Multiscale quantum-to-device simulation framework for Ca₃AsBr₃ perovskite solar cells: engineering efficient electron transport layers

Abstract

The development of stable, environmentally benign, and high-performance perovskite solar cells (PSCs) has increasingly focused on innovative inorganic absorber materials. In this study, we conduct a detailed evaluation of the optoelectronic and mechanical properties of Ca₃AsBr₃, a promising non-toxic halide perovskite, using density functional theory (DFT) alongside SCAPS-1D simulations. The DFT results indicate that Ca₃AsBr₃ possesses a direct bandgap of 1.66 eV, along with good mechanical stability and strong optical absorption, making it well-suited for photovoltaic applications. To further investigate device performance, four electron transport layers (ETLs)—WS₂, SnS₂, CdS, and TiO₂ were incorporated into HTL-free FTO/ETL/Ca₃AsBr₃/Au architecture, allowing analysis of energy band alignment, defect tolerance, and overall efficiency. Among these configurations, the WS₂-based device demonstrated superior performance, achieving a power conversion efficiency (PCE) of 20.50%, with an open-circuit voltage (Voc) of 1.165 V, a short-circuit current density (Jsc) of 20.55 mA/cm², and a fill factor (FF) of 85.64%. Further simulation results highlight that an optimal absorber thickness of 1200 nm, along with reduced bulk and interface defect densities (≤ 10¹⁵ cm⁻³ and ≤ 10¹³ cm⁻²), plays a crucial role in minimizing non-radiative recombination losses and improving charge carrier collection. Overall, this work identifies Ca₃AsBr₃ as a viable eco-friendly absorber material and emphasizes the importance of ETL optimization in achieving efficient, stable, and scalable PSC devices.

Introduction

A major transition is underway in the global energy domain, necessitated by the increasing urgency to mitigate climate change, lessen dependence on fossil fuels, and ensure long-term energy sustainability^{1, 2, 3}. In this context, photovoltaic (PV) technologies have emerged as pivotal solutions, offering direct conversion of sunlight into electricity with minimal environmental impact. Among these, perovskite solar cells (PSCs) have received widespread recognition, on account of their rapidly increasing power conversion efficiencies (PCEs), ease of fabrication, and material versatility. While hybrid organic–inorganic lead halide perovskites have achieved efficiencies increasing from 3.9% to over 20%, with predictions suggesting they could reach up to 31%, their widespread adoption is hindered by intrinsic instability, toxicity (primarily due to lead content), and scalability limitations^{4, 5}.

To overcome these challenges, the research community is intensively investigating lead-free, fully inorganic perovskite materials that can combine stability, environmental safety, and competitive efficiency^{6, 7, 8}. One such promising candidate is Ca₃AsBr₃, an emerging inorganic halide perovskite that has not yet been extensively explored in the photovoltaic domain. Ca₃AsBr₃ corresponds to the A₃MX₃ group of inorganic perovskites, where ‘A’ represents alkaline earth metals like Ba, Sr, Ca, and Mg; ‘M’ includes elements such as Sb, As, N, and P; and ‘X’ consists of halogens like I, Br, Cl, and F. Its non-toxic elemental composition, structural robustness, and potentially favorable optoelectronic properties position it as an attractive alternative to conventional lead-based perovskites⁹. Moreover, these perovskites have a tunable band gap from 0.8 to 2.0 eV, making them suitable for high-performance solar cells⁹. Simultaneously, these perovskites structure demonstrates superior stability, higher efficiency, and enhanced structural, mechanical, and optical properties compared to those of other inorganic perovskites within the more common ABX₃ group. However, to validate and exploit its photovoltaic potential, a comprehensive investigation into its fundamental material characteristics and device-level performance is essential^{10, 11, 12}. Effective optical management is crucial for achieving the high efficiencies predicted in simulations under real conditions by reducing reflection losses at device interfaces. Anti-reflection coatings (ARC), such as single-layer MgF₂, improve refractive index matching and enhance light transmission. Multilayer dielectric stacks (e.g., SiO₂/TiO₂) provide broadband anti-reflection through interference effects. Additional strategies like surface texturing, light trapping, and graded refractive index layers further enhance photon absorption, improving light coupling and helping maintain high device performance¹³. Recent advances in SCAPS-based modeling and inorganic absorber engineering have significantly accelerated the development of high-efficiency photovoltaic devices. For instance, DFT and SCAPS-1D guided optimization of GaInP solar cells has demonstrated notable improvements in device performance¹⁴. Similarly, temperature-resistant and highly doped Ag₂BeSnSe₄ kesterite solar cells have shown promising efficiency and stability for sustainable energy harvesting¹⁵. Moreover, the integration of machine learning with SCAPS-1D simulations has enabled efficient optimization of N719/BEHP-co-MEH-PPV bilayer dye-sensitized solar cells, highlighting the growing importance of multi-scale and data-driven approaches in photovoltaic research¹⁶. Recent studies have also explored advanced absorber and device configurations, such as Cs₂SnBr₆/Cs₂TeI₆ bilayer absorbers¹⁷, SWCNT-based carbon solar cells¹⁸, and radiation-tolerant perovskite devices¹⁹, demonstrating enhanced performance through combined DFT, SCAPS, and machine learning approaches. Despite these advancements, many previously reported systems either rely on complex multilayer architectures, expensive materials, or exhibit limited environmental stability, even when higher power conversion efficiencies (PCEs) are achieved. In this context, there remains a strong need for simple, stable, and lead-free absorber materials with efficient device configurations. To address this gap, the present work focuses on Ca₃AsBr₃, a relatively unexplored inorganic perovskite, and investigates its potential as an efficient and stable absorber for next-generation photovoltaic applications. On the other hand, ML further accelerates optimization by identifying critical parameters such as defect density, thickness, and mobility. The approach enables efficient, low-power, high-performance device design, making it well-suited for IoT applications and advancing scalable, lead-free perovskite technologies²⁰.

Following the material-level analysis, we design and simulate solar cell architectures incorporating Ca₃AsBr₃ using the SCAPS-1D (Solar Cell Capacitance Simulator) software within the PSC architecture FTO/ETL(WS₂, SnS₂, TiO₂, and CdS)/Ca₃AsBr₃/Au. The effectiveness of PSCs heavily depends on electron transport layers (ETLs), which enable rapid charge collection and reduce carrier recombination^{21, 22}. WS₂, SnS₂, TiO₂, and CdS were selected as electron transporting layers (ETLs) because of their demonstrated effectiveness in other perovskite solar cells and their suitable band alignment with Ca₃AsBr₃, which facilitates efficient electron transport^{23, 24}. These materials are noted for their strong chemical and thermal stability, which enhances the overall durability of the device^{25, 26}. Each ETL presents distinct characteristics, such as high electron mobility or appropriate band gaps, enabling a thorough examination of their impact on device performance^{24, 27, 28, 29}. Their compatibility with Ca₃AsBr₃ improves electron extraction and reduces recombination losses, making them excellent choices for this study^{28, 30, 31, 32, 33}. Selecting an appropriate ETL is therefore critical to boost both efficiency and long-term stability of emerging PSC technologies³⁴. Additionally, our paper presents a novel structure, specifically an HTL-free design based on Ca₃ASBr₃ (FTO/ETL(WS₂, SnS₂, CdS and TiO₂)/Ca₃AsBr₃/Au). Traditional HTLs often involve expensive materials or complex production techniques, making large-scale production challenging. This new configuration could particularly benefit such large-scale production. Moreover, the susceptibility of certain HTL materials to degradation over time due to environmental exposure poses a risk to the prolonged operational stability and performance of solar cells. Considering the potential to enhance device stability, adopting an HTL-free configuration becomes a viable option. On the other hand, ML further accelerates optimization by identifying critical parameters such as defect density, thickness, and mobility. The approach enables efficient, low-power, high-performance device design, making it well-suited for IoT applications and advancing scalable, lead-free perovskite technologies²⁰.

