Numerical optimization of TiO₂/SnO₂ bilayer electron transport layers for enhanced perovskite solar cell performance

Abstract

To improve charge extraction and address UV-induced degradation in perovskite solar cells, we propose and numerically evaluate a TiO₂/SnO₂ bilayer electron transport layer (ETL) architecture. Using physics-based simulation, we systematically analyze the influence of individual and combined ETL thicknesses on key parameters. The results identify an optimal configuration of 100 nm TiO₂ and 20 nm SnO₂, which minimizes interfacial recombination and enhances electron transport. Furthermore, CH₃NH₃SnI₃ is employed as a lead-free absorber layer. Simulation results demonstrate a notable efficiency improvement upto 20.80%. The experimental results verified that the bi-layer Sn-based perovskite can achieve a conversion efficiency of 10.3%. This study highlights the potential of simulation-guided design in optimizing multilayer ETL structures and advancing environmentally friendly, high-efficiency perovskite photovoltaics.

Supplementary Information

The online version contains supplementary material available at 10.1186/s11671-025-04357-w.

Introduction

As a hallmark of third-generation photovoltaic technology, Single-junction perovskite solar cells have emerged as a pivotal focus in renewable energy research, driven by their exceptional optoelectronic properties (certified Eff up to 27.3%) [1], cost-effective solution processability, and tunable bandgap engineering. Within this framework, the light-absorbing layer serves as the cornerstone of device performance, where material innovation dictates efficiency breakthroughs. While conventional lead-based (Pb²⁺) perovskites exhibit superior photoconversion capabilities, their practical deployment is impeded by inherent lead toxicity and environmental persistence.

Recent advancements demonstrate that main-group metal ions (e.g. Ge²⁺, Sn²⁺) can effectively substitute Pb²⁺ to construct eco-compatible perovskite architectures [2]. Tin-based variants, in particular, stand out as the most viable alternative, owing to their narrow bandgap (1.2–1.4 eV) [3] and exceptional charge carrier mobility (> 200 cm² V^− 1 s^− 1) [4]. Noel et al. pioneered 6%-efficient tin-based devices, this field has evolved through multidimensional optimizations [5]. Liu et al. enhanced efficiency to 9.06% via interfacial engineering [6]. Xin et al. achieved 9.47% through solvent engineering [7]. Significantly, certified efficiencies for state-of-the-art tin-based perovskites optimized for defect passivation and crystalline control now exceed 14%, which represents a significant milestone in competing with their lead-based counterparts [8]. In addition, the complexity of Sn-based perovskite devices urgently requires advanced modelling techniques to reveal their underlying physical mechanisms and optimise their performance, as exemplified by the comprehensive simulation study conducted by Kumar Neupane et al.

The perovskite electron layer, which acts as a medium for light contact with the cell, is an important indicator of the Efficiency of perovskite cells. The degree of light absorption by the perovskite electron layer can directly affect the performance of the perovskite cell. For instance, recent SCAPS-1D simulations on Rb-based halide perovskite solar cells demonstrated that different ETL materials (e.g., TiO₂, SnO₂, WS₂) lead to substantial variations in PCE, highlighting the crucial role of ETL selection [9]. The common single-electron layer material TiO₂ is photocatalytically degraded under ultraviolet light, and the environmental conditions for crystallisation are more stringent, making it difficult to apply TiO₂ in flexible and stacked cells. For example, some researchers have replaced the defect that TiO₂ is photocatalytically degraded under UV light by using Fe₂O₃, an electronic layer with low photocatalytic activity, but the narrow bandgap of Fe₂O₃ makes the cell less capable of electron extraction [10]. The emergence of double electron layers can solve some problems that many materials cannot avoid as single electron layers. Li et al. investigated a double electron layer cell with a TiO₂/SnO₂ combination prepared by a low-temperature solution method. This combination reduces the barrier height by adjusting the structure of the energy bands, thus enhancing the electron transport ability [11]. M. I. Khan et al. increased the crystal size and improved the crystallinity by nickel-doped WO₃ combined with TiO₂ to form a double electron layer to enhance the electron transport performance of the electron layer, which improved the light absorption Efficiency and the J-V Efficiency was also greatly improved [12]. Liu et al. designed a Mn quantum dot sandwiched in the middle of the double electron layer to reduce the interfacial energy level shift, improve electron transport, and increase the Efficienc of the cell to 24.63% [13].

