Perovskite Tandem Solar Cells: A Review of Recent Progress and Future Perspectives

Abstract

Perovskite tandem solar cells (TSCs) have emerged as a transformative photovoltaic technology, offering a viable pathway to surpass the Shockley-Queisser limit of single-junction devices by enabling broader solar spectrum utilization and reduced thermalization losses. This review provides a comprehensive overview of recent progress in perovskite-based TSCs, covering four major device architectures: perovskite/silicon, perovskite/CIGS, all-perovskite, and perovskite/organic TSCs. We systematically discuss the fundamental working principles, including bandgap engineering, charge generation and separation, and current-voltage matching, followed by an in-depth analysis of strategies for perovskite layer regulation, interface engineering, and transport-layer optimization. Key advancements, such as compositional engineering, defect passivation, crystallization control, and optical management, have synergistically pushed power conversion efficiencies (PCEs) beyond 34% for perovskite/silicon TSCs and over 28% for all-perovskite and perovskite/organic configurations. Despite these achievements, critical challenges remain, including material instability, halide phase segregation, lead toxicity, scalable fabrication, and cost-effective integration. This review also outlines future perspectives, emphasizing the development of lead-free perovskites, novel charge-transport materials, advanced encapsulation techniques, and large-area manufacturing processes. With continued interdisciplinary efforts, perovskite TSCs hold great promise for driving the global transition toward sustainable and low-carbon energy systems.

1. Introduction

The demand for clean and renewable energy sources has become increasingly urgent in the face of global climate change and the depletion of fossil fuels. Solar energy, as an abundant and sustainable resource, holds great promise for meeting the world’s energy needs. Photovoltaic (PV) technology, which converts sunlight directly into electricity, has witnessed significant progress in recent years. Among the various PV technologies, perovskite solar cells (PSCs) have emerged as a highly promising candidate due to their exceptional optoelectronic properties, low cost, and ease of fabrication. Perovskite materials, with the general formula ABX₃, have shown remarkable performance in single-junction devices, achieving PCEs exceeding 26%. However, the theoretical PCE of single-junction solar cells is fundamentally restricted to 29.4% by the Shockley-Queisser (S-Q) limit [1]. To overcome this limitation, TSCs have been developed, which combine two or more semiconductor materials with complementary bandgaps to capture a broader spectrum of sunlight and minimize thermalization losses [2]. Perovskite materials are uniquely suited for TSCs due to their highly tunable bandgaps (1.2–2.3 eV), allowing them to function as ideal top-cell absorbers [3]. Beyond bandgap optimization and current matching, the long-term reliability of perovskite-based top cells is another critical factor determining their practical viability. Wide-bandgap (WBG) mixed-halide perovskites often suffer from light-induced halide phase segregation, which induces bandgap instability and substantial voltage losses during operation. Recent efforts have focused on enhancing their stability through compositional engineering, such as triple-halide alloying and incorporation of inorganic cations, as well as defect passivation and low-dimensional structural modulation to suppress ion migration pathways. In addition, methylammonium-free compositions and lattice reinforcement strategies based on strong intermolecular interactions have further improved thermal robustness. These advances demonstrate that engineered WBG perovskites can achieve enhanced operational stability under continuous illumination and elevated-temperature conditions, strengthening their role as promising top-cell absorbers for tandem architectures.

Among the various tandem configurations, perovskite/silicon TSCs have attracted the most significant research attention and have demonstrated the most impressive progress. Silicon PV technology, with its well-established manufacturing infrastructure, high efficiency, and proven long-term stability, provides an ideal bottom-cell platform. The combination of a WBG perovskite top cell with a silicon bottom cell has yielded remarkable results, with certified PCEs reaching 34.85% as of early 2026, substantially exceeding the theoretical limit of single-junction silicon devices [4]. Beyond conventional monofacial designs, bifacial tandem configurations that capture albedo light from the rear side have gained increasing interest, offering not only enhanced energy yield but also potential pathways for improved device stability through current-matching optimization [5]. In parallel, perovskite/Cu(In,Ga)Se₂ (CIGS) TSCs offer a thin-film alternative with flexible substrate compatibility and high efficiency potential, while all-perovskite TSCs present the advantage of a unified material system, potentially simplifying fabrication and reducing costs. Furthermore, perovskite/organic TSCs have emerged as compelling low-temperature-processed alternatives, with recent studies demonstrating that rational sub-cell engineering, including molecular design of novel acceptors to form energetic alloys, can significantly optimize charge carrier dynamics and reduce energy losses within the tandem stack [6].

Despite these remarkable achievements, several critical challenges must be addressed before perovskite-based TSCs can be successfully commercialized. A major bottleneck lies in the WBG perovskite top cells, which often suffer from severe non-radiative recombination and light-induced halide phase segregation. These issues lead to substantial open-circuit voltage (V_oc) deficits, limiting the device performance, and compromise the long-term operational stability of the entire tandem device [7]. Additionally, the design of efficient and robust interconnecting layers (ICLs) remains essential for ensuring low electrical resistance and high optical transparency between the sub-cells, as any inefficiencies in charge recombination or transport at these interfaces can result in significant performance losses [2]. The stability of perovskite materials under operational conditions, particularly in the presence of moisture, oxygen, heat, and continuous illumination, remains a formidable challenge for commercialization. Moreover, the scalability of fabrication processes from laboratory-scale devices to large-area modules, the toxicity concerns associated with lead-containing perovskites, and the economic competitiveness with established PV technologies all represent substantial hurdles that require innovative solutions.

This review aims to provide a comprehensive overview of the current status and recent progress of perovskite TSCs, including their fundamental working principles, diverse device architectures, and State-of-the-Art optimization strategies. We systematically examine the critical aspects of materials engineering, including compositional tuning, defect passivation, interface modification, and crystallization control for high-quality perovskite absorber layers. We also discuss the design and optimization of charge-transport layers (CTLs), ICLs, and transparent electrodes that are crucial for efficient carrier extraction and optical management. Furthermore, we highlight the key challenges associated with device stability, scalability, and lead toxicity, and offer perspectives on future research directions and commercialization pathways. Through this comprehensive analysis, we aim to provide valuable insights for researchers and engineers working toward the development of next-generation high-efficiency, stable, and cost-effective perovskite tandem PVs that can contribute meaningfully to the global transition toward sustainable energy. Figure 1 summarizes the working principles, device architectures, optimization strategies, and challenges of perovskite TSCs.

Figure 1. Overview of perovskite TSCs, including working principles, device architectures, optimization strategies, and challenges.

2. Working Principles of Perovskite TSCs

2.1. Bandgap Engineering

The key principle behind TSCs is the utilization of multiple semiconductor materials with different bandgaps to absorb different portions of the solar spectrum, thereby breaking through the S-Q efficiency limit of single-junction devices. In a single-junction solar cell, the S-Q limit restricts efficiency to ~33% due to transmission of low-energy photons and thermalization of high-energy photons. Photons with higher energy than the bandgap are absorbed but lose excess energy as heat, a loss mechanism called thermalization. TSCs minimize thermalization by using multiple absorbers, each responsible for a section of the solar spectrum. However, while tandem architectures intrinsically reduce thermalization losses through improved spectral utilization, additional loss pathways, such as interfacial recombination, may arise from the introduction of multiple sub-cell interfaces. Therefore, minimizing interfacial losses through optimized device architectures and interface engineering is essential to fully realize the theoretical efficiency advantages of TSCs. In a perovskite TSC, the top cell, typically made of a WBG perovskite material, absorbs the high-energy photons, while the bottom cell, made of a narrow-bandgap (NBG) material (such as Si or CIGS), absorbs the low-energy photons that are transmitted through the top cell. Theoretically, an optimal combination of ~1.9 eV for the top cell and ~1.0 eV for the bottom cell can achieve a PCE of up to ~42% (Figure 2c–e). By carefully selecting the bandgaps of the two materials, it is possible to maximize the absorption of sunlight across the entire solar spectrum, thereby increasing the PCE of the tandem cell. The bandgap of perovskite materials can be tuned by varying the composition of the A, B, and X components. For example, by substituting different organic or inorganic cations for the A-site, different divalent metal cations for the B-site, or different halide anions for the X-site, the bandgap of the perovskite can be adjusted within a wide range (Figure 2g). This tunability of the bandgap makes perovskite materials highly suitable for use in TSCs, as it allows for precise optimization of the optical absorption characteristics of the top cell to match the requirements of the tandem architecture [8].

2.2. Charge Generation and Separation

When sunlight is incident on a perovskite TSC, photons are absorbed by both the top and bottom cells, generating electron–hole pairs (excitons). In such a configuration, the perovskite top cell absorbs visible light but transmits infrared and near-infrared light to the bottom cell, enabling a large portion of the solar spectrum to be utilized. In perovskite materials, which exhibit steep absorption edge and exceptionally low sub-gap absorption, due to their strong optical absorption coefficients and high carrier mobility, the generated excitons can quickly dissociate into free electrons and holes. The electrons and holes are then separated by the built-in electric field at the heterojunction between the perovskite and the CTLs and are transported toward the respective electrodes. In a tandem cell, the WBG top cell generates a higher voltage from high-energy photons, while the NBG bottom cell generates a lower voltage from low-energy photons. The total voltage is the sum of the two sub-cells, but the current is limited by the lower of the two. Therefore, the charge transfer between the two sub-cells is a crucial process. To ensure efficient charge transfer, a recombination layer (RL) or an interconnecting layer is typically inserted between the two sub-cells. In a two-terminal monolithic tandem, directly connecting the n-layer of one sub-cell to the p-layer of the other would form a blocking p-n junction (Figure 2a,b). To resolve this, an RL or tunnel junction is inserted, which supports the flow of electrons from one sub-cell to the other. This layer serves to facilitate the transfer of holes from the bottom cell to the top cell, and electrons from the top cell to the bottom cell, while minimizing recombination losses. The design and properties of the RL play a vital role in determining the performance of the tandem cell, as any inefficiencies in charge transfer can lead to significant losses in PCE [9]. In particular, interfacial recombination at the connecting layers represents an additional non-radiative loss pathway, where photogenerated carriers are lost as heat rather than contributing to photocurrent. This loss mechanism is distinct from thermalization losses associated with photon-energy mismatch and must be minimized through careful interface engineering and RL optimization.

2.3. Current and Voltage Matching

In a TSC, the two sub-cells are connected in series, meaning that the current flowing through both sub-cells must be the same [10]. To achieve this, the optical absorption and charge generation in each sub-cell need to be carefully balanced to ensure that they produce the same amount of current under the same illumination conditions (Figure 2f). This process is known as current matching. If the current generated by one sub-cell is significantly higher than that of the other sub-cell, a phenomenon called current mismatch occurs, which can lead to a reduction in the PCE of the tandem device [11]. To avoid current mismatch, various strategies can be employed, such as optimizing the thickness and composition of the absorber layers in each sub-cell, adjusting the optical properties of the device through the use of antireflective coatings or light-trapping structures, and carefully selecting the bandgaps of the materials to ensure efficient absorption of different portions of the solar spectrum. In addition to current matching, voltage matching is also an important consideration in TSCs. The V_oc of a tandem cell is the sum of the V_oc values of the individual sub-cells. However, due to losses at the interfaces between the sub-cells and in the CTLs, the actual V_oc of the tandem cell is usually lower than the sum of the V_oc values of the sub-cells. To maximize the V_oc of the tandem cell, it is essential to minimize these losses through the use of high-quality materials, optimized device architectures, and efficient CTLs [12].

Figure 2. Schematic illustration of (a) the SHJ cell and (b) the test configuration. Reprint with permission [13]. Copyright 2026, Elsevier. Schematic illustration of (c) single-junction and (d) tandem solar cell. (e) Response of the top and bottom cells to the spectrum of light. Reprint with permission [9]. Copyright 2018, Elsevier. (f) Sketch of the current-matching machine (CMM) experimental setup. Reprint with permission [11]. Copyright 2013, IEEE. (g) Advantage of a TSC. Reprint with permission [11]. Copyright 2013, IEEE.

Figure 2. Schematic illustration of (a) the SHJ cell and (b) the test configuration. Reprint with permission [13]. Copyright 2026, Elsevier. Schematic illustration of (c) single-junction and (d) tandem solar cell. (e) Response of the top and bottom cells to the spectrum of light. Reprint with permission [9]. Copyright 2018, Elsevier. (f) Sketch of the current-matching machine (CMM) experimental setup. Reprint with permission [11]. Copyright 2013, IEEE. (g) Advantage of a TSC. Reprint with permission [11]. Copyright 2013, IEEE.

3. Device Architectures of Perovskite TSCs

3.1. Perovskite/Silicon TSCs

Perovskite/silicon TSCs are one of the most widely studied and promising architectures in the field of perovskite TSCs. In this architecture, a WBG perovskite top cell is stacked on top of an NBG silicon bottom cell. The silicon bottom cell, which is a well-established and highly efficient PV technology, provides a stable and reliable foundation for the tandem cell, while the perovskite top cell enables the absorption of high-energy photons that are not efficiently captured by the silicon cell. There are two main types of perovskite/silicon TSCs: monolithic and mechanically stacked. In monolithic perovskite/silicon TSCs, the perovskite and silicon sub-cells are grown directly on top of each other, sharing a common RL. This architecture offers the advantage of a simpler device structure and potentially higher efficiency, as it allows for more efficient charge transfer between the two sub-cells. However, the fabrication process of monolithic tandem cells is more challenging, as it requires precise control over the growth and interface properties of the two materials. In mechanically stacked perovskite/silicon TSCs, the perovskite and silicon sub-cells are fabricated separately and then mechanically stacked on top of each other. This architecture offers greater flexibility in terms of material selection and device design, as the two sub-cells can be optimized independently. However, the mechanical-stacking process can introduce additional optical and electrical losses, which can reduce the efficiency of the TSCs.

3.1.1. Perovskite-Layer Regulation

The performance of perovskite/silicon TSCs relies critically on the systematic regulation of the perovskite layer, which encompasses multiple interrelated dimensions: from material composition design, defect and interface passivation, and crystallization-process control to the final optical and electrical performance matching. These aspects are interconnected and progressively built upon one another, determining the ultimate efficiency and stability of the tandem devices.

Compositional engineering: A central objective of compositional engineering is the precise tuning of the perovskite bandgap by regulating ionic components. Early studies demonstrated that adjusting the Br⁻/I⁻ ratio in FA_0.9Cs_0.1Pb(I_1−xBr_x)₃ could widen the bandgap from 1.53 eV to 1.7 eV, enhancing stability and enabling efficient spectral filtering for four-terminal reflective tandem devices. Theoretical calculations further guided bandgap optimization: detailed balance limit analysis determined ideal top-cell bandgaps for different tandem architectures, while optoelectronic modeling enabled rational tuning of A-site cations and halide anions, supporting efficient current matching and high-efficiency tandem design. Guided by these theoretical and experimental foundations, mixed-cation/mixed-halide perovskites with precisely tuned bandgaps (e.g., 1.61 eV) were developed. Combined with buried interface modification using Al₂O₃ nanoparticles, these compositions improved crystallinity, suppressed interfacial recombination, and delivered high V_oc and efficient performance under current-matched conditions [14,15,16,17]. A critical challenge for WBG perovskites is intrinsic halide phase segregation that degrades device performance. The fundamental origin of light-induced halide phase segregation in WBG mixed-halide perovskites arises from two synergistic driving forces: photogenerated charge accumulation distorts the Pb-X octahedral framework to generate local lattice strain, while thermodynamically favored halide ion migration redistributes I⁻ and Br⁻ across grain boundaries under continuous illumination. Light excitation weakens Pb-Br bonds preferentially, creating mobile halide vacancies that act as fast diffusion channels, eventually forming Br-rich and I-rich segregated domains with mismatched bandgaps. This spatial band fluctuation introduces severe non-radiative recombination centers and permanent V_oc loss, which can be suppressed via lattice rigidification and halide ion immobilization strategies. The close correlation between photoinduced lattice strain, ionic mobility, and long-term phase instability has been systematically validated via seasonal outdoor stability measurements of n-i-p perovskite devices [18]. Two effective strategies have been established to address this issue. Lowering Br⁻ content to tune the bandgap to 1.59–1.62 eV enables efficient current matching under bifacial irradiation by utilizing albedo to boost the silicon sub-cell current, which suppresses light-induced phase segregation and achieves certified efficiencies over 25% while balancing efficiency and stability. For applications requiring WBG perovskites, triple-halide (Cl/Br/I) alloying by incorporating MAPbCl₃ into CsFA-based precursors promotes Cl incorporation via Br-mediated lattice contraction, significantly suppressing phase segregation, lowering defect density, and delivering a certified efficiency of 27% in 1 cm² perovskite/silicon tandem devices [5,19,20]. Su et al. optimized MA-free WBG perovskites for 4-T perovskite/silicon TSCs. They screened Cs-FA-Br-I compositions to obtain a stable 1.65 eV bandgap system. Rb⁺ doping suppressed lattice strain and phase segregation, reduced defects, and prolonged carrier lifetime. The 4-T tandem device achieved 29.98% PCE [21]. Shen et al. fabricated a WBG CPRA layer via cesium formate spin coating, which enhances light in-coupling, pins the absorption edge at 1.68 eV for ideal current matching, and yields a champion PCE of 28.53% [22]. To mitigate toxicity and spectral mismatch in lead-free perovskites, gradient compositional tuning of X-site halides (I, Br, and Cl) modulates the CsSnX₃ bandgap, enabling spectral complementarity between sub-cells for optimal current matching, achieving 33.03% efficiency [23]. Zhang et al. employed a halogen anion pre-homogenization strategy via co-evaporated PbIBr eutectic template to reduce nucleation disparity between iodide and bromide, enabling phase-separation-free WBG tuning (1.59–1.78 eV) and promoting homogeneous crystallization with minimized interface defects, achieving 30.83% PCE [24]. De Wolf et al. incorporated methylenediammonium dichloride to generate THTZ-H⁺ cations in situ, which reinforce lattice interactions through multidirectional N-H^…I hydrogen bonding, thereby enhancing phase stability under light/heat and achieving 33.7% certified PCE [25]. Table 1 summarizes the compositional engineering strategies for perovskite/silicon TSCs. In summary, triple-halide multi-cation schemes outperform simple Br/I binary tuning, as they rigidify lattices to suppress phase segregation fundamentally. MA-free designs gain thermal stability but demand stricter deposition control, representing a trade-off between stability and fabrication tolerance.

Table 1. Compositional engineering strategies and the PV parameters for perovskite/silicon TSCs.

Device Architecture	Active Area (cm2)	PCE (%)	Jsc (mA/cm2)	Voc (V)	FF (%)	Stability	Year [Ref.]
ITO/PTAA/Al2O3-NPs/RbCsMAFAPb(IBr)3/C60/BCP/AgAg/(n)a-Si/(i)a-Si/(n)c-Si/(i)a-Si/(p)a-Si/ITO	/	23.85	17.91	1.73	77.77	/	2024 [17]
ITO/PTAA/Al2O3-NPs/RbCsMAFAPb(IBr)3/C60/BCP/AgAg/(n)a-Si/(i)a-Si/(n)c-Si/(i)a-Si/(p)a-Si/ITO	1.03	25.2	22.5	1.84	78	/	2021 [19]
MgF2/IZO/SnO2/C60/perovskite(bandgap:1.59–1.70 eV)/SnO2/ITO(RJ) ITO(RJ)/a-Si:H(i)/c-Si(n)/a-Si:H(i)/ITO	1.00	27.13	19.12	1.886	75.3	Retained >96% of the initial PCE after 1000 h of continuous MPP operation at 60 °C in a N2 atmosphere.	2020 [20]
ITO/NiOx/poly-TPD/Cs0.22FA0.78Pb(I/Br/Cl)3/C60/BCP/SnO2-ZTO/ITO//c-Si(n)/a-Si:H(i)/a-Si:H(p+)/ITO/Ag	0.1	19.21	21.12	1.14	79.74	Maintained 84% of its initial PCE after 186 h of continuous 1-sun illumination.	2025 [21]
ITO/NiOx/spiro-OMeTAD/MA-freeRb-dopedCs0.20FA0.80Pb(I0.80Br0.20)3(~1.65 eV)/PCBM/PEI/IZO ITO/a-Si:H(p+)/a-Si:H(i)/ c-Si(n)/a-Si:H(i)/a-Si:H(n+)/ ITO/Ag	0.35	28.53	20.19	1.73	81.93	Retained ~90% of its initial PCE after 400 h under continuous 1-sun illumination at 85 °C and 85% RH.	2024 [22]

Defect and interface passivation: While compositional engineering optimizes the intrinsic electronic structure of perovskites, the inevitable defects (e.g., halide vacancies and grain boundary states) and interfacial recombination still limit tandem performance. Thus, targeted defect and interface passivation has become a key complementary strategy. The central objective of defect and interface passivation is to suppress non-radiative recombination and enhance operational stability, primarily achieved through strategic defect passivation and meticulous interface engineering. This approach focuses on precise defect targeting at the molecular level, with strategies ranging from small-molecule hydrogen bond donors to functionalized 2D additives, all designed to interact specifically with defect sites. Wolf et al. used carbazole molecules to form hydrogen-bonded networks at grain boundaries (Figure 3b), specifically passivating defects to suppress halide phase separation, which led to a certified PCE of 28.2% for perovskite/silicon TSCs [26]. For stubborn deep-level traps induced by iodine oxidation (forming I₂), Yu et al. incorporated ammonium diethyldithiocarbamate (ADDC) into the perovskite precursor, which effectively reduced I₂ to I⁻, suppressing deep-level traps and ion migration [27]. Shin et al. employed an anion-engineered 2D additive (PEA(I_0.25SCN_0.75)). The SCN⁻ induces a conductive PbI₂-based 2D phase at grain boundaries that simultaneously passivates defects and enhances charge transport, tackling both major limitations of WBG perovskites [28]. Aspartic acid (ASP) was employed as a double-charged additive to coordinate with PbI₂, through which defects were passivated and non-radiative recombination was suppressed, yielding 30.68% PCE with 95% retention after 1900 h [29]. BD-Boc was introduced to coordinate with undercoordinated Pb²⁺ via carbonyl groups, by which crystallization kinetics were modulated and grain-boundary defects were passivated. Consequently, non-radiative recombination was suppressed and 28.47% efficiency was achieved in perovskite/silicon TSCs [30]. Overall, dual-functional molecular passivation is superior to single additives, repairing both bulk and grain-boundary defects. 2D/3D heterojunction passivation delivers the best long-term stability yet adds extra processing steps, limiting low-cost production.

