Ruddlesden–Popper Perovskites: Synthesis and Optical Properties for Optoelectronic Applications

Abstract Ruddlesden–Popper perovskites with a formula of (A′)2(A)n −1BnX3 n +1 have recently gained widespread interest as candidates for the next generation of optoelectronic devices. The variations of organic cation, metal halide, and the number of layers in the structure lead to the change of crystal structures and properties for different optoelectronic applications. Herein, the different synthetic methods for 2D perovskite crystals and thin films are summarized and compared. The optoelectronic properties and the charge transfer process in the devices are also delved, in particular, for light‐emitting diodes and solar cells.


Introduction
Recently, 3D organometallic halide perovskites with a general formula of ABX 3 (where A is a monovalent organic ammo nium cation such as MA + (CH 3 NH 2 + ) or FA + (CH(NH 2 ) 2 + ), B is a divalent cation such as Pb 2+ or Sn 2+ , and X is a monova lent halide anion) have been widely studied in lightemitting binding energy, [13] strong quantum confinement effect, [14] and superior stability to moisture [15] compared with 3D perovskites. Different selection of cationic ligands, metal halides, and the number of layers of inorganic materials will lead to the change of crystal structure and optical properties of 2D perovskite materials, realizing the bandgap tunability, [16] narrowband emission, [15b,17] and broadband emission wavelength. [11,18] Moreover, these organic ligands have a great influence on the electronic properties of inorganic layer by twisting the soft inorganic framework. [19] Therefore, it is important to under stand the relationship among the material constituents, crystal structures, and optoelectronic properties, so that one can tune the bandgap, transport performance, and charge carrier dynamics and eventually fabricate excellent optoelectronic devices.
The first 2D layered lead halide perovskite (C 9 H 19 NH 3 ) 2 PbI 4 was obtained by Dolzhenko et al. in 1986. [20] (C 9 H 19 NH 3 ) 2 PbI 4 shows the ability to intercalate with appropriate organic sol vents through weak interaction. Another 2D perovskite (C 10 H 21 NH 3 ) 2 PbI 4 was reported by Ishihara et al., [21] which is similar to a quantum well (QW) structure where inorganic layers are separated by insulating organic layers. Over the years, 2D organic-inorganic hybrid halide perovskites have been known by researchers, but they have not achieved the same attention as their 3D analogues.
While a number of reviews have been published about 2D perovskites regarding their structure and application in solar cells, [12,22] we here summarize the recent advances in the syn thesis of 2D metal halide perovskites, highlight their unique tunable bandgap, narrowband fluorescence, and white light radiation properties, discuss the charge carriers' transport, and focus more on LEDs. We hope that this review will stimulate more efforts in this field, including materials' preparation and optoelectronic device fabrications.

Design and Synthesis of 2D Perovskites
Dimensions of perovskites can be controlled by selecting different organic ligands and metal halides. The orientation of the inorganic thin layers is dependent on the geometry and noncovalent bond interaction of spatial cations, and the number of inorganic layers is directly determined by the stoi chiometric ratio. [13a] [9] Copyright 2018, Annual Reviews. b,c) Adapted with permission. [11] Copyright 2014, American Chemical Society.

Single-Crystal Growth Methods
Single crystal is the most useful state to analyze the struc ture and physical properties of materials. 2D singlecrystalline perovskites have been synthesized by a variety of solution methods. Liquidphase crystallization involves dissolving the divalent metal halide (MX 2 ) and organic amine halide (RNH 2 ·HX) at high temperatures in solvent, then mixing them together to start the crystal growth, and later cooling them to room temperature at a certain cooling rate to quench the fur ther growth; (mixed) solvent evaporation is similar to the liquidphase crystallization. The single crystals are obtained by evaporating solvent(s) at a relatively slow rate, while the solvent evaporation can be accelerated by adding another solvent; the layered solution method involves dissolving the two reactants in two solvents with certain mutual solubility, and the two solu tions have obvious density difference, so as to form a clear inter face between the two solutions due to the different solubility and density. With a slow diffusion, largesized single crystal is precipitated at the interface. These solution methods have some advantages of stability, low cost, and easy operation. The main solutionprocessed methods for synthesizing 2D organicinorganic hybrid perovskites are summarized in Table 1.
2D perovskites can be synthesized by enormous alkylammo nium cations with different lengths that mainly act as the struc tural guides to regulate the interlayer spacing among inorganic layers. [37] Leng et al. [24a] reported a temperatureprogrammed crystallization method to achieve a series of 2D perovskites (BA) 2 (CH 3 NH 3 ) n−1 Pb n I 3n+1 (n = 1, 2, 3, 4) (BA = C 4 H 9 NH 3 + ). Typically, the uniform solution including varied mass ratios of PbO, BAI, MAI, and HI (containing H 3 PO 2 ) was heated to boiling with magnetic stirring. Then, the largesized monolayer perovskites could be separated out from the solutions after cooling down from 110 °C to room temperature at a rate of 3 °C h −1 (Figure 2a-d). Atomic force microscopy (AFM) images indicate that the monolayer's thickness and the n value have a good agreement with the caxis lattice constants (Figure 2e-h) of monounit cell 2D RPPs. By this method, largersize and higherquality single crystals of 2D hybrid perovskite structures can be obtained.
In addition to single ammonium cations, diammonium cat ions are also introduced to synthesize 2D perovskites. Diammo nium cations have an advantage that complex R′ is more liable to form 2D layers and diammonium cations can eliminate van der Waals gaps and directly connect the layers together. [10b,33] 2D diammonium singlecrystal NH 3 (CH 2 ) n NH 3 PbI 4 (n = 4, 6,8) perovskites were prepared by the solvent evaporation method. [41] Although the inorganic layers are slightly distorted by the spa tial constraints imposed by the diammonium cations, these perovskites have a typical cornersharing structure. Through the crystallography in these 2D materials, the welldefined cation positions showed slower cation movement and migration than MAPbI 3 , which is capable of overcoming stability problems.
Organic layers containing functional groups have also been adopted to synthesize 2D perovskites and bring new function ality. [42] By temperaturecontrolled crystallization, when the solution cools to room temperature at certain rates, single crys tals of 2D hybrid perovskite PEA 2 PbI 4 ·(MAPbI 3 ) n−1 (n = 1, 2, 3, 4) (PEA = C 8 H 9 NH 3 + ) were prepared. [29] The thicknesses of the single crystals obtained are between 20 and 100 µm, and the spin coherence lifetime is affected by Rashba splitting and phonon scattering, both depending on the layer thickness. When n = 1, the sample has a larger recombination rate con stant due to the large exciton binding energy compared with n = 2, 3, 4 samples, which is beneficial to lightemitting applica tions. Peng et al. believe that the decrease of the level of self doping and the decrease of the crystal sizes are the result of the defectinhibiting crystallization process by introducing large organic cation PEA. [16b] The diffraction patterns of PEA 2 PbI 4 ·(MAPbI 3 ) n−1 were indexed as shown in Figure 2j. They calculated the lattice distance of the first diffraction peaks of different n values to realize that the increment is the thickness of the singlelayer PbI 6 (0.6 nm).
Kamminga et al. used four phenyl alkylammonium cations with different alkyl chains of one to four carbons to prepare singlecrystal perovskites at room temperature by a layered solution technique. [27a,43] The obtained products have good stability and can be stored in low humidity for several months without damage. It is interesting that two compounds with PMA (C 6 H 5 CH 2 NH 3 + ) and PEA cations have 2D perovskite structure where inorganic layers are linked by cornersharing PbI 6 octahedra isolated by bilayers of organic cations. [42c,44] However, with longer carbon chains, the 1D perovskites with inorganic layers consisting of cornersharing and facesharing PbI 6 octahedra are obtained. Subsequently, Ye's group syn thesized (PMA) 2 PbI 4 , (PEA) 2 PbI 4 , and (PBA) 2 PbI 4 (PBA = C 6 H 5 (CH 2 ) 4 NH 3 + ) perovskites by the same method. The source of fluorescence and the behavior of excitons were confirmed by experiments, and the quantum confinement effect caused by the structural reorganization was demonstrated by calculation. [27b] A series of phenyl and naphthylcontaining amine 2D perov skites with noncentrosymmetric structures were achieved through a simple and highyielding liquidphase crystalliza tion. [26] These 2D perovskites possess broad white fluorescence emission in the longwavelength region resulting from the inorganic layer distortion induced by the introduction of large organic cations.
In addition to lead 2D perovskites, many efforts have been applied for the synthesis of nonlead 2D perovskites. [45] In 1994, Mitzi prepared 2D Snbased perovskite (BA) 2 (MA) n−1 Sn n I 3n+1 by liquidphase crystallization in the argon atmosphere to prevent oxidation. Unlike oxide perovskites, which were syn thesized at high temperatures, these materials could be pro duced at lower temperatures. When the precursor solution containing SnI 2 , C 4 H 9 NH 3 I, and CH 3 NH 3 I was cooled down at a rate of 2-5 °C h −1 from 90 to 10 °C, the platelike products  Reproduced with permission. [25a] Copyright 2016, American Chemical Society. j) XRD patterns by lock-in coupled θ-2θ scan of the freshly cleaved single crystals of PEA 2 PbI 4 · (MAPbI 3 ) n−1 (n = 1, 2, 3) (left), and schematic illustrations of the layered structure and the corresponding orientation of 2D perovskite crystals showing that the thickness of a single perovskite sheet is ≈0.6 nm (right). Adapted with permission. [16b] Copyright 2017, American Chemical Society.
were obtained. When n = 3, the orthorhombic structure was obtained. [46] Later, Mitzi reported a layered Gebased perovskite BA 2 GeI 4 , [38] whose crystal structure and optical properties were studied. Recently, Han's group reported a 2D leadfree (PEA) 2 GeI 4 perovskite, prepared by the liquidphase crystalliza tion, which was precipitated by cooling HI and H 3 PO 2 mixed solution containing stoichiometric GeO 2 and PEAI. [47] Its direct bandgap is 2.12 eV. They also found that the introduction of PEA cation for the layered structure could actually improve the perovskite stability.

