Hybrid Perovskites and 2D Materials in Optoelectronic and Photocatalytic Applications

Abstract

Metal halide perovskites, emerging innovative and promising semiconductor materials with notable properties, have been a great success in the optoelectronic and photocatalytic fields. At the same time, two-dimensional (2D) materials, including graphene, transition metal dichalcogenides (TMDCs), black phosphorus (BP) and so on, have attracted significant interest due to their remarkable attributes. While substantial advancements have been made in recent decades, there are still hurdles in enhancing the performance of devices made from perovskites or 2D materials and in addressing their stability for reliable use. Recently, heterostructures combining perovskites with cost-effective 2D materials have exhibited significant advancements in both efficiency and stability, attributed to the unique properties at the heterointerface. In this review, we provide a thorough overview of perovskite and 2D material heterostructures, spanning from synthesis to application. We begin by detailing the diverse fabrication techniques, categorizing them into solid-state and solution-processed methods. Subsequently, we delve into the applications of perovskite and 2D material heterostructures, elaborating on their use in photodetectors, solar cells, and photocatalysis. We conclude by spotlighting existing challenges in developing perovskite and 2D material heterostructures and suggesting potential avenues for further advancements in this research area.

1. Introduction

Metal halide perovskites have surged to the forefront of research over the past decade. The cutting-edge optoelectronic devices built on these perovskites, such as photodetectors (PDs), light-emitting diodes (LEDs), perovskite solar cells (PSCs), are setting new benchmarks and leading current technological trends due to the intriguing properties of perovskites [1,2,3,4]. Since the birth of the first perovskite-sensitized solar cell with a power conversion efficiency (PCE) of 3.8% in 2009 [5], the best PSC exhibited 25.6% certified power efficiency, with 80% of that efficiency after 500 h of operation at 85 °C [6]. This exceptional device performance is primarily due to the unique features of perovskites, including superior optical absorption coefficient, high carrier mobility, long carrier diffusion length and slow carrier recombination rate [7,8,9]. Additionally, the solubility of perovskite materials offers a promising avenue for producing cost-effective, large-scale solar cell modules, giving them a competitive edge over conventional silicon-based alternatives. While the impressive physical attributes linked with their structural and functional versatility have propelled perovskite materials to success in optoelectronic fields, there are still hurdles to overcome in enhancing their efficiency and addressing stability concerns for upcoming commercialization [10]. Hybridizing perovskites with other functional materials, e.g., two-dimensional (2D) materials, has proven to be one promising way to achieve device performance and higher stability [11,12].

Two-dimensional materials such as graphene, transition metal dichalcogenides (TMDCs), black phosphorus (BP), and transition-metal carbides and nitrides (MXenes) have garnered extensive attention because of their exceptional qualities derived from distinct van der Waals (vdWs) configurations [13,14,15,16,17,18]. Unlike traditional bulk structures, where 2D materials are composed of distinct layers held together by strong in-plane covalent bonds, atomic-scale 2D materials manifest distinct and remarkable optical, electrical, thermal, chemical, and mechanical characteristics [19]. Moreover, the distinctive 2D nature of these materials allows for easy conversion into consistently thin films with highly aligned microstructures, making them ideal for optoelectronic applications. However, due to their atomically thin nature, 2D materials typically face challenges with limited light absorption [20].

Recently, the fusion of perovskites and 2D materials into heterostructures, blending two leading research areas, has captivated interest in multiple fields [21]. Integrating perovskites with 2D materials not only boosts the light absorption capacity of the extremely thin 2D materials, leveraging the exceptional optical properties of perovskites, but also strengthens the stability of perovskites, especially when 2D nanosheets act as protective encapsulation layers [22]. Additionally, the perovskite and 2D material heterostructures display distinctive characteristics such as ultrafast charge and energy transfer [23], photogating effect [24], strong interlayer coupling [25], and so on, which result in significantly improved optoelectronic and photoelectric properties compared to their components. More importantly, the vast array of perovskites and 2D materials offer excellent opportunities to design and tailor the compositions and functionalities of perovskite and 2D material heterostructures.

In this review, different fabrication techniques for perovskite and 2D material heterostructures are discussed, classifying them into solid-state approaches and solution-based processes. Then, a detailed overview of their applications in photodetectors, solar cells and photocatalysis is provided. Lastly, we discuss the obstacles to creating high-quality, stable, eco-friendly perovskite and 2D material heterostructures for broad-scale use. This review intends to offer researchers a structured update on novel fabrication methods and recent applications of perovskite and 2D material heterostructures, thereby advancing the progress of perovskite and 2D material heterostructures.

2. Fabrication Techniques of Perovskite and 2D Material Heterostructures

Perovskite and 2D material heterostructures can be prepared using either solid-state or solution techniques. Typically, solid-state approaches involve the use of vapor deposition to form these heterostructures, followed by dry-wet transfer onto designated substrates. The solid-state method is suitable for fabricating individual devices, studying the effects of the composition, thickness, etc., of different layers of the device on the device, and producing novel heterostructures for foundational studies. On the other hand, solution techniques, such as spin coating, ex situ hybridization and in situ growth, are acknowledged for their simplicity and cost-effectiveness, making them ideal for the large-scale production of these heterostructures. Moreover, using low-temperature processing in these solution methods can minimize energy use during fabrication and can lead to devices with mechanical flexibility. In this section, we summarize the techniques related to solid-state and solution methods for the fabrication of perovskite and 2D material heterostructures.

2.1. Solid-State Methods

Vapor deposition stands out as a prominent solid-state technique in creating perovskite and 2D material heterostructures. In this method, materials or volatile precursors are initially vaporized at elevated temperatures. Subsequently, this vapor phase is condensed onto a substrate, resulting in the desired layered structure where high-quality single-crystal and continuous thin films can be formed. The typical fabrication process for perovskite and 2D material heterostructures through vapor deposition involves initially laying down consistent and flat 2D material flakes using the chemical vapor deposition (CVD) approach, which is followed by a one- or two-step of vapor deposition of a perovskite layer. Niu and colleagues pioneered the use of the vapor deposition technique to create three distinct heterostructures: perovskite/hexagonal boron nitride (h-BN), perovskite/MoS₂, and perovskite/graphene [26]. Initially, they laid down 2D monolayers on SiO₂/Si platforms using CVD, followed by layering PbI₂ nanosheets onto the 2D monolayer via physical vapor deposition (PVD). In the final stage, they transformed the PbI₂ on the 2D monolayer into CH₃NH₃PbI₃ by facilitating a reaction with the vapor of methylammonium iodide. The scanning electron microscope (SEM) images in Figure 1a–c show that the perovskite crystals are located on the 2D material monolayer. The surface roughness ≈ 2.0 nm indicates the high quality of the perovskite and 2D material heterostructure. Intriguingly, the use of photolithography allows us to produce these heterostructures in varying shapes, catering to diverse requirements, and high-quality perovskite and 2D material heterostructures were successfully constructed via vapor deposition, facilitating their application in optoelectronics. Furthermore, heterostructures produced via the vapor deposition technique displayed enhanced photodetection capabilities. A notable case is the work by Lu et al., where they applied focused laser annealing to CVD-generated WSe₂ monolayers during the fabrication of CH₃NH₃PbI₃/WSe₂ heterostructures [27]. This post-treatment procedure allowed chalcogenide gaps in WSe₂ to be filled through oxygen substitution, boosting the conductivity of WSe₂. Consequently, the resultant PD demonstrated remarkably high photoresponsivity and quantum efficiency.

The dry/wet transfer method presents an alternative solid-state approach for creating perovskite and 2D material heterostructures, which involves transitioning a sample from one substrate to another, utilizing either a solid or liquid medium. The widely used transfer techniques are based on polymer thin films such as polymethyl methacrylate (PMMA) [29]. Due to the adhesive force between the polymer film and the sample, it is possible to detach the sample and relocate it onto a designated substrate in a specific order. Once transferred, the polymer film can be dissolved using typical organic solvents, placing the sample neatly on the target substrate. Using several cycles of the transfer process, a layered heterostructure can be constructed. In 2016, graphene/perovskite/graphene and graphene/WSe₂/perovskite/graphene heterojunctions were prepared for the first time via multiple dry/wet transfers [30]. For graphene/perovskite/graphene devices, the bottom graphene is first exfoliated onto an SiO₂/Si substrate, and the PbI₂ is then mechanically exfoliated and dry-transferred on top of graphene through a polymethyl methacrylate/polypropylene carbonate (PMMA/PPC). Then, PbI₂ is converted to CH₃NH₃PbI₃ after reaction with MAI steam, and the obtained CH₃NH₃PbI₃ perovskite is immediately covered by another graphene flake via PMMA. The fabrication process for graphene/WSe₂/perovskite/graphene devices parallels that of graphene/perovskite/graphene devices, with a distinction: an added dry transfer phase wherein WSe₂ flakes are layered onto the primary graphene monolayer using a PMMA/PPC stack. Additionally, to bolster protection for the device, an extra boron nitride (BN) layer can be introduced through dry-transfer. These multi-layered heterostructures prepared using solid-state techniques such as vapor deposition and dry-wet transfer are surface-cleaned, positioning them as promising platforms for delving into new physical, optical, and electronic characteristics. Nonetheless, these solid-state approaches come with inherent challenges. The intricacies of the fabrication procedures, which might encompass techniques such as CVD, PVD, mechanical extraction, numerous wet and dry transfer phases, and precise alignment, are considerable. Coupled with their energy-intensive prerequisites, these factors constrain the spectrum of applicable perovskites and 2D materials, and they also curb the mass production of high-grade structures.

2.2. Solution Methods

Spin coating stands as a facile technique favored in fabricating optoelectronics. Notably cost-effective and adaptable, it aligns seamlessly with a broad spectrum of 2D materials when paired with perovskites. By spin-coating perovskite and 2D material precursor solutions onto designated substrates, layered structures of perovskites and 2D materials can be constructed. The spin-coating approach paves the way for mass-producing these heterostructures economically. Three-dimensional perovskites on many two-dimensional material substrates, such as mechanically exfoliated MoS₂, mechanically exfoliated black phosphorus (BP), CVD graphene, CVD WS₂, and chemically exfoliated MoS₂, are predominantly realized via one-step spin-coating using perovskite precursor solutions. Beyond just layering structures, a one-step spin-coating method can be employed to introduce 2D materials like graphene derivatives, MXene, and BP directly into perovskite films. Studies have shown that integrating 2D materials with perovskites leads to larger grain sizes in the films as a result of a slower crystallization process [31]. Additionally, these 2D materials can help in reducing surface defects and enhance the conductivity of the perovskite films, thereby improving their performance in optoelectronic applications [32].

