WorksPhysics and Astronomy
Three-Dimensional Optical-Electrical Simulation of Cs2AgBiBr6 Double Perovskite Solar Cells
Abstract
Despite significant advances in lead-free perovskite photovoltaics, achieving a balance among environmental safety and high optoelectronic performance remains challenging. The inorganic double perovskite Cs2AgBiBr6 has emerged as a promising candidate owing to its robust three-dimensional crystal structure and suitable visible-range bandgap. However, best power conversion efficiencies (PCEs) for Cs2AgBiBr6 solar cells reported so far - 6.37% experimentally and 27.78% in numerical studies - remain below the theoretical performance potential, largely due to suboptimal charge transport layers, and interface-related recombination losses. Here, we address this gap using a 3D finite-element method (FEM) implemented in COMSOL Multiphysics, which couples optical simulations with semiconductor drift-diffusion transport. To our knowledge, this work represents the first comprehensive 3D FEM-based study of a double halide perovskite solar cell. Screening of 25 electron transport layer (ETL)-hole transport layer (HTL) combinations identifies CeO2 and P3HT as the optimal ETL and HTL respectively. Device performance is further analyzed through systematic variation of layer thicknesses, doping concentrations and defect densities within the FTO/CeO2/Cs2AgBiBr6/P3HT/Au architecture. Under optimized parameters, the simulated device achieves a PCE of 31.76%, representing the theoretical upper bound predicted by the model. Overall, this work demonstrates 3D physics-based device engineering as a decisive pathway for overcoming efficiency bottlenecks in lead-free double perovskite photovoltaics.
Introduction
The urgent need to mitigate climate change and reduce dependence on fossil fuel–based power generation has accelerated the global transition toward clean and renewable energy technologies [1], [2]. Among the available renewable resources, solar energy is the most abundant and universally accessible, making photovoltaic (PV) conversion a central pillar of future sustainable energy systems [3]. While crystalline silicon solar cells continue to dominate the commercial market due to their technological maturity, their energy-intensive manufacturing and high processing temperatures motivate continued exploration of alternative absorber materials capable of delivering high efficiency through scalable, low-temperature fabrication routes [4][5]. Third-generation photovoltaic technologies—including dye-sensitized, organic, quantum dot, and perovskite solar cells—have therefore attracted extensive research interest [6]. In particular, lead (Pb)-based hybrid organic–inorganic halide perovskites have achieved a remarkable rise in power conversion efficiency (PCE), from 3.8% in 2009 [7] to 26.95% in 2025 [8], approaching crystalline silicon benchmarks. These materials typically adopt the perovskite structure (Figure 1a), where A denotes a larger organic or inorganic cation (, , , , etc.), B corresponds to a divalent metal cation—most commonly —and X represents a halide anion (, , or ). This crystal model enables compositional tunability while supporting the outstanding optoelectronic properties of perovskite absorbers, including strong optical absorption, long carrier diffusion lengths, low exciton binding energies, and balanced charge transport [9]. However, despite these exceptional properties, large-scale deployment remains constrained by Pb toxicity and long-term instability associated with volatile organic cations [10, 11, 12]. These concerns have driven the search for environmentally benign, structurally stable, lead-free alternatives.

Figure 1. Schematic crystal structures showing the atomic site occupation of (a) perovskite and (b) double perovskite materials.
Early substitution strategies replacing with or were motivated by chemical similarity but suffer from rapid oxidation and severe defect formation, leading to instability and performance degradation [13, 14, 15, 16]. Alternatively, substitution using heterovalent cation combinations has led to the emergence of several families of lead-free halide perovskite derivatives formed to satisfy charge neutrality. These materials span a range of structural motifs, including , and related low-symmetry or vacancy-ordered variants, where A-site cations are typically monovalent (e.g., , , , or organic cations such as ) and B-site cations include trivalent or monovalent metals (e.g., , , , , , ) [17], [18]. While these compositions incorporate environmentally benign cations, their crystal structures are typically zero- to two-dimensional in nature. This reduced dimensionality inherently limits carrier mobility and enhances recombination, resulting in suboptimal photovoltaic performance [19, 20, 21].
Three-dimensional halide double perovskites (HDPs) with general formula offer a promising pathway to overcome both toxicity and dimensionality limitations [10]. In these materials, is replaced by a monovalent (B′) and a trivalent (B″) cation, while preserving the three-dimensional perovskite framework (Figure 1b). This design strategy offers substantial compositional flexibility through diverse choices of A, B′, and B″-site ions, allowing systematic tuning of band structure, carrier transport, and thermodynamic stability [22, 23, 24, 25].
Among the reported HDPs, — crystallizing in the cubic elpasolite-type double perovskite structure (space group ) — has attracted particular attention as a benchmark inorganic lead-free perovskite absorber [11], [26]. The cooperative incorporation of and at the B′ and B″-sites enhances lattice stability, resulting in a high decomposition energy and exceptional resistance to thermal and moisture-induced degradation [27]. In addition to its robust structural stability, exhibits a suitable visible-range bandgap, long radiative and nonradiative carrier lifetimes, relatively low carrier effective masses, and notable defect tolerance—properties that are essential for efficient photovoltaic operation [10], [22], [28]. Importantly, several of these optoelectronic
characteristics approach those of archetypal Pb-based perovskites, while offering the decisive advantages of low toxicity and long-term environmental compatibility [29, 30, 31].
Since its initial incorporation into photovoltaic devices by Greul et al. [32] using a planar FTO//CsAgBiBr/Spiro-OMeTAD/Au architecture, which delivered a PCE of 2.43% and established the experimental feasibility of this absorber, CsAgBiBr has been explored extensively across diverse device architectures, contact layer combinations, and deposition routes, including both solution-processed and vapor-deposited films [33], [34]. Continuous experimental progress has led to the current record efficiency of 6.37%, achieved by Zhang et al. via optimized planar architectures (ITO/SnO/CsAgBiBr/Spiro-OMeTAD/Au), demonstrating excellent operational stability (1440h at 85°C) as well [28]; however, the performance remains modest relative to Pb-based counterparts. In parallel, numerical studies suggest that significantly higher efficiencies may be achievable through appropriate selection of charge transport layers and careful optimization of material and interfacial parameters [35, 36, 37, 38]. For instance, Danladi et al. reported a PCE of 25.56% for the FTO/ZnO/CsAgBiBr/CFTS/Au structure [36]. Hakami et al. proposed an all-inorganic device architecture based on FTO/TiO/CsAgBiBr/NiO/Au, achieving a simulated PCE of 26.71% [37]. Similarly, Srivastava et al. identified an optimized planar FTO/AZO/CsAgBiBr/ZnTe configuration, predicting a PCE of 26.89% [35]. More recently, Raj et al. reported an optimized efficiency of 27.78% for the FTO/TiO/CsAgBiBr/CuO/Au structure [38]. Collectively, these insights motivate a systematic device-level optimization strategy in which suitable electron transport layer (ETL)–hole transport layer (HTL) combinations are first identified, followed by targeted tuning of layer thicknesses, doping concentrations, and bulk and interfacial defect densities, with the aim of identifying realistic pathways toward performance enhancement in CsAgBiBr-based solar cells.
Despite the advances, most prior modeling efforts rely on one-dimensional drift–diffusion solvers that assume simplified generation profiles and neglect spatial optical redistribution and multidimensional interfacial effects. Therefore, realistic modeling of optical field distribution and its coupling to carrier transport is essential for identifying true performance limits and dominant loss mechanisms.
In this context, physics-based finite-element modeling offers a powerful framework for resolving coupled optical and electrical phenomena in heterogeneous device stacks. COMSOL Multiphysics enables self-consistent solutions of Maxwell’s equations and semiconductor transport equations within multidimensional geometries, allowing spatially resolved analysis of photogeneration, band bending, quasi-Fermi level splitting, and recombination processes [39, 40, 41, 42]. However, such a coupled three-dimensional (3D) treatment remains largely unexplored for CsAgBiBr-based solar cells.
Motivated by this gap, the present work develops a comprehensive 3D finite-element opto-electrical framework to investigate CsAgBiBr solar cells across multiple charge transport layer combinations. Twenty-five ETL–HTL pairings are systematically screened to identify optimal contact materials. Building upon the most favorable configuration, layer thicknesses, doping concentrations, bulk and interfacial defect densities, intrinsic recombination coefficients, temperature, and illumination intensity are rigorously examined. By explicitly calculating optical generation profiles and coupling them to drift–diffusion transport, this integrated framework establishes physically grounded performance bounds, identifies critical design tolerance windows, and explains the fundamental loss mechanisms governing CsAgBiBr photovoltaics, thereby providing quantitative guidance toward experimentally realizable, high-efficiency lead-free double perovskite solar cells.
Simulation Framework
Device Structure and Material Parameters
All simulations were carried out using a three-dimensional finite-element method in COMSOL Multiphysics framework. The investigated device architecture adopts a layered heterojunction configuration of FTO/ETL//HTL/Au (Figure 2a). A schematic of the equilibrium energy-band alignment of the complete device stack is shown in Figure 2b.
Figure 2. (a) Schematic illustration of the perovskite (PVK) absorber-based solar cell architecture, with the individual layer thicknesses; detailed material parameters are listed in Table 1-Table 3. (b) Energy band alignment diagram showing the conduction band minima (CBM) and valence band maxima (VBM) of the constituent layers; for each material, the upper and lower numerical values denote the CBM and VBM energy levels, respectively. Five electron transport layers (ETLs)—, AZO, PCBM, , and C60—and five hole transport layers (HTLs)—P3HT, CuO, NiO, PEDOT:PSS, and —were investigated. The conduction and valence band offsets (CBO, VBO) at the absorber/ETL and absorber/HTL interfaces are summarized in Table 4 and Table 5. The symbols “e” and “h” in panel (b) represent electrons and holes. Under illumination, photogenerated electron–hole pairs are created within the absorber; electrons are transported toward the FTO front contact through the ETL, while holes are collected at the Au back contact via the HTL.
| Parameters | ETL/CsAgBiBr interface | CsAgBiBr/HTL interface | Bulk CsAgBiBr |
| Defect type | Neutral | Neutral | Neutral |
| Electron capture cross section (cm) | |||
| Hole capture cross section (cm) | |||
| Energetic distribution | Single | Single | Single |
| Thermal velocity (cm/s) | |||
| Impurity energy level | Midgap | Midgap | Midgap |
| Total density (cm) | (Variable) | (Variable) | cm (Variable) |
Table 3. Input parameters for interface and bulk defects
| ETL | AZO | CeO | PCBM | WS | C |
| CBO | 0.35 | 0.4 | -0.3 | -0.25 | -0.3 |
Table 4. CBO between CsAgBiBr and ETL materials: CBO =
Fluorine-doped tin oxide (FTO) is employed as the transparent front electrode due to its high optical transmittance in the visible region and excellent chemical and thermal stability, making it a well-established and reliable choice for perovskite and inorganic thin-film photovoltaic devices [43].
An ETL is introduced between FTO and the absorber to enable selective extraction of photogenerated electrons while suppressing hole transport toward the front contact. To assess the influence of transport-layer properties, five representative ETL materials − AZO (aluminum-doped zinc oxide), (cerium (IV) oxide), (tungsten disulphide), C60 (buckminsterfullerene) and PCBM ([6, 6]-phenyl-C61-butyric acid methyl ester) − were examined. These materials span a broad range of electronic characteristics, including wide-bandgap metal oxides, layered inorganic semiconductors, and organic fullerene derivatives [44][45]. Such diversity allows a comprehensive evaluation of band alignment, carrier mobility, optical transparency, and interfacial recombination behavior within the device.
is employed as the absorber layer, where photon absorption and electron–hole pair generation occur under illumination. In the simulated architecture, carrier transport within the absorber
is governed by bulk material properties, defect-assisted recombination, and interfacial charge transfer at the ETL and HTL interfaces.
On the rear side of the absorber, a HTL is incorporated to facilitate efficient extraction of photogenerated holes while blocking electron transport toward the back contact. Five HTL materials − P3HT (poly(3-hexylthiophene)), PEDOT:PSS (Poly(3,4-ethylenedioxythiophene) polystyrene sulfonate), CuO (copper (II) oxide), NiO (nickel (II) oxide) and MoO (molybdenum trioxide) − were considered, representing polymeric, oxide, and transition-metal-oxides, with distinct valence-band alignments and hole transport characteristics. These materials have been extensively reported as effective HTLs in perovskite and thin-film solar cells and provide a suitable basis for systematic transport-layer engineering [46, 47].
