Shedding light on the energy applications of emerging 2D hybrid organic-inorganic halide perovskites.

Unique performance of the hybrid organic-inorganic halide perovskites (HOIPs) has attracted great attention because of their continuous exploration and breakthrough in a multitude of energy-related applications. However, the instability and lead-induced toxicity that arise in bulk perovskites are the two major challenges that impede their future commercialization process. To find a solution, a series of two-dimensional HOIPs (2D HOIPs) are investigated to prolong the device lifetime with highly efficient photoelectric conversion and energy storage. Herein, the recent advances of 2D HOIPs and their structural derivatives for the energy realms are summarized and discussed. The basic understanding of crystal structures, physicochemical properties, and growth mechanisms is presented. In addition, the current challenges and future directions to provide a roadmap for the development of next generation 2D HOIPs are prospected.

Introduction

Since the advancement of industrialized society and the increasing demand for energy, the fossil fuel-based energy structures faced unprecedented challenges. Currently, more than 80% of the world's energy demand is mainly produced from the burning of fossil fuels, promoting massive flux of natural environmental threats such as global warming, air pollution, water pollution, and the depletion of natural resources, which are receiving urgent attention (Wang et al., 2015a). Concerted efforts have been made to minimize the effect of environmental pollution and to reduce the consumption of traditional energy (Ng et al., 2021, Ong et al., 2020a, Ong and Shak, 2020b, Peng et al., 2019, Zhang et al., 2020a). To name a few, solar energy is an ideal candidate for a clean energy source that is readily available, abundant in reserves, sustainable, easily-accessible, and environmentally-friendly. Therefore, the exploration of new materials has become a focal area in the context of energy research.

To meet the energy challenge, many researchers have studied how to enhance the conversion of solar energy, they have focused on the traditional perovskite oxide ABO3, where “A” stands for rare earth or alkaline earth metal, and “B” represents a first-row transition metal. In similar fashion, the crystal structure of the halide perovskite material is similar to that of the general formula ABX3, where A site is Cs or organic cation, the B site is a smaller cation (Pb, Sn, and Ge), and the “X” is a halide anion that bonds to the “A” and “B” cations. This material has attracted people's attention in the field of catalysis (Chen et al., 2020; Zejian et al., 2021). Ever since CH3NH3PbI3 (MAPbI3) was first applied to dye-sensitized solar cells by Miyasaka’ group (Kojima et al., 2009), a great progress has been shown in the evolution of the power conversion efficiencies (PCE) of lead halide perovskite solar cells within a few years. The recorded PCE of MAPbI3-based DSSC increased rapidly from 3.8% in 2009 to 22.1% in 2016 (Kojima et al., 2009; Chen et al., 2015; Jeon et al., 2014; Jeon et al., 2015; Lee et al., 2012; Liu et al., 2013; McMeekin et al., 2016; Saliba et al., 2016a; Saliba et al., 2016b; Shao et al., 2016; Shin et al., 2017; Yang et al., 2015; Zhou et al., 2014), which fascinates the development of photovoltaic industry. The results show that the excellent and distinctive photovoltaic characteristics of halide perovskites are the main reason for its rapid improvement, such as a favorable band gap, extremely high optical absorption coefficients, a strong defect tolerance, and shallow point defects, ultra-long photogenerated carrier lifetime, as well as the unprecedented long diffusion length compared to the other polycrystalline thin film solar cell materials. In addition, halide perovskites also have good application prospects in other energy fields (Kim et al., 2018, 2020a). For example, halide perovskite has current-voltage (I-V) hysteresis behavior, which makes it applicable to resistance switching memory (Choi et al., 2017; Kim et al., 2019a, 2020b).

However, the bulk (three-dimensional, 3D) halide perovskite still has two major problems, namely thermodynamics instability and lead toxicity, which are the main obstacles on the road to commercialization. Therefore, researchers are continuously trying to solve the fundamental problems that shrouding the limiting factors of bulk halide perovskites. The low-dimensionality of materials has become a choice for many researchers to explore material modification (Li et al., 2021; Xu et al., 2018a). Then, 2D organic-inorganic halide perovskites (2D HOIPs) nanomaterials have garnered great attention because of their unprecedented physical properties and their hold great promise compared to their bulk counterparts (Jiang et al., 2019). As shown in Figure 1, the number of citations has increased rapidly each year, and the growth rate of 2D perovskites is obviously higher than that of 3D perovskites, which shows that more and more researchers have discovered the research prospects of this material.

Figure 1

Research trends of halide perovskites for energy in the past five years

The growth rate (compared to the first year) of (A) publications and (B) citations of articles on "2D halide perovskites for energy", "3D halide perovskites for energy" in the last 5 years (the data is from the database Web of Science as of December 2021).

Unlike 3D halide perovskites, 2D HOIPs is able to incorporate bigger, less-volatile, and hydrophobic organic cations that promote the active materials with better thermal stability and chemical stability. To this end, we have summarized and compared the relevant properties of 2D and 3D halide perovskites in Table 1, using solar cells as an example. Moreover, 2D HOIPs feature the strong light-matter interaction, tunable thickness-dependent properties, structural relaxation, and photoluminescence shift. This structure can be considered as natural multiple quantum wells, where the insulating organic layers served as potential “barriers” and the semiconducting inorganic layers served as potential “wells” (Kataoka et al., 1993; Muljarov et al., 1995). In this configuration, the electronic confinement emerges in subnanometer layers that lead to the formation of stable excitons with exceptionally high binding energy. Moreover, the Bohr radius could expand beyond the limit of a single layer, because of not only dimensional confinement but also the dielectric properties modulated by the organic material. Consequently, the 2D HOIPs are useful in the field of electronic and optical devices, such as field-efiect transistors (FETs) (Kagan et al., 1999) and light-emitting diodes (LEDs) (Ishihara et al., 1989; Shang et al., 2019; Wang et al., 2016) for their unique optoelectronic performance. Studies have also shown that the photovoltaic performance of 2D HOIPs can be enhanced by adjusting non-covalent interactions (such as “CH···π” interaction) (Yan et al., 2020).

