high device performance can be obtained with defect density from 10 15 to 10 16 cm −3 in the CsSnI 3 layer. densities in low-quality perovskite layers raise carrier recombination rates because d trap charge-carriers, as seen in Figure 4b. It is currently challenging to significantly r the defect density in the CsSnI3 layer due to the fact that the CsSnI3 perovskite is un in the air and requires sophisticated fabrication techniques [36,45,73]. According simulation results, high device performance can be obtained with defect density fro to 10 16 cm −3 in the CsSnI3 layer. In addition, the performance of PSCs is significantly influenced by the characte of the interface layers. It is widely speculated that low-quality interfaces result in surface/interface defect densities, which in turn increase the recombination of charg riers [74]. We assume that the defect densities of IDL1 and IDL2 range between 10 10 22 cm −3 in order to discuss the impact of the defect density of the interface layers PSC performance. The J-V characteristic curves and PCEs of the as-simulated hom tion-based HTM-free PSCs as a function of defect densities in IDL1 and IDL2 are s in Figure S2a-d. These graphs demonstrate that lowering defect density in the int In addition, the performance of PSCs is significantly influenced by the characteristics of the interface layers. It is widely speculated that low-quality interfaces result in larger surface/interface defect densities, which in turn increase the recombination of chargecarriers [74]. We assume that the defect densities of IDL1 and IDL2 range between 10 18 and 10 22 cm −3 in order to discuss the impact of the defect density of the interface layers on the PSC performance. The J-V characteristic curves and PCEs of the as-simulated homojunction-based HTM-free PSCs as a function of defect densities in IDL1 and IDL2 are shown in Figure S2a-d. These graphs demonstrate that lowering defect density in the interface layer will boost the PCE. Evidently, the effect of the defect density on the J-V characteristics becomes noticeable when the defect density is more than 10 18 cm −3 , which deteriorates the performance of the homojunction-based HTM-free PSCs. It is clear that IDL1 has a higher impact on PCE than IDL2, which can be explained by the possibility that defects may occur during the deposition of the CsSnI 3 layer on FASnI 3 thin-film. Based on simulation results, the defect densities of IDL1 and IDL2 should be less than 10 19 cm −3 , in order to attain the desired efficiency of homojunction-based HTM-free PSCs.
The homojunction-based FASnI 3 /CsSnI 3 is utilized in this study as an absorber layer in the lead-free HTM-free PSCs. It is revealed that the thickness, doping density, and defect density of the CsSnI 3 layer should be around 100 nm, 10 17 cm −3 , and 10 16 cm −3 , respectively, for the HTM-free PSCs to perform at their best. The optimal performance of the assimulated HTM-free PSCs is shown in Figure 5 and the photovoltaic parameters are listed in Table 2. The improvement of the overall performance of the homojunction-based HTM-free PSC (Figure 5a) is attributed to the increased photo-generated charge-carrier generation rate ( Figure 5b) and extended absorption of the homojunction-based FASnI 3 /CsSnI 3 (Figure 5c). The electric field within the PSC might be impacted by unintentional defects in the bulk or at device interfaces. As shown in Figure 5d, the electric field generated by the FASnI 3 /CsSnI 3 homojunction with a defect density of 10 16 cm −3 and doping density of 10 17 cm −3 enables directed charge-carrier transfer. Free holes will rise as electron concentrations decrease in the p-type perovskite region, exhibiting carrier-directed transport [55]. It should be noted that when the doping density is equal to or lower than the defect density, the perovskite layers become semi-insulating and fail to form the desired p-n homojunction [75]. Based on the simulation results, the PCE of FASnI 3 /CsSnI 3 homojunction-based HTM-free PSCs significantly increased from 14.62 to 19.03% when compared to the devices based on a bare FASnI 3 perovskite layer. 10 17 cm −3 enables directed charge-carrier transfer. Free holes will rise as electron concentrations decrease in the p-type perovskite region, exhibiting carrier-directed transport [55]. It should be noted that when the doping density is equal to or lower than the defect density, the perovskite layers become semi-insulating and fail to form the desired p-n homojunction [75]. Based on the simulation results, the PCE of FASnI3/CsSnI3 homojunction-based HTM-free PSCs significantly increased from 14.62 to 19.03% when compared to the devices based on a bare FASnI3 perovskite layer.   Meanwhile, the performance of the as-simulated HTM-free PSCs is tested as a function of operating temperature. For PSC simulation, 300 K is typically the testing temperature used. However, under experimental conditions, the operational temperature is typically higher than the simulation. Here, under the conditions stated in Table 1, the working temperature for the efficient homojunction-based HTM-free PSC was maintained at 300 K. It is important to note that high temperatures may cause devices to deform and undergo higher stress [76], which will increase interfacial defects and poor interconnectivity between the layers, lowering the photovoltaic parameters, as depicted in Figure S3. Despite the modeling findings demonstrating reduction in the photovoltaic parameters of the assimulated HTM-free PSCs, the stability of the Sn-based PSCs can be increased, though, by modifying the active layer morphology [77,78], using a controlled environment for film formation [36,47], and adding tin halides (SnF 2 or SnCl 2 ) to the perovskite precursor solution [79]. According to some studies, FASnI 3 and CsSnI 3 exhibit significantly better thermal and photochemical stability with regard to both light and high temperatures [45,80,81].