Gold (Au), with a high work function of approximately 5.1 eV, is selected as the back contact to ensure favorable energetic alignment with the HTL and efficient hole collection while minimizing contact resistance. The top and bottom electrodes of the device were modeled as Ohmic contacts by applying Dirichlet boundary conditions for the electrostatic potential. In COMSOL, these conditions directly fix the potential at the contact nodes, ensuring that the applied voltage is correctly enforced during the solution of the semiconductor equations. The applied voltage bias was introduced at the back contact, while the front contact was grounded. All other external boundaries were treated as electrically insulating to prevent unintended current flow. An air domain was included above the FTO front electrode to represent the surrounding optical medium and to correctly impose the incident
Table 2. Input parameters of ETL materials
electromagnetic boundary conditions. All the material parameters – including bandgaps, electron affinities, dielectric constants, and carrier mobilities – were adopted from previously reported literature.
Unless otherwise stated, all simulations were performed under standard AM1.5G illumination with an incident power density of at an operating temperature of . The complete set of material, and interface parameters employed in the simulations is summarized in Table 1-Table 5 and Figure 3.
Table 1. Input parameters of FTO, CsAgBiBr, and HTL materials: band gap (), electron affinity (), relative dielectric permittivity (), effective density of states in the conduction band (), effective density of states in the valence band (), electron mobility (), hole mobility (), acceptor density (), donor density (), and defect density ()
| HTL | P3HT | CuO | NiO | PEDOT-PSS | MoO |
| VBO | 0.04 | -0.34 | -0.02 | 0.24 | -0.71 |
Table 5. VBO between CsAgBiBr and HTL materials: VBO =
Figure 3. (a,b) Refractive index (n) and (c,d) extinction coefficient (k) spectra of the FTO, absorber, ETL and HTL materials used in the simulations, extracted from Refs. [48], [57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67].
Mesh Generation
A user-controlled, structured mesh was employed for the six-layer device architecture. Each domain was discretized individually with minimum and maximum element sizes of and , respectively. The maximum element growth rate was set to with a curvature factor of and narrow region resolution of to ensure accurate resolution of sharp interfacial gradients. A hybrid meshing scheme combining free triangular elements (surface) and swept quadrilateral elements (volume) was implemented. The top and bottom faces were meshed using free triangular elements, while remaining faces were discretized using a swept quadrilateral sequence. Thickness-dependent element distributions were applied across domains using a combination of linear and exponential growth profiles. For example, the ETL layer was discretized into 80 elements with symmetric linear grading, whereas the perovskite absorber was initially resolved with 50 elements using exponential grading (growth ratio = 5), ensuring refined meshing near the ETL/perovskite heterointerface. Tessellation was set to automatic. Adaptive mesh refinement was enabled to dynamically resolve regions exhibiting large gradients in electrostatic potential, carrier concentration, and recombination rate. Depending on the device configuration, the total number of degrees of freedom (DOFs) solved ranged from approximately for the initial structure to for the optimized device. Mesh independence was confirmed by progressively increasing mesh density until the current–voltage characteristics exhibited negligible variation. Mesh quality was evaluated using skewness and growth rate metrics; the resulting distributions satisfied standard quality criteria, confirming numerical stability and solution convergence.
Theoretical Formulation
The simulations were performed by coupling the Electromagnetic Waves, Frequency Domain (EWFD) interface with the Semiconductor interface within COMSOL Multiphysics. The coupled approach enables the calculation of optical absorption and photogeneration using full-wave electromagnetic analysis, followed by charge carrier transport using a drift–diffusion formalism.
Optical Simulation Using EWFD Interface
The optical response of the device under illumination was computed using the EWFD interface, which solves Maxwell’s equations in three dimensions in the frequency domain to obtain the wavelength-dependent electric field distribution, [68]:
where denotes the electric field vector, and are the relative permittivity and permeability of the materials, respectively, and is the free-space wave number. and are expressed as [69],
where and are the real and imaginary parts of , respectively; and are refractive index and extinction coefficient, respectively; is the angular frequency of the incident radiation, and is the speed of light in vacuum.
A normally incident plane wave corresponding to the AM1.5G solar spectrum was applied at the FTO side to represent solar irradiation. The wavelength-dependent and of each layer were incorporated to account for optical dispersion and absorption. The local volumetric absorbed power density at each wavelength was calculated as:
Here, is the vacuum permittivity, and represents the magnitude of the electric field. The photogeneration rate at each wavelength was then computed as [69],
Here is Planck’s constant and is the photon frequency. To account for the full spectral range of illumination, the wavelength-dependent generation rate was integrated over the relevant wavelength interval for each layer:
The spectrally integrated generation profile was then volume-averaged and applied to the corresponding layers in Semiconductor physics module, where it served as the photogeneration term in the carrier continuity equations for drift–diffusion simulations.
Charge Transport Modeling in Semiconductor Interface
Electrical modeling was performed using the Semiconductor interface, which solves Poisson’s equation coupled with the electron and hole continuity equations within the drift–diffusion approximation.
The electrostatic potential is governed by Poisson’s equation:
where is the relative permittivity, is the elementary charge, and are the electron and hole concentrations, and and represent ionized donor and acceptor densities.
The continuity equations for electrons and holes are given by:
where G is the total carrier generation rate, including photogeneration, and R is the recombination rate. The subscripts n and p correspond to electron and hole, respectively.
The electron and hole current densities are expressed as:
where is the carrier mobility, and D is the diffusion coefficient, related to the mobilities through the Einstein relation:
Where k is the Boltzmann constant and T is temperature. Carrier recombination within the device was modeled using three primary mechanisms: Shockley–Read–Hall (SRH), radiative (direct), and Auger recombination. SRH recombination accounts for defect-mediated, nonradiative recombination:
where is the intrinsic carrier concentration, and and are the electron and hole lifetimes, respectively. Radiative (direct) recombination describes band-to-band recombination:
where is the radiative recombination coefficient. Auger recombination models three-particle interactions, significant at high carrier densities:
where and are the Auger coefficients for electrons and holes, respectively.
Performance Parameters
The coupled optical–electrical problem was solved using a stationary solver. From the converged solutions, current density–voltage (–) characteristics under illumination were extracted, and key photovoltaic parameters—including short-circuit current density (), open-circuit voltage (), fill factor (FF), and PCE—were calculated.
Where is the reverse saturation current density, and refer to the voltage and current density respectively at the maximum power point, and is the input power.
Results and Analysis
Model Validation and Benchmarking
To establish the reliability of the proposed simulation framework, the device model was first validated by reproducing experimentally reported results of a previously published CsAgBiBr-based solar cell with the configuration: ITO/SnO/CsAgBiBr/Spiro-OMeTAD/Au [28]. The simulation parameters, including layer thicknesses and material properties were carefully adopted from the experimental study to ensure a meaningful comparison.
Figure 4 presents a direct comparison between the simulation and experimental results in terms of – characteristics under AM1.5G illumination. The simulated curve closely reproduces the experimental response across the entire bias range, capturing both the slope near open-circuit conditions and the curvature around the maximum power point. This agreement indicates that the model accurately describes the dominant carrier transport and recombination processes governing device operation.
Figure 4. Benchmarking and validation of the simulation framework presented in this work through comparison with experimentally reported J–V characteristics and associated photovoltaic parameters (, , FF, and PCE) of -based solar cell [28].
In addition to the - characteristics, the extracted photovoltaic performance parameters—, , FF, and PCE—are summarized in the Table inside the figure and compared with the values reported in Ref. [28]. The simulated results show close quantitative agreement with the measured data, with only minor deviations that can reasonably be attributed to experimental uncertainties, material inhomogeneities, and processing-induced variations not explicitly captured in the numerical model.
Overall, the strong agreement between simulation and experiment confirms the validity of the adopted physical models and parameter sets. This validation provides a solid foundation for systematic transport-layer engineering, followed by broader material and device-level parametric investigations aimed at optimizing the performance of the -absorber-based solar cell architecture.
Performance Analysis of ETL-HTL Pairings
A comprehensive charge transport-layer (ETL and HTL) study with 25 distinct ETL/HTL combinations was carried out to identify optimal carrier-selective contacts for the absorber. The selected transport materials satisfy established selection criteria reported in prior literature [70], including suitable band alignment, sufficient carrier mobility, optical transparency, and chemical stability. , , FF, and PCE were extracted for each combination (Figure 5), a pronounced dependence of device performance on transport-layer pairing is observed, underscoring the critical role of interfacial energetics and carrier selectivity.
Figure 5. Comparison of performance parameters for 25 ETL/HTL combinations (5 ETLs × 5 HTLs), showing (a) , (b) , (c) FF, and (d) PCE. All the simulations were performed using the following layer thicknesses along the -axis: FTO = 100 nm, ETL = 50 nm, CsAgBiBr = 330 nm, HTL = 350 nm, and Au = 33 nm.
Higher exceeding V are achieved for AZO/P3HT, AZO/NiO, CeO/PEDOT:PSS, WS/CuO, and WS/MoO configurations (Figure 5a). These higher voltages can be attributed to favorable interfacial band alignment and effective suppression of interfacial recombination, which enhance quasi-Fermi level splitting within the absorber. In contrast, combinations with suboptimal energy-level matching exhibit lower , highlighting the sensitivity of voltage output to interface energetics.
Highest of 16.45 mA cm is obtained for CeO/P3HT combination (Figure 5b). This enhanced current response reflects the excellent optical transparency of CeO in the visible spectrum, combined with efficient electron extraction and minimal parasitic absorption in the ETL. Simultaneously, the favorable hole mobility and interfacial compatibility of P3HT facilitate efficient charge collection, leading to superior photogenerated current compared with alternative transport-layer pairings.
FF also varies considerably with transport-layer pairing (Figure 5c). High FF exceeding 85% are achieved for PCBM/MoO, WS/MoO, WS/CuO, and WS/P3HT configurations. These combinations benefit from reduced series resistance, efficient carrier selectivity, and suppressed recombination losses near the maximum power point, resulting in improved diode rectification and more ideal – characteristics.
As a cumulative metric, the PCE reflects the combined trends in , , and FF (Figure 5d). The highest PCE of 15.15% is obtained for the CeO/P3HT configuration, which simultaneously offers the largest along with competitive and FF values. Other combinations exhibiting high voltage or fill factor alone do not achieve comparable efficiencies due to trade-offs in current generation or resistive losses. These results indicate that balanced optimization of charge extraction, optical transmission, and interfacial recombination is essential for maximizing device efficiency.
Based on this comprehensive comparative assessment, CeO and P3HT emerge as the most promising ETL and HTL, respectively, for CsAgBiBr absorber-based solar cell. Accordingly, this ETL/HTL pair is selected for subsequent wavelength-resolved optical analysis and detailed material and device-level parametric optimization.
Optical Response and Carrier Generation Analysis
Wavelength-Dependent Optical Field, Absorbed Power, and Carrier Generation
In this section, the spatial distributions of the optical electric field, local volumetric absorbed power density (), and carrier generation rate within the full device stack with the optimal charge transport layers (CeO and P3HT), obtained from EWFD module (Figure 6-Figure 8), have been analyzed. 3D visualizations provide a global perspective on light propagation, field localization, and absorption across all layers (Figure 6-Figure 8, a-e). To enable clearer layer-resolved interpretation, two-dimensional (2D) cross-sectional maps were subsequently extracted in the yz-plane at a fixed lateral position ( nm), which intersects all functional layers from the FTO front contact to the Au back electrode (Figure 6-Figure 8, f-j).

Figure 6. (a–e) 3D and (f–j) corresponding two-dimensional (2D) cross-sectional distributions of the optical electric field at representative wavelengths. The 2D cross sections are taken in the yz plane at a fixed lateral position of x = 125 nm. Panels (a,f), (b,g), (c,h), (d,i), and (e,j) correspond to λ = 330, 500, 630, 710, and 730 nm, respectively. The color scales represent the electric-field magnitude in V/m. PVK corresponds to the absorber layer.

Figure 8. (a–e) 3D and (f–j) corresponding 2D cross-sectional distributions of the carrier (electron-hole pair) generation rate at representative wavelengths. The 2D cross sections are taken in the yz plane at a fixed lateral position of nm. Panels (a,f), (b,g), (c,h), (d,i), and (e,j) correspond to , 500, 630, 710, and 730 nm, respectively. The color scales represent the generation rate in cms. PVK corresponds to the absorber layer.