Interface reconstruction and structured passivation: This category addresses both surface and interfacial defects through structural modification, including surface reconstruction, low-dimensional phase induction, and targeted surface coating, all aimed at forming a protective or passivating layer that eliminates defect sites. Wet chemical nano-polishing was employed to remove the defect-rich surface layer, and subsequent passivation with piperazinium iodide was applied, through which defects were eliminated, lattice strain was released, and interface energy alignment was optimized, thereby improving the perovskite/electron transport layer (ETL) interface (Figure 3a) [31]. By passivating bromine vacancies via PEABr surface coating, non-radiative recombination was suppressed, enabling 31% PCE in TSCs [32]. 4M-PEACI was used to induce a quasi-2D phase at grain boundaries, enabling regional defect encapsulation, which yielded 23.7% steady-state efficiency and 30.3% for 4T perovskite–silicon TSCs [33]. DMPESI reacted with PbI₂ to form a 1D DMPESPbI₃ phase, which accumulated at interfaces as a hydrophobic barrier, thereby suppressing moisture-induced decomposition (Figure 3f–g) [34].

Synergistic passivation strategies: These strategies integrate multiple mechanisms for comprehensive performance improvement. A KSCN-MAI synergistic strategy was employed, with K⁺ passivating defects, SCN⁻ enlarging grains, and MAI removing residual SCN⁻ via double displacement to suppress phase segregation, yielding 15.0% (single-junction) and 26.4% (triple-junction) PCEs (Figure 3j) [35]. Chen et al. adopted one-step post-treatment with mixed OAmI and GACl to simultaneously generate RP and ACI 2D perovskite phases, constructing gradient 2D/3D heterojunctions as described in the original work (Figure 3c). Macromolecular OAm⁺ cations only remain on the surface of the perovskite film to form Ruddlesden–Popper (RP) phases, while small GA⁺ ions can penetrate into the film interior at a depth of 150–300 nm to produce Alternating-Cation-Interlayer (ACI) phases. The content of the two phases gradually changes along the film thickness, enabling simultaneous passivation of both surface and bulk defects [36]. Zhang et al. incorporated MFC into the antisolvent. MFC formed strong hydrogen bonds with the perovskite, slowing crystallization and passivating defects, which enlarged grains and induced an n-type surface, yielding 30.65% PCE and 1.951 V V_oc with enhanced stability (Figure 3i) [37].

Process-integrated passivation: Integrating passivation into fabrication processes has emerged as an efficient pathway. Vacuum-assisted solution treatment and GuaBr interface modification were combined to coordinately manage bulk and surface defects, yielding 19.67% PCE (1.22 V V_oc) and 24.01% for TSCs [38]. Z907 was added into the antisolvent to form a self-assembled monolayer modulating crystallization (Figure 3a). -SCN/-COOH groups passivate defects and align interfacial energy, while hydrophobic chains enhance moisture resistance [39]. Post-treatment with FAI/mF-PEAI mixed solution was applied to slot-die-coated films, passivating iodine vacancies, reducing non-radiative recombination, and removing excess PbI₂ to optimize charge transport, raising top-cell PCE from 16.14% to 20.79% [40]. Perovskite and CuSCN were co-deposited on textured silicon wafers. Cross-sectional TEM-EDS mapping (Figure 3e) directly demonstrates CuSCN enrichment at perovskite buried grain boundaries to build local hole-collecting contacts, while density functional theory (DFT) interfacial models (Figure 3d) confirm strong binding between CuSCN and uncoordinated Pb/I defects for suppressed non-radiative recombination and accelerated hole transport. This design delivers a certified efficiency of 31.46%, with 93.8% initial efficiency maintained after 1200 h (maximum power point) MPP tracking [41]. Table 2 summarizes the defect and interface passivation strategies for high-performance perovskite/silicon TSCs.

Table 2. Defect and interface passivation strategies and the PV parameters for high-performance perovskite/silicon TSCs.

Device Architecture	Active Area (cm2)	PCE (%)	Jsc (mA/cm2)	Voc (V)	FF (%)	Stability	Year [Ref.]
ITO/2PACz/Cs0.05FA0.8MA0.15Pb(I0.75Br0.25)3(1.68 eV)/LiF/C60/BCP/Ag n-Si/α-Si(i)/α-Si(n+)/α-Si(p+)/ITO/Ag	1.03	28.2	19.6	1.859	78.2	Maintained >93% of its initial PCE after outdoor operation for 43 days in a hot desert climate.	2021 [26]
ITO/SnO2/1.77-eV mixed-halide perovskite/spiro-OMeTAD/MoOX/IZO/Ag/MgF2 ITO/a-Si:H(p+)/a-Si:H(i)/c-Si(n)/a-Si:H(i)/a-Si:H(n+)/ITO/Ag	/	30.24	18.82	1.24	79.58	Maintained >90% of its initial PCE after 500 h of continuous MPPT under simulated illumination in a N2 atmosphere.	2022 [27]
ITO/PTAA/FA0.85MA0.2Cs0.15Pb(I0.8Br0.2) with 2D PEA(I0.25SCN0.75)/C60/PEIE/ITO c-Si(n)/a-Si:H(i)/a-Si:H(p+)/ITO	0.188	26.7	19.2	1.756	79.2	Retained >80% of its initial PCE after 1000 h of continuous illumination in a N2 atmosphere (unencapsulated).	2020 [28]
ITO/NiOX/Me-4PACz/FA0.8Cs0.2Pb(I0.8Br0.2)3 + BD-Boc/EDAI2/C60/SnO2/IZO/Ag n-FZ-Si(textured)/a-Si:H(i)/a-Si:H(n)/a-Si:H(p)/ITO/Ag	1.00	28.47	19.07	1.87	79.48	Retained >92% of its initial PCE after 2000 h of continuous 1-sun illumination in a N2 atmosphere.	2025 [30]
ITO/NiOx/Me-4PACz/Cs0.05FA0.80MA0.15Pb(I0.75Br0.25)3/C60/SnO2/IZO/Ag grid/MgF2Glass/MgF2/a-Si:H(i)/c-Si(n)/POLO-SiO2(n+)/a-Si:H(i)/µc-Si:H(p+)/ITO/Ag	1.0	20.39	20.54	1.23	80.65	Maintained 80% of its initial PCE under continuous MPPT operation in N2 for 1505 h.	2024 [31]
FTO/c-TiO2/SnO2/Cs0.17FA0.82Pb(I0.75Br0.25)3/PEABr/PTAA/VOx/ITO	1.05	31.0	19.67	1.94	81.2	Maintained >80% of its initial PCE after 1000 h of thermal aging at 85 °C (T80).	2025 [32]
Glass/ITO/SnO2/MA-free perovskite (FAPbI3+CsCl+4M-PEACI)/spiro-MeOTAD/MoO3/IZO/Au FZ n-Si (PERL)	4.0	30.3	24.0	1.131	0.842	Retained >93% of its initial PCE after 1000 h of continuous operation under 1-sun illumination (white LED) in a N2 atmosphere (cell temperature maintained at 25 °C).	2023 [33]
Ag/MgF2/IZO/SnO2/C60/FA0.8Cs0.2Pb(I0.8Br0.2)3(1.67 eV)/NiOX/SAMS/ITO/nc-Si:H(n+/p+)/a-Si:H(i/n)/c-Si/a-Si:H(i)/nc-Si:H(p)/ITO/Ag	1.21	29.8	20.33	1.895	79.13	Maintained~84% (MPP tracking, ~1800 h) and ~80% (damp heat, 85 °C/85% RH, 723 h) of the initial PCE.	2025 [34]
ITO/2PACz/CsFA(Br,I)3/C60/SnO2/IZO/NiOX/MeO-2PACz/RbCsFAPbI3/C60/SnO2/IZO/a-Si:H(n,i,p)/c-Si(n)/a-Si:H(i,p)/IZRO/Ag	1.00	26.4	11.9	3.04	72.9	88.5% after 1000 h at 85 °C	2024 [35]
ITO/PTAA/Cs0.235FA0.768Pb(I0.862Br0.133)3/C60/BCP/Ag ITO/p+a-Si:H/ia-Si:H/c-Si(n)/i a-Si:H/n+ a-Si:H/ITO/Ag	1.00	30.65	20.54	1.899	80.59	Retained 98% of its initial PCE after 1000 h of damp-heat testing at 85 °C/85% RH.	2025 [36]
ITO/HTL/Perovskite/C60/SnO2/IZO/Ag Si (n-type c-Si)	1.05	30.65	1.951	18.96	82.82	Retained 97.3% of its initial PCE after 2000 h of storage in a N2 atmosphere at room temperature.	2025 [37]
ITO/SnO2/Cs0.15FA0.85Pb(I0.7Br0.3)3(GuaBr)/spiro-OMeTAD/MoOX/IZOa-Si(p/i)/c-Si(n)/a-Si(i/n)/Ag	0.509	24.01	17.33	1.92	72.3	Maintained 80% of its initial PCE after 1000 h of MPP tracking at 25 °C in an argon atmosphere (unencapsulated).	2021 [38]
ITO/SnO2/Cs0.22FA0.78Pb(I0.85Br0.15)3/spiro-OMeTAD/MoOX/IZO/Ag ITO/a-Si:H(i)/c-Si(n)/a-Si:H(i)/a-Si:H(n+)/ITO/Ag	0.09	32.07	21.08	1.833	83.0	Retained 97.2% of its initial PCE after 1000 h at 25 °C/30–40% RH; maintained >95.5% and >92.6% after 10 h at 85 °C and 100 °C, respectively.	2024 [39]
ITO/CuSCN-embedded(Cs0.05FA0.8MA0.13PbI2.85Br0.15)/MgFX/SnO2/C60/IZO/Ag Ag/ITO/nc-Si:H(p)/a-Si:H(i)/n-Si wafer/a-Si:H(i)/nc-SiOX:H(n)/ITO	1.0	31.46	20.6	1.89	77	Maintained 93.8% of its initial PCE after 1200 h of maximum power-point tracking at 45 °C; maintained 90.2% after >1000 h of damp-heat testing at 85 °C and 85% RH.	2024 [41]

Figure 3. (a) Schematic diagram of the changes in perovskite film before and after nano-polishing treatment. Reprint with permission [31]. Copyright 2024, Springer Nature. (b) Schematic illustration of interaction of carbazole with perovskite films. Reprint with permission [26]. Copyright 2021, Elsevier. (c) Schematic illustration of B-2D strategy. Reprint with permission [36]. Copyright 2024, Springer Nature. (d) Interfacial models of DFT calculations for CuSCN and FAIor PbI₂-rich FAPbI₃ (001) surfaces. Reprint with permission [41]. Copyright 2024, Springer Nature. (e) Cross-sectional high-angle annular dark-field TEM image. Reprint with permission [41]. Copyright 2024, Springer Nature. (f) Schematic of the hybrid two-step deposition method. Reprint with permission [34]. Copyright 2025, American Chemical Society. (g) Schematic diagram of the dynamic evolution of DMPESI molecules throughout the perovskite conversion pathway. Reprint with permission [34]. Copyright 2025, American Chemical Society. (h) Experiment procedures for the Z907 SAM-capped crystallization strategy. Reprint with permission [39]. Copyright 2024, Elsevier. (i) Schematic diagram of the interaction between MFC and the perovskite. Reprint with permission [37]. Copyright 2025, American Chemical Society. (j) Schematic diagram of the working mechanism of KSCN + MAI additive engineering. Reprint with permission [35]. Copyright 2024, Elsevier.

Figure 3. (a) Schematic diagram of the changes in perovskite film before and after nano-polishing treatment. Reprint with permission [31]. Copyright 2024, Springer Nature. (b) Schematic illustration of interaction of carbazole with perovskite films. Reprint with permission [26]. Copyright 2021, Elsevier. (c) Schematic illustration of B-2D strategy. Reprint with permission [36]. Copyright 2024, Springer Nature. (d) Interfacial models of DFT calculations for CuSCN and FAIor PbI₂-rich FAPbI₃ (001) surfaces. Reprint with permission [41]. Copyright 2024, Springer Nature. (e) Cross-sectional high-angle annular dark-field TEM image. Reprint with permission [41]. Copyright 2024, Springer Nature. (f) Schematic of the hybrid two-step deposition method. Reprint with permission [34]. Copyright 2025, American Chemical Society. (g) Schematic diagram of the dynamic evolution of DMPESI molecules throughout the perovskite conversion pathway. Reprint with permission [34]. Copyright 2025, American Chemical Society. (h) Experiment procedures for the Z907 SAM-capped crystallization strategy. Reprint with permission [39]. Copyright 2024, Elsevier. (i) Schematic diagram of the interaction between MFC and the perovskite. Reprint with permission [37]. Copyright 2025, American Chemical Society. (j) Schematic diagram of the working mechanism of KSCN + MAI additive engineering. Reprint with permission [35]. Copyright 2024, Elsevier.

Deposition and crystallization control: Effective passivation only alleviates defects. The fundamental issue of poor film quality (small grains, pinholes, and uneven crystallization) on textured silicon substrates requires advanced deposition and crystallization control. Thus, process control to obtain high-quality, uniform perovskite films is essential for high-efficiency tandem devices, as it addresses the root cause and enables scalable fabrication. To achieve high-efficiency perovskite/silicon TSCs, scalable deposition techniques were developed for textured silicon substrates. Solvent engineering with nitrogen-knife-assisted blade coating enables void-free filling (Figure 4a), while slot-die coating with DMF-NMP optimized annealing reduces PbI₂ aggregation, improving carrier lifetime and quasi-Fermi splitting to 25.2%. A two-stage vacuum flash method suppresses pinholes by sequential DMSO/DMF removal under controlled vacuum, yielding pinhole-free films and 22.5% PCE. For low-pressure CVD, Cs⁺ ions spontaneously migrate to achieve uniform vertical distribution regardless of precursor order. CsBr-first deposition accelerates conversion and widens the bandgap, enhancing charge transfer, leading to 18.7% (semi-transparent) and 26.9% (monolithic tandem) PCEs [42,43,44,45].

Additive design or template-guided growth was employed to optimize film morphology and crystallinity. Zhang et al. introduced FAI and RbAc into the PbI₂/PbBr₂ precursor to form a lead halide complex that regulates crystallization and promotes (100)-oriented grain growth (Figure 4c) [46]. Ballif et al. used a hybrid evaporation–spin-coating process. Co-evaporated PbI₂/CsBr formed an inorganic template, followed by spin coating of FAI/MAI/MACl, inducing preferential (001) orientation with ~1 μm grains and reducing non-radiative recombination losses by 120 mV [47]. Zhang et al. co-evaporated CsCl with PbI₂ to generate a δ-Cs_0.3Pb₄I₈ intermediate phase that acts as a template for preferential black-phase growth, yielding denser films and 29.22% PCE (Figure 4e) [48]. Qiao et al. combined Pb(SCN)₂ additive with air annealing, where Pb(SCN)₂ promotes grain growth and air annealing, removes PbI₂ impurities, and enhances crystallinity via moisture-assisted Ostwald ripening [49]. Chu et al. proposed DMSO-mediated solution aging with pure DMSO, nitrogen quenching, and Marangoni reflux to slow crystallization, producing uniform, low-stress, low-roughness WBG films that suppress halide segregation and improve transparency and carrier transport (Figure 4b) [50]. For crystal-nucleus control, Chen et al. employed OAmI-mediated nucleation with vacuum quenching to promote 3C cubic-phase nucleation and suppress Br-rich and 2H hexagonal phases, yielding 32.5% tandem PCE [51]. Schulze et al. introduced urea to lower the nucleation barrier for α-phase growth at low temperature, while excess urea migrated to the surface to passivate the perovskite/C₆₀ interface, achieving 30.0% [52]. Xu et al. introduced PbCl₂ via triple-source co-evaporation to form a longitudinal homogenized intermediate that slows crystallization and prevents capping-layer formation, enabling uniform interdiffusion under ambient air (Figure 4d) [53]. Table 3 summarizes the deposition and crystallization control for high-quality perovskite films. Vacuum evaporation yields large low-defect grains but costs more, while blade coating balances performance and scalability for industrialization. Additive-induced oriented growth is the most accessible low-cost route to improve film quality.

Table 3. Deposition and crystallization control and the PV performance for high-quality perovskite films TSCs.

Device Architecture	Active Area (cm2)	PCE (%)	Jsc (mA/cm2)	Voc (V)	FF (%)	Stability	Year [Ref.]
PDMS/ITO/SnO2/C60/Cs0.1MA0.9Pb(I0.9Br0.1)3/PTAA a-Si:H(n)/a-Si:H(i)/c-Si(n)/a-Si:H(i)/a-Si:H(p)/ITO/SiO2/Ag	0.4225	26.2	19.2	1.82	75.3	Retained 92% of its initial PCE after 100 h under constant illumination in ambient conditions without encapsulation.	2020 [42]
Glass/ITO/SAM(2PACz)Cs0.22FA0.78Pb(I0.85Br0.15)3 + 5 mol% MAPbCl3/C60/LiF/SnO2/IZO//n-Cz-Si(SDE)/a-Si:H(i)/a-Si:H(p+)/TCO/Ag	1.0	25.2	18.5	1.87	73.1	Stable MPP output over 300 s.	2022 [43]
ITO/NiOx/MeO-2PACz/Cs1−XFAXPb(IγBr1−γ)3(CVD)/C60/ALD-SnO2/ITO TOPCon c-Si(n)	1.00	26.9	20.2	1.71	0.78	Retained ~80% of its initial PCE after 200 h of continuous maximum power-point tracking under AM 1.5G illumination (100 mW cm−2) in ambient atmosphere.	2025 [45]
ITO/PTAA/(FAI+RbAc)-modified perovskite (1.63 eV)/PCBM/BCP/Cu (n-type) c-Si/(i) a-Si:H/(p+) a-Si:H/ITO	0.5003	27.64	18.99	1.83	79.46	Maintained >70% of its initial PCE after 135 d in ambient storage (RH > 40%) without encapsulation.	2022 [46]
MgFx/Ag grid/IZrO/SnOx/C60/Perovskite (FAPbI3-based)/SAM (MeO-2PACz)/ITO/(SHJ bottom cell) textured Si with a-Si:H(i)/a-Si:H(n)/nc-Si:H(n) on front and a-Si:H(i)/nc-Si:H(p) on rear/ITO/Ag	1.21	24.68	17.53	1.714	82.3	Maintained 60% of initial PCE after 288 h under Voc at 45 °C in N2 without encapsulation.	2024 [47]
ITO/spiro-TTB/CsCl-passivation/PbI2 + CsCl + FAI + FABr + MACI/C60/ZnO/SnO2(ALD)/IZO/MgF2 Ag/ITO/a-Si:H(n+)/a-Si:H(i)/c-Si(n)/a-Si:H(i)/a-Si:H(p+)/ITO	0.5003	29.22	19.72	1.857	79.80	Maintained 80% of its initial efficiency after 270 h of maximum power point (MPP) tracking under continuous 1-sun illumination in a N2 atmosphere (<0.001% RH, 25 ± 1 °C) without encapsulation.	2024 [48]
ITO/SnO2/Cs0.22FA0.78Pb(I0.85Br0.15)3/spiro-OMeTAD/MoOx/IZO/Ag grid TOPCon/c-Si(n)	0.09	30.32	21.22 40.56	1.199 0.66	0.82 0.83	Maintained ~95% of its initial PCE after being stored in ambient air with ~40% RH.	2023 [49]
ITO/NiOx/SAM/FA0.6Cs0.3DMA0.1Pb(I0.7Br0.3)3/PCBM/AZO/ALD-SnO2/IZO ITO/a-Si:H(p+/i)/c-Si(n)/a-Si:H(i/n+)/ITO/Ag	0.1017	28.09	17.90 17.13	1.24 1.27	82.07 81.01	The unencapsulated device retained 95% of the initial PCE after 900 h of continuous illumination and MPP tracking at 55 °C in a N2 environment.	2024 [50]
ITO/2PACzCs0.14FA0.86Pb(I0.76Br0.22)3/C60/SnOx/ITO/Ag/ARC Ag/ITO/a-Si:H(i/n+)/c-Si(n)/a-Si:H(i/p+)/ITO	1.00	30.0	~19.8	~19.2	~78.9	Retained a comparable performance over 6000 h of storage in a dark N2-rich atmosphere.	2024 [53]
ITO/NiOx/Me-4PACz/LHI perovskite/C60/SnO2/IZO/Ag grid textured Si(n)/a-Si:H(i)/a-Si:H(p+)/ITO/Ag	1.002	28.55	19.37	1.88	78.40	Maintained 92% of its initial PCE after 1200 h of continuous MPP tracking under 1-sun illumination at ~40 °C.	2024 [53]

Figure 4. (a) Mechanism of dense perovskite film formation. Reprint with permission [42]. Copyright 2020, Elsevier. (b) The preparation of WBG perovskite films. Reprint with permission [50]. Copyright 2024, Wiley-VCH GmbH. (c) PbI₂/PbBr₂ crystals and control perovskite film. Reprint with permission [46]. Copyright 2022, American Chemical Society. (d) Schematic of the LHI process. Reprint with permission [53]. Copyright 2024, American Chemical Society. (e) Preparation process of cesium salt/PbI₂-based perovskite thin films. Reprint with permission [48]. Copyright 2024, Elsevier. (f) The schematic illustration of the low-pressure vapor deposition process for the perovskite film. Reprint with permission [45]. Copyright 2025, Wiley-VCH GmbH.

Figure 4. (a) Mechanism of dense perovskite film formation. Reprint with permission [42]. Copyright 2020, Elsevier. (b) The preparation of WBG perovskite films. Reprint with permission [50]. Copyright 2024, Wiley-VCH GmbH. (c) PbI₂/PbBr₂ crystals and control perovskite film. Reprint with permission [46]. Copyright 2022, American Chemical Society. (d) Schematic of the LHI process. Reprint with permission [53]. Copyright 2024, American Chemical Society. (e) Preparation process of cesium salt/PbI₂-based perovskite thin films. Reprint with permission [48]. Copyright 2024, Elsevier. (f) The schematic illustration of the low-pressure vapor deposition process for the perovskite film. Reprint with permission [45]. Copyright 2025, Wiley-VCH GmbH.

Optical management and current matching: Early strategies included nanophotonic tuning of perovskite thickness on textured silicon (optimal pyramid period, 450 nm) to match current density (20.6 mA/cm²), and bifacial design considering albedo, requiring thicker perovskite (0.55 μm) and 1.65–1.75 eV bandgap for matching, yielding >31.9% PCE [54,55]. Innovative electrode structures and bifacial illumination were employed. A quasi-finger-back-contact (QFBC) electrode (period, 750 nm; width ratio, 0.4) acted as a diffraction element to enhance short-wavelength absorption and transmit long-wavelength light, reducing series resistance and eliminating shading losses, thereby enabling 29.65% four-terminal PCE [56]. Furthermore, current-matching loss (CML) was quantified under dual-side illumination. Optimizing bandgap (~1.65 eV) and albedo minimized CML, achieving 1692 kW·h/m² annual generation, 12.07% higher than single-sided devices [57].