This study integrates first-principles DFT calculations with device-level SCAPS-1D simulations to systematically evaluate the structural, electronic, optical, and photovoltaic performance of Ca₃AsBr₃-based PSCs. The results highlight the critical role of ETL engineering and identify WS₂ as the most promising electron transport layer. This work not only establishes Ca₃AsBr₃ as a viable lead-free absorber but also clearly differentiates itself from previous studies by combining material design, device optimization, and simplified HTL-free architecture within a unified framework, thereby providing a robust foundation for future experimental and technological development.

Methodology

This study employs a dual approach combining first-principles density functional theory (DFT) with numerical simulations using SCAPS-1D to investigate the performance of a new inorganic halide PSC incorporating Ca₃AsBr₃ as the absorber layer. The photovoltaic device structure under investigation follows the configuration: FTO/ETL (WS₂, SnS₂, CdS, TiO₂)/Ca₃AsBr₃/Au.

First-principles calculations

To explore the fundamental behaviors of Ca₃AsBr₃, DFT-based simulations were performed with the aid of the Quantum ESPRESSO package. The electronic and structural characteristics were computed using the plane-wave pseudopotential (PWPP) formalism³⁵, applying the Perdew–Burke–Ernzerhof (PBE) functional³⁶ within the generalized gradient approximation (GGA). Ultrasoft norm-conserving pseudopotentials were utilized to analyze core-valence interactions^{37, 38}. The kinetic and charge density cutoffs were set to 225 and 30 Rydbergs, respectively to ensure convergence of total energies and forces. Brillouin zone integrations were accomplished by means of an 8 × 8 × 8 Monkhorst–Pack k-point mesh to ensure precise sampling of the electronic structure. Geometry optimization continued until the atomic forces remained under 0.01 eV/Å, with the electronic SCF convergence threshold assigned a value of 10⁻⁶ atomic units. Band configuration and partial density of states (PDOS) analyses were conducted to confirm the direct band gap and assess orbital contributions near the Fermi level.

Device configuration and simulation

Following material characterization, the optoelectronic behavior of the Ca₃AsBr₃-based device was simulated using SCAPS-1D. It solves the coupled Poisson, continuity, and drift-diffusion equations under steady-state conditions, enabling precise modeling of charge carrier dynamics, recombination processes, and current-voltage characteristics³⁹. The layered device architecture incorporates a p-type Ca₃AsBr₃ absorber (band gap: 1.66 eV; thickness: 1.2 μm), overlaid by different n-type ETLs: WS₂, SnS₂, CdS, and TiO₂. The front contact is composed of fluorine-doped tin oxide (FTO) with a thickness of 100 nm and a band gap of ~ 3.60 eV, while gold (Au) serves as the rear electrode.

The ETLs were chosen for their varied electronic and optical characteristics. WS₂ (1.80 eV) offers excellent band alignment and high electron mobility; SnS₂ (2.24 eV) is chemically stable with favorable transparency; CdS (2.40 eV) is widely used but has parasitic absorption concerns; TiO₂ (3.20 eV) is historically dominant but limited by low carrier mobility and high interfacial recombination. Each ETL’s suitability was assessed based on its energy band alignment with the Ca₃AsBr₃ absorber and its role in governing charge extraction processes and recombination mechanisms. Band diagrams were constructed to visualize heterojunction alignments.

Material parameters, such as dielectric constants, carrier mobilities, band gaps, and electron affinities, were sourced from experimental literature and validated theoretical databases (Table 1). The thermal velocities of the carriers were standardized at 10⁷ cm/s across all layers. The simulation also considered interface defect densities, absorber doping concentrations, and variations in layer thickness. Neutral defect states were introduced at interfaces to model Shockley–Read–Hall recombination realistically (Table 2)⁴⁰. Series and shunt resistances were kept at 1 Ω·cm² and 10⁶ Ω·cm², respectively. The high shunt resistance value was adopted to represent near-ideal device conditions with minimal leakage current⁴¹, which is a commonly used assumption in SCAPS-1D simulations of high-performance perovskite solar cells. All computational analyses assumed AM 1.5G solar spectrum conditions (100 mW/cm²) and an ambient temperature of 300 K.

Table 1.

Parameters used for simulating the PSC.

Parameters	FTO42–44	WS2 45	SnS246	TiO245	CdS47	Ca3AsBr3
Thickness (nm)	50	100	100	50	50	1200
Band gap, Eg (eV)	3.6	1.80	2.240	3.20	2.42	1.66
Electron affinity, χ (eV)	4.5	3.95	4.240	4.0	4.30	4.10
Dielectric permittivity (relative), ϵr	10	13.60	10.0	9.0	9.35	4.80
CB effective density of states, NC (1 /cm3)	2.0 × 1018	1.0 × 1018	2.2 × 1018	2.0 × 1018	2.2 × 1018	6.01 × 1018
Electron thermal velocity (cm/s)	1 × 107	1 × 107	1 × 107	1 × 107	1 × 107	1 × 107
VB effective density of states, NV (1 /cm3)	1.8 × 1019	2.4 × 1019	1.8 × 1019	1.8 × 1019	1.8 × 1019	1.76 × 1019
Hole thermal velocity (cm/s)	1 × 107	1 × 107	1 × 107	1 × 107	1 × 107	1 × 107
Electron mobility, µn (cm2 V− 1 s− 1)	100	100	50	20	100	100
Shallow uniform donor density, ND (1 /cm3)	1 × 1018	1 × 1018	1 × 1017	9.0 × 1016	1 × 1017	0
Shallow uniform acceptor density, NA (1 /cm3)	0	0	0	0	0	1 × 1017
Hole mobility, µh (cm2 V− 1 s− 1)	20	100	50	10	25	50

Table 2.

Interface parameters employed in the Ca₃AsBr₃-based PSC simulations.

Parameters	Ca3AsBr3/WS2	Ca3AsBr3/SnS2	Ca3AsBr3/CdS	Ca3AsBr3/TiO2
σe (cm2)	1 × 1019	1 × 1019	1 × 1019	1 × 1019
σh (cm2)	1 × 1019	1 × 1019	1 × 1019	1 × 1019
Total defect density (cm− 1)	1013	1013	1013	1013
Er	0.6	0.6	0.6	0.6

Chemical reaction of Ca₃AsBr₃ perovskite

The formation of the inorganic perovskite compound Ca₃AsBr₃ involves the combination of three elemental components: Calcium (Ca), Arsenic (As), and Bromine (Br). These elements react in specific stoichiometric ratios to yield the ternary halide perovskite structure. The synthesis process typically follows a balanced chemical reaction, which illustrates the precise proportions of reactants required to produce Ca₃AsBr₃. An example of such a reaction is provided below,

1

Results and discussion

Analysis of DFT results

Structural characteristics

Figure 1(a) illustrates the crystallographic configuration of Ca₃AsBr₃ and taht fits to the cubic space group Pm−3 m (No. 221) and contains a unit cell with seven atoms^{48, 49, 50, 51, 52, 53}. In this arrangement, Br atoms are sited at the 3d Wyckoff locations (0.5, 0, 0), As occupies the 1b site (0.5, 0.5, 0.5), and Ca atoms are positioned at the 3c sites (0.5, 0.5, 0). The optimized lattice constant for this configuration is calculated to be 5.98 Å. To assess the structural viability of creating a stable perovskite phase, the Goldschmidt tolerance factor (t) is applied and is expressed as follows^{48, 49},

2

Fig. 1.

(a) Crystal structure of Ca₃AsBr₃, (b) proposed device architecture incorporating ETL, (c) different ETL configurations, and (d) optimized Ca₃AsBr₃-based PSC.

Here, r_A, r_M, and r_X signify the ionic radii of the Ca²⁺ cation, As³⁻ anion, and Br⁻ anion, respectively. Perovskite compounds are typically stable once the t has the value from 0.8 to 1.0, using the same as those in the 0.9 to 1.0 range exhibiting increased stability and structural perfection^{48, 49}. For Ca₃AsBr₃, the tolerance factor is calculated to be 0.949, indicating its stability. Table 3 provides information on the tolerance factor for Ca₃AsBr₃.