Colloidal Synthesis
Solutionprocessed methods that strongly rely on stoichiometric ratios are simple to operate, but take longer time to crystallize. Colloidal synthesis as a mature method has previously been widely used to synthesize inorganic quantum dots (QDs). [48] Schmidt et al. first synthesized MAPbBr 3 QD colloidal disper sions, which resulted in high luminescence and good disper sion due to the surface ligand capping. [49] Recently, a number of 2D perovskites have been obtained by this method.
Feldmann's group [50] realized 3D to 2D conversion of halide perovskites with varied thickness through regulating the proportion of octylamine by modifying Schmidt's method (Figure 3a). First, MABr and OABr (OA = octylamine) were obtained by adding HBr to a solution of methylamine and octylamine in ethanol, respectively. Excess acid was used to ensure that the amines were completely protonated, and a rotary evaporator was utilized to help the crystallization of ammonium salts. Then, PbBr 2 , OABr, and MABr were mixed in dimethylformamide (DMF) with desired proportions and underwent heating to form a uniform solution. Finally, under vigorous agitation, this precursor solution was dropwise added into toluene. The product was precipitated by centrifugation and was redispersed in toluene. As the ratio of OA increases, the thickness of the nanosheet gradually shrinks until reaching a monolayer.
In order to understand the effect of ligands in the formation of 2D perovskites, a series of (C 8 H 17 NH 3 ) 2 (CH 3 NH 3 ) 2 Pb 3 (I x Br 1−x ) 10 2D perovskite nanorods were pre pared. On one hand, sufficient octylamine can stabilize the perovskite surface; on the other hand, enough oleic acid can ensure the control of morphology through the molar ratio of OAI/OAc (OAc = oleic acid). [51] Weidman et al. achieved fully tunable colloidal 2D perovskite L 2 [ABX 3 ] n−1 BX 4 through dif ferent organic cation, metal, and halide components. They found that the changes of absorption and emission wavelengths were the result of the change of B or X, while A species can greatly affect the photoluminescence quantum yield (PLQY) and stability. [52] Zhang's group selected toluene as a solvent and obtained (PEA) 2 PbX 4 perovskite nanosheets. Then, they studied the effect of three solvents (chlorobenzene, chloroform, and dichloromethane) on the crystallization process. [15d] The results proved that lateral size of 2D perovskites is tunable through changing solvents. More importantly, singlelayer (PEA) 2 PbI 4 is more stable under light irradiation and ambient conditions than the conventional 3D MAPbI 3 QDs. To synthesize corru gated (EDBE)PbBr 4 halide perovskite (EDBE = 2,2′(ethylen edioxy)bis(ethylammonium)), PbBr 2 was dissolved in nonpolar hexane containing octanoic acid; later, a turbid solution formed Adv. Sci. 2019, 6,1900941 Figure 3. a) Raw materials and synthetic route. Reproduced with permission. [50] Copyright 2015, American Chemical Society. b) One-pot synthetic procedure and its reagents and products. Reproduced with permission. [53] Copyright 2016, Wiley-VCH. by injecting EDBE. The reaction solution was strongly stirred continuously for 24 h until a white colloidal solution was achieved ( Figure 3b). [53] White LEDs were then obtained from (EDBE)PbBr 4 aroused by a 365 nm UV LED chip.
Hot injection is frequently used to prepare conventional inor ganic QDs [54] and perovskite QDs. [55] Lately, Zhang et al. synthe sized 2D RPP (C 18 H 35 NH 3 ) 2 SnBr 4 through this method. [56] The product was obtained by swiftly injecting a preheated SnBr 2 -TOP solution to an ODE solution containing quantitative OAc and oleylamine ligands at 180 °C protected by N 2 gas. The reac tion continued for 10 s, and then an ice bath was used to stop it. Finally, the product was obtained by adding hexane and then centrifuging. Xray diffraction (XRD) confirms a periodic dif fraction pattern with a regular interval of 2.3° at small angles derived from the periodic 2D structure, similar to the previous reports. [15a,25a,52] This perovskite material with high fluores cence efficiency was used to make LEDs. [56]