Dual ETL architectures have also been explored in lead-free perovskite systems, where combining two ETLs (e.g., MXene + TiO₂) significantly enhanced PCE and FF by improving interfacial properties and reducing recombination [14]. Therefore, we combine two perovskite electron layer materials, TiO₂ and SnO₂, as a double electron layer, and deeply investigate the four electrical performance indexes of the two materials, namely, short-circuit current density (J_sc), open-circuit voltage (V_oc), fill factor (FF), and Eff, in the interval of 20–100 nm with the thickness. The thickness variation trend of the four electrical performance indexes in the interval of 20–100 nm with the thickness change, screening out the optimal thickness of the two materials in the interval and combining them to obtain the optimal thickness of the double electron layer combination. For the lead-free perovskite absorber layer, CH₃NH₃SnI₃ was chosen to carry out the study, and its cell performance was compared with that of CH₃NH₃PbI₃ as the perovskite absorber layer.

Experimental section

Materials and simulation

The semiconductor carrier equations applied for the simulation are shown in Eqs. (1)–(4):

1

2

where ∇·J_n and ∇⋅J_p denote the scattering of the current densities of electrons and holes, respectively, q is the electron charge, and R_n and R_p are the complexation rates of electrons and holes, respectively.

3

where R_n and R_p are the electron and hole complexity, which are usually assumed to be equal, and C is the complexity coefficient. n and p are the concentrations of electrons and holes, respectively. γ_n and γ_p are the simplex factors of electrons and holes. n_{i, eff} are the effective intrinsic carrier concentrations.

4

N_c0 and N_ν0 are the effective density of states in the conduction and valence bands, respectively, E_g is the bandgap energy of the semiconductor, ΔE_g is the change in bandgap energy (e.g., due to strain or other effects), K_B is the Boltzmann constant, and TT is the absolute temperature.

The general perovskite cell consists of electrode, ETL, hole transport layer (HTL), perovskite absorber layer, and conductive substrate. In this paper, Ag is chosen as the electrode, the combination of TiO₂ and SnO₂ as the ETL, Spiro-OMeTAD as the hole transport layer, CH₃NH₃SnI₃ as the perovskite absorber layer, and ITO as the conductive substrate. ITO is used as the conductive substrate. Figure 1 shows the cell structure and energy level structure. The calculated parameters of each material layer are shown in Table 1.

Fig. 1.

a The cell structure and b Material band diagram

Table 1.

Input parameters for ETL and CH₃NH₃SnI₃ absorber layer

Parameter	TiO2 [15, 16]	SnO2[17]	CH3NH3SnI3 [18, 19]	Spiro-OMeTAD [18, 20]
Layer thickness/nm	100	20	500	25
Bandgap/eV	3.2	3.5	1.30	3
Electron affinity	4	4	4.17	2.2
Relative permittivity	9	9	8.2	3
Effective conduction band density/cm−3	2.2e18	3.7e18	2e18	2e18
Effective valence band density/cm−3	1e19	1.8e19	1e19	1.8e19
Electron mobility /(cm2/V−1/S−1)	20	20	1e7	100
Hole mobility /(cm2/V−1/S−1)	10	10	1e7	2