The performance of the homojunction-based PSCs was assessed, and then FASnI 3 and FASnI 3 /CsSnI 3 perovskite thin-films were experimentally fabricated. To make the FASnI 3 thin-film, a one-step spin-coating was taken. Contrarily, a two-step technique was used to fabricate the CsSnI 3 layer on top of a FASnI 3 layer (made through a one-step spin-coating process), as shown in Figure 2. Both perovskite (FASnI 3 and FASnI 3 /CsSnI 3 ) layers are compact and pinhole-free, as shown by surface morphologies and cross-sectional SEM images (Figure 6a-d), which will aid in charge-carrier transport and prevent shunting paths. Figure 6e shows the optical absorption of the FASnI 3 and FASnI 3 /CsSnI 3 thinfilms. According to the results, the FASnI 3 /CsSnI 3 sample exhibits improved optical absorption, which is advantageous for the generation of photo-induced charge-carriers in the homojunction-based HTM-free PSCs. The charge-carrier dynamics within as-prepared samples were investigated by steady-state photoluminescence (PL) and time-resolved photoluminescence (TRPL) spectroscopies as depicted in Figure 6. Compared to FASnI 3 , the PL of the FASnI 3 /CsSnI 3 sample decreases, as shown in Figure 6f. The higher quenched PL spectra shows that in the final device, the FASnI 3 /CsSnI 3 film has a substantially higher charge extraction and transfer efficiency. The FASnI 3 /CsSnI 3 sample has a shorter decay lifespan than the FASnI 3 sample, according to the TRPL spectra (Figure 6g), indicating that the charge-carriers were extracted quickly at the FASnI 3 /CsSnI 3 interface. In order to maximize the efficiency of charge-carrier transport within homojunction-based HTM-free PSCs, the FASnI 3 /CsSnI 3 layer could reduce charge losses. We fabricated two different types of PSCs (FTO/c-TiO2/m-TiO2/FASnI3/Au and FTO/c-TiO2/m-TiO2/FASnI3/CsSnI3/Au) while taking into consideration the aforementioned simulation results and optoelectronic characteristics of the perovskite thin-films. Figure 7a,b shows the cross-sectional SEM images of the PSCs prepared using FASnI3 and FASnI3/CsSnI3 layers. The surface and cross-sectional SEM results show that the functional layers in both PSCs were compact and free of pinholes. The photovoltaic parameters are presented in Table 3. Figure 7c Figure 7d. The simulation results showed higher PCEs than those for experimentally prepared HTM-free PSCs, which we believed might be due to the defects in the perovskite layer. In this scenario, it is possible to further improve performance of the homojunction-based HTM-free PSCs by incorporating vari- We fabricated two different types of PSCs (FTO/c-TiO 2 /m-TiO 2 /FASnI 3 /Au and FTO/c-TiO 2 /m-TiO 2 /FASnI 3 /CsSnI 3 /Au) while taking into consideration the aforementioned simulation results and optoelectronic characteristics of the perovskite thin-films. Figure 7a,b shows the cross-sectional SEM images of the PSCs prepared using FASnI 3 and FASnI 3 /CsSnI 3 layers. The surface and cross-sectional SEM results show that the functional layers in both PSCs were compact and free of pinholes. The photovoltaic parameters are presented in Table 3. Figure 7c displays the J-V characteristic curves for the FASnI 3 and FASnI 3 /CsSnI 3 devices under AM1.5G illumination. The average PCE of the FASnI 3 /CsSnI 3 HTM-free PSC was 11.77%, with a V oc of 0.84 V, a J sc of 22.06 mA cm −2 , and an FF of 63.50%. In comparison, the PCE of the FASnI 3 device was significantly lower with PCE of 8.94%, V oc of 0.79 V, J sc of 21.20 mA cm −2 , and FF of 53.35%. The efficient extraction of charge-carriers and improved light harvesting are attributable for the improved performance of FASnI 3 /CsSnI 3 -based PSCs, as depicted in Figure 7d. The simulation results showed higher PCEs than those for experimentally