Another promising approach for the large-scale production of perovskite and 2D material heterostructures is through direct solution-based synthesis. By mixing pre-fabricated perovskite with 2D materials in solution, the two components can be combined through either chemical bonds or non-covalent interactions, resulting in heterogeneous micro or nano-structures. In 2018, Wu et al. prepared MAPbI₃/Reduced graphene oxide (RGO) micro-composites by blending MAPbI₃ and GO in a strongly acidic water-based solution [33]. Upon being subjected to visible light for a few hours, the mixture of MAPbI₃ and GO naturally transitioned into MAPbI₃/RGO micro-assemblies by the reduction process of GO to RGO. Given the rich presence of functional groups like hydroxyl, carbonyl, and carboxyl on the RGO surface and edges, a formation of Pb-O-C linkages are established between MAPbI₃ and RGO. In addition to chemical bonds induced by surface functional groups, non-covalent interactions also offer a strategy to construct the perovskite and 2D material heterostructures. However, producing perovskite crystals on the surface of TMDCs is intricate, primarily due to the pronounced difference in the surface chemistry between perovskites and TMDCs. Hassan and his team introduced a method that can be processed in a solution to produce MoSe₂-CsPbBr₃ nanocomposites, using 4-aminothiophenol (4-ATP) as the bridging agent between MoSe₂ and CsPbBr₃, as depicted in Figure 1d [28]. Initially, 4-ATP was added to MoSe₂ nanosheets rich in defects, enabling the passivation of free Mo atoms via chemical bonding. Subsequently, pre-fabricated CsPbBr₃ nanocrystals were integrated into the solution containing MoSe₂ nanosheets functionalized with 4-ATP. Note that the 4-ATP not only serves as a binder between TMDCs and CsPbBr₃ nanocrystals through electrostatic attraction but also ensures that MoSe₂ and CsPbBr₃ are closely in contact, promoting efficient charge transfer across the MoSe₂-CsPbBr₃ boundary.

Thermal injection offers a straightforward and rapid technique for producing perovskite nanocrystals, typically operating at elevated temperatures (around 120–200 °C) under inert gas conditions, boasting high in situ growth efficiency [34,35]. In the improved thermal injection technique for constructing heterostructures of perovskites and 2D materials, 2D nanosheets are typically introduced to the Pb precursor solution beforehand, resulting in a well-mixed solution. Following the addition of cesium oleate, the desired heterostructure of perovskites and 2D materials forms easily in the subsequent chemical reaction. Notwithstanding the advancements in this method, the synthesis demands high temperatures, often ranging from 120 to 200 °C, and the protection of an inert gas like nitrogen. These requirements contribute to significant costs and increased energy usage. As an alternative, the ligand-assisted reprecipitation (LARP) method offers a more efficient approach to synthesize perovskites in ambient conditions [36]. Typically, in LARP synthesis, a mixture of perovskite precursors, including lead halides, alkylammonium halides, and surfactants, is gradually introduced to a non-polar solvent, such as toluene, under agitation. This change in solvent polarity triggers the precipitation of perovskite nanocrystals. For the fabrication of heterostructures involving perovskites and 2D materials, 2D nanosheets can either be incorporated into the perovskite precursor mix or directly into the non-polar solvents. Later, the perovskites and 2D materials heterostructure can be formed when the perovskite precursor solution is merged with the non-polar solvent under stirring.

Achieving atomically precise epitaxial interfaces is crucial for enhancing device performance. Yet, the epitaxial growth of halide perovskites with such sharp interfaces remains unachieved. The inherent high ion mobility in these materials causes issues like interdiffusion, which broadens junction widths. Additionally, their chemical instability often results in the decomposition of layers during subsequent layer preparations [37,38]. Shi et al. report a method to considerably curb in-plane ion diffusion in 2D halide perovskites through the inclusion of rigid π-conjugated organic ligands [39]. They constructed two distinct lateral 2D halide perovskite heterostructures using varied organic ligands: one employed a conjugated ligand based on bithiophenylethylammonium (2T⁺), while the other utilized a more common alkyl ligand, butylammonium (BA⁺). Notably, the Br–I heterostructure in 3D perovskites lacks thermal stability owing to the high diffusivity of Br⁻ and I⁻ anions, and even a mild heating can initiate and speed up the interdiffusion of these halide anions across the heterojunction [40]. However, with the introduction of conjugated 2T ligands, the authors observed a marked reduction in interdiffusion between Br⁻ and I⁻ within the (2T)₂PbI₄–(2T)₂PbBr₄ heterostructures. Employing these rigid conjugated ligands in 2D perovskite frameworks has been shown to dramatically suppress ionic interdiffusion across 2D halide perovskite junctions. By inhibiting ion migration, they were able to produce stable interfaces with near-atomic precision through continuous epitaxial growth.

This section provides a brief overview of the various techniques used to fabricate heterostructures combining perovskites and 2D materials. These methods can be broadly classified into solid-state and solution-based methods. The solid-state method includes vapor deposition and the dry/wet transfer method. While these processes can yield high-quality heterostructures with precision, their prohibitive costs and complex procedures make them less suited for large-scale manufacturing and broader applications. On the other hand, the spin-coating technique offers a more practical alternative, given its simplicity, low-temperature requirements, and energy efficiency. This makes it an attractive method for producing heterostructures of perovskites and 2D materials. For those seeking to synthesize these heterostructures directly in solution, there are two main strategies available: ex situ hybridization and in situ growth. Ex situ hybridization is a direct and uncomplicated method that combines perovskite crystals with 2D materials either through chemical bonding or non-covalent interactions. However, this sometimes results in weaker contacts between the two components. Meanwhile, the in situ growth method allows for the reprecipitation of perovskite crystals onto 2D nanosheets in the solution, leading to a more integrated and optimized heterointerface. This method is especially promising for achieving high-quality heterostructures at an affordable cost and with a high yield.

Table 1. Summary of advantages and disadvantages of different fabrication techniques.

Fabrication Techniques	Advantages	Disadvantages
Solid-state methods	High-quality High surface cleanness Good thickness control Immediate device fabrication	Complex fabrication processes High consumption conditions Small-scale production
Solution methods	Facile processing Low temperature Low energy consumption Good scalability	Low quality and uniformity Limited thickness control Residual solvents and impurities

compares the advantages and disadvantages of different fabrication techniques.

3. Perovskite and 2D Material Heterostructure Photodetectors

In recent years, with the development of the optoelectronic information industry, there has been a heightened demand for enhanced performance of optoelectronic devices to cater to diverse applications [41]. Photodetection operates on the principle of translating absorbed light signals into electron–hole pairs. This conversion is facilitated by semiconductor materials possessing appropriate band gaps. Under the influence of either an inherent or externally applied electric field, these electron–hole pairs are collected via metal electrodes, resulting in the generation of electrical signals. The main indicators for evaluating photodetection are responsivity (R), detectivity (D*), on/off ratio, external quantum efficiency (EQE), photoconductive gain (G), noise equivalent power (NEP), and response speed (τ_rise, τ_decay). R is the ratio of the photocurrent to the incident light power, expressed as R = I_ph/P_in, where I_ph is the photocurrent and P_in is the incident light power. Its value depends on the wavelength of the light, as the absorption of photons is related to the band gap of the semiconductor. EQE is the number of electron–hole pairs collected divided by the number of incident photons, that is, EQE = (I_ph/e)/(P_in/E_ph), where e is the charge of the electron and E_ph is the energy of the photon. Photoconductive gain evaluates the ability of a photodetector (PD) to generate charge carriers in response to a single photon. The response speed is usually characterized by the light pulse’s rise and decay/fall times on and off, i.e., the time interval required for the output signal to increase from 10% to 90% and decrease from 90% to 10% of its on-state maximum value, respectively. D is defined as the inverse of the noise equivalent power, and its normalized result to the active detection area and detection bandwidth, called D*, has been widely used to quantify the sensitivity of PDs to noise signals. D* = R/(2eJ_d)^1/2, where J_d is the dark current density in Jones or cm Hz^1/2 W⁻¹.

Perovskites have developed tremendously in the past time, and the performance of PDs made of a single material has not been greatly improved. Two-dimensional materials have the advantages of high carrier mobility, a rich and tunable band gap, fast photoresponse speed, and wide spectral response range [41]. Therefore, researchers use perovskite and two-dimensional material heterostructures to construct PDs, and the performance of PDs has been greatly improved. In this section, we will introduce perovskite heterostructure PDs from the discussion of perovskite/graphene, perovskite/TMDCs, perovskite/BP, and perovskite/MXene.

3.1. Perovskite/Graphene Heterostructures

Perovskites offer significant advantages such as a high light absorption coefficient, extended carrier diffusion length, and superior carrier mobility, positioning them as prime candidates for light harvesting in PDs [42]. However, a primary limitation of these PDs is the rapid recombination of photo-induced electron–hole pairs within few picoseconds in perovskites, preventing their collection by the electrodes. To address this, introducing graphene, known for its quick response time and exceptional carrier mobility, can serve as an efficient carrier transport material in PDs. Graphene has attracted enormous attention since its discovery. The carrier mobility of graphene is as high as 15,000 cm² v⁻¹s⁻¹, and due to its zero-bandgap nature, graphene has an ultrabroad absorption spectrum from ultraviolet to the terahertz region [43]. However, its relatively low absorption cross-section can restrict the broader use of pure graphene in practical optoelectronic applications [44]. To improve the performance of perovskite-based PDs, Lee et al. pioneered the concept of a perovskite/graphene heterostructure PD, paving the way for the development of high-performance devices [45]. The device boasts impressive photosensitive characteristics across the entire visible light spectrum and the n-octadecyltrimethoxysilane layer serves as a surface modifier, minimizing surface charge traps on the SiO₂. There is a notable decrease in the PL intensity of the CH₃NH₃PbI₃ perovskite/graphene heterojunction, ensuring efficient electron and hole separation. Due to graphene’s ultra-high electron mobility, holes are transported multiple times within the channel before recombining with trapped electrons. Consequently, an impressive photoresponsivity of 180 A W⁻¹ and effective quantum efficiency 5 × 10⁴% are attained under a substantial illumination power of 1 µW.

While integrating a high-efficiency carrier transport layer enhances performance, it also inadvertently creates a leakage pathway, thereby increasing the device dark current. To address the rise in dark current, improve the on–off current ratio, and maintain efficient carrier transport, Tan et al. proposed a graphene/perovskite/graphene structure perovskite PD, which has its source and drain graphene electrodes covered, showcases a minimal dark current (≈10⁻¹⁰ A) and high on/off current ratio (≈10³) [46]. Due to the smaller active area, its photocurrent is comparably lower (~100 nA). To further optimize the dark current and enhance the PD performance, Chen et al. introduced a specific channel structure on graphene. This design ensures an effective carrier separation within perovskites by curtailing the leakage current [47]. Due to the different work functions of perovskite and graphene, the difference in the Fermi level makes the perovskite band at the interface bend upward. Since a Schottky junction is formed at the interface of perovskite/graphene, and a large built-in electric field is introduced, which causes electrons from graphene to migrate into the valence band of perovskite, occupying its vacant states and effectively suppress the recombination of electron–hole pairs within the perovskite. Schottky junctions are not only formed in the vertical alignment of the graphene/perovskite interface but also emerge on either side of the channel. Consequently, PDs structured with the graphene/perovskite/graphene configuration display an impressive on/off switch ratio of 2.64 × 10³ and responsivity of 22 mA W⁻¹ when illuminated with 452 nm light. To elucidate the effect of horizontally structured Schottky junctions on photocurrent, one, five, and nine-channel graphene devices are discussed. Despite the formation of additional Schottky junctions, the photocurrent diminishes with an increase in channel numbers due to constrained carrier transfer in multi-channel devices. Additionally, a rise in the recombination and quenching of carriers further reduces the photocurrent. Notably, introducing a single graphene channel yields the most optimal detector performance.