2D Perovskite Thin Films
Appropriate thin film deposition technology is of great sig nificance for obtaining highquality optoelectronic devices. Two common methods are spin coating and chemical vapor deposition. For obtaining perovskite thin films by spin coating, organic halide AX and bivalent metal halide BX 2 (PbI 2 , PbBr 2 , or PbCl 2 ) are dissolved in organic solvents to form precursor solutions, which are then spin casted or dropped onto different matrices and annealed to form perovskite thin films. It is very important to choose the appropriate processing time and tem perature based on different precursor compositions for the needed crystallinity, phase state, and morphology of perovskite films. [57] Some important research results are presented here.
PEAI [(C 6 H 5 C 2 H 4 NH 3 ) 2 I] and PbI 2 were dissolved in DMF and then the solution was spin coated on a quartz substrate to form a (PEA) 2 PbI 4 thin film. [58] The film thickness varies from 3 to 100 nm, affected by the precursor concentration and the spincoating speed. Atomically thin uniform 2D square perov skite (BA) 2 PbBr 4 was reported by Yang's group in 2015.
[17c] A very dilute precursor solution was dropped onto a silica sub strate and heated to dry under 75 °C. When a mixed solvent of DMF and chlorobenzene was used to dissolve BABr and PbBr 2 , the obtained products were thick and randomly distributed on the substrate. When acetonitrile was introduced to form a ter nary mixed solvent, uniform square perovskite sheets grew on the substrate because of a faster evaporation (Figure 4b). AFM image shows that the thicknesses of single and double layers were 1.6 and 3.4 nm, respectively (Figure 4c,d).
Butterflyshaped (BA) 2 PbI 4 2D perovskites with different sizes and thicknesses were synthesized by Fang et al. through (blue balls for lead atoms, large orange balls for bromine atoms, red balls for nitrogen atoms, and small orange balls for carbon atoms; H atoms were omitted for clarity). b) Optical image of the 2D square sheets. Scale bar is 10 mm. c,d) AFM images and height profiles of several single/double layers with thickness of 1.6/3.4 nm (±0.2 nm). Reproduced with permission. [17c] Copyright 2015, American Association for the Advancement of Science (AAAS). e) Quasi-2D perovskite/PEO composite thin film by spin coating followed by thermal annealing. Reproduced with permission. [60] Copyright 2018, Wiley-VCH. f) Proposed crystallization process and BAI concentration as a function of distance from the substrate. Reproduced with permission. [67] Copyright 2018, American Chemical Society. g) Measured thicknesses and optical images of initial PbI 2 nanoplatelets and corresponding CH 3 NH 3 PbI 3 platelets. Reproduced with permission. [73] Copyright 2014, Wiley-VCH. growth control with temperature and mass ratio. [59] A quasi 2D perovskite (BA) 2 Cs n−1 Pb n I 3n+1 /PEO composite film (BA = benzyl ammonium, PEO = poly(ethylene oxide)) was used as lightemitting layer to assemble efficient red light LEDs ( Figure 4e). [60] A lower temperature of 70 °C was enough for the phase transition of CsPbI 3 perovskite from yellow phase to black phase, due to the confinement of inorganic layer of BA cation. More importantly, the introduction of PEO not only helps form nanoscale perovskites with smooth thin films due to its viscidity, but also promotes the charge transfer in the perovskite-PEO composite for good PLQYs because of its good ion conductivity.
It is well known that the PCE of perovskite solar cells depends heavily on the quality and morphology of thin films. Snaith's group introduced BA into 3D doublecation perov skite FA 0.83 Cs 0.7 Pb(I 0.6 Br 0.4 ) 3 . They obtained fully crystallized 2D/3D BA x (FA 0.83 Cs 0.7 ) 1−x Pb(I 0.6 Br 0.4 ) 3 perovskite films by annealing ascast precursor films in air for 80 min at 175 °C. The presence of BA not only accelerated and evolved the high crystallinity of thin films, but also induced the change of lattice parameters of the 3D perovskite phase. The heterostructures between 2D and 3D perovskite phases passivated the interfa cial grain boundary, thus inhibiting nonradiative recombina tion and achieving enhancement of performance and stability of perovskite solar cells. [61] Very recently, Zhu's group fabri cated a 2D perovskite (BA) 2 (Cs 0.02 MA 0.64 FA 0.34 ) 4 PbI 6 film with Cs + -MA + -FA + triple cations by a simple spin coating at room temperature. Compared with 2D perovskite film with a monoca tion, the 2D triplecation perovskite has smoother, denser sur face morphology, larger apparent grain size, and smaller grain boundary, leading to a longer carrier life and a higher conduc tivity. [62] Recently, Gao's group reported a simple method for high quality of RPP films by incorporating dimethyl sulfoxide (DMSO) and MACl in the precursor solution, followed by one step spin coating and solvent annealing process. During crystalli zation, the synergistic effect of DMSO and MACl led to uniform morphology, good crystallinity, and reduced energy disorder. [63] A novel hotcoating technique has been proposed to achieve highquality RPP films with favorable orientation for charge transfer eventually. [35,45d,64] In order to obtain highquality films, Tsai et al. reported that 2D perovskite (BA) 2 MA n−1 Pb n I 3n+1 single crystals were dissolved in DMF, and the solution was under con tinuous stirring for 30 min at 70 °C before the coating. Then, FTO/PEDOT:PSS substrates were preheated for 10 min from 30 to 150 °C, and the precursor solution was dropped on the hot substrate and spin coated at a speed of 5000 rpm for 20 s. [35] The core of this technology is the precise control of the temperature of the substrates. From AFM and scanning electron microscopy (SEM) observations, the films obtained by hot coating not only have larger grains, leading to a more compact and uniform film, but also have lower pinhole density, compared with films obtained by roomtemperature coating. From the synchronous diffraction data, the main growth direction of perovskite is along (101) plane parallel to the q z direction. [35] Recent studies of Snbased RPP films also suggest that the preferential orientation can be con trolled by precursor solvents through hotcoating method. [45d] Sim ilar results were found in (BA) 2 (MA) 4 Pb 5 I 16 RPP film, which also was highly oriented from DMF/DMSO mixtures by hot coating. [65] A twostep consecutive deposition [66] was presented to grow quasi2D perovskite (BA) 2 (MA) n−1 Pb n I 3n+1 , a hierarchical structure with 2D perovskite on a 3D perovskite film. [67] The growth mecha nism of this hierarchical structure is a spatially limited nucleation of the nanosheets on 3D perovskite film, due to the respective con centration of BAI and MAI and their ratio. Especially, the vertical growth of perovskite nanosheets on a thin film is closely related to the concentration gradient of BAI as shown in Figure 4f.
A profound understanding of the growth mechanism is important for regulating the orientation of materials through precursors. Since a perfect band alignment naturally exists in the materials, this special structure can facilitate electron and hole transfer, which may further promote efficient emission and photovoltaic performance. [68] The spincoating method is charac terized by simple operation, low cost, and easy to form large area, but it is not easy to select the appropriate solvent that not only dis solves hybrid perovskite crystal precursors, but also has good wet tability to the substrate. Molecular orientation degree and carrier mobility of the asprepared films are not high, and the thickness, uniformity, and surface morphology of the films are difficult to control, limiting the application range of the spincoating method.
Chemical vapor deposition is also widely applied to pre pare 2D materials, such as graphene [69] and transition metal sulfides. [70] The materials obtained with this method have the advantages of higher crystallinity and fewer defects, but the yield is often low and the performance is not very reproducible, apparently not suitable for large scales. Liu et al. found that MAPbI 3−x Cl x perovskite thin film was much more uniform if prepared by a onestep dualsource (PbCl 2 and MAI) vapor dep osition than the one obtained through the solution process. [71] In addition, aerosolassisted chemical vapor deposition has also been used to prepare perovskite films. [72] Twostep vapor deposi tion was used to equip MAPbI 3 perovskites. [73] First, with van der Waals force epitaxial growth, lead halide nanoplatelets were achieved on the muscovite mica; then, a gas-solid heterogeneous reaction was employed to convert the grown nanoplatelets to perovskites with methylammonium halide molecules. The lateral dimension was controlled from 5 to 10 µm. Figure 4g presents the relationship between CH 3 NH 3 I and PbI 2 platelets-the perov skite platelet thickness was achieved by adjusting the thickness of the relevant lead halide platelets. Similar work has been reported by Shi's group. They demonstrated that weak van der Waals force played an important role in the growth of largesized singlecrystal 2D perovskites. Ionic crystals with delocalized bonds are more likely to form ultrathin structures than covalent compounds with localized bonds. [74] By adjusting the pressure, temperature, and other conditions during the conversion process, it is expected to produce 2D mixed lead halide perovskites and realize a broad range adjustment of wavelength. In addition, there are other methods to prepare 2D perovskites, such as mechanical exfolia tion [75] and soft lithography. [76] A good understanding of the exper imental condition control on the material properties is essential to realize practical optoelectronic applications.