Table 3.

The tolerance factor of Ca₃AsBr₃.

Material	r A	r M	r X	Tolerance factor, t	Refs.
Ca3AsBr3	1.12	0.335	1.96	0.949	This work
Sr3PI3	1.26	0.170	2.20	1.000	48
Sr3AsCl3	1.21	0.335	1.81	1.000	49

Phonon dispersion curves alongside the phonon density of states

Realization of phonon behavior is essential for investigating crystalline compounds. It is possible to understand phase transitions, a material’s thermal, structural stability, and the impact of vibrations on charge transference characteristics by using phonon dispersion spectra (PDS)^{54, 55}. Each distinct energy level’s possible energy levels are quantified by the DOS^{46, 56}. Referring to Fig. 2, the finite displacement method (FDM) and density functional perturbation theory (DFPT) are used in this study to calculate phonon characteristics. The phonon DOS is revealed in Fig. 2(a), whereas the phonon dispersion plot is demonstrated in Fig. 2(b). The frequency spectrum contains no negative values near the gamma point, as seen in Fig. 2(a). At the gamma point, a negative frequency generally indicates chemical instability. On the other hand, since the gamma point is devoid of negative frequencies, Ca₃AsBr₃ is dynamically stable.

Fig. 2.

The (a) phonon density of states and (b) the phonon dispersion curve for Ca₃AsBr₃.

Electronic properties of Ca₃AsBr₃

The characteristics of a compound’s electrons, known as its electronic properties are crucial in estimating the material’s optical, magnetic and electrical behaviors. These attributes are of great importance across diverse fields such as condensed matter physics, electronics, and materials science. This discussion covers key terms like band structure, TDOS, and PDOS, which are instrumental in identifying the nature of a substance, whether it is metallic, insulating, or semiconducting^{57, 58}.The investigation focused on the optical characteristics of an inorganic Ca₃AsBr₃ perovskite by examining its electronic band structures (EBS).

Figure 3(a) demonstrates the electronic band structure of Ca₃AsBr₃ along high-symmetry directions (Γ–X–M–R–Γ) in the first Brillouin zone, extending across an energy spanning from − 3 eV to 3 eV. At 0 eV, the dashed line specifies the Fermi level, which is located between the conduction band (CB) and valence band (VB). The Fermi level indicates where the VB and CB intersect, in accordance with the band structures. Figure 3(a) also highlights the valence band maximum (VBM) as well as conduction band minimum (CBM) close to the Γ-point. The computed direct band gaps are obtained to be 2.45 eV and 1.66 eV at the HSE and PBE function, respectively, as indicated in Fig. 3(a). It is noteworthy that the widely recognized drawback of GGA-based methods (such as PBE and LDA) is their tendency to underestimate the bandgap value during evaluation^{57, 59, 60, 61}. The use of the GGA-PBE functional in this study is motivated by its computational efficiency and its well-established reliability in capturing structural and electronic trends in perovskite materials. Although it is well known that GGA-PBE tends to underestimate the bandgap compared to hybrid functionals such as HSE, it provides a consistent framework for comparative analysis. Moreover, hybrid functionals like HSE require significantly higher computational cost, particularly for evaluating optical and mechanical properties, making them less feasible for large-scale simulations. Since the present work primarily focuses on understanding relative electronic behavior and device-level performance through SCAPS-1D simulations, the use of GGA-PBE is considered appropriate and widely adopted in similar studies.

Fig. 3.

(a) Electronic band structure of Ca₃AsBr₃ calculated utilizing both PBE and HSE functionals, and (b) PDOS highlighting the orbital contributions to the electronic structure of the Ca₃AsBr₃ perovskite.

The study’s results clearly indicate that structures based on Ca₃AsBr₃ feature a broader bandgap, along with larger atomic sizes and an increased number of nucleons. A decreased electrical interaction between the atomic nucleus and surrounding electrons in the highest occupied energy level (HOEL) is the cause of this bandgap widening. The space between the CB and VB gets wider as the cohesive energy drops and the electrical force of attraction becomes less strong. The measured energy band gap values at the attained level propose that Ca₃AsBr₃ is a potential material for PSC use because of its larger band gap^{49, 57, 58, 62}. The DOS signifies the count of unique electron states possible at a specified energy level. Figure 3(b) illustrates the calculated TDOS and PDOS for Ca₃AsBr₃. The energy states of PDOS for various Ca₃AsBr₃ configuration are detailed in Fig. 3(b) across the energy spectrum ranging from − 3 eV to 3 eV.

The TDOS for Ca₃AsBr₃ have a Fermi level (EF) of about 27 states/eV in Fig. 3(b). It is important to highlight that the As 2p orbital is the main contributor to the CB, particularly around ~ 0.7 eV (about 13 states per eV), whereas the Br-2p orbital contributes to the VB around ~ − 0.5 and − 3 eV (about 26 as well as 24 states per eV). Due to Ca₃AsBr₃’s significant influence on both bands, there is minimal overlap between the CB as well as VB near the Fermi level. This significant contribution establishes Ca₃AsBr₃ as a semiconductor.

Mechanical performance of Ca₃AsBr₃

The elastic constants of solid-state compounds act as key parameters that interconnect various mechanical characteristics. For the Ca₃AsBr₃ compound, the independent elastic constants C₁₁​, C12, and C44 were estimated by utilizing slight strains to the optimized cubic unit cell using the PBE functional within the DFT framework. From these primary constants, key mechanical indicators, for instance, the bulk modulus (B), Young’s modulus (E), shear modulus (G), Poisson’s ratio (ν), elastic anisotropy index (A), Cauchy pressure (Cʺ), and Pugh’s ratio (B/G)—were computed to assess the compound’s mechanical stability and deformation behavior. These parameters, which are critical for evaluating the material’s potential in device applications, are comprehensively presented in Table 4.

Table 4.

The obtained elastic constants and the mechanical parameters.

Materials	C11	C12	C44	Cʺ	ν	B	Y	A	G	B/G
Ca3AsBr3	78.55	12.07	17.41	-5.34	0.23	34.22	55.62	0.82	22.63	1.51
Ba3SbI355	53.82	7.02	8.67	-1.65	0.235	22.62	32.89	0.37	13.08	1.73
Ba3AsI364	56.31	6.58	8.01	-1.43	0.27	23.16	32.57	0.14	12.87	1.80

These constants must meet the standards of certain disparities, known as the Born criteria, which are employed to verify the mechanical stability of crystalline materials⁶³.

3

Cauchy pressure

4

The B_H, G_H, and Y of Ca₃AsBr₃ have been assessed using the Voigt–Reuss–Hill (VRH) technique.

Bulk modulus⁶⁴:

5

Young’s modulus⁶⁵:

6

Shear modulus⁶⁶:

7

The Born stability requirement is satisfied by the Ca₃AsBr₃ material shown in Table 4 since it has notable mechanical stability. The material is brittle when the Cauchy pressure is negative, and ductile when it is positive⁶⁷. Cauchy pressure values are negative, so Ca₃AsBr₃ materials are brittle. The ν as well as B/G, which are usually set at 0.26 as well as 1.75, respectively, provide an important criterion for identifying the nature of a material’s mechanical response as brittle or ductile. A compound is considered brittle (ductile) when the ν is greater (smaller) than 0.26 and the B/G is more (smaller) than 1.75 ⁵¹. Table 4 provides specifics regarding the mechanical stability of Ca₃AsBr₃. The computed Cauchy pressure for Ca₃AsBr₃ is -5.34 GPa. Furthermore, for Ca₃AsBr₃, the Poisson’s as well as Pugh’s ratios are both less than 0.26 and 1.75, respectively. This results in the material displaying ductility.