Figure 7. (a–e) 3D and (f–j) corresponding 2D cross-sectional distributions of the local absorbed power density at representative wavelengths. The 2D cross sections are taken in the yz plane at a fixed lateral position of nm. Panels (a,f), (b,g), (c,h), (d,i), and (e,j) correspond to , 500, 630, 710, and 730 nm, respectively. The color scales represent the absorbed power density in W/cm. PVK corresponds to the absorber layer.
Thickness-Dependent Carrier Generation in ETL, Absorber, and HTL
Following the wavelength-resolved optical analysis, the influence of individual layer thickness on the volume-averaged carrier generation rate was next examined (Figure 9). The thicknesses of the absorber, ETL, and HTL were varied independently while keeping the remaining layers fixed, allowing the role of each layer in governing overall photogeneration to be isolated.
Figure 9. Wavelength-integrated volume-averaged carrier generation rate inside the absorber (PVK in the legend), ETL, and HTL as a function of layer thickness: (a) absorber, (b) ETL, and (c) HTL. For (a), ETL and HTL thicknesses were fixed at 50 nm and 350 nm, respectively; for (b), absorber and HTL thicknesses were fixed at 800 nm and 350 nm; and for (c), absorber and ETL thicknesses were fixed at 800 nm and 100 nm.
When the thickness of the absorber was increased, a monotonic decrease in the generation rate within the absorber was observed (Figure 9a). This trend arises because optical absorption in is strongly front-loaded, with most photons absorbed near the ETL/absorber interface, as established by the earlier and generation profiles. Increasing absorber thickness therefore enlarges the optically weakly excited rear region, reducing the average generation rate when normalized over the full absorber volume. In contrast, the generation rates within the ETL and HTL remain nearly unchanged and negligibly small, confirming that photogeneration is predominantly confined to the absorber layer and is insensitive to absorber thickness variations in the adjacent transport layers.
A similar but weaker trend is observed when varying the ETL thickness (Figure 9b). Increasing the thickness from 30 nm to 700 nm leads to a slight reduction in the generation rate within the absorber, accompanied by a continuous decrease in generation within the ETL itself. The latter reflects the dilution of absorbed power over a thicker wide-bandgap layer, while the former originates from increased optical attenuation and parasitic absorption in the ETL, which reduce photon flux reaching the absorber. Throughout this thickness range, the generation rate within the HTL remains essentially zero, signifying negligible absorption in the investigated spectral range.
In contrast, variation of the HTL thickness has a negligible influence on photogeneration within the device (Figure 9c). As the thickness is increased from 50 to 400 nm, the carrier generation rate inside the absorber remains nearly unchanged, with only a shallow dip observed around 200 nm. The generation rate within the ETL likewise remains essentially constant, while photogeneration inside the HTL itself is negligible across the entire thickness range. These behaviors reflect the fact that photon absorption occurs predominantly within the absorber, while , owing to its wider bandgap relative to the absorber, does not contribute appreciably to absorption over the relevant spectral range. The minor reduction in generation at intermediate HTL thickness is attributed to thickness-dependent optical field redistribution within the multilayer stack, which slightly modifies the standing-wave pattern near the absorber/HTL interface rather than introducing additional parasitic absorption.
Overall, these thickness-dependent trends reinforce the earlier wavelength-resolved findings: photogeneration in the architecture is dominated by absorption within the absorber and is primarily governed by optical field distribution and interference effects rather than absorption in the transport layers. Variations in ETL and HTL thickness mainly modulate the spatial redistribution of the optical field and photon flux reaching the absorber, whereas changes in absorber thickness predominantly affect the volumetric averaging of the generation rate rather than the total absorbed photon number. These thickness-dependent generation rates are fed into the Semiconductor module for the subsequent charge transport analysis, allowing a self-consistent evaluation of how optical-geometrical interplay influences device performance – an aspect often simplified or neglected in prior studies.
Electrical Performance
Effect of Absorber Thickness and Doping Density
To explore the influence of absorber-layer properties on overall device performance, the thickness and acceptor doping density of the absorber were varied while keeping all other structural and material parameters fixed (Figure 10a-Figure 10d). increases with absorber doping density for a given absorber thickness, with a noticeably stronger rate of increase beyond a doping density of approximately (Figure 10a). When the absorber thickness is varied, increases only weakly at low doping densities (), whereas a more pronounced thickness-dependent enhancement is observed for highly doped absorbers. These trends originate from the evolution of band bending and quasi-Fermi level splitting within the absorber. At low doping densities, weak internal electric fields limit the potential that can be sustained across the absorber, so increasing thickness primarily extends the quasi-neutral region and has little impact on . With increasing doping density, stronger band bending at both the and interfaces enhances field-assisted carrier separation
Figure 10. Effect of thickness and doping density of the absorber, on , , FF, and PCE of the device. All other material parameters were kept fixed at their default values (Tables 1-3).
and suppresses recombination, allowing quasi-Fermi level splitting to be maintained deeper into the absorber and resulting in a stronger thickness dependence of .
exhibits more complex dependencies (Figure 10b). For a fixed absorber thickness, remains nearly independent of doping density up to approximately and decreases at higher doping levels. With increasing absorber thickness, increases at low and moderate doping densities due to enhanced optical absorption, while at the highest doping density studied in this work (), a non-monotonic thickness dependence is observed, characterized by an initial increase followed by a reduction at intermediate thickness and partial recovery at larger thicknesses. These behaviors reflect the competition between absorption-driven current enhancement and doping-induced recombination losses. At high doping densities, increased free-carrier concentration and defect-assisted recombination shorten carrier lifetimes, reducing current collection efficiency, particularly in thicker absorbers where carriers need to traverse longer transport paths.
For a fixed thickness, FF remains nearly constant with increasing doping density up to approximately and increases thereafter (Figure 10c). When absorber thickness is varied, FF generally decreases with thickness, with the reduction being more pronounced at low doping densities and becoming weaker at higher doping levels. At the highest doping density, FF exhibits a sharp enhancement at an intermediate thickness (), followed by degradation at larger thicknesses. These trends arise from the interplay between series resistance, recombination, and field-assisted carrier extraction. At low doping densities, limited conductivity leads to increased transport resistance and carrier accumulation in thicker absorbers, reducing FF. Increasing doping density improves absorber conductivity and strengthens the internal electric field, reducing resistive losses and enabling more efficient carrier extraction, which explains the FF enhancement and its partial recovery at high doping and intermediate thickness.
PCE reflects the combined evolution of , , and FF with absorber doping density and thickness (Figure 10d). For a given thickness, PCE remains nearly constant at low doping densities and increases at higher doping levels, with the onset of improvement occurring at lower doping densities for thicker absorbers. When absorber thickness is varied, PCE generally increases with thickness, and the rate of enhancement becomes more pronounced at higher doping densities. These trends result from the balance between voltage enhancement, photocurrent generation, and fill-factor evolution. At low doping densities, thickness-induced absorption gains are partially offset by transport and recombination losses, limiting PCE improvement. At higher doping densities, stronger internal electric fields and improved conductivity allow thickness-related absorption gains to translate more effectively into increases in efficiency.
Effect of ETL Thickness and Doping Density
Following the absorber-layer analysis, the role of the ETL was examined by varying the thickness and donor doping density of the ETL while maintaining all other device parameters constant (Figure 11a-Figure 11d). For thick ETLs (), increases with ETL doping density for a given thickness, with the enhancement becoming pronounced once the doping density exceeds approximately (Figure 11a). The increase continues up to , beyond which a reduction in is observed. For thin ETLs (), remains nearly constant with ETL doping density up to , followed by a noticeable increase at and a slight degradation at higher doping densities. When ETL thickness is varied, increases with thickness up to approximately for all doping densities. Beyond this thickness, decreases at low ETL doping densities (), whereas it continues to increase for highly doped ETLs ().
Figure 11. Effect of thickness and doping density of the ETL (CeO₂) on , , FF, and PCE of the device. The absorber thickness and acceptor doping density were fixed at 800 nm and cm, respectively, which yielded the maximum PCE in section 3.4.1. All other material parameters were kept fixed at their default values (Tables 1–3).
These trends originate from the evolution of band bending, electron selectivity, and recombination at the interface. At low ETL doping densities, weak band bending and limited electron conductivity reduce the internal electric field and increase interfacial recombination, so increasing ETL thickness beyond an optimal value primarily introduces transport resistance and recombination losses, leading to reduced . As ETL doping density increases, stronger band bending enhances electron selectivity and suppresses carrier backflow into the absorber, allowing quasi-Fermi level splitting to be preserved across thicker ETLs and resulting in a stronger thickness-dependent enhancement of . At excessively high doping densities, increased free-carrier concentration and defect-assisted recombination within the ETL counteract these benefits, producing the observed reduction in .
exhibits a comparatively weak dependence on ETL doping density (Figure 11b). For a fixed ETL thickness, remains nearly independent of ETL doping over most of the investigated range, with only slight reductions observed at high doping densities when the ETL is thick. The maximum variation corresponds to a decrease from approximately 15.84 to 14.5 mA cm at an ETL thickness of 0.4 μm. This limited variation indicates that photocurrent generation is primarily governed by optical absorption in the CsAgBiBr absorber rather than by ETL electronic properties. When ETL thickness is varied, decreases monotonically with increasing thickness, primarily due to enhanced parasitic absorption and increased transport resistance in the CeO layer, which limit carrier extraction.
FF increases monotonically with ETL doping density for all ETL thicknesses studied (Figure 11c). When ETL thickness is varied, FF increases up to an intermediate thickness of approximately 0.4 μm and decreases at larger thicknesses.
These trends arise from the interplay between series resistance, recombination, and field-assisted carrier extraction. Increasing ETL doping density enhances electron conductivity and reduces series resistance, facilitating more efficient carrier extraction and improving FF. Moderate increases in ETL thickness improve junction uniformity and reduce leakage pathways, resulting in initial FF enhancement. Beyond the optimal thickness, however, increased transport resistance and carrier accumulation dominate, leading to FF degradation despite improved ETL conductivity.
For a given ETL thickness, PCE increases monotonically with ETL doping density (Figure 11d). When ETL thickness is varied, PCE exhibits a non-linear dependence, characterized by an initial decrease at low and moderate thicknesses, followed by a partial recovery at larger thicknesses.
These trends result from the balance between voltage enhancement, photocurrent collection, and fill-factor evolution. At small ETL thicknesses, efficiency gains are obtained due to reduced series resistance and enhanced junction selectivity. As ETL thickness increases, losses in associated with parasitic absorption and transport resistance outweigh gains in , leading to PCE degradation. At very large ETL thicknesses and high doping densities, strengthened band bending and improved electron selectivity partially compensate for current losses, resulting in the observed recovery in PCE.
Effect of HTL Thickness and Doping Density
Similarly, the influence of HTL properties on device performance was investigated by independently varying the thickness and acceptor doping density of the P3HT HTL while keeping all other parameters fixed (Figure 12a–Figure 12d). increases monotonically with HTL doping density for all HTL thicknesses studied, with a noticeably stronger rate of increase at larger HTL thicknesses ( µm) (Figure 12a). When HTL thickness is varied, exhibits a non-monotonic dependence: it increases initially up to ~0.1 µm, decreases slightly at ~0.15 µm, rises again up to ~0.3 µm, followed by a reduction at ~0.35 µm and a subsequent recovery at larger thicknesses.
These trends originate from the evolution of band bending, hole selectivity, and interfacial recombination at the CsAgBiBr/P3HT interface. At low HTL doping densities, limited hole conductivity and weak band bending restrict efficient hole extraction, making sensitive to thickness-induced transport losses. Increasing HTL doping strengthens band bending and improves

(a)

(b)

(c)

(d)
Figure 12. Effect of thickness and doping density of the HTL (P3HT) on , , FF, and PCE of the device. The absorber and ETL thicknesses were fixed at 800 nm and 100 nm respectively, while their acceptor and donor doping densities were maintained at cm and cm, respectively, which yielded the maximum PCE in section 3.4.2. All other material parameters were kept fixed at their default values (Tables 1-3).
hole selectivity, suppressing interfacial recombination and allowing quasi-Fermi level splitting to be better preserved, particularly in thicker HTLs. The non-monotonic thickness dependence reflects the competing effects of improved junction formation at moderate thicknesses and increased transport resistance and carrier accumulation in thicker P3HT layers. At sufficiently large thicknesses, improved interface quality and stabilized band alignment partially offset transport losses, leading to recovery.
shows a comparatively weak dependence on HTL doping density (Figure 12b). For HTL thicknesses up to µm, remains nearly independent of doping density, whereas a reduction is observed at higher doping levels ( cm) when the HTL is thick ( µm). When HTL thickness is varied, generally decreases with increasing thickness; however, a noticeable recovery in occurs at larger thicknesses ( µm) when doping density is cm.