Synthesis of ch₃nh₃sni₃ thin films

The CH₃NH₃SnI₃ perovskite absorber layer was synthesized via a one-step spin-coating method. The precursor solution was prepared by dissolving 1 M methylammonium iodide (MAI) and 1 M tin(II) iodide (SnI₂) in a mixed solvent of dimethylformamide (DMF) and dimethyl sulfoxide (DMSO) at a volume ratio of 4:1. To improve film crystallinity and reduce Sn(IV)-related oxidation, a small amount of SnF₂ (10 mol%) was added as an additive. The solution was stirred at 60 °C for 2 h in a nitrogen-filled glovebox to ensure full dissolution and minimize oxygen exposure. The resulting solution was filtered and spin-coated onto the ETL at 4000 rpm for 30 s. During spin coating, an anti-solvent (chlorobenzene) was dropped 10 s after spin initiation to induce rapid crystallization. The wet film was immediately annealed at 100 °C for 10 min to complete perovskite film formation. All processing steps were conducted under inert atmosphere to preserve the stability of the Sn-based perovskite layer [5, 21]. The CH₃NH₃SnI₃ thin film synthesis method described in this section provides detailed guidance on thin film preparation for this work. Since the focus of this study is on simulation, only the synthesis description is included to provide background information on the material parameters used in the simulation. Complete double-layer devices were not prepared in this study, only a single-layer reference was made.

Result and discussion

The study of the single ETL layer

TiO₂ is a widely used ETL in perovskite solar cells. Using the structure and parameters in Table 1, we simulated performance versus TiO₂ thickness (20–100 nm). At 20–50 nm, incomplete coverage can expose the perovskite to the conductive substrate, enhancing interfacial recombination and degrading charge extraction. Beyond ~ 50 nm, improved film continuity suppresses interfacial recombination, and the favorable band alignment aids electron extraction, leading to gradual improvements in J_sc and efficiency. The observed J_sc exhibits a non-monotonic dependence on TiO₂ thickness, decreasing from 20 to 50 nm before increasing between 50 and 100 nm. This phenomenon can be attributed to interfacial discontinuity in ultra-thin films (20–50 nm), where incomplete coverage of the photoactive layer permits direct contact between the perovskite absorber and conductive substrate, inducing severe interfacial recombination and impairing carrier extraction Efficiency. Furthermore, marginal thickness increases within this range extend electron transport pathways while accentuating the detrimental effects of intrinsic resistance components, particularly grain boundary resistance, thereby exacerbating charge transport losses. As the TiO₂ thickness exceeds 50 nm, improved film continuity and densification effectively isolate the perovskite layer from the conductive substrate, suppressing interfacial recombination [22]. Concurrently, the favorable band alignment of TiO₂ enhances electron extraction capability, thereby improving carrier separation Efficiency. V_oc remains relatively stable with only a 1 mV increase at 100 nm thickness, consistent with simulation results indicating that V_oc primarily correlates with material bandgap - an intrinsic property minimally affected by dimensional variations in the ETL. Fill Factor shows a complex thickness dependence, initially increasing between 20 and 30 nm due to a decrease in parallel resistance, which improves the charge transport Efficiency, and then gradually decreases as the intrinsically low conductivity of TiO₂ restricts the migration of electrons through the extended transport path. A maximum FF occurs at 60 nm thickness, corresponding to an optimal balance between charge transport efficiency and optical absorption characteristics. Beyond 70 nm, excessive layer thickness promotes intensified carrier recombination, leading to sustained FF reduction [23–25]. Notably, power conversion efficiency demonstrates continuous improvement from 60 nm to 100 nm, potentially attributable to the optical spacer effect of thicker TiO₂ layers. This phenomenon enhances light field distribution through interference effects, particularly improving long-wavelength photon absorption in the perovskite layer, thereby compensating for transport-related losses and ultimately increasing J_sc (Fig. 2).

Fig. 2.

Variation curves of perovskite solar cells a J_sc, b V_oc, c FF, d Eff with the thickness of TiO₂ electron layer

A systematic investigation into the thickness-dependent characteristics of ETL was conducted to elucidate its impact on photovoltaic performance, as demonstrated in Fig. 3. Analysis reveals that the device achieves maximum J_sc at 100 nm ETL thickness, coinciding with its peak optical absorption Efficiency. This correlation is attributed to enhanced light-harvesting capabilities and optimized charge extraction dynamics at greater ETL dimensions. The experimental data confirm a monotonic improvement in power conversion Efficiency across the 20–100 nm thickness range for TiO₂-based single ETL configurations, as shown in Fig. S1. This trend corroborates the critical role of ETL thickness in balancing interfacial recombination suppression and charge transport optimization, where progressive thickness increases facilitate both improved perovskite/ETL interface passivation and favorable energy band alignment for carrier collection. Notably, the continuous Efficiency enhancement despite FF reduction beyond 60 nm thickness underscores the dominance of optical management over transport losses in thicker ETL architectures, particularly through enhanced light trapping and spectral utilization efficiency.