Optical properties of Ca₃AsBr₃

Investigating a compounds optical characteristic is essential to comprehend how it interacts with light. A thorough comprehension of the diverse optical features of Ca₃AsBr₃ material is vital for maximizing its utilization in optoelectronic devices. The dielectric function is expressed as ε(ω), with two parts: the real part (signified as ε₁ (ω)) and the imaginary portion (symbolized as ε₂(ω)).

8

The ε₁ (ω) is obtained utilizing the Kramers-Kronig transformation⁶⁸, whereas the ε₂(ω) is calculated utilizing the momentum matrix’s ⁶⁹.

Analyzing the ε₁ (ω) contributes significantly to understanding the compound’s polarization response and dispersion characteristics. For the cubic phase of Ca₃AsBr₃, the static dielectric constant ε₁(0) is calculated to be around 4.80. As exemplified in Fig. 4 (a), the ε₁(ω) of the dielectric function shows a characteristic trend: it initially rises from ε₁(0), reaches several pronounced peaks—most notably around photon energies of 2.2 eV and 6.0 eV—before gradually declining. This behavior reflects the energy-dependent polarizability of the material. These variations are fundamentally linked to interband electronic transitions.

Fig. 4.

The dielectric function (a) real parts, (b) imaginary part, (c) the absorption coefficient, and (d) loss function of Ca₃AsBr₃.

In contrast, the ε₂(ω), which is directly linked to the electronic band configuration, reflects the compound’s light absorption capabilities. Figure 4 (b) reveals a prominent absorption peak in ε₂(ω) at a light energy of almost 1.5 eV, with the highest optical response reaching 4.6 eV. These peaks in the imaginary spectrum stem from transitions of electrons from the VB to the CB, highlighting the strong light-matter interaction within the visible to near-infrared range. The observed absorption characteristics can be ascribed to the relatively narrow bandgap of Ca₃AsBr₃. Furthermore, the shifting of absorption peaks correlates with modifications in the bandgap and lattice parameters, indicating a sensitive dependence of optical properties on structural variations. Notably, beyond a photon energy of 9.9 eV, the ε₂(ω) reaches zero, revealing enhanced optical transparency and minimal absorption at higher energies.

The ability of a compound to capture incoming photon serve a key function in finding its effectiveness for high solar PCE, with the optical absorption coefficient serving as an extremely important parameter in this assessment⁵². Throughout every configuration studied, a strong association emerges between the ε₂(ω) and the shape of the optical absorption spectra, indicating a consistent optical response behavior. For Ca₃AsBr₃, the absorption spectrum exposes a pronounced peak within the visible region, underscoring its strong interaction with photons in this range. As described in Fig. 4(c), the absorption coefficient attains the highest value near 7.8 eV, with significant absorption occurring over a broad spectral range spanning approximately 3.8 to 9 eV. The absorption onset begins at 2.0 eV and gradually diminishes, approaching zero near 10 eV. This wide absorption bandwidth suggests that Ca₃AsBr₃ is capable of efficiently capturing high-energy photons, a behavior attributed to its relatively extended optical bandgap. The relationship between the photon energy (eV) and the loss function, L(ω), is illustrated in Fig. 4(d). This figure depicts Ca₃AsBr₃ as a semiconductor and shows how the material responds energetically to electromagnetic radiation. Less energy is available for electronic transitions at low photon energies, as indicated by the L(ω) being almost zero. As photon energy increases, the semiconductor absorbs incident light, with a sudden increase after 9 eV and a return to zero after 10.4 eV. signifying a rapid decline absorption coefficient in Ca₃AsBr₃. These optical characteristics suggest that Ca₃AsBr₃ is a favorable choice for applications involving optoelectronic and photovoltaic devices. In addition to photovoltaic applications, the favorable optoelectronic properties of Ca₃AsBr₃, including its direct bandgap and strong optical absorption, indicate its potential for multifunctional applications such as photodetectors and hybrid optoelectronic devices. These properties enable efficient light detection and carrier generation, making the material suitable for broadband photodetection and advanced optoelectronic integration.

Analysis of SCAPS-1D simulation results

Band alignment of Ca₃AsBr₃-based solar cells with different ETL

Figure 5 illustrates the band diagrams of Ca₃AsBr₃-based PSCs integrated with different electron transport layers (ETLs). Amongst the four structures, the WS₂-based device demonstrates the most satisfactory band orientation with the Ca₃AsBr₃ active layer, facilitating actual electron collection and lessening the interfacial recombination due to its suitable CB offset (~ 0.14 eV) and high electron mobility. This optimal alignment directly contributes to its superior PCE exceeding 20%. SnS₂, with a moderate band offset (~ 0.28 eV) and good chemical stability, offers relatively efficient charge transfer but with slightly higher recombination losses compared to WS₂. CdS, despite its historical usage, shows a less ideal band offset (~ 0.34 eV), which results in parasitic absorption and increased interfacial resistance, thus limiting the performance. TiO₂, although widely used in conventional PSCs, suffers from a relatively large CB mismatch (~ 0.46 eV) and lower electron mobility, leading to enhanced charge recombination and the lowest overall performance among the ETLs studied. These variations in energy band alignment, carrier mobility, and optical losses fundamentally explain the hierarchy in device efficiency observed across the four ETLs as exhibited in Fig. 5.

Fig. 5.

The energy band configurations with (a) WS₂, (b) SnS₂, (c) CdS, and (d) TiO₂ ETL of the Ca₃AsBr₃-based PSCs.

Effects of active layer thickness and shallow acceptor concentration

The output parameters of a PSC is largely influenced by the active layer thickness and shallow acceptor concentration^{32, 70}. Figure 6 reveals the role of changing thickness of active layer (ranging from 100 nm to 2500 nm) as well as shallow acceptor concentration (spanning from 10¹⁴ to 10²¹ cm^− 3) on the output of our four simulated (FTO/ETL(WS₂, SnS₂, TiO₂, CdS) /Ca₃AsBr₃/Au) PSCs structures. Under conditions where the active layer is highly thin (0.1–2.5 μm) at a shallow acceptor concentration of 10¹⁷ cm^− 3, light absorption is limited, as revealed in Fig. 6(a). Although the Voc remains nearly constant with elevating absorber thickness, the J_SC, FF, and PCE display a substantial enhancement as the thickness of the Ca₃AsBr₃ layer enhances in the PSC design incorporating an ETL layer²⁸. Figure 6(b) shows how changes in a shallow acceptor density affect the situation. The reciprocal correlation between the carrier density and reverse saturation current results in a notable is rised in V_OC when the shallow acceptor density is raised from 10¹⁴ to 10²¹ cm^− 3. Higher carrier concentrations cause the J_SC to drastically drop, which raises the V_OC⁷¹. In the construction of solar cells (SCs) incorporating an electron transport layer (ETL), a higher shallow acceptor density typically results in an enhanced Voc, while concurrently causing a reduction in the Jsc.

Fig. 6.

Role of (a) thickness and (b) shallow acceptor density of Ca₃AsBr₃ active layer on the output parameters of the PSCs with several ETLs.

Concurrently, both the FF and PCE tend to improve. However, this trend does not apply when using TiO₂ as the ETL, where different behavior is observed. Keeping a constant shallow acceptor density of 10¹⁷ cm^− 3 with an abctive layer thickness of 1200 nm, the PSC shows optimized PCE. It should be noted that very high absorber thickness (1200 nm) may introduce practical limitations such as increased series resistance and reduced charge transport efficiency in real devices. For the efficiency of the designated structural device, a doping concentration of 10¹⁷ cm^− 3 is thought to be ideal.

Role of acceptor density along with defect density variation of Ca₃AsBr₃ active layer on performance

The PSC performance markedly drops owing to severe carrier recombination attributed to increased levels of defect density (N_t) and acceptor density (N_A)^{72, 73, 74}. Increased recombination of photogenerated charge carriers significantly lowers the output of the cell^{75, 76, 77}. This work examines the impacts of varying N_t and N_A values in Ca₃AsBr₃ absorber within definite span of 10¹⁰–10¹⁷ cm^− 3 and 10¹⁴–10²⁰ cm^− 3, respectively, affect the output parameters. The objective is to assess PSC output utilizing accurate and dependable data to reflect accurate behavior.

