Design and Optimization of CsPbI₃ Perovskite Solar Cells Using SCAPS-1D

3.1 Validation of simulation

Before proceeding with the detailed parameter optimization, the baseline simulation results of our unoptimized CsPbI3 solar cell model were compared with experimental data of similar CsPbI3 architectures available in recent literature. This direct comparison demonstrates that our initial model parameters yield highly realistic device performance metrics, thereby validating the fundamental physical models and material definitions utilized within our SCAPS-1D simulation framework.

3.2 Thickness of absorber layer

Figure 2 is the J-V curve obtained from the simulation. Figure 3 shows the simulation results for the effect of the absorber layer thickness on efficiency (PCE), Voc, Jsc, and FF with a thickness variation of 100–2000 nm. As the absorber layer thickness increases, efficiency increases, reaching a maximum of 17.66% at 1600 nm. The CsPbI3 material reaches its fundamental absorption depth limit around 1 μm, meaning it absorbs nearly all incident photons with energies above its bandgap. Consequently, increasing the thickness beyond this optimal point does not significantly contribute to additional photogeneration, but instead unnecessarily increases the travel distance for charge carriers, which subsequently raises the probability of non-radiative recombination before they can be effectively collected at the electrodes. This correlation between device efficiency and absorber thickness is inherently non-linear. The non-linearity arises from the fundamental physical competition between two opposing processes within the semiconductor: the generation of charge carriers and their subsequent recombination. Initially, the rapid increase in efficiency is heavily dominated by enhanced photon absorption and carrier generation; however, as the thickness continues to grow, the recombination rate begins to dominate due to longer carrier transit times, causing the efficiency curve to flatten and eventually decline [27]. Therefore, finding the optimal absorber thickness is necessary to achieve a good balance between light absorption and charge-carrier transport [28,29].

Moreover, the increase in absorber thickness causes the Voc value to decrease. This decrease occurs because the reverse saturation current increases as charge accumulates on the electrodes. When charge builds up on the electrodes, current can no longer flow through the device. The value of Jsc increases with thickness because the spectral response at longer wavelengths increases with thickness. The increase in electron-hole pairs is related to photon absorption. Therefore, Jsc increases with the absorber thickness. The value of the Fill Factor (FF) also decreased. This occurred due to photon absorption and charge-carrier recombination. The optimum thickness value to achieve maximum efficiency is obtained at a thickness of 1.6 μm with a PCE value of 17.66%, a Voc value of 1.13 Volts, a Jsc value of 20.79 mA/cm2, and FF of 74.96%.

Furthermore, it is important to contrast this theoretically optimized thickness of 1.6 μm with typical experimental data. In practical device fabrication, the optimal thickness for CsPbI3 absorber layers is frequently found to be significantly lower, typically in the range of 300 to 500 nm [15,16]. This discrepancy arises because our ideal SCAPS-1D simulation does not fully account for the macroscopic morphological defects, severe strain accumulation, and high series resistance that typically plague experimentally deposited perovskite films when their thickness exceeds 500 nm. Thus, while our simulation identifies the fundamental optical limit for maximum photogeneration, practical fabrication must balance optical absorption against the severe morphological deterioration associated with thick films.

3.3 Defect density of absorber layer

The absorber defect density strongly influences the solar cell efficiency. Photoelectrons are produced when the absorber layer interacts with light. As the absorber thickness increases, efficiency increases, but defect density also increases, thereby increasing recombination. PCE, Voc, Jsc, and FF values decreased when the Nt value of the absorber was increased from 1014 to 1017 cm−3. The PCE decreased from 16.4% to 9.1% due to non-radiative Shockley-Read-Hall (SRH) recombination [30,31,32], carrier recombination [30,33], and device performance degradation [34,35,36]. Higher defect states act as active recombination centers deep within the bandgap, which severely diminishes the minority carrier lifetime and diffusion length according to Shockley-Read-Hall (SRH) statistics. One way to reduce the value of Nt is by using an ETL layer. This ETL layer can prevent direct reactions between the perovskite absorber and the electrode and can also increase the crystallinity of the perovskite absorber. After optimization, an efficiency of 16.4%, Voc 1.19 V, Jsc 18.33 mA/cm2, and FF = 75.27% are obtained. Comparatively, while our simulation demonstrates that reducing the defect density to 1014 cm−3 maximizes performance, typical experimental CsPbI3 films exhibit defect densities in the range of 1015 to 1016 cm−3 due to inherent iodine vacancies and imperfect crystallization [15,16]. Achieving the theoretically optimal 1014 cm−3 level in practical devices remains a significant challenge that requires advanced surface passivation and controlled crystallization techniques.