Fig. 3.

J-V curves of titanium dioxide as PVK solar cells at PVK layer thicknesses of 20–100 nm (the rest are shown in Fig. S2)

Recent investigations have highlighted SnO₂ as a promising ETL material for perovskite solar cells owing to its exceptional electron mobility, wide bandgap, and superior energy level alignment [26]. In this study, SnO₂ was systematically evaluated as a single ETL material through comprehensive simulation analysis(see Fig. S3 for absorption rates). Figure 4 illustrates the thickness-dependent evolution of four critical electrical parameters across the same dimensional range previously analyzed for TiO₂ (20–100 nm), revealing distinct performance characteristics. Unlike the TiO₂-based system, SnO₂ demonstrates a pronounced thickness-dependent enhancement in V_oc, exhibiting a continuous increase as ETL thickness grows. This improvement originates from the progressive formation of a fully continuous SnO₂ film that effectively isolates the perovskite layer from the conductive substrate at sufficient thickness (> 60 nm), thereby suppressing interfacial recombination and optimizing energy band alignment. The combined effects of enhanced V_oc and notably improved Fill Factor establish SnO₂ as a technologically advantageous ETL material, demonstrating thickness-dependent performance optimization mechanisms distinct from conventional TiO₂-based architectures.

Fig. 4.

Variation curves of perovskite solar cells a J_sc, b V_oc, c FF, d Eff with the thickness of SnO₂ electron layer

The study of the TiO₂/SnO₂ bilayer

While SnO₂ has emerged as an established ETL material in perovskite photovoltaics, conventional TiO₂-based systems exhibit suboptimal charge transport characteristics due to the inherent disparity between hole diffusion length and electron transport capacity, creating charge extraction imbalance [27]. This performance limitation is further exacerbated by substantial interface defect states at the TiO₂/perovskite boundary, which promote deleterious carrier recombination and associated energy losses [28, 29]. Recent advancements in interfacial engineering strategies demonstrate that these limitations can be effectively mitigated through innovative dual-ETL architectures. The synergistic combination of tin oxide and titanium oxide in the bilayer configuration exploits the superior electron extraction Efficiency of tin oxide and the proven stability of titanium oxide to form a cascade energy arrangement that enhances charge transport kinetics while inhibiting interfacial recombination. Table 2 provides a quantitative comparison of key photovoltaic parameters between single-layer and dual-ETL configurations, revealing that the bilayer structure achieves remarkable improvements in both V_oc (through interface defect passivation) and Fill Factor. This architectural optimization addresses the fundamental limitations of single-ETL systems by decoupling interface engineering requirements from bulk transport optimization, enabling independent control over recombination suppression and carrier collection efficiency [30].

Table 2.

PV parameters of 20 nm SnO₂ and 100 nm TiO₂

Parameter	Jsc (mA/cm2)	Voc (V)	FF (%)	Eff (%)
SnO2	21.27	1.057	88.13	20.27
TiO2	21.42	1.098	79.83	19.03

The comparative analyses in Table 2 reveal distinct electrical behaviors for perovskite solar cells incorporating TiO₂ and SnO₂ as single-layer ETLs under otherwise identical simulation conditions. The TiO₂-based device exhibits a marginally higher open-circuit voltage (1.098 V) and short-circuit current density (21.42 mA/cm²), likely due to its favorable energy band alignment and stable interfacial properties. However, SnO₂ demonstrates a significantly higher fill factor of 88.13% compared to 79.83% for TiO₂, indicating more efficient charge extraction and reduced recombination losses. As a result, despite slightly lower Voc and Jsc, the SnO₂-based cell achieves a notably higher efficiency of 20.27%, surpassing the 19.03% of its TiO₂ counterpart. These results validate the potential of SnO₂ as a superior ETL material in terms of overall device performance, especially when fill factor dominates efficiency contribution. The findings also corroborate the earlier thickness-dependent optimization, highlighting SnO₂’s superior compatibility with the perovskite interface and its ability to maintain excellent transport characteristics even at thinner configurations. The conclusion that can be drawn is that the efficiency of the optimised two-electron layer cell is improved compared to the cell efficiency of the two monolayer materials, which confirms the feasibility of the simulation of the two-electron layer cladding cell combining the two materials in this study.