The heterojunction can be structured both horizontally and vertically to enhance device performance. For instance, Chang et al. fabricated graphene-MAPbI₃ perovskite hybrid phototransistors by vapor deposition [48]. This device exhibited an astounding responsivity of 1.73 × 10⁷ A W⁻¹ and a superb detectivity of 2 × 10¹⁵ Jones, with an effective quantum efficiency approximating 108% across the visible light spectrum. Zou et al. proposed a vertical-structure PD with perovskite single-crystal MAPbBr₃ as the light absorption layer and graphene as the transport layer [49]. Due to the mismatch in the graphene and perovskite work functions and the formed interfacial electric field, holes are transferred to the graphene, while electrons are bound to the perovskite, acting as additional phototunable gates. The electric field created by these electrons induces hole generation in the graphene. This, in turn, depresses the Fermi level of the graphene, further amplifying the photocurrent. Under the conditions of a 3 V bias and 532 nm laser exposure, the device exhibits remarkable photoresponse performance at a low light intensity (0.66 mW cm⁻¹), displaying values ≈ 1017.1 A W⁻¹ for photoresponse, approximately 2.02 × 10¹³ Jones for photodetection and an ultra-high photoconductive gain of ≈2.37 × 10³. In comparison to the individual perovskite device, the performance is greatly improved.

By combining perovskites with high-mobility 2D materials like graphene and localized surface plasmon resonance (LSPR) metal nanostructures to create heterojunctions, one can bolster carrier transport and photoresponsivity, subsequently enhancing device performance [50]. Notably, perovskites produce a substantial volume of photo-induced electron–hole pairs, while gold nanostructures possess light-trapping capabilities and amplify electromagnetic fields. Sun et al. found that, upon incorporating plasmonic gold nanoparticles (NPs) (approximately 40 nm in diameter) just beneath the graphene in a perovskite/graphene heterojunction PD, there was a near two-fold increase in responsivity by a factor enhancement of 1.96 [51] (Figure 2a). As shown in Figure 2b, a significantly larger photoresponse can be observed in perovskite/graphene/Au NPs devices compared to perovskite/graphene PDs. Figure 2c describes the generation, diffusion, and transfer of photo-induced carriers in the perovskite layer, comparing scenarios with and without the presence of gold NPs. The majority of photo-induced carriers in the perovskite/graphene/Au NPs PD are produced near the interface of perovskite and graphene, attributed to the near-field enhancement caused by Au NPs. Since these carriers originate close to the perovskite/graphene interface, they have a reduced possibility of undergoing recombination during their transport to graphene compared to carriers that are generated at a distance from the interface. This phenomenon is associated with the gradual reduction in responsivity observed with increased light intensity in perovskite/graphene/Au NPs devices, as shown in Figure 2d.Wang et al. achieved a 59% improvement in the photoresponsivity of vertical perovskite PDs by incorporating gold nanorods into the perovskite layer [52]. Similarly, Liu and colleagues presented devices that combined perovskite, graphene, and Ag nanoparticles, where R was increased by a factor of 7.45 due to the incorporation of Ag nanostructures [53]. Feng et al. fabricated novel broadband quasi 2D perovskite (BA)₂(FA)_n−1PB_nI_3n+1 hybrid-structure PD with good stability by combining both monolayer graphene and Au square nanoarrays, as shown in Figure 2e,f [54]. Due to the near-field enhancement of the gold array, most of the photogenerated carriers are generated near the graphene–perovskite interface. Reduced carrier recombination, combined with the efficient extraction of electrons from the perovskite layer via graphene, enhances the photoelectric response of the Graphene-Au-perovskite PDs Therefore, a hybrid system employing both graphene and gold square nanoarrays can markedly boost carrier mobility and light absorption. This synergistic effect, optimizing optical trapping and promoting light-induced carrier extraction, results in a significant surge in photocurrent throughout the visible and near-infrared spectrum. The graphene-Au array-perovskite-based PDs had a low dark current of 10⁻¹⁰ A, large on/off ratio of 10⁴, high responsivity of 18.71 A W⁻¹, and detectivity of 2.21 × 10¹³ Jones. The responsivity and detectivity were two orders of magnitude higher than those of PDs based only on perovskites, as shown in Figure 2g. Regarding PDs based on perovskite/graphene heterostructures, several limitations remain, including a high dark current, reduced on/off ratio, and detectivity stemming from graphene’s inherent zero bandgap characteristic. In this sense, the electron blocking layer can be added to improve the on/off ratio using poly(3-hexylthiophene) (P3HT) and [6,6]-phenyl-C₆₁-butyric acid methyl ester (PCBM) can be added at the interface, acting as an electron trapping site to enhance the charge separation efficiency and thus the effective gain and responsivity enhancement [55,56]. However, with the incorporation of PCBM, the fall time is observed to be much longer than the rise time, likely because of the slower electron detrapping process from PCBM aggregates. So, effective gain and responsivity enhancement occurs at the price of much lower fall and rise time, which is still not desirable for practical application. Also, the deposition of perovskite on graphene causes the Dirac point of the graphene to shift to a more positive level, leading to a decrease in the mobility of the graphene.

3.2. Perovskite/TMDC Heterostructures

Although perovskite has a significant light absorption coefficient and generates substantial photogenerated carriers after illumination, how to effectively collect them through electrodes is a problem. This issue can be solved using graphene with high electron mobility as a transport layer. Apart from graphene, TMDCs are also a good solution due to their excellent optical properties, electrical properties, and abundant band gaps [57]. Although 2D TMDCs have excellent properties, it is still difficult for them to fully absorb incident light due to the atomic layer thickness and weak light absorption coefficient, limiting the performance of PDs. Hence, developing heterojunction devices with perovskite and TMDCs emerges as a promising strategy to enhance the performance of PDs. Using two materials with suitable energy band structures to form a heterojunction to create a type II energy band arrangement can significantly speed up the separation of carriers, increase the photocurrent, reduce the dark current, and improve the response speed for the device. In addition, the composite structure of different materials can also obtain a wider photoresponse range.

Kang et al. reported the first perovskite/MoS₂ heterojunction PD [58]. In this configuration, MAPbI₃ serves as the photoactive layer, while multilayer MoS₂ functions as the transport layer, as illustrated in Figure 3a. As a laser strikes the perovskite/MoS₂ junction, the predominant generation of electron–hole pairs occur in the perovskite absorber layer, given its superior absorption property, and type I heterojunction facilitates the transfer of both photo-induced electrons and holes from the perovskite to MoS₂. In the presence of an external electric field, these electrons and holes move in opposite directions within the MoS₂ channel, producing a photocurrent. Notably, the MAPbI₃/MoS₂/APTES device reached a responsivity of 2.11 × 10⁴ A W⁻¹ and a specific detectivity of 1.38 × 10¹⁰ Jones, respectively. (Figure 3b,c) The performance of the perovskite/MoS₂ PD was further improved by Sun et al., with the device showcasing rapid, stable, and sensitive response features even at minimal operating potentials. At an incident power of 2.2 pW (520 nm) and a bias voltage of 2 V, the responsivity and specific detection rate are 342 A W⁻¹ and 1.14 × 10¹² jones. Without the package, the device demonstrated high stability in transient on/off tests, with response and recovery times of 27 ms and 21 ms, respectively [59].

The noble metal disulfide compound PtSe₂ stands out as an emerging TMDC material. In recent times, it has garnered substantial research attention, primarily because, compared to traditional TMDCs, it has high electron mobility and a narrower band gap (1.25 eV), with good air stability [61]. It is a good material for photoelectric detection. Zhang et al. developed a self-driven, air-stable, ultrafast PD by combining perovskite with multilayer PtSe₂, and a schematic diagram of the device structure is shown in Figure 3d [61]. Zhang et al. fabricated large-scale PtSe₂ thin films by selenization of a pre-deposited Pt layer and employed cesium-doped FACsPbI₃ perovskite as the light-absorbing medium, which demonstrated a notably higher electron mobility in comparison to the more commonly used MAPbI₃ perovskite. Beyond this, it also proved to be more stable when exposed to air. As showcased in Figure 3e, the energy bands of the two materials form a type II band alignment, which facilitates the rapid separation of electrons and holes near the heterojunction interface, driven by an inherent electric field. The efficient separation of photogenerated electrons and holes is further demonstrated by the markedly diminished photoluminescence (PL) intensity of the PtSe₂/perovskite hybrid compared to that of the perovskite film alone (Figure 3f). This mechanism is responsible for the impressively swift rise and fall times recorded at 78 and 60 ns, respectively (Figure 3g). At zero bias, the light/dark ratio is 5.7 × 10³, the responsivity is 117.7 mA W⁻¹, and the specific detection rate is 2.91 × 10¹² Jones. A notable observation from this study was the PD resilience; even after being stored in standard ambient conditions for three weeks, it managed to maintain most of its photoresponsive characteristics. In another study, CsPbBr₃/MoS₂-based PD was reported to have excellent thermal stability and operational durability in air [62]. By manipulating the characteristic parameters, particularly the solution concentration and the surface ligand content of colloidal CsPbBr₃, one can control and modulate the optoelectronic properties of these detectors, which are closely related to the light absorption capability of the perovskite layer and the charge transfer efficiency at the heterojunction interface. Oxygen adsorbents play a significant role in this system. Their typical p-type behavior enhances the reaction rate of the PD. In the air, when the gate voltage is zero, the photoresponsivity is 6.40 × 10⁵ mA W⁻¹, the external quantum efficiency is 1.50 × 10⁵%, and the specific detection rate is 3.38 × 10¹¹ Jones.

