In this section, we present all obtained simulation results and a discussion of each finding. Numerical simulation allows several studies to receive an optimized structure of the solar cell. The section is divided into different subsections. Firstly, there are the calibration of this simulation work and the published experimental work to make sure the simulation results meet the practical applications. The obtained current density, voltage curve and main performance parameters, short circuit current density, open circuit voltage, fill factor, and power conversion efficiency, are compared with the published experimental work. Secondly, the codesign concept is applied to the calibrated cell to enhance its output performance parameters. Moreover, the optimization of the basic parameters of the calibrated cell to enhance its performance is introduced. The optimization starts with optimizing the thickness and doping of the electron transport layer. After that, the thickness, the doping, and the bulk defect density of the absorber layer are optimized to enhance the output performance parameters.
3.1. Calibration versus Experimental Work
Measuring calibrating simulation results against experimental data is an essential step in ensuring the accuracy and reliability of computer models. This process involves adjusting the parameters of the simulation model to match the results of real-world experiments, with the aim of minimizing the discrepancy between the two. Notably, experimental data provide a benchmark against which the accuracy of simulation models can be assessed. By comparing the results of the simulation model with experimental data, researchers can identify discrepancies and refine the model to improve its accuracy. In addition, calibration helps to ensure that simulation models are not based on incorrect assumptions or oversimplifications. Thus, this work started by calibrating the simulated solar cell with a published experimental work [
29]. As mentioned before, SCAPS is used to analyze the performance metrics:
VOC,
JSC, FF, and PCE. The JV curve is the most significant output metric to measure solar cell performance. As a sequence, the JV curves of the simulated cell and the reported experimental work [
29] are illustrated in
Figure 3. It can be deduced from
Figure 3 that the simulated device using SCAPS-1D agrees with the experiment [
29]. By neglecting the series and shunt resistances of the cell in the simulation, the performance metrics of the two cells are given in
Table 3 for the purpose of comparison with the same thickness and doping as in the experimental [
29].
The performance factors obtained from the experiment and simulation results in
Table 3 show a slightly higher
JSC than the experimental results, which is most probably due to a better photo-absorption in the absorber layer. It is worth mentioning that the shading loss in the front metal contact on the top is ignored due to the one-dimensional geometry considered, which will influence it by reducing
JSC [
32]. A very small difference between
VOC measures can be noticed, which might result from the multiple reflections from the interfaces, mechanisms which are not included in SCAPS [
33]. The fill factor and the power conversion efficiency have a subtle difference of less than 0.1%. The output performance parameters,
VOC,
JSC, FF, and PCE, besides the JV curves of the experimental work [
29], and the simulated solar cell show the validity of the used materials parameters and design.
3.2. Codesign and Optimization
Theoretical studies give physical insight with a thorough physical explanation of the output performance. The codesign problem could be solved utilizing practical combinatorial methods, such as with an orthogonal composition gradient in the utilized layers [
34]. CuSbCh
2 [
35], Cu
2ZnSnCh
4 [
36], CIGS [
37], and CdTe [
38] absorbers or CdS [
39], Cu
xZn
1-xS [
40], Zn–Ni–Co–O [
41], and Mg
xZn
1-xO transport layers are being compositionally graded and investigated using combinatorial methods [
42].
In this part of the design, the composition
x of the Cu
2ZnSn(S
xSe
1-x)
4 absorber is tuned to adjust the electron affinity and the energy bandgap, and composition
y of the Cd
yZn
1-yS ETL is concurrently adjusted to achieve the optimum efficiency.
Figure 4a illustrates the PCE variation with the composition
x (of the absorber) and composition
y (of the ETL). As illustrated in the figure, the optimum PCEs are achieved in two points: point 1 (
x = 0.8,
y = 0.52) and point 2 (
x = 0.7,
y = 0.14), with a PCE exceeding 14%. The illuminated JV characteristics of the initial design and the two optimum case cells are illustrated in
Figure 4b. The simulation results reveal a boost in the
VOC of the codesign compared to the initial cell. This necessitates the usefulness of the codesign technique in alleviating the
VOC deficit encountered in this type of thin-film solar cell. The performance metrics (
VOC,
JSC, FF, and PCE) of the initial design and the two optimum cases are listed in
Table 4 for comparison.
To give a physical insight into the previous results, the issue of band alignment, which is responsible for such results, is highlighted. Notably, reducing the charge recombination at the interface between ETL and the absorber is very important to enhance carrier extraction. To accomplish this, the CBO at the ETL/absorber interface should be 0–0.3 eV [
43], where the CBO is given with Equation (6),
where
and
are the affinities of the photoactive layer and ETL, respectively.
Figure 5 illustrates the CBO dependence on different x and y values. The affinities of the ETL and the absorber with their corresponding energy gaps are listed in
Table 5 for the three scenarios.