prepared HTM-free PSCs, which we believed might be due to the defects in the perovskite layer. In this scenario, it is possible to further improve performance of the homojunction-based HTM-free PSCs by incorporating various additives into the perovskite precursor, such as N,N'-methylenebis(acrylamide) and pentafluorophen-oxyethylammonium iodide [82,83]. The diffusion of the Au metal, which may create serious deep centers, could be another possible cause of the inferior performance of the as-prepared HTM-free PSCs. In this context, we suggest interfacial engineering to promote favorable charge transport in perovskites [84]. Another potential method to increase the efficacy of the HTM-free PSCs is to introduce a plasmonic effect, which has been shown to promote the optoelectronic characteristics of the perovskites [85,86].  In addition, the statistical data of PSCs based on FASnI3 and FASnI3/CsSnI3 (33 devices for each) were obtained, as shown in Figure S4a. The statistical data show that HTMfree PSCs prepared with the FASnI3/CsSnI3 homojunction have higher PCEs and better reproducibility than the FASnI3-based HTM-free devices. The PCEs for FASnI3/CsSnI3 homojunction-based HTM-free PSCs fall within a narrow range, whereas FASnI3-based devices exhibit PCEs that are widely distributed. The average PCE for the FASnI3-based devices reached 5.78%. The average PCE for FASnI3/CsSnI3 homojunction-based HTMfree PSCs, in comparison, reaches up to 10.83%. The high PCE and reproducibility of the HTM-free PSCs are two important characteristics, but another is their stability. The steady-state Jsc and PCE of the FASnI3/CsSnI3-based HTM-free PSCs were initially tracked at a maximum power point at a bias voltage of 0.9 V at room temperature under AM 1.5G illumination, as shown in Figure S4b. After 140 s of continuous illumination, the FASnI3/CsSnI3-based HTM-free PSC exhibits a stabilized PCE of 11.58%, which is considerably near to the maximum PCE of 11.77% for the identical PSC. Further, unsealed PSCs made using FASnI3 and FASnI3/CsSnI3 perovskites had their long-term stability tested every 45 min for 500 min under AM 1.5G illumination, as depicted in Figure S4c. After each stability test, the measured devices were kept in a glass oven at 25 °C and 55% humidity. Figure S4c illustrates the superior stability of the FASnI3/CsSnI3-based HTM-free PSCs over FASnI3-based devices. After 500 min of testing, the efficiency of FASnI3/CsSnI3based HTM-free PSCs slightly decreases from 11.73 to 10.43% while maintaining 89% of  In addition, the statistical data of PSCs based on FASnI 3 and FASnI 3 /CsSnI 3 (33 devices for each) were obtained, as shown in Figure S4a. The statistical data show that HTMfree PSCs prepared with the FASnI 3 /CsSnI 3 homojunction have higher PCEs and better reproducibility than the FASnI 3 -based HTM-free devices. The PCEs for FASnI 3 /CsSnI 3 homojunction-based HTM-free PSCs fall within a narrow range, whereas FASnI 3 -based devices exhibit PCEs that are widely distributed. The average PCE for the FASnI 3 -based devices reached 5.78%. The average PCE for FASnI 3 /CsSnI 3 homojunction-based HTM-free PSCs, in comparison, reaches up to 10.83%. The high PCE and reproducibility of the HTMfree PSCs are two important characteristics, but another is their stability. The steady-state J sc and PCE of the FASnI 3 /CsSnI 3 -based HTM-free PSCs were initially tracked at a maximum power point at a bias voltage of 0.9 V at room temperature under AM 1.5G illumination, as shown in Figure S4b. After 140 s of continuous illumination, the FASnI 3 /CsSnI 3 -based HTM-free PSC exhibits a stabilized PCE of 11.58%, which is considerably near to the maximum PCE of 11.77% for the identical PSC. Further, unsealed PSCs made using FASnI 3 and FASnI 