Moreover, 2D Ruddlesden–Popper perovskites demonstrate significant promise for optical and optoelectronic applications [63,64]. Their superior environmental stability over 3D perovskites positions them as potential replacements for 3D variants in light-absorbing layers, aiming to enhance the efficiency of monolayer TMDC PDs. However, studies on the optoelectronic properties of TMDC monolayer/2D perovskite vertical heterostructures are still at an early stage. For example, Fang et al. reported PDs based on n-type MoS₂ and p-type lead-free 2D Ruddlesden–Popper perovskite (PEA)₂SnI₄ heterostructures [65]. This combined device can detect light across the full visible to near-infrared spectrum, offering adjustable photoresponse peaks. Incorporating a few layers of graphene flakes as electrical contacts further improves the device performance, recording a response rate of 1100 A W⁻¹ at 3 V bias and ~40 ms without any bias. Wang et al. reported graphene/perovskite/WS₂/graphene heterojunction devices [66] (Figure 4a). Under laser illumination, the photogenerated electrons move from the perovskite layer to the WS₂ monolayer, whereas the photogenerated holes travel in the reverse direction. This results in an accumulation of extra electrons in the WS₂ monolayer, which then combine with the photogenerated electron–hole pairs in WS₂, leading to the formation of negatively charged trions, depicted in Figure 4b. Compared with the photoresponse of the WS₂ monolayer region, there is a notable surge in photocurrent, as seen in Figure 4c. Figure 4d compares the photoresponsivity of the WS₂/PEPI heterostructure with that of the WS₂ monolayer, both measured at the same bias voltage V_ds of 4 V. It is observed that the photoresponsivity of heterointerface device surpasses that of the WS₂ monolayer device. For instance, at a laser power of 50 μW, the heterostructure photoresponsivity achieves 0.13 mA W⁻¹, which is five times greater than the 0.027 mA W⁻¹ responsivity of the WS₂ monolayer device. This enhance-ment is attributed to the superior light absorption capacity of the PEPI layer, coupled with the augmented charge carrier extraction efficiency facilitated by the innate potential at the WS₂ and PEPI interface. Additionally, the photoresponsivity and EQE of the self-driven PD reach 24.2 μA W⁻¹ and 5.7 × 10⁵, respectively.

Over the past few years, there has been notable advancement in perovskite/TMDCs heterostructures PDs. However, most of the photoelectric responses of 2D perovskite PDs are restricted to the ultraviolet-visible spectrum [68]. Given the substantial intrinsic optical band gap of 2D perovskites, their sensitivity to near-infrared (NIR) wavelengths is practically nonexistent. However, the type II staggered band structure observed in the TMDC-based vdW heterojunction facilitates the inter-layer charge transition between two distinct materials [69]. This means that under light exposure, electrons can shift from the valence band (VB) of one material to the conduction band (CB) of another, creating an energy divergence broader than the inherent bandgaps of both materials individually. This phenomenon broadens the detection capability of heterojunction PDs to encompass longer wavelengths. Zhou et al. realized a high-performance near-infrared PD using 2D perovskite/MoS₂ [67]. (Figure 4e) Within the heterostructure region, the interlayer coupling-dependent energy gap is effectively narrowed to less than 0.8 eV, as shown in Figure 4f. Experimental findings reveal that the MoS₂/2D perovskite vdW heterostructure responds distinctly to 1550 nm laser exposure, achieving a peak photoresponsivity of 121 A W⁻¹ and a detectivity of 4.3 × 10¹⁴ Jones at a wavelength of 860 nm (Figure 4g,h).

In summary, numerous perovskite/TMDC heterostructure PDs have been developed, taking advantage of the exceptional absorption of perovskite and the advanced electronic features of 2D TMDCs. It may be observed that there is a trade-off among the performance metrics for various PDs. Specifically, PDs based on perovskite/TMDC heterostructures showcase exceptionally high responsivity, surpassing that of typical commercial PDs. However, they lag in response speed. In a PD, the gain is intrinsically linked to carrier mobility and carrier lifetime. The remarkable responsivity seen in perovskite/TMDC PDs is often attributed to the significantly high gain due to extended carrier lifetimes, which majorly influences the speed of PDs. However, this prolonged carrier lifetime poses a significant challenge to their market adoption. Therefore, efforts to enhance responsivity should focus on increasing gain through boosting carrier mobility instead of lengthening carrier lifetime.

3.3. Perovskite/BP Heterostructures

In addition to using graphene and TMDC to build heterostructures, black phosphorus is also a good choice due to its unique anisotropic structure, excellent light absorption ability, high carrier mobility, and unique layer-controlled direct bandgap [70]. Combining black phosphorus with perovskites enables the performance improvement of PDs.

Zou et al. introduced an innovative Schottky barrier-controlled phototransistor where the channel current is primarily governed by the Schottky barrier at the source electrode [71]. The device is based on a few-layer BP crystal channel decorated by adding a MAPbI_3-xCl_x perovskite layer. Due to the band alignment of the BP/perovskite heterojunction, holes migrate to the BP channel while electrons amass within the perovskite layer. These gathered electrons induce an added electric field adjacent to the Schottky barrier, which subsequently lowers the barrier height. The pristine BP device without the perovskite layer demonstrates a bipolar nature with an on–off ratio exceeding 10⁴. After coating MAPbI_3-xCl_x, the transfer curves shift towards a higher gate voltage, highlighting the p-type doping effect on the BP channel brought about by the perovskite layer. The peak responsivity reaches 4 × 10⁶ A W⁻¹ at a gate voltage of approximately −30 V. Distinct from traditional phototransistors, this innovative device utilizes a field-assisted mechanism to release photocarriers within its channel. Consequently, the device exhibits high responsivity up to 10⁸ A W⁻¹ and a high specific detection rate up to 9 × 10¹³ Jones.

Up until now, the majority of 2D PDs reported in the literature operate under forward bias. This mode of operation often resembles a photoconductor with a gating effect, rather than adhering to the intended photovoltaic mechanism. This observed gating effect in photosensitive devices is frequently termed trap-assisted photoconductivity, where photogenerated carriers circulate multiple times within the device channel filled with trapped charges before recombination occurs. The short time necessitates a device channel with high conductivity, which inadvertently leads to a significant dark current. Faced with these challenges, there is an urgent need for a more efficient device structure that is not limited by the above-mentioned operational mechanisms so that the PDs are either fast and insensitive or sensitive and slow. Wang et al. proposed and developed a hybrid device based on perovskite and BP/MoS₂ PDs, and the device structure is given in Figure 5a [72]. It is noted in Figure 5b that type I and type II heterojunctions are formed at the perovskite/BP and BP/MoS₂ interfaces, respectively. When the perovskite film is exposed to light, photocarriers are spontaneously generated, being attributed to the intrinsic electric field present at the BP/MoS₂ junction. These photocarriers subsequently migrate into the BP layer where they undergo separation and collection. As shown in Figure 5c, it is observed that the pure perovskite film will display a pro-nounced PL peak at 775 nm, related to the bandgap transition of MAPbI₃. However, nota-ble PL quenching effects are observed in the perovskite/BP, perovskite/MoS₂, and perov-skite/BP/MoS₂ heterostructures, which is attributed to the efficient exciton dissociation and interlayer charge transfer between layers. Figure 5d presents the I–V curve for a typical p–n PD with BP/MoS₂ heterostructure, displaying clear rectifying behavior with a rectifica-tion ratio of 20. When capped with the perovskite layer, the device dark current reduces by roughly two orders of magnitude compared to the original BP/MoS₂ device. Distinguishing it from traditional detectors that rely on photovoltaic and gating mechanisms, this hybrid device stands out due to its elevated responsivity and swift response time. When the reverse bias voltage is −2 V, the responsivity of the device can be significantly improved to 11 A W⁻¹, and the response time can be optimized to hundreds of microseconds. The corresponding detective rate is as high as 1.3 × 10¹² Jones. Moreover, the combination of perovskites’ efficient light absorption and the substantial electric field created by the BP/MoS₂ junction makes this hybrid device exceptionally compatible for self-driven, broadband photodetection. Under 457 nm laser irradiation and zero bias, the device has a high detectivity of 3 × 10¹¹ Jones and an ultra-high light/dark value of 3 × 10⁷.

3.4. Perovskite/MXene Heterostructures

In recent years, two-dimensional Ti₃C₂T_x (MXene) has attracted much attention due to its distinctive properties such as metallic conductivity, adjustable work function, commendable light transmittance, and the ability to form stackable conductive films, making it well-suited for flexible electronic devices [74,75]. Deng showcased a novel approach by introducing an all-spray processable, wide-area, flexible PD array that combined 2D MXene and 2D CsPbBr₃. This was implemented on ordinary paper, serving as electrodes and the primary active materials and the detailed structure of this device can be visualized in Figure 5e [73]. Notably, the authors realized the optical communication function by fabricating a large-area array of 1665 pixels within 72 cm². Once the illumination energy is larger than the semiconductor bandgap, a large number of photogenerated electron–hole pairs in the CsPbBr₃ nanosheets are rapidly separated by the reverse local electric field and have a low recombination rate, resulting in a free carrier concentration increase, as shown in Figure 5f. Conversely, during the charge transfer process at the interface between CsPbBr₃ and MXene, charge accumulation and band bending create a depletion layer. This layer plays a crucial role in enhancing the separation efficiency of photogenerated excitons. The light response speed is faster than 18 ms, and the on/off ratio reaches 2.3 × 10³. The optoelectronic characteristics of the MXene-based PD at various bending degrees can be observed in Figure 5g. As the bending angle progressively increases, there is a slight decline in the photocurrent, which eventually stabilizes. This behavior underscores the fact that both MXene nanosheets and CsPbBr₃ nanosheets possess excellent layered structures, coupled with an impressive ability to relieve stress. This research not only showcases a cost-effective technique for PD fabrication but also hints at promising applications in the realms of wearable electronics and optical communication.

The electromagnetic field near the surface is naturally intensified due to the oscillation of electrons within metal nanoparticles, which can enhance PD performance even when incident light is scarce. However, the enhanced electric field can only be distributed near the surface of metal nanoparticles, limiting the light absorption capacity of the entire photosensitive layer [76]. Two-dimensional materials, with their extensive surface area and minimal thickness, allow for more interaction with the photosensitive layer without adding stress to the film. Among MXenes, Ti₃C₂T_x nanosheets are noteworthy. They exhibit outstanding features such as high conductivity (10⁴ S cm⁻¹), broad spectral response, and a work function that aligns well with CsPbBr₃, and abundant free electrons offer great possibilities to improve the optoelectronic response of perovskite films instead of metal nanostructures. Li et al. proposed a simple method to prepare high-performance all-inorganic halide perovskite CsPbBr₃ quantum dot PDs by uniformly mixing two-dimensional Ti₃C₂T_x nanosheets with different concentrations [77]. Electrons at the conduction band tend to be spontaneously injected into the 2D Ti₃C₂T_x nanosheets after excitation, and ions can be partially enhanced due to the enhanced carrier transport. Simulations using the finite-difference time-domain (FDTD) method demonstrate that the enhanced near-surface electromagnetic field of Ti₃C₂T_x nanosheet notably increases the light efficiency of the entire CsPbBr₃ QD film. In comparison with unmodified CsPbBr₃ QD devices, the presence of the nanosheets hot electrons considerably enhances the photocurrent, leading to an impressive 300% surge in EQE. The photogenerated carriers naturally migrate from the CsPbBr₃ QDs to the Ti₃C₂T_x nanosheets due to the creation of downward band bending, which is also beneficial for enhancing the photocurrent. Under 490 nm illumination, the CsPbBr₃ QD/MXene nanosheet PDs have a response rate of 97 μA W⁻¹ at a power of 2.9 mW cm⁻². This peak performance was sustained for four months in normal atmospheric conditions, suggesting a promising approach to achieve advancements in the mass production of perovskite optoelectronic devices.