Diverse Properties of 2D Perovskites
The 2D RPPs are made of a series of alternately arranged inor ganic and organic layers. They have a quantum well structure: the inorganic layer "well" is composed of metal halide, and the organic cation insulating layer acts as "barrier" to isolate the inorganic layer. The inorganic layers have large quantum con finement effect due to the small dielectric shielding effect from the organic cations, which confines the charge in the inorganic layer and is more conducive to charge recombination. Further, the number of stacked inorganic layers reflects the intensity of the quantum confinement effect. Single layer shows the strongest quantum confinement effect.
The change of each component of the structure will influ ence the properties. For example, the selection of cations is an important factor affecting the lattice orientation of inorganic layers and the number of inorganic layers is related to the reac tion stoichiometric ratio. All these factors will change the phys ical and optical properties of 2D perovskites.

Excitons and Electronic Structure Properties
The exciton binding energy (E b ) and bandgap of 2D perovskites are more significantly affected by dielectric and quantum con finement effects than those of 3D perovskites. [77] In general, the dielectric constant of organic layers is much smaller than that of inorganic layers. In this case, the Coulomb interaction between the electron and the hole will be stronger because of the small shielding effect. So, the exciton binding energy of 2D perovskites is almost five times as high as that of 3D ana logues. [78] The bandgap (E g ) and exciton binding energy (E b ) of (quasi) 2D organic-inorganic halide perovskites are summa rized in Table 2.

Effect of Organic Layer
In 1990, Ishihara et al. reported the reflection spectra of (C n H 2n+1 NH 3 ) 2 PbI 4 2D perovskites with n = 4, 6, 8, 9, 10, and 12 in the region of 248-540 nm. The lattice spacing between the PbI 4 layers increases as the number of carbon chain increases from 15.17 Å for n = 4 to 24.51 Å for n = 12. The E b values of these compounds are nearly the same despite the different spacings. Among them, the E b of (C 10 H 21 NH 3 ) 2 PbI 4 is 320 meV, which is higher than that of 3D lead iodide perovskites. [78] Later, C 6 H 5 C 2 H 4 NH 3 (PhE) with a greater dielectric constant due to its aromatic ring was introduced to replace decylammonium. Through the optical absorption spectra at T = 300 and 10 K (Figure 5a), [77b] the groundstate excitons of (C 10 H 21 NH 3 ) 2 PbI 4 and (PhE) 2 PbI 4 are both measured at photon energy of 2.4 eV at room temperature. When the temperature decreases to 10 K, the exciton absorption peak suddenly became sharper; more importantly, the bandgap E g of (PhE) 2 PbI 4 is identified as 2.58 eV. By the formula E b = E g − (exciton peak energy), [78,94] E b of (PhE) 2 PbI 4 is 220 meV. As expected, a smaller exciton binding energy is obtained because of the larger dielectric con finement effects.
As described in Figure 5b, the fluorescence emission peak of (C 6 H 13 NH 3 ) 2 PbI 4 perovskite becomes sharper and more sym metrical at 127 K compared with that at 290 K. Through fem tosecond vibrational spectroscopy, it is found that excitons of (BA) 2 PbI 4 couple to phonons dominantly at 100 cm −1 , while in (C 6 [80] , 290 [81] (BA) 2 MAPb 2 I 7 1.99 [15a] , 2.17 [25a] 170-270 [80,81] (BA) 2 MA 2 Pb 3 I 10 1.85 [15a] , 2.03 [25a] 220 [80] , 130 [81] (BA) 2 MA 3 Pb 4 I 13 1.56 [15a] , 1.91 [25a] 220 [80] (BA affected the exciton-phonon coupling. [95] Moreover, strong exciton-phonon coupling may lead to a wider PL peak, which is undesirable for monochromatic LEDs. However, in some other applications, strong exciton-phonon coupling is desir able and may be beneficial for the white light emission [28] and broadband shortpulse lasers. The biexciton binding energy of (BA) 2 PbBr 4 is found to be 60 meV and that of (C 6 H 13 NH 3 ) 2 PbI 4 is 44 meV. They are relatively larger values compared to other semiconductors. This is so because the value of biexciton binding energy depends on the gap energy difference between "well" and "barrier" in the quantum well structure. [96] Recently, Sanvitto's group reported that biexciton also influenced exciton confinement and spectral response, in terms of affecting the outofplane exciton-photon interaction. [97] Different organic cations cause different bandgap energies and lead to different biexciton binding energies. Yan et al. reported that replacing MA with FA in (BA) 2 (MA) n−1 Pb n I 3n+1 perovskites not only effectively reduced the bandgap of the 2D perovskites, but also improved their ambient stability. [82] Particularly, the bandgap of (BA) 2 (FA) 2 Pb 3 I 10 film is only 1.51 eV, which is much smaller than that of (BA) 2 (MA) 2 Pb 3 I 10 (1.89 eV) [15a] and contributes to a good PCE of 6.88%. Quarti et al. proved that the electronic properties of the 2D perovskites were influenced by the length of the organic alkyl chain, and a longer chain led to an increase in the bandgap. [98] Theoretical calculation of (C 6 H 13 NH 3 ) 2 PbI 4 and (C 12 H 25 NH 3 ) 2 PbI 4 indicates that this effect is caused by the distortion of the PbI 6 octahedral structure due to the long alkyl chains. So, the size of organic cations plays an important role in adjusting the inorganic layer structure, thus leading to the regulation of the bandgap. [4b,27b,99] According to a recent report, [26] the length of the alkyl chain between the aromatic ring and the ammonium group, rather than the number of aro matic rings, is vital in the bandgap of 2D perovskites containing aromatic cations (Figure 5c). In addition to the chain length and dielectric constant of cations discussed earlier, perovskite phase transition can also affect the structure and optical proper ties. Through in situ highpressure XRD, the shift of exciton  [95] Copyright 2017, American Chemical Society. c) UV-vis absorption spectra for 2D hybrid organic-inorganic perovskite films. Reproduced with permission. [26] Copyright 2017, American Chemical Society. d) Tauc plots of (PEA) 2 Ge 1−x Sn x I 4 (x = 0, 0.125, 0.25, 0.5). Photographs of the compounds with different Sn content are shown as inset. Reproduced with permission. [102] Copyright 2018, American Chemical Society. e,f) Optical absorbance spectra of PhE-PbI 4(1−x) Br 4x and PhE-PbBr 4(1−x) Cl 4x . Reproduced with permission. [103] Copyright 2014, American Chemical Society. g) Electronic band structure of the polar configurations of selected (BA) 2 (MA) n−1 Pb n I 3n+1 perovskites. Reproduced with permission. [25a] Copyright 2016, American Chemical Society. h) Absorption of the exfoliated crystals. Reproduced with permission. [80] Copyright 2017, AAAS.
bandgap of (BA) 2 PbI 4 resulted from the change of PbI bond length and PbIPb bond angle derived from the pressure induced phase transition has been studied. [24b] The influence of structure phase transition of (C m H 2m+1 NH 3 ) 2 PbI 4 (m = 4, 8, 9, 10, and 12) perovskites on bandgap was also observed between 235 and 310 K. [78]