Theoretical simulations guided parameter optimization. DFT and SCAPS-1D were used to determine CsGeI₃’s suitable 1.6 eV bandgap and optimal thickness (265 nm), while controlling bulk/interface defects to minimize recombination, yielding 28.43% tandem PCE (Figure 5c) [58]. Similarly, SCAPS-1D optimized a tin-based perovskite (GA_0.06(FA_0.8Cs_0.2)_0.94SnI₂Br) by adjusting thickness and defect density and selecting CTLs to suppress non-radiative recombination, achieving 32.12% (Figure 5f) [59]. Coordinated regulation of bandgap and thickness (e.g., 1.7 eV bandgap with increased thickness) enabled current matching and 32.71% efficiency [60]. Finally, a multiscale optical model with a genetic algorithm tuned photonic crystal parameters (lattice constants and side-wall angles) and perovskite thickness (465.5 nm) to enhance light trapping, achieving 32.07% efficiency, surpassing the original thick-device configuration (Figure 5e) [61]. Bifacial optical design optimizes current matching better than single-sided light trapping. Numerical simulation-guided bandgap-thickness tuning achieves theoretical current balance and higher energy output.

Figure 5. (a,b) Influence of the period of the surface texture. Reprint with permission [54]. Copyright 2018, The Royal Society of Chemistry. (c) Schematic structure of CsGeI₃-on-Si TSC (right). Reprint with permission [58]. Copyright 2024, Elsevier. (d) J_sc of top and bottom sub-cells. Reprint with permission [60]. Copyright 2024, Elsevier. (e) Variation in the perovskite layer thickness and maximum achievable photocurrent density (JMAPD). Reprint with permission [61]. Copyright 2025, Elsevier. (f) The energy-level alignment of the absorber layer. Reprint with permission [59]. Copyright 2024, Elsevier.

3.1.2. Interface-Regulation Strategies

Interface-regulation strategies were developed to suppress non-radiative recombination and optimize carrier extraction. A hybrid two-step deposition with an nc-Si:H recombination junction improved shunt resistance, achieving 25.2% [62]. NiI₂ was used to fill iodine vacancies via stronger Ni-I bonds, reducing interfacial recombination [63]. An ultrathin ITO interlayer (1.7 nm) enabled efficient carrier recombination and suppressed shunting, yielding 27.2% PCE [64]. LiF (2 nm) acted as a physical barrier and improved wettability, leading to 25.1% on 24 cm² (Figure 6c) [65]. Mesoporous Al₂O₃ regulated ion distribution and converted the buried interface to p-type, forming a homojunction that enhanced the built-in field, giving 29.04% PCE [66]. A NiO_x/MeO-2PACz hybrid improved SAM uniformity and energy alignment, resulting in 28.47% PCE with 81.8% FF (Figure 6d) [67]. A dual-layer LiF + EDAI combined field-effect and chemical passivation, enabling a certified PCE of 33.89% [68]. Oblique-angle deposited SiO_x provided chemical (Pb=O) and field-effect passivation via an n/n⁺ homojunction, achieving 30.2% PCE [69]. AZO nanoparticles replaced atomic layer deposition (ALD), protecting the perovskite and promoting charge transport, yielding 25.3% PCE [70].

Vacancy defects and grain boundary traps within the perovskite layer, as well as energy-level offsets between perovskite and transport layers/silicon substrates, often lead to severe non-radiative recombination and hindered carrier extraction, directly limiting device efficiency and stability. To address this, research teams have developed diversified strategies, ranging from self-assembled monolayers (SAMs) and multifunctional additives to gradient interface modification, centered on two core goals: “chemical passivation of defects” and “physical optimization of energy bands.” By precisely regulating the interfacial chemical state and electronic structure, these strategies achieve the dual effects of reduced defect density and optimized energy-level alignment. Me-4PACz SAM optimized hole extraction and passivation, certifying 29.15% PCE [71]. Two studies focused on the passivation effect of phenformin hydrochloride (PhenHCl). PhenHCl is the hydrochloride salt of phenformin, a biguanide derivative containing multiple nitrogen-containing functional groups. Its hydrophilic biguanide moiety provides abundant coordination sites, while the phenethyl group introduces moderate hydrophobic characteristics, enabling interactions with both perovskite defects and film surfaces. In combination with Me-4PACz SAM, PhenHCl provides complementary defect regulation. Me-4PACz forms a well-organized monolayer on the ITO substrate to facilitate hole extraction and energy-level alignment, whereas PhenHCl mainly targets bulk and surface defects through coordination with undercoordinated Pb²⁺ and hydrogen-bond interactions with halide species. This synergistic strategy enables more comprehensive defect passivation from the buried interface to the perovskite surface. On the one hand, the innovation in perovskite layer regulation in this study lies in the use of PhenHCl to achieve synergistic passivation of cationic and anionic defects. PhenHCl has both electron-rich (amino and imino) and electron-deficient (ammonium) domains. When added to the precursor solution, it enables grain boundary passivation (GBP). When spin-coated on the surface, it realizes top surface passivation (TSP). This adaptability to perovskites with different surface chemical compositions reduces non-radiative recombination, while suppressing photoinduced halide segregation in WBG mixed-halide perovskites (Figure 6f). Additionally, it is combined with 2PACz SAM to optimize hole contact. Based on this strategy, the efficiency of the single-junction PSC reached 20.5%, and the final efficiency of the monolithic tandem cell integrated with textured silicon solar cell was 27.4% [72].

On the other hand, a study designed and applied the multifunctional molecule PhenHCl to simultaneously passivate Pb²⁺ cationic and I⁻/Br⁻ anionic defects in the perovskite layer. Confirmed via DFT calculations, PhenHCl contains electron-rich amine/imine groups and electron-deficient ammonium groups, allowing it to adaptively bind to and neutralize unsaturated bonds at different chemical terminations on the perovskite surface. This overcomes the limitation of traditional single-functional molecules (e.g., PEAI), which exhibit poor passivation performance on specific lead-rich surfaces. In experiments, this molecule was added to the perovskite precursor solution to achieve GBP and used as a post-treatment coating to realize TSP. This synergistic approach can comprehensively suppress non-radiative recombination and defect-driven halide phase separation. Results show that for the 1.68 eV WBG perovskite top cells passivated by PhenHCl, the V_oc increased by an average of approximately 100 mV, with single-junction efficiency exceeding 20%. When this strategy was applied to monolithic tandem devices with textured silicon bottom cells, the average efficiency of the cells increased from 23.3% to 26.5%, and the certified efficiency of the champion device reached 27.4% [73]. TEABr, with its conjugated structure, enabled conductive passivation without sacrificing FF, giving 28.51% PCE [74]. MASCN + PEAI induced grain regrowth and double-sided passivation, with PEAI penetrating to the buried interface, achieving 29.8% PCE [75].

In addition, Albrecht et al. significantly improved the performance of perovskite/silicon TSCs by introducing piperazinium iodide (PI) as an interfacial modification molecule between the perovskite and the ETL. The team spin-coated this PI molecular solution onto a WBG (1.68 eV) triple-halide perovskite absorber layer. The aim was to optimize energy-band alignment by forming an interfacial dipole, rather than relying on traditional chemical passivation. Through characterizations such as photoelectron spectroscopy, the study confirmed that PI induced a positive dipole on the perovskite surface. This effectively increased the work function of the C₆₀ ETL, thereby significantly reducing the energy-level offset between the perovskite conduction band minimum and the C₆₀ lowest unoccupied molecular orbital, from −0.31 eV (unfavorable for electron extraction) to +0.04 eV (nearly ideal). This optimized band alignment reduced non-radiative recombination losses at the interface and accelerated electron extraction. The results showed that the tandem cell modified with the PI interface achieved a PCE of up to 32.5% [76]. FPEAI passivated grain boundaries and increased the iodide migration barrier, suppressing phase separation and yielding 29.57% PCE (Figure 6a) [77]. pFBPA combined with SiO₂ nanoparticles to passivate defects and repair pinholes, certifying 30.9% PCE [78]. A C₆₀ SAM + F-PEABr composite improved conductivity, passivation, and inhibited ion migration, giving 27% PCE (Figure 6b) [79].

On the other hand, Fang’s research team constructed a 2D/3D heterojunction passivation layer on the surface of WBG perovskite using isopropylammonium iodide (i-PAI), which effectively improved the interface quality and optoelectronic performance of the perovskite top layer. They utilized i-PAI post-treatment to form a (i-PA)₂PbI₄ 2D phase on the surface of 1.67 eV WBG perovskite, aiming to suppress halide phase separation, reduce interface defect density, and enhance the binding energy within the perovskite layer. This passivation strategy increased grain size, prolonged carrier lifetime, and reduced trap state density. Finally, the WBG perovskite single-junction solar cell achieved a PCE of 22.4%, and the four-terminal perovskite/silicon TSC obtained a champion PCE of 31.1% [80]. Meanwhile, Qin et al. designed and synthesized SAMs with open-shell diradical properties (diradical SAMs) as hole-selective contacts. They constructed two types of diradical SAMs, RS-1 and RS-2, using a donor–acceptor coplanar conjugation strategy. Leveraging the diradical properties, these SAMs provide additional unpaired electrons and favorable energy-level orbitals to enhance hole-transport capability. Meanwhile, the introduction of sterically hindered groups improves molecular stability and uniformity in large-area solution processing (Figure 6h). Precise characterization via scanning electrochemical cell microscopy–thin layer cyclic voltammetry (SECCM-TLCV) showed that the perovskite single-junction solar cell based on the optimal diradical SAM (RS-2) achieved an efficiency of 26.3%, the mini-module (10.04 cm²) reached an efficiency of 23.6%, and the perovskite/silicon TSC obtained a certified PCE as high as 34.2% [81].

The composition of perovskite (such as cation ratio and halogen component) directly determines its bandgap width, crystallization quality, and ionic stability. Interface modification, however, needs to adapt to the surface chemical state of specific compositions to efficiently address defect recombination and carrier transport issues. Single-component regulation is prone to limitations such as interfacial energy-level mismatch or residual defects, while pure interface modification can hardly compensate for the impact of bulk crystallization defects in perovskite. Therefore, research teams have precisely designed perovskite cation/halogen components to optimize bandgap and crystallinity, and simultaneously paired them with targeted interface passivation and defect repair strategies, achieving the dual goal of “bulk performance improvement and interfacial loss reduction.” Me-4PACz doped in Al₂O₃ improved wettability and coverage, increasing tandem PCE to 30.92% [82]. CuSCN replaced PEDOT:PSS to reduce parasitic absorption, with AgNWs optimizing transmittance/conductivity, yielding 26.7% and 25.2% PCEs in 4T TSCs [83]. NiO_x/SAM composite improved wettability and energy alignment, giving 27.6% PCE [84]. 4PADCB spontaneously migrated to form a hole-transport layer (HTL), optimizing energy levels and achieving 31.71% certified PCE (Figure 6e) [85]. SiN_x reduced reflection, and Cr/Ag/Sn ohmic contacts lowered resistance, yielding 23.11% PCE [86]. Hierarchical micro–nano-textures guided conformal coating via capillary forces, delivering 30.1% with 84.5% FF [87]. On the other hand, a study adopted an innovative interface engineering strategy: in p-i-n structured devices, a thin 2D perovskite layer (PEA₂MA_n−1PbnI_3n+1, with energy-level alignment optimized by adjusting the dimensionality n value) was inserted between the perovskite absorption layer and the HTL. Meanwhile, blade-coating and gas-quenching techniques were combined to achieve rapid crystallization of 3D perovskite. This approach aimed to reduce interface defects and improve energy-level matching, thereby enhancing hole-extraction efficiency and suppressing non-radiative recombination (Figure 6g). Results showed that the optimized 2D-3 (n = 3) underlayer structure boosted the efficiency of single-junction PSCs to 22.6%. For perovskite/silicon TSCs, this interface-regulation strategy enabled the device to achieve a certified efficiency of 31.2% [88]. Table 4 summarizes the interface engineering strategies and the photovoltaic performance of the corresponding TSCs. SAM-based interface modification achieves optimal energy alignment with minimal loss. Multilayer composite interlayers further cut recombination, yet overly complex structures raise series resistance and fabrication difficulty.

Table 4. Interface engineering strategies and the PV performance of the corresponding TSCs.

Device Architecture	Active Area (cm2)	PCE (%)	Jsc (mA/cm2)	Voc (V)	FF (%)	Stability	Year [Ref.]
MgF2/Ag grid/IZO/SnO2/C60/LiF/CsxFA1−xPb(I,Br)3 (1.6 eV perovskite, 440 nm, hybrid co-evaporation/spin coating)/spiro-TTB/nc-Si:H(p+)/nc-Si:H(n+)/a-Si:H(n)/intrinsic a-Si:H/double-side textured n-type c-Si/intrinsic a-Si:H/a-Si:H(p)/ITO/Ag	1.42	25.24	19.53	1.78	73.1	Retains 91.7% of initial PCE after 61 h for unencapsulated devices under 1-sun MPP tracking at 25 °C and 20–30% RH, with a linear degradation rate of −0.15% rel/h.	2018 [62]
PDMS/Au grid/ITO/MoO3/spiro-OMeTAD/CsPbI3−xBrx (1.80 eV perovskite, NiI2-passivated)/SnO2 (CoCl2-doped)/1.7 nm ultra-thin ITO/(p)a-Si:H/intrinsic a-Si:H/n-Si/a-Si:H(n+)/ITO/Ag	0.5003	22.95	14.34	2.04	78.37	Encapsulated device retained 95.7% initial efficiency after 300 h MPPT (1-sun illumination; N2 atmosphere).	2022 [63]
PDMS/Ag grid/ITO/MoO3/spiro-OMeTAD/(FAPbI3)0.83(MAPbBr3)0.17 perovskite (480 nm)/SnO2/1.7 nm ultra-thin ITO interlayer/(p)a-Si:H/intrinsic a-Si:H/n-Si/a-Si:H(n+)/ITO/Ag	1.0	27.2	18.1	1.819	82.4	Retains 98% of initial efficiency after 600 h MPPT for encapsulated devices under 1-sun illumination, 60 ± 20% RH. and room temperature.	2023 [64]
PDMS/Ag/ITO/SnO2/C60/Cs0.1FA0.2MA0.7Pb(I0.83Br0.17)3 (WBG perovskite)/LiF/PTAA/ITO/a-Si:H(n+)/intrinsic a-Si:H/c-Si(n)/intrinsic a-Si:H/a-Si:H(p+)/ITO/silver	24	25.1	18.1	1.89	73.6	Retained 91% initial PCE after 300 h MPP operation (encapsulated, 25 °C; 1-sun illumination).	2023 [65]
Top cell: Silicon heterojunction substrate/NiOx/Me-4PACz SAM/mesoporous Al2O3/Cs0.1MA0.25FA0.65Pb(I0.8Br0.2)3 (1.67 eV perovskite, 1.2 μm)/C60/ALD SnO2/IZO. Bottom cell: n-type c-Si/intrinsic a-Si:H/p-type a-Si:H/ITO/Ag. Recombination junction: ITO. Encapsulation: MgF2/Ag grid	1.00	29.04	19.68	1.78	83.0	Retains >85% of initial PCE after 1200 h for unencapsulated devices under 1-sun illumination, demonstrating good operational stability.	2024 [66]
Ag/ITO/a-Si:H(p+)/intrinsic a-Si:H/c-Si(n)/intrinsic a-Si:H/a-Si:H(n+)/ITO/NiOx/MeO-2PACz (SAM)/CsFAMAPb(I,Br)3 (1.71 eV perovskite)/PEAI/C60/SnO2/IZO/Ag	1	28.47	18.39	1.87	81.8	Retained 77% initial PCE after 620 h MPPT (encapsulated, ambient air, 40–50% RH, and 1-sun illumination).	2024 [67]
MgF2/Ag grid/IZO/SnO2/C60/LiF/EDAI/FA0.8MA0.15Cs0.05Pb(I0.76Br0.24)3 (1.69 eV perovskite, 1.6 M, 5% excess PbX2)/MeO-4PACz SAM/IZO/nc-SiOx:H(n)/intrinsic a-Si:H/n-type CZ c-Si/intrinsic a-Si:H/nc-Si:H(p)/IZO/Ag; asymmetric texture	1.0	33.89	20.76	1.966	83.0	Retains ~80% of initial PCE after 1200 h 1-sun MPPT for encapsulated devices in nitrogen atmosphere at room temperature.	2024 [68]
Ag/IZO/BCP:Ag/BCP/C60/SiOx/Cs0.05(FA0.77MA0.23)0.95Pb(I0.77Br0.23)3 (1.68 eV perovskite)/MeO-2PACz/IZO/recombination junction/poly-Si(n)/SiOx/TOPCon silicon bottom cell/poly-Si(p)/SiOx/Cr/Ag	0.1375	30.2	20.0	1.84	83.4	Retained 95% initial PCE after 150 h MPPT (unencapsulated, ISOS-L-1, 25–35 °C, and 40–70% RH air).	2025 [69]
MgF2/ITO/AZO N-21X/PC61BM/Cs0.12FA0.64MA0.24Pb (I0.94Br0.06)3/PEACl/PTAA/ITO/n-type nc-SiOx/intrinsic a-Si:H/p-type nanocrystalline silicon/p-type Si wafer/AZO/Ag	0.32	25.3	18.74	1.76	76.93	Retained minimized sputter-induced damage (PLQ = 0.1) to perovskite/transport layers after transparent electrode sputtering (attributed to AZO N-21X buffer layer).	2025 [70]
LiF/Ag/IZO/SnO2/C60/LiF/Cs00.05(FA0.77MA0.23)0.95Pb(I0.77Br0.23)3 (1.68 eV perovskite)/Me-4PACz SAM/ITO/nc-SiOx:H(n)/intrinsic a-Si:H/n-type c-Si/intrinsic a-Si:H/a-Si:H(p+)/AZO/Ag	1.064	29.15	19.23	1.90	79.4	Retains 95.5% of initial PCE after 300 h 1-sun MPP tracking at 25 °C and 30–40% RH for non-encapsulated devices with Me-4PACz (no LiF).	2020 [71]
MgF2/Ag grid/IZO/SnOx/C60/Cs0.15MA0.15FA0.70Pb(I0.80Br0.20)3 (1.68 eV perovskite, PhenHCl GBP + TSP passivated)/2PACz SAM/ITO/a-Si:H(p+)/intrinsic a-Si:H/a-Si:H(n+)/double-side textured n-Si/intrinsic a-Si:H/a-Si:H(p+)/ITO/Ag	1.03	27.4	19.6	1.84	76.0	Stabilized PCE maintained for 10 min MPPT tracking; no obvious decay.	2021 [72]
MgF2/Ag/IZO/VOx/TPBI/spiro-TTB:F6-TCNNQ/iBAI:FAI passivated Cs0.05MA0.15FA0.8Pb(I0.75Br0.25)3 (1.68 eV perovskite)/C60-SAM anchored a-NbOx/ITO/a-Si:H(n+)/intrinsic a-Si:H/a-Si:H(p+)/double-side textured n-Si/SiNx/IZRO/Ag	1.03	27.1	19.5	1.823	72.8	Retained 90% initial Voc and FF after 150 h MPPT (glass/glass encapsulated, 40 °C, and 1-sun illumination).	2021 [73]
MgF2/Al/IZO/SnO2/C60/TEABr/Cs0.1FA0.9PbI2.74Br0.16Cl0.1 (1.626 eV perovskite)/spiro-TTB/nc-Si:H recombination junction/textured silicon heterojunction (SHJ) bottom cell/Al	0.5036	28.51	19.35	1.849	79.62	Retained >90% initial PCE after 130 h MPPT (unencapsulated, 25 °C, and 30% RH air).	2022 [74]
LiF/Ag grid/IZO/SnO2/C60 (16 nm)/piperazinium iodide/Cs0.22FA0.78Pb(I0.85Br0.15)3 + 5% MAPbCl3 (1.68 eV triple-halide perovskite)/Me-4PACz SAM/ITO/dielectric buffer layer reflector (RDBL)/intrinsic a-Si:H/a-Si:H(p+)/n-type c-Si heterojunction/intrinsic a-Si:H/a-Si:H(n+)/TCO/Ag	1.0	32.5	20.24	1.98	81.18	Reaches T80 (80% initial PCE retention) after 347 h 1-sun MPP tracking at 23–32 °C and ambient air for non-encapsulated devices, retaining 75.7% after 478 h.	2023 [76]
Glass/ITO/PolyTPD/PFN/Cs0.17FA0.83Pb(I0.80Br0.20)3/C60/SnO2/ITO (FPEAI introduced in antisolvent step)/Si heterojunction	0.25	29.57	19.9	1.869	79.5	Retained 80% of initial PCE after 250 h of encapsulated MPPT at 20 °C under 1 sun.	2024 [77]
LiF/Ag grid/IZO/SnOx/C60/Cs0.05[FA0.9MA0.1]0.95Pb(I0.8Br0.2)3 (1.65–1.70 eV perovskite, pFBPA-doped)/SiO2-NPs/Me-4PACz SAM/ITO/a-Si:H(n+)/intrinsic a-Si:H/nc-SiOx:H(n)/intrinsic a-Si:H/a-Si:H(p+)/nc-Si:H(p)/SiOx/Ag/SiOx/Ag/textured n-Si	1.17	30.9	19.81	1.952	79.9	Retained 85–95% initial PCE after 84 h operation (12 h light/12 h dark cycle, ambient air, 25 °C, 30–50% RH, and unencapsulated).	2024 [78]
Ag/WOx/spiro-TTB:F6TCNQ/PEABr/Cs0.17FA0.83Pb(I0.80Br0.20)3 (1.68 eV perovskite)/F-CPE1 (C60-SAM + F-PEABr)/a-NbOx/ITO/textured silicon sub-cell	1	27.0	19.8	1.91	71.5	Retained over 85% initial PCE after 350 h storage in nitrogen atmosphere.	2024 [79]
Anti-reflection coating/transparent electrode/WBG perovskite/ETL (C60/SnO2)/RS-2 diradical SAM/ITO/silicon heterojunction/back electrode	1.0	34.2	20.7	1.99	83.0	Retains 97% of initial PCE after 2000 h MPPT for encapsulated devices under 1-sun LED illumination at 45 °C and 60% RH.	2025 [81]
MgF2/Ag grid/IZO/SnO2/C60/Cs0.05FA0.8MA0.15Pb(I0.75Br0.25)3 (1.68 eV perovskite)/PDADI/Me-4PACz-doped Al2O3/Me-4PACz SAM/NiOx/ITO/intrinsic a-Si:H/a-Si:H(p+)/double-side textured n-Si/intrinsic a-Si:H/a-Si:H(n+)/ITO/NiOx	1.05	30.92	19.75	1.94	80.62	Me-4PACz-doped Al2O3 improves wettability (contact angle reduced to 22.75°), enhancing perovskite film coverage and uniformity.	2025 [82]
Two inverted perovskite top cells (Glass/IWO/NiOx/2PACz SAM/Cs0.05(FA0.84MA0.16)0.95Pb(I0.85Br0.15)3 (1.63 eV perovskite)/PCBM/BCP/Ag) + bifacial SHJ silicon bottom cell (a-Si:H(p+)/intrinsic a-Si:H/c-Si(n)/intrinsic a-Si:H/a-Si:H(n+)/Ag), with perovskite cells placed at 45° to silicon cell	0.16	27.6	21.7	1.120	78.2	Unencapsulated perovskite cell retained 70% initial PCE after 10,000 s light soaking (1-sun, 30–40% RH, and ambient air).	2022 [84]
MgF2/Ag/IZO/SnO2/C60/BCP/Cs0.05(FA0.84MA0.16)0.95Pb(I0.75Br0.25)3 (1.66 eV perovskite, 4PADCB-doped)/NiOx/ITO/recombination junction/n-Poly Si/SiOx/n-Si substrate/SiOx/p+-Si/SiNx/Ag	1.1	31.71	20.42	1.925	80.65	Retained nearly 100% initial PCE after 600 h storage in nitrogen atmosphere.	2025 [85]
Textured PDMS/IZO/ZnO nanoparticles/PCBM/Cs0.05FA0.75MA0.2Pb(I0.75Br0.25)3 perovskite/PFN-Br/PTAA/ITO/Cr/Ag/Sn/SiNx/n+-emitter/c-Si/Al-BSF/Ti/Ag	0.25	23.11	16.48	1.75	80.13	Encapsulated device retains >90% initial PCE after 800 h under 85 °C/85% RH (vs. ~80% for reference without SiNx).	2025 [86]
MgFx/IZO/BCP:Ag/C60/Cs0.17FA0.83PbI0.4Br0.6 (1.65 eV perovskite)/MeO-2PACz/IZO/n-TOPCon (ultrathin SiOx + n-type poly-Si)/n-Si/p-TOPCon/Cr/Ag	0.1375	30.1	20.0	1.80	85.4	Exhibits excellent angle-independent light harvesting, with reflectance suppressed across 300–1200 nm until the angle of incidence (AOI) exceeds 60°.	2024 [87]
MgF2/Ag grid/IZO/SnOx/C60/PDAI-2D top layer/Cs0.22FA0.63MA0.15Pb(I0.83Br0.14Cl0.03) (1.66 eV perovskite, blade-coated, Pb(SCN)2-doped)/PEA2MA2Pb3I10 (2D-3 bottom layer)/MeO-2PACz:Me-4PACz SAM/ITO/nc-Si:H(n+)/intrinsic a-Si:H/a-Si:H(p+)/double-side textured n-Si/intrinsic a-Si:H/a-Si:H(p+)/ITO/meso-Al2O3/Ag	1.04	32.1	20.7	1.92	80.9	Encapsulated device retains 80% initial PCE (T80) for 1700 h under 1-sun continuous MPPT (ambient temperature; glass–glass encapsulation).	2025 [88]

Figure 6. (a) The complete surface of the sample. Reproduced with permission [77]. Copyright 2024, American Chemical Society. (b) The a-NbO_x surface-treatment process. Reproduced with permission [79]. Copyright 2024, American Chemical Society. (c) Schematic of the CMP process of FZ silicon wafer. Reproduced with permission [65]. Copyright 2023, Cell Press. (d) The roles of hybrid ICLs in perovskite/silicon TSCs. Reproduced with permission [67]. Copyright 2024, Cell Press. (e) Schematic diagrams of doping strategy. Reproduced with permission [85]. Copyright 2025, Elsevier. (f) Illustration of the p-i-n device structure. Reproduced with permission [72]. Copyright 2021, Elsevier. (g) Illustration of the 3D WBG perovskite (1.66 eV) blade-coating process on top of the 2D target film. Reproduced with permission [88]. Copyright 2025, Elsevier. (h) Design concept of open-shell diradical SAMs. Reproduced with permission [81]. Copyright 2025, AAAS.