This section mainly introduces the research progress of perovskite and 2D material heterostructures in PDs. PD devices based on perovskite/2D heterostructures generally perform better than single devices due to the ultrafast charge transfer and unique energy band mechanism at the heterointerface. Heterostructures combining perovskites and TMDCs mainly absorb light in the visible to near-IR range, a characteristic stemming from the bandgaps of the base materials and the way their energy bands align. When targeting the crucial mid-to-far IR spectrum, which has specific resonance with many molecules, researchers have turned to materials like graphene and other 2D semiconductors with narrow gaps, such as BP. Nonetheless, these materials present challenges. For instance, graphene-integrated PDs often have significant dark current, while BP-based ones suffer from reduced resilience in ambient conditions. The distinct characteristics of interlayer excitons in 2D TMDCs pave the way for creating intriguing and novel devices [78]. Perovskites have great advantages as light-absorbing layers in heterostructures. Two-dimensional materials have rich and tunable band gaps, extremely high electron mobility, ultra-thin and transparent properties, and can obtain PDs with excellent performance by constructing heterojunctions with perovskites. The PD parameters in this section are summarized in

Table 2. Summary of the key parameters for PDs based on perovskite and 2D material heterostructures.

Device Structure	R [AW−1]	D* [Jones]	On/Off Ratio	EQE [%]	Response Time [trise,tfall.ms]	Test Conditions	Reference
CH3NH3PbI3/Graphene	180	109	-	5 × 104	87, 540	λ:520 nm, P:1 uw, VDS = 0.1 V, VGS = 0 V	[45]
Graphene/(C4H9NH3)2PbBr4/Graphene	2.1 × 103	-	103	-	-	λ:470 nm, P:10 uw, Vbias = 0.5 V	[46]
Graphene/CH3NH3PbI3/Graphene	22 × 10−3	-	2.6 × 103	-	-	λ:452 nm, P:2.1 mWcm−2, Vbias = 2 V	[47]
MAPbBr3/Graphene	1.017 × 103	2.02 × 1013	-	-	50.9, 26	λ:532 nm, P:200 mWcm−2, Vbias = 3 V	[49]
CH3NH3PbI3/Graphene	1.73 × 107	2 × 1015	-	108	-, 879	White light, VDS = 1.5 V, VGS = 40 v	[48]
graphene/Au nanoarray/(BA)2(FA)n−1PbnI3n+1	18.71	2.21 × 1013	6 × 104	-	0.33 × 103, 0.27 × 103	λ:405 nm, P:10.4 mWcm−2, Vbias = 2 V	[54]
CH3NH3PbI3/Graphene/Ag nanoparticle	495.3	-	-	-	7 × 103, -	λ:585 nm, P:1.27 uW, Vbias = 0.1 V	[53]
CH3NH3PbI3/MoS2	21.1 × 103	1.38 × 1010	-	-	6.17 × 103, 4.5 × 103	λ:520 nm	[58]
CH3NH3PbI3/MoS2	342	1.14 × 1012	-	-	27, 21	λ:520 nm, P:2.2 pW, Vbias = 2 V	[59]
CH3NH3PbI3/WSe2	110	2.2 × 1011	-	2.5 × 104	143, 2113	λ:532 nm, P:2.8 mW, Vbias = 2 V	[27]
FACsPbI3/PtSe2	0.1177	2.91 × 1012	5.7 × 103	-	78 × 10−6, 60 × 10−6	λ:808 nm, Vbias = 0 V	[79]
CsPbBr3/MoS2	640	3.38 × 1011	≈104	1.5 × 105	92, -	λ:532 nm, P:4.86 mWcm−2, VDS = 1 V	[62]
Graphene/(PEA)2SnI4/MoS2/Graphene	0.121	8.09 × 109	500	38.2	34, 38	λ:532 nm, P:5.46 nW, VDS = 0 V, VGS = 0 V	[65]
Graphene/(C6H5C2H4NH3)2PbI4/WS2/Graphene	24.2 × 10−6	-	1.5 × 103	5.7 × 10−5	200, 200	λ:532 nm, Vbias = 0 V	[66]
(BA)2MA3Pb4I13/MoS2	121	4.3 × 1014	-	-	8.2 × 10−3, 28.6 × 10−3	λ:860 nm, P:5 uw, Vbias = 2 V,	[67]
MAPbI3-xClx/BP	106–108	9 × 1013	-	-	8, 17	-	[71]
MAPbI3/BP/MoS2	11	1.3 × 1012	6 × 104	80	0.15, 0.24	λ:450 nm, Vbias = −2 V	[72]
CsPbBr3/Ti3C2Tx	44.9 × 10−3	6.4 × 108	2.3 × 103	-	48, 18	λ:450 nm, P:7.07 mWmm−2, Vbias = 10 V	[73]
CsPbBr3/Ti3C2Tx nanosheet	97 × 10−6	-	-	-	46.2, 24.6	λ:490 nm, P:2.9 mWcm−2	[77]

.

4. Perovskite and 2D Material Heterostructure Solar Cells

Solar cells are essential devices that convert light energy into electrical energy. As the global population grows and the demand for energy increases, ensuring a stable supply of energy becomes even more critical. At present, there have been many problems in the way of obtaining energy by burning fossil fuels, such as insufficient fossil resources and environmental pollution [80]. The use of efficient and clean solar energy to replace fossil energy is a very effective method.

The solar energy conversion efficiency of perovskite solar cells (PSCs) has reached 25.2% in the past decade, which has attracted great interest, and the device structure of PSCs can be divided into the n-i-p structure and p-i-n structure [81]. PSC devices typically consist of a transparent bottom electrode, an electron transport layer (ETL), a perovskite photoactive layer, a hole transport layer (HTL), and a top electrode. The fundamental operation of a solar cell can be described in these steps: First, when exposed to light, electron–hole pairs form in the perovskite layer. Then, driven by the inherent electric field, the generated electrons are captured by the ETL, while the holes move into the HTL. These photo-induced charge carriers are then gathered by the electrodes and converted into electrical energy for output. The key parameters to evaluate the performance of PSCs are open-circuit voltage (V_oc), short-circuit current density (J_sc), fill factor (FF), and power conversion efficiency (PCE). The charge transport material dramatically influences the efficiency and stability of PSCs [82]. HTL and ETL materials promote efficient carrier extraction and recently suppress recombination between perovskite films and electrodes. Two-dimensional materials are often used as charge transport layers for PSCs to collect electrons and holes to improve performance efficiently. Note that graphene and other 2D materials can also be used as additives in the photoactive layer inside PSCs; this part of the review is not included, but it can be found in previous review paper [22]. Additionally, perovskite itself as a 2D additive within 3D perovskite and its potential for photovoltaics is also not included here in this part.

4.1. Perovskite/Graphene

For PSCs, the ideal interfacial layers are those that can be processed from solutions, are affordable, highly conductive, and maintain stability either chemically or thermally. Also, these layers should align their energy bands appropriately with the photoactive layer to ensure selective charge extraction [83,84]. Graphene, with its high electron mobility, stands out as an optimal material for extracting and transferring charges to electrodes. For instance, to extract electrons from the perovskite, a layer of graphene quantum dots (GQDs) was strategically positioned between the perovskite photoactive layer and the TiO₂, serving as the ETL to extract electrons from the perovskite [85] (Figure 6a). The improvement in efficiency is mainly attributed to the increase in photocurrent since the incident photon-to-current conversion efficiency (IPCE) indicates a significant amplification in the conversion of absorbed photons to current within the visible to near-infrared spectrum due to the incorporation of GQDs. Utilizing ultrafast transient absorption spectroscopy, the researchers observed that the electron extraction time for the PGT film at 90~106 ps, substantially quicker than the PT film (260~307 ps), which can effectively compete with carrier trapping, as shown in Figure 6b. With the addition of GQDs, the PCE of the solar cell was significantly improved from 8.81% to 10.15% (Figure 6c). Agresti et al. further optimized the electron extraction performance of the ETL layer [86]. Li-neutralized graphene oxide was used as the ETL in FTO/cTiO₂/G+mTiO₂/GO-Li/CH₃NH₃PbI₃/spiro-OMeTAD/Au PSCs, and the PCE is increased from 11.6% to 12.6%.

In addition to being used as ETL, graphene has also been used as the HTL to improve the performance of solar cells. Wang et al. formed strong Pb-Cl and Pb-O bonds between [CH(NH₂)₂]_x[CH₃NH₃]_1-xPb_1+yI₃ films with Pb-rich surfaces and chlorinated graphene oxide layers [87]. Serving as the HTL, Cl-GO exhibits an efficient extraction of holes. Under a forward scan, the perovskite/Cl-GO cell achieved notable metrics: a PCE of 21.08%, a J_sc of 23.82 mA cm⁻², a V_oc of 1.12 V, and a fill factor (FF) of 0.79. Furthermore, forming strong chemical bonds on the surface of soft perovskite films stabilizes the perovskite heterostructure, which considerably curtails the deterioration of perovskite components, subsequently minimizing harm to the organic HTLs. As a result, the stability of PSC was dramatically improved, maintaining 21% of the initial value after 1000 h of operation at 60 °C under AM1.5G solar lamp and 1000 h of operation at a maximum power point of 90% of the efficiency.

4.2. Perovskite/TMDCs

Apart from graphene, TMDCs have exhibited notable benefits as charge transport materials in PSCs. For example, Zhao et al. reported the low-temperature processing of 2D SnS₂ nanosheets as a novel ETL in PSCs and achieved a PCE of 13.63% for the device, a J_sc of 23.70 mA cm⁻², a V_oc of 0.95 V, and a FF of 0.61 [88]. Similarly, 2D SnS₂ was used as an ETL for flexible PSCs, and the structure is shown in Figure 7a [89]. As the ETL of PSCs, SnS₂ can align its conduction band minimum with that of the perovskite and photoanode to ensure smooth carrier transfer. In addition, the authors optimized the SnS₂ film thickness. As shown in Figure 7b, with a SnS₂ ETL of 60 nm, the PCE of the device reached 13.2% due to the best coverage, electron extraction and larger charge recombination resistance, with a V_oc of 1.011 V, a J_sc of 1.011 V, a J_sc of 21.70 mA cm⁻², and a FF of 0.60. In addition, since MAPbI₃ can destroy ZnO, the introduction of the SnS₂ buffer layer also improves the stability of PSCs: The device using ZnO ETL degrades very seriously at 40~50% humidity, resulting in an initial after 30 days. PCE lost about 30%. In contrast, SnS₂-based devices exhibited good long-term stability with almost unchanged PCE with only 7% loss in PCE after 30 days of storage (Figure 7c). Zhao et al. used 2D SnS₂ ETL prepared via the self-assembly stack deposition method for efficient planar PSCs [90]. The Pb-S intermolecular interaction between perovskite and SnS₂ helps in passivating the interface trap states. This interaction reduces charge recombination and facilitates electron extraction at the interface of both the ETL/perovskite and the HTL/perovskite, ensuring a balanced charge transport throughout the device. As a result, a PCE of more than 20.12% was achieved, with an outstanding V_oc of 1.161 V, a J_sc of 23.55 mA cm⁻², and a high FF of 0.73 s.