This behavior indicates that photocurrent generation is primarily governed by optical absorption in the absorber, while charge extraction is limited by hole transport in the P3HT layer. The initial reduction in with HTL thickness arises from increased transport resistance and carrier accumulation in thicker layers, which enhance recombination losses. At intermediate doping densities, improved hole conductivity and reduced interfacial recombination partially mitigate these losses at larger thicknesses, leading to the observed recovery in .
FF exhibits a moderate dependence on both HTL doping density and thickness (Figure 12c). For thin HTLs ( µm), FF increases slightly with increasing doping density. At higher thicknesses (0.3–0.35 µm), FF decreases beyond a doping density of cm, whereas at the largest thickness studied (0.4 µm), FF shows saturation at high doping densities following an initial increase. When thickness is varied, FF increases up to µm, decreases at µm, and increases again at larger thicknesses.
These trends reflect the balance between series resistance, recombination, and junction uniformity. Moderate increases in HTL thickness improve interfacial coverage and suppress shunt pathways, enhancing FF. However, thicker P3HT layers introduce higher transport resistance and carrier accumulation, leading to FF degradation at intermediate thicknesses. Increasing HTL doping improves hole conductivity and reduces resistive losses, stabilizing FF at large thicknesses and producing saturation behavior at high doping densities.
For HTL thicknesses µm, PCE increases slightly with doping density, whereas at larger thicknesses ( µm), a reduction in PCE is observed at high doping densities ( cm) (Figure 12d). When HTL thickness is varied, PCE exhibits a strongly non-monotonic behavior, characterized by alternating regions of enhancement and degradation.
At small HTL thicknesses, efficiency improvements are driven mainly by the improved hole selectivity and reduced series resistance. As thickness increases, losses associated with transport resistance and recombination dominate, reducing PCE. At specific thickness–doping combinations, partial recovery occurs when improved junction quality and stabilized band alignment compensate for transport limitations.
Effect of Absorber Defect Density
Prior experimental and theoretical studies have identified a diverse population of defect species in , including halide vacancies (), cation vacancies (, ), and antisite defects (e.g., , ) that introduce trap states within the bandgap and act as non-radiative recombination centers [71], [72]. Therefore, the absorber bulk defect density was treated as a key variable in the present study and varied from to cm to explore the impact on device performance (Figure 13). All the photovoltaic parameters exhibit a monotonic degradation with increasing defect density. At relatively low defect concentrations ( cm), the parameters decrease only gradually, indicating that nonradiative recombination remains weak and does not yet dominate the overall carrier dynamics. However, once the defect density reaches cm and above, device performance degrades rapidly across all key metrics.

Figure 13. Effect of bulk defect density in the Cs₂AgBiBr₆ on , , FF, and PCE of the device. The absorber, ETL and HTL thicknesses were fixed at 800 nm, 100 nm and 50 nm, respectively, while their doping densities were maintained at cm, cm and cm, respectively, which yielded the maximum PCE in § 5.2.1. All other material parameters were kept fixed at their default values (Table 1-Table 3).
This transition reflects the onset of defect-assisted SRH recombination as the dominant recombination pathway within the absorber. As the defect density () increases, the carrier lifetimes ( and ) and diffusion lengths ( and ) are significantly reduced (equation (19)-(20), [73]), leading to enhanced nonradiative recombination within the absorber.
where is the carrier capture cross-section, and is the thermal velocity. The resulting reduction in and directly suppresses the quasi-Fermi level splitting, giving rise to a rapid decline in . Simultaneously, intensified bulk recombination reduces the collection probability of photogenerated carriers, causing a measurable decrease in . FF is also adversely affected at high defect densities, as increased recombination near the maximum power point degrades diode ideality and flattens the J–V characteristics. As a cumulative outcome of these losses, the PCE deteriorates sharply beyond the critical defect density threshold.
Overall, these results suggest that maintaining defect concentrations below cm is essential to preserve long carrier lifetimes, efficient charge extraction, and high photovoltaic performance in the device.
Effect of CTL/Absorber Interface Defect Density
In addition to bulk defects, defects at the CeO/CsAgBiBr and P3HT/CsAgBiBr interfaces can significantly influence device performance. These interface defects primarily originate from interfacial disorder, surface roughness, imperfect band alignment, and incomplete chemical passivation during layer deposition. They act as recombination centers and are commonly described within the SRH formalism through an interface recombination velocity, reflecting their critical influence on carrier extraction and voltage losses.
For an interface defect density , the SRH interface recombination rate can be expressed as
where and are the electron and hole interface recombination velocities, respectively, denotes the capture cross section, is the energy level of the interface trap, is the intrinsic Fermi level, is the intrinsic carrier concentration, is the Boltzmann constant and is the temperature. An increase in therefore enhances interface recombination and reduces carrier lifetimes near the interface.
ETL/Absorber Interface Defect
shows a weakly non-monotonic dependence on CeO/CsAgBiBr interface defect density (Figure 14). A slight increase in is observed at very low interface defect densities (–) (Figure 14a). Within the numerical model, this behavior is attributed to minor electrostatic redistribution and changes in local quasi-Fermi level splitting under extremely low interface recombination conditions, rather than to an explicit interface passivation effect. Beyond this narrow regime, at higher defect densities, decreases gradually as enhanced SRH recombination at the interface reduces the quasi-Fermi level splitting.
Figure 14. Effect of defect density at CeO/CsAgBiBr interface on , , FF, and PCE of the device. The absorber, ETL and HTL thicknesses were fixed at 800 nm, 100 nm and 50 nm, respectively, while their doping densities were maintained at , and , respectively, which yielded the maximum PCE in section 3.4.3. All other material parameters were kept fixed at their default values (Tables 1-3).
decreases monotonically with increasing defect density, although the total reduction remains small ( up to ) (Figure 14a). This weak sensitivity indicates that electron extraction through CeO is comparatively resilient to moderate interface recombination. Under short-circuit conditions, carrier collection is dominated by field-assisted transport and favorable conduction-band alignment at
the interface, which promote rapid carrier sweep-out and limit the residence time of electrons near interfacial recombination centers.
FF displays a more complex response (Figure 14b). FF initially decreases with increasing interface defect density up to approximately , reflecting enhanced recombination losses near the maximum power point. Interestingly, a partial recovery in FF is observed at intermediate defect densities, before a renewed decline occurs once the defect density exceeds . This non-monotonic trend suggests a competition between recombination-induced losses and changes in interfacial charge transport, where moderate defect densities may transiently modify local electric fields or carrier selectivity, while high defect densities ultimately dominate through recombination-driven degradation.
As a cumulative consequence of these effects, the PCE decreases continuously with increasing ETL/absorber interface defect density (Figure 14b), confirming that interface recombination at the electron-extraction side progressively undermines device performance.
HTL/Absorber Interface Defect
In contrast, defects at the interface exert a stronger influence on device performance (Figure 15). decreases steadily with increasing defect density and saturates beyond (Figure 15a), consistent with recombination-limited quasi-Fermi level splitting once interface SRH recombination dominates.
Figure 15. Effect of defect density at interface on , , FF, and PCE of the device. The absorber, ETL and HTL thicknesses were fixed at 800 nm, 100 nm and 50 nm, respectively, while their doping densities were maintained at , and , respectively, which yielded the maximum PCE in section 3.4.3. All other material parameters were kept fixed at their default values (Tables 1-3).
also decreases with increasing defect density, with a pronounced drop emerging once the interface defect density reaches (Figure 15a). The reduction in is attributed to enhanced trap-assisted recombination at the interface, where photogenerated holes accumulate under operation, increasing their residence time near interfacial defect states and suppressing efficient carrier extraction.
FF follows a non-monotonic trend, exhibiting an initial decline at low defect densities, a weak recovery at intermediate densities, and a sharp degradation beyond (Figure 15b). At high defect concentrations, enhanced interface recombination significantly distorts the diode ideality and increases recombination losses near the maximum power point.
Consequently, the PCE decreases with increasing HTL/absorber interface defect density (Figure 15b), remaining nearly constant only over a narrow intermediate regime before undergoing a sharp decline at high defect densities.
The results of this section highlight that effective control of interfacial defect densities is a key requirement for maintaining efficient carrier extraction and achieving high-performance -based solar cells.
Effect of Direct and Auger Recombination Coefficient
When the radiative (direct) and Auger recombination coefficients of the absorber are varied simultaneously, a clear threshold-dependent degradation of device performance is observed (Figure 16). For low recombination coefficients, all photovoltaic parameters remain nearly unchanged, indicating that recombination losses are not rate-limiting in this regime. However, once either recombination coefficient increases beyond a critical level, a pronounced deterioration in device performance occurs.
Figure 16. Effect of radiative and Auger recombination coefficients on , , FF, and PCE of the device. All other material parameters were kept fixed at their default values (Tables 1-3), and the Auger coefficients were assumed identical for electrons and holes. The absorber, ETL and HTL thicknesses were fixed at 800 nm, 100 nm and 50 nm, respectively, while their doping densities were maintained at cm, cm and cm, respectively, which yielded the maximum PCE in section 3.4.3. All other material parameters were kept fixed at their default values (Tables 1-3).
At low radiative and Auger recombination coefficients, photogenerated carriers in the absorber possess sufficiently long lifetimes to be efficiently separated and extracted by the built-in electric fields at the / and /P3HT interfaces. Under these conditions, quasi-Fermi level splitting is well preserved, leading to high , while efficient carrier collection sustains both and FF.
As the radiative recombination coefficient increases beyond its threshold, band-to-band recombination losses become significant. Although radiative recombination is an intrinsic process, an excessively high radiative coefficient reduces the effective carrier lifetime and limits the achievable quasi-Fermi level splitting, resulting in a noticeable reduction in (Figure 16a). The accompanying increase in bulk recombination also lowers carrier collection efficiency, contributing to simultaneous declines in and FF (Figure 16b,Figure 16c).
Auger recombination has an even more detrimental impact at higher carrier densities due to its strong dependence on carrier concentration. Once the Auger recombination coefficient exceeds , three-carrier recombination processes dominate, particularly in regions of strong band bending near the junction where carrier densities are highest. This leads to rapid carrier annihilation, severely reducing carrier lifetime and diffusion length. As a consequence, the performance parameters degrade markedly.
The cumulative effect of these recombination losses is clearly reflected in the PCE (Figure 16d). PCE exhibits a pronounced decline once either recombination coefficient exceeds its critical threshold. Based on the observed trends, optimal device performance is achieved when the radiative recombination coefficient remains below and the Auger recombination coefficient remains below .
Effect of Temperature
The temperature dependence of the photovoltaic performance of the device was investigated over a wide temperature range from 200 to 500 K to assess the interplay between thermal activation, charge transport, and recombination (Figure 17). All device metrics exhibit non-monotonic temperature dependence, indicating a transition between transport-limited and recombination-limited operating regimes.
Figure 17. Effect of temperature () on , , FF, and PCE of the device. The absorber, ETL and HTL thicknesses were fixed at 800 nm, 100 nm and 50 nm, respectively, while their doping densities were maintained at cm, cm and cm, respectively, which yielded the maximum PCE in § 5.2.1. All other material parameters were kept fixed at their default values (Table 1-Table 3).
initially increases with temperature up to approximately 300 K, followed by a gradual decrease up to K and saturation at higher temperatures. This behavior is consistent with the temperature dependence of the quasi-Fermi level splitting (equation 16). At low temperatures (200–300 K), enhanced thermal activation of the carriers and reduced interfacial barriers lead to a positive temperature coefficient, . At higher temperatures, the exponential increase of dominates [74],
resulting in a negative temperature coefficient, , as recombination and effective bandgap narrowing reduce the attainable voltage. The eventual saturation of above K indicates a recombination-controlled regime in which further thermal activation no longer alters the quasi-Fermi level splitting.
shows a similarly non-monotonic temperature dependence. increases slightly from 200 K to K due to enhanced carrier mobility and improved charge collection efficiency. With further temperature increase up to K, decreases as carrier–phonon scattering and thermally activated recombination increasingly limit carrier transport. At higher temperatures, becomes nearly temperature independent, suggesting that optical generation and transport losses reach a dynamic balance that is weakly sensitive to further thermal variation.