Figure 6. (a) The complete surface of the sample. Reproduced with permission [77]. Copyright 2024, American Chemical Society. (b) The a-NbO_x surface-treatment process. Reproduced with permission [79]. Copyright 2024, American Chemical Society. (c) Schematic of the CMP process of FZ silicon wafer. Reproduced with permission [65]. Copyright 2023, Cell Press. (d) The roles of hybrid ICLs in perovskite/silicon TSCs. Reproduced with permission [67]. Copyright 2024, Cell Press. (e) Schematic diagrams of doping strategy. Reproduced with permission [85]. Copyright 2025, Elsevier. (f) Illustration of the p-i-n device structure. Reproduced with permission [72]. Copyright 2021, Elsevier. (g) Illustration of the 3D WBG perovskite (1.66 eV) blade-coating process on top of the 2D target film. Reproduced with permission [88]. Copyright 2025, Elsevier. (h) Design concept of open-shell diradical SAMs. Reproduced with permission [81]. Copyright 2025, AAAS.

3.1.3. Transport-Layer Strategies

Transport layers play a crucial role in regulating carrier extraction, minimizing interfacial recombination, and improving energy-level alignment in perovskite/silicon TSCs. Continuous optimization of transport-layer materials and device architectures has therefore become an important strategy for improving device efficiency and stability. In particular, recent studies have focused on tailoring HTLs, ETLs, intermediate recombination layers (IRLs), and transparent electrodes to establish efficient charge-transport pathways and reduce parasitic optical and electrical losses across the tandem structure.

Early work by Albrecht et al. demonstrated a fully low-temperature (<120 °C) transport-layer process compatible with SHJ bottom cells, where an ALD-SnO₂ ETL and a spiro-OMeTAD/MoO₃/ITO HTL stack enable efficient carrier transport, yielding a stabilized efficiency of 18.1% [89]. Subsequently, Chen et al. addressed film inhomogeneity on textured silicon by employing nitrogen-assisted blade coating of PTAA, producing pinhole-free HTLs (Figure 7a) and achieving a tandem PCE of 26.2% [42]. To further improve interfacial energetics and carrier transport, composite and bilayer HTLs have been developed. Zhang et al. studied a NiO_x/SAM HTL incorporating 2PACz (Figure 7b), which passivates surface defects and improves energy alignment, while Ma et al. optimized NiO_x/PTAA bilayers via solvent engineering to enhance polymer ordering and hole mobility. Both strategies effectively suppress recombination and improve device performance [84,90]. More recently, molecular engineering of SAM-based HTLs has become a key direction. Ho-Baillie et al. developed Ph-2PACz with extended π-conjugation to reduce energy-level offset and improve conductivity (Figure 7c). Golobostanfard et al. demonstrated that MeO-2PACz promotes oriented perovskite growth and suppresses defect formation, both leading to reduced voltage loss and improved film quality [47,91]. In parallel, advanced HTL architectures have also been explored. Borchert et al. achieved conformal coverage on textured substrates using thermally evaporated HTLs with precise thickness control, while Shen et al. designed a dopant-free poly-TPD/spiro-TTB bilayer enabling efficient hole extraction via cascaded energy alignment, resulting in efficiencies approaching 30% [92,93].

In addition, Zhang et al. explored LiCl-doped SnO₂ ETLs, in which Cl⁻ passivates surface defects, while Li⁺ enhances conductivity and adjusts the conduction band position, thereby increasing V_oc to 1.244 V and enabling a tandem efficiency of 25.42% [94]. Wei et al. developed two low-temperature solution modification strategies for metal-oxide ETLs to optimize interfacial energy matching and film quality. First, they proposed UV-assisted TiI₄ surface augmentation for TiO₂. Iodides merely chemisorb on the TiO₂ surface without altering its bulk lattice, which reversibly tunes work function, uplifts conduction band, expands perovskite grains, and lifts single-junction PCE from 18.4% to 19.4%, with better repeatability [95]. Second, Saeki et al. proposed a sequential diluted spin-coating strategy to fabricate pinhole-free, ultrasmooth SnO₂ films, which suppress non-radiative recombination and deliver a champion 4T tandem PCE of 22.99%. Unencapsulated devices retain 90% efficiency after 1000 h aging. Both mild-process techniques are compatible with large-area roll-to-roll tandem module manufacturing [96]. From a fabrication perspective, mitigating sputtering-induced damage has become a key consideration. Huang et al. investigated an ALD-SnO₂ buffer layer to replace BCP, which improves interfacial adhesion and resists IZO sputtering damage, achieving a PCE of 26.0%. In a related approach, Du et al. introduced a highly transparent In₂O₃ buffer layer with favorable alignment to C₆₀, enabling efficient electron extraction while protecting the perovskite during ITO deposition [97,98]. Beyond planar layer optimization, structural design of ETLs has also emerged as an effective route. Abbasiyan et al. investigated a partially passivated TiO₂ nanorod array modified with a PMMA:PCBM blend, where the structured interface suppresses interfacial recombination and reduces series resistance. When integrated with an optimized Si₃N₄ antireflection layer, the four-terminal tandem device achieved an efficiency of 30.73% [99]. Overall, ETL engineering has progressed from conventional material selection toward coordinated defect passivation, interface protection, and structural design, collectively improving carrier extraction and device stability in perovskite/silicon TSCs.

The intermediate recombination layer (IRL) must enable efficient vertical carrier recombination while minimizing parasitic absorption and lateral shunting. McDonald et al. used in situ grown n-type nc-Si:H via PECVD, whose low-resistance contact with SnO₂ arises from favorable band alignment, and refractive index-matching function reduces optical reflection (Figure 7d) [100]. To overcome parasitic absorption and lateral leakage in nc-Si:H, Li et al. designed nc-SiO_x:H tunnel junctions, where vertically aligned nanocrystalline Si domains embedded in an amorphous matrix promote directional tunneling, while the insulating oxide matrix suppresses lateral conduction [101]. Liu et al. employed p⁺/n⁺ poly-Si junctions for perovskite/tunnel oxide passivated contact (TOPCon) TSCs. A two-step annealing prevents dopant interdiffusion, yielding an atomically steep junction that lowers contact resistivity and raises implied V_oc via enhanced band-to-band tunneling [102]. Alternative IRLs exploit distinct transport mechanisms. Zhang et al. used spray-coated single-walled carbon nanotube networks (Figure 7f), where the porous architecture provides percolative conductivity and high transmittance, while the replaceable top-cell design leverages the mechanical decoupling of the junction [103]. Chozas-Barrientos et al. introduced an all-organic n-n-p molecular junction (doped C₆₀/TaTm), whose low lateral conductivity arises from the organic matrix, effectively blocking shunting paths on textured Si, while vertical transport is maintained through molecular orbital overlap [104]. In a word, IRL design has evolved from inorganic silicon-based tunnel junctions toward diversified architectures incorporating oxide-modified, carbon-based, and fully organic RLs, aiming to balance optical transparency, electrical conductivity, and recombination efficiency in high-performance tandem devices.

Three representative interconnect/recombination systems have been widely exploited in monolithic perovskite tandem devices, exhibiting distinct trade-offs in electrical transport, optical performance, and fabrication compatibility. The nanocrystalline silicon (nc-Si:H) tunnel junction stands out with low series resistance and well-matched band structures that facilitate efficient carrier tunneling, together with mature deposition techniques compatible with industrial silicon manufacturing. Nevertheless, this architecture suffers from unavoidable parasitic light absorption across visible and near-infrared spectra, and lateral shunting defects tend to emerge when deposited on textured silicon wafers, deteriorating device fill factors. As an alternative, ultrathin ITO interlayers deliver outstanding full-spectrum optical transmittance and can be readily fabricated via magnetron sputtering, yet high-energy sputtering particles would inflict irreversible structural damage to the underlying perovskite absorber layers, accompanied by moderately high contact resistance that aggravates voltage loss. Differing from inorganic tunneling materials, conductive SAM-based all-organic molecular junctions completely eliminate parasitic optical absorption and enable damage-free low-temperature solution deposition, which efficiently restrains lateral leakage paths. Unfortunately, such organic interconnects are restricted by insufficient vertical carrier conductivity, and their molecular frameworks tend to decompose under prolonged thermal and illumination stress, resulting in unsatisfactory long-term operational stability.

For transparent electrodes, the core challenge is balancing free-carrier absorption and sheet resistance. Zheng et al. replaced ITO with W-doped In₂O₃ (IWO). The large valence difference between W⁶⁺ and In³⁺ allows low resistivity at reduced carrier density, cutting near-infrared (NIR) parasitic absorption from 12.4% to 0.6% (Figure 7e) [105]. In addition, the optimization of transparent conductive oxides has focused on deposition process control. The Sittinger team studied sputtering parameter engineering (substrate temperature, oxygen partial pressure, and target power) to improve ITO and IZO films, achieving a carrier mobility of 50 cm²/V·s and a weighted absorption below 3.1%, which enabled tandem efficiencies of 24.3% [106]. Similarly, Kabaklı et al. studied oxygen flow-controlled DC sputtering to fabricate low-defect ITO films with a mobility of 53.4 cm²/V·s. By constructing a bilayer ITO with different oxygen contents, the interfacial barrier with SnO_x was reduced, increasing the average tandem efficiency from 22% to 25% [107]. Wang et al. tuned IZO thickness to modulate optical interference within the perovskite, thereby adjusting sub-cell photocurrent balance via cavity effects [108]. Hatt et al. used ALD-Al₂O₃ as a protective mask during electroplated Cu metallization, ensuring chemical stability for large-area processing [109]. Finally, Zhang et al. designed bilayer IZrO with a columnar-equiaxed dual microstructure. The columnar phase enhances vertical transport, the equiaxed phase improves lateral spreading, and the graded refractive index minimizes reflection, achieving certified 30.12% PCE (Figure 7g) [110].

In general, IRL design has shifted from simple inorganic tunnel diodes toward oxide-modified, carbon-based, and organic junctions, each exploiting distinct carrier-transport and leakage-suppression mechanisms. Transparent electrodes have evolved from conductivity-transparency trade-offs to integrated optical–electrical–structural engineering, enabling better current matching, reduced losses, and scalable high-performance TSCs. Table 5 summarizes the transport-layer strategies and the photovoltaic parameters of perovskite/silicon TSCs. nc-Si tunnel junctions offer high conductivity but cause optical loss. All-organic interconnects reduce parasitic absorption but suffer from poor heat resistance. Composite oxide transport layers strike the best balance of optical and electrical performance. Perovskite/silicon systems benefit most from lattice rigidification and SAM passivation due to inert silicon substrates. This set of strategies cannot be fully replicated in flexible thin-film tandem architectures.

Table 5. Transport-layer strategies and the PV parameters of perovskite/silicon TSCs.

Device Architecture	Active Area (cm2)	PCE (%)	Jsc (mA/cm2)	Voc (V)	FF (%)	Stability	Year [Ref.]
Textured c-Si/SnO2/WBG perovskite/C60/SnO2/ITO (monolithic 2T)	/	25.4	18.8	1.82	74.1	Stabilized power output measured (18.1%); no degradation during stabilization test.	2016 [89]
N-I-P monolithic: textured SHJ Si/SnO2–LiCl/WBG perovskite/spiro-OMeTAD/MoO3/IZO	0.5	25.42	16.99	1.92	77.95	Continuous illumination stability test for several hours; no significant degradation.	2019 [105]
Monolithic 2T perovskite/Si tandem (blade-coated perovskite on sub-µm textured SHJ Si, ITO RL, PDMS light-scattering layer)	0.42	26.2	19.3	1.8	75	Unencapsulated device retained ~92% of initial efficiency after 100 h continuous AM1.5G illumination in ambient air.	2020 [42]
Monolithic 2T N–I–P perovskite/SHJ Si tandem with SnO2–LiCl composite ETL (inorganic ETL engineering for recombination junction)	0.5	25.42	16.99	1.92	77.95	Exhibited negligible decay during MPP tracking under AM 1.5G illumination (~2000 s).	2021 [94]
Monolithic 2T perovskite/Si tandem with ITO stack front electrode	/	25.6	19.8	1.73	73	MPP tracking at maximum power point for 90 s; stable power output with negligible hysteresis.	2023 [107]
Monolithic 2T perovskite/Si tandem with Ph-2PACz self-assembled hole-selective layer	1.03	28.9	19.1	1.91	79.1	Continuous 1-sun illumination for 680 h with negligible degradation; damp-heat test at 85 °C/85% RH for 280 h.	2023 [91]
Glass/ITO/PTAA/FAPbI3-based perovskite/C60/SnO2/IZO/SHJ silicon bottom cell/ITO	/	24	18.5	1.85	70	Continuous light soaking for several hours; negligible degradation reported.	2024 [47]
Monolithic 2T perovskite/TOPCon tandem with p+/n+polysilicon tunneling IRL	/	27.6	18.1	1.84	82.9	/	2024 [102]
2T detachable perovskite/Si tandem (n–i–p WBG perovskite/SWCNT RL/TOPCon Si)	/	23.2	16.82	1.837	75.1	Retained >95% of initial PCE after 11 cycles of water soaking, detachment, and reassembly.	2024 [103]
Monolithic 2T: textured SHJ Si/multifunctional interlayer/WBG perovskite/C60/SnO2/ITO	1	28.1	19.4	1.88	77.1	MPP tracking for 1000 h under 1 sun at 25 °C, retained over 90% of initial PCE.	2025 [92]

Figure 7. (a) Schematic depiction of the evolution of a perovskite layer on a textured surface. Reprint with permission [42]. Copyright 2020, Elsevier. (b) Schematic diagram of an inverted planar PSC. Reprint with permission [84]. Copyright 2022, Royal Society of Chemistry. (c) Schematic of the p-i-n silicon–perovskite TSC. Reprint with permission [91]. Copyright 2023, Elsevier. (d) Device performance of perovskite–Si TSCs with various RJs. Reprint with permission [100]. Copyright 2022, American Chemical Society. (e) Schematic diagram and the FESEM images of the bifacial HIT silicon solar cell. Reprint with permission [105]. Copyright 2019, Royal Society of Chemistry. (f) The detachment and reassembly of 2-T perovskite/Si tandem via the accelerated aging procedure. Reprint with permission [103]. Copyright 2024, American Chemical Society. (g) Structure and carrier transport path schematic of a perovskite/Si TSC. Reprint with permission [110]. Copyright 2025, American Chemical Society.

Figure 7. (a) Schematic depiction of the evolution of a perovskite layer on a textured surface. Reprint with permission [42]. Copyright 2020, Elsevier. (b) Schematic diagram of an inverted planar PSC. Reprint with permission [84]. Copyright 2022, Royal Society of Chemistry. (c) Schematic of the p-i-n silicon–perovskite TSC. Reprint with permission [91]. Copyright 2023, Elsevier. (d) Device performance of perovskite–Si TSCs with various RJs. Reprint with permission [100]. Copyright 2022, American Chemical Society. (e) Schematic diagram and the FESEM images of the bifacial HIT silicon solar cell. Reprint with permission [105]. Copyright 2019, Royal Society of Chemistry. (f) The detachment and reassembly of 2-T perovskite/Si tandem via the accelerated aging procedure. Reprint with permission [103]. Copyright 2024, American Chemical Society. (g) Structure and carrier transport path schematic of a perovskite/Si TSC. Reprint with permission [110]. Copyright 2025, American Chemical Society.

3.2. Perovskite/CIGS TSCs

Perovskite/CIGS TSCs are another promising architecture that combines the high efficiency of CIGS solar cells with the tunable bandgap and low-cost fabrication of perovskite materials. In this architecture, a WBG perovskite top cell is stacked on top of an NBG CIGS bottom cell. The CIGS bottom cell, which is known for its high PCEs and excellent stability, provides a reliable base for the tandem cell, while the perovskite top cell allows for the absorption of high-energy photons that are not efficiently captured by the CIGS cell. Similar to perovskite/silicon TSCs, perovskite/CIGS TSCs can be fabricated in both monolithic and mechanically stacked configurations. Monolithic perovskite/CIGS tandem cells offer the advantage of a more efficient charge transfer between the two sub-cells, but the fabrication process is more complex and requires precise control over the growth and interface properties of the two materials. Mechanically stacked perovskite/CIGS TSCs, on the other hand, offer greater flexibility in terms of material selection and device design, but the mechanical-stacking process can introduce additional optical and electrical losses.

3.2.1. Perovskite-Layer Regulation

Surface defects and composition mismatch are the core issues limiting the performance of perovskite/CIGS TSCs. To address these problems, strategies such as organic salt modification, ion doping, and additive regulation have been developed, which can precisely passivate defect sites while optimizing the crystal structure and band alignment of perovskites. Energy-band alignment at interfaces is achieved by selecting transport layers (Cu:NiO_x for hole extraction, and bis-C₆₀ for electron barrier reduction) and adjusting the CIGS composition (increasing S/Se ratio raises CIGS conduction band to reduce the CdS barrier) [111]. In p-i-n structures, MDADI supplies I⁻ to fill iodine vacancies and coordinate uncoordinated Pb²⁺, suppressing non-radiative recombination [112]. FABr treatment induces synchronous FA⁺/Br⁻ diffusion. FA⁺ adjusts the tolerance factor to stabilize the cubic phase, while Br⁻/Cl⁻ exchange passivates halogen vacancies and removes secondary phases [113]. In CsFAPb(IBr)₃, Cl doping combined with PDI surface treatment increases ion-migration activation energy and reduces phase/surface defects (Figure 8b) [114]. For lead-free tin perovskites, the Pd(dppf)Cl₂ additive acts multifunctionally. The benzene ring delays crystallization via weak interactions, Cl⁻ coordinates uncoordinated Sn²⁺, and Fe²⁺/Pd²⁺ synergistically reduces Sn⁴⁺ to Sn²⁺, inhibiting oxidation (Figure 8a) [115]. The preparation process of perovskite films directly affects their crystallization quality and coverage uniformity. Especially for the irregular and rough surface of CIGS bottom cells, process innovation is required to achieve conformal deposition and defect control, while balancing large-area compatibility and high performance. Ultrasonic spray coating disperses precursor into micron droplets for conformal deposition, while SAMs and LiF/C₆₀/BCP layers passivate interfaces [116]. Stepwise DMSO annealing first applies 100 °C to break precursor-DMSO coordination for rapid nucleation and large grains, and then 80 °C to slowly release residual solvent, avoiding pinholes and suppressing volatile-halide sublimation that forms PbI₂ impurities [117].

In addition, with the aid of numerical simulation tools and multi-junction structure design, the band alignment, thickness parameters, and charge-transport paths of perovskite/CIGS TSCs can be systematically optimized. This maximizes light absorption and carrier-utilization efficiency, breaking through the limitations of experimental optimization. Multi-junction designs exploit complementary absorption bands (400–800 nm vs. 800–1200 nm) and concentration effects (linear J_sc increase; logarithmic V_oc rise) to reduce thermal and series-resistance losses [118]. On the other hand, a study employed SCAPS-1D numerical simulation to optimize a monolithic 2-terminal 2D-3D perovskite/CIGS TSC. Ultimately, the top cell achieved a PCE of 21.41%, the bottom cell had a PCE of 12.98%, and the tandem cell PCE exceeded 30%. The study optimized the top cell by adjusting the perovskite bandgap and thickness. Increasing the bandgap enhanced the quasi-Fermi level difference, and increasing the thickness improved light absorption. For the bottom cell, the study optimized the thickness of p-CIGS and n-GaSe to enhance long-wave light absorption and reduce interface defects, and controlled the doping concentration to optimize charge transport [119]. In addition, Y. K. Singh et al. first determined Cu₂O as the optimal HTL and ZnO as the optimal ETL, thereby optimizing the charge-transport path and reducing the transmission obstacles. Subsequently, they optimized the thickness of the active layer (1 μm for the top battery, and 3 μm for the bottom battery) and the defect density (1 × 10¹⁰ cm⁻³ for the top cell body defect density, and 1 × 10¹⁴ cm⁻³ for the bottom cell), thereby reducing the non-radiative recombination of charge carriers. Then, they combined with filtering spectroscopy analysis and current-matching adjustment to avoid current loss. The combined effect of these three factors significantly improved the performance of the tandem battery, ultimately achieving a PCE of 34.52%, V_oc of 2.29 V, J_sc of 17.59 mA/cm², and FF of 85.69% [120]. Table 6 summarizes the defect modification and component regulation strategies and the photovoltaic performance of the corresponding perovskite/CIGS TSCs. For CIGS bottom cells, mild halide tuning avoids lattice mismatch damage. Spray coating conformal perovskite films works better than spin coating on rough CIGS surfaces for flexible devices.