Meanwhile, the HTL also plays an essential role in PSCs. The efficiency with which holes are separated directly influences the device performance [92]. Due to their high charge mobility, 2D TMDCs are also suitable candidates for the HTL. Notably, while impressive PCEs exceeding 20% have been realized using Spiro-OMeTAD as HTLs in PSCs, the migration of iodine from the perovskite active layer through Spiro-OMeTAD to the metal electrodes poses a challenge to the efficiency and durability of the devices [93]. To eliminate this effect, Capasso et al. added a few-layer MoS₂ flakes as a buffer in PSCs [94]. The MoS₂ layer serves a dual purpose. It acts as a protective barrier, preventing any interaction between the perovskite and the electrode. This protective measure avoids efficiency losses that can occur due to the direct contact of the perovskite with the electrode. Simultaneously, the MoS₂ sheet functions as an HTL, facilitating the effective transfer of holes from the perovskite to the electrode. The PCE of the as-obtained PSCs was 13.3%. Additionally, the stability of the device is significantly improved: the PCE of PSCs without MoS₂ decreased by 14.2% in the 550 h durability test, while that of PSCs with MoS₂ added decreased to 12.4% after 550 h. Wang et al. mixed MoS₂ nanosheets in a PEDOT:PSS layer to form a hybrid HTL layer embedded in inverted PSCs [91] (Figure 7d). MoS₂ nanosheets enhance the charge extraction efficiency of the HTL, mitigating recombination at the interface and substantially decreasing electrode polarization and hysteresis effects. Impedance spectroscopy (IS) analysis reveals a 50% rise in the composite resistance when MoS₂ is combined with PEDOT:PSS in the HTL. As shown in Figure 7e, the average Jsc for six samples of the device, enhanced with a 10% MoS₂ blend, at 20.7 mA cm⁻², outperforming the pristine PEDOT:PSS HTL devices. How-ever, the Voc of the modified HTL-based device is marginally lower than that of the un-modified PEDOT:PSS HTL device. Devices with the MoS₂ hybrid PEDOT:PSS HTL witnessed a significant boost in their PCE, achieving approximately 14.7% (Figure 7f). Moreover, these modified HTL-based devices demonstrated impressive durability, maintaining over 95% of their initial PCE even after a duration of 4 weeks.

4.3. Perovskite/BP

Liquid-exfoliated, few-layered 2D BP nanosheets have been explored as HTLs for PSCs by Muduli et al. [95]. Photoelectron spectroscopy (PESA) measurements carried out in ambient conditions verified that the valence band level of BP nanosheets, at −5.2 eV, is optimally suited for the hole injection of CH₃NH₃PbI₃. The electrical properties of the best-performing PSC incorporating BP nanosheets as HTM. When the BP nanosheets are fully utilized as HTM, a PCE of 7.88% is obtained, an improvement over the PCE value of 4% for the HTM-free device. For comparison, the authors mixed BP nanosheets with the commonly used Spiro-OMeTAD as HTL, achieving a PCE of 16.4%. In contrast, using Spiro-OMeTAD exclusively as HTL yielded a PCE of just 13.1%, indicating that BP can be vital in facilitating hole extraction.

Photostability remains a significant concern for the widespread adoption of PSCs. A key degradation pathway, especially under continuous illumination, is the reduction of the Pb²⁺ ions in the commonly used MAPbI₃ perovskite material to elemental metallic Pb⁰ during continuous illumination. Pb⁰, as a deep defect state within the perovskite structure, not only degrades the performance of PSC devices but also reduces the long-term stability of the devices [96]. The introduction of 2D BP into MAPbI₃ can not only delay hot carrier recombination but also suppress the Pb⁰ defect formation, improving the stability of PSCs. Wang et al. added BP to MAPbI₃ perovskite to fabricate PSCs (Figure 8a) [97]. The PCE of the device is 20.62%, with a J_sc of 23.31 mA cm⁻², V_oc of 1.082 V and FF of 0.82. Both photostability and photovoltaic performance were significantly enhanced due to the application of BP (Figure 8b,c). The addition of BP has also greatly improved the stability of PSC and provided people with excellent ideas. In particular, in a dry N₂ glove box, the initial efficiency of MAPbI₃/BP-based PSCs remained at 94% after 1000 h of continuous white LED irradiation, while the initial efficiency of PSCs without BP addition dropped to 30%.

4.4. Perovskite/MXene

MXenes possess outstanding optoelectronic characteristics, superior electrical conductivity, and enhanced carrier mobility, making them prime candidates for fabricating PSC interfacial layers. Utilizing 2D Ti₃C₂T_x through spin-coating, Yang et al. developed flat and uniform ETL surfaces for ITO/ETL/CH₃NH₃PbI₃/Spiro-OMeTAD/Ag PSCs [100]. A brief UV ozone treatment on metallic Ti₃C₂T_x enhances its surface bonding with Ti-O without affecting its inherent properties, thereby improving its suitability as an ETL. This process improved electron transfer and suppressed the recombination at the ETL/perovskite interface, elevating the PCE of untreated Ti₃C₂T_x films from 5.00% to a notable 17.17% after being subjected to UV-ozone for 30 min. Following their initial findings, Yang et al. continued to optimize the performance of PSCs by introducing in situ oxidized Ti₃C₂T_x (O-Ti₃C₂T_x) MXene with SnO₂, creating a nanoscale heterojunction and using SnO₂ as the ETL of MAPbI₃ PSCs [98] (Figure 8d). The oxidation process transitioned the Ti₃C₂T_x from being predominantly metallic to semiconducting, and this change allowed for the tuning of Ti₃C₂T_x properties to better align with the energy levels of perovskites. Adding O-Ti₃C₂T_x also enhanced the electron mobility of SnO₂ ETL. As a result, the O-Ti₃C₂T_x/SnO₂ heterojunction excelled in electron extraction and minimized the recombination between the ETL and perovskite, resulting in a significant improvement in the performance of PSC, with PCE increasing from 17.68% to 20.09% and enhanced stability in air (Figure 8e). After being stored in ambient air for over 1000 h, the O-Ti₃C₂T_x/SnO₂ heterojunction device maintained 82% of its original PCE, showing great long-term stability, as seen in Figure 8f. Furthermore, MXenes can serve as HTLs within PSCs. Chen et al. introduced 2D Ti₃C₂-MXene as an interlayer into all-inorganic PSCs as HTL [99] (Figure 8g). Between the CsPbBr₃ layer and the carbon electrode, the presence of 2D Ti₃C₂-MXene improves the energy level alignment, accelerates hole extraction, and achieves efficient carrier transport and low interfacial recombination. Compared with the one without MXene, PCE increased from 8.21% to 9.01% (Figure 8h,i). Moreover, the functional groups within Ti₃C₂-MXene effectively passivated perovskite grains, diminished trap defects in CsPbBr₃ films, and elevated the overall quality of the perovskite films. Impressively, the film showcased excellent hygrothermal stability, lasting for 1900 h and extending beyond 600 h.

This section presents the recent research progress of perovskite and 2D material heterostructures in PSCs. ETL and HTL are the critical parts of PSCs, and whether the material can efficiently transfer charges directly affects the overall performance of PSCs. So far, significant research has been directed towards transition metal oxide ETLs, including titanium dioxide (TiO₂), zinc oxide (ZnO), and tin dioxide (SnO₂), but they come with multiple limitations that restrict the device performance [101]. For example, TiO₂, widely used as an ETL, results in subpar device stability and an elevated density of trap states at the TiO₂/perovskite boundary [102]. Conversely, Spiro-MeOTAD, which is the predominant HTL material used in n-i-p PSCs, faces stability issues and requires complex doping processes. Additionally, the material PEDOT:PSS is compromised due to its acidic nature and its vulnerability to moisture [103]. Two-dimensional materials are characterized by their exceptionally high electron mobility and diverse, tunable energy bands. When selected appropriately for interface layers, these materials facilitate ultrafast charge extraction. Furthermore, 2D materials serve as efficient barriers, protecting the perovskite photoactive layer from moisture. This makes them promising candidates for both ETL and HTL in high-performance PSCs. In general, PSCs that incorporate 2D materials as interfacial layers (like HTL and ETL) tend to exhibit superior performance and stability compared to traditional configurations. These 2D layers not only enable rapid charge extraction but also act as effective shields against moisture for the perovskite photoactive layer. Though perovskite and 2D material heterostructures have been incorporated into PSCs to enhance both PCE and stability, the intricacies of the interface dynamics and interactions between perovskites and 2D materials remain largely unexplored. Delving deeper into the interface structure is essential to gain clearer insights into the formation mechanism and the combined benefits of perovskite and 2D material heterostructures.

Table 3. Summary of the key parameters for solar cells based on perovskite and 2D material heterostructures.

Device Structure	Voc [V]	Jsc [mA cm−2]	FF	PCE [%]	Reference
FTO/TiO2/GQDS/CH3NH3PbI3/Spiro-OMeTAD/Au	0.937	17.06	0.635	10.15	[85]
FTO/TiO2/G+mTiO2/GO-Li/CH3NH3PbI3/spiro-OMeTAD/Au	8.57	-	0.646	12.6	[86]
ITO/SnO2/[CH(NH2)2]x[CH3NH3]1−xPb1+yI3/Cl-GO/PTAA/AU	1.12	23.82	0.79	21.08	[87]
FTO/SnS2/CH3NH3PbI3/Spiro-OMeTAD/Au	0.95	23.70	0.61	13.63	[88]
FTO/SnS2/CH3NH3PbI3/Spiro-OMeTAD/Au	1.011	21.70	0.60	12.20	[89]
ITO/SnO2/SnS2/FA0.75MA0.15Cs0.1PbI2.65Br0.35/Spiro-OMeTAD/Au	1.161	23.55	0.73	20.12	[90]
FTO/TiO2/TiO2/CH3NH3PbI3/MoS2/Spiro-OMeTAD/Au	0.943 ± 0.026	−18.11 ± 0.65	0.628 ± 0.034	10.7 ± 0.4	[94]
ITO/PEDOT:PSS, MoS2/MAPbI3/PC71BM/BCP/Ag	0.9522 ± 0.001	72.3 ± 0.9	0.723 ± 0.009	14.2 ± 0.39	[91]
FTO/TiO2/Meso-TiO2/MAPbI3/BP, Spiro-OMeTAD/AU	1.06	20.22	0.761	16.4	[95]
FTO/TiO2/SnO2/CH3NH3PbI3,BP/Spiro-OMeTAD/Ag	1.082	23.31	0.8225	20.65	[97]
ITO/Ti3C2Tx/CH3NH3PbI3/Spiro-OMeTAD/Ag	1.08	22.63	0.70	17.17	[100]
FTO/O-Ti3C2Tx/SnO2/CH3NH3PbI3/Spiro-OMeTAD/Ag	1.073	23.88	0.784	20.09	[98]
FTO/CsPbBr3/Ti3C2/Au	1.444	8.54	0.7308	9.01	[99]

compares the key parameters of the PSCs discussed in this section.

