1. Introduction
Traditional energy sources that include coal, fossil fuels, natural gas, etc., have detrimental effects on human civilization and the environment [1] . Research efforts are therefore focused on finding and developing alternative sources of energy to address long-term energy needs while minimizing negative impacts on the economy and environment. Solar energy, being a renewable resource with an inexhaustible supply, exhibits significant potential to meet daily energy needs. In addition, solar energy is inexpensive, and the establishment of more efficient solar cells in recent years has demonstrated enormous potential [2] . CZTS, a kesterite and stannite structured material, has been extensively utilized as one of the most widely adopted SC materials in endeavors to enhance SC efficiency. CZTS solar cells are attractive to researchers due to their low cost, non-toxicity, and bounty of core elements in nature. CZTS has a direct band gap energy of about 1.4 eV to 1.5 eV and an optical absorption coefficient of 104 cm−1. The cheapness of CZTS materials and their high melting point (990˚C) indicate their potential for practical manufacture using an affordable solution technique, hence ensuring long-term operational durability [3] [4] [5] [6] . Spin-coating and sputtering processes are commonly used for CZTS SC fabrication. Nevertheless, the CZTS material exhibits certain limitations, with its primary drawback being the disorder of Cu-Zn cations, leading to a substantial deficit in the VOC due to electron trapping. However, temperature treatment can overcome this limitation significantly.
Several novel investigations have already been undertaken to explore the efficacy of CZTS solar cells. Cadmium sulfide (CdS), zinc selenide (ZnSe), and indium sulfide (In2S3) are commonly employed as buffer layers in conjunction with CZTS as the absorber layer in thin film solar cells [7] [8] . However, the Shockley-Queisser limit states that the optimal conversion efficiency for CZTS solar cells is 32.2% [9] . A research investigation successfully attained an efficiency of 18.66% with a 2000 nm thick CZTS absorber layer and a 100 nm thick CdS buffer layer [10] . Additionally, a different approach obtained 18.68% efficiency by employing FTO as the window layer, In2S3 as the buffer layer, CZTS as the absorber layer and Mo as back contact [11] , while another study on CZTS SC recorded 19.23% efficiency with a ZnO window layer, a 50 nm thick In2S3 buffer layer, and a 1000 nm thick CZTS absorber layer [12] . A photovoltaic device composed of the Al-ZnO/CdS/CZTS/MoO3/Au structure achieved a PCE of 22.28% during an investigation into the utilization of MoO3 as the back-surface field. The enhanced efficiency observed in this case can be attributed to the utilization of MoO3 BSF, which effectively facilitates carrier transportation, promotes their accumulation at the electrodes, and minimizes carrier recombination at the interface [9] . Hence, it is imperative to consider the implementation of a BSF layer with the intention of enhancing the performance of conventional CZTS SCs. These developments signify promising progress in the utilization of CZTS solar cells for improved energy conversion and renewable energy applications.
In this paper, a comprehensive investigation of proposed CZTS-based SCs performance parameters with and without CuSbS2 BSF was carried out employing SCAPS 1D simulation software. The investigation was conducted to inspect the impact of various factors, including the thickness, doping density, and defect density of the absorber and buffer layer, as well as the thickness and doping density of the back surface field (BSF), and the operating temperature, on the output parameters of a solar cell. The objective of this investigation was to optimize the device structure to enhance photo conversion efficiency.
2. Device Configuration and Material Parameters
The numerical analysis of the SC was conducted utilizing SCAPS-1D software specifically designed for SC analysis. This software has been programmed by the esteemed Department of Electronics and Information Systems at the University of Gent, Belgium [13] . The structure may incorporate a maximum of seven distinct layers, six interface layers and two electrodes. SCAPS 1D is a software tool that effectively models and presents an extensive array of parameters related to renewable energy. These parameters encompass crucial aspects such as PCE, FF, Voc, Jsc, QE, and J-V characteristics. This software utilizes the Poisson equation, continuity equation, and current density equation to accurately simulate and analyze the various parameters [14] . The schematic representations of CZTS- based SC are visually depicted in Figure 1, showcasing the intricate structures. Additionally, the energy band diagram of these structures is thoughtfully illustrated in Figure 2, providing a comprehensive understanding of the energy dynamics.
The window layer incorporated in the aforementioned cell composition comprises the intrinsic zinc oxide (i-ZnO) and aluminium doped zinc oxide (Al: ZnO) layer, which is chosen for its cost-effectiveness and exceptional optical transparency [15] .