Influence of acceptor and defect densities on V_OC

Figure 7 presents the influence of N_A and N_t on the V_OC, with N_A varying from 10¹⁴ to 10²⁰ cm⁻³ and N_t spanning from 10¹⁰ to 10¹⁷ cm⁻³. The V_OC response varies across the different ETL materials used in the FTO/ETL (WS₂, SnS₂, CdS, TiO₂)/Ca₃AsBr₃/Au solar cell configurations. Specifically, V_OC increases from 0.918 V to a top of 1.35 V for WS₂ (Fig. 7a), from 0.910 V to 1.34 V for SnS₂ (Fig. 7b), decreases from 1.14 V to 0.867 V for TiO₂ (Fig. 7c), and declines from 1.35 V to 0.912 V for CdS (Fig. 7d). Notably, in the WS₂-based ETL structure, the peak V_OC of 1.35 V is achieved when N_A surpasses 10¹⁹ cm⁻³ and N_t remains under 10¹⁵ cm⁻³. However, a significant decline in V_OC to approximately 0.8095 V is observed when N_A falls below 10¹⁸ cm⁻³. Analogous tendencies are evident for SnS₂ and TiO₂ ETLs, as showed in Fig. 7b and c, respectively. In contrast, the CdS-based ETL demonstrates its highest V_OC of 1.14 V when N_A is below 10¹⁸ cm⁻³, with minimal sensitivity to N_t (Fig. 7d). Raised defect densities in the absorber layer intensify SRH recombination, adversely affecting device performance^{76, 77}. Based on these findings, the optimal parameters for the Ca₃AsBr₃ absorber are identified as an N_A of 10¹⁷ cm⁻³ and N_t of 10¹⁵ cm⁻³, yielding a topmost V_OC of 1.165 V in the proposed device configurations.

Fig. 7.

Impact of varying the N_A and N_t of the Ca₃AsBr₃ layer on the PSCs output with (a) WS₂, (b) SnS₂, (c) CdS, and (d) TiO₂ ETL.

Role of acceptor and defect densities on Jsc

Figure 8 depicts the change in J_SC as a function of N_A (10¹⁴–10²⁰ cm⁻³) and N_t (10¹⁰–10¹⁷ cm⁻³) across Ca₃AsBr₃-based solar cells with different ETLs. A general trend is observed where J_SC declines with increasing N_A and N_t, mainly due to intensified non-radiative SRH recombination, which limits the collection of photogenerated carriers. For the WS₂ ETL configuration (Fig. 8a), J_SC peaks at 22.90 mA/cm² under N_A (≤ 10¹⁷ cm^− 3) and N_t (≤ 10¹⁵ cm^− 3), but drops significantly to 14.28 mA/cm², once N_t surpasses 10¹⁵ cm⁻³ owing to increased trap-assisted recombination. A similar pattern is evident in the SnS₂-based device (Fig. 8b), where J_SC decreases from 22.90 to 16.76 mA/cm² under is the same condition, highlighting SnS₂ is slightly improved tolerance to defects compared to WS₂.

Fig. 8.

Simultaneous effect of N_A and N_t in the Ca₃AsBr₃ active layer on the J_SC for different ETL configurations, (a) WS₂, (b) SnS₂, (c) CdS, and (d) TiO₂.

In the TiO₂ ETL structure (Fig. 8c), the J_SC drops more drastically from 22.90 to 9.47 mA/cm² as N_t rises, which can be ascribed to the higher CB offset between TiO₂ and Ca₃AsBr₃ that hinders effective charge extraction and exacerbates recombination losses. For the CdS-based device (Fig. 8d), J_SC is declines from 19.85 to 6.65 mA/cm² under elevated defect densities, again due to intensified SRH recombination within the absorber layer. These observations underscore that excessive doping and increased trap density deteriorate device performance by reducing carrier lifetime and collection efficiency^{78, 79}. Based on these results, the optimal doping and defect parameters for Ca₃AsBr₃ are identified as N_A≈ 10¹⁷ cm⁻³ and N_t≈ 10¹⁵ cm⁻³, yielding maximum photocurrent and overall device output.

Effect of acceptor and defect densities on FF

Figure 9 illustrates the influence of N_A (10¹⁴ − 10²⁰ cm⁻³) and Nₜ (10¹⁰ − 10¹⁷ cm⁻³) in the Ca₃AsBr₃ absorber on the FF. The FF varies notably across different ETLs: from 87.35% to 69.15% for WS₂ (Fig. 9a), 90.60% to 64.50% for SnS₂ (Fig. 9b), 82.30% to 55.90% for TiO₂ (Fig. 9c), and 81.00% to 31.00% for CdS (Fig. 9d). A consistently high FF is observed when N_A surpasses 10¹⁷ cm⁻³ and N_t remains under 10¹⁵ cm⁻³, indicating minimized SRH recombination under these conditions. For instance, the peak FF of 87.35% in the WS₂-based device (Fig. 9a) is attained under optimal doping and low defect conditions, while increased N_t beyond 10¹⁵ cm⁻³ triggers a substantial drop to 69.15%, due to enhanced trap-assisted recombination. A similar pattern is observed in SnS₂-based cells (Fig. 9b), where FF declines from 90.60% to 64.50% with rising defect densities. In TiO₂ ETL systems (Fig. 9c), FF falls from 82.30% to 55.90% as Nₜ surpasses 10¹⁶ cm⁻³, and for CdS (Fig. 9d), a sharp reduction from 81.00% to 31.00% is observed beyond Nₜ = 10¹⁵ cm⁻³. These reductions are primarily ascribed to the accumulation of defect states that enable non-radiative recombination, hence impeding charge carrier extraction and reducing photovoltaic efficiency. Collectively, the analysis identifies Nₐ ≈ 10¹⁷ cm⁻³ and Nₜ ≈ 10¹⁵ cm⁻³ as optimal parameters for maintaining high FF, approaching 86% in Ca₃AsBr₃-based heterojunction devices.

Fig. 9.

Role of varying N_A and N_t in the Ca₃AsBr₃ active layer on the FF across different ETLs, (a) WS₂, (b) SnS₂, (c) CdS, and (d) TiO₂.

Influence of acceptor and defect densities on PCE

Figure 10 presents the impact of N_A, and N_t on the PCE of Ca₃AsBr₃-based PSC. The PCE varied significantly across different ETL configurations: from 22.26% to 13.82% for WS₂ (Fig. 10a), 23.45% to 12.85% for SnS₂ (Fig. 10b), 18.65% to 3.20% for CdS (Fig. 10c), and 18.60% to 3.66% for TiO₂ (Fig. 10d). Notably, PCE initially increases with rising N_A owing to enhanced built-in electric field and enhanced carrier separation, but beyond an optimal threshold, further increases in N_A or N_t induce severe non-radiative recombination losses, primarily via SRH and Auger processes⁷⁵, thereby degrading overall device efficiency. The highest PCE of 20.50% was obtained in the WS₂-based ETL structure, corresponding to a V_OC of 1.165 V, J_SC of 20.55 mA/cm², and FF of 85.64%, below optimal active layer doping conditions (Nₐ ≈ 10¹⁷ cm⁻³ and Nₜ ≤ 10¹⁵ cm⁻³).

Fig. 10.

Role of N_A and N_t in the Ca₃AsBr₃ active layer on the PCE of solar cell configurations employing different ETLs, (a) WS₂, (b) SnS₂, (c) CdS, and (d) TiO₂.

The effect of the defect density and thickness of Ca₃AsBr₃ layer on the PV performance

The photovoltaic performance of Ca₃AsBr₃-based devices is strongly influenced by film quality and defect densities introduced during fabrication^{80, 81}. As shown in Figs. 11, 12, 13 and 14, defect density critically affects device efficiency, particularly for donor-type defects, which form more readily due to their lower formation energies^{70, 80, 82}. This study evaluates device behavior over active layer thicknesses sopanning from 100 to 2500 nm and defect densities between 10¹⁰ and 10¹⁹ cm⁻³ to reflect realistic conditions and understand their impact on recombination and charge transport.