Table 6. Defect modification and component regulation strategies and the PV performance of the corresponding perovskite/CIGS TSCs.

Device Architecture	Active Area (cm2)	PCE (%)	Jsc (mA/cm2)	Voc (V)	FF (%)	Stability	Year [Ref.]
Top cell: Glass/ITO/Cu:NiOx/MAPbI3/C60/bis-C60/ITO. Bottom cell: Glass/Mo/Cu(In,Ga)(S,Se)2/CdS/i-ZnO/TCO	0.1	18.8	15.8	1.07	74.0	Perovskite top cell shows efficiency improvement under illumination (12.5%→13.6% after 600 s).	2017 [111]
Top cell: Glass/ITO/NiOx/Me-4PACz SAM/Rb0.05Cs0.05[(FA0.83MA0.17)0.9]Pb(I0.83Br0.17)3 (1.63 eV perovskite, MDADI surface passivation)/C60/SnOx/ITO/LiF. Bottom cell: Glass/Mo/Ag-alloyed CIS/CdS/ZnO/AZO/Ni/Al/MgF2	0.055	29.9	21.5	1.24	81.9	LiF optical layer suppresses NIR interference, maintaining 81.5% average NIR transmittance (800–1300 nm).	2023 [112]
Top cell: Glass/FTO/ML-TAPC/(FAxMA1−x)Pb(I3–x−γBrxClγ (1.65 eV triple-halide perovskite)/C60/PEIE/IZO. Bottom cell: Glass/Mo/CIGS/CdS/i-ZnO/ITO	/	24.5	19.37	1.11	75.9	Retains 80% initial PCE after 40 days (25 °C, 25% RH) vs. 54% for control perovskite cell.	2023 [113]
Top cell: Glass/IZO/SnOx/C60/(FA0.78Cs0.22)Pb(I0.82Br0.15Cl0.03 (1.67 eV perovskite, PDI surface treatment)/Poly-TPD/ITO. Bottom cell: Glass/Mo/CIGS/CdS/i-ZnO/AZO/Ni/Al/Ni	0.076	28.4	19.5	1.20	81.42	Retains 100% of initial PCE after 600 h for unencapsulated perovskite top cells under MPPT, 50 °C, and N2 atmosphere, and 90% after 1300 h.	2023 [114]
Ag/IZO/ALD SnOx/C60/PEAFACsSnI2Br (1.66 eV WBG Sn perovskite, 1.2 M, Pd(dppf)Cl2 additive)/PEDOT:PSS/IZO/CIGS/Mo	0.42	14.51	15.5	1.20	78.0	Retains 88% initial PCE after 1000 h in a N2 glovebox (vs. 60% for control Sn PSC).	2025 [115]
Glass/Mo/Cu(In,Ga)SSe/CdS/ZnO/ITO/NiOx/Al2O3 nanoparticles/Me-4PACz SAM/Cs0.18FA0.82Pb(I0.94Br0.06)3 (1.6 eV perovskite, ~300 nm, ultrasonic spray-coated)/LiF/C60/BCP/ITO/Ag	0.45	20.2	24.8	1.59	51.1	Me-4PACz SAM reduces non-radiative losses, boosting Voc.	2024 [116]
Top cell: Glass/ITO/SnO2/Cs0.15FA0.60MA0.20PbI2.4Br0.6 (1.68 eV perovskite, stepwise 100→80 °C DMSO annealing)/spiro-OMeTAD/MoOx (buffer layer)/IZO/Ag fingers. Bottom cell: Opaque CIGS solar cell	0.2025	29.36	21.80	1.204	81.80	Retains stable photocurrent for 5 h under MPPT for opaque perovskite cells, with a stabilized PCE of 21.26% and no obvious decay.	2025 [117]
Glass/Ag/Al2O3/CIGS/CdS/ZnO/BZO/IZO/TiO2/FAxCs1−xPb(I,Br)3 (1.62 eV perovskite, 660 nm)/CuSCN/ITO/MgF2	/	34.39	21.12	1.89	86.0	PCE increases by 5% at 10 suns, maintaining efficiency up to 36%.	2022 [118]
FTO/SnO2/MPA0.2MA0.8PbI3 (1.65 eV 2D–3D perovskite, 500 nm)/NiO/ZnO/n-GaSe/p-CIGS/Mo/glass	/	32.7	22.3	1.704	84.1	Optimized bandgap matching (perovskite 1.65 eV/CIGS 1.15 eV) minimizes spectral loss and enhances long-term performance.	2024 [119]
Top cell: FTO/Cu2O/CH3NH3PbI3−xClx (1.65 eV perovskite, 288 nm)/ZnO/IZO. Bottom cell: ZnO:Al/i-ZnO/CdS/CIGS/Mo/glass with current-matching	/	34.52	17.59	2.29	85.69	Cu2O/ZnO CTLs enable favorable band alignment, reducing non-radiative losses.	2025 [120]

3.2.2. Interface Regulation

Interface defects and energy-level mismatch are the core bottlenecks restricting the carrier transport efficiency of perovskite/CIGS TSCs. Some research teams have achieved the synergistic optimization of defect passivation, energy-level alignment, and crystallization quality improvement through strategies such as organic molecule modification and dopant regulation. GABr in an IPA–toluene mixed solvent reduces IPA erosion while enabling GA⁺–iodine coordination and Br⁻ filling of halogen vacancies. It also promotes α-phase crystallization and optimizes energy alignment (Figure 8c) [121]. The Me-4PACz self-assembled monolayer conforms to rough CIGS, while PEAI forms a 2D/3D heterojunction that passivates surface traps. LiF acts as a dipole layer to lower the C₆₀/perovskite barrier, and RbF post-treatment increases Cu-vacancy density, enhancing p-type conductivity [122]. On NiO_x, 4PADCB binds via phosphonic acid, reducing the Ni³⁺/Ni²⁺ ratio, adjusting valence-band position for hole extraction, and improving wettability for conformal perovskite growth [123]. Intermediate layers serve as both chemical barriers and transport bridges. A 20 nm IZO layer between AZO and perovskite blocks deprotonation reactions (AZO alkaline sites vs. MA⁺ cations) and inhibits Zn²⁺/I⁻ interdiffusion, while its conduction band (~−4.5 eV) ensures efficient electron transfer; it also induces oriented perovskite crystallization [124]. In addition, the interfacial stability and mechanical properties of flexible perovskite/CIGS TSCs are highly dependent on substrate modification and CIGS-layer growth regulation. Synergistic breakthroughs in the efficiency and stability of flexible devices can be achieved through PI substrate modification, gradient composition design of the CIGS layer, and doping optimization. On PI, a graded Ga profile and trace Ag seeding improve CIGS quality [125]. Alternatively, a PI interlayer on SLG suppresses Na⁺/K⁺ diffusion via steric/chemical adsorption, preventing K⁺ substitution for V_Cu, and matches thermal expansion to shift the neutral plane, reducing bending stress (Figure 8d) [126]. Interface regulation involves molecular scale passivation chemistry, dipolar energy-level modulation, diffusion-blocking interlayers, and stress-managed substrate design. All of these strategies aim to reduce non-radiative recombination, improve band alignment, and enhance mechanical integrity. Table 7 summarizes the interface modulation and interlayer design strategies and the photovoltaic performance of the corresponding perovskite/CIGS TSCs. Buffer interlayers block ion interdiffusion between perovskite and CIGS, outperforming single-molecule passivation alone. SAM modification fits uneven CIGS substrates most effectively.

Table 7. Interface modulation and interlayer design strategies and the PV performance of the corresponding perovskite/CIGS TSCs.

Device Architecture	Active Area (cm2)	PCE (%)	Jsc (mA/cm2)	Voc (V)	FF (%)	Stability	Year [Ref.]
Top cell: Glass/MgF2/ITO/SnO2/Cs0.05MA0.10FA0.85Pb(I0.97Br0.03)3 (1.58 eV perovskite, GABr mixed solvent post-treatment)/spiro-OMeTAD/MoOx/ITO/Ag fingers. Bottom cell: Glass/Mo/CIGS/CdS/i-ZnO/AZO	0.07	25.5	22.1	1.03	79.7	Retains 88% of initial PCE after 1000 h for unencapsulated top cells under storage at room temperature and 10% RH.	2022 [121]
Glass/Mo/CIGS/CdS/i-ZnO/ZnO:Al (recombination layer)/Me-4PACz SAM/Cs0.05(MA0.23FA0.77)Pb1.1(I0.77Br0.23)3 (1.68 eV perovskite, PEAI additive)/LiF/C60/ALD SnO2/IZO/LiF/Ag frame	1.04	24.2	18.8	1.77	71.2	Me-4PACz SAM enables conformal coverage on rough CIGS surface, suppressing interfacial recombination.	2022 [122]
Polyimide substrate/Mo/CIGS/CdS/i-ZnO/AZO/NiOx/4PADCB SAM/Cs0.05(FA0.77MA0.23)0.95Pb(I0.77Br0.23)3 (WBG perovskite, p-F-PEAI additive)/C60/BCP/thin Ag/ITO/Ag/MgF2	0.07	22.8	16.8	1.82	74.1	Retains 93% of initial PCE after 300 h for devices under thermal aging at 80 °C and N2 atmosphere, and 90% after 150 thermal cycles at 65 °C.	2025 [123]
Flexible metal substrate/Mo/CIGS/CdS/i-ZnO/AZO/IZO/SAM/NiOx/WBG perovskite/LiF/C60/SnO2/IZO/LiF	0.12	21.24	17.08	1.829	68.0	Retains 80% of initial PCE after 120 min MPP operation for unencapsulated devices under high-temperature conditions (90 °C, ambient air), outperforming control (fails after ~25 min).	2025 [124]
Top cell: Flexible PI substrate/ITO/SnO2/perovskite/spiro-OMeTAD/MoOx/ITO/Ag fingers. Bottom cell: Flexible PI substrate/Mo/Ag/CIGS/CdS/i-ZnO/AZO/Ni/Al grid/MgF2	0.07	21.56	20.45	1.015	73.74	Retains >90% of initial PCE after 200 h for devices under storage in a dry cabinet or glovebox.	2023 [125]
Flexible PI substrate/Mo/CIGS/CdS/i-ZnO/IZO/Br-2EPT SAM/Cs0.05(FA0.83MA0.17)0.95Pb(I0.83Br0.17)3 (1.64 eV perovskite)/C60/ALD SnO2/ITO/Ag/MgF2	0.16	22.8	17.94	1.752	75.24	Retains 86.5% of initial PCE after 500 h light soaking and 85.0% after 280 h MPPT for unencapsulated devices at 35 °C and N2 atmosphere.	2025 [126]

3.2.3. Transport-Layer Strategies

The material selection of transport layers and buffer layers directly affects carrier transport efficiency and device stability. Some research teams have addressed core issues such as ion diffusion, parasitic absorption, and process damage through strategies including replacing traditional materials, removing redundant layers, and constructing protective buffer layers. During ITO sputtering, high-energy Ar⁺ ions oxidize amino groups of spiro-OMeTAD to C=O bonds, reducing the energy offset at the spiro-OMeTAD/ITO interface and eliminating the hole injection barrier. Removing the MoO_x buffer layer avoids near-infrared parasitic absorption and interface defects from its thermal decomposition, while K doping fills lattice vacancies and suppresses I⁻ migration, improving crystallinity and transport [127]. A ZnO nanoparticle buffer layer deposited by a water-free spin-coating process disperses Ar⁺ impact during IZO sputtering, blocks In diffusion into the perovskite, and prevents the CsPbI₃ phase transition caused by water. DMAI in the precursor regulates crystallization, enlarging grains and reducing defect density, while P3CT and PCBM align with the valence and conduction bands to lower the charge-extraction barrier (Figure 8f) [128]. Numerical simulations enable precise screening of transport layers and interface structures. A 1 nm Me-4PACz layer passivates perovskite surface defects through its phosphonic acid groups, and its HOMO matches the perovskite valence band to enhance hole extraction and reduce parasitic absorption. The Me-4PACz/ITO tunnel recombination junction provides efficient recombination and low-resistance transport between sub-cells [129]. In an HTL-free architecture with BaSnO₃ as the ETL, a favorable conduction band offset with the perovskite suppresses electron backflow and promotes extraction, while the high electron mobility of BaSnO₃ reduces series resistance. The absence of NiO removes the associated interface barrier. The triple-cation perovskite further improves stability by suppressing phase separation [130]. For damage-free integration via lamination, a transparent conductive adhesive (TCA) exploits Al₂O₃ nanoparticles to disrupt PVDF-HFP crystalline domains, creating more transport channels for the ionic liquid. Hydrogen bonds between PVDF-HFP C-F groups and ITO hydroxyls, together with EMI⁺ binding to oxygen vacancies on ITO, ensure strong adhesion. Low-temperature (80 °C) dry transfer prevents thermal degradation and solvent erosion of the perovskite, and the weak TCA–perovskite interaction eliminates lead residue during peeling (Figure 8e) [131]. Low-temperature organic transport materials match flexible CIGS fabrication, while metal oxide layers deliver higher charge mobility but require high-temperature treatment incompatible with plastic substrates. CIGS-based TSCs prioritize conformal deposition and stress-relief interlayers, restricted by rough surfaces and flexible substrate limits. High-temperature lattice engineering is less effective here.

Figure 8. (a) Schematic diagram of possible chemical reactions in solution. Reproduced with permission [115]. Copyright 2025, American Chemical Society. (b) Schematic diagram of the Cl bulk incorporation and PDI surface treatment. Reproduced with permission [114]. Copyright 2023, Royal Society of Chemistry. (c) Schematic fabrication process of a planar PSC. Reproduced with permission [121]. Copyright 2022, Wiley-VCH GmbH. (d) Lift-off process for f-CSCs. Reproduced with permission [126]. Copyright 2024, Elsevier. (e) Circular loop of laminated perovskite–CIGS PSC. Reproduced with permission [131]. Copyright 2024, American Chemical Society. (f) Impact of buffer-layer deposition. Reproduced with permission [128]. Copyright 2024, Elsevier.

Figure 8. (a) Schematic diagram of possible chemical reactions in solution. Reproduced with permission [115]. Copyright 2025, American Chemical Society. (b) Schematic diagram of the Cl bulk incorporation and PDI surface treatment. Reproduced with permission [114]. Copyright 2023, Royal Society of Chemistry. (c) Schematic fabrication process of a planar PSC. Reproduced with permission [121]. Copyright 2022, Wiley-VCH GmbH. (d) Lift-off process for f-CSCs. Reproduced with permission [126]. Copyright 2024, Elsevier. (e) Circular loop of laminated perovskite–CIGS PSC. Reproduced with permission [131]. Copyright 2024, American Chemical Society. (f) Impact of buffer-layer deposition. Reproduced with permission [128]. Copyright 2024, Elsevier.

3.3. All-Perovskite TSCs

All-perovskite TSCs are a relatively new and exciting architecture that uses two different perovskite materials with different bandgaps to form a tandem cell. In this architecture, a WBG perovskite top cell is stacked on top of an NBG perovskite bottom cell. The advantage of all-perovskite tandem cells is that they can be fabricated using a single material system, which simplifies the fabrication process and potentially reduces costs. Additionally, the tunability of the bandgap of perovskite materials allows for precise optimization of the optical absorption characteristics of the two sub-cells to maximize the efficiency of the tandem cell. However, the development of all-perovskite TSCs also faces several challenges. One of the main challenges is the instability of perovskite materials, especially under high humidity, temperature, and light-exposure conditions. To address this issue, various strategies have been proposed, such as the use of stable perovskite compositions, the incorporation of protective layers, and the optimization of the device architecture. Another challenge is the development of efficient and stable RLs for all-perovskite TSCs, as the interface between the two perovskite materials can introduce significant recombination losses.

3.3.1. Perovskite-Layer Regulation

Mechanisms for suppressing V_oc loss and photoinduced halide segregation in WBG perovskites center on defect passivation, ion-immobilization, and energy-level tuning. DMA induces octahedral tilting through steric hindrance, shortening Pb-I bonds and raising the VBM to widen the bandgap and hinder halide migration, while Cl⁻ fills iodine vacancies and lowers the CBM [132]. PbCl₂ supplies Cl⁻ to form a trihalide phase that reduces halide vacancies and inhibits migration, while PMACl forms a surface 2D perovskite that passivates defects and blocks moisture [133]. PDA uses bidentate coordination to uncoordinated Pb²⁺ and hydrogen bonds to fill halide vacancies, optimizing energy alignment with C₆₀, strengthening the built-in field, and suppressing PbI₂/Br-rich phase separation [134]. DBSA coordinates Pb²⁺ via its sulfonate group, and the long alkyl chain sterically restricts grain tilting to guide (001) vertical growth. Its oxygen atoms form Pb-O and H-bonds with FA⁺, passivating bulk defects and homogenizing the surface electric field to suppress I/Br phase segregation (Figure 9d) [135]. DDPA employs phosphonic acid groups to strongly coordinate with FA⁺, halides, and uncoordinated Pb²⁺, suppressing V_FA, V_I, and V_Pb, while the alkyl chain restricts tilting and promotes (011) orientation, and its hydrophobicity enhances moisture resistance [136]. F-PEAI spin coating forms a 2D FPEA₂PbI₄ capping layer concentrated in the top 10 nm. The large organic cations increase the activation energy for halide migration, inhibit I/Br phase separation, passivate grain-boundary/surface defects, and enhance the built-in potential without introducing a transport barrier (Figure 9a) [137]. Piracetam preferentially adsorbs on the (110) plane to induce oriented growth, then reacts with residual PbI₂ to form 1D nanoneedles and hydrogen-bonds with FA⁺ to suppress vacancy defects (Figure 11a) [138]. Furthermore, Du et al. added 2,5-dibromothieno [3,2-B]thiophene (DBrT) as an additive to the perovskite precursor of FA_0.75Cs_0.25PbI_2.4Br_0.6. For the inverted WBG PSCs, the PCE reached 22.40%. The semi-transparent WBG device achieved a PCE of 20.39%. And the four-terminal tandem device broke through a PCE of 28.31%. This is because after crystallization, DBrT diffuses to the grain boundaries as well as the upper and lower surfaces of the perovskite. It passivates the uncoordinated Pb²⁺ defects, inhibits the phase separation of iodine/bromine, optimizes the energy-level alignment between the perovskite and C₆₀, and prolongs the carrier lifetime simultaneously (Figure 9c) [139].

On the other hand, Fu et al. proposed an in situ molecular compensation optimization method by introducing 5-aminovaleric acid hydroiodide (5-AVAI) into the 1.79 eV WBG perovskite precursor. By adding 5-AVAI to the precursor of WBG perovskite (Cs_0.2FA_0.8Pb(I_0.6Br_0.4)₃), the single-junction WBG PSCs achieved a champion PCE of 20.23%. After integrating with an NBG bottom cell (PCE 22.01%), the champion PCE of the all-perovskite TSCs (APTSCs) reached 28.9%. Owing to its high dipole moment (2.57 Debye, higher than 1.60 Debye of Me-4PACz) and small horizontal diameter (3 Å), 5-AVAI migrates spontaneously to the buried interface during perovskite deposition. It compensates for the defect sites of SAMs through strong coordination and hydrogen bonding, reducing interfacial energy loss. Meanwhile, its amphoteric property, possessing both amino and carboxyl groups, enables interaction with the perovskite lattice via hydrogen bonding, passivating grain boundary defects, inhibiting ion migration caused by V_I and V_Cs/V_FA, and alleviating halide segregation (Figure 10a). Additionally, 5-AVAI can regulate perovskite crystallization, improve film compactness, extend carrier lifetime, and reduce non-radiative recombination [140]. p-OHPEAI adsorbs parallelly via synergistic -NH₃⁺ and -OH, avoiding an insulating 2D phase. -NH₃⁺ binds uncoordinated Pb²⁺ and I⁻ fills Pb vacancies, while -OH increases the electron density on the benzene ring to suppress metallic Pb and forms a dipole layer at the perovskite/C₆₀ interface to accelerate carrier extraction [141].

Figure 9. (a) The schematic illustration of 2D wrapped 3D structural perovskite. Reproduced with permission [137]. Copyright 2024, American Chemical Society. (b) Schematics illustrating membrane filtration and magnetic filtration processes. Reproduced with permission [142]. Copyright 2024, Wiley-VCH GmbH. (c) Schematic diagram of adding DBrT in the growth process of perovskite crystals. Reproduced with permission [139]. Copyright 2024, Wiley-VCH GmbH. (d) Schematic diagrams of a possible (001) facet-stacking mode. Reproduced with permission [135]. Copyright 2023, Wiley-VCH GmbH. (e) Schematic illustration of the operational mechanism for TFPHCl-induced self-assembly. Reproduced with permission [143]. Copyright 2024, Royal Society of Chemistry. (f) Schematic evolution of perovskite films. Reproduced with permission [144]. Copyright 2025, Royal Society of Chemistry.

Figure 9. (a) The schematic illustration of 2D wrapped 3D structural perovskite. Reproduced with permission [137]. Copyright 2024, American Chemical Society. (b) Schematics illustrating membrane filtration and magnetic filtration processes. Reproduced with permission [142]. Copyright 2024, Wiley-VCH GmbH. (c) Schematic diagram of adding DBrT in the growth process of perovskite crystals. Reproduced with permission [139]. Copyright 2024, Wiley-VCH GmbH. (d) Schematic diagrams of a possible (001) facet-stacking mode. Reproduced with permission [135]. Copyright 2023, Wiley-VCH GmbH. (e) Schematic illustration of the operational mechanism for TFPHCl-induced self-assembly. Reproduced with permission [143]. Copyright 2024, Royal Society of Chemistry. (f) Schematic evolution of perovskite films. Reproduced with permission [144]. Copyright 2025, Royal Society of Chemistry.