Effect of Inorganic Layer
Unlike conventional semiconductors, where their valence bands consist of p orbitals and conduction bands consist of s orbitals, the valence bands of 2D and 3D perovskites are mainly composed of p orbitals of halogens hybridized with the s orbitals of metals, while the conduction bands are emphati cally made of the p orbitals of metals. In lead iodide-based perovskites, the valence band is relevant to the orbitals of I 5p and Pb 6s, and Pb 6p orbitals for conduction band. [44b] There fore, both metal substitution and halogen doping can affect the bandgap of perovskites to achieve the desired properties.
2D perovskites with Sn and Ge are known to have smaller bandgaps than Pbbased perovskites. [15d,38,52,83,100] For example, the bandgap of PEA 2 SnI 4 is 2.19 eV, while that of PEA 2 PbI 4 is 2.62 eV. [101] The binding energies of excitons have also been reported to decrease from 230 to 160-190 meV for these two 2D perovskites. [101] Recently, Zeng's group obtained the bandgap of BA 2 MI 4 (M = Ge, Sn, and Pb) by theoretical calculations. The bandgaps of BA 2 GeI 4 , BA 2 SnI 4 , and BA 2 PbI 4 are 1.74, 1.45, and 1.96 eV, respectively. BA 2 GeI 4 is more affected than BA 2 SnI 4 and BA 2 PbI 4 by the distorted MI 6 octahedra, which resulted from the reduced coordination symmetry around the cations by unbonded lone pair electrons. [83] A series of mixed Ge-Sn halide-based 2D perovskites (PEA) 2 Ge 1−x Sn x I 4 were synthesized by Han's group. [102] It can be seen from Figure 5d that the bandgap reduces with the increase of Sn component. When x = 0.5, the smallest bandgap is 1.95 eV. A partial substi tution of Sn not only reduces the bandgap, but also improves the conductivity, and the improvement of moisture stability of (PEA) 2 Ge 0.5 Sn 0.5 I 4 is caused by the addition of PEA with hydrophobic groups, which is more helpful as a lightabsorbing material in solar cells.
The bandgap of 2D perovskites can also be changed by halide substitution. Replacing iodide with bromide and chloride will increase the bandgap of perovskite, because the maximum value of valence band (p orbitals) is lowered by the introduction of relatively high electronegative elem ents. [15d,16a,36,52,103] 2D perovskites with mixed halide such as (PEA) 2 PbZ 4(1−x) Y 4x , where Z and Y stand for I, Br, or Cl, have been reported. [103] From the optical absorbance spectra shown in Figure 5e,f, strong absorption peaks are observed with narrow bandwidths, at 2.4 eV (Ionly), 3.1 eV (Bronly), and 3.7 eV (Clonly), respectively, corresponding to a previous report. [2b] The absorption bands come from exciton formed by the transition from the Pb 2+ 6s orbital to the Pb 2+ 6p orbital, and the continuous regulation of bandgap is therefore real ized. Compared with the single halide perovskites, the mixed halide perovskites have inhomogeneous broader absorption peak due to the disordered distribution of halides in inorganic layers. [104] Recently, the same halide regulation of (PEA) 2 PbX 4 was reported by Zhang's group with a highest PLQY of 46.5% for (PEA) 2 PbBr 4 . [15d] The bandgap of perovskites can also be tuned through dif ferent number of inorganic layers. [25a,80,83,93,105] In the series of (BA) 2 (MA) n−1 Pb n I 3n+1 perovskites, the optical absorption band energies are 2.43 eV (n = 1), 2.17 eV (n = 2), 2.03 eV (n = 3), 1.91 eV (n = 4), and 1.50 eV (n = ∞, it actually becomes MAPbI 3 ). The bandgap decreases with the increase of n value depending on the stoichiometric ratio, which is attributed to the reduc tion of dielectric and quantum confinement effects. [25a] The corresponding fluorescence emission wavelengths also have a redshift with the increase of n. The (BA) 2 (MA) n−1 Pb n I 3n+1 2D perovskites are all semiconductors, with a clear direct bandgap shown in Figure 5g, where the valence band is mainly com posed of I 5p and a small number of Pb 6s, while the conduc tion band is composed of Pb 6p orbital. The consistent results of exfoliated crystals (BA) 2 (MA) n−1 Pb n I 3n+1 are also reported, [80] and the bandgap is in the range of 2.42 eV (n = 1) to 1.85 eV (n = 5) (Figure 5h). Importantly, the exciton binding energy decreases from 380 meV (n = 1) to an average 220 meV (n ≥ 2), which also results from the quantum confinement effects. Peng et al. reported that the bandgap of (PEA) 2 MA n−1 Pb n I 3n+1 single crystal reduced continuously from 2.4 eV (n = 1) to 2.2 eV (n = 2) and 2.0 eV (n = 3), resembling other 2D perovskites. [16b]

Narrow Emission
The recombination of free exciton is the source of narrow emis sion and small Stokes shift about (001) 2D lead halide perov skites. Stimulated by light, electrons transit from the ground state to the excited state, leaving holes in the ground state, and then the recombination of free exciton releases energy in the form of fluorescence, as shown in Figure 6a. [106] The merits of tunable color in the visible range and high PLQY of 2D perov skites are demonstrated. The luminescence with different wave lengths is realized through the regulation of metals (Pb and Sn) and halogens (Cl, Br, I) of L 2 [ABX 3 ] n−1 BX 4 (n = 1 and 2) perovskites as depicted in Figure 6b. [52] By replacing bromine with iodine of leadbased perovskites, the emission peak moves to the direction of lower energy from 3.08 to 2.41 eV for n = 1, and from 2.82 to 2.16 eV for n = 2. By replacing Pb with Sn, the emission peak shifts to much lower energy of 1.97 (n = 1) and 1.80 eV (n = 2). 2D metal halide perovskites (BA) 2 PbX 4 with high PLQY and adjusted band edge emission were reported by Dou et al., [17c] as demonstrated in Figure 6c. Zhang's group reported tunable emission of ultrathin monolayer (PEA) 2 PbX 4 2D perovskites with halogen substitution. [15d] It can be seen from Figure 6d that the emission peak of (PEA) 2 PbI 4 is located at 524.0 nm with a full width at half maximum (FWHM) of 14.7 nm. When the proportion of Br increases, the emission peak gradually shifts to blue, until (PEA) 2 PbBr 4 forms with a highest PLQY of 46.5% at 409.1 nm and a narrow FWHM of 10.6 nm. Figure 6e shows the color change of (PEA) 2 PbX 4 from violet to blue and finally green, under a 365 nm light excitation.
The PLQY of 2D perovskites is generally low due to the phase impurity and traps on the film surface from the solution process. In order to improve the luminescence efficiency, some surface modifications have been carried out. Bolink's group reported an impressive PLQY of (BA) 2 (MA) 4 Pb 5 Br 16 thin film at 515 nm exceeding 80% with a molar ratio of 3:3 between BA and MA through the introduction of an electron donor SPPO1 (Figure 6f). [107] The surface defects are effectively pas sivated to reduce the nonradiative recombination, so the radia tive efficiency is greatly improved. Another example is to coat the surface of quasi2D PEA 2 (FA) n−1 Pb n Br 3n+1 (n ≥ 2) perovskite film with trioctylphosphine oxide (TOPO). [108] According to the report, all of the PEA 2 (FA) n−1 Pb n Br 3n+1 perovskites have two emission peaks: a stronger green emission centered at about 532 nm from larger n phase and a weaker blue emission located at ≈440 nm from the n = 2 phase. Among them, the highest PLQY of 57.3% is from PEA 2 (FA) 2 Pb 3 Br 10 (n = 3), which can be greatly improved to 73.8% with TOPO passivation. Moreover, the fluorescence decay time is also extended from 0.17 to 2 µs, as shown in Figure 6g,h. [108]