3.4 Acceptor concentration (N_a) of absorber layer

Figure 6 and Figure 7 result from simulating a CsPbI3-based solar cell by varying the Na value from 1015 cm−3 to 1018 cm−3. There is an increase in the PCE, Voc, and FF values and a decrease in the Jsc value when the Na value increases. When the Na value is increased, there is a decrease in the energy level of the hole, which increases the Voc value. This increase also occurs because the built-in potential increases. This increases the value of Voc because the charge separation also increases. The optimal Na value was obtained (1018 cm−3) with an efficiency value of 20.90%, Voc = 1.31 V, Jsc = 18.21 mA/cm2, and FF = 87.48%. It is worth noting that experimentally achieving such a high intentional doping concentration of 1018 cm−3 without inducing structural disorder or secondary non-perovskite phases is highly challenging. Most experimental high-efficiency CsPbI3 devices rely on unintentional self-doping or mild extrinsic doping that reaches approximately 1016 cm−3 to 1017 cm−3 [16]. Therefore, this theoretical optimum serves as an upper bound, highlighting the potential gains if highly effective, non-destructive doping strategies can be experimentally realized.

3.5 Thickness of ETL and HTL

ETL is an important layer that reduces recombination losses and should exhibit high optical transmittance to allow efficient light absorption in the perovskite layer. The ZnO thickness varied from 50 to 500 nm. Figure 8(a) shows how ETL thickness affects efficiency, Voc, Jsc, and FF. There was an insignificant decrease in Voc, Jsc, FF, and PCE with increasing ETL thickness. The thickness of the ETL partially absorbs light, inhibiting the rate of charge generation and collection. This decrease in performance can also result from a reduction in transmittance. The thickness of Spiro-OMeTAD was varied from 50 to 500 nm to find the optimal value, as shown in Figure 8(b). As with ETL thickness variations, no significant influence was observed during HTL thickness modulation. Therefore, the HTL thickness was chosen to be 170 nm for further optimization, which was the initial setting of this study. The thinner the ETL and HTL layers, the higher the efficiency value. The HTL layer needs to be thicker than the ETL layer to reduce the possibility of recombination, as it facilitates prompt transport of the same number of charge carriers to the terminal. Thicker HTL than ETL can make the light-absorbing layer more capable of absorbing photons. These simulated optimal ranges—approximately 50 nm for the ZnO ETL and 170 nm for the Spiro-OMeTAD HTL—are in excellent agreement with state-of-the-art experimental fabrications. In high-efficiency experimental CsPbI3 solar cells, the ETL is typically deposited at thicknesses around 30–50 nm to minimize series resistance, while the HTL is coated at 150–200 nm to ensure complete coverage of the rough perovskite surface while maintaining efficient hole extraction [15,19].

3.6 Donor density (N_d) of ETL

Figure 9 shows the influence of the donor concentration (Nd) of the ETL layer with variations of 1017−1019 cm−3. The results obtained are similar to the results of variations in the acceptor concentration of the absorber, where there is an increase in the efficiency, Voc, and FF values and a decrease in the Jsc value. However, the Nd variation did not experience a significant increase or decrease, so it did not significantly affect the performance of the CsPbI3 perovskite solar cell. The optimum value of Nd is 1019 cm−3 because the Nd value must be greater than the Na value. After all, the greater the donor concentration, the more free electrons, so that it can absorb photons with an efficiency value of 16.38%, Voc=1.19 V, Jsc=18.33 mA/cm2, and FF=75.20%. This optimally simulated donor density of 1019 cm−3 aligns perfectly with experimental realities. ZnO thin films intrinsically possess a high concentration of shallow donor states, primarily arising from oxygen vacancies and interstitial zinc atoms, which typically yield carrier concentrations in the 1018 to 1019 cm−3 range [25]. Thus, our simulation accurately reflects the inherent electrical properties of experimental ZnO ETLs.