NBG Pb-Sn perovskites are essential for bottom cells in TSCs but suffer from poor vertical uniformity, easy oxidation of Sn²⁺, and severe interfacial recombination. Xing et al. employed a molecular anchoring strategy using L-alanine methyl ester (AME) as additive, a PEA₂PbI₃SCN buried layer, and EDA(I/Cl)₂ post-treatment to prepare vertically uniform Cs_0.1MA_0.3FA_0.6Pb_0.5Sn_0.5I₃ films with optimized interfaces, achieving a champion single-junction PCE of 22.3% and a 4-T all-perovskite tandem PCE of 27.1% [145]. Sun et al. introduced benzyl viologen (BV) for surface charge-transfer doping, which shifted the Fermi level upward by 0.29 eV, suppressed Sn²⁺ oxidation, and strengthened the built-in field, yielding single-junction PCE of 22.58% and 2-T tandem PCE of 26.33% [146]. 4-Trifluoromethylphenylhydrazine hydrochloride (TFPHCl) was added into a FA_0.7MA_0.3Sn_0.5Pb_0.5I₃ precursor. The TFPH⁺ cations floated to the surface and formed an antioxidant p/p⁻ homojunction that improved stability and charge separation (Figure 9e), resulting in single-junction PCE of 23.09% and tandem PCE of 27.42% [143]. Introducing canolol (CNL), a natural reactive-oxygen-species scavenger, into a (FASnI₃)_0.6(MAPbI₃)_0.4 precursor balanced crystallization reduced defects, and extended carrier lifetime. The champion single-junction PCE reached 23.7% (22.08% for 1.035 cm²), and the 2-T tandem PCE attained 28.71% (27.2% for large area) [147]. Succinic acid (AA) and thiomalic acid (TA) were used as additives. AA strengthens Sn-I bonds and suppresses iodine migration (single-junction PCE 21.95%), while TA adds reductive thiol groups that directly reduce Sn⁴⁺ and relieve tensile strain (single-junction PCE 22.7%). A TA-modified bottom cell enabled a 2-T tandem PCE of 28.6% [148]. Replacing Sn particles with Ni particles as the reducing agent (Figure 9b) takes advantage of the lower reduction potential of Ni and its ferromagnetism for magnetic removal, yielding single-junction PCE of 22.29% and a monolithic tandem PCE of 28.13% [142]. Zhao’s team introduced K₂SnO₃ into the precursor and PEDOT:PSS, in situ generating PbSnO₃ seeds that induce vertical grain growth and synergistically passivate defects (Figure 10c). Single-junction NBG PCE reached 23.12%, 2-T tandem 28.12%, and 4-T tandem 28.81% [149].

Wang et al. developed a confined-space annealing (CSA) strategy that slows solvent release to improve crystallinity and reduce defects (Figure 10d), achieving NBG single-junction PCE of 21.51% and tandem PCEs of 25.15% (4-T) and 25.05% (2-T) [150]. In situ growth of a BABr quasi-2D layer on CsPbI₂Br formed a 2D/3D heterostructure that passivates defects and enhances hole extraction (Figure 10b), giving a single-junction PCE of 16.57% and a monolithic tandem PCE of 10.22% [151]. Another study proposed an optimization method for photonic radiative annealing, which regulates crystal growth through rapid top-down heating and has been applied to all-perovskite TSCs. In traditional vacuum-assisted processes, the strong solvent extraction at the interface between the wet film and the vacuum leads to the formation of a thin perovskite layer on the top surface and residual solvents at the buried interface, ultimately causing film delamination and voids. However, the rapid top-down heating of photonic radiative annealing can precisely control the solvent evaporation rate, avoid uneven solvent distribution at the interface, and guide the unidirectional growth of crystals. As a result, high-quality Sn-Pb perovskite films with more uniform composition, fewer grain boundary defects, no obvious delamination, and no buried voids are obtained, reducing bulk and interfacial non-radiative recombination (Figure 9f). The performance of the optimized devices has achieved a significant breakthrough. The single-junction vacuum-assisted Sn-Pb perovskite solar cell has achieved a PCE of 22.31%, and the efficiency of the all-perovskite TSC has been further increased to 27.62% [144]. Table 8 summarizes the WBG and NBG perovskite optimization strategies and the photovoltaic performance of corresponding all-perovskite TSCs. WBG triple-halide passivation restrains I/Br segregation. Sn-Pb NBG antioxidant additives are mandatory. Dual bulk-surface treatment simultaneously lifts the performance of both sub-cells.

Table 8. WBG and NBG perovskite optimization strategies and the PV performance of corresponding all-perovskite TSCs.

Device Architecture	Active Area (cm2)	PCE (%)	Jsc (mA/cm2)	Voc (V)	FF (%)	Stability	Year [Ref.]
ITO/NiOx/VNPB/WBG perovskite (1.8 eV, DMA0.1Cs0.4FA0.5Pb(I0.75Br00.25)2.85Cl0.15)/C60/BCP/tunnel recombination junction (TRJ)/NBG perovskite (1.2 eV, MA0.3FA0.7Pb0.5Sn0.5I3)/C60/BCP/Cu	/	26.2	16.0	2.046	80.1	Retains 90% of initial PCE after 1045 h under encapsulated condition, ambient air, AM 1.5G illumination (100 mW/cm2), and MPP tracking.	2022 [132]
ITO/PTAA/PFNBr/WBG perovskite (1.73 eV, FA/Cs-based mixed-halide, PbCl2 + PMACl co-doped)/C60/SnO2/ITO/PEDOT:PSS/NBG perovskite (1.25 eV, Sn-Pb mixed)/C60/BCP/Cu	/	26.68	15.29	2.115	82.5	Retains 85% of initial PCE after 300 h for unencapsulated devices under MPP tracking, 1-sun LED illumination, and inert atmosphere.	2023 [133]
ITO/NiOx/Me-4PACz/WBG perovskite (1.79 eV, Cs0.2FA0.8Pb(I0.6Br0.4)3, PDA-treated)/C60/ALD-SnO2/1 nm Au/PEDOT:PSS/NBG perovskite (1.22 eV, Cs0.05FA0.7MA0.25Pb0.5Sn0.5I3)/C60/ALD-SnO2/Ag	0.049	27.4	15.05	2.19	83.1	Retains 86% of initial PCE after 500 h for encapsulated devices under MPP tracking, AM 1.5G illumination, and ambient air.	2023 [134]
Top cell (semi-transparent WBG): Glass/ITO/MeO-2PACz/WBG perovskite (1.66 eV, MA-free, DBSA-doped)/C60/BCP/Cu. Bottom cell (NBG): Glass/ITO/PEDOT:PSS/NBG perovskite (1.25 eV, Sn-Pb mixed)/C60/BCP/Cu. Mechanically stacked	/	28.06	19.78	1.23	82.79	Retains over 85% of initial PCE after 650 h for unencapsulated WBG single-junction devices under MPP tracking, AM 1.5G illumination (100 mW/cm2), and N2 atmosphere at ≈55 °C.	2024 [135]
Glass/ITO/NiOx/Me-4PACz/WBG perovskite (1.77 eV, FA/Cs-based mixed-halide, DDPA-doped)/C60/BCP/PEDOT:PSS/NBG perovskite (1.25 eV, Sn-Pb mixed)/C60/BCP/Cu	/	27.41	15.53	2.14	82.28	Retains over 85% of initial PCE after 16 days unencapsulated storage in air (40–80% RH) due to DDPA’s hydrophobic long-alkyl group.	2024 [136]
ITO/NiOx/2PACz/WBG perovskite (1.78 eV, FA0.8Cs0.2Pb(I0.6Br0.4)3, 2D/3D structure with 7.5 mM F-PEAI)/C60/SnO2/ITO nanocrystals/ALD-SnO2/PEDOT:PSS/NBG perovskite (1.22 eV, FA0.8Cs0.2Pb0.5Sn0.5I3, 980 nm)/C60/SnO2/Cu	/	27.0	15.8	2.11	81.0	Retains 94.0% of initial PCE after 500 h for unencapsulated devices under 1-sun MPP tracking and N2 atmosphere, outperforming control (retains 51.1%).	2024 [137]
ITO/NiOx/Me-4PACz/WBG perovskite (1.77 eV, FA0.8Cs0.2Pb(I0.6Br0.4)3, piracetam-doped)/C60/ALD-SnO2/1 nm Au/PEDOT:PSS/NBG perovskite (1.25 eV, FA0.7MA0.3Sn0.5Pb0.5I3)/C60/BCP/Cu	0.07	28.71	16.19	2.18	81.35	Retains 90% of initial PCE after 663 h for unencapsulated devices under MPP tracking, 1-sun illumination, and N2-filled glovebox at ~55 °C.	2025 [138]
Top cell (semi-transparent WBG): Glass/ITO/MeO-2PACz/WBG perovskite (1.66 eV, FA0.75Cs0.25PbI2.4Br0.6, DBrT-doped)/C60/BCP/Cu. Bottom cell (NBG): Glass/ITO/PEDOT:PSS/NBG perovskite (1.26 eV, Sn-Pb mixed)/C60/BCP/Cu. Mechanically stacked	/	28.31	19.79 (top cell); 11.72 (bottom cell, filtered)	1.23 (top cell); 0.83 (bottom cell, filtered)	83.77 (top cell); 81.82 (bottom cell, filtered)	Retains 90% of initial PCE after 500 h for unencapsulated WBG single-junction devices under 1-sun MPP tracking and N2 atmosphere, outperforming control (retains 80%).	2025 [139]
ITO/NiOx/Me-4PACz/WBG perovskite (1.79 eV, FA/Cs-based mixed-halide, 5-AVAI-doped)/C60/ALD-SnO2/Au/PEDOT:PSS/NBG perovskite (1.25 eV, Sn-Pb mixed)/C60/BCP/Ag	/	28.9	15.74	2.183	83.6	Retains 87.3% of initial PCE after 800 h for encapsulated devices under MPPT, AM 1.5G illumination, and ambient conditions.	2025 [140]
ITO/4PADCB/WBG perovskite (1.77 eV, FA0.8Cs0.2Pb(I0.6Br0.4)3, p-OHPEAI-doped)/C60/ALD-SnO2/IZO/PEDOT:PSS/LBG perovskite (1.25 eV, Sn-Pb mixed)/C60/BCP/Cu	/	28.50	16.07	2.155	82.3	Retains 90% of initial PCE after 350 h for unencapsulated devices under MPP tracking, 100 mW/cm2 LED illumination, and inert atmosphere.	2025 [141]
Top cell (semi-transparent WBG): Glass/ITO/NiOx/MeO-2PACz/WBG perovskite (1.78 eV, Cs0.2FA0.8Pb(I0.6Br0.4)3)/C60/SnO2/ITO. Bottom cell (NBG): Glass/ITO/PEDOT:PSS/PEA2PbI3SCN/Pb-Sn perovskite (1.27 eV, Cs0.1MA0.3FA0.6Pb0.5Sn0.5I3, AME-doped, EDA(I/Cl)2 post-treated)/PCBM/BCP/Ag	/	27.1	16.9	1.23	82.14	Retains stable PCE during 1000 s MPP tracking under AM 1.5G illumination, with negligible decay compared to control device’s rapid drop to <40% initial value.	2024 [145]
TO/PTAA/WBG perovskite (1.78 eV, Cs0.2FA0.8PbI1.8Br1.2)/C60/ALD-SnO2/1 nm Au/PEDOT:PSS/NBG perovskite (1.25 eV, FA0.7MA0.3Sn0.5Pb0.5I3, BV post-treated)/C60/BCP/Ag	0.06	26.33	15.78	2.12	78.7	Retains 88.3% of initial PCE after 1000 h for unencapsulated devices under MPP tracking, AM 1.5G illumination, and argon-filled glovebox.	2024 [146]
Glass/ITO/NiOx/Me-4PACz/WBG perovskite (1.78 eV, Cs0.2FA0.8Pb(I0.6Br0.4)3)/C60/ALD-SnO2/Au/PEDOT:PSS/NBG perovskite (1.25 eV, FA0.7MA0.3Sn0.5Pb0.5I3, TFPHCl-doped)/C60/BCP/Ag	/	27.42	15.64	2.11	83.08	Retains 80% of initial PCE after 450 h for unencapsulated devices under MPP tracking, continuous illumination, and 25 °C.	2024 [143]
Glass/ITO/NiOx/SAM (Me-4PACz:MeO-2PACz = 1:1)/WBG perovskite (1.79 eV, FA0.8Cs0.2Pb(I0.62Br0.38)3)/C60/ALD-SnO2/1 nm Au/PEDOT:PSS/NBG perovskite (1.25 eV, (FASnI3)0.6(MAPbI3)0.4, CNL-doped)/C60/ALD-SnO2/Ag	0.077	28.48	15.77	2.153	84.58	Retains >90% of initial PCE after 750 h for encapsulated devices under MPPT, AM 1.5G illumination, and 25 °C in ambient air.	2024 [147]
ITO/NiOx/Me-4PACz/WBG perovskite (1.78 eV, Cs0.2FA0.8Pb(I0.6Br0.4)3)/C60/SnO2/Au/PEDOT:PSS/NBG perovskite (1.27 eV, FA0.6MA0.3Cs0.1Pb0.5Sn0.5I3, TA-doped)/C60/BCP/Ag	/	28.6	16.75	2.10	81.57	Retains 80.5% of initial PCE after 300 h for encapsulated devices under MPP tracking, AM 1.5G illumination, and N2 environment.	2025 [148]
FTO/2PACz/WBG perovskite (1.77 eV, FA0.6MA0.4Pb(I0.6Br0.4)3, 400 nm)/C60/ALD SnO2/graphene oxide (GO)/PEDOT:PSS/NBG perovskite (1.25 eV, FA0.6MA0.4Sn0.6Pb0.4I3, 900 nm, Ni-reduced)/C60/BCP/Cu	/	28.13	16.81	2.038	82.10	Retains ~98% of initial PCE after 300 s for encapsulated devices under 1-sun MPP tracking and ambient air.	2025 [142]
ITO/NiOx/Me-4PACz/WBG perovskite (1.77 eV, FA0.8Cs0.2Pb(I0.6Br0.4)3, K2SnO3-doped)/C60/ALD-SnOx/0.8 nm Au/PEDOT:PSS (K2SnO3-modified)/NBG perovskite (1.25 eV, FA0.7MA0.3Pb0.5Sn0.5I3, K2SnO3-doped)/C60/BCP/Cu	0.070225	28.20	27.60	2.14	83.07	Retains 95.73% of initial PCE after 4038 h for unencapsulated devices under storage in N2-filled glovebox.	2025 [149]
Glass/ITO/PTAA/WBG perovskite (1.75 eV, FA0.8Cs0.2Pb(I0.7Br0.3)3, CSA-processed)/C60/ALD-SnO2/ITO/PEDOT:PSS/NBG perovskite (1.25 eV, (FASnI3)0.6(MAPbI3)0.4, CSA-processed)/C60/BCP/Ag	0.0625	25.05	15.41	1.956	83.1	Retains 90% of initial PCE after 450 h for unencapsulated devices under MPP tracking, AM 1.5G illumination, and N2-filled glovebox.	2022 [150]
ITO/SnO2/2D/3D CsPbI2Br (1.91 eV, BABr in situ modified)/spiro-OMeTAD/PEDOT:PSS/C60/MAPbI3 (1.59 eV)/spiro-OMeTAD/Ag	/	10.22	8.04	2.33	54.54	Retains 84% of initial PCE after 300 h for unencapsulated devices under dark storage, 85 °C, and N2 atmosphere.	2022 [151]
Glass/ITO/NiOx/4PADCB/WBG perovskite (1.78 eV, DMA0.1Cs0.5FA0.4PbI0.85Br0.15)/C60/ALD-SnO2/1 nm Au/PEDOT:PSS/NBG perovskite (1.25 eV, Cs0.1FA0.6MA0.3Pb0.5Sn0.5I3, LRA-annealed)/C60/BCP/Ag	-	27.62	15.98	2.143	80.65	Retains 80% of initial PCE after 100 h for encapsulated devices under MPP tracking, AM 1.5G illumination, and ambient air (vs. 19 h for HPA device).	2025 [144]

Figure 10. (a) Molecule structures and schematic of interface compensation. Reproduced with permission [140]. Copyright 2025, The Royal Society of Chemistry. (b) Schematic diagram of the preparation process and 2D/3D heterostructure of CsPbI₂Br solar cells. Reproduced with permission [151]. Copyright 2022, The Royal Society of Chemistry. (c) Schematic diagram illustrating the seed-induced oriented crystallization of K₂SnO₃-treated perovskite films. Reproduced with permission [149]. Copyright 2025, Springer Nature. (d) Scheme of normal annealing, solvent annealing and CSA processes. Reproduced with permission [150]. Copyright 2022, Springer Nature. (e) Schematic diagram of defect passivation and dipole moment generated by GABr/F-PEAI. Reproduced with permission [152]. Copyright 2024, Elsevier.

Figure 10. (a) Molecule structures and schematic of interface compensation. Reproduced with permission [140]. Copyright 2025, The Royal Society of Chemistry. (b) Schematic diagram of the preparation process and 2D/3D heterostructure of CsPbI₂Br solar cells. Reproduced with permission [151]. Copyright 2022, The Royal Society of Chemistry. (c) Schematic diagram illustrating the seed-induced oriented crystallization of K₂SnO₃-treated perovskite films. Reproduced with permission [149]. Copyright 2025, Springer Nature. (d) Scheme of normal annealing, solvent annealing and CSA processes. Reproduced with permission [150]. Copyright 2022, Springer Nature. (e) Schematic diagram of defect passivation and dipole moment generated by GABr/F-PEAI. Reproduced with permission [152]. Copyright 2024, Elsevier.

3.3.2. Interface Regulation

WBG I/Br mixed perovskites suffer from abundant interfacial defects, poor crystallinity, and photoinduced halide segregation. Various interfacial modification strategies have been developed to simultaneously passivate defects, align energy levels, and suppress ion migration. A four-source vacuum deposition technique produced FA_0.7Cs_0.3Pb(I_xBr_1−x)₃ films with precise bandgap control (1.62–1.80 eV). Combining MeO-2PACz as HTM and EDAI₂ surface passivation yielded a 2-T all-perovskite tandem PCE of 24.1% [153]. Zhou et al. adopted a synergistic bulk (MASCN) and surface (PEAI) passivation for 1.71 eV WBG perovskite made by blade coating with vacuum quenching. Together with a metal-free ALD-SnO₂ RL (93% transmittance, Figure 11b), the tandem device achieved 23.65% PCE [154]. Zhang et al. employed binary cations GABr and F-PEAI together with Al₂O₃ nanoparticles to improve wettability and passivate the WBG perovskite. The 1.80 eV WBG cell reached 19.0% PCE and the tandem device 27.2% PCE (Figure 10e) [152]. Chen et al. inserted 6-hydroxy-2-naphthalenecarboxylic acid (HNA) at the WBG/C₆₀ interface, building a cross-linked conjugated network that passivates defects and accelerates charge extraction, leading to a tandem PCE of 28.25% [155]. Modifying the buried HTL/perovskite interface with the silane coupling agent CN-TMOS gave a 1.77 eV WBG cell with 19.69% PCE and a tandem PCE of 28.45% [156]. Guo et al. used PACl at the SnO₂/perovskite buried interface to stabilize 1.68 eV MAPbI_2.4Br_0.6 through low-dimensional perovskite formation, relieving strain and suppressing halide segregation. The 4-T tandem reached 27.21% PCE [157]. A dual-molecule approach with BzMIMBr additive and DMFP surface passivation formed an A-B-A sandwich structure, yielding a WBG PCE of 22.77% and a 4-T tandem PCE of 26.18% [158].

For NBG Pb-Sn perovskites, interface strategies target severe non-radiative recombination, high carrier barriers, and Sn²⁺ oxidation. Sun et al. constructed an internal self-anchored dual-passivation (ISADP) layer by spin coating EDAI₂, followed by PCBM. The synergistic effect boosted the NBG cell PCE to 24.02% and the monolithic tandem to a certified 28.12% (28.61% champion) [159]. Song et al. modified the PEDOT:PSS/Pb-Sn perovskite interface with ammonium sulfide (AS), generating PbS nucleation sites, reducing acidity, and optimizing the energy landscape, achieving a single-junction PCE of 24.23% and a tandem PCE of 27.48% [160]. Benzyl viologen (BV) was introduced for surface charge transfer doping during the perovskite spin-coating process via anti-solvent engineering, utilizing the electron-donor property of BV to compensate for excess holes on the perovskite surface. BV molecules have an extremely low reduction potential, enabling them to transfer electrons to the perovskite surface and compensate for excess holes, converting the perovskite from heavily p-type doped to lightly p-type. BV forms coordination interactions with uncoordinated Sn²⁺ on the perovskite surface, inhibiting the oxidation of Sn²⁺ to Sn⁴⁺ and reducing Sn vacancy defects. Additionally, BV optimizes the energy-level alignment at the perovskite/C₆₀ interface, enhances the interfacial built-in electric field, and accelerates carrier extraction. Benefiting from these effects, the champion PCE of the single-junction NBG PSCs reached 22.58%. The champion PCE of the all-perovskite tandem device reached 26.33% [146]. Table 9 summarizes the interface modification strategies for WBG and NBG perovskites and the photovoltaic performance of the corresponding all-perovskite TSCs. 2D/3D heterojunctions suppress WBG interfacial loss, while reductive interface layers mitigate Sn²⁺ oxidation in NBG perovskites. Combined dual-side passivation gains the highest tandem V_oc.

Table 9. Interface modification strategies for WBG and NBG perovskites and the PV performance of the corresponding all-perovskite TSCs.