FF shows an initial improvement with temperature, reaching a maximum near K, followed by a gradual decline at higher temperatures and eventual saturation above K. The initial increase in FF can be attributed to reduced series resistance and improved carrier transport with increasing temperature. At higher temperatures, however, increased recombination and enhanced resistive losses reduce the squareness of the current–voltage characteristics, leading to FF degradation.
As the cumulative metric, PCE closely tracks the combined evolution of , , and FF. PCE increases initially with temperature, reaching a maximum near K, driven by the simultaneous enhancement of , a modest increase in , and improved FF. At higher temperatures, the concurrent reduction in and dominates, resulting in a progressive decrease in PCE up to K, followed by a plateau at higher temperatures. This saturation reflects the establishment of a thermally limited operating regime in which further increases in temperature do not substantially alter the balance between photogeneration, transport, and recombination.
Overall, the temperature-dependent behavior underscores the critical role of thermally activated recombination in limiting device performance at elevated temperatures, while also highlighting an optimal operating window near room temperature where transport, recombination, and interfacial energetics are most favorably balanced in the device.
Effect of Incident Light Intensity
The operational response of the device under varying illumination conditions was evaluated by examining the dependence of the photovoltaic parameters on incident light intensity (Figure 18), an essential consideration given the substantial fluctuations in irradiance between outdoor and indoor environments and under different weather conditions. exhibits a logarithmic increase with increasing light intensity, in agreement with the theoretical relationship described by equation (16). Since scales linearly with illumination, increases logarithmically. Physically, higher photon flux generates a larger population of photocarriers, leading to enhanced quasi-Fermi level splitting within the absorber and a corresponding rise in .
Figure 18. Effect of incident light intensity on , , FF, and PCE of the device. The absorber, ETL and HTL thicknesses were fixed at 800 nm, 100 nm and 50 nm, respectively, while their doping densities were maintained at , and , respectively, which yielded the maximum PCE in § 5.2.1. All other material parameters were kept fixed at their default values (Table 1-Table 3).
increases linearly with incident light intensity , reflecting the direct proportionality between photon flux and the rate of photogenerated electron–hole pairs [75]. This proportionality indicates that photocarrier generation and collection remain efficient over the investigated intensity range, with no evidence of current saturation or transport-induced limitations.
A slight yet systematic increase in the FF is observed with increasing illumination intensity. This trend can be attributed to improved diode rectification and a reduced relative influence of resistive losses at higher photogenerated current densities, which together enhance the curvature of the J–V characteristics near the maximum power point.
Consequently, the PCE increases monotonically with illumination intensity, driven primarily by the linear enhancement of and the logarithmic rise in , with secondary contributions from FF improvement. Notably, the PCE remains substantial under low-intensity illumination and exhibits
stable behavior at higher irradiances, indicating the absence of recombination-induced rollover or resistive degradation across the investigated range.
Optimized Device
Based on the preceding optical and electrical analyses, the device parameters were systematically tuned to identify an optimized device configuration under standard operating conditions (AM1.5G illumination at 1000 W m and 300 K). The optimized FTO/CeO/CsAgBiBr/P3HT/Au device delivers a of 1.96 V, of 18.75 mA cm, FF of 86.44%, and PCE of 31.76%. The corresponding – and – characteristics and external quantum efficiency (EQE) response are presented in Figure 19.

Figure 19. (a) – and (power)– characteristics, and (b) EQE of the optimized FTO/CeO/CsAgBiBr/P3HT/Au device. EQE was calculated using the formula [76]: , where is the electron flux, in the form of electrical current (at short circuit) and is the incident photon flux. In panel (b), the AM1.5G spectrum is shown in the background in light brown, with values corresponding to normalized spectral irradiance. The UV (ultra-violet) region (300–380 nm), visible range (380–750 nm), and IR (infra-red) region (>750 nm) are indicated for reference. Thickness of different layers were set as: CeO = 100 nm, CsAgBiBr = 800 nm, P3HT = 50 nm; doping densities were set as: CeO = cm, CsAgBiBr = cm, P3HT = cm. All other material parameters were kept fixed at their default values, as listed in (Table 1–Table 3).
To place these results in context, the optimized device performance was compared with recently reported CsAgBiBr-based solar cells from both experimental and numerical studies published between 2022 and 2026 (Table 6). Experimentally reported efficiencies span a range from 0.19% to 6.37%, while simulation studies predicted efficiencies between 5.56% and 27.78%. The simulated optimal efficiency obtained in the present work exceeds the values reported in these studies, underscoring the potential of the proposed device architecture and the comprehensive 3D physics-based optimization strategy. It should be emphasized that the reported value represents the best-case simulated performance under idealized conditions.
| Device structure | (V) | (mA/cm) | FF (%) | PCE (%) | Reference | Year |
| Experimental works | ||||||
| FTO///Carbon | 0.69 | 1.71 | 47.9 | 0.54 | [77] | 2023 |
| FTO///Carbon | 1.22 | 3.04 | 62.0 | 2.03 | [78] | 2023 |
| ITO///Spiro-OMeTAD/Au | 0.89 | 11.4 | 55.6 | 6.37 | [28] | 2022 |
| //Spiro-OMeTAD | 0.92 | 11.4 | 60.93 | 6.37 | [28] | 2022 |
| PTB7/Au/FTO/-/-//Spiro-OMeTAD/Ag | 1.03 | 5.14 | 55.4 | 3.07 | [79] | 2022 |
| FTO/-//Carbon | 1.29 | 0.49 | 30.0 | 0.19 | [80] | 2022 |
| FTO/-/-//(PEA)AgBiBr/Spiro-OMeTAD/Au | 1.07 | 3.5 | 66.0 | 2.47 | [81] | 2022 |
| //Spiro-OMeTAD | 1.08 | 3.55 | 64.0 | 2.46 | [81] | 2022 |
| FTO/-/-//Carbon | 1.2 | 2.61 | 71.0 | 2.22 | [82] | 2022 |
| FTO/-/-//MoS/carbon | 1.1 | 6.12 | 72.0 | 4.17 | [83] | 2022 |
| FTO/-/-//poly[[4,8-bis[5-(2-ethylhexyl)-2-thienyl]benzo[1 2-b:4,5-b′] | 1.28 | 3.34 | 77.5 | 3.31 | [84] | 2022 |
| FTO/-/-//Spiro-OMeTAD/Ag | 1.05 | 1.77 | 71.7 | 1.33 | [85] | 2022 |
| FTO/-/-//Spiro-OMeTAD/Ag | 1.23 | 2.38 | 71.0 | 2.09 | [86] | 2022 |
| FTO/-/-/D419-@TiCT/Spiro-OMeTAD/Ag | 0.72 | 8.85 | 70.1 | 4.47 | [87] | 2022 |
| FTO/-//SnS QDs/carbon | 1.02 | 3.74 | 51.0 | 1.95 | [88] | 2022 |
| FTO//CsAgZnBiBr/Carbon | 1.0 | 4.23 | 51.0 | 2.16 | [89] | 2022 |
| ITO///P3HT/MoO/PTAA/Au | 1.06 | 2.8 | 65.0 | 1.94 | [90] | 2022 |
| Simulation works | ||||||
| FTO///CuO/Au | 1.71 | 23.33 | 76.24 | 27.78 | [38] | 2026 |
| FTO///P3HT/Al | 1.79 | 13.26 | 75.41 | 17.56 | [91] | 2025 |
| FTO////CuI/Au | 1.04 | 8.54 | 62.37 | 5.56 | [92] | 2025 |
| FTO///CuO/Au | 1.33 | 8.78 | 64.96 | 7.6 | [92] | 2025 |
| FTO///CuSCN/Au | 1.14 | 8.6 | 63.71 | 6.28 | [92] | 2025 |
| FTO///GQDs/Au | 0.814 | 24.55 | 76.69 | 15.33 | [93] | 2025 |
| FTO///Spiro-OMEOTED/Au | 1.06 | 8.55 | 62.53 | 5.68 | [92] | 2025 |
| FTO/ZnO//CBTS/Au | 1.42 | 19.978 | 89.958 | 25.562 | [36] | 2025 |
| FTO/ZnO//CFTS/Au | 1.13 | 22.08 | 62.93 | 15.65 | [36] | 2025 |
| ITO///CuO/Au | 1.11 | 16.54 | 82.43 | 15.11 | [94] | 2025 |
| FTO/WS/CsAgBiBr/Spiro-OMeTAD/Au | 1.5 | 18.86 | 85.02 | 24.16 | [95] | 2025 |
Table 6. Comparison of photovoltaic performance of the proposed absorber–based solar cell with recent experimental and simulation studies (2022–2026).
| Device structure | (V) | (mA/cm) | FF (%) | PCE (%) | Reference | Year |
| (gradient)/C | 1.69 | 12.11 | 88.79 | 18.21 | [96] | 2025 |
| 1.375 | 21.38 | 79.93 | 23.5 | [48] | 2024 | |
| 1.372 | 21.41 | 78.82 | 23.15 | [48] | 2024 | |
| 1.377 | 21.42 | 79.09 | 23.34 | [97] | 2024 | |
| 1.373 | 21.4 | 79.72 | 23.42 | [48] | 2024 | |
| 1.376 | 21.41 | 78.87 | 23.23 | [48] | 2024 | |
| 1.04 | 11.79 | 66.41 | 8.15 | [98] | 2024 | |
| 1.37 | 21.38 | 79.93 | 23.5 | [48] | 2024 | |
| 1.5 | 21.35 | 83.87 | 26.89 | [35] | 2024 | |
| 1.787 | 8.648 | 36.38 | 5.62 | [99] | 2024 | |
| -doped C | 1.4 | 20.0 | 74.76 | 20.93 | [100] | 2024 |
| 0.84 | 32.28 | 85.77 | 23.49 | [101] | 2023 | |
| 2.49 | 8.17 | 76.16 | 15.49 | [1] | 2023 | |
| 2.49 | 11.49 | 77.03 | 21.92 | [1] | 2023 | |
| 1.3 | 17.44 | 62.59 | 14.29 | [102] | 2023 | |
| 0.86 | 16.56 | 53.84 | 7.66 | [103] | 2023 | |
| 1.29 | 20.69 | 81.72 | 21.88 | [103] | 2023 | |
| 1.642 | 4.913 | 88.74 | 7.16 | [104] | 2023 | |
| 1.278 | 23.3 | 88.21 | 26.3 | [105] | 2023 | |
| 1.527 | 10.42 | 75.66 | 12.08 | [106] | 2023 | |
| -doped | 1.68 | 14.81 | 80.54 | 20.06 | [107] | 2023 |
| 1.39 | 14.8 | 77.47 | 16.07 | [107] | 2023 | |
| 1.66 | 14.81 | 80.42 | 19.86 | [107] | 2023 | |
| 1.45 | 14.8 | 78.43 | 16.84 | [107] | 2023 | |
| 1.24 | 14.79 | 73.64 | 13.51 | [107] | 2023 | |
| 1.44 | 14.8 | 78.0 | 16.72 | [107] | 2023 | |
| 0.9 | 11.09 | 66.61 | 6.68 | [108] | 2022 | |
| 1.33 | 21.46 | 88.53 | 25.38 | [109] | 2022 | |
| 1.29 | 20.69 | 81.72 | 21.88 | [103] | 2022 | |
| 1.96 | 18.75 | 86.44 | 31.76 | This work | 2026 |
Table 7.