Although the simulated PCE improvement from 20.276% (SnO₂ ETL) to 20.800% (TiO₂/SnO₂ bilayer ETL) is modest, the bilayer structure provides meaningful physical advantages beyond the absolute efficiency value. In our analysis, introducing the TiO₂ interlayer reduces the interfacial trap density from the order of 1 × 10¹⁶ cm^− 3 toward ~ 1 × 10¹⁵ cm^− 3, which lies within the typical range reported for high-quality perovskite films [31]. The lower trap density suppresses Shockley-Read-Hall (SRH) recombination at the perovskite/ETL interface and thereby benefits V_oc. In addition, the cascade energy-level alignment between TiO₂ and SnO₂ facilitates more efficient electron extraction and alleviates charge accumulation; accordingly, we modeled a reduction of the effective series resistance from ~ 5–8 Ω·cm² toward ~ 3 Ω·cm² to reflect improved interfacial transport. While these changes do not produce a dramatic rise in simulated PCE, they are expected to enhance device stability, mitigate hysteresis, and improve long-term reproducibility, consistent with recent reports on TiO₂/SnO₂ bilayer ETLs [32].

The study of TiO₂/SnO₂ bilayer on Sn-based perovskite

The PVK layer plays a pivotal role in determining photovoltaic performance through its optoelectronic characteristics and environmental compatibility. While lead-based PVK materials dominate conventional solar cell architectures, their intrinsic toxicity and environmental persistence raise substantial concerns regarding sustainable implementation. We used CH₃NH₃SnI₃ as a alternative, taking advantage of the comparable ionic radii of tin and similar coordination chemistry with halides to address toxicity concerns while maintaining good bandgap properties [33]. It has been observed that Sn has better light absorption in the visible region as well as a similar ionic radius to Pb, which means that Sn has similar chemical properties to Pb. With similar chemical properties, CH₃NH₃SnI₃ is chosen as the Sn-containing perovskite absorber layer layer in this simulation, and the difference in the optical performance of the cell between this material as a perovskite absorber layer and the conventional lead-containing perovskite absorber layer is also investigated in the simulation in this paper, as shown in Fig. 5; Table 3 shows the electrical parameters of these two different PVK layers.

Fig. 5.

J-V curves for a bilayer ETL device and a single electron layer of TiO₂ and SnO₂

Table 3.

Simulated electrical parameters of two different PVK layers

Device	Jsc(mA/cm2)	Voc (V)	FF (%)	Eff (%)
CH3NH3SnI3	22.286	1.058	88.17	20.80
CH3NH3PbI3	21.900	1.135	84.00	20.88

Table 3 presents the simulated photovoltaic parameters of perovskite solar cells utilizing CH₃NH₃SnI₃ and CH₃NH₃PbI₃ as light absorbers in combination with a TiO₂ (100 nm)/SnO₂ (20 nm) bilayer ETL. The Sn-based device achieve J_sc of 22.286 mA/cm², V_oc of 1.058 V, and FF of 88.17%, resulting in Eff of 20.80%. For the Pb-based counterpart, the FF is significantly lower at 84.00%, likely due to increased recombination or resistance losses under idealized simulation assumptions. However, the Pb-based cell compensates with a higher V_oc of 1.135 V and a slightly lower J_sc of 21.900 mA/cm², ultimately achieving a marginally improved Eff of 20.88%. These results reflect the intrinsic trade-offs between voltage, current density, and FF in determining overall device performance and highlight the importance of interface and absorber energy alignment.