Broad Emission
Different from the (001) 2D perovskites with a narrow emis sion, the corrugated (110) perovskites show a broad emission that spans the entire visible region. According to the corruga tion length, these structures can be defined as "n × n", where n stands for the number of octahedra in one unit. In the pre sent study, the most common structure is 2 × 2, as shown in Figure 1c. The corrugation lengths with "3 × 3" [109] or "4 × 4" [110] have also been achieved, although they are just rare. The first corrugated (110) perovskite with 3 × 3 structure is α(DMEN) PbBr 4 prepared by the Kanatzidis's group. [  . Colored arrows indicate absorption or PL, and the black arrow represents nonradiative relaxation. Reproduced with permission. [106] Copyright 2018, American Chemical Society. b) Solution-phase absorption (dotted lines) and PL (solid lines) spectra for n = 1 and n = 2 nanoplatelets in toluene, highlighting the changes that occur when the halide (X) changes from Cl to Br to I and when the metal changes from Pb to Sn. Reproduced with permission. [52] Copyright 2016, American Chemical Society. c) PL of different 2D hybrid perovskites and the corresponding optical PL images. Scale bars are 2 mm for (i-v) and 10 mm for (vi). Reproduced with permission. [17c] Copyright 2015, AAAS. d) PL spectra of (PEA) 2 PbX 4 NSs (X = Cl, Br, I) with different compositions. e) Photograph of solutions of (PEA) 2 PbX 4 NSs with different composition under the irradiation of a 365 nm UV lamp. Reproduced with permission. [15d] Copyright 2017, Wiley-VCH. f) PL spectra under excitation of 330 nm for the quasi-2D compound with equimolar BA:MA ratio, without and with solvent evaporation, and with solvent evaporation in the presence of SPPO1. Reproduced with permission. [107] Copyright 2017, Royal Society of Chemistry. g,h) PLQY and time-resolved photoluminescence of the (PEA) 2 FA 2 Pb 3 Br 10 perovskite films with and without TOPO passivation. Reproduced with permission. [108] Copyright 2018, Springer Nature. distorted structure results from the special "chelating effect" of hydrogen bond interactions. α[NH 3 (CH 2 ) 5 NH 3 ]SnI 4 and α(HA)SnI 4 with 4 × 4 structure are only two examples reported.
The fluorescence spectrum hardly changes by changing the morphology and crystallinity, so the surface defects are not the cause of these wide emissions. According to the emission dependence on excitation intensity diagram, the PL inten sity increases linearly with the increase of excitation inten sity, and there is no PL saturation. Both indicate that the broadband emission is not from the permanent defects of the materials. [111a] Then, this wide fluorescence emission of (110) 2D perovskites is confirmed to be from the "excitedstate defects" formed from transient lattice distortions, which are induced by the coupling of photogenerated electrons/holes with the lattice. The intrinsic selftrapping states of 2D perovskites can be explained by a model depicted in Figure 7d(A). The elec tron or hole is regarded as a hard ball. When the ball falls on the elastic sheet (soft lattice), the sheet is twisted, and then the sheet will return to its original state in the absence of the ball. This is different from the permanent defect trapping in that the distortion is already present before the ball drops onto the sheet, and the ball will sink with different indentation depths, as shown in Figure 7d(B). However, the extrinsic selftrapping is related to lattice with local heterogeneity (Figure 7d(C)). [106] Transient absorbance measurement is one of the most direct evidences for the exciton selftrapping. Under the excitation of a Adv. Sci. 2019, 6,1900941  . Reproduced with permission. [11] Copyright 2014, American Chemical Society. c) X-ray crystal structure of the (110) perovskite (EDBE)PbBr 4 , and its emission spanning the entire visible spectrum. Inset: photographs of an (EDBE)PbBr 4 crystal. Reproduced with permission. [111a] Copyright 2014, American Chemical Society. d) Self-trapping (A), trapping at permanent defects (B), and self-trapping influenced by permanent defects (C) represented by a ball interacting with a rubber sheet. Reproduced with permission. [106] Copyright 2018, American Chemical Society. e) Schematic of the adiabatic potential energy curves of the ground state (G), free-exciton state (FE), free-carrier state (FC), and various excited states (STEs) in a configuration space. The horizontal dashed line shows possible nonradiative decay processes of the STEs. Reproduced with permission. [18a] Copyright 2016, American Chemical Society. f) Normalized absorption (Abs), PL excitation (PLE, monitored at 620 nm), and PL (excited by 365 nm) spectra of (OAm) 2 SnBr 4 perovskite film. Inset: photograph of the colloidal suspension of (OAm) 2 SnBr 4 perovskites under UV light. Reproduced with permission. [56] Copyright 2019, American Chemical Society. nearUV light, (NMEDA)PbBr 4 shows a broad absorption in the range of visible spectrum, which is consistent with the forma tion of shortlived, lightinduced defect states. [18a] In addition, for (NMEDA)PbBr 4 , the wavelengthdependent PL shows that the onset time of broad emission is dependent on wavelength, and the decay time also shows the emission wavelength dependence due to the selftrapped states. [18a] On the whole, these measure ments prove the mechanism of the broad emission depicted in Figure 7e. [18a] After photon excitation, free excitons are formed in picoseconds, and then selftrapped excitons formed by lattice distortion begin to generate broad emission, and the deeper the selftrapped states, the lower the energy and the longer the PL wavelength. Zhang et al. reported that (OAm) 2 SnBr 4 2D perov skites emitted a wide orange light with a PLQY of 88%, which is the highest value among the known leadfree 2D perovskites. [56] Different from the white light emission with two peaks from the PbBr 2D perovskite, (OAm) 2 SnBr 4 has only one PL peak located at 620 nm with an FWHM of 140 nm upon 365 nm excitation (Figure 7f). The emission is only from the exciton selftrapping state, because the Sn 2+ lone pair with higher chemical activity leads to stronger excited state structure distortion and coupling of photogenerated electrons/holes with the lattice of tin halide. [112] The selftrapping reflects the bulk properties of the lattice, so the broad PL emission can be regulated by changing the crystal structure through synthesis with various organic amine cations that are typically small, highly symmetric, or flexible ditopic, based on the indepth understanding about the relationship between the selftrapping states and the crystal structure of perovskites.

