Fig. 2 shows how the photovoltaic parameters (V_oc, J_sc, FF, and η) change as the ZnO electron transport layer (ETL) thickness increases from 100 nm to 800 nm in steps of 50 nm, while the absorber layers CH₃NH₃PbI_3−xCl_x and CH₃NH₃SnI₃ are fixed at 800 nm each, FTO at 200 nm, and ZnTe at 100 nm.

The (J_sc) shows only a marginal increase from 20.59 mA/cm² at 100 nm to 20.62 mA/cm² at 800 nm, as illustrated in Fig. 2a. The slight improvement is likely due to minor reductions in resistive losses or small optical effects, but the variation is too small to be considered significant. In contrast, the results reveal that (V_oc) remained constant at 1.964 V across the entire thickness range, as shown in Fig. 2b, indicating that ZnO thickness has a negligible impact on the built-in potential under the simulated conditions. This suggests that interface energetics and absorber properties dominate V_oc rather than ETL thickness.

The (FF) remains stable at approximately 93.1% across all thicknesses. Otherwise, the (η) followed the trend of J_sc, increasing marginally from 37.66% at 100 nm to 37.72% at 800 nm, as shown in Fig. 2c,d. The extremely small variation < 0.06% absolute change confirms that ZnO thickness beyond a certain minimum (350 nm) has little influence on overall device performance.

The reported open-circuit voltage (V_oc ≈ 2.0 V) and fill factor (FF > 93%) may initially appear unusually high for conventional single-junction perovskite solar cells, which typically exhibit V_oc values below 1.2 V and FFs in the range of 75–85%. However, these values become physically plausible in the context of the tandem (or multi-absorber) architecture employed in this work. The device integrates two perovskite subcells with complementary bandgaps approximately 1.55–1.60 eV for CH₃NH₃PbI_3−xCl_x and 1.2–1.3 eV for CH₃NH₃SnI₃. In a properly connected two-terminal tandem configuration, the total V_oc is expected to be close to the sum of the individual subcell V_oc values (typically ~0.95–1.10 V for the wide-gap cell and ~0.75–0.85 V for the narrow-gap cell), yielding a combined V_oc in the range of 1.7–1.95 V. A value near 2.0 V, while at the upper end, is not implausible under optimized conditions (e.g., low non-radiative recombination, excellent interfacial passivation, and balanced photocurrent).

Similarly, a fill factor exceeding 93% is rare but has been reported in high-performance tandem cells, particularly those with low series resistance, high shunt resistance, and minimal carrier recombination losses at interlayers [9,36].

The use of highly conductive, heavily doped p⁺⁺ZnTe and high-mobility ZnO likely contributes to reduced resistive losses, while the graded perovskite heterojunction may suppress interface recombination.

In the proposed device architecture, the electron transport layer (ETL) based on ZnO exhibits negligible dependence on thickness once it exceeds a minimum threshold of approximately 350–600 nm. Below this range, incomplete surface coverage of the FTO substrate leads to shunting paths and reduced electron extraction efficiency. However, for thicknesses ≥ 350 nm, the ZnO film forms a continuous, pinhole-free layer that fully covers the underlying FTO, ensuring efficient electron collection and effective hole blocking.

Physically, this behavior arises because ZnO functions primarily as a charge-selective contact, not a light-absorbing layer. Once the film is thick enough to prevent direct contact between the perovskite and FTO (which would cause recombination and leakage), further increases in thickness do not significantly alter device performance provided the material maintains high crystallinity and low defect density. Additionally, ZnO’s high electron mobility (~10–100 cm²V⁻¹s⁻¹) ensures that even relatively thin layers can transport extracted electrons efficiently to the electrode without introducing substantial series resistance.

Thus, the minimum effective thickness is ~350 nm, dictated by the need for morpho-logical continuity and interfacial isolation rather than optical or transport limitations. However, a thickness of 600 nm provides good device efficiency, making it the recommended standard for optimal performance.

Fig. 3 presents the variation of key photovoltaic parameters as the ZnTe hole-transport layer (HTL) thickness was varied from 100 nm to 350 nm in steps of 50 nm, while keeping the CH₃NH₃PbI_3−xCl_x and CH₃NH₃SnI₃ absorber layers fixed at 800 nm each, FTO at 200 nm, and ZnO at 100 nm.

The (J_sc) remained essentially constant at 20.57–20.59 mA/cm² throughout the thickness range, as observed in Fig. 3a, implying that optical absorption and carrier generation are unaffected by ZnTe thickness in this configuration. Fig. 3b shows that (V_oc) a gradual increase with ZnTe thickness, rising from 1.964 V at 100 nm to 1.996 V at 350 nm. This indicates that a thicker ZnTe layer improves the built-in potential, possibly due to reduced interface recombination. This behavior provides important insights into the optical and electrical design of the optimized device architecture.

ZnTe has a bandgap of ~2.26 eV, which renders it transparent to photons with energy below this value—i.e., wavelengths longer than ~550 nm. However, in the present solar stack, all above-bandgap photons are already absorbed within the two perovskite absorber layers:

The wide-gap top cell (CH₃NH₃PbI_3−xCl_x, E_g ≈ 1.55 eV) absorbs photons in the 300–780 nm range.

Since ZnTe lies behind both absorbers (adjacent to the reflective Au back contact), and since it does not absorb significantly in the spectral region where the bottom cell operates, increasing its thickness introduces negligible parasitic absorption. Consequently, the photogenerated carrier density remains unchanged, and J_sc stays constant.

The p⁺⁺ZnTe layer is heavily doped (N_A > 10¹⁹ cm⁻³), resulting in: high hole conductivity, short Debye length, and strong built-in field at the perovskite/ZnTe interface.