Conclusion
In this work, a physics-based three-dimensional numerical framework, grounded in Maxwell’s electrodynamics and semiconductor drift-diffusion theory, was developed to investigate the performance-limiting factors and optimization pathways for -based lead-free double perovskite solar cells. A systematic screening of 25 ETL–HTL combinations, constructed from five representative electron transport layers (, AZO, , PCBM, and ) and five hole transport layers (P3HT, CuO, NiO, PEDOT:PSS, and ), identified the /P3HT pairing as the most promising architecture, delivering the highest power conversion efficiency among all evaluated configurations. Building on the FTO///P3HT/Au stack, comprehensive optical and electrical analyses were conducted, covering optimization of layer thicknesses and doping densities, together with quantitative assessment of bulk and interfacial defect states, to resolve the coupled photogeneration, carrier transport, and recombination dynamics governing device operation. Optical analyses revealed strongly front-loaded photogeneration, with peak absorption localized near the ETL/absorber interface and generation rates on the order of across the 400–800 nm spectral range. Importantly, generation profiles were calculated for each absorber, ETL, and HTL thickness variation, enabling realistic capture of optical redistribution and interference effects that are commonly neglected in typical device modeling. Systematic variation of absorber thickness and doping density demonstrated that and FF dominates efficiency gains at higher doping levels (), whereas photocurrent improvements associated with increased thickness are progressively offset by transport and recombination losses at excessive doping or thickness. Engineering of the layer revealed an optimal doping window () that enhances electron selectivity and device efficiency, while excessive ETL thickness () introduces parasitic absorption and transport resistance, reducing and PCE. Optimization of the P3HT layer exposed a pronounced non-monotonic dependencies of , FF, and PCE on both thickness and acceptor density, with higher efficiencies achieved at higher doping levels () and reduced thicknesses ( nm). Defect physics emerged as a dominant determinant of achievable performance. Bulk defect densities in exceeding drive a transition toward recombination-limited operation, sharply degrading photovoltaic metrics through reduced carrier lifetime and diffusion length. Interface defects exhibit asymmetric impact: the / interface shows relative tolerance to moderate trap densities (), whereas the P3HT/ interface demonstrates pronounced sensitivity, with continuous PCE degradation across the investigated defect range. Analysis of intrinsic recombination pathways revealed distinct threshold behavior. Device performance remains stable for radiative recombination coefficients below and Auger coefficients below , beyond which rapid efficiency loss occurs due to severe carrier lifetime shortening and suppressed carrier collection. Thermal analyses indicate that device performance peaks near room temperature ( K), decreases afterwards and gradually saturates at higher temperatures, reflecting a dynamic balance between thermally induced carrier generation and recombination processes. Under varying illumination, the device exhibited linear scaling (187.52 mA/W) and logarithmic enhancement. PCE remains high across the examined range, showing no collapse at higher irradiances due to recombination or resistive effects, while maintaining substantial efficiency even under low-intensity conditions. By integrating optimized material, geometrical, and recombination parameters within physically realistic bounds, a peak simulated performance of V, , FF = 86.44%, and PCE = 31.76% was achieved. These values represent theoretical performance projections under the current physically consistent modeling framework; practical devices may experience additional loss mechanisms—such as microstructural inhomogeneity, ion migration, and environmental degradation—that could moderate experimentally achievable efficiencies. Overall, this study establishes a physically grounded device-engineering framework for solar cells, demonstrating how tightly coupled optical–electrical modeling can convert material-level characteristics into quantitatively guided device optimization strategies. The framework established here is broadly applicable to emerging lead-free perovskite systems and offers a scalable pathway toward experimentally realizable, high-efficiency photovoltaic technologies.
References
- [1]S. Beriha, A. Dikhit, and S. K. Tripathy, “Improving the efficiency of a FTO/PCBM/Cs2AgBiBr6/NiOx/Au lead-free double perovskite solar cell using numerical simulation through optimizing the absorption layer thickness and work function of electrodes,” Results Opt., vol. 12, p. 100467, Jul. 2023, doi: 10.1016/j.rio.2023.100467.DOI
- [2]B. Roose et al., “Local manufacturing of perovskite solar cells, a game-changer for low- and lower-middle income countries?,” Energy Environ. Sci., vol. 15, no. 9, pp. 3571–3582, 2022, doi: 10.1039/D2EE01343F.DOI
- [3]N. Kannan and D. Vakeesan, “Solar energy for future world: - A review,” Renew. Sustain. Energy Rev., vol. 62, pp. 1092–1105, Sep. 2016, doi: 10.1016/j.rser.2016.05.022.DOI
- [4]K. T. Arockiadoss, A.-D. Rasu Chettiar, M. F. Rahman, M. K. Hossain, and L. Marasamy, “Architecture guidelines for Cu2SrSnS4 solar cells using chalcogenide and oxide hole transport layers by SCAPS-1D simulation,” J. Phys. Chem. Solids, vol. 203, p. 112732, Aug. 2025, doi: 10.1016/j.jpcs.2025.112732.DOI
- [5]M. Aliaghayee, “Optimization of the Perovskite Solar Cell Design with Layer Thickness Engineering for Improving the Photovoltaic Response Using SCAPS-1D,” J. Electron. Mater., vol. 52, no. 4, pp. 2475–2491, Apr. 2023, doi: 10.1007/s11664-022-10203-x.DOI
- [6]P. Kung et al., “A Review of Inorganic Hole Transport Materials for Perovskite Solar Cells,” Adv. Mater. Interfaces, vol. 5, no. 22, Nov. 2018, doi: 10.1002/admi.201800882.DOI
- [7]A. Kojima, K. Teshima, Y. Shirai, and T. Miyasaka, “Organometal Halide Perovskites as Visible-Light Sensitizers for Photovoltaic Cells,” J. Am. Chem. Soc., vol. 131, no. 17, pp. 6050–6051, May 2009, doi: 10.1021/ja809598r.DOI
- [8]“Best Research-Cell Efficiencies, National Renewable Energy Laboratory [Online].” [Online]. Available: https://www.nrel.gov/pv/cell-efficiency.html
- [9]T. A. Chowdhury, M. A. Bin Zafar, M. Sajjad-Ul Islam, M. Shahinuzzaman, M. A. Islam, and M. U. Khandaker, “Stability of perovskite solar cells: issues and prospects,” RSC Adv., vol. 10, no. 3, pp. 1787–1810, 2023, doi: 10.1039/D2RA05903G.DOI
- [10]S. Ullah, T. Alshahrani, F. Khan, and F. Rasheed J., “Investigating the potential of lead-free Cs2AgBiI6 and Cs2AgBiBr6 double perovskites for photovoltaic applications,” Mater. Today Commun., vol. 38, p. 108514, Mar. 2024, doi: 10.1016/j.mtcomm.2024.108514.DOI
- [11]A. Al-Ahmed, M. Afzaal, F. Khan, and M. P. U. Haris, “Double-Perovskite Materials: Possibilities and Reality for a Better Solar Cell Device,” IEEE J. Photovoltaics, vol. 15, no. 3, pp. 380–392, 2025, doi: 10.1109/JPHOTOV.2025.3551499.DOI
- [12]A. A. Hassan et al., “Investigating the physical characteristics and photovoltaic performance of inorganic Ba3NCl3 perovskite utilizing DFT and SCAPS-1D simulations,” Mater. Sci. Eng. B, vol. 308, p. 117559, Oct. 2024, doi: 10.1016/j.mseb.2024.117559.DOI
- [13]M. Pitaro, E. K. Tekelenburg, S. Shao, and M. A. Loi, “Tin Halide Perovskites: From Fundamental Properties to Solar Cells,” Adv. Mater., vol. 34, no. 1, Jan. 2022, doi: 10.1002/adma.202105844.DOI
- [14]M. Liu et al., “B-Site Co-Alloying with Germanium Improves the Efficiency and Stability of All-Inorganic Tin-Based Perovskite Nanocrystal Solar Cells,” Angew. Chemie Int. Ed., vol. 59, no. 49, pp. 22117–22125, Dec. 2020, doi: 10.1002/anie.202008724.DOI
- [15]S. W. Park, J. H. Heo, H. J. Lee, H. Kim, S. H. Im, and K.-H. Hong, “Compositional Design for High-Efficiency All-Inorganic Tin Halide Perovskite Solar Cells,” ACS Energy Lett., vol. 8, no. 12, pp. 5061–5069, Dec. 2023, doi: 10.1021/acsenergylett.3c02032.DOI
- [16]W. Ke and M. G. Kanatzidis, “Prospects for low-toxicity lead-free perovskite solar cells,” Nat. Commun., vol. 10, no. 1, p. 965, Feb. 2019, doi: 10.1038/s41467-019-08918-3.
- [17]I. López-Fernández et al., “Lead-Free Halide Perovskite Materials and Optoelectronic Devices: Progress and Prospective,” Adv. Funct. Mater., vol. 34, no. 6, Feb. 2024, doi: 10.1002/adfm.202307896.
- [18]Y. Gao, Y. Pan, F. Zhou, G. Niu, and C. Yan, “Lead-free halide perovskites: a review of the structure–property relationship and applications in light emitting devices and radiation detectors,” J. Mater. Chem. A, vol. 9, no. 20, pp. 11931–11943, 2021, doi: 10.1039/D1TA01737C.DOI
- [19]S. H. Cho et al., “Fabrication Strategies for 2D Halide Perovskite Towards Next-Generation Optoelectronic Applications,” Int. J. Precis. Eng. Manuf. Technol., vol. 12, no. 1, pp. 349–380, Jan. 2025, doi: 10.1007/s40684-024-00663-3.DOI
- [20]W. Yu et al., “Breaking the bottleneck of lead-free perovskite solar cells through dimensionality modulation,” Chem. Soc. Rev., vol. 53, no. 4, pp. 1769–1788, 2024, doi: 10.1039/D3CS00728F.DOI
- [21]T. Song, Q.-X. Ma, Q. Wang, and H.-L. Zhang, “Design of two-dimensional halide perovskite composites for optoelectronic applications and beyond,” Mater. Adv., vol. 3, no. 2, pp. 756–778, 2022, doi: 10.1039/D1MA00944C.
- [22]J. Huang, H. Xiang, R. Ran, W. Zhou, W. Wang, and Z. Shao, “Fundamental understanding in the performance-limiting factors of Cs2AgBiBr6-based perovskite photovoltaics,” Renew. Sustain. Energy Rev., vol. 191, p. 114187, Mar. 2024, doi: 10.1016/j.rser.2023.114187.DOI
- [23]T. Zhang, Z. Cai, and S. Chen, “Chemical Trends in the Thermodynamic Stability and Band Gaps of 980 Halide Double Perovskites: A High-Throughput First-Principles Study,” ACS Appl. Mater. Interfaces, vol. 12, no. 18, pp. 20680–20690, May 2020, doi: 10.1021/acsami.0c03622.DOI
- [24]P. Li, W. Gao, C. Ran, H. Dong, X. Hou, and Z. Wu, “Post-Treatment Engineering of Vacuum-Deposited Cs 2 NaBiI 6 Double Perovskite Film for Enhanced Photovoltaic Performance,” Phys. status solidi, vol. 216, no. 23, Dec. 2019, doi: 10.1002/pssa.201900567.DOI
- [25]L. Zhang et al., “Tuning Emission and Electron–Phonon Coupling in Lead-Free Halide Double Perovskite Cs 2 AgBiCl 6 under Pressure,” ACS Energy Lett., vol. 4, no. 12, pp. 2975–2982, Dec. 2019, doi: 10.1021/acsenergylett.9b02155.DOI
- [26]V. Murgulov et al., “Double perovskite Cs2AgBiBr6 radiation sensor: synthesis and characterization of single crystals,” J. Mater. Sci., vol. 57, no. 4, pp. 2758–2774, Jan. 2022, doi: 10.1007/s10853-021-06847-5.DOI
- [27]C. N. Savory, A. Walsh, and D. O. Scanlon, “Can Pb-Free Halide Double Perovskites Support High-Efficiency Solar Cells?,” ACS Energy Lett., vol. 1, no. 5, pp. 949–955, Nov. 2016, doi: 10.1021/acsenergylett.6b00471.DOI
- [28]Z. Zhang et al., “Hydrogenated Cs2AgBiBr6 for significantly improved efficiency of lead-free inorganic double perovskite solar cell,” Nat. Commun., vol. 13, no. 1, p. 3397, Jun. 2022, doi: 10.1038/s41467-022-31016-w.DOI
- [29]A. H. Slavney, T. Hu, A. M. Lindenberg, and H. I. Karunadasa, “A Bismuth-Halide Double Perovskite with Long Carrier Recombination Lifetime for Photovoltaic Applications,” J. Am. Chem. Soc., vol. 138, no. 7, pp. 2138–2141, 2016, doi: 10.1021/jacs.5b13294.DOI
- [30]G. Volonakis et al., “Lead-Free Halide Double Perovskites via Heterovalent Substitution of Noble Metals,” J. Phys. Chem. Lett., vol. 7, no. 7, pp. 1254–1259, Apr. 2016, doi: 10.1021/acs.jpclett.6b00376.DOI
- [31]M. Jakir Hossen et al., “Recent progress on the efficiency and stability of lead-free Cs2AgBiBr6 double halide perovskite solar cells,” Phys. Scr., vol. 100, no. 1, 2025, doi: 10.1088/1402-4896/ad9b59.DOI
- [32]E. Greul, M. L. Petrus, A. Binek, P. Docampo, and T. Bein, “Highly stable, phase pure Cs 2 AgBiBr 6 double perovskite thin films for optoelectronic applications,” J. Mater. Chem. A, vol. 5, no. 37, pp. 19972–19981, 2017, doi: 10.1039/C7TA06816F.DOI
- [33]Y. Zou et al., “Surface Reconstruction for Efficient Cs 2 AgBiBr 6 Solar Cells,” J. Phys. Chem. C, vol. 127, no. 37, pp. 18301–18309, Sep. 2023, doi: 10.1021/acs.jpcc.3c04362.