5. Perovskite and 2D Material Heterostructure Photocatalysis

The advancement of photocatalysis is viewed as one of the most promising approaches to address the growing energy and environmental problems, during which H₂ can be produced using photocatalytic techniques and CO₂ can be converted [104]. The photocatalytic process typically involves three stages: (1) under light exposure, the photocatalyst produces free charges, (2) these charges migrate to the photocatalyst’s surface, and (3) a redox reaction takes place with the adsorbed substances [105]. From a thermodynamic perspective, the potential of the conduction band ought to be more negative, while the potential of the valence band should be more positive than the reaction potential of particular species. For example, in order to facilitate photocatalytic H₂ evolution, the conduction band potential of an optimal photocatalyst should lie more negative than the H⁺ reduction potential. When assessing photocatalysis, metrics such as CO₂ reduction rate, H₂ production rate, and pollutant degradation rate are used, varying based on the specific photocatalysis reaction. Conventional photocatalysis materials such as TiO₂, ZnO, and ZnS usually suffer from narrow light absorption, short photoinduced charge lifetime, and poor stability [106]. As a new type of semiconductor material, perovskite has a high absorption coefficient and can fully use light energy; however, the carrier recombination in perovskite seriously affects the performance of the catalyst. Constructing a heterojunction with 2D materials can quickly separate the charges and improve the catalyst’s performance. Next, some developments in perovskites and 2D materials for photocatalysis will be introduced.

5.1. Perovskite and 2D Material Heterostructures for Hydrogen Production

Hydrogen is increasingly viewed as a potential successor to fossil fuels in the upcoming years given its exceptional purity, high energy density per unit mass, and non-polluting combustion [107]. Currently, hydrogen production predominantly relies on the steam reforming of natural gas. This method necessitates the use of non-renewable fossil fuels, operates at elevated temperatures, incurs high production expenses, and contributes to rising greenhouse gas emissions. Using solar energy for photocatalytic hydrogen production offers a compelling alternative to traditional steam reforming. However, the development of photocatalytic hydrogen evolution technology has been curtailed due to the slow reaction rates and instability of semiconductor photocatalysts [108]. Perovskite materials with high absorption coefficients, abundant band gaps, and long charge diffusion lengths are suitable candidates for photocatalytic hydrogen evolution [109]. Nevertheless, the rapid charge recombination observed in perovskites impedes efficient charge transport, consequently restricting H₂ production [110]. To enhance the photocatalytic activity of perovskites, noble metals such as platinum are usually used as cocatalysts, but the rarity and elevated cost hinder its widespread use in large-scale photocatalytic H₂ production [111]. In recent times, 2D materials have gained attention as cocatalysts for perovskites, given their cost-effectiveness, robust stability, and impressive carrier mobility, thereby enhancing photocatalytic performance.

In 2016, Park et al. first used 3D bulk MAPbI₃ powder in conjunction with a saturated HI aqueous solution to stably achieve photocatalytic decomposition of HI, releasing hydrogen in the process [111]. Notably, splitting HX presents certain advantages over direct water splitting as HX reduction engages only two electrons, in contrast to the four-electron process of water splitting. Moreover, perovskites boast a substantial absorption coefficient for visible light, thus fully harnessing light energy. Reduced graphene oxide (rGO) acts as both an electron acceptor and transporter. It enhances the performance of semiconductor photocatalysts by facilitating the transfer of photogenerated electrons within semiconductors and also serves as a framework, augmenting the surface area [112]. To further improve the efficiency, Wu et al. reported a novel MAPbI₃/rGO composite structure as a photocatalyst for hydrogen evolution [33]. The MAPbI₃/rGO composite structure can be obtained by dispersing GO powder in a high-acid aqueous solution saturated with MAPbI₃ and then exposing the solution to visible light (λ ≥ 420 nm) for several hours of photoreaction. Figure 9a illustrates that the MAPbI₃/rGO composite exhibits superior catalytic activity, with a hydrogen evolution rate of 93.9 μmol h⁻¹. Under visible light exposure (λ ≥ 420 nm) and intensity below 120 mW cm⁻², this rate is an impressive 67 times faster than that of unmodified MAPbI₃. Furthermore, the composite demonstrates high stability, with consistent hydrogen evolution results even after 200 h, as depicted in Figure 9b. To investigate the charge transfer properties of MAPbI₃ and MAPbI₃/rGO, the corresponding photoelectrodes were assembled, and electrochemical impedance spectroscopy (EIS) analysis was conducted in a dichloromethane solution. The Nyquist plot, depicted in Figure 9c, reveals that the charge transfer resistance in MAPbI₃/rGO films is markedly reduced compared to MAPbI₃. This observation is corroborated by the notably smaller semicircle evident in the plot for MAPbI₃/rGO, underscoring the enhanced charge transport ability offered by the incorporated rGO, thus facilitating swifter charge transfer. Consequently, during the photocatalytic activity, photogenerated electrons within MAPbI₃ are swiftly relayed to rGO via the Pb-O-C bond, which ensures their segregation from the photogenerated holes. At the rGO location, H⁺ gets reduced to H₂. Simultaneously, the photogenerated holes participate in the oxidation of I⁻ to ${\mathrm{I}}_3^{-}$.

Wang et al. pioneered the utilization of Cs₂AgBiB₆ (CABB) double perovskite for cracking hydrobromic acid (HBr) when exposed to visible light [115]. RGO, with good electron transfer properties and acid stability, was introduced to enhance the hydrogen production performance further. Using the CABB/RGO composite as a photocatalyst under visible light (λ ≥ 420 nm, via a 300 W Xe lamp), they were able to generate a remarkable 489 μmol g⁻¹ H₂ for 10 h. Notably, the exemplary photocatalyst of CABB combined with 2.5% RGO showcased robust stability, persisting through 120 h of ongoing photocatalytic hydrogen evolution. In a saturated solution, CABB perovskite undergoes a dynamic precipitation-dissolution equilibrium process, and RGO can attach CABB particles. As a proficient visible light absorber, CABB produces electrons and holes when irradiated. These generated electrons, in turn, transfer to the conductive RGO via the M-O-C bonds, which reduces H⁺ to generate H₂ at the active site of RGO. Moreover, BP is renowned for its outstanding optoelectronic characteristics, serving as an efficient charge-transfer material. Li et al. synthesized a BP/MAPbI₃ heterostructure in a MAPbI₃-saturated HI aqueous solution system by an electrostatic coupling method and used it as a photocatalyst for HI splitting for hydrogen production [116]. The type I heterojunction on the BP/MAPbI₃ interface captures the photogenerated electrons from MAPbI₃, significantly diminishing the recombination of these electrons. As a result, the engineered BP/MAPbI₃ photocatalyst showcases an impressive capability for H₂ production when exposed to visible light, achieving a photocatalytic rate of 3742 μmol h⁻¹g⁻¹. Remarkably, this rate of hydrogen evolution is a hundred times faster compared to MAPbI₃ at 35 μmol h⁻¹g⁻¹. Furthermore, the BP/MAPbI₃ photocatalyst maintains its stability in a MAPbI₃-saturated HI solution, with its performance remaining largely consistent throughout the photoreaction period.

MoS₂, another non-noble metal layered semiconductor, has garnered significant research interest because of its outstanding optoelectronic characteristics [117]. Wang et al. prepared a multilayer MoS₂/MAPbI₃ composite catalyst using a facile in situ coupling method, combining synthesized MAPbI₃ powder with MoS₂ nanosheet powder [118]. Compared with pristine MAPbI₃, the composite structure has a higher photocatalytic rate of 2061 µmol h^−1, a 121-fold increase. The composite maintained stability throughout the 156-h hydrogen evolution reaction. By employing in situ rapid crystallization on a monolayer of MoS₂, a new type II heterojunction photocatalyst was introduced, leading to the creation of MoS₂/MAPbI₃ heterojunction catalysts [119]. The robust built-in electric field in the MoS₂/MAPbI₃ enhances charge separation, which is crucial for efficient HI splitting and photocatalytic hydrogen generation. A type II heterojunction, along with this strong electric field, is established between the precisely structured MAPbI₃ and MoS₂ nanosheets. This has been directly validated using Kelvin probe force microscopy (KPFM). As a result, this takes the hydrogen production of perovskite photocatalysts to unprecedented levels, achieving a hydrogen production efficiency of 1.09% and an activity rate of 13,600 µmol g⁻¹h⁻¹ when exposed to visible light. Compared with the single MAPbI₃ catalyst, the catalytic efficiency of the heterojunction is 220 times higher. Guan et al. also adopted the MAPbI₃/MoS₂ composite structure, and the MoS₂ nanoflowers with abundant active sites were assembled with MAPbI₃ crystallites to form a heterostructure [113]. Remarkably, its rate of hydrogen evolution reaches up to 30,000 μmol g⁻¹h⁻¹, a rate more than 1000 times greater than that of pure MAPbI₃ under the visible light. (as depicted in Figure 9d) Importantly, the solar splitting efficiency peaks at 7.35%, positioning it among the highest efficiencies recorded so far. By incorporating MoS₂, which has a suitable band configuration and unsaturated entities, it can effectively promote charge separation and provide more active sites for hydrogen production.