Device Architecture	Active Area (cm2)	PCE (%)	Jsc (mA/cm2)	Voc (V)	FF (%)	Stability	Year [Ref.]
ITO/MeO-2PACz/WBG perovskite (1.77 eV, FA0.7Cs0.3Pb(I0.64Br0.36)3, vacuum-deposited + EDAI2-passivated)/C60/ALD-SnOx/1 nm Au/PEDOT:PSS/NBG perovskite (1.28 eV, FA0.75Cs0.25Pb0.5Sn0.5I3, solution-processed + EDAI2-passivated)/C60/BCP/Cu	/	24.1	15.2	2.06	76.9	Retains 92% of initial PCE after 26 days for encapsulated devices under storage in ambient air.	2023 [153]
ITO/PTAA/WBG perovskite (1.71 eV, FA0.9Cs0.1Pb(I0.7Br0.3)3, MASCN-doped + PEAI-passivated)/PCBM/ALD-SnO2/PEDOT:PSS/NBG perovskite (1.27 eV, Sn-Pb mixed)/PCBM/BCP/Ag	/	23.65	14.66	2.05	78.7	Retains 80% of initial PCE after 1000 h for unencapsulated devices under storage in ambient air (50% RH).	2023 [154]
ITO/NiOx/Me-4PACz/WBG perovskite (1.78 eV, FA/Cs-based mixed-halide, HNA-modified)/C60/ALD-SnO2/Au/PEDOT:PSS/NBG perovskite (1.25 eV, Sn-Pb mixed)/C60/BCP/Ag	/	28.25	16.11	2.13	82.11	Retains 80% of initial PCE after 500 h for unencapsulated devices under MPP tracking, 1-sun LED illumination, and Ar-filled glovebox at 25 °C.	2025 [155]
ITO/4PADCB/Al2O3/CN-TMOS/WBG perovskite (1.77 eV, FA0.8Cs0.2Pb(I0.6Br0.4)3, PDAI2-passivated)/C60/ALD-SnO2/IZO/PEDOT:PSS/NBG perovskite (1.25 eV, Cs0.1FA0.6MA0.3Sn0.5Pb0.5I3, EDAI2-passivated)/C60/BCP/Ag	/	28.45	16.13	2.134	82.66	Retains 90% of initial PCE after 500 h for encapsulated devices under MPP tracking, 1-sun illumination, and N2-filled glovebox.	2025 [156]
Top cell (semi-transparent WBG): FTO/SnO2/PACl/WBG perovskite (1.68 eV, MAPbI2.4Br0.6, PACl-modified)/MoO3/Au/MoO3. Bottom cell (NBG): FTO/P3CT/NBG perovskite (1.23 eV)/TFB/MoO3/Ag	/	27.21	18.83	1.194	81.21	WBG sub-cell retains 90% of initial PCE after 2000 h for encapsulated devices under storage in N2-filled glovebox (O2/water < 0.01 ppm, 25 °C).	2025 [157]
Top cell (semi-transparent WBG): ITO/MeO-2PACz:Me-4PACz/WBG perovskite (1.68 eV, Cs/FA/MA-based mixed-halide, BzMIM Br-doped + DMFP-passivated)/PC60BM/BCP/Ag. Bottom cell (NBG): ITO/PEDOT:PSS/NBG perovskite (1.22 eV, Pb-Sn mixed)/C60/BCP/Ag	/	26.18	19.72	1.203	82.44	WBG sub-cell retains 95% of initial PCE after 1500 h for unencapsulated devices under MPP tracking, 1-sun illumination, and N2-filled glovebox.	2025 [158]
ITO/NiOx/SAM/WBG perovskite (1.77 eV, FA0.83Cs0.17Pb(I0.6Br0.4)3, PDAI2-passivated)/C60/ALD-SnO2/Au/PEDOT:PSS/NBG perovskite (1.24 eV, (FASnI3)0.6(MAPbI3)0.4, ISADP-modified)/C60/ALD-SnO2/Ag	0.07634	28.12	15.68	2.129	84.25	Retains 90% of initial PCE after 726 h for encapsulated devices under MPP tracking, AM 1.5G illumination, and ambient air.	2024 [159]
FTO/NiOx/Me-4PACz/WBG perovskite (mixed-halide)/C60/ALD-SnO2/Au/PEDOT:PSS (AS-modified)/NBG perovskite (1.25 eV, Pb-Sn mixed)/C60/BCP/Au	/	27.48	16.02	2.082	82.4	NBG sub-cell (AS-modified) retains 82% of initial PCE after 450 h for encapsulated devices under MPP tracking, 1-sun illumination, and ambient conditions.	2024 [160]
TO/PTAA/WBG perovskite (1.78 eV, Cs0.2FA0.8PbI1.8Br1.2)/C60/ALD-SnO2/1 nm Au/PEDOT:PSS/NBG perovskite (1.25 eV, FA0.7MA0.3Sn0.5Pb0.5I3, BV post-treated)/C60/BCP/Ag	0.06	26.33	15.78	2.12	78.7	Retains 88.3% of initial PCE after 1000 h for unencapsulated devices under MPP tracking, AM 1.5G illumination, and argon-filled glovebox.	2025 [146]

3.3.3. Transport-Layer Strategies

Traditional transport layers (e.g., PEDOT:PSS and PTAA) suffer from issues such as acid corrosion, strong parasitic absorption, and poor stability. Thus, strategies including the development of novel molecular transport layers and polymer modification can be adopted to upgrade the performance of transport layers while optimizing the interfacial compatibility with perovskites. Jiang et al. developed a novel solution-processed charge-recombination layer (CRL) with the architecture spiro-OMeTAD/PEDOT:PSS/PEI/PCBM:PEI, fabricated entirely via orthogonal-solvent low-temperature processing. Chlorobenzene served as the solvent for spiro-OMeTAD and PCBM:PEI, isopropanol for PEI, and the PEDOT:PSS layer was dry-transferred to prevent perovskite damage from aqueous solutions. The bottom perovskite sub-cell (with an Ag electrode) achieved a PCE of 11.7%, the top perovskite sub-cell (also with an Ag electrode) reached 11.4%, and the tandem device delivered a V_oc of 1.89 V. In this design, spiro-OMeTAD functions as the hole-collecting layer and PCBM:PEI as the electron-collecting layer. PEI modifies the PEDOT:PSS surface to establish a work-function gradient that creates efficient charge-recombination sites, while the phase separation between PEI and PCBM enhances electron transport. Moreover, the low-temperature process eliminates thermal damage to the bottom device (Figure 11e) [161]. Bi et al. introduced the single-molecule HTL 4dp3PACz, achieving a 1.77 eV WBG cell PCE of 17.17% and a tandem PCE of 26.47% [162]. PEDOT:PSS was replaced with Pb-doped P3CT (P3CT-Pb). Its high-pK_a carboxyl groups suppress deprotonation and inhibit acid erosion and iodine volatilization, and the Pb doping optimizes energy alignment and provides nucleation sites, giving an NBG cell PCE of 22.7% and a tandem PCE of 27.8% [163]. Replacing the Au layer with graphene oxide (GO) and PEDOT:PSS with 2PACz forms an SnO₂/GO/2PACz interconnect that reduces parasitic absorption, avoids acidic degradation, and improves stability, delivering a tandem PCE of 23.3% [164].

The interface matching and light transmittance between electrodes and transport layers directly affect the light absorption and charge-transport efficiency of TSCs. Strategies such as electrode material replacement, buffer-layer design, and HTL-free structure construction can be adopted to achieve the synergistic improvement of optical and electrical properties. Zhao et al. employed a 4-T mechanically stacked configuration with a semi-transparent 1.75 eV WBG top cell and a 1.25 eV LBG bottom cell, using MoO_x/ITO electrodes to minimize parasitic absorption, and achieved 23.1% [165]. In NBG (1.26 eV) Sn-Pb mixed PSCs, traditional HTLs suffer from parasitic absorption and significant NIR light loss in transparent electrodes. An HTL-free structure with hydrogen-doped indium oxide (IO:H) electrodes eliminated HTL parasitic absorption and enhanced near-infrared transmittance, yielding a 4-T tandem PCE of 24.8% [166]. To protect the C₆₀ and perovskite layers during ITO sputtering, Dong et al. spin-coated a PEI buffer layer, forming a C₆₀/PEI/ITO structure that converts into a conductive coordination compound, reduces damage, and optimizes energy alignment (Figure 11d), resulting in a 4-T tandem PCE of 26.89% [167]. Hu et al. developed a lamination process using a transparent conductive adhesive (TCA) with Ag-coated PMMA microspheres as the recombination junction, enabling independent sub-cell fabrication and self-encapsulation (Figure 11c), achieving a 2-T tandem PCE of 18.2% [168]. SCAPS-1D simulations for all-lead-free CsGeI₃/FASnI₃ tandem cells, with optimized transport-layer parameters and absorber thicknesses, predicted a theoretical PCE of 30.42% [169]. Table 10 summarizes the transport-layer, electrode, and interconnect-layer optimization strategies and the photovoltaic performance of the corresponding perovskite TSCs. Metal-free organic RLs avoid optical loss for all-perovskite stacks, while inorganic tunneling layers offer higher conductivity but risk damaging the underlying NBG film. All-perovskite devices require dual optimization of WBG segregation and NBG Sn oxidation, making composite passivation indispensable, while single-sided modification only brings limited efficiency gains.

Table 10. Transport-layer, electrode, and interconnect-layer optimization strategies and the PV performance of the corresponding perovskite TSCs.

Device Architecture	Active Area (cm2)	PCE (%)	Jsc (mA/cm2)	Voc (V)	FF (%)	Stability	Year [Ref.]
Glass/FTO/c-TiO2/m-TiO2/bottom perovskite (CH3NH3PbI3)/spiro-OMeTAD/PEDOT:PSS/PEI/PCBM:PEI/top perovskite (CH3NH3PbI3)/spiro-OMeTAD/hc-PEDOT:PSS (top electrode)	0.04	7.0	6.61	1.89	0.56	Low-temperature processing (<100 °C) minimizes thermal damage to underlying layers, maintaining structural stability.	2016 [161]
ITO/4dp3PACz/WBG perovskite (1.77 eV, FA0.8Cs0.2PbI1.8Br1.2)/C60/ALD-SnO2/IZO/PEDOT:PSS/NBG perovskite (1.25 eV, Cs0.025FA0.475MA0.5Sn0.5Pb0.5I2.925Br0.075)/C60/BCP/Ag	/	26.47	17.53	1.77	85.3	Retains 95% of initial PCE after 2000 h for unencapsulated devices under storage in N2-filled glovebox.	2023 [162]
ITO/NiOx/Me-4PACz/WBG perovskite (1.78 eV, FA0.8Cs0.2PbI1.8Br1.2, PDAI2-passivated)/C60/ALD-SnO2/Au/P3CT-Pb/LBG perovskite (1.26 eV, Cs0.05FA00.7MA0.25Pb0.5Sn0.5I3, EDA-passivated)/C60/BCP/Ag	0.073	27.8	15.7	2.147	82.6	Retains 97% of initial PCE after 1000 h for encapsulated devices under MPPT, 0.9-sun illumination, and ambient air (temperature < 30 °C).	2024 [163]
ITO/2PACz/WBG perovskite (1.74–1.75 eV, Cs0.25FA0.75Pb(I0.73Br0.27)3, PDAI2-passivated)/C60/ALD-SnO2/GO/2PACz/LBG perovskite (1.27–1.28 eV, Cs0.25FA0.75Pb0.5Sn0.5I3, EDAI2-passivated)/C60/ALD-SnO2/Cu	0.12	23.3	15.4	1.94	77.7	Retains initial PCE after 100 h for unencapsulated devices under MPP tracking, 1-sun illumination, and N2 atmosphere at 25 °C.	2025 [164]
Top cell (semi-transparent WBG): FTO/SnO2/C60-SAM/WBG perovskite (1.75 eV, FA0.8Cs0.2Pb(I0.7Br0.3)3)/spiro-OMeTAD/MoOx/ITO. Bottom cell (NBG): ITO/PEDOT:PSS/NBG perovskite (1.25 eV, (FASnI3)0.6(MAPbI3)0.4)/C60/BCP/Ag.	/	23.1	17.6	1.204	74.6	Maintains steady-state efficiency of 22.9% during MPP tracking under 100 mW/cm2 AM 1.5G illumination.	2018 [165]
Top cell (semi-transparent WBG): Glass/IO:H/np-SnO2/WBG perovskite (1.62 eV, Cs0.1(FA0.83MA0.17)0.9Pb(I0.83Br0.17)3)/spiro-MeOTAD/MoOx/IZO/MgF2. Bottom cell (HTL-free NBG): Glass/IO:H/NBG perovskite (1.26 eV, Cs0.025(FA0.8MA0.2)0.975Sn0.5Pb0.5I3)/PCBM/C60/BCP/Ag	/	24.8	20.8	1.159	78.7	IO:H electrode reduces Sn4+ formation at the interface, suppressing oxidation-induced degradation of NBG perovskite.	2021 [166]
FTO/TiO2 (ETL)/CsGeI3 (WBG perovskite, 1.6 eV)/spiro-OMeTAD (HTL)/ZnO (recombination layer)/FASnI3 (NBG perovskite, 1.41 eV)/spiro-OMeTAD (HTL)/Au	/	30.42	16.71	2.12	85.87	/	2024 [169]

Figure 11. (a) Schematic diagram of the growth process of WBG perovskite films. Reproduced with permission [138]. Copyright 2025, Springer Nature. (b) Schematic diagram of the vacuum-assisted blade coating. Reproduced with permission [154]. Copyright 2022, American Chemical Society. (c) Schematic of the fabrication process of the laminated architecture. Reproduced with permission [168]. Copyright 2025, American Chemical Society. (d) Optimized configuration of ITO capped on PEI fragment. Reproduced with permission [167]. Copyright 2024, Wiley-VCH GmbH. (e) Schematic of the as-prepared two-terminal bottom-up perovskite/perovskite TSC and chemical structures of the spiro-OMeTAD, PEDOT:PSS, PEI, and PCBM comprising the charge-recombination layer. Reproduced with permission [161]. Copyright 2016, The Royal Society of Chemistry.

Figure 11. (a) Schematic diagram of the growth process of WBG perovskite films. Reproduced with permission [138]. Copyright 2025, Springer Nature. (b) Schematic diagram of the vacuum-assisted blade coating. Reproduced with permission [154]. Copyright 2022, American Chemical Society. (c) Schematic of the fabrication process of the laminated architecture. Reproduced with permission [168]. Copyright 2025, American Chemical Society. (d) Optimized configuration of ITO capped on PEI fragment. Reproduced with permission [167]. Copyright 2024, Wiley-VCH GmbH. (e) Schematic of the as-prepared two-terminal bottom-up perovskite/perovskite TSC and chemical structures of the spiro-OMeTAD, PEDOT:PSS, PEI, and PCBM comprising the charge-recombination layer. Reproduced with permission [161]. Copyright 2016, The Royal Society of Chemistry.

3.4. Perovskite/Organic Solar Cell (OSC) TSCs

3.4.1. Perovskite-Layer Regulation

Perovskite-layer modulation constitutes the core of efficiency advancement in perovskite/organic TSCs, encompassing strategies such as compositional engineering, crystallization control, defect passivation, and phase-segregation suppression to optimize the optoelectronic properties of WBG front sub-cells. Wu et al. fabricated all-inorganic perovskite/organic TSCs using benzylmethylammonium chloride (PMACl) passivation on a CsPbI_2.1Br_0.9 top sub-cell, which effectively mitigated non-radiative recombination and, combined with orthogonal solvent processing, enabled a tandem efficiency of 18.06% with excellent stability [170]. Subsequently, a monolithic two-terminal perovskite/CIS TSC was developed, achieving current matching by modulating perovskite thickness and employing a low-bromine composition to suppress phase segregation, which enhanced photostability and carrier transport efficiency. Together with a NiO_x/2PACz HTL that prevented methylammonium degradation, the device reached a certified PCE of 23.5% [171].

Building upon these foundational efforts, Tang et al. employed solvent engineering via single-solvent NMP blade coating combined with 2D/3D heterojunction passivation. The NMP facilitated stable adduct formation to regulate crystallization, while the 2D/3D structure reduced non-radiative recombination (Figure 12f), enabling a 1.8 eV perovskite single-junction device to achieve 18.92% efficiency and a tandem device to reach 22.25% [172]. SCN⁻ was introduced via Pb(SCN)₂ to suppress halide phase segregation in WBG perovskite. SCN⁻ slowed crystallization to reduce grain boundaries, formed an I/Br/SCN alloy within the lattice, occupied iodine vacancies, and utilized steric hindrance to block halide migration, resulting in a certified tandem PCE of 25.06% and demonstrating a powerful approach to overcoming the long-standing phase segregation challenge [173]. Mali et al. fabricated n-i-p tandem cells by dynamic hot-air deposition using Rb-doped γ-Cs_0.99Rb_0.01PbI_2.2Br_0.8 and a PM6:Y6:PC₆₁BM rear cell. PTACl and PDIN passivation suppressed interfacial recombination (Figure 12g,h), achieving 23.07% PCE [174]. An imidazolium cation alloying strategy was employed by Chen et al. to regulate the crystallization of mixed-halide perovskites. With this approach, I/Br nucleation rates were balanced, the halogen distribution was homogenized, and defects were passivated by the IA cations, yielding a 1.79 eV perovskite sub-cell with a remarkable PCE of 19.50% [175]. The organic additive AIDCN was introduced to suppress halide phase segregation in a 1.86 eV WBG perovskite. The highly polar AIDCN molecules form hydrogen bonds with I⁻, which raises the energy barrier for photoinduced iodine escape and preserves the perovskite structure, thereby enabling a tandem device to achieve a certified PCE of 23.40% [176].

A PDADI post-treatment was employed to form high-n-value Dion–Jacobson (DJ)-phase quasi-2D perovskites on WBG absorbers. This treatment passivated shallow and deep defects, established a Type-II energy alignment, shortened the Pb-I interlayer spacing to strengthen electronic coupling, and weakened quantum confinement to promote carrier transport, ultimately yielding a tandem device with a PCE of 25.37% and an impressive FF of 82.82% [177]. Removable additives, specifically 1,4-dichlorobenzene, 1,4-dibromobenzene, and 1,4 -diiodobenzene, were employed to modulate the crystallization process of a 1.85 eV WBG perovskite. These additives slowed solvent evaporation, widened the processing window, reduced defects, and suppressed phase segregation (Figure 12i), enabling a perovskite single junction to achieve a V_oc of 1.370 V and a tandem PCE of 23.55% [178]. Moroxydine hydrochloride (MODCl) was introduced into a 1.85 eV I/Br WBG perovskite. Through its interactions with Pb²⁺ and halide ions, MODCl optimized the crystallization kinetics, reduced defects, and suppressed phase segregation (Figure 12j,k), yielding a WBG PSC with an efficiency of 18.76% and a tandem device with an outstanding PCE of 25.98% [179].

In summary, the evolution of perovskite-layer modulation strategies in perovskite/organic TSCs follows a clear trajectory, progressing from initial compositional optimization and empirical passivation to sophisticated mechanistic interventions. These include targeted suppression of phase segregation, regulation of crystallization kinetics through removable additives, and quasi-2D heterojunction engineering, which have propelled tandem efficiencies beyond 25% while simultaneously addressing the critical stability challenges associated with WBG mixed-halide perovskites. Low-temperature solvent engineering is the most suitable for organic bottom sub-cells. SCN additives effectively restrain halide segregation without harming organic active layers.

Figure 12. (a) The interfacial passivation mechanism and the molecular structure of PTB7-Th. (b) Device-architecture diagram. Reprint with permission [180]. Copyright 2023, Royal Society of Chemistry. (c) SEM cross-sectional images. (d) Perovskite film on the textured silicon substrate coated with TaTm/CuPc. Reprint with permission [181]. Copyright 2024, American Chemical Society. (e) Schematic diagram of material assembly and morphology characteristics. Reprint with permission [182]. Copyright 2025, Wiley-VCH. (f) Schematic diagram of vacuum-assisted blade coating. Reprint with permission [172]. Copyright 2023, Elsevier. (g,h) Schematic diagram of conventional all-perovskite inverted 2T monolithic tandem devices. Reprint with permission [174]. Copyright 2023, Royal Society of Chemistry. (i) Schematic illustration of the film formation processes for the control and the target film. Reprint with permission [178]. Copyright 2025, Royal Society of Chemistry. (j) Process diagram of a one-step spin-coating method of preparing WBG films. (k) The calculated electrostatic potentials of MOD. Reprint with permission [179]. Copyright 2025, Royal Society of Chemistry.

Figure 12. (a) The interfacial passivation mechanism and the molecular structure of PTB7-Th. (b) Device-architecture diagram. Reprint with permission [180]. Copyright 2023, Royal Society of Chemistry. (c) SEM cross-sectional images. (d) Perovskite film on the textured silicon substrate coated with TaTm/CuPc. Reprint with permission [181]. Copyright 2024, American Chemical Society. (e) Schematic diagram of material assembly and morphology characteristics. Reprint with permission [182]. Copyright 2025, Wiley-VCH. (f) Schematic diagram of vacuum-assisted blade coating. Reprint with permission [172]. Copyright 2023, Elsevier. (g,h) Schematic diagram of conventional all-perovskite inverted 2T monolithic tandem devices. Reprint with permission [174]. Copyright 2023, Royal Society of Chemistry. (i) Schematic illustration of the film formation processes for the control and the target film. Reprint with permission [178]. Copyright 2025, Royal Society of Chemistry. (j) Process diagram of a one-step spin-coating method of preparing WBG films. (k) The calculated electrostatic potentials of MOD. Reprint with permission [179]. Copyright 2025, Royal Society of Chemistry.

3.4.2. Interface Regulation

Interfacial engineering plays an equally vital role in determining the performance of perovskite/organic TSCs by minimizing non-radiative recombination, optimizing energy-level alignment, and facilitating efficient charge extraction across heterojunctions. An optoelectronic model combining Monte Carlo simulations with sensitivity analysis was developed to investigate two-terminal perovskite/organic TSCs, identifying key performance parameters and recombination losses. By optimizing the perovskite hole mobility and layer thicknesses, interfacial recombination was effectively reduced, yielding a device efficiency of 23.6% and outlining a pathway toward 28% efficiency [183]. High-color-fidelity semi-transparent perovskite/organic TSCs were developed by using a systematic selection method, in which the visible EQE fluctuation factor was employed to select CsPbI₂Br as the front sub-cell and the NIR absorption selection factor was used to evaluate PTB7-Th:IEICO-4F as the rear sub-cell. Notably, the sulfur atoms in PTB7-Th bonded with Pb²⁺ to passivate halogen vacancies in CsPbI₂Br (Figure 13a), effectively reducing interfacial recombination and aiding charge extraction, ultimately delivering a PCE of 24.3% with enhanced color fidelity [180].

Moving toward more sophisticated interfacial modulation strategies, a multilayer HTL structure composed of TaTm/CuPc/Al₂O₃ was developed for PSCs and textured perovskite/silicon tandem cells. In this architecture, CuPc ensured conformal coating and improved wetting, TaTm enhanced electron blocking, and the ultrathin Al₂O₃ layer effectively suppressed non-radiative recombination while reducing surface recombination without compromising carrier extraction (Figure 13b–d). This synergistic interfacial engineering approach enabled a p-i-n PSC to achieve a PCE of 24.5% with an impressive FF [181]. Most recently, plant Photosystem I (PSI) was integrated with C₆₀ fullerene into a perovskite/organic TSC to address weak red-light absorption. PSI enhanced photon harvesting in the red spectral region, while C₆₀ improved charge extraction and reduced non-radiative recombination losses, yielding a tandem PCE of 20.24% [184]. Concurrently, two pyrazine-based molecules, EPCN and PCN, were designed, and the assembly of PCN was regulated via a metal coordination strategy. Coordination of PCN with zirconium enhanced charge transport and light scattering in the organic sub-cell, while PCN also effectively passivated perovskite defects and modulated the Fermi level at the interface (Figure 13e). This multifunctional interfacial layer enabled a four-terminal perovskite/organic TSC to achieve a PCE of 25.43%, representing one of the highest efficiencies reported in this category [182].

In other words, the progression of interfacial engineering strategies in perovskite/organic TSCs reflects a clear evolution from empirical optimization toward rational design, encompassing optoelectronic modeling, defect passivation, multilayer charge-transport architectures, and even bio-inspired integration, all of which contribute to pushing device efficiencies beyond 25% while simultaneously addressing stability and functional requirements such as color fidelity. Thin buffer layers prevent solvent erosion of organic films, while conjugated SAMs optimize energy alignment without sacrificing near-infrared transmission.