3 /CsSnI 3 perovskites had their long-term stability tested every 45 min for 500 min under AM 1.5G illumination, as depicted in Figure S4c. After each stability test, the measured devices were kept in a glass oven at 25 • C and 55% humidity. Figure S4c illustrates the superior stability of the FASnI 3 /CsSnI 3 -based HTM-free PSCs over FASnI 3based devices. After 500 min of testing, the efficiency of FASnI 3 /CsSnI 3 -based HTM-free PSCs slightly decreases from 11.73 to 10.43% while maintaining 89% of the original PCE. The PCE of FASnI 3 -based PSCs under the identical testing conditions declined sharply from 8.58 to 1.04%, demonstrating poor stability and quick degradation. The protection of the CsSnI 3 layer, which lessens the interaction of degradation factors such as moisture and oxygen present in the environment, can be attributed to the higher stability of FASnI 3 /CsSnI 3based devices.

Conclusions
In this research, numerical analysis was first used to determine the optimum conditions for fabricating highly efficient homojunction-based HTM-free PSCs. There were a number of factors that have all been carefully considered and the influences on the performance of the HTM-free PSCs were studied. The simulation results revealed the broad absorption spectrum by the inclusion of CsSnI 3 , and the high hole mobility facilitated the chargecarrier transportation. For effective photon harvesting in highly efficient HTM-free PSCs, a CsSnI 3 layer thickness of approximately 100 nm was required. The optimal doping density for the CsSnI 3 layer was between 10 15 and 10 18 cm −3 . The simulation results also showed that the PSC operates better when its defect density was suitable, but that it performs poorly and has a greater charge-carrier recombination rate when it has too many defects. The simulation method indicated that the FASnI 3 /CsSnI 3 homojunctionbased HTM-free PSCs displayed PCEs up to 19.03% under optimal conditions. Taking into account the aforesaid simulation results, an experimental FASnI 3 /CsSnI 3 homojunctionbased HTM-free PSC with a configuration of FTO/c-TiO 2 /m-TiO 2 /FASnI 3 /CsSnI 3 /Au was fabricated. Compared to the FASnI 3 /CsSnI 3 -based HTM-free PSC, which delivered a PCE of 11.77%, the PCE of the FASnI 3 device was much lower (8.94%). The FASnI 3 /CsSnI 3based HTM-free PSC additionally showed long-term stability while retaining 89% of its initial efficiency. Our study is crucial for providing new viewpoints on using leadfree homojunction perovskites and for further enhancing the performance of lead-free homojunction-based HTM-free PSCs.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/nano13060983/s1, Figure S1: The current-voltage (J-V) characteristic curves (a), power conversion efficiencies (PCEs) (b), energy levels (c) and zoomed-in energy levels (d) of the as-simulated homojunction-based HTM-free PSCs as a function of doping density of CsSnI 3 . Note that the FASnI 3 /CsSnI 3 region is where the zoomed-in energy levels were taken; Figure S2: The current-voltage (J-V) characteristic curves and power conversion efficiencies (PCEs) of the as-simulated homojunction-based HTM-free PSCs as a function of defect density in IDL1 (a, b) and IDL2 (c, d); Figure S3: Photovoltaic performance of the as-simulated homojunction-based PSCs as a function of temperature; Figure S4: Statistics of the HTM-free PSCs with FASnI 3 or FASnI 3 /CsSnI 3 (a). Efficiency and current density of FASnI 3 /CsSnI 3 homojunction-based HTM-free PSC at maximum power point tracking (b), long-term stability of the FASnI 3 -based devices and FASnI 3 /CsSnI 3 -based devices at 25 • C in air (c).  Data Availability Statement: All the data presented in the manuscript can be obtained from the corresponding authors by reasonable request.