The chosen buffer and absorber layers for this particular application were CdS and CZTS, respectively. Copper antimony sulfide (CuSbS2) has been carefully selected as the back surface field (BSF) layer due to its exceptional band gap and remarkable optical absorption coefficient [16] . SC performance at 300 K is simulated under 100 mW/cm2 of incident light, 1 MHz of radiation frequency, and AM 1.5G of the solar spectrum. To keep things straightforward, the impact of resistance is not taken into account. Table 1 shows the important parameters for the various layers, whereas Table 2 shows the interface and bulk defect parameters.
![]()
(a) ![]()
(b)
Figure 1. CZTS solar cell configuration (a) with BSF layer (b) without BSF layer.
![]()
(a) ![]()
(b)
Figure 2. Energy band diagram (a) with BSF layer (b) without BSF layer.
![]()
Table 1. Input parameters that were used in simulation.
![]()
Table 2. Bulk and interface defect used in simulation [21] .
3. Results and Discussion
3.1. Impact of Absorber Layer Thickness and Doping Density Variation on PV Cell
Figure 3 exemplifies the simultaneous effect of varying the thickness of the absorber layer and acceptor doping (NA) on PV parameters. The thickness and NA were varied from 0.5 μm to 2.5 μm and 1013 cm−3 to 1019 cm−3, respectively. More photons are absorbed by a thick absorber layer, which also produces many electron-hole pairs [22] . Consequently, Jsc rises from 28.78 mA/cm2 at 0.5 μm CuSbS2 absorber layer thickness to 31.72 mA/cm2 at 1 μm, as shown in Figure 3(b). The effect of increasing the thickness of the absorber layer from 0.5 μm to 2.5 μm is seen in Figure 3(a), where Voc marginally decreases from 1.039 V to 0.9750 V while recombination rises with the thickness of the absorber layer. Voc and Jsc, on the contrary, displayed a reversed characteristic for rising doping density. According to Figure 3(a) and Figure 3(b), the greatest Voc and Jsc values were 1.233 V and 29.76 mA/cm2 at NA 1020 cm−3, respectively. The excessive absorption of free carriers, which rises linearly along with the number of carriers, might be blamed for the decrease in short circuit current at greater doping concentrations [23] . Figure 3(c) demonstrates the influence of simultaneous variations in NA and thickness on the Fill Factor (FF). The greatest FF was achieved when the NA was 10−19 cm−3 and the thickness was 0.5 µm. The range of FF from 82.88% to 89.65% was observed when the NA was greater than 1015 cm−3 and the thickness was between 0.5 µm and 2.5 µm. Similarly, FF was obtained ranging 80.35% to 81.51% when NA was less than 1015 cm−3 and thickness was equaled or surpassed 1.5 µm. Additionally, FF was more affected by the variation of NA. Figure 3(d) shows the impact on PCE due to varying NA and thickness simultaneously. When the NA is less than 1019 cm−3 and the thickness is between 0.5 µm to 2.5 µm, the PCE is achieved from 28.58% to 32.14%. It was found that PCE was less affected by thickness variations. When thickness was altered from 1 µm to 2.5 µm and NA was above 1018 cm−3, the maximum PCE was achieved from about 33.23% to 34.32%. The maximum value of PCE was obtained when NA was 1019 cm−3 and thickness was 2 µm. Therefore, the thickness and acceptor doping density both exert a substantial implication on the solar cell’s overall performance. In this study, the optimal thickness and the absorber doping density of absorber layer were kept at 2 µm and 1 × 1018 cm−3 considering the structure size and fabrication cost for further optimization.
![]()
Figure 3. Contour plot visualization of the effect on PV parameters with the variation of thickness and doping density.
3.2. Impact of Absorber Defect Density on PV Cell
Figure 4 depicts the effect of defect density of the absorber layer with and without the BSF layer. The defect density was varied from 1010 to 1016 cm−3 in both structures. All PV parameters were stable up to a defect density of 1015 cm−3. When the defect density rises above that limit, a reduction in all parameters is observed. With the BSF layer, the Voc, Jsc, FF, and PCE all dropped from 1.06 V to 0.94 V, 32.45 mA/cm2 to 31.29 mA/cm2, 86.45% to 82.86%, and 28.85% to 22.87%, respectively. The structure without the BSF had similar effects, but the range of reduction was different. Voc drops from 0.897 to 0.874 V, Jsc from 32.45 mA/cm2 to 31.29 mA/cm2, FF from 85.43% to 81.83%, and PCE from 24.25% to 22.39% at the same range of variation. Because electron-hole pair formation is hindered by an excessive defect level, Jsc decreases as defect density increases. Additionally, the process of Shockley-Read-Hall (SRH) carrier recombination causes a decline in Voc with an improvement in the dark current [22] [23] [24] [25] . Therefore, the optimal value of defect density was set to 1014 cm−3.