3.4.3. Transport-Layer Strategies

Transport-layer engineering plays a critical role in perovskite/organic TSCs by regulating interfacial energetics, suppressing parasitic absorption, facilitating carrier recombination, and enabling efficient current matching between sub-cells. Early efforts focused primarily on transparent conductive electrodes and RL design. Hydrogenated indium oxide (IO:H) electrodes were introduced via magnetron sputtering in four-terminal perovskite/CIS TSCs, where H⁺ passivates oxygen vacancies to reduce carrier scattering, and O₂ suppresses metallic indium formation to enhance transparency (Figure 13a). The resulting IO:H films exhibit reduced near-infrared parasitic absorption compared with IZO, enabling improved current matching and a PCE of 24.6% [185]. Subsequently, polyTPD was utilized as the hole-transport material in the interconnection layer of monolithic perovskite/organic TSCs. Its low optical absorption, appropriate hole mobility, and quasi-Ohmic contact reduce interfacial barriers, suppress hole accumulation (Figure 13b,c), and minimize voltage loss, leading to a PCE of 21.1% in CsPbI₂Br/PM6:Y6-BO devices [186].

Building on RL optimization, a metal-free interconnection layer was constructed by combining ALD-grown SnO₂ with diluted PEDOT:PSS, which avoids metal-induced parasitic absorption, ensures Ohmic contact, and allows the perovskite layer to filter ultraviolet light, achieving a tandem efficiency of 17.6% [187]. In parallel, from the perspective of HTL design, the dopant-free polymer PTQ10 was introduced as an HTL with interconnection functionality in CsPbI₂Br-based devices. Its deep HOMO improves energy-level alignment, while S, N, and F atoms passivate surface defects and maintain efficient carrier transport (Figure 13d), thereby reducing voltage loss and improving current matching [188]. Further molecular engineering employed D-π-A polymers (PE51–PE53) as HTLs, where the longer side chains in PE53 promote face-on packing, enhance hole mobility (Figure 13e), and suppress recombination, enabling a tandem PCE of 23.07% [189].

With increasing emphasis on synergistic electrical and optical regulation, both transport layers and active layer morphology were optimized in inorganic perovskite/organic TSCs. Cl@MZO was employed as the ETL to improve energy-level alignment and reduce defect density, while the PM6:BTP-eC9 ratio (1:1.8) was tuned to enhance absorption and balance current between sub-cells, resulting in a PCE of 24.07% [190]. More recently, transport-layer design has evolved toward integrated optical–electrical modulation. Evaporated MoO_x was replaced with solution-processed PEDOT:F in the interconnection layer, where its low surface energy controls Au diffusion to form smaller and more uniform nanoparticles (Figure 13f), shifting plasmonic absorption away from the near-infrared region and reducing optical loss while maintaining efficient recombination, achieving a PCE of 25.34% [191].

Molecular-level interfacial engineering has further improved transport properties. A volatile additive (TCB) strategy was developed to regulate SAM-based HTLs (Figure 13h), where noncovalent interactions induce ordered face-on stacking after additive evaporation, enhancing hole mobility, reducing interfacial recombination, and optimizing surface energy for improved upper-layer crystallinity and phase separation, leading to a tandem efficiency of 26.09% [192]. Meanwhile, PEDOT:PSS-based interconnection layers were optimized via 2PACz modification to tune surface coverage, which enhances charge extraction, balances recombination within the ICL (Figure 13g), and reduces parasitic absorption, resulting in a PCE of 25.26% with improved device stability [193]. These advances highlight the importance of simultaneously optimizing interfacial energetics and optical properties within transport and RLs.

Overall, transport-layer strategies in perovskite/organic TSCs have evolved from initial electrode and RL optimization toward molecular-level interfacial engineering and optical–electrical synergistic regulation, enabling efficient carrier transport, reduced energy loss, and tandem efficiencies exceeding 25%. Table 11 summarizes of device performance parameters of typical perovskite/organic TSCs. Dopant-free polymer transport materials match low-temperature organic processing best. Inorganic oxide layers improve mobility but introduce thermal-incompatibility issues. Perovskite/organic TSCs are limited by organic thermal instability, so low-temperature crystallization and mild interface treatment are the only viable high-performance strategies.

Table 11. Summary of device performance parameters of typical perovskite/organic TSCs.

Device Architecture	Active Area (cm2)	PCE (%)	Jsc (mA/cm2)	Voc (V)	FF (%)	Stability	Year [Ref.]
Glass/IO:H/SnO2/(Cs0.1FA0.9PbI3)0.97(MAPbBr3)0.03/spiro-MeOTAD/MoOx/IZO/Ni-Al grid/MgF2	0.3	17.3	22.9	1.09	70.09	/	2020 [185]
Glass/ITO/SnO2/perovskite (≈1.73 eV)/C60/SnO2/IZO/CIS/CdS/i-ZnO/AZO/Ni/Al/Ni	0.16	24.6	14.7	2.05	81.5	/	2022 [171]
ITO/SnO2/CsPbI2Br/PTQ10/MoO3/Ag	/	17.8	15.24	1.4	83.2	Unencapsulated device retained 98.3% after 1112 h in glovebox and 94.7% after 532 h in air (RH < 10%).	2022 [188]
ITO/SnO2/WBG perovskite (1.84 eV, solvent-engineered)/C60/BCP/ Ag/MoO3/PM6:Y6:PC61BM/PDINN/Ag	1.02	22.1	13.4	2.05	80.4	Retained ~90% after 300 h storage in N2.	2023 [172]
ITO/SnO2/CsPbI2Br/PE53/MoO3/Ag	/	17.65	15.07	1.42	82.48	Unencapsulated device retained ~94% of initial PCE after 840 h (RH < 10%).	2023 [189]
ITO/SnO2/WBG perovskite/C60/BCP/Ag/MoO3/PEDOT:PSS/PM6:Y6/MoO3/Ag	/	24.3	14.7	2.05	80.6	Unencapsulated device retained ≈90% of initial PCE after 1000 h storage in N2.	2023 [176,180]
ITO/SnO2/WBG perovskite (≈1.8 eV, stabilized composition)/C60/BCP/Ag/MoO3/PM6:Y6/PDINN/Ag	/	25.3	14.4	2.18	80.5	Maintained >85% after 1000 h MPP tracking (encapsulated, 1 sun).	2024 [176]
ITO/Cl@MZO/CsPbI2Br/PM6/MoO3/Ag/PFN-Br/PM6:BTP-eC9/MoO3/Ag	/	24.07	13.02	2.15	81.36	Unencapsulated device retained 93% of initial PCE after 500 h storage in air (<15% RH).	2024 [190]
ITO/SnO2/WBG perovskite/C60/BCP/Ag/MoO3/organic active layer/MoO3/Ag	/	25.82	15.0	2.19	78.6	Maintained >85% of initial PCE after 1000 h MPP tracking.	2024 [173]
Glass/FTO/c-TiO2/mp-TiO2/g-Cs0.99Rb0.01PbI2.2Br0.8(PTACl)/PBDB-T/MoO3/Ag/ZnO/PDIN/PM6:Y6:PC61BM/MoO3/Ag	0.09	23.07	13.65	2.11	80.1	Retained ~90% after 200 h MPP tracking (encapsulated, 1 sun).	2024 [174]
ITO/SnO2/WBG perovskite/C60/BCP/Ag/MoO3/organic active layer/MoO3/Ag	/	22.4	14.2	2.01	78.6	Retained ~85% of initial PCE after 500 h storage in N2.	2024 [175]
Glass/ITO/ZnO/SnO2/CsPbI2Br/PEDOT:F/Au/ZnO/PFN-Br/PM6:L8-BO:BTP-eC9/SAM//Ag	0.038	25.34	14.98	2.23	75.75	Retained 97.9% of initial PCE after >4000 h storage in N2 (dark).	2025 [191]
ITO/Me-4PACz/WBG perovskite/C60/BCP/Ag/MoO3/TCB-treated SAM/PM6:BTP-eC9/C60/BCP/Ag	/	26.09	14.95	2.131	81.9	/	2025 [192]
ITO/Me-4PACz/Al2O3/Cs0.2FA0.8PbI1.7Br1.3/C60/SnO2/Ag/PEDOT:PSS/2PACz/BHJ(PM6:BTP-eC9)/PNDIT-F3N/Ag	0.04	25.26	14.93	2.05	82.33	Retained over 86% of the initial PCE after 1600 h of storage in a nitrogen atmosphere.	2025 [193]
ITO/SnO2/Perovskite/Photosystem I-modified interlayer/PEDOT:PSS/organic active layer/MoO3/Ag	/	21.8	13.6	2.03	78.9	Retained ~90% of initial PCE after 500 h storage in N2.	2025 [184]
ITO/NiOx/2PACz/perovskite/C60/BCP/Ag/MoO3/2PACz/PM6:D18:BTP-eC9/C60/BCP/Ag	/	27.3	15.1	2.19	82.4	Retained ~90% of initial PCE after 1000 h continuous illumination (N2).	2025 [177]
ITO/4PADCB/Al2O3/WBG perovskite (1.85 eV, with DIB additive)/PCBM/C60/SnOx/PEDOT:PSS/NBG organic (PM6:Y6:PCBM)/PDINN/Ag	/	23.55	13.87	2.147	79.08	Retained ~80% after 350 h MPP tracking (ambient, 25 °C, 50–80% RH).	2025 [178]
FTO/SAM/WBG perovskite (1.85 eV, MODCl-doped)/C60/ALD-SnOx/Au/PEDOT:PSS/PM6:BTP-eC9/PNDIT-F3N/Ag	/	25.98	14.56	2.21	80.76	T80 after 500 h MPP tracking (unencapsulated, 1 sun).	2025 [179]

Figure 13. (a) Scheme of the 4-T perovskite/CIS TSCs. Reprint with permission [185]. Copyright 2020, American Chemical Society. (b) Device structure, cross-sectional scanning electron microscopy image. (c) Hole-transport materials: polyTPD, PBDB-T-Si, and D18. Reprint with permission [186]. Copyright 2021, American Chemical Society. (d) Device structure of CsPbI₂Br PSCs and molecule structures of PTQ10 and P3HT. Reprint with permission [188]. Copyright 2022, Elsevier. (e) Device structure of TSCs and the molecule structure of D18 and Y6. Reprint with permission [189]. Copyright 2022, Elsevier. (f) Schematic illustration of possible mechanism. Reprint with permission [191]. Copyright 2025, Springer Nature. (g) Schematic illustration of the coverage state and in situ washing of 2PACz. Reprint with permission [193]. Copyright 2021, American Chemical Society. (h) Schematic plot showing that a more ordered stacking with a face-on feature formed after TCB treatment. Reprint with permission [192]. Copyright 2025, Royal Society of Chemistry.

Figure 13. (a) Scheme of the 4-T perovskite/CIS TSCs. Reprint with permission [185]. Copyright 2020, American Chemical Society. (b) Device structure, cross-sectional scanning electron microscopy image. (c) Hole-transport materials: polyTPD, PBDB-T-Si, and D18. Reprint with permission [186]. Copyright 2021, American Chemical Society. (d) Device structure of CsPbI₂Br PSCs and molecule structures of PTQ10 and P3HT. Reprint with permission [188]. Copyright 2022, Elsevier. (e) Device structure of TSCs and the molecule structure of D18 and Y6. Reprint with permission [189]. Copyright 2022, Elsevier. (f) Schematic illustration of possible mechanism. Reprint with permission [191]. Copyright 2025, Springer Nature. (g) Schematic illustration of the coverage state and in situ washing of 2PACz. Reprint with permission [193]. Copyright 2021, American Chemical Society. (h) Schematic plot showing that a more ordered stacking with a face-on feature formed after TCB treatment. Reprint with permission [192]. Copyright 2025, Royal Society of Chemistry.

4. Challenges, Opportunities, and Future Perspectives of Perovskite TSCs

4.1. Core Challenges

Material-related challenges: Despite significant progress in perovskite TSCs, several material-related challenges persist. A primary concern is the instability of perovskite materials, which can cause device performance to degrade over time. Perovskites are inherently sensitive to humidity, temperature, and light exposure. Therefore, developing stable compositions and robust device architectures is essential for commercialization. Another challenge is the toxicity of certain constituents, most notably lead (Pb). As a toxic heavy metal, lead raises environmental and health concerns. To address this, substantial efforts have been directed toward lead-free perovskites based on tin (Sn), bismuth (Bi), or antimony (Sb). However, the performance of these lead-free alternatives still lags behind that of their lead-based counterparts. For example, recent reviews have noted that, while tin-based perovskites offer favorable bandgaps, their efficiency remains limited by large V_oc and FF losses [194], and further research is needed to improve their optoelectronic properties and stability. Scalability of perovskite synthesis represents an additional hurdle. Current laboratory-scale methods, such as solution processing and vapor deposition, are not readily transferable to high-volume production. The development of scalable, cost-effective routes to high-quality perovskite materials is therefore critical for commercial viability.

Intrinsic fundamental limitations of four tandem architectures: From the perspective of intrinsic structural and material limitations, the four mainstream perovskite tandem architectures each face distinct fundamental bottlenecks. Perovskite/silicon TSCs are constrained by bandgap mismatch (~1.6–1.7 eV vs. 1.12 eV of Si) causing current mismatch, and textured Si substrates inevitably induce uneven film growth and sputtering damage. All-perovskite TSCs suffer from light-induced halide segregation in WBG perovskites and rapid Sn²⁺ oxidation in NBG Pb-Sn perovskites, while the ~0.5 eV bandgap offset limits theoretical efficiency. Perovskite/CIGS devices are hindered by lattice mismatch and dense interfacial traps, with the narrow CIGS bandgap (~1.0 eV) forcing a current–voltage trade-off. Perovskite/organic TSCs are fundamentally limited by weak NIR absorption, low carrier mobility of organic layers, and thermal incompatibility between solution-processed perovskites and vacuum-deposited organic films. These architecture-specific constraints cannot be completely resolved by conventional passivation or compositional tuning, and remain the root obstacles to further efficiency breakthroughs. Beyond these inherent limitations, additional material-, fabrication-, and cost-related challenges persist.

Device fabrication and integration challenges: Fabricating and integrating perovskite TSCs also presents considerable difficulties. One major challenge is the development of efficient, stable RLs to connect the two sub-cells. The inter-sub-cell interface can introduce significant recombination losses, diminishing tandem efficiency. As highlighted by Li et al. in their review of perovskite/silicon TSCs, charge-transport inefficiencies in interconnecting layers remain a key obstacle to optimal device performance [195]. High-quality RLs with low resistance and excellent charge-transfer properties are therefore essential. Another fabrication challenge is controlling the thickness and uniformity of the perovskite and charge-transport layers. Device performance is highly sensitive to these parameters, and even small variations can cause substantial degradation. Achieving precise thickness and uniformity over large areas is especially demanding for mass production. Advanced deposition techniques such as atomic layer deposition and slot-die coating show promise, but further optimization is required to ensure consistent performance across large-area devices [196]. Module-level integration introduces further complexity. Connecting multiple cells in series or parallel can generate additional losses from interconnections, edge effects, and non-uniformity. Reliable and efficient module designs and fabrication processes are therefore needed to translate the high efficiencies of small-area tandem cells into practical large-scale devices.

Cost and commercialization barriers: Although perovskite TSCs offer impressive efficiency potential, their commercialization faces several cost-related barriers. Materials, fabrication, and module integration must be cost-competitive with established photovoltaic technologies such as silicon. While perovskite materials themselves can be low-cost and solution-processable, the need for high-purity materials, precise fabrication, and robust encapsulation to ensure long-term stability may increase overall costs. The absence of standardized, low-cost manufacturing protocols for large-area tandem devices remains a major obstacle [197]. Additionally, the relatively high cost of certain charge-transport materials (e.g., spiro-OMeTAD) and transparent electrodes (e.g., ITO) further hinders cost reduction [198,199]. Addressing these barriers requires not only material substitution but also process simplification and yield improvement.

Cross-system stability comparison of four tandem structures: The four perovskite tandem architectures exhibit obvious differences in long-term operational stability under unified ISOS standard test conditions. Perovskite/silicon TSCs possess the best overall stability. Silicon substrates have stable chemical properties and low ion-migration risks. After triple-halide passivation and interface optimization, encapsulated devices can retain over 80% initial efficiency after 1000 h damp heat aging. Perovskite/CIGS TSCs rank second. Flexible PI substrates introduce interfacial stress, accelerating ion diffusion, so unencapsulated devices degrade rapidly under high humidity. All-perovskite TSCs face the worst stability bottlenecks. WBG layers suffer light-induced halide segregation, while NBG Sn-Pb perovskites are prone to oxidation, leading to obvious efficiency decay within hundreds of hours without full inert encapsulation. Perovskite/organic TSCs are limited by organic-layer thermal phase separation. High temperature and continuous illumination cause irreversible absorption loss of organic sub-cells, restricting long-term output stability. In addition, for any single tandem system, composite modification combining component regulation and interface passivation always achieves far better stability improvement than adopting only one single optimization strategy. Different tandem architectures show completely different responses to identical modification strategies, determined by substrate thermal stability, lattice matching, and processing-temperature windows.

4.2. Future Perspectives

Material development: Future advances in perovskite TSCs will depend heavily on the development of new materials. A key focus area is lead-free perovskites with high efficiency and stability. Although progress has been made, their performance still trails that of lead-based systems. For instance, Li et al. recently observed that tin-based perovskites continue to suffer from poor crystallization quality, high defect densities, and severe non-radiative recombination at buried interfaces [200]. Further fundamental understanding and strategies to enhance carrier mobility and reduce defect densities are needed. Another direction is the exploration of perovskite compositions with tunable bandgaps beyond the current range, enabling more flexible tandem architectures and better current matching with diverse bottom-cell materials. The development of new charge-transport materials with high conductivity, transparency, and stability is equally important for minimizing losses in tandem devices [201].

Device engineering: Advances in device engineering will also play a vital role. Novel architectures, such as multi-junction TSCs with three or more sub-cells, could further boost efficiency by capturing a broader portion of the solar spectrum. Realizing this potential, however, demands precise bandgap control and efficient charge transfer between sub-cells. Interface optimization is another critical area. Improving interfacial quality can reduce recombination losses and enhance charge transport. Techniques such as surface passivation and interface engineering with SAMs are promising for addressing these issues [202].

Scale-up and manufacturing: To realize the commercial promise of TSCs, substantial progress in scale-up and manufacturing is necessary. High-throughput, low-cost large-area fabrication techniques must be developed. Roll-to-roll processing, widely used in flexible electronics, is a promising route for large-scale manufacturing, but challenges related to uniformity and reproducibility must be overcome. Reliable encapsulation techniques for large-area modules are equally critical for long-term stability. Encapsulation materials and processes must be cost-effective, compatible with high-volume production, and capable of providing effective protection against environmental degradation.

Integration and commercialization pathways: Integrating perovskite TSCs with other technologies will open new application opportunities. For example, combining tandem cells with energy storage systems could enable self-sustaining energy solutions, while integration with smart-grid technologies can facilitate more efficient solar energy management and distribution. Building-integrated photovoltaics (BIPV) represents another impactful area. The flexibility, light weight, and potential semi-transparency of perovskite TSCs make them ideal for facades, windows, and rooftops, turning buildings into active energy generators. As noted by Tiwari in a review of self-powered IoT devices, emerging photovoltaics including silicon/perovskite tandem cells can offer superior solutions for applications requiring energy autonomy and sustainable operation [203]. Beyond BIPV, portable electronics, drones, and space technologies constitute attractive niches where the high specific power of perovskite TSCs could be particularly advantageous. Although current manufacturing costs remain relatively high, continued progress in scalable fabrication and encapsulation may eventually enable cost-competitive utility-scale deployment. Furthermore, integrating perovskite top cells into existing silicon bottom-cell production lines offers a promising pathway toward commercial entry.

4.3. Standardization of Photovoltaic Stability Characterization Protocols

The current literature adopts diverse stability test conditions (MPP tracking, damp-heat aging, light soaking, and thermal cycling), leading to inconsistent T80/T90 lifetime metrics that hinder cross-study quantitative comparison. The International Stability Testing Operating Standards (ISOS) series provides unified standardized testing frameworks for perovskite photovoltaic devices, which should be uniformly followed in future experimental reports.

Key recommended reporting specifications are summarized as follows: Mandatory full environmental parameters: Test temperature, relative humidity (RH), illumination intensity (1 sun/AM1.5G), inert/air atmosphere, and encapsulation architecture (glass–glass, foil encapsulation, and unencapsulated) must be explicitly listed alongside PCE retention data (T80/T90, MPP retention hours).

For unified stability evaluation, ISOS tests are classified into three standard categories. ISOS-L is a continuous light soaking test conducted under operational MPP tracking, which is the test protocol most representative of real-world working conditions. ISOS-D refers to the dark damp-heat aging test performed at 85 °C and 85% relative humidity, serving as an accelerated test to characterize humidity-induced degradation. ISOS-T is the thermal cycling test, designed to simulate day-night temperature fluctuations that devices would experience in practical operation.

5. Conclusions

Perovskite TSCs offer a compelling pathway to surpass the Shockley–Queisser limit of single-junction photovoltaics. This review has systematically surveyed recent progress across four major configurations, namely perovskite/silicon, perovskite/CIGS, all-perovskite, and perovskite/organic TSCs, covering their working principles, optimization strategies, and remaining challenges.

Perovskite/silicon TSCs have achieved certified efficiencies beyond 34%, driven by compositional engineering, multifunctional defect passivation, crystallization control on textured substrates, and interface engineering with SAMs. These approaches have also improved device stability. Perovskite/CIGS TSCs have approached 30% efficiency through bandgap matching and interfacial modification, with promising results on flexible substrates. All-perovskite TSCs have surpassed 28% by effectively suppressing halide segregation in WBG cells and mitigating Sn-Pb oxidation in NBG cells. Perovskite/organic TSCs, though less mature, have rapidly advanced beyond 25% efficiency, benefiting from innovative interconnection layers, quasi-2D/3D heterojunctions, and molecular additives that stabilize mixed-halide perovskites.

Nevertheless, critical challenges persist. The inherent instability of perovskite materials under moisture, oxygen, thermal stress, and prolonged illumination demands robust encapsulation and intrinsically stable compositions. Halide phase segregation continues to impair the performance and lifetime of WBG perovskites. Lead toxicity necessitates the exploration of tin-based and other alternative materials, although their efficiencies and stabilities currently lag behind. Moreover, scaling up from laboratory cells to large-area modules, developing efficient and transparent RLs, and achieving cost competitiveness with established silicon technology remain essential for commercialization.

Future research should prioritize lead-free perovskites with improved optoelectronic properties, advanced encapsulation techniques such as atomic layer deposition, and novel charge-transport materials that minimize energy losses. Scalable manufacturing methods, including blade coating, slot-die coating, and roll-to-roll processing, must be refined for high-throughput production. Integrating tandem cells with energy storage systems, building-integrated photovoltaics, and portable electronics could further broaden their application scope.

In summary, perovskite TSCs represent one of the most promising routes to next-generation high-efficiency photovoltaics. Sustained interdisciplinary efforts in materials science, device engineering, and manufacturing will be key to overcoming existing barriers and accelerating the global transition to sustainable energy. The ultimate commercial success will depend on the synergistic advancement of efficiency, stability, and scalability.