Fig. 11.

Influence of Ca₃AsBr₃ absorber thickness and defect density on the V_OC in Ca₃AsBr₃-based PSCs with various ETLs (a) WS₂, (b) SnS₂, (c) CdS, and (d) TiO₂.

Fig. 12.

Role of Ca₃AsBr₃ absorber layer thickness and defect density on the J_SC for several ETLs, (a) WS₂, (b) SnS₂, (c) CdS, and (d) TiO₂.

Fig. 13.

Impact of Ca₃AsBr₃ absorber layer thickness and intrinsic defect density on the FF of devices incorporating various ETLs, (a) WS₂, (b) SnS₂, (c) CdS, and (d) TiO₂.

Fig. 14.

The role of absorber layer thickness and defect density in Ca₃AsBr₃ on PCE for PSC architectures using various ETLs, (a) WS₂, (b) SnS₂, (c) CdS, and (d) TiO₂.

Role of defect density and thickness on V_OC

Boosting the thickness of the Ca₃AsBr₃ active layer generally enhances the V_OC by improving light-harvesting efficiency. However, beyond a certain threshold, thicker absorber layers can intensify bulk recombination, particularly in the presence of moderate-to-high defect densities, which introduce trap-assisted recombination pathways and reduce carrier lifetimes, thereby diminishing V_OC. As depicted in Fig. 11(a–d), V_OC declines sharply with increasing defect density (10¹⁰ to 10¹⁹ cm⁻³) ²⁹ and varies with absorber thickness (100 to 2500 nm). V_OC goes from 1.19 V to 0.867 V for WS₂ (Fig. 11a), 1.18 V to 0.872 V for SnS₂ (Fig. 11b), 1.158 V to 0.9815 V for CdS (Fig. 11c), and 1.19 V to 0.998 V for TiO₂ (Fig. 11d) ETL as defect concentrations increase. Optimal V_OC values are achieved when defect densities are maintained at or below ~ 10¹⁵ cm^− 3—representing a moderate defect level consistent with previously reported experimental and simulation ranges for perovskite solar cells—and the absorber thickness exceeds 1200 nm, yielding 1.165 V (WS₂), 1.1599 V (SnS₂), 1.1424 V (CdS), and 1.1657 V (TiO₂), confirming the serious role of defect suppression and active layer thickness in rising device voltage output.

Effect of defect density and thickness on J_SC

The J_SC the highest current produced when the solar cell operates under zero voltage is governed by the efficient collection of photogenerated carriers. It is controlled by multiple contributing factors, including the device’s active area, incident photon flux, spectral distribution of the light, and the absorption characteristics of the active compound⁸³. Enhancing the thickness of the Ca₃AsBr₃ absorber generally improves light absorption, thereby increasing J_SC. However, excessively thick layers may hinder carrier extraction due to recombination losses before collection, resulting in a decline in J_SC. Furthermore, elevated defect densities inside the absorber layer act as non-radiative recombination centers, dropping carrier lifetime and overall collection efficiency, which also degrades J_SC⁸⁴.

As demonstrated in Fig. 12(a–d), the J_SC of the Ca₃AsBr₃-based PSC varies with absorber defect density and thickness. Specifically, J_SC declines from 22.30 to 12.70 mA/cm² for WS₂ (Fig. 12a), from 21.80 to 8.95 mA/cm² for SnS₂ (Fig. 12b), from 21.75 to 7.55 mA/cm² for TiO₂ (Fig. 12c), and from 21.70 to 6.90 mA/cm² for CdS (Fig. 12d) as the defect density upsurges from 10¹⁰ to 10¹⁹ cm⁻³ and thickness varies from 100 to 2500 nm. The highest J_SC values are observed once the active layer thickness exceeds 1500 nm and the defect concentration remains below 10¹⁵ cm⁻³. Under optimized conditions at a thickness of 1200 nm and a defect concentration of 10¹⁵ cm⁻³, the peak J_SC values attained are 20.55 mA/cm² for WS₂, 19.993 mA/cm² for SnS₂, 19.783 mA/cm² for CdS, and 19.73 mA/cm² for TiO₂.

Roll of defect density and thickness on FF

The most effective thickness for the Ca₃AsBr₃ absorber represents a harmony between efficient photocarrier collection and minimized recombination losses. Greater thickness can contribute to a rise in series resistance and lessened carrier extraction efficiency, thereby lowering the FF. Additionally, elevated defect densities in the active layer enhance non-radiative recombination and hinder charge transport, further degrading FF^85,86. As displayed in Fig. 13, the FF exhibits a trend analogous to that of the V_OC⁴⁶, declining as defect density increases and thickness deviates from the optimal range. The FF of FTO/ETL(WS₂, SnS₂, CdS, and TiO₂)/Ca₃AsBr₃/Au architectures decreases across different ETLs: from 86.80% to 69.56% for WS₂ (Fig. 13a), 86.95% to 68.60% for SnS₂ (Fig. 13b), 79.60% to 36.00% for TiO₂ (Fig. 13c), and 82.85% to 64.30% for CdS (Fig. 13d). Maximum FF values were attained at a thickness of 1200 nm and a defect concentration of 10¹⁵ cm⁻³, specifically: 85.64% for WS₂, 86.34% for SnS₂, 82.26% for CdS, and 77.84% for TiO₂.

Role of defect density and thickness on PCE

The interaction of V_OC, J_SC, and FF influences the PSC’s performance as measured by PCE. Therefore, adjusting the active layer’s thickness is necessary to get the best PCE, which calls for balancing elements like recombination, carrier collection, and light absorption. Increased defect densities have an adverse effect on each of the separate output parameters, which lowers PCE⁸⁷. The PCE of FTO/ETL(WS₂, SnS₂, CdS, and TiO₂)/Ca₃AsBr₃/Au configurations exhibits a decreasing trend across various ETLs: from 23.05% to 7.650% for WS₂ (Fig. 14a), 22.35% to 5.40% for SnS₂ (Fig. 14b), 20.50% to 4.850% for TiO₂ (Fig. 14c), and 20.50% to 2.450% for CdS (Fig. 14d). Figure 14 shows the optimal PCE attained at a thickness of 1200 nm and a defect density of 10¹⁵ cm⁻³: 20.50% for WS₂, 20.02% for SnS₂, 18.59% for CdS, and 17.80% for TiO₂.

Optimization of interface defect densities

Interfacial trap states act as recombination centers, limiting photogenerated carrier collection and thereby reducing device performance⁸⁸. Figures 15, 16, 17 and 18 illustrate how variations in Ca₃AsBr₃ absorber thickness (100–2500 nm) and interfacial defect concentration (10¹⁰–10¹⁹ cm⁻²) influence the output parameters of the PSC. The interface quality between Ca₃AsBr₃ and ETLs (WS₂, SnS₂, CdS, TiO₂) critically governs PV performance^{70, 89, 90}. In the present study, the interface defect states were modeled in SCAPS-1D using a simplified approach by assuming a uniform distribution of trap states at the ETL/Ca₃AsBr₆ interface. While this method effectively captures the general trends of interfacial recombination, it does not fully represent the complex nature of real interfaces, where defect energetics, spatial distribution, and interface chemistry may vary significantly. In practical devices, interfacial recombination is one of the dominant loss mechanisms, strongly influencing carrier lifetime and extraction efficiency. Recent device-level optimization studies have emphasized the importance of advanced interface engineering strategies, including defect passivation, band alignment tuning, and interface modification techniques to suppress recombination losses and enhance device performance⁹¹. Therefore, more realistic interface modeling and experimental validation are required for accurate prediction of device behavior. Once interfacial defect densities exceed 10¹³ cm⁻², sharp declines in efficiency occur due to enhanced carrier recombination from high-density interfacial states^{80, 92}.