5.2. Perovskite and 2D Material Heterostructures for CO₂ Reduction

Using solar energy for the photocatalytic reduction of CO₂ to transform it into renewable fuels is seen as a potential remedy for the escalating CO₂ levels in the atmosphere [120,121]. Unfortunately, most pure perovskites have low photocatalytic efficiency due to their fast charge recombination. Consequently, the creation of composite structures has been pursued to enhance the performance of perovskite-based catalysts. Graphene and its offshoots serve as superior charge transport mediums, facilitating the increase of electron extraction and transportation, thereby boosting catalytic efficiency. Xu et al. constructed a CsPbBr₃ QD/graphene oxide (GO) composite as a catalyst for CO₂ reduction, as shown in Figure 9e [114]. Compared with single CsPbBr₃ QDs, the electron consumption rate was increased from 23.7 μmol g⁻¹h⁻¹ to 29.8 μmol g⁻¹h⁻¹ after the introduction of GO. Photoluminescence and photoelectrochemical impedance, as showcased in Figure 9f, g prove that the boost in photocatalytic activity is attributed to the electrical connectivity offered by conductive GO. This ensures the effective conveyance of photogenerated carriers from the perovskite. Contrary to pure graphene, rGO nanosheets often display defects on their basal planes and edges due to lingering oxygen-related impurities [122]. Using the combination of perovskite with rGO-like defects, Wang et al. obtained an efficient catalyst for CO reduction [123]. Due to the composite structure of Cs₄PbBr₆/rGO, Cs₄PbBr₆/rGO exhibited higher selectivity (~94.6%) and 6.2 times higher CO₂ conversion (11.4 μmol g⁻¹h⁻¹) than pristine Cs₄PbBr₆. One of the key reasons is that these surface irregularities in rGO serve as trapping centers, enabling electrons from the excited Cs4PbBr6 to be adeptly channeled to the defect spots within rGO. The other is that oxygen deficiency acts as an active center, assisting CO₂ adsorption, activation and reduction of CO. Among these oxygen-based defects, the OH group emerges as the prime site for the photoreduction of CO to CO, due to its minimal energy barrier for CO adsorption, the creation of CO*, and its subsequent release. This is the first study to improve CO₂ photoreduction activity and selectivity by introducing OH defects, providing a new way to construct high-performance, low-cost perovskite-type solar energy conversion composites. While the heterostructures of perovskites and 2D materials have risen to prominence as potent materials in the realm of photocatalysis, the issue of lead toxicity continues to hinder their widespread adoption. Wang et al. synthesized a novel tin atom-sharing Cs₂SnI₆ perovskite nanocrystal/SnS₂ nanosheet heterojunction by a facile solution method [124]. A joint study combining DFT calculations and transient absorption and KPFM characterizations shows that holes are transferred from SnS₂ to Cs₂SnI₆, and electrons are transferred from SnS₂ to Cs₂SnI₆. Further investigations through photocatalytic CO₂ reduction and photo-electrochemical tests confirmed that the build-up of electrons in SnS₂ effectively extends the charge lifetime, enhancing efficiency by 5.4-fold and 10.6-fold. Given its non-toxic nature and ease of fabrication, this catalyst shows immense promise for widespread application.

Two-dimensional black phosphorus (BP) has also been widely used as a metal-free candidate catalyst for photocatalytic purposes and when combined with other semiconductors to form heterojunctions, BP exhibits superior performance in CO₂ photoreduction [125]. For the first time, Wang et al. synthesized a CsPbBr₃ heterostructure onto 2D BP, showing its efficacy as a photocatalyst for CO₂ reduction [126]. In comparison to the single-component perovskite CsPbBr₃, the introduced BP greatly enhances the photocatalytic activity, and the conversion rates of CO₂ to CO and CH₄ are 44.7 and 10.7 μmol g⁻¹ h⁻¹, respectively. CO₂ photocatalytic reduction of CH₄ and CH₄. The photocatalytic CO₂ reduction yields for CO and CH₄ rose by factors of 2.4 and 4.4, respectively. The CsPbBr₃/BP composition, with its structural and interfacial interactions, offers a higher number of active sites that facilitate the adsorption and activation of CO₂. Concurrently, the integrated BP serves as a channel for electron transfer, enhancing the effective use of charge carriers that reach the surface. On the other hand, MXene is also an excellent conductive material. Pan et al. synthesized CsPbBr₃/Ti₃C₂T_x nanocomposites by in situ growth of CsPbBr₃ on the surface of monolayer Ti₃C₂T_x nanosheets [127]. The pronounced quenching of fluorescence in CsPbBr₃ nanocrystals, alongside time-resolved photoluminescence lifetime and photoconductivity assessments, confirmed the effective charge transfer between Ti₃C₂T_x and CsPbBr₃ nanocrystals. When exposed to simulated sunlight, the CO production rate for CsPbBr₃/Ti₃C₂T_x clocked in at 26.32 μmol g⁻¹ h⁻¹. This is significantly superior to the yield from standalone CsPbBr₃ nanocrystals, which was less than 4.4 μmol g⁻¹ h⁻¹.

This section mainly describes the application of perovskite/two-dimensional material conforming structure in photocatalysis, including CO₂ emission reduction and H₂ production. People are facing huge problems with energy and the environment. Photocatalytic CO₂ emission reduction combined with H₂ production presents an effective solution to these challenges. While the exploration of perovskites and 2D materials in heterostructures for photocatalysis is still in its early stages, the current study opens promising prospects for visible light-driven photocatalysts based on perovskite and 2D material heterostructures.

Table 4. Summary of the key parameters for photocatalysis based on perovskite and 2D material heterostructures.

Heterostructure Type	Photoreaction Type	Reaction Solution	Activity	Light Source	Reference
MAPbI3/RGO	H2 production	HI	H2 (93.9 µmolh−1)	120 mWcm−2 visible-light (λ ≥ 420 nm)	[33]
Cs2AgBiBr6/rGO	H2 production	HBr	H2 (498 µmolg−1)	visible light irradiation (λ ≥ 420 nm, 300 W Xe lamp)	[115]
MAPbI3/BP	H2 production	HI	H2 (3742 µmolh−1g−1)	visible light (λ ≥ 420 nm)	[116]
MAPbI3/MoS2	H2 production	HI	H2 (206.1 µmolh−1)	visible light	[118]
MAPbI3/MoS2	H2 production	HI	H2 (13,600 µmolh−1g−1)	visible light (λ ≥ 420 nm)	[119]
MAPbI3/MoS2	H2 production	HI	H2 (30,000 µmolh−1g−1)	visible light (λ ≥ 420 nm)	[113]
CsPbBr3 QD/GO	CO2 reduction	Ethyl acetate	CO (58.7 µmol g−1) CH4 (29.6 µmol g−1) H2 (1.58 µmol g−1)	100 W Xe lamp with an AM 1.5 G filter	[114]
Cs4PbBr6/rGO	CO2 reduction	Ethyl acetate	CO (11.4 µmol g−1h−1)	300 W Xe lamp equipped with 420 nm filter	[123]
Cs2SnI6 NC/SnS2	CO2 reduction	-	CO (6.09 µmol g−1)	150 mWcm−2 Xe lamp with 400 nm filter	[124]
CsPbBr3 NC/BP	CO2 reduction	Ethyl acetate	CO (44.7 µmol g−1h−1) CH4 (10.7 µmol g−1h−1)	200 mWcm−2 Xe lamp	[126]
CsPbBr3 NCs/Ti3C2Tx	CO2 reduction	Ethyl acetate	CO (26.32 µmol g−1h−1) CH4 (7.25 µmol g−1h−1)	300 W Xe lamp with a cut off filter >420 nm	[127]

summarizes the relevant parameters of photocatalysts based on perovskite and 2D material heterostructures in this section.

6. Conclusions and Perspectives

This review covers the synthesis methods of perovskite and 2D material heterostructures and their applications in photodetectors, solar cells and photocatalysis. First, the fabrication techniques are introduced for perovskite and 2D material heterostructures, including solid-state and solution methods. These can be broadly classified into solid-state and solution methods. The solid-state techniques are suitable for producing high-quality perovskite and 2D material heterostructures characterized by sharp and clean interfaces. In contrast, solution-processed techniques are more apt for mass-scale manufacturing of these heterostructures. We then partition the exploration of applications into three distinct segments: photodetectors, solar cells, and photocatalysis, detailing the advancements and potentialities in each domain. Due to fast charge transfer and transport at the interface, combining 2D materials with perovskite facilitates better figure-of-merits in hybrid photodetectors than those with a single material. In addition, the addition of 2D materials inside PSCs leads to more stable perovskite thin films and more efficient charge separation and collection when 2D materials are used as ETLs and HTLs additives. In addition, perovskite and 2D material heterostructures in photocatalysis exhibit great potential as high-performance and stable photocatalysts due to ultrafast charge separation and transport in perovskites and 2D materials. Numerous studies in materials and physics still await, with a wealth of potential yet to be unearthed. Some foundational characteristics of perovskite and 2D material heterostructures have already been examined, including PL quenching, enhancement, and shifts attributable to interlayer charge transfers, as well as interface-driven broadband emissions and selective spin injections. Nevertheless, other prospective properties remain unverified in experimental settings. A deeper investigation into the optical properties would be beneficial, especially at lower temperatures. Analyzing these properties in relation to temperature can offer valuable insights into the recombination mechanism and phonon interactions, and such temperature-dependent studies can help identify fine spectral components that cannot be discernible at room temperature.

While perovskite and 2D material heterostructures hold immense promise, several hurdles persist, notably concerning stability, scalability, and material safety. To begin with, the stability of perovskite and 2D material heterostructures is often compromised. The inherent sensitivity of perovskite to external factors such as moisture, oxygen, light, and heat largely contributes to its poor stability [128]. Therefore, it is inevitable to utilize appropriate passivation strategies for fabricating devices with high performance and long-term stability. Chemical passivation primarily addresses intrinsic point defects through chemical bonding, while physical passivation relies on physical processes like strain relaxation or treatments such as temperature adjustments, lasers and adhesive tape [129]. Therefore, enhancing the stability of perovskite and 2D material heterostructures using both chemical and physical passivation techniques could offer a viable path forward. At the same time, novel perovskites have been developed for stability enhancement. Notably, the 2D layered Ruddlesden–Popper perovskites, which incorporate hydrophobic organic chains, have displayed superior photostability when juxtaposed with their traditional 3D counterparts [130]. Similarly, 2D layered Dion−Jacobson perovskites have emerged as a potential material with excellent environmental stability. The unique structure of Dion−Jacobson perovskites features adjacent inorganic layers connected by diammonium ligands. These ligands serve to eradicate the van der Waals (vdWs) gaps and reduce the inter-slab distance. Consequently, they offer enhanced stability when compared with the 2D layered Ruddlesden–Popper perovskites [131]. These novel 2D layered perovskites are promising candidates for fabricating stable perovskite and 2D material heterostructures.

Secondly, there is a trade-off between high-quality perovskite and 2D material heterostructures and high yields. Though solution-processed methods facilitate mass production of perovskite and 2D material heterostructures, the heterostructure quality is adversely affected by the poor contact and large lattice mismatch at the heterointerface. It has been reported that seed engineering is a successful approach to the controlled synthesis of hierarchical heterostructures, which may also be used for the highly controlled epitaxial growth of perovskite and 2D material heterostructures [132]. It also has been demonstrated that the dangling-bond-free surfaces of 2D material nanosheets as synthetic templates may trigger the solution-phase epitaxial growth of other materials [133]. Therefore, the solution-processed epitaxy technique may provide a promising construction of perovskite and 2D material heterostructures.

Lastly, the most extensively studied perovskites in heterostructures contain toxic and dangerous heavy metals, especially Pb, which is concerning both for environmental reasons and human health. This toxicity issue of lead-related perovskite and 2D material heterostructures can be solved by replacing Pb with other elements such as Sn, Mn, and Zn [134]. However, performances of optoelectronics or photocatalysis are still unsatisfactory compared with their Pb-based counterparts. For instance, devices based on Sn underperformed when compared to those using Pb. This inferior performance can be attributed to a high defect density within the perovskite lattice and the significant oxidation of Sn, transitioning from a +2 to a +4 state, which hampers their prolonged stability. Therefore, the development of perovskites based on non-toxic substances is required for future advancement of perovskite and 2D material heterostructures optoelectronics.