Charge Carrier Transport
Solar cells and LEDs have different requirements for the charge transfer process. The charge transfer process is mainly deter mined by the interplay between carrier mobility (µ) and exciton binding energy (E b ), so they play a guiding role in the design of efficient optoelectronic devices. [113] Here, mobility refers to the velocity of charge carriers moving through conductive media under the electric field, and the binding energy of excitons is a representation of the strength of the binding force between an electron and a hole. Solar cells need fast charge separa tion, where both carrier radiative recombination and nonradia tive recombination caused by defects need to be suppressed. In general, weak exciton binding and fast carrier mobility are required. To some extent, high mobility can reduce the contact time between the carrier and the trap, thus speeding up the escape rate from the shallow trap. However, in most cases, high mobility will actually speed up the trapping, so an appropriate mobility value is very important. LEDs ask for effective charge injection and radiative recombination, so nonradiative recombi nation resulting from defects should be avoided. For lumines cent materials, strong exciton binding energy and low mobility increase the radiative recombination rate. However, in LED devices, low mobility can make charge injection unbalanced, and leads to charge accumulation, so that electrons accumulate at one side and holes at the other side of the device, resulting in lower device efficiency. A profound review about the design and construction of heterogeneous structures to improve the efficiency of charge transfer in semiconductor optoelectronic devices has been presented by Sargent's group. [113] It is indis putable that charge transfer, which plays a governing role in dif ferent optoelectronic devices, is a key to the device design and needs careful study. [114] In quasi2D perovskites with the exist ence of multiple phases, charge transfer is still controversial.
In order to achieve an efficient charge transfer, the thin film preparation has to be improved. Recently, an orderly aligned orientation of (BA) 2 (MA) 3 Pb 4 I 13 2D perovskite was achieved through cationinduced recrystallization process (CIRP). [36] Com pared with the random orientation without the CIRP treatment, the cations under design are evenly distributed, so the width of the quantum well is narrowed, which promotes the separa tion of charges and thus reduces the charge accumulation. This 2D perovskite is applied in a TiO 2 /Al 2 O 3 /NiO/C framework for solar cells, showing a fine PCE of 8.2%. The XRD of the film of (BA) 2 (MA) n−1 Pb n I 3n+1 (n = 1-4) shows that the texture of crystal changes gradually, and the proportion of [101] tex tured domains increases, as the thickness n of lead iodide layer increases (Figure 8a). Different from the lead iodide layer in the monolayer (n = 1) compound, which is preferentially arranged parallel to the substrate, it is almost completely perpendicular in the n = 4 compound. [115] The same controlled orientation of (BA) 2 (MA) n−1 Pb n I 3n+1 film is used to assemble LEDs (Figure 8b). Making inorganic layers perpendicular to the substrate, electrons and holes can be injected and transported to the deeper center of the film, rather than across the barrier of organic cations, com pared with inorganic layers that are parallel to the substrate, thus improving the probability of radiative recombination. [116] The complex multiphase distribution in quasi2D perovskites has caused a controversy of electronic energy band pattern in dif ferent optoelectronic devices. For solar cells based on 2D perov skites, the typeII band alignment is always used (Figure 8c). It is described that both the conduction band and valence band have higher energies with a smaller number of n, compared with those of larger n phase of perovskites. Thus, the electrons transfer from the smaller n to the larger n domains, while the holes move in the opposite direction. Such a separation of electrons and holes enables efficient solar cells. [15a] However, the typeI band alignment is proposed in the LEDs based on quasi2D perovskites (Figure 8d). The conduction band energy is lower when the n increases, while the valence band energy is almost the same. In this way, generated electrons and holes can be concentrated into the highn region where charge radiative recombination can be achieved. At the same time, it also effec tively inhibits the nonradiative recombination, thus achieving LEDs with high efficiencies. [15b] In the complex multiphase dis tribution of quasi2D perovskites, the distribution ratio of each phase is often affected by the synthesis process in different laboratories, which may lead to the low reproducibility on the synthesis and structure of materials. This could be the source of the controversy for different device applications. [117]

LEDs
Recently, organic-inorganic hybrid perovskites have been used in LEDs due to their highly efficient PL and wide color modulation in the visible and nearinfrared ranges. In addition to the abovementioned characteristics, (quasi) 2D perovskites show fascinating prospects in LED application with the large exciton binding energy. From the point of view of fabrication, layered 2D perovskites have good film processability with excel lent optical properties. The published (quasi) 2D perovskitebased LEDs are summarized in Table 3.
In order to achieve higher efficiency in LEDs, maximized radiative recombination is desired, while nonradiative recom bination should be suppressed, which requires the regulation of basic material properties, such as defect state density, car rier mobility, and exciton binding energy. Quasi2D perovskites have large exciton binding energy, so electrons and holes can stay in a limited region for a period longer than decay time, providing a greater possibility for the radiative recombination (the radiative recombination rate depends on the overlapping ratio of wave functions of electrons and holes). [113] It is well known that quasi2D perovskites are usually mul tiphase structures, and the phase impurity and disordered materials increase the likelihood to trap carriers, causing nonradiative recombination, thereby reducing emission effi ciency. Thus, the precise regulation of components is crucial for effective radiation. However, the defects of 3D perovskites are caused by the halide vacancies in the material, or the sur face dangling bonds made by the falling off of the surface ligands. [146] In addition, the perovskite film obtained through the solution process, whether 3D or 2D perovskites, is treated at a low temperature, meaning that there is a great possibility to form surface defects. [113] Therefore, improving the crystal linity of thin films is an important way to reduce defects. In addition to quenching temperature, the quasi2D perovskite films obtained by the solution process have a smaller crystal size than 3D perovskite films, which will increase the defect concentration on the surface and grain boundary of the films and the defects become the centers of nonradiative recombina tion, thereby reducing the emission efficiency. In addition, high surface area and porosity will also provide a greater probability of oxygen or water penetration. Recent reports have confirmed that 2D perovskites have a lower defect trap density due to the presence of ligands (organic amine cation) compared with 3D perovskites. [122] In 3D perovskite LEDs and solar cells, surface passivation and doping are extensively studied as effective ways to reduce surface defect density. However, for quasi2D perov skite, these reports are relatively few. Stability is also an impor tant indicator of device performance. The reported (quasi) 2D perovskites' electroluminescent lifetimes are listed in Table 4.
Era et al. fabricated LEDs using (PEA) 2 PbI 4 layered perov skite with a device structure of ITO/perovskite/OXD7/Mg/ Ag in 1994. [118] The strong electroluminescence (EL) peak at 520 nm was observed at liquid nitrogen temperature, and the maximum brightness of the device was up to 10 000 cd m −2 . The effective EL is attributed to the introduction of OXD7, Adv. Sci. 2019, 6, 1900941   Figure 8. a) Specular X-ray diffraction spectra of (BA) 2 (MA) n−1 Pb n I 3n+1 for n = 1-4, illustrating an increase in the fraction of crystals textured along [101] with increasing n (left). Schematic of parallel and perpendicular texturing of lead iodide sheets (blue layers) along with their respective crystallographic axes (right). Reproduced with permission. [115] Copyright 2018, American Chemical Society. b) Schematic illustration of the charge injection/recombination process in oriented film. Reproduced with permission. [116] Copyright 2018, Wiley-VCH. c) Comparative band energy diagram of (BA) 2 (MA) n−1 Pb n I 3n+1 perovskite compounds. Reproduced with permission.