Table 1

Comparison of characteristics between 2D and 3D halide perovskites in solar cell

Perovskite	Dimension	Band gap	PCE(%)	Voc(V)	Stability	Ref
MAPbI3	3D	1.55	21.09	1.076	NA	(Chen et al., 2019a)
MAPbI3	3D	1.55	14.8	1.04	NA	(Guo et al., 2017a)
MAPbI3	3D	1.55	16.79	1.04	86% after 50 days in air	(Guo et al., 2019)
MAPbICl2	3D	1.65	9.3	0.82	NA	(Mehdi et al., 2020)
FAPbI3	3D	1.53	19.5	1.07	NA	(Jo et al., 2016)
α-FAPbI3	3D	1.53	25.21	1.174	65% for reference cell and 90% for target cell after 1000 h in the dark at 25°C RH 20	(Jeong et al., 2021)
BA2MA3Pb4I13	quasi-2D	1.52–2.43	12.51	1.01	70% after 2250 h in RH 65 under light	(Tsai et al., 2016)
PEI2PbI4-MAPbI3	quasi-2D	2.3	16.06	1.08	90% after 200 h in RH 50 under dark	(Yao et al., 2015)
MAPbI3-PEA2PbI4	quasi-2D	NA	19.89	1.17	64% after 30 days in RH 20–30 under dark	(Bai et al., 2017)
GA2MA4Pb5I16	2D	NA	19.3	1.17	94% after more than 3000 h storage under ambient conditions	(Huang et al., 2021)
FAxPEA1-xPbI3	2D	1.52	17.7	1.04	Stability measured at maximum power point up to 2 weeks	(Li et al., 2017)
(BA)2 (MA)n-1PbnI3n+1	2D	NA	19.56	1.11	96.5% after 100 h at 95°C under dark	(Lin et al., 2018)
(PEI)2 (MA)6Pb7I22	2D	1.62	10.08	1.10	90% after 100 h under light	(Yao et al., 2016)
(CA)2 (MA)n−1PbnI3n+1	2D	1.63	9.03	0.88	NA	(Koh et al., 2016)

In recent studies, researchers have studied the semiconductor physics of layered perovskites (Blancon et al., 2020, Mauck and Tisdale, 2019). Some researchers reviewed the properties, preparation, and application of low-dimensional perovskites. Although some researchers have reviewed 2D HOIPs applications in the direction of optoelectronic devices (Zhang et al., 2020b), a systematic review of the application of 2D HOIPs in energy has not yet appeared. At the same time, there is an increase in number of literatures everyday, which deserves another focused review. In this review, the progress of 2D HOIPs materials in recent years is summarized, including some 3D layered HOIPs materials, and in particular current advances in terms of their applications for photoelectric conversion, energy storage, photovoltaics, LEDs, batteries, supercapacitors, and thermoelectrics are reviewed. Meanwhile, the corresponding basic theoretical developments, which are beneficial to understand how 2D HOIP's excellent and unique photovoltaic characteristics are produced, are also reviewed. As shown in Figure 2, we summarize an overview of outstanding developments in the field of 2D HOIPs(Kagan et al., 1999; Mitzi et al., 1994; Even et al., 2013; Smith et al., 2014; Dou et al., 2015; Ha et al., 2016; Stoumpos et al., 2017; Liu et al., 2017a; Mao et al., 2018; Shao et al., 2018; Zhang et al., 2019a; Lin et al., 2019). Finally, the perspectives and outlooks are involved in this review of the emerging 2D HOIPs materials, i.e., designing the highly anticipated lead-free perovskite to resolve the Pb toxicity, air-instability, and temperature-instability issues.

Figure 2

Research timeline of the developments in 2D HOIPs

Reproduced with permission (Mitzi et al., 1994). Copyright 1995 AAAS. Reproduced with permission (Kagan et al., 1999). Copyright 1999 AAAS. Reproduced with permission (Even et al., 2013). Copyright 2013 American Chemical Society. Reproduced with permission (Smith et al., 2014). Copyright 2014 Wiley-VCH. Reproduced with permission (Dou et al., 2015). Copyright 2015 AAAS. Reproduced with permission (Ha et al., 2016). Copyright 2016 Nature Publishing Group. Reproduced with permission (Stoumpos et al., 2017). Copyright 2017 Cell Press. Reproduced with permission (Liu et al., 2017a). Copyright 2017, American Chemical Society (Mao et al., 2018). Copyright 2018 American Chemical Society. Reproduced with permission (Shao et al., 2018). Copyright 2018 Wiley-VCH. Reproduced with permission (Zhang et al., 2019a). Copyright 2019 Wiley-VCH. Reproduced with permission (Lin et al., 2019). Copyright 2019 American Chemical Society. Reproduced with permission (Shi et al., 2020). Copyright 2020 Springer Nature Ltd.

Rational design of 2D HOIPS

2D HOIPs structure

S.N. Ruddlesden and P. Popper first reported the crystal structure of 2D layered inorganic perovskites in 1957 (Ruddlesden and Popper, 1958). The periodical arrangements consist of 2D perovskite layers interleaved with cations and therefore named after these two pioneers as Ruddlesden-Popper (RP) perovskites. The 2D HOIPs materials are analogues of RP phase structure and represent the general formula of L2An-1BnX3n+1. In the molecular formula, L stands for a large cation, A represents a regular cation such as Cs+ or MA+, B stands for a divalent metal cation such as Pb2+ or Sn2+, and X is a halide. In this formula, n denotes the number of the octahedral stacking layers of metal halide sheets. The intergrowth of two L cation layers and n metal halide octahedral layers formed the 2D HOIPs structure. In this structure, apart from the van der Waals interaction between different units, the inorganic perovskite layers can be combined with organic perovskite layers by either ionic bonding or hydrogen bonding. Therefore, the 2D HOIPs structures are stabilized by different types of interactions compared to the bulk perovskite structure.

The basic units of 2D HOIPs and its morphological structure are shown in Figures 3A and 3B, respectively. The single-layer 2D HOIP is an ideal quantum well (QW) material. Typically, the “A” site is occupied by methylammonium (MA+) ligands; the organic layers can be single or double layers depending on the ammonium compounds (Stoumpos et al., 2017). The RP phase perovskite is such a structure with a double-layered organic layer. The weak van der Waals interactions between the organic layers of this structure affect the structural stability of the perovskite. In contrast, the Dion-Jacobson (DJ) phase composed of single organic layer, and its organic cations and inorganic perovskite layers are connected by hydrogen bonds instead of weak van der Waals interactions between the organic layers, and thus exhibit better structural stability (Mao et al., 2018; Mao et al., 2019; Tremblay et al., 2019). We have provided common organic cations in 2D HOIPs in Table 2. In addition, the 2D halide perovskites can be obtained by reducing the thickness of the corresponding perovskite through delamination. The top and bottom surfaces are terminated by “A” cations instead of “L” cations (Shi et al., 2018).

Figure 3

Structure information of 2D HOIPs

(A) Structures and (B) schematic of the 2D HOIPs.