Fig. 15.

Three-dimensional representation of the variation in V_OC with absorber thickness and interfacial defect density at the ETL/Ca₃AsBr₃ interface for different ETLs, (a) WS₂, (b) SnS₂, (c) CdS, and (d) TiO₂.

Fig. 16.

Three-dimensional analysis of J_SC variations as influenced by active thickness and interfacial defect density at the ETL/Ca₃AsBr₃ interface for different ETLs, (a) WS₂, (b) SnS₂, (c) CdS, and (d) TiO₂.

Fig. 17.

Concurrent effect of interfacial defect density and active layer thickness on FF for FTO/ETL (a) WS₂, (b) SnS₂, (c) CdS, and (d) TiO₂/Ca₃AsBr₃/Au architectures.

Fig. 18.

Three-dimensional visualization of the PCE response in Ca₃AsBr₃-based PSCs as a function of interface defect density and absorber thickness for different ETLs, (a) WS₂, (b) SnS₂, (c) CdS, and (d) TiO₂.

Impact of interface defect densities and thickness on V_OC

The V_OC of Ca₃AsBr₃-based PSC exhibits a substantial reduction with raising interfacial defect density, mainly owing to elevated non-radiative recombination at the ETL/ Ca₃AsBr₃ interface. As evidenced by Fig. 15, V_OC declines for various ETLs: from 1.18 V to 0.998 V for WS₂ (Fig. 15a), 1.189 V to 0.65 V for SnS₂ (Fig. 15b), 1.20 V to 0.551 V for CdS (Fig. 15c), and 1.186 V to 0.8815 V for TiO₂ (Fig. 15d), as the interface defect density rises from 10¹⁰ to 10¹⁹ cm⁻². This output deterioration becomes significantly more pronounced beyond 10¹³ cm^− 2 due to enhanced SRH recombination and the consequent reduction in carrier lifetime. It is important to note that an interfacial defect density of 10¹³ cm^− 2 falls within the moderate range rather than being considered low. Such values are commonly reported for experimentally fabricated thin-film interfaces, where imperfections arising from lattice mismatch, surface roughness, and fabrication-induced defects are unavoidable. Optimized device performance is achieved at a Ca₃AsBr₃ absorber thickness of approximately 1200 nm combined with an interfacial defect density of 10¹³ cm^− 2, resulting in V_OC values of 1.165 V for WS₂, 1.1599 V for SnS₂, 1.1424 V for CdS, and 1.1659 V for TiO₂. These results are consistent with realistic interfacial defect density ranges (Nit ≈ 10¹⁰–10¹³ cm^− 2), indicating that the selected values represent practically achievable moderate defect conditions rather than idealized low-defect scenarios⁸⁸. Notably, the WS₂-based device maintains the uppermost V_OC under these conditions, underscoring its favorable band alignment and superior carrier-selective interface properties.

Effect of interface defect densities and thickness on J_SC

Figure 16(a–d) illustrates the variation in PV output as a function of both interfacial defect density (ranging from 10¹⁰ to 10¹⁸ cm⁻²) and active layer thickness (0.1 to 3.0 μm) at the WS₂/Ca₃AsBr₃, SnS₂/Ca₃AsBr₃, CdS/Ca₃AsBr₃, and TiO₂/Ca₃AsBr₃ interfaces. As the interfacial defect density is increases, J_SC drops significantly from 22 mA·cm⁻² to 6 mA·cm⁻² for WS₂ and SnS₂, 22 mA·cm⁻² to 5.5 mA·cm⁻² for CdS, and 22 mA·cm⁻² to 6.5 mA·cm⁻² for TiO₂ because of enhanced trap-assisted recombination at the defective interfaces. The reduction in carrier collection efficiency is further compounded at lower absorber thicknesses. Optimal J_SC are attained at a thickness of 1200 nm and an interfacial defect density of 10¹³ cm^− 2: 20.55 mA/cm² for WS₂, 19.976 mA/cm² for SnS₂, 19.78 mA/cm² for CdS, and 19.71 mA/cm² for TiO₂. Among all configurations, the WS₂-based ETL yields the highest J_SC, ascribed to its better interface quality and favorable band alignment facilitating efficient carrier extraction.

Influence of interface defect densities and thickness on FF

The FF of the Ca₃AsBr₃-based PSC exhibits a significant decline across different ETLs as the interfacial defect density increases. Specifically, FF decreases from 86% to 62.16% for WS₂ (Fig. 17a), 88% to 66.28% for SnS₂ (Fig. 17b), 88% to 62.10% for CdS (Fig. 17c), and 80% to as low as 37% for TiO₂ (Fig. 17d).

This degradation is largely caused by elevated trap-assisted recombination and impaired charge extraction at the ETL/absorber interface, which is exacerbated by elevated defect densities and suboptimal carrier transport pathways. Interfacial traps act as non-radiative recombination centers, meaningfully reducing the quality of the charge transport junction and suppressing the power output of the device. Figure 17 highlights that the optimal FF values are achieved at an active layer thickness of 1200 nm and an interfacial defect concentration of 10¹³ cm⁻². Within these operational settings, the FF reaches 85.64% for WS₂, 86.34% for SnS₂, 82.26% for CdS, and 77.84% for TiO₂. Among these, the WS₂ ETL demonstrate superior FF performance because of their advantageous energy level configuration and lower interfacial recombination rates, promoting more efficient charge extraction and transport.

Influence of interface defect densities and thickness on PCE

The PCE of Ca₃AsBr₃-based PSC exhibits significant dependence on the choice of ETL and the interface quality. As illustrated in Fig. 18, rising the interfacial defect density from 10¹⁰ to 10¹⁹ cm⁻² leads to a substantial decay in PCE due to intensified trap-assisted recombination and hindered carrier extraction at the ETL/absorber junction. Specifically, PCE decreases from 22% to 4% for WS₂ (Fig. 18a), 22% to 3.5% for SnS₂ (Fig. 18b), 22% to 3% for CdS (Fig. 18c), and 20.5% to 3.21% for TiO₂ (Fig. 18d). The optimal performance is reached at an interface defect concentration of 10¹³ cm^− 2 and active thickness of 1200 nm, yielding peak PCEs of 20.55% for WS₂, 20.02% for SnS₂, 18.59% for CdS, and 17.80% for TiO₂.

Spatial distribution profiles of electron and hole concentrations, along with carrier generation and recombination rates

Figure 19 explains the impact of active layer thickness on carrier concentration and the generation–recombination (G–R) dynamics within Ca₃AsBr₃-based optimized device architecture. The G–R profile, evaluated over a thickness range from 0 to 1.35 μm, reveals the spatial evolution of photogenerated carriers within the absorber layer. Upon illumination, incident photons promote electrons from the VB into the CB, generating electron–hole pairs. The peak generation rates occur within the 1.17–1.28 μm region, indicating the zone of most efficient optical absorption and carrier excitation. Notably, devices employing WS₂ as the ETL demonstrate the highest generation rate within this depth, attributed to enhanced photon absorption and superior charge extraction at this position compared to other configurations. Beyond this optimal region, the generation rate declines progressively due to increased optical losses and reduced absorption efficiency. The spatially resolved generation rate G(x) is estimated using SCAPS-1D, incorporating the position-dependent photon flux N_phot(λ, x) as,

9

Fig. 19.

Spatial distribution of (a) hole concentration, (b) electron concentration, (c) total photogeneration rate, and (d) total recombination rate across PSCs incorporating different ETLs.