3.7 Working temperature

Figure 10 and Figure 11 present the effect of operating temperature (275–475 K) on the photovoltaic performance of the CsPbI3-based perovskite solar cell, including the J-V characteristics and key parameters such as efficiency (PCE), open-circuit voltage (Voc), short-circuit current density (Jsc), and fill factor (FF). While standard testing conditions are near room temperature (300 K), deployed solar cells in real-world applications often operate at significantly elevated temperatures under concentrated sunlight or in hot, arid climates, easily reaching 350 K or more. Therefore, simulating temperatures up to 475 K is absolutely critical for rigorously evaluating the extreme thermal stability limits of the CsPbI3 device and understanding the failure mechanisms associated with increased intrinsic carrier concentrations and enhanced phonon scattering. As the temperature increases, a clear degradation in device performance is observed, with efficiency, Voc, and FF decreasing significantly, while Jsc shows a slight increase. The reduction in Voc with increasing temperature is primarily attributed to an increase in intrinsic carrier concentration, which enhances the reverse saturation current. This leads to higher recombination rates and a reduction in the quasi-Fermi level splitting, ultimately lowering the open-circuit voltage. Since Voc is highly sensitive to recombination, it becomes the most affected parameter under elevated temperatures. The fill factor (FF) also decreases with temperature due to increased carrier scattering and recombination losses, as well as degraded diode quality. Higher temperatures can reduce carrier mobility and increase resistive losses, which contribute to a less ideal J-V curve shape.

In contrast, Jsc slightly increases with temperature. This behavior can be attributed to the narrowing of the semiconductor bandgap at higher temperatures, which allows the absorption of lower-energy photons, thereby increasing the photogenerated current. However, this improvement is relatively small compared to the losses in Voc and FF. As a result, the overall efficiency decreases with increasing temperature, since the reduction in Voc and FF dominates over the slight increase in Jsc. The J-V curves further confirm this trend, where the curve shifts toward lower voltage as temperature increases, indicating a reduction in Voc and degradation in device performance. Overall, these results highlight that temperature is a critical factor affecting the stability and performance of perovskite solar cells. The degradation at higher temperatures is mainly governed by increased recombination, bandgap narrowing, and deterioration of diode characteristics rather than defect generation alone. From an experimental perspective, prolonged exposure to temperatures approaching 400 K can trigger a detrimental phase transition in CsPbI3, where the photoactive black perovskite phase (α or γ) degrades into the non-photoactive yellow phase (δ) [15,16]. While our simulation captures the electronic degradation mechanisms such as enhanced recombination and intrinsic carrier generation, practical device failure at these elevated temperatures is often dominated by this structural instability, underscoring the critical need for phase-stabilization engineering in real-world applications.

3.8 Optimized J–V characteristic and quantum efficiency (QE)

Before optimizing the CsPbI3 perovskite solar cell device, the efficiency was 16.4%, the Voc value was 1.19 V, the Jsc value was 18.34 mA/cm2, and the FF value was 75.2%. After optimization of the FTO/ZnO/CsPbI3/Spiro-OMeTAD device, the thickness and Na were from 0.5 μm to 1.6 μm and 1015 to 1018 cm−3. The Nd value of ETL was optimized from 1017 to 1019 cm−3. After final optimization, efficiency increased to 23.1%, Voc to 1.39 V, Jsc to 20.5 mA/cm2, and FF 87.4%. Figure 12 shows the comparison of the J-V curve and quantum efficiency before and after optimization. Detail of optimized parameters is summarized in Table 3. The theoretically optimized PCE of 23.1% provides a compelling target for future experimental efforts. Currently, the record experimental efficiency for all-inorganic CsPbI3 solar cells hovers around 20–21% [16]. The gap between our theoretical limit and current experimental records suggests that while significant progress has been made, there remains substantial room for improvement through rigorous control of interface energetics, suppression of deep-level defects, and precise tailoring of charge transport layer properties as guided by our simulation.

It is seen that as the voltage increases, current density also increases, and the correlation between wavelength and QE is optimized. QE is the ratio of the number of charge carriers produced photoelectrically to the number of photons entering the solar cell. The QE value increases when the wavelength is 300-710 nm. The highest QE is found in ∼360 nm (100%). The QE is gradually decreased by shifting the wavelength to ∼710 nm. After optimization, the QE value is almost maximum (100%) in the 300-710 nm range.