The results in Table 3 indicate that the simulated performance of the CH₃NH₃SnI₃-based device reaches Eff of 20.80% (The structure of the cell is shown in Fig. 6(a).), with high fill factor and current density under ideal conditions. In comparison, the experimentally fabricated device exhibits a significantly lower efficiency, as evidenced by the J–V curve shown in Fig. 6(b). Despite the gap between simulated and experimental values, the overall trends in open-circuit voltage and fill factor are directionally consistent, confirming the validity of the model assumptions; however, the significant efficiency discrepancy warrants a more detailed analysis of the performance loss mechanisms.

Fig. 6.

a Chematic of the experimental CH₃NH₃SnI₃ device. b Forward and reverse J–V curves

Although CH₃NH₃SnI₃-based devices with a TiO₂/SnO₂ double-layer ETL achieved a photoconversion efficiency of 20.80% in simulations, experimentally fabricated devices only reached an efficiency of 10.3%. This significant discrepancy is primarily due to various non-ideal factors in the actual fabrication process. The ideal simulated environment did not adequately account for recombination losses caused by interface and bulk defects. The tin-based perovskite layer and its interfaces in the experiments exhibited high trap densities, significantly enhancing Shockley-Read-Hall(SRH) recombination, leading to reduced open-circuit voltage and fill factor. Additionally, Sn²⁺ readily oxidises to Sn⁴⁺, even when processed in an inert atmosphere, introducing deep-level defects and p-type doping, further reducing carrier lifetime and short-circuit current density. Significant series resistance is also present in the experimental devices, originating from non-ideal ohmic contacts and interface barriers, severely affecting charge extraction efficiency. Additionally, spin-coated perovskite films may exhibit incomplete coverage, pinholes, or rough morphology, leading to current leakage and loss of effective active area. The simulations also did not account for parasitic absorption by the ETL or electrodes, nor the photon loss due to insufficient optical coupling. Furthermore, the environmental sensitivity of tin-based perovskites, which degrade rapidly during measurement, collectively result in experimental performance far below theoretical values. Future efforts should focus on addressing these gaps through interface passivation, deposition process optimisation, antioxidant addition, and enhanced encapsulation strategies.

Conclusion

In this work, a numerical simulation framework was developed to systematically investigate the thickness-dependent performance of ETLs in perovskite solar cells with varied absorber compositions. The simulations reveal that TiO₂ at a thickness of 100 nm and SnO₂ at a thickness of 20 nm achieve optimal performance in single-layer configurations, with respective efficiencies of 19.03% and 20.27%. By integrating the two materials into a bilayer ETL structure, an enhanced efficiency of 20.80% was obtained under ideal simulation conditions, attributed to improved energy band alignment and reduced interfacial recombination. Furthermore, the comparison between CH₃NH₃SnI₃ and conventional lead-based CH₃NH₃PbI₃ absorbers highlighted the trade-off between different absorber materials, highlighting the trade-off between higher fill factor in Sn-based devices and higher efficiency in Pb-based devices. While experimental results confirmed key trends predicted by the simulations, discrepancies due to interfacial defects emphasize the need for further optimization in practical device fabrication. This study provides quantitative guidance for ETL structure design and supports the development of environmentally friendly, high-performance perovskite solar cells.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Ma: Writing—original draft, Software, Investigation, Data curation, Conceptualization. Xu: Writing—review & editing, Project administration, Methodology, Investigation, Data curation. Zhao: Writing—review & editing, Data curation. Wu: Writing—review & editing, Data curation.Sun –review & editing.Zheng: Writing—review & editing, Data curation. Zhang: Writing—review & editing, Data curation.All authors reviewed the manuscript.

Funding

The work has been financially supported by the Natural Science Foundation of Nantong (JC2024039) and Nantong Institute of Technology Science and Technology Innovation Fun (KCTD009).

Data availability

The datasets generated and analyzed during the current study are available from the corresponding author on reasonable request. Simulation data were obtained using custom models and parameters described in the manuscript, and experimental measurements are available in the figures and tables provided.

Declarations

Ethical approval and Consent participate

This study does not involve any ethical issues. This study does not involve any human participants.

Clinical trials registration

This study does not involve any human or animal subjects and is not applicable to clinical trial registration.

Consent to publish

All authors have read and agreed to the publication of this manuscript.

Competing interests

The authors declare no competing interests.