- [34]P. Fan et al., “Single-Source Vapor-Deposited Cs2AgBiBr6 Thin Films for Lead-Free Perovskite Solar Cells,” Nanomaterials, vol. 9, no. 12, p. 1760, Dec. 2019, doi: 10.3390/nano9121760.DOI
- [35]S. Srivastava, R. Srivastava, D. Punetha, A. K. Sharma, and Y. K. Prajapati, “Enhancing the efficiency of lead- free Cs2AgBiBr6 based double perovskite solar cells with optimizing ETLs and HTLs using SCAPS-1D,” Chem. Phys. Lett., vol. 854, p. 141557, Nov. 2024, doi: 10.1016/j.cplett.2024.141557.
- [36]E. Danladi, E. E. Oguzie, and F. I. Ezema, “Highly efficient 25.562% Cs2AgBiBr6 double perovskite solar cell with copper barium tin sulfide and ZnO as charge transport channels: an intuition from a theoretical study using SCAPS-1D,” J. Photonics Energy, vol. 15, no. 02, Mar. 2025, doi: 10.1117/1.JPE.15.025401.DOI
- [37]J. Hakami, “A comprehensive study to optimize organic/inorganic ETL and HTL materials on double perovskite layer Cs2AgBiBr6 solar cells with SCAPS 1D simulator,” Results Eng., vol. 25, p. 104044, Mar. 2025, doi: 10.1016/j.rineng.2025.104044.DOI
- [38]M. Raj, A. Makhija, A. Kushwaha, and N. Goel, “Integrated First-Principles, SCAPS-1D, and ML Framework for Engineering CsAgBiBr₆ based Perovskite Solar Cell for IoT-Integrated Applications,” Mater. Sci. Eng. B, vol. 324, p. 119014, Feb. 2026, doi: 10.1016/j.mseb.2025.119014.DOI
- [39]A. Kowsar et al., “An overview of solar cell simulation tools,” Sol. Energy Adv., vol. 5, p. 100077, 2025, doi: 10.1016/j.seja.2024.100077.DOI
- [40]“COMSOL Multiphysics Reference Manual (Version 6.0).” www.comsol.com, pp. 1–2026. [Online]. Available: https://doc.comsol.com/6.0/doc/com.comsol.help.comsol/COMSOL_ReferenceManual.pdf
- [41]K. G. Beepat, D. P. Sharma, A. Mahajan, D. Pathak, and V. Kumar, “Simulation of multijunction solar cell interfaces for enhancement of the power conversion efficiency,” Discov. Appl. Sci., vol. 6, no. 6, p. 283, May 2024, doi: 10.1007/s42452-024-05930-1.DOI
- [42]J. Zhao et al., “Simulation of Crystalline Silicon Photovoltaic Cells for Wearable Applications,” IEEE Access, vol. 9, pp. 20868–20877, 2021, doi: 10.1109/ACCESS.2021.3050431.DOI
- [43]J. Kim, S. Wong, G. Kim, Y.-B. Park, J. van Embden, and E. Della Gaspera, “Transparent electrodes based on spray coated fluorine-doped tin oxide with enhanced optical, electrical and mechanical properties,” J. Mater. Chem. C, vol. 8, no. 41, pp. 14531–14539, 2020, doi: 10.1039/D0TC03314F.DOI
- [45]T. Kim, J. Lim, and S. Song, “Recent Progress and Challenges of Electron Transport Layers in Organic–Inorganic Perovskite Solar Cells,” Energies, vol. 13, no. 21, p. 5572, Oct. 2020, doi: 10.3390/en13215572.DOI
- [46]M. Ferdous Rahman, M. Al Jajul Islam, M. Chowdhury, L. Ben Farhat, S. Ezzine, and A. T. M. Saiful Islam, “Efficiency improvement of CsSnI3 based heterojunction solar cells with P3HT HTL: A numerical simulation approach,” Mater. Sci. Eng. B, vol. 307, p. 117524, Sep. 2024, doi: 10.1016/j.mseb.2024.117524.DOI
- [47]R. S. Singh, R. D. Patidar, K. Deshmukh, A. Gautam, and A. Kumar, “Influence of CuO Layer on the Performance of Thin-Film Copper Indium Gallium Selenide Solar Cells: A Numerical Analysis,” J. Electron. Mater., vol. 54, no. 1, pp. 609–619, Jan. 2025, doi: 10.1007/s11664-024-11588-7.DOI
- [48]M. S. Uddin et al., “An In-Depth Investigation of the Combined Optoelectronic and Photovoltaic Properties of Lead-Free Cs 2 AgBiBr 6 Double Perovskite Solar Cells Using DFT and SCAPS-1D Frameworks,” Adv. Electron. Mater., vol. 10, no. 5, May 2024, doi: 10.1002/aelm.202300751.DOI
- [49]P. Zhou et al., “Efficient and stable mixed perovskite solar cells using P3HT as a hole transporting layer,” J. Mater. Chem. C, vol. 6, no. 21, pp. 5733–5737, 2018, doi: 10.1039/C8TC01345D.DOI
- [50]M. K. Hossain et al., “An extensive study on multiple ETL and HTL layers to design and simulation of high-performance lead-free CsSnCl3-based perovskite solar cells,” Sci. Rep., vol. 13, no. 1, p. 2521, Feb. 2023, doi: 10.1038/s41598-023-28506-2.DOI
- [51]M. K. Hossain, M. H. K. Rubel, G. F. I. Toki, I. Alam, M. F. Rahman, and H. Bencherif, “Effect of Various Electron and Hole Transport Layers on the Performance of CsPbI 3 -Based Perovskite Solar Cells: A Numerical Investigation in DFT, SCAPS-1D, and wxAMPS Frameworks,” ACS Omega, vol. 7, no. 47, pp. 43210–43230, Nov. 2022, doi: 10.1021/acsomega.2c05912.DOI
- [52]H. Alipour and A. Ghadimi, “Optimization of lead-free perovskite solar cells in normal-structure with <math altimg="si1.svg"> <mrow> <msub> <mtext>WO</mtext> <mn>3</mn> </msub> </mrow> </math> and water- free PEDOT: PSS composite for hole transport layer by SCAPS-1D simulation,” Opt. Mater. (Amst.), vol. 120, p. 111432, Oct. 2021, doi: 10.1016/j.optmat.2021.111432.
- [53]Zulfanzam Abdul Wahab et al., “Insights of MoO3 HTL in Perovskite Solar Cells from a Simulation Perspective,” Int. J. Nanoelectron. Mater., vol. 17, no. December, pp. 31–38, Dec. 2024, doi: 10.5895/ijneam.v17iDecember.1604.DOI
- [54]H. El-assib et al., “High-performance optimization and analysis of Cs2CuSbCl6-Based lead-free double perovskite solar cells with theoretical efficiency exceeding 27 %,” Renew. Energy, vol. 239, p. 122092, Feb. 2025, doi: 10.1016/j.renene.2024.122092.
- [55]Y. Gan et al., “Numerical Investigation Energy Conversion Performance of Tin-Based Perovskite Solar Cells Using Cell Capacitance Simulator,” Energies, vol. 13, no. 22, p. 5907, Nov. 2020, doi: 10.3390/en13225907.DOI
- [56]Naureen, Sadanand, S. Rai, R. K. Yadav, P. Lohia, and D. K. Dwivedi, “A simulation study of quantum dot solar cells using two distinct ETL of WO3 and WS2,” Opt. Quantum Electron., vol. 55, no. 6, p. 541, Jun. 2023, doi: 10.1007/s11082-023-04809-6.DOI
- [57]S. Ben Ameur et al., “Physical investigations on undoped and Fluorine doped SnO2 nanofilms on flexible substrate along with wettability and photocatalytic activity tests,” Mater. Sci. Semicond. Process., vol. 61, pp. 17–26, Apr. 2017, doi: 10.1016/j.mssp.2016.12.019.
- [58]H. J. Mohamad, Y. M. Abdul-Husaain, and U. A. S. Al-Jarah, “The structural, optical, and morphological properties of NiO-Cu prepared by spray pyrolysis technique,” Microw. Opt. Technol. Lett., vol. 62, no. 11, pp. 3519–3526, Nov. 2020, doi: 10.1002/mop.32500.DOI
- [59]“Refractive Index Database: PEDOT:PSS.” [Online]. Available: https://refractiveindex.info/?shelf=other&book=PEDOT-PSS&page=Chen
- [60]Q. H. Li, D. Zhu, W. Liu, Y. Liu, and X. C. Ma, “Optical properties of Al-doped ZnO thin films by ellipsometry,” Appl. Surf. Sci., vol. 254, no. 10, pp. 2922–2926, Mar. 2008, doi: 10.1016/j.apsusc.2007.09.104.DOI
- [61]M. Ullah, A. M. Rana, U. Mehtab, and M. Farooq, “Study of structural, electronic and optical properties of co-doped CeO2 using the density functional theory (DFT) method,” Mater. Sci. Semicond. Process., vol. 130, p. 105800, Aug. 2021, doi: 10.1016/j.mssp.2021.105800.DOI
- [62]F. Gami, I. Guizani, M. A. Sebak, A. A. Abuelwafa, and M. M. Mostafa, “Investigation of structural, optical and electrical properties of PCBM/ZnOEP thin films,” Opt. Mater. (Amst.), vol. 134, p. 113093, Dec. 2022, doi: 10.1016/j.optmat.2022.113093.DOI
- [63]“Refractive Index Database: WS2.” [Online]. Available: https://refractiveindex.info/?shelf=main&book=WS2&page=Jung
- [64]D. Datta and S. Kumar, “Growth and ellipsometric studies on C60 thin films for solar cell applications,” J. Appl. Phys., vol. 106, no. 7, Oct. 2009, doi: 10.1063/1.3239853.DOI
- [65]L. Hrostea, M. Girtan, R. Mallet, and L. Leontie, “Optical and Morphological Properties of P3HT and P3HT:PCBM Thin Films Used in Photovoltaic Applications,” IOP Conf. Ser. Mater. Sci. Eng., vol. 374, p. 012015, Jun. 2018, doi: 10.1088/1757-899X/374/1/012015.DOI
- [66]M. Nesa, M. A. Momin, M. Sharmin, and A. H. Bhuiyan, “Structural, optical and electronic properties of CuO and Zn doped CuO: DFT based First-principles calculations,” Chem. Phys., vol. 528, p. 110536, Jan. 2020, doi: 10.1016/j.chemphys.2019.110536.DOI
- [67]“Refractive Index Database: MoO3.” [Online]. Available: https://refractiveindex.info/?shelf=main&book=MoO3&page=Stelling
- [68]C. Multiphysics, “Introduction To Wave Optics Module.” [Online]. Available: https://doc.comsol.com/6.0/doc/com.comsol.help.woptics/IntroductionToWaveOpticsModule.pdf
- [69]A. A. Zaky, S. Elewa, S. Alyahya, M. Al-Dhaifallah, H. Rezk, and B. Yousif, “Mitigation of Temperature Effects and Performance Enhancement of Perovskite Solar Cells Using Nano-Pyramids Grating,” IEEE Access, vol. 11, no. April, pp. 36399–36408, 2023, doi: 10.1109/ACCESS.2023.3266240.DOI
- [70]M. T. Bin Kashem and S. A. Esha, “Advancing lead-free all-inorganic NaSnCl3 based perovskite solar cell for high efficiency: Computational optimization of charge transport layers and material parameters,” Mater. Today Commun., vol. 47, p. 113192, Jul. 2025, doi: 10.1016/j.mtcomm.2025.113192.