Conversely, a considerable amount of electron-hole concentration is noted in the area that is farthest from the ETL/absorber interface, which is explained by a extended diffusion length and a shorter carrier lifetime. Therefore, across the entire cell area, both diffusion length and carrier density strongly affect the recombination rate. Defect states in the absober of each PSC significantly contribute to electron-hole recombination. The highest recombination rates, analogous to the generation rate, were observed in the 1.17 to 1.28 μm range, with WS₂ ETL exhibiting the highest rate. This occurs when high-energy electrons in the CB overcome the energy barrier to the VB, stabilizing by occupying holes with minimal energy loss. Elevated levels of acceptor doping result in a slight enhancement of hole concentration in the Ca₃AsBr₃ area, establishing an effective DOS in the VB, reliable with findings from other studies on defect density and carrier concentration^{93, 94, 95}. Thus, integrating Ca₃AsBr₃-based heterojunctions with WS₂ as the ETL offers a promising pathway toward high-efficiency PSCs by effectively minimizing electron–hole recombination.

Analysis of quantum efficiency (Q-E) and current density-voltage (J-V) characteristics

Figure 20 presents the J–V behavior of Ca₃AsBr₃-based inorganic PSCs incorporating various ETLs across active thicknesses changing from 0.1 to 2.5 μm. The performance of both J–V and QE characteristics is highly dependent on absorber thickness. Thicker Ca₃AsBr₃ layers enhance photon absorption and photogenerated carrier density, leading to increased J_SC. However, excessive thickness also promotes bulk recombination and reduces charge extraction efficiency, thereby decreasing V_OC and FF. Figures 20 (a, c, e, and g) illustrate the J–V responses for the various ETL arrangements, while Fig. 20 (i) shows the optimized J–V profiles at an active layer thickness of 1200 nm, N_A of 10¹⁷ cm⁻³, N_t of 10¹³ cm⁻³, and N_int of 10¹³ cm⁻². The corresponding photovoltaic parameters were: V_OC= 1.165 V and J_SC= 20.55 mA/cm² for WS₂, 1.161 V and 19.99 mA/cm² for SnS₂, 1.166 V and 19.73 mA/cm² for TiO₂, and 1.143 V and 19.78 mA/cm² for CdS. Among them, the WS₂-based device achieved the highest output due to its efficient carrier extraction and reduced recombination losses, enabling a peak PCE of 20.5% at an optimized absorber thickness of 1200 nm. Additionally, QE enhancement at longer wavelengths is directly correlated with increased optical path length and photon absorption in thicker Ca₃AsBr₃ layers.

Fig. 20.

J-V and Q-E characteristics of proposed PSCs (a,b) WS₂, (c,d) SnS₂, (e,f) TiO₂, (g,h) CdS, and (i,j) Optimal ETL, as a function of thickness of Ca₃AsBr₃ active layer.

However, an excess of thickness can result in profound photon absorption, leading to reduced efficiency in charge carrier collection as well as a decrease in Q-E. For thicknesses of the active layer varying from 100 nm to 2500 nm, Fig. 20(b, d, f, and h) displays analogous impacts on the Q-E response of identical PSCs, as shown in Jsc. Furthermore, thicker active layers result in increased cell resistance and a longer diffusion length, both of which lower FF.

However, the rise of thickness gradually decelerates rate of enhancement of QE and attains a saturated value beyond 1200 nm. The J-V behaviors were tuned for the investigated ETL, as shown in Fig. 20(J). WS₂ had better QE characteristics than SnS₂, TiO₂, and CdS, which is comparable to the J-V characteristics. In Fig. 20(J), J–V performance was evaluated under optimized conditions. Analogous to the J-V response, WS₂ exhibited the better QE response compared to SnS₂, TiO₂, and CdS. Figures 20(i) and 20(j) reveal the projected J-V and Q-E characteristics of the Au/Ca₃AsBr₃/WS₂/FTO solar cell, revealing a PCE of 20.50%, J_SC of 20.55 mAcm^− 2, FF of 85.64%, and V_OC of 1.165 V.

Comparative performance analysis

To further validate the effectiveness of the proposed device, a comparative performance analysis with recently reported SCAPS-based photovoltaic systems is presented in Table 5. The Ca₃AsBr₃-based HTL-free device exhibits a maximum PCE of 20.50%, along with a high FF of 85.64% and an V_OC of 1.165 V. These values indicate efficient charge transport, reduced recombination losses, and favorable band alignment, particularly for the WS₂-based configuration. In comparison, previously reported devices such as FTO/TiO₂/RbGeI₃/CuI demonstrate a comparable PCE of 18.97%, but with lower V_OC and FF values, indicating relatively higher recombination losses. On the other hand, the FTO/TiO₂/Cs₂SnBr₆/N-doped TiO₂/Au structure achieves a significantly higher PCE of 38.70%, primarily due to enhanced charge transport and optimized multilayer interface engineering. However, such high-performance devices often involve more complex architectures and additional functional layers, which may increase fabrication complexity and cost. In contrast, the present work emphasizes a simplified HTL-free architecture using a lead-free Ca₃AsBr₃ absorber, which offers a balanced combination of efficiency, structural simplicity, and potential stability. The elimination of the hole transport layer not only reduces fabrication complexity but also minimizes degradation pathways associated with commonly used HTL materials. Therefore, despite not achieving the highest reported PCE, the proposed device demonstrates competitive performance with practical advantages, making it a promising candidate for scalable and stable next-generation photovoltaic applications.

Table 5.

Performance comparison of the proposed Ca₃AsBr₃-based HTL-free perovskite solar cell with recently reported SCAPS-based photovoltaic devices.

Device	VOC (V)	JSC (mA/cm2)	FF (%)	PCE (%)	Reference
FTO/WS2/Ca3AsBr3/Au	1.165	20.55	85.64	20.50	This work
FTO/SnS2/Ca3AsBr3/Au	1.159	19.99	86.34	20.02	This work
FTO/CdS/Ca3AsBr3/Au	1.143	19.78	82.26	18.59	This work
FTO/TiO2/Ca3AsBr3/Au	1.166	19.73	77.84	17.80	This work
FTO/TiO2/ RbGeI3/ CuI	0.777	33.969	71.93	18.97	96
FTO/TiO2/Cs2SnBr6/N-doped TiO2/Au	1.29	33.34	90.21	38.70	97

Limitation

This study considers absorber thicknesses up to 1200 nm to explore the theoretical performance limits of the device. However, such thicknesses may not be practically feasible for perovskite solar cells, as they can lead to increased series resistance and inefficient charge transport. In realistic device fabrication, thinner absorber layers (typically 300–900 nm) are preferred to ensure efficient carrier extraction and minimize resistive losses. Future work should focus on optimizing device architectures, improving carrier mobility, and engineering interfaces to achieve high efficiency at reduced thicknesses.

Conclusions

This work introduces an integrated design-to-performance approach for Ca₃AsBr₃-based perovskite solar cells by combining DFT-driven material analysis with SCAPS-1D device simulations. The results confirm that Ca₃AsBr₃ exhibits promising mechanical, optical, and electronic properties, such as a direct bandgap of 1.66 eV, strong dielectric behavior, and dynamic structural stability. At the device level, the study reveals that effective passivation of bulk and interfacial defects, along with the careful selection of electron transport layers (ETLs), plays a vital role in reducing recombination losses and enhancing overall efficiency. Among the ETLs considered, WS₂ provides the most favorable band alignment and improved charge transport, leading to superior device performance. The optimized HTL-free architecture (FTO/WS₂/Ca₃AsBr₃/Au) delivers high power conversion efficiency (PCE) of 20.50%, highlighting Ca₃AsBr₃ as a promising candidate for next-generation, lead-free perovskite solar cells. This study not only offers key insights into designing efficient and environmentally friendly PSCs but also establishes a theoretical foundation to guide future experimental synthesis and device development.

Author contributions

Md. Ferdous Rahman, Md. Azizur Rahman, Mst. Marufa Sabrin: Methodology, validation, software, Conceptualization, investigation, formal analysis, data curation, visualization, writing-original draft, and review and editing.Md. Faruk Hossain, Noureddine Elboughdiri, Nacer Badi, Aijaz Rasool Chaudhry, Ahmad Irfan: Investigation, validation, software, formal analysis, data curation, writing-originaldraft, and review and editing.

Data availability

All data are available from the corresponding author upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.