The insensitivity of J_sc to ZnTe thickness indicates that:

The device is optically optimized: all usable photons are harvested in the perovskite stack; no significant light reaches or is lost in the HTL.

J_sc often decreases with HTL thickness due to low mobility and parasitic absorption. The robustness of J_sc here underscores the advantage of inorganic, heavily doped HTLs like p⁺⁺ZnTe in high-performance tandem PSCs.

The constant J_sc confirms that the architecture achieves complete photon harvesting and lossless hole extraction, validating the choice of ZnTe as an ideal HTL for all perovskite tandem solar cells.

The (FF) exhibits slight fluctuations between 93.00% and 93.41%. The highest FF is observed at 200 nm thickness, suggesting an optimal balance between reduced series resistance and minimal recombination. In addition, the (η) increases from 37.66% at 100 nm to a maximum of 38.24% at 350 nm, as depicted in Fig. 3c,d. The enhancement is primarily attributed to the combined effects of increased V_oc and marginal improvements in FF, while J_sc remained stable.

When comparing the two layers, ZnO-ETL thickness variation had minimal impact on performance, while ZnTe-HTL thickness variation produced measurable improvements in V_oc and η. The ETL performance plateau suggests that a thickness of 600 nm is adequate for optimal operation. The HTL, on the other hand, benefits from moderate thickening of 350 nm, which improves device efficiency without introducing significant resistive losses. This highlights the importance of targeted thickness optimization based on layer function within the device architecture. Comparison of the effects of ZnO-ETL and ZnTe-HTL thickness on photovoltaic performance parameters V_oc, J_sc, FF, and η for the simulated double-heterojunction PSC with fixed absorber thicknesses of 800 nm for CH₃NH₃PbI_3−xCl_x and CH₃NH₃SnI₃ is determined in Fig. 4.

The finalized device configuration consisted of an FTO layer (200 nm), ZnO-ETL (600 nm), CH₃NH₃PbI_3−xCl_x and CH₃NH₃SnI₃ absorber layers (800 nm each), and a ZnTe-HTL (350 nm). Simulation outputs include the J-V characteristics, quantum efficiency (QE) spectrum, Electric field distribution, and energy band diagram for this configuration, as shown in Fig. 5.

The simulated J-V curve for the optimized structure, Fig. 5a, shows an ideality factor of 0.999, a series resistance of 0.0076 Ω·cm², a J_sc of approximately 20.60 mA/cm², and a V_oc of 1.996 V. The FF remains above 93%, leading to a power conversion efficiency of about 38.32%. The near-rectangular J-V curve shape indicates minimal resistive losses. More importantly, it confirms efficient carrier extraction, a key factor in the device’s high performance.

Fig. 5b demonstrates the QE spectrum with a strong photoresponse across a broad spectral range. Quantum efficiency begins at 32% at 200 nm, increases steadily with wavelength, and reaches a maximum exceeding 99% at 360 nm. It then averages around 80% in the visible range, from 400 to 700 nm. The QE drops sharply beyond about 780 nm, corresponding to the bandgap limitations of the absorber layers. The high, flat QE plateau across most of the visible spectrum confirms efficient photon absorption and carrier collection in both absorber layers.

The distribution of the simulated electric field (E) across the optimized solar cell’s thickness is shown in Fig. 5c. A sharp spike in E of 3.4 × 10⁵ V·cm⁻¹ occurs at the transparent contact/ETL interface, creating the initial built-in field and facilitating electron collection into the ETL. A second, distinct maximum of 3 × 10⁴ V·cm⁻¹ indicates the main electron-selective junction, where the field quickly dissociates photocarriers generated near the front absorber. A moderate peak of 5 × 10⁴ V·cm⁻¹ appears at the internal absorber/absorber junction, aiding the movement of carriers across the band-gap step and reducing interfacial recombination between the two photoactive sub-layers. The most significant field enhancement, reaching about 2 × 10⁶ V·cm⁻¹ at the hole-selective junction, ensures efficient hole extraction and effective electron blocking. Between these peaks, the field levels off to a nearly uniform plateau of 10⁴ V·cm⁻¹, sufficient to support drift-assisted transport across the absorber layers while minimizing resistive losses.

Fig. 5d displays the energy band diagram of the optimized double heterojunction solar cell. The conduction band edge (E_c), valence band edge (E_v), and the quasi-Fermi levels for electrons (E_fn) and holes (E_fp) are plotted across the whole device thickness (0–2.75 μm).

In the TCO region, E_c remains nearly flat at 0.02 eV, and E_v is positioned around −3.58 eV, confirming the wide bandgap and low recombination probability in this layer. At the TCO/ETL interface 0.20 μm, a sharp step in both E_c and E_v is observed, producing the internal potential barrier that drives carrier separation. Within the perovskite absorber region, E_c gradually rises from 0.3 eV to over 1.2 eV, while E_v shifts upward from −3.3 eV to −0.2 eV, showing strong band bending associated with photogenerated charge separation. Toward the HTL, the conduction band increases steeply to 2.3 eV, while the valence band aligns near −0.12 eV, which is favorable for hole transport and electron blocking.

The quasi-Fermi levels split substantially across both absorber sub-cells. The maximum E_fn − E_fp separation approaches about 2.0 eV, consistent with the measured V_oc of 1.99 V at open circuit. The flat E_fn near the ETL and the close tracking of E_fp with the HTL valence band reflect efficient, selective contacts at both sides.

Table 2 compares the photovoltaic performance parameters of the optimized device with representative perovskite solar cells reported in the literature.