(C) Summary of bandgap energy of typical 2D HOIPs. (BA)2PbI4 (Stoumpos et al., 2016), BA2SnI4(Chen et al., 2017a), (FC2H4NH3)2PbCl4 (Lermer et al., 2016), PEA2PbI4(Peng et al., 2017a), PEA2PbI4·(MAPbI3)2 (Peng et al., 2017b), Cs4SnBi2Br12(Xu et al., 2018b), EA4Pb3Br10(Mao et al., 2017a), EA4Pb3Cl10(Mao et al., 2017a), (C9H19NH3)2PbI2Br2(Abid et al., 2017), Cs2[C(NH2)3]Pb2Br7 (Nazarenko et al., 2017), Cs[C(NH2)3]PbI4 (Nazarenko et al., 2017), Rb2PbI4(Guo et al., 2017b), (BA)2SnI4(Wang et al., 2018)

Table 2

Common organic cations

Organic cation	Molecular structure	Organic cation	Molecular structure
MA		EA
PA		BA
i-BA		BYA
FA		GA
N-MPA		N-MEDA
3AMP		4AMP
PEA		PMA
CEA		BEA

A property that is closely related to the crystal structure of 2D HOIPs is their electronic bandgaps. It is widely accepted that the flexibility of both crystal structure and electronic properties gives these 2D HOIPs great opportunities for diverse physicochemical properties such as crystal phase, bandgap energy, and effective mass of carriers. As such, a wide range of studies have been devoted to choosing different cations to achieve band gap regulation, and finding the most suitable cation for their required performance is also a major challenge for researchers. A summary of bandgap energy of typical 2D HOIPs is shown in Figure 3C. In addition, there are some other challenges in the application of 2D HOIP, such as stability, lead toxicity, and the possibility of large-scale production. Therefore, an extended overview will be introduced in the following synthesis and theoretical section.

Synthesis of 2D HOIP

Solution methods

In recent years, some representative techniques have been used to prepare 2D HOIPs materials. The most efficient and convenient preparation methods can be divided into two categories, namely solution and vapor methods. The solution method is one of the main preparation methods of 2D HOIPs that are widely accessible using low temperature deposition techniques. The basic procedure is essentially to gently add metal halides and organic ammonium salts into the cosolvent, and after sufficient reaction time, 2D HOIPs samples can be obtained by dry casting or aging procedure, as shown in Figure 4A. Dou et al. obtained atomically thin 2D HOIPs [(C4H9NH3)2PbBr4] and derivatives by growing directly from solution (Dou et al., 2015). A ternary cosolvent was used to synthesize uniform square-shaped 2D crystals with high yield and excellent reproducibility on a flat substrate, as shown in Figures 4B and 4C. Following the same recipe, Chen and coworkers employed a chlorobenzene-dimethylformamide-acetonitrile ternary solvent method to prepare the 2D (C4H9NH3)2PbBr4 perovskites (Chen et al., 2017a). A larger lateral dimension and smaller thickness of a few nanometers were obtained. The critical factors, including the influence of solvent volume ratio, solvent polarity, and crystallization temperature on the growth dynamics were systematically studied. It is found that the combined effects of c-axis suppression and diffusion-mechanism-dominated branched growth determine the 2D HOIPs framework.

Figure 4

Preparation of 2D HOIPs using a solution method

(A) Solution-based synthesis schematic of 2D HOIPs.

(B) TEM image of (C4H9NH3)2PbBr4 sheets.

(C) Optical image of the 2D square sheets. Reproduced with permission (Dou et al., 2015). Copyright 2015 American Association for the Advancement of Science. And TEM images of the CH3NH3PbBr3 nanodots (D) and the resultant nanoplates (E). Reproduced with permission (Liu et al., 2017b). Copyright 2017 Wiley-VCH

Apart from the solution methods mentioned above, colloidal synthesis can be an alternative approach reported by Schmidt et al. (Schmidt et al., 2014). In addition, Tyagi et al. managed to obtain 2D monolayer methylammonium lead bromide perovskite (CH3NH3PbBr3) crystal using colloidal method (Tyagi et al., 2015). In their works, lead bromide and methylammonium bromide were dissolved in dimethylformamide; subsequently, the mixture was added into a toluene solution to obtain a colloidal solution of CH3NH3PbBr3 nanostructures. It turns out that the PL emission peak was blue-shifted about 0.5 eV from the sharp excitonic absorption feature of the bulk MAPbBr3 phase, confirming the feature of 2D nanostructure.

Perovskite nanoplatelets with different thicknesses can be acquired via adjusting the ratio of organic cations used in the synthesis, namely octylammonium and the shorter methylammonium. The layer thickness of the product was decreased as the content of octylammonium in the precursor suspension increased. Similar method was employed to synthesize CH3NH3PbBr3 nanosheets with different layers (Huang et al., 2017). Zhong and his coworkers proposed the synthesis of uniform CH3NH3PbBr3 nanoplatelets through the self-organization method of preformed colloidal CH3NH3PbBr3 nanodots (Figures 4D and 4E) (Liu et al., 2017b). Maintaining the obtained colloidal nanocrystals at 3.5 mg/mL (a high concentration) for three days or combining the synthesis of nanodots with self-organization for tailoring the geometric transformation from spherical nanodots to square or rectangular nanoplatelets. These findings highlight successful examples in the preparation of 2D HOIPs nanoplatelets.

2D HOIPs can be grown by the noncolloidal approaches to tailor a well-defined thickness (the value of “n”) and large lateral dimensions. However, the drawback of this approach is lack of process ability after material deposition. Moreover, it is difficult to transfer the as-deposited samples to other desired substrates. In comparison to the noncolloidal chemistry, the colloidal methods offer facile and fast routes to synthesize 2D HOIPs. However, the as-synthesized 2D HOIPs based on colloidal solution method can be difficult to control their thickness and lateral dimensions size. Therefore, significant attempts have been made to improve the output of colloidal approaches. For instance, by changing the HBr amount, the ligand amount, and their ratio, the temperature of the synthesis, the reaction time, as well as the length of ligands, the morphologies of 2D HOIPs can be varied, so that the thickness-dependent and size-dependent properties are characterized. In the aforementioned wet chemistry approaches, surfactant plays an important role to accommodate morphology improvement during the synthesis process. However, a surfactant could contaminate the surface materials and eventually negatively influences the intrinsic electronic properties of 2D materials.

Vapor methods

The vapor method includes diverse and versatile deposition techniques, namely, the chemical vapor deposition (CVD), and van der Waals epitaxy is one of the most widely explored approach for other 2D materials such as grapheme (Lang et al., 2015; Wang et al., 2011), transition metal dichalcogenides (TMDs) (Jiang et al., 2017; Qi et al., 2017; Suleiman et al., 2020; Zhang et al., 2019b), and 2D transition-metal carbides/nitrides, and carbonitrides (MXenes) (Xu et al., 2015). It is also applied in the growth of perovskite thin films (Lewis and O'Brien, 2014, Leyden et al., 2014). Ha and his coworkers synthesized the organic-inorganic lead halide perovskite nanoplatelets with a two-step vapor method (Ha et al., 2014). First, the PbI2 nanoplatelets on the surface of muscovite mica were prepared in a vapor transport chemical deposition system utilizing van der Waals epitaxy.

The as-grown platelets were transformed to CH3NH3PbI3 perovskites through a gas-solid hetero-phase reaction with methylammonium halide molecules. The schematic of this method with the vapor-transport system is shown in Figure 5A. The color of the lead halide platelets can be tuned as a function of thickness. The thickness of PbI2 platelets had a linear relationship with the CH3NH3PbI3 platelets, whereby the ratio follows the c lattice constant variation. Liu group has made significant contributions to the development of the vapor synthesis of 2D HOIPs(Niu et al., 2015). In 2015, organic/inorganic van der Waals solids that consist of perovskite and some other 2D materials such as graphene, MoS2, and hexagonal boron nitride (h-BN) with CVD were fabricated (Figure 5B) (Niu et al., 2015).

Figure 5

Preparation of 2D HOIPs using a vapor method

(A) Schematic of the preparation of 2D HOIPs by vapor methods.

(B) Morphologies of the perovskite/2D van der Waals (VDW) solids probed by scanning electron microscopy. Reproduced with permission (Niu et al., 2015). Copyright 2015 Wiley-VCH

In the following years, 2D CH3NH3PbI3 microplatelet arrays and CH3NH3PbI3 nanoplatelets on the silicon substrate with single layer h-BN film as the buffer layer were successfully synthesized (Liu et al., 2016; Niu et al., 2016). The h-BN film acted as not only the intermediate layer, but also as the growth nuclei site. By changing the synthesis condition, the thickness of nanoplatelets is controllable between a few layers to hundred nanometers. Wang et al. prepared 2D MAPbCl3 perovskite with operating temperature at 220°C using the dual precursor sources CVD method (Wang et al., 2015b). In their study, the fractal evolution of MAPbCl3 thin films and the VDW epitaxy mechanism was explained in combination with Monte Carlo simulations. It is the weak interaction of the VDW film with the substrate and the low cohesive energy of the perovskite halide that make large-scale growth of single crystal ultrathin 2D crystals possible.

When it comes to the all-inorganic metal halide perovskites, it is not feasible to use vapor methods for the synthesis, because of their strong covalent bonding and non-van-der-Waals characterization. However, Zhang et al. successfully synthesized CsPbX3 (X = Cl, Br, I) via VDW epitaxy method (Zhang et al., 2016). This serves as a generic recipe of vapor methods in all-inorganic metal halide perovskites fabrications (Hu et al., 2017; Wang et al., 2017a; Zhou et al., 2017). Compared with solution methods, the 2D HOIPs structures prepared by CVD or VDW epitaxy method have shown good crystallinity and high purities. It is noted that Chen et al. developed a self-doping approach for preparing ultrathin and large sized 2D homologous (C4H9NH3)2(CH3NH3)n-1PbnBr3n+1 perovskites by using 2D (C4H9NH3)2PbBr4 perovskites as a template with CH3NH3+ dopant by incorporating both solution and vapor methods (Chen et al., 2017b). In the first step, the 2D (C4H9NH3)2PbBr4 perovskites were obtained by ternary solvent method. Transformation of 2D (C4H9NH3)2PbBr4 into 2D homologous (C4H9NH3)2(CH3NH3)n-1PbnBr3n+1 perovskites was implemented by CVD of CH3NH3Br. Interestingly, it took the advantages of both solution and vapor methods by changing the direction of source vapors and carrier gas flow, which can selectively oxidize and etch the unstable nuclei and edges. Hence, realizing continuous domain size and sequential edge-epitaxy growth is required to promote the controllable synthesis of high-quality 2D materials.

Among the two main types of preparation methods, the solution method, in which the reactants are added directly to the cosolvent for reaction, can synthesize two-dimensional crystals with high yields and good reproducibility because of its less demanding reaction conditions. In addition, by changing the reaction conditions to influence the growth kinetics, larger size, and smaller thickness of the material can be obtained. However, the addition of solvents inevitably contaminates the surface material and affects the properties. In contrast, good crystallinity and high purity of 2D crystals can be obtained by the vapor method, but the thin film materials prepared by the conventional vapor method tend to be thicker, so researchers have made it possible to synthesize thinner and larger 2D crystals by adding VDW epitaxy method.

Theoretical designing

Material design by DFT calculations holds a predominant role in the study of the physicochemical properties of hybrid perovskites. From theoretical constructions and calculations of the expected crystal structure, we can effectively predict their possible performances, and provide a systematic guidance on how to synthesize the desired structures, explain the growth process, and to analyze the intrinsic feature of the device performance.

The strong spin-orbit coupling effect cannot be neglected in these halide perovskites as they could manifest as about 1 eV to the bandgap, while the underestimated bandgap energy through the LDA method has similar value. In 2016, Kanatzidis et al. (Stoumpos et al., 2016) first reported the crystal, electronic structures and optical properties of (BA)2(MA)n-1PbnI3n+1 HOIPs with varied n value. The results revealed that by increasing n value, the bandgap energy decreases. The particular perovskites exhibit promising PL features when the value of n is smaller than 2, on the other hand, this characteristic is not appealing for photoelectric applications. When the n is larger than 3, these perovskites cover the visible range of solar energy, leading to a huge potential in the conversion of visible light irradiation into electric energy. There are several classes of hybrid perovskites that are potentially promising candidates for photovoltaics. Fundamental criteria of the electronic properties of monolayer HOIPs of BA2MI4 (M = Pb, Ge, or Sn) have been investigated by DFT with Heyd-Scuseria-Ernzerhof (HSE) functional level (Ma et al., 2017). In the calculation result, the bandgap of 2D BA2MI4 is found to be 1.5–2.0 eV, whereas the ultrathin 2D BA2GeI4 and BA2SnI4 show the bandgap energy of 1.7 and 1.5 eV, respectively.

In 2017, Stoumpos et al. (Stoumpos et al., 2017) have successfully synthesized (CH3(CH2)3NH3)2(CH3NH3)4Pb5I16 in high yield and purity. In addition, a computational calculation of the corresponding electronic structure is presented. The inclusion of spin-orbital coupling (SOC) gives rise to the conduction band lifting that consequently lowered the bandgap of 0.28 eV located at Г(Stoumpos et al., 2017). In addition, the flat dispersions along the Z-Г and Y-Г directions in the reciprocal space suggest the nature of layered structure and the absence of electronic coupling between interlayers, reflecting the 2D nature of the hybrid materials.

In the process of preparing 2D CH3NH3PbBr3 by colloidal approach, Zhong et al. (Liu et al., 2017b) used DFT to study the self-assembly process of hybrid perovskites (Figure 6A). Two and four coalescent nanodots (NDs) have the most stable structure during the self-organization process by taking into consideration of the variation of dipolar vectors (Liu et al., 2017b). As such, a wide range of studies systematically investigated the function of organic groups by the DFT method. Since the organic ligands could serve as a dominant role in determining the dynamic structure of halide perovskites, it is quite difficult to unambiguously determine the precise position of organic groups because of the random orientation. Defect can be unintentionally generated during structural transformation from bulk perovskites into 2D HOIPs. Owing to the nature of layered structure, 2D HOIPs introduce more defects that can be related to the grain boundaries, which act as recombination centers and thereby decrease the PL quantum yield and charge carrier lifetime (Iyikanat et al., 2017). Therefore, extended studies have been focused to understand the implication of defects in the 2D HOIPs-based device performance.

Figure 6

Defect state analysis of halide perovskite

(A and B) The potential field distribution in a two dot alignment of CH3NH3PbBr3 NCs(Liu et al., 2017b). (C) Electronic structure of typical III-V, II-VI, or group IV defect intolerant semiconductors compared to (D) the defect tolerant lead halide perovskite (Brandt et al., 2017).

(C and D) are reproduced with permission from ref (Brandt et al., 2017). Copyright 2017, American Chemical Society.

(E) Normalized orbital overlap (NOO) to describe the similarity of the states near the band edges (Pandey et al., 2016). (E) is adapted with permission from ref (Chen et al., 2019b). Copyright 2016, American Chemical Society

The nature of defect-tolerance of bulk halide perovskites has been well studied (Brandt et al., 2017, Shi and Du, 2014, Yin et al., 2014). Specifically, the contribution of defect-tolerance to the perovskite performance is investigated. The following factors are important to consider: the mixed ionic-covalent bonding (related to the large Born effective charges and high dielectric constant) (Du, 2014, Du and Singh, 2010, Park et al., 2019), excess of halogen (Srimath Kandada et al., 2016), insufficient cation orbital hybridization at halogen vacancies (with large interionic distance for cation and low coordination numbers for anion) (Shi and Du, 2014), small capture cross sections (with low charge states) (Brandt et al., 2017). The cations with the outermost electron configuration of ns2 (i.e. Tl+ and Pb2+) has a significant contribution to the shallow defect states in halide perovskites (Figures 6C and 6D) (Brandt et al., 2017). Using the normalized orbital overlap (NOO), Pandey quantitatively described the tendency to form mid-gap defect states and this feature resembles the electronic states near the band edges (Pandey et al., 2016). As shown in Figure 6E, the NOO values of different perovskites are significantly less than 1, which indicates the shallow defect levels within the bandgap for optimizing device performance (Pandey et al., 2016).

Energy applications

At present, there are many challenges in the development of various subfields in the energy field, such as the improvement of photoelectric conversion efficiency in the photovoltaic field, the balance between the reaction rate and reaction speed of photodetectors, and the joint improvement of supercapacitor capacity and discharge efficiency. In addition, there are some common problems such as stability. Among many exciting characteristics of 2D HOIP, the excellent optoelectronic properties, such as the large absorption coefficient, high carrier mobility, and long diffusion length, make them promising for energy applications. Up to now, great achievements have been made in various energy utilizations. Here, we summarize the recent progress of 2D HOIP for energy applications, including photovoltaics, and photodetectors, supercapacitor, LEDs, and Thermoelectricity.

Photovoltaics and photodetectors

In a short period of time, HOIP based solar cells have been significantly improved in terms of power conversion efficiency (PCE) from 3.8% to 25.6% (Jeong et al., 2021). After the boom of research in photovoltaic, breakthroughs were subsequently achieved in a range of energy conversion and storage fields such as light-emitting diodes (LEDs), supercapacitor, thermoelectric. Despite the recent spectacular progress and development, most of the devices based on HOIP with 3D crystal structures suffer from degradation caused by ion migration, high temperature, UV irradiation etc., which poses a challenge for the future commercialization. 2D HOIPs have significant advantages in comparison with their 3D counterpart not only in the superior stability but also in the higher structural diversity. However, 2D HOIPs with n = 1 have a considerably large bandgap (>2.0 eV), high exciton binding energy (∼300 meV), and strong electron–phonon coupling, which hampered its photovoltaic applications.

Herz et al. (Herz, 2017) described the underlying mechanism that defined the basic upper limit to charge-carrier mobility values in MHPs. This parameter can be adjusted by changing their stoichiometric number, n. They reported that the effective charge-carrier mobility generally decreased with increasing 2-phenylethylammonium (PEA+) content, with a trade-off between electronic confinement and layer orientation, the effective charge-carrier mobility of quasi-2D HOIP MA(PEA)2PbI7 can reach a maximum of 11 cm2/(V.s).

Recently, significant progress have been made in terms of the efficiency of quasi-2D HOIP based solar cells (Zhang et al., 2021), i.e., over 19% for the Ruddlesden–Popper phase, 18% for the Dion–Jacobson phase and the alternating cations in the interlayer space phase. Controlling the crystallographic planes of the inorganic layer with a preferential vertical alignment, and optimization of the thickness distributions of QWs within the film are two most effective ways to improve the efficiency of quasi-2D perovskite solar cells. For example, cation engineering on mixed halide perovskite is shown in Figures 7C–7F n-Butylammonium cation was introduced into a mixed-cation bulk perovskite FA0.83Cs0.17Pb(IyBr1-y)3, which was synthesized by Wang and coworkers (Wang et al., 2017b).

Figure 7

The application and performance of 2D HOIPs in the field of photovoltaics

(A) Schematic of 2D HOIPs solar cells.

(B) Cross-sectional SEM image of PSCs with 8 mg/mL PMABr/IPA solution treatment. Reproduced with permission (Bu et al., 2020). Copyright 2020 Elsevier Ltd.

(C) J-V characteristics of optimized perovskite solar cells with different active layers.

(D) Stabilized power output (SPO) of the champion cells.

(E) J-V characteristic of the best performing perovskite solar cells, and (F) the distribution of PCE. (E-F) are adapted and modified with permission from ref (Wang et al., 2017b). Copyright 2017 Nature Publishing Group

The formation of 2D HOIPs, interspersed among highly orientated bulk perovskite grains, which resulted in suppression of the nonradiative charge recombination. Moreover, the solar cells measurements indicated that an optimum butylammonium content led to average stabilized PCE of 18% with a 1.61 eV-bandgap perovskite, as shown in Figures 7C–6F. The stability under simulated lighting was also improved. In addition, after 1,000 h use of solar cells in the air, their ‘post burn-in’ efficiency can be well-maintained at about 80% of the initial, and the lifespan after encapsulating is close to 4,000 h.

Xiao and coworkers used materials synthesis, characterization, and DFT calculation methods to explore the photovoltaic characteristics of the 2D MA2Pb(SCN)2I2 (MA = CH3NH3+) perovskite of MA2Pb(SCN)2I2. These hybrid perovskites are significantly different from those reported in the previous literature (Xiao et al., 2016). Another work by Chen and coworkers has shown a similar theory that a strong vertical orientation is necessary to achieve a reliable solar cell performance (Chen et al., 2018). They fabricated thin films of BA2MA3Pb4I13 (BA = butylammonium, MA = methylammonium) by use of a simple solution treatment on various substrates at a temperature ranging from 25 to 140°C. The ex situ grazing incidence wide angle X-ray scattering (GIWAXS) results show, depending on the processing conditions, there are different degrees of preferential vertical orientation. The solar cells performance test indicated that BA2MA3Pb4I13 with higher degree of vertical orientation manifest higher PCE than those perovskites with randomized orientations. Furthermore, systematically designed in situ GIWAXS experiments confirmed that the heterogeneous nucleation and growth of BA2MA3Pb4I13 happen to the liquid-air interface, forming a top-crust with strong vertical direction. As such, insights into the self-assembly of 2D metal halide perovskite and its influence on the performance of optoelectronic devices will remarkably promote the wide application of 2D metal halide perovskites devices.

In spite of the notable stability enhancement and efficiency improvement, the PCEs of PSCs based on 2D or quasi-2D perovskite are far from performance of the 3D perovskite and still not sufficient for commercial applications. Recently, a 3D/2D bilayer perovskite structure, in which 2D perovskite is deposited on the 3D bulk film in the form of an extremely thin layer (Figure 7B) have been comprehensively employed to ensure a superb performance in terms of both the stability and PCE of PSCs. Huang et al. achieved an efficiency of 18.5% with negligible hysteresis and a stabilized efficiency of 17.9% for a PSCs based on wide bandgap (∼1.74 eV). In addition, the long-term stability of the devices is also remarkably enhanced as no degradation was observed after the devices are stored under ambient air conditions with 20% RH without encapsulation over 2000 h (Bu et al., 2020).

The first photodetector based on 2D HOIPs was successfully manufactured by Tan et al. (2016) Figure 8C shows the current-voltage (ISD-VSD) curves of the device in darkness and different illumination intensities. Here, the photocurrent of the device was gradually increased with the incident power. This 2D HOIPs crystal photodetector exhibits extremely low dark current with graphene gapped electrodes, indicating the good reliability of the perovskite photodetector performance. However, the electroluminescence efficiency of 2D HOIPs is limited by the presence of undesired nonradiative recombination effect, which is related to imperfection and leakage current caused by incomplete surface coverage.

Figure 8

2D HOIPs in Light-emitting devices

(A) Schematic of 2D HOIPs LED device.

(B) Current-voltage (ISD-VSD) curves under different illumination intensities. Reproduced with permission (Tan et al., 2016). Copyright 2016 American Chemical Society.

(C) PL spectra of corresponding 2D HOIPs. Reproduced with permission (Papadatos et al., 2017). Copyright 2017 Elsevier.

(D) PL spectra of (BA)4AgBiBr8 at high pressure. Reproduced with permission (Fang et al., 2019). Copyright 2019 Wiley-VCH.

Light-emitting devices

Light emission is a key property that needs to be controlled to achieve high performance in a variety of energy conversion devices including solar cells, LEDs, and lasers. HOIP have superior electrical and optical properties, which offer them great potential for use in LEDs. Their narrow emission linewidths (full width at half maximum (FWHM) ≤ 20 nm) can achieve ultra-high color purity (color gamut ≥140% in National Television Standards Committee (NTSC) TV color standard), which are superior than traditional light-emitting technologies such as inorganic quantum-dot LEDs (QLEDs, FWHM >40 nm; color gamut <100% in NTSC) and organic LEDs (OLEDs, FWHM ≈30 nm; color gamut ≈100% in NTSC) (Kim et al., 2019b).

In addition to the property needed for high performance solar cells, such as excellent charge transport properties and large mobilities entail, a key requirement for high performance LEDs is that the charges must be locally confined within the perovskite emissive layer to enhance radiative recombination. Owing to the strong quantum and dielectric confinement that offered by the disruption of the normally periodic perovskite lattice with bisecting organic layers, 2D HOIP can possess exceptionally high exciton binding energy and radiative recombination rate that for high performance LEDs (Walters et al., 2020). Therefore, the PL efficiency of 2D HOIP solids has been increased to near unity and EQEmax over 20% of pero-LEDs in the green, red, and infrared light regions (Wei and Xing, 2019).

The structure of 2D HOIPs optoelectronic devices is in principle similar to the aforementioned discussion on solar cell devices. Typically, a light-emitting hybrid perovskite emission diode consists of the positive and negative electrodes, protons and electron transport layers, and 2D HOIPs situated at the middle of device structure as “reaction vessels”. The confinement of the 2D structure allows protons and electrons to recombine more efficiently between layers to produce photons as shown in Figure 8A(Kim et al., 2015).

Wang et al.(Ishihara et al., 1989) demonstrated a solution-processed perovskite light-emitting diode (LED) device based on self-organized multiple quantum wells (MQWs). It has eminent film morphology, and its external quantum efficiency is as high as 11.7%, which is extremely stable. When the current density is 100 mA cm−2, the energy conversion efficiency is 5.5%, signifying excellent high-power performance. The perovskite MQWs with higher energy gap effectively drive these excellent optoelectronic performances, thereby producing effective radiation decay characteristics.

Recently, it was discovered that because of the self-trapped excitons (STEs), within the visible range, 2D HOIPs exhibited white-light and broadband emission. Mao and coworkers incorporated the bifunctional ammonium di-cations to 2D lead bromide perovskites that serve as templates for white light emitter applications (Mao et al., 2017b). Moreover, the corresponding PL research shows that the inorganic layer is directly related to the geometrical distortion of the "PbBr6" octahedron. The PL emission broadening indicates that the octahedral distortion of the inorganic layer becomes larger; therefore, the bandwidth of the PL emission is widened. For instance, the most distorted perovskite structure has the widest emission and the service life (T = 1.39 ns) (Tan et al., 2016). Among these three compounds, an intriguing “3 × 3” type of corrugated (110)-oriented 2D HOIPs, α-(DMEN)PbBr4, is fixed by “chelating effect”, a special effect of hydrogen bonding interactions. Owing to this influence, the distorted structure of α-(DMEN)PbBr4 is the most obvious compared to the other two homologue compounds, resulting in wide white-light emission, and the color rendering index (CRI) is 73. This result is comparable to the fluorescent light source and correlated color temperature (CCT) of 7863 K, and producing "cold" white light. As a whole, this work not only shows a highly efficient white light luminescent material, but also this correlation can be used as a breakthrough for PL emission characteristics in other related 2D HOIPs systems.

Papadatos and coworkers have presented a single layer LED that operates based on the concept of energy transfer effects (Papadatos et al., 2017). Compared to the traditional manufacturing strategy, the proposed work offers an instant, simple, and low cost method (Papadatos et al., 2017). The active LED layer is basically a mixture of 0D (CH3NH3)4PbI6, 2D (F-C6H4CH2CH2NH2)2PbI4 and 3D-bulk (CH3NH3)PbI3 HOIPs that transformed into a quasi-2D HOIPs for the first time. The OA spectra of this quasi-2D structure are shown in Figure 8C. During the electroluminescence behavior, it exhibited peculiar excitonic recombination bright yellow light that centered at a peak of 592 nm. In this scenario, utilizing the proper mixtures ratio to create films containing these phases allows us to demonstrate a multicolor EL emission. Several combinations can be found: yellow/green or yellow/red/deep-red emission. An LED device made under air is beneficial since it can be operated at low voltages and room temperature, whereas the active layer showed considerable film continuity under any deposition method. In addition, the high-efficiency quasi-2D structure perovskite light-emitting diodes (4.90 cd A−1) have been proved to be a mixture of a bulk methylammonium lead bromide and a 2D phenylethyl ammonium lead bromide. Hence, the excellent optoelectronic performance could be attributed to enhanced exciton confinement, excellent film uniformity, and reduced trap density (Byun et al., 2016).

In the investigation of the luminescent properties of 2D HOIPs, the pressure-induced luminescent phenomenon (PIE) has also attracted considerable attention besides the electroluminescent properties. Mao et al. (Fang et al., 2019) have systematically studied the luminescent properties of (BA)4AgBiBr8 (BA = CH3(CH2)3NH3+) under certain pressure. The relationship between the PL intensity and frequency is shown in Figure 8D. It was found that with increasing pressure, a broad emission band with FWHM of ∼180 nm suddenly appeared at 2.5 GPa. Another excellent work in this field has been reported by Zhang et al. (Zhang et al., 2019a). By exploring layered (001) Pb-Br perovskite pressure-induced broadband emission under different pressures, a new route for adjusting the optical properties of 2D HOIPs was established.

Supercapacitor

Compared to the conventional parallel-plate capacitors, the design of supercapacitors has shown a tremendous capacity output and a faster discharge rate relative to the Li-based batteries. In recent years, researchers have attempted to use 2D HOIPs since their characteristics are promising for supercapacitor applications. Figure 9A shows the device structure using 2D HOIPs as a supercapacitor material. Compared with conventional electrode materials, the 2D HOIPs own higher surface area and better ionic mobility, which make these materials promising for the candidate of electrodes in the design of supercapacitors.

Figure 9

The application and performance of 2D HOIPs in supercapacitors (A) Schematic of 2D HOIPs supercapacitors

(B) Ionic behavior of the perovskite MAPbI3 exhibited prominent interface charge density at various biases. Reproduced with permission (Zhang et al., 2015). Copyright 2015 Wiley-VCH.

(C) Hysteretic charge density (QH/A) as measured by the DWM on a CH3NH3PbI3 thin film.

(D and E) Frequency dependence of the QV hysteresis as a function of 1/f. (F) DWM measurement plotted as J-V curves. Reproduced with permission (Beilsten-Edmands et al., 2015). Copyright 2015 AIP Publishing

Through impedance spectroscopy, the ionic conductivity in MAPbI3 (MA, methylammonium) can be measured as a value of ∼10−8 S cm, indicating that the perovskite MAPbI3 was a solid electrolyte with both the transport capacity of ion carrier and electron carrier. Figure 9B shows the measured value of charge densities caused by the ions accumulation at the interfaces under different biases, and the capacitance could be calculated from the slope of the curve. Having a large capacitance (∼100 μF cm−2) and EMF (open-circuit electromotive force ∼0.5 V), CH3NH3PbI3-xBrx is regarded as a potential energy-storage material that is applicable for the field of solid-state supercapacitors (Zhang et al., 2015).

Based on Beilsten-Edmands’ research (Beilsten-Edmands et al., 2015), the ionic motion within the material can be used to determine the capacitance value of ∼1.8 mF cm−2 in a 350-nm-thick CH3NH3PbI3 material in a parallel-plate capacitor (electrostatically). Figure 9C shows the hysteretic loop of thin films at 2 Hz, and the changes of the hysteretic charge density with frequency are shown in Figures 9D and 9E. The shape of the presented loop resembles the ferroelectric material; however, the hysteresis charge density is an order of magnitude greater than that of the perovskite ferroelectric material. To determine the true cause of the QV hysteresis, pulse-based measurements were carried out to verify that the majority of the frequency depends on the current density rather than the accumulated charges (Figure 9F). This finding suggests that the current - voltage hysteresis is not driven by the ferroelectricity, which further highlights the significance of ionic migration to improve the efficiency of CH3NH3PbI3 devices.

Thermoelectricity

Thermoelectricity is another form of energy conversion to transfer waste heat into electricity. To date, various metal halide perovskites have been employed as active material in the thermoelectric fields. However, the research of HOIPs in this field is still in its primary stage. Singh et al. found that superlattices had an impact on thermoelectric properties. They study the electronic transport properties of HOIPs (MAPbI3 and MASnI3) and their superlattices by using the first-principles computations, material synthesis, and experimental measurements (Singh et al., 2017).

As shown in Figure 10A, five different models are proposed to illustrate the framework of the study. According to the Boltzmann transport theory, the Seebeck coefficient (S), electron contribution to thermal conductivity (k), and electrical conductivity per unit relaxation time (σ/τ) could be defined from their calculations. Compared to superlattices, the calculated values of S, k, and σ/τ are smaller, which can be understood because of the fact that the superlattice structures have less average energy than the simple perovskite lattices. According to Filippetti’s calculations and the ultralow lattice thermal transport measured recently (Filippetti et al., 2016), HOIPs appeared to become excellent materials for low-cost thermoelectric and solar-thermoelectric applications. For MAPI3, its power factor (σS2) is about 0.8 × 10−3 W/mK2, which is not comparable to tellurides and skutterudites (five or six times larger value than the former case) (Filippetti et al., 2016).

Figure 10

The thermoelectric performance of HOIPs

(A) Molecular representation of the different HOIPs structures. Reproduced with permission (Singh et al., 2017). Copyright 2017 Royal Society of Chemistry.

(B) Comparison of the thermoelectric properties of n-doped SC and PC MAPI3 with other thermoelectric materials with the best performance. Reproduced with permission (Filippetti et al., 2016). Copyright 2016 American Chemical Society.

(C) Seebeck coefficients of centimeter-sized CH3NH3PbI3 at different average temperatures. Reproduced with permission (Ye et al., 2017). Copyright 2017 the Royal Society of Chemistry.

(D) Seebeck coefficients (S) and (E) the predicted values of the material figure of merit (ZT) as a function of carrier concentration. Reproduced with permission (He and Galli, 2014). Copyright 2014 American Chemical Society

However, the relatively small thermal conductivity (k) renders MAPI3 potentially promising thermoelectric (Figure 10B). Ye and coworkers obtained a centimeter-sized of HOIPs CH3NH3PbI3 using an improved method of rapid growth (Filippetti et al., 2016). They studied the thermoelectric properties of the MAPbI3 single crystal. They found that the See-beck coefficient of 920 ± 91 mV K−1 basically remained unchanged from room temperature to 330 K, and it progressively increased and reached the value of 1693 ± 146 mV K−1 at 351 K as shown in Figure 10C.

In contrast, the thermal conductivities have no tendency to be observed as a function of temperature. The thermal conductivities remain between 0.30 and 0.42 W m K−1 from 0 to 127°C. To unravel the role of different metal at the B-site in 2D HOIPs on thermoelectric properties, He et al. performed ab-initio calculations to explore thermal transport characteristics of CH3NH3PbI3 and CH3NH3SnI3 perovskites through the kane single band model (He and Galli, 2014). The figure of merit (ZT) of Pb-based perovskites was greater than that of Sn-perovskite. In addition, the ZT of p-type perovskites was smaller than that of n-type for both perovskites. When the carrier concentration was ca. 1018 cm−3, the Seebeck coefficient was large (Figure 10D) and the value of ZT was in the range of one and 2 (Figure 10E). These results indicated that CH3NH3PbI3 and CH3NH3SnI3 are not only useful in the photovoltaic applications, but in the solar thermoelectric fields.

Mettan’s group attempted to increase the ZT of the 2D HOIPs by employing the photo-induced doping (M = Pb) and chemical doping (M = Sn) (Mettan et al., 2015). It turns out that the ZT of CH3NH3SnI3 is more than three orders of magnitude higher than the pristine case at room temperature. Temperature-dependent thermal conductivity of CH3NH3PbI3 and CsPbI3 is investigated in the range of 7 and 300 K and compare them to the Debye model via the Callaway method. Kovalsky et al. discovered extremely low thermal conductivity of these two materials over the entire temperature range. Moreover, the CH3NH3PbI3 thermal conductivity was lower than that of CsPbI3 (Kovalsky et al., 2017). Extended fitting analyses showed a resonant phonon scattering term (where the frequency associated is in the range of ∼15-30 cm−1) linking to the rotational degree of freedom of the organic ion. Therefore, this feature is a responsible factor that suppresses the thermal conductivity of CH3NH3PbI3 compared with CsPbI3.

Summary and future prospects

In summary, as 2D HOIPs, have a good foundation for the study of future energy applications, and have been extended to detectors, lasers, transistors, and other technical fields. This review describes the structural, electronic properties, and typical energy applications of 2D HOIPs. Figure 11 depicts the summary of 2D HOIPs for the energy realm. Although 2D HOIPs have good performance and application, the current research is still in its infancy, and there are still many difficulties to be overcome to realize high-performance 2D HOIPs devices. To better use these 2D HOIPs devices in the near future, there are some scientific questions that need to be solved systematically.

Figure 11

Summary of 2D hybrid organic-inorganic halide perovskites (2D HOIPs) for energy conversion

Stability of 2D HOIPs. Compared to their bulk phase, the stability of 2D HOIPs is relatively better. However, the most challenging part is the long-term stability, which must be resolved before putting it into the application in energy conversion. There are some ways to extend the long lifetime. First, modifying the material itself, such as doping in A or X to improve entropy contents to form a much more stable phase; Second, modifying the interface, such as increase different bonding to form a stable interfacial layer which can resist moisture, oxygen or UV; Third, modifying the distance, such as using carbon electrode to reduce the penetration into the perovskite layer. Apart from that, heterostructuring of 2D HOIPs would be essential to not only improve the stability, but also enhance physicochemical properties of the hybrid nanocomposites (Shi et al., 2020). Hence, this would be the future direction in 2D HOIPs energy applications.

Artificial intelligence for designing novel 2D HOIPs. In the research field of HOIPs, with the continuous advancement of experimental investigations, data mining effort by using artificial intelligence such as machine learning has also made substantial achievements in recent years (Im et al., 2019; Kirman et al., 2020; Lu et al., 2019). Lu et al. (2018) employed machine learning models from 212 reported HOIPs' band gap values, and then successfully screened six different types of lead-free HOIPs with appropriate electronic band gap and room temperature thermal stability from 5158 untapped potential HOIPs, two of which exhibited direct band gap and excellent environmental stability in the visible light region. Several reasonable prospects are formulated to boost up the progress of 2D HOIPs materials.

The machine learning that assists the compelling design of 2D HOIPs structures is an imperative mainstream trend for the past three years. The state-of-the-art computing technologies envisage important applications to achieve high-throughput screening represented by machine learning in the exploration of newly designed 2D HOIPs structures. At the same time, the development of the Internet of Things (IoT) has promoted the integration of material design and theoretical research. The Internet of things can connect and exchange data with other devices and systems over the Internet, which will be an opportunity for the development of mechanical learning. Combined with first-principles calculations, one can foresee that the large set of 2D HOIPs new materials with highly efficient performance can be constructed using the "bottom-up" approach in the upcoming years.

Novel synthetic approaches for air-stable 2D HOIPs. The existing preparation methods such as spin coating and CVD methods are impractical for large-scale manufacturing processes. It is even more difficult to avoid the problem arising from excessive structural defects when translating the lab-scale to industrial readiness. For this, inspiration can be found from the large-scale manufacture of other emerging 2D materials, such as the stripping and dispersion of bulk materials, taking the perspective of interface and surface effects into consideration.

Lead-free 2D HOIPs. Lead-free development has been gaining in momentum recently to seek a new candidate of hybrid material. Although the current efficiency of lead-free 2D HOIPs is relatively low, researchers have gained a deep understanding of the charge transfer mechanism in these materials through various theoretical methods such as first-principles calculations. This facilitates the construction of highly efficient materials by surface modification and forming heterojunctions with other materials.

Structure-property relationship on the 2D HOIPs system is fundamental toward the rational design and tunable manipulation of their energy-related application. There is an urgent need for systematic research to reveal the correlation between device performance and material properties such as the chemical composition, surface morphology, and crystallinity of 2D HOIPs. Considering that there are a large number of possible combinations between metal cations and organic functional groups within the hybrid materials, using high-throughput screening and synthesis to accelerate the development of new materials in the family of 2D HOIPs. This could trigger efficient solutions to the current challenges in the community of halide perovskite research.

Limitations of the study

Although the current articles on 2D HOIPs are increasing year by year, there are still some unsolved problems such as: (1) the problem of long-term stability must be solved before applying it to energy conversion, (2) the commonly used preparation methods cannot meet the requirements of industrial production At the same time, it is difficult to avoid problems caused by excessive structural defects in industrial production. (3) The efficiency of lead-free 2D HOIPs is relatively low, and there is still a lot of room for development. These issues are major challenges in the development of 2D HOIPs waiting to be crossed by researchers.