3.3. Impact of Buffer Layer Thickness, Doping and Defect Density on Solar Cell
A solar cell’s buffer layer is essential for eliminating electrons and holes from both sides of the cell’s structure. A better electron-hole pair formation is achieved by increased photon absorption when a larger band gap buffer substance is utilized instead of the absorber material to transmit incoming light to the junction
![]()
Figure 4. Defect density variation of CZTS absorber layer.
region. Additionally, an effective gathering of carriers produced by photons was provided by the controlled carrier (electron) flow from the photo-active area of the cell to the exterior metal electrode (front contact) [20] . To enable the most incoming light to pass easily a thickness that is as thin as feasible is needed. However, a very thin thickness might result in an audible leakage current [26] . Figure 5(a) shows how changing the thickness of the buffer layer (CdS) affects the PV parameters with and without the BSF layer. The thickness of buffer layers ranged from 0.01 µm to 0.1 µm. Changes in thickness had no effect on PV parameters for either case with or without the BSF layer due to the thick absorber layer. The structure including the BSF layer performs better than the other one. Figure 5(b) shows that the Voc curve remains nearly the same for both with and without the BSF layer at 1.01 V and 0.90 V when the doping concentration of the CdS layer is increased from 1012 cm−3 to 1019 cm−3. In the presence of a BSF layer, Jsc is 32.54 mA/cm2 regardless of doping concentration. The Jsc value stays at 31.14 mA/cm2 till a doping density of 1016 without the BSF layer. Beyond this point, the Jsc steadily increases with increased doping density until it hits a saturation point at roughly 31.5 mA/cm2 at 1017 cm−3 doping density. The fill factor of the two structures with and without the BSF layer remains around 85.30% and 67.30%, respectively, up to a doping concentration of 1014 cm−3, after which the FF of the structure without BSF layer gradually rises to a maximum value of 85.41% as the doping density increases. The structure with the BSF layer, on the other hand, exhibits a slight boost in the FF after doping concentration 1014 cm−3, with a maximum value of 87.37%. Similar to FF, PCE has a similar behavior, with values remaining nearly constant until 1014 for both the presence of the BSF layer and the case without the BSF. Thereafter, further increasing the doping density causes the PCE to increase to maximum values of 28.74% and 24.17%, respectively. Figure 5(c) illustrates the repercussions of the defect density of the buffer layer on the PV parameters, both in the presence and absence of the BSF layer. The range of defect density observed was between 1010 cm−3 and 1016 cm−3. The observed variations in defect density did not have discernible effects on the PV characteristics, regardless of the presence or absence of the BSF layer. Due to its smaller thickness, concentration, and high bandgap which was demonstrated in prior research, CdS buffer layer defect density has a negligible impact on performance parameters [20] . Hence, it can be inferred that the implementation of the BSF layer enables the achievement of improved PV parameters at a lower doping density of the buffer layer. In its absence, even with higher doping concentration, comparable performance remains unattainable. The optimum thickness, doping density, and defect density were therefore adjusted to 50 nm, 1018 cm−3, and 1014 cm−3, respectively.
3.4. Impact of BSF Layer Thickness and Doping Density on Solar Cell
The effect of CuSbS2 BSF thickness and doping concentration on device performance has been investigated in Figure 6(a) and Figure 6(b) noticeably
![]()
(a) ![]()
(b) ![]()
(c)
Figure 5. Variation of (a) thickness (b) doping concentration and (c) defect density of buffer layer with and without BSF layer and its effect on PV parameters.
![]()
(a)![]()
(b)
Figure 6. Variation of (a) thickness and (b) doping concentration of BSF layer and its effect on PV parameters.
demonstrates that all the PV parameters remained steady in relation to the thickness of BSF. There could be no denying that infected absorption rises as BSF layer thickness is enhanced [27] . The simulation shows that, increasing the thickness of the BSF layer could not contribute to increasing the built-in- potential at the interface in the structure. However, all the PV parameters showed significant variation based on doping concentration. With the increase of doping concentration from 1012 to 1018 cm−3, Voc was increased from 0.503 to 1.012 V, Jsc from 32.03 mA/cm2 to 32.53 mA/cm2, FF from 79.43% to 87.35%, and efficiency from 12.82% to 28.74%. It is found that all the parameters rose rapidly up to a doping density of 1015 cm−3 after that it became almost saturated. The enhancement might be caused by the impact of adding additional dopants, which increases the concentration of free carriers and acceptors, which reduces the interdiffusion of grains inside the BSF layer and passivates flaws [28] . Further increase in doping retained almost the same result of parameters. Therefore, the optimal value of thickness and doping density was determined at 0.05 μm and 1015 cm−3 to reduce fabrication cost.
3.5. Effect of Temperature Variation
The detrimental effect of operational temperature from 270 K to 330 K on performance parameters is shown in Figure 7. According to simulation, Voc, FF,
![]()
Figure 7. Variation of temperature and its effect on PV parameters.
and PCE decreased as temperature increased, whereas Jsc nearly remained constant for the recommended structure with as well as without BSF. At 270 K and 330 K, the efficiency of the cell without a BSF layer was found to be 26.30% and 22.17%, respectively, while the PCE of the suggested SC with a BSF layer ranged from 30.54% to 26.88% over the same temperature range. These simulation results show that a CZTS SC with a BSF layer can have better thermal stability than a device without a BSF layer. Similar findings have been observed in previous studies [19] .
3.6. C-V Characteristics Curve
Figure 8 depicts the correlation between capacitance and applied voltage within the specified range of −0.8 V to 0.8 V for the configuration that includes a back surface field (BSF) layer. The experiment was carried out at a frequency of 1 MHz to mitigate the impact of deep level traps. The image provided demonstrates that an enhancement in the supply voltage results in a significant exponential expansion in capacitance. The figure exhibits a non-linear shape as a result of the presence of several junctions. However, it can be deduced that the device reaches the depletion region when the bias is adjusted to a value of zero. The implementation of forward bias leads to a notable growth in capacitance. The utilization of reverse bias results in a significant decrease in capacitance.
3.7. Impact of J-V & QE Characteristics
Figure 9(a) demonstrates the J-V characteristics of CZTS SC with the presence of a BSF layer and without the BSF layer. It is observed from the Figure that CZTS SC without BSF layer produced a PCE of 24.23% with Jsc of 31.56 mA/cm2, Voc of 0.8980 V, and an FF of 85.47%, whereas the addition of BSF layer produced PCE of 28.74% with Jsc of 32.53 mA/cm2, a Voc of 1.0123 V, and FF of 87.27%. That means, the PV parameters determined by numerical simulation of the SC containing CuSbS2 are significantly greater than the structure without BSF. The improvement in the parameters due to the addition if the BSF layer is
![]()
(a)![]()
(b)
Figure 9. (a) J-V characteristic curve and (b) QE curve.
already identical with the previous research works [29] . The quantum efficiency of the recommended structure with and without BSF has been demonstrated in Figure 9(b). The visible light spectrum is mostly covered by both structures but the recommended solar cell with BSF layer has showed higher performance compared to that of without BSF. This is because of enhancement in absorption due to the inclusion of BSF layer. The similar tendency of QE has already been reported in the previous studies [2] .
3.8. Output of SC Parameters
A comprehensive analysis and summary of previous research work on CZTS solar cell architectures are presented in Table 3. The numerical simulation of CZTS SC used CuSbS2 as the BSF layer showed improved SC performance over earlier research.
![]()
Table 3. Performance evaluation for existing PSCs based on CZTS.
This research has explored the potential benefits of combining CZTS SC with a BSF layer made of CuSbS2. The CuSbS2 BSF layer facilitates carrier transportation, promotes their accumulation at the electrodes, and minimises carrier recombination at the interface which in turns, enhancing the performance of PV parameters.
4. Conclusion
This research work employed SCAPS-1D software to numerically analyze the impact of the CuSbS2 BSF layer on the PV parameters of CZTS solar cell. During the simulation, the impacts of thickness, defect density, doping concentration, quantum efficiency, and temperature on solar cell output parameters are studied. Incorporating CuSbS2 as a BSF layer led to a PCE of 28.76%, Voc of 1.0123 V, Jsc of 32.53 mA/cm2, and FF of 87.35%, while without BSF layer-based CZTS structure yielded a PCE of 24.21%, Voc of 0.898 V, Jsc of 31.56 mA/cm2, and FF of 85.32%. This implies that the overall performance of the SC is greatly improved by using CuSbS2 as the BSF layer. The optimized structure yielded the highest efficiency for CZTS, CdS, and CuSbS2 layer thicknesses of 2 μm, 0.05 μm, and 0.05 μm with carrier concentrations of 1 × 1016 cm−3, 1 × 1018 cm−3, and 1 × 1018 cm−3 at 300 K. This numerical simulation shows that CZTS-based thin film solar cells can perform better with an appropriate BSF layer.
Acknowledgements
The authors express their thankfulness to Dr. Marc Burgelman from the University of Gent, Belgium for generously providing the SCAPS 1-D simulation software.