Based on the theoretical calculations of Shockley and Queisser who reported that the maximum PCE can be achieved for an energy gap of 1.34 eV [
44], it is expected that a higher PCE is achieved (with respect to the calibrated cell) when the energy gap ≅ 1.33 eV (which corresponds to
x = 0.7). However, the highest PCE is obtained for a slightly higher bandgap of about 1.38 eV (which corresponds to
x = 0.8), as indicated in
Table 4. This can be attributed to the crucial role of the CBO as illustrated in
Figure 6, which shows the energy band diagrams of these two cases compared to the initial structure. Although the initial design reveals a favorable spike-like band offset as
Figure 6a signifies, the low bandgap results in a lower PCE. For the case of
Eg = 1.33, a CBO of −0.073 eV is obtained (
Figure 6b and
Table 5), which corresponds to an unfavorable cliff-like band. Conversely, a CBO of 0.175 eV is observed for
Eg = 1.38 eV (see
Figure 6c and
Table 5), signifying a favorable low spike-like band offset. These interesting results show the importance of the alloy codesign as the sole design of the absorber parameters may not be sufficient to optimize the cell performance. Instead, the ETL should be simultaneously designed with the absorber to meet both superior absorption and band alignment at the same time.
Furthermore, the optimization after the alloy composition codesign of the ETL and absorber compositions started with optimizing the thickness and doping of the ETL of the two optimum cases as illustrated in
Figure 7.
Figure 7a demonstrates that the performance improves at a certain thickness for both cases of composition. Since most of the photons with short wavelength are absorbed near the surface, it is desirable to have a thinner ETL. As can be deduced from
Figure 7a, the optimum thickness is around 0.02 µm with PCEs of ≅14.13%, and 14.3% for cases 1 and 2, respectively. Additionally, one can observe in
Figure 7b that varying the doping concentration from 10
18 cm
−3 to 10
20 cm
−3 results in an increase in the PCE. The highest efficiency is achieved approximately in the range 5 × 10
19 cm
−3 to 10
20 cm
−3 due to an increase in conductivity. The PCE increases to about 14.3% for the two cases, as shown with a slightly higher PCE for
x = 0.7.
The influence of the thickness, the doping concentration, and the defect density of the absorber material on the PCE has been investigated as illustrated in
Figure 8. The thickness of the photoactive layer is a critical factor since it is responsible for absorbing the incident photons. However, it should not be excessively thick to allow for an effective collection of the photogenerated carriers before recombination. In the simulation, the thickness of the absorber layer varied from 0.3 μm to 2 μm, as depicted in
Figure 8a. As can be inferred from the figure, the PCE does not show a significant effect from increasing absorber thickness beyond 1 µm. Further, decreasing the doping of the absorber material makes a gradient doping at the interface between the absorber and the ETL, which raises the diffusion current and enhances the PCE, as can be inferred from
Figure 8b. Additionally, the higher doping concentration leads to an improved built-in electric field, facilitating better charge-carrier separation and ultimately enhancing the overall performance of the cell. However, it is important to note that further increasing the doping concentration can result in a higher recombination rate. This phenomenon is not conducive to an increase in the open-circuit voltage and, consequently, the PCE. Therefore, there is a delicate balance to be maintained, ensuring optimal doping concentrations for achieving the desired cell performance. As can be deduced from
Figure 8b, the optimum doping level for the absorber layer in both scenarios is
cm
−3.
Moreover, the bulk defects of the absorber material have a great influence on the PCE. As can be inferred from
Figure 8c, there is no significant improvement in the PCE below the defect density of 10
14 cm
−3. The optimum factors of the previous optimization are listed in
Table 6. Combining these optimization steps yields a high
VOC of 0.8 V, a
JSC of 28.97 mA/cm
2, a FF of 73.72%, and a PCE of 17.16%. The JV and EQE curves of the initial and optimum cells are illustrated in
Figure 9 for comparison.
To compare this work with some recent publications in the literature, a group of similar structures has been collected and illustrated in
Table 7. In [
45], the PCE of the CZTSSe solar cell device has been notably increased from 7.4% to 8.5% by adding Zn to the CdS buffer layer. In [
46], a study aimed to investigate the effect of three different buffer layers, namely CdS, ZnS, and CdZnS, on the performance parameters of CZTS solar cells through numerical simulations. The simulation results showed that a CdZnS buffer layer achieves high efficiency CZTS solar cells. PCEs utilizing these buffer layers were around 11.20%. In [
47], the utilization of a CdZnS buffer layer has resulted in a 13.23% conversion efficiency in CZTS solar cells, which is an improvement over the conventional buffer layer, CdS, that only yielded 12.61%. In [
48], compared to the CZTS/CdS device, the CZTS/CdZnS device exhibited an increase in both the conversion efficiency and fill factor. Specifically, the PCE increased from 2.31% to 4.88%.
Finally, the co-design of the elemental compositions of the electron transport layer (ETL) and the absorber represents a new field, and its widespread application across various materials has not been widely explored. One of these few efforts was devoted to the codesign of alloy compositions of CdSe
xTe
1_x (as an absorber material) and Mg
yZn
1-yO (as an ETL) [
49]. In their experimental study, the authors revealed that the solar cell performance is a strongly coupled function of both x and y compositions. They achieved a PCE of up to 17.7% at specific atomic compositions (x = 0.04 and y = 0.18) [
49]. As this field of research is advancing, a need for simulation studies is mandatory to investigate and evaluate new material systems and to provide design guidelines for the corresponding solar cells. This study serves as a significant contribution by shedding light on this emerging field, paving the way for further advancements.