www.advancedscience.com
Adv. Sci. 2019, 6,1900941 [145] which not only acts as an appropriate electron transport layer, but also acts as a barrier layer to confine the holes in the emit ting layer. This heterostructure provides an approach to achieve efficient LEDs. Sargent's group prepared LEDs with efficient multilayered quasi2D perovskite PEA 2 (MA) n−1 Pb n I 3n+1 . The device structure is ITO/TiO 2 /perovskite/F8/MoO 3 /Au, where TiO 2 and F8 [poly(9,9′dioctylfluorence)] are electron and hole injection layers, respectively, as shown in Figure 8d. The LEDs with PEA 2 (CH 3 NH 3 ) 4 Pb 5 I 16 perovskite display a high external quantum efficiency (EQE) of 8.8% in the nearinfrared region, and the maximum radiance is 80 W sr −1 m −2 when the perovskite film is 200 nm. The good performance resulted from the efficient accumulation and recombina tion of electrons and holes occurring at the lowest bandgap of the multiphase quasi2D perovskites. [15b] A series of LEDs with perovskite films that contain 1naphthylmethylamine iodide (NMAI), FAI/FABr, and PbI 2 with a molar ratio of 2:1:2 was reported by Huang's group. [121] A wide range of EL was achieved by adjusting the proportion of halogen com ponents in the precursor solution, and the highest EQE of up to 11.7% at 763 nm with a radiance of 82 W sr −1 m −2 was achieved by (NMA) 2 Pb 2 I 6 Br (NFPI 6 Br) perovskite. Good device performance resulted from the com plete surface coverage of the film, which reduces defects and leakage current, thereby suppressing nonradiative recombination. More importantly, the LEDs show improved lifetime due to the device's high efficiency and the perovskite film's superior sta bility. As shown in Figure 9a, after 2 h continuous working at a current density of 10 mA cm −2 , the EQE only decreases to half of its initial value. In order to reduce the efficiency roll off at high current density, Huang's group tuned the QW width by increasing the proportion of FA cation (the molar ratio of NMAI, FAI, and PbI 2 changed from 2:1:2 to 2:1.9:2 in the pre cursors). The formation of wider QWs is proved by a 5.6 nm redshift of the PL peak. The wider QWs also suppress the lumi nescence quenching, so the EQE of LEDs is further improved to 12.7%. Additionally, the efficiency rolloff is greatly reduced, Adv. Sci. 2019, 6,1900941 Figure 9. a) Stability data for a NFPI 7 EL device tested at a constant current density of 10 mA cm −2 . Reproduced with permission. [121] Copyright 2016, Springer Nature. b) EQE versus current density. For the 2:1.9:2 multiple quantum well LEDs, a peak EQE of 12.7% is achieved at a current density of 80 mA cm −2 . The EQE of the 2:1.9:2 device remains ≈10% at 500 mA cm −2 due to a significantly suppressed EQE roll-off. Reproduced with permission. [132] Copyright 2018, Springer Nature. c) Schematic of LED device structure. d) Electroluminescence spectral stability under 3.5 V continuous voltage operation; insets: photographs of devices at 4 V. Reproduced with permission. [60] Copyright 2018, Wiley-VCH. e) Typical EL spectra of (PEA) 2 FA 2 Pb 3 Br 10 -based LEDs under different voltage biases. Inset shows the electroluminescence image of LEDs. Reproduced with permission. [108] Copyright 2018, Springer Nature. f) J-V-L data and current efficiency of devices based on CsPbBr 3 perovskite films with the introduction of 0% PEABr, 40% PEABr, and 40% PEABr-crown. Reproduced with permission. [134] Copyright 2018, Springer Nature. g) Normalized EL spectra of the CsPb(Br/Y) 3 RPP devices at the turn-on voltage; inset: photographs of the blue light LEDs. Reproduced with permission. [143] Copyright 2018, American Chemical Society. h) Development trend of EQE of 2D and 3D organic-inorganic hybrid perovskite LEDs. and the efficiency is still maintained at about 10% under a current density of 500 mA cm −2 (Figure 9b). The device has a highest radiance of 254 W sr −1 m −2 in solutionprocessed near infrared LEDs. [132] The highest EQE of 20.1% of 2D perovskite LEDs to date in the nearinfrared range was reported by Friend's group, [136] which is based on (NMA) 2 (FA)Pb 2 I 7 2D perovskite and polyHEMA (HEMA = 2hydroxyethyl methacrylate). This excellent EQE results from the ultrafast migration of excitons, which takes only ≈1 ps. It makes nonradiative recombination uncompetitive in dynamics and thus greatly suppresses bulk and interfacial nonradiative recombination. Ma's group reported a series of efficient red light LEDs with emission peaks of 638, 664, and 680 nm based on quasi2D perovskite (BA) 2 Cs n−1 Pb n I 3n+1 /PEO composite with a device structure of ITO/PEDOT:PSS/polyTPD/perovskite/TPBi/LiF/ Al (Figure 9c). The LEDs have a highest EQE of 6.23% and a brightness of 1293 cd m −2 at 680 nm emission peak, and show exceptional EL spectral stability under continuous operation (Figure 9d). [60] Green light 2D perovskite LEDs based on (PEA) 2 (FA) n−1 Pb n Br 3n+1 with EQE of 14.36% were reported by Yang et al. [108] The reason for such a high EQE is the increased PL efficiency of the film due to less surface defect states caused by TOPO passivation; thus, nonradiative recombination at the surface and grain boundaries is reduced, as shown in Figure 6g. The EL spectra of (PEA) 2 (FA) 2 Pb 3 Br 10 based LEDs under different operating voltages are shown in Figure 9e. Dif ferent from the PL peak, there is only one single green EL peak located at 532 nm, while the blue light peak is not observed. The reason for this phenomenon is that the driving force for PL is the energy difference only, while the driving force of EL is the combination of energy difference and applied electric field, so most charges are injected into the smallest bandgap region and then recombine.
In order to achieve more efficient green light LEDs, the films are made of 2D organic-inorganic hybrid perovskite nanosheets and CsPbBr 3 nanocrystals so as to provide an effective energy channel for the injection of excitons into the radiative recom bination centers. However, there are still some problems with thin films, including crystallite distribution of CsPbBr 3 nanocrystals and phase separation between the organic and inorganic phases. [134,141] Ban et al. demonstrated that the intro duction of a crown molecule accurately controlled phase sepa ration and improved film quality. Compared with CsPbBr 3 LEDs, the leakage current of CsPbBr 3 with 40% PEABr is lower and the turnon voltage of CsPbBr 3 with 40% PEABrcrown is further decreased. In addition, the current density and bright ness are greatly improved (Figure 9f). So, the final EQE of these LEDs reaches 15.5% at 510 nm. [134] Compared with the efficient nearinfrared, red, and green light LEDs, blue LEDs based on perovskites still have inferior performance. The ways to achieve blue light emission of LEDs are composition engineering and dimensional engineering. In 3D perovskites, Br is replaced by Cl to widen the perovskite bandgap to achieve blue light emission. In dimensional engi neering, a reduced dimension enhances the quantum confine ment effect, so the PL peak is shifted to blue light.
Recently, Cao's group reported an EQE of 5.7% for quasi 2D perovskite LEDs with 480 nm blue emission. [145] The introduction of PEABr into 3D perovskite CsPbCl 0.9 Br 2.1 effectively passivates the surface trap of the film, where the trap density of perovskite film dramatically decreased from ≈4.1 × 10 17 to 3.0 × 10 16 cm −3 with the increase of PEABr ratio from 0 to 100%, and the PLQY increases from 0.15% to 27%. It can be seen that the effective inhibition of nonradiative combi nation is crucial to the PL efficiency of perovskites.
2D/3D mixed halide perovskites BA 2 Cs n−1 Pb n (Br/X) 3n+1 (X = Cl, I) were reported to make LEDs with tunable color across the whole visible spectrum (Figure 9g). [143] It is worth mentioning that the highest EQE for blue light at 486 nm is up to 6.2% with a luminance of 3340 cd m −2 at 8 V and the EQE reaches 10.1% at 506 nm. The first article on leadfree 2D perovskite (PEA) 2 SnI x Br 4−x -based LEDs with a structure of ITO/PEDOT:PSS/EM/F8/LiF/Al was reported by Lanzetta et al. in 2017. [45b] Although the EQE of these LEDs is very low and the luminance is only 0.15 cd m −2 , it indicates a possible way to fabricate LEDs from lowdimensional leadfree perovskites. Zhang's group recently reported improved 2D Snbased perov skite LEDs with an EQE of 0.1% and a maximum luminance of 350 cd m −2 , which is the highest brightness of leadfree perov skite LEDs to date and opens up their promising display appli cation potentials. [56] We summarized the annually reported highest efficiency of 2D and 3D organic-inorganic hybrid perovskite-based LEDs in recent years (Figure 9h). The EQE of 2D perovskite-based LEDs has a rapid development from 9.6% to 20.1% for green and nearinfrared emissions in just 3 years, and the current efficiency already approaches the level of 3D organic-inorganic hybrid perovskite-based LEDs. Apparently, 2D perovskites have a good prospect in LEDs.

Solar Cells
Today, 3D perovskites as light absorption layer for solar cells have reached a very good PCE as high as 24.2%. [7b] However, their sensitivity to the environment, especially moisture, is a major barrier to commercialization. Considerable efforts have been made to improve their stability. [147] Compared to 3D perovskites, 2D perovskites have larger exciton binding energy and better stability in the ambient environment. However, 2D layered perovskites also bring some bad characteristics. First, the existence of longchain organic amine cation insulating layer and the unsatisfactory orientation of the inorganic layer structure will cause charge transfer problems, including charge accumulation and more charge recombination, so that the elec trons and holes