- [71]J. Chen et al., “Improving the performance of lead-free Cs 2 AgBiBr 6 double perovskite solar cells by passivating Br vacancies,” J. Mater. Chem. C, vol. 12, no. 35, pp. 14074–14084, 2024, doi: 10.1039/D4TC02339K.DOI
- [72]D. Liu, C. M. Perez, A. S. Vasenko, and O. V. Prezhdo, “Ag–Bi Charge Redistribution Creates Deep Traps in Defective Cs 2 AgBiBr 6: Machine Learning Analysis of Density Functional Theory,” J. Phys. Chem. Lett., vol. 13, no. 16, pp. 3645–3651, Apr. 2022, doi: 10.1021/acs.jpclett.2c00869.DOI
- [73]A. Alghafis and M. K. S. Bin Rafiq, “Impact of interface defects and doping levels on bismuth-based double perovskite solar cells: a numerical modeling approach,” Opt. Quantum Electron., vol. 57, no. 4, p. 245, Apr. 2025, doi: 10.1007/s11082-025-08158-4.DOI
- [74]P. Singh and N. M. Ravindra, “Temperature dependence of solar cell performance - An analysis,” Sol. Energy Mater. Sol. Cells, vol. 101, no. October, pp. 36–45, 2012, doi: 10.1016/j.solmat.2012.02.019.DOI
- [75]J. H. Kim, K. J. Moon, J. M. Kim, D. Lee, and S. H. Kim, “Effects of various light-intensity and temperature environments on the photovoltaic performance of dye-sensitized solar cells,” Sol. Energy, vol. 113, pp. 251–257, Mar. 2015, doi: 10.1016/j.solener.2015.01.012.DOI
- [76]J. Bisquert, The physics of solar cells: Perovskites, organics, and photovoltaic fundamentals. 2017. doi: 10.1201/b22380.DOI
- [77]S. Liu, H. Zhang, Y. Xing, Q. He, S. Han, and H. Zhu, “Thermodynamically regulated preparation of high performance for carbon-based Cs2AgBiBr6 perovskite solar cells,” Mater. Sci. Semicond. Process., vol. 154, p. 107193, Feb. 2023, doi: 10.1016/j.mssp.2022.107193.DOI
- [78]J. Duan et al., “MACl enhanced electron extraction in all-inorganic Cs 2 AgBiBr 6 perovskite photovoltaics,” Chem. Commun., vol. 59, no. 9, pp. 1173–1176, 2023, doi: 10.1039/D2CC06432D.DOI
- [79]X. Yang et al., “Thiourea with sulfur-donor as an effective additive for enhanced performance of lead-free double perovskite photovoltaic cells,” J. Colloid Interface Sci., vol. 628, pp. 476–485, Dec. 2022, doi: 10.1016/j.jcis.2022.07.165.DOI
- [80]F. Schmitz et al., “High Open-Circuit Voltage Cs 2 AgBiBr 6 Carbon-Based Perovskite Solar Cells via Green Processing of Ultrasonic Spray-Coated Carbon Electrodes from Waste Tire Sources,” ChemSusChem, vol. 15, no. 22, Nov. 2022, doi: 10.1002/cssc.202201590.DOI
- [81]M. T. Sirtl et al., “2D/3D Hybrid Cs 2 AgBiBr 6 Double Perovskite Solar Cells: Improved Energy Level Alignment for Higher Contact-Selectivity and Large Open Circuit Voltage,” Adv. Energy Mater., vol. 12, no. 7, Feb. 2022, doi: 10.1002/aenm.202103215.DOI
- [82]J. Li et al., “Pinning Bromide Ion with Ionic Liquid in Lead-Free Cs 2 AgBiBr 6 Double Perovskite Solar Cells,” Adv. Funct. Mater., vol. 32, no. 25, Jun. 2022, doi: 10.1002/adfm.202112991.DOI
- [83]B. Pang et al., “Improved charge extraction and atmospheric stability of all-inorganic Cs2AgBiBr6 perovskite solar cells by MoS2 nanoflakes,” Sol. Energy Mater. Sol. Cells, vol. 246, p. 111932, Oct. 2022, doi: 10.1016/j.solmat.2022.111932.DOI
- [84]B. Li et al., “Efficient and stable Cs2AgBiBr6 double perovskite solar cells through in-situ surface modulation,” Chem. Eng. J., vol. 446, p. 137144, Oct. 2022, doi: 10.1016/j.cej.2022.137144.DOI
- [85]D. Zhao et al., “Overcoming the Limitation of Cs 2 AgBiBr 6 Double Perovskite Solar Cells Through Using Mesoporous TiO 2 Electron Extraction Layer,” ENERGY Environ. Mater., vol. 5, no. 4, pp. 1317–1322, Oct. 2022, doi: 10.1002/eem2.12249.DOI
- [86]A. Yang et al., “VOC over 1.2 V for Cs2AgBiBr6 solar cells based on formamidinium acetate additive,” J. Mater. Sci. Mater. Electron., vol. 33, no. 23, pp. 18758–18767, Aug. 2022, doi: 10.1007/s10854-022-08724-6.DOI
- [87]L. Yang et al., “Performance improvement of dye-sensitized double perovskite solar cells by adding Ti3C2T MXene,” Chem. Eng. J., vol. 446, p. 136963, Oct. 2022, doi: 10.1016/j.cej.2022.136963.DOI
- [88]Y. Ou et al., “Efficiency improvement of Cs2AgBiBr6 perovskite solar cells with modification of SnS quantum dots,” Mater. Lett., vol. 312, p. 131672, Apr. 2022, doi: 10.1016/j.matlet.2022.131672.DOI
- [89]Y. Ou et al., “Boosting the stability and efficiency of Cs2AgBiBr6 perovskite solar cells via Zn doping,” Opt. Mater. (Amst.), vol. 129, p. 112452, Jul. 2022, doi: 10.1016/j.optmat.2022.112452.DOI
- [90]Z. Zhang et al., “Efficient Nonleod Double Perovskite Solar Cell with Multiple Hole Transport Layers,” ACS Appl. Energy Mater., vol. 3, no. 10, pp. 9594–9599, Oct. 2020, doi: 10.1021/acsaem.0c01066.DOI
- [91]M. Z. Arif et al., “Comprehensive Analysis of Cesium Silver Bismuth Bromide Perovskite Solar Cells: Enhancing Performance through Optimization and Modeling,” Phys. status solidi, vol. 222, no. 16, Aug. 2025, doi: 10.1002/pssa.202500159.DOI
- [92]A. Raj, M. Kumar, D. V. Singh, B. Singh, D. K. Dwivedi, and A. Anshul, “Physical parameter optimization and band alignment approach for efficiency improvement in Cs2AgBiBr6 based lead-free perovskite solar cells,” Sci. Rep., vol. 15, no. 1, p. 32868, Sep. 2025, doi: 10.1038/s41598-025-02203-8.DOI
- [93]M. A. Saifee, M. Ali, F. F. A. Khan, A. K. Srivastava, J. Ali, and M. S. Khan, “Design and performance optimization of a lead-free Cs2AgBiBr6 perovskite solar cell with graphene quantum dot hole transport layer using SCAPS-1D and machine learning,” J. Opt., Feb. 2025, doi: 10.1007/s12596-025-02539-4.DOI
- [94]M. Mehrabian, M. Taleb-Abbasi, and O. Akhavan, “Comparing the performances of Cs2TiBr6, Cs2AgBiBr6, and Cs2PtI6 halide compositions in double perovskite photovoltaic devices,” Mater. Renew. Sustain. Energy, vol. 14, no. 2, p. 38, Aug. 2025, doi: 10.1007/s40243-025-00311-z.DOI
- [95]M. A. Hachimi et al., “Combined experimental and numerical approach for the optimization of double Cs2AgBiBr6 perovskite solar cells: Ab initio study, simulation and performance comparison,” Solid State Sci., vol. 169, p. 108101, Nov. 2025, doi: 10.1016/j.solidstatesciences.2025.108101.DOI
- [96]W. Yang, W. Li, Q. Liu, and Y. Jin, “Design and simulation of gradient-structured Cs2AgBiBr6 carbon-based double perovskite solar cell for boosting photovoltaic performance,” Sol. Energy, vol. 290, p. 113347, Apr. 2025, doi: 10.1016/j.solener.2025.113347.DOI
- [97]S. Rawat, J. Madan, and A. Verma, “Investigating the Efficiency of Photovoltaic Cells Using Cs 2 AgBiBr 6 Perovskite: Analyzing the Impact of Absorber Layer Thickness,” in 2024 Second International Conference on Microwave, Antenna and Communication (MAC), IEEE, Oct. 2024, pp. 1–4. doi: 10.1109/MAC61551.2024.10837548.DOI
- [98]M. Z. Arif, M. Rahman, N. K. Chowdhury, S. Islam, M. Z. Alam, and M. S. Reza, “Numerical investigation and design optimization for enhanced efficiency of Cs 2 AgBiBr 6 perovskite solar cell,” Phys. Scr., vol. 99, no. 10, p. 105110, Oct. 2024, doi: 10.1088/1402-4896/ad7545.DOI
- [99]S. Das et al., “Electron Transport Layer Material Optimization for Cs2AgBiBr6 Based Solar Cell Using SCAPS,” J. Nano- Electron. Phys., vol. 16, no. 1, pp. 01014-1–01014-4, 2024, doi: 10.21272/jnep.16(1).01014.DOI
- [100]A. S. M. Mosabbir et al., “Optimizing lead-free Cs 2 AgBiBr 6 double perovskite solar cells: insights from SCAPS and FDTD simulations,” Sustain. Energy Fuels, vol. 8, no. 18, pp. 4311–4323, 2024, doi: 10.1039/D4SE00958D.DOI
- [101]N. Jaiswal, D. Kumari, R. Shukla, and S. K. Pandey, “Design and Performance Optimization of Eco-friendly Cs2AgBiBr6 Double Perovskite Solar Cell,” J. Electron. Mater., vol. 52, no. 12, pp. 7842–7849, Dec. 2023, doi: 10.1007/s11664-023-10705-2.DOI
- [102]T. I. Alanazi, “Design and Device Numerical Analysis of Lead-Free Cs2AgBiBr6 Double Perovskite Solar Cell,” Crystals, vol. 13, no. 2, p. 267, Feb. 2023, doi: 10.3390/cryst13020267.DOI
- [103]H. I. Alkham mash, M. Mottakin, M. M. Hossen, M. Akhtaruzzaman, and M. J. Rashid, “Design and defect study of Cs 2 AgBiBr 6 double perovskite solar cell using suitable charge transport layers,” Semicond. Sci. Technol., vol. 38, no. 1, p. 015005, Jan. 2023, doi: 10.1088/1361-6641/aca42b.DOI
- [104]I. Chabri, Y. Benhouria, A. Oubelkacem, A. Kaiba, I. Essaoudi, and A. Ainane, “Cs2AgBiBr6-based perovskite solar cell: A novel combination of ITO/CdS/ Cs2AgBiBr6/ CuAlO2/Pt, with inorganic charge transport layers,” Optik (Stuttg.), vol. 274, p. 170560, Mar. 2023, doi: 10.1016/j.ijleo.2023.170560.DOI
- [105]H. Sabbah, Z. Abdel Baki, R. Mezher, and J. Arayro, “SCAPS-1D Modeling of Hydrogenated Lead-Free Cs2AgBiBr6 Double Perovskite Solar Cells with a Remarkable Efficiency of 26.3%,” Nanomaterials, vol. 14, no. 1, p. 48, Dec. 2023, doi: 10.3390/nano14010048.DOI
- [106]N. Shrivastav et al., “Optimizing the performance of Cs2AgBiBr6 based solar cell through modification of electron and hole transport layers,” Mater. Today Commun., vol. 36, p. 106761, Aug. 2023, doi: 10.1016/j.mtcomm.2023.106761.DOI
- [107]P. Kumar and A. Kumar, “High-performance optimization and analysis of Cs2AgBiBr6-based lead-free double perovskite solar cells,” J. Mater. Sci. Mater. Electron., vol. 34, no. 26, p. 1810, Sep. 2023, doi: 10.1007/s10854-023-11225-9.DOI
- [108]A. Raj, M. Kumar, A. Kumar, A. Laref, and A. Anshul, “Investigating the potential of lead-free double perovskite <scp> Cs 2 AgBiBr 6 </scp> material for solar cell applications: A theoretical study,” Int. J. Energy Res., vol. 46, no. 10, pp. 13801–13819, Aug. 2022, doi: 10.1002/er.8099.DOI
- [109]M. R. Iftekhar, M. G. Rabbani, A. Hosen, M. S. Islam, M. S. Mian, and S. R. Al Ahmed, “Simulating the electrical characteristics of a highly efficient Cs 2 AgBiBr 6 -based perovskite solar cell with NiO x hole transport layer,” in 2022 International Conference on Advancement in Electrical and Electronic Engineering (ICAEEE), IEEE, Feb. 2022, pp. 1–4. doi: 10.1109/ICAEEE54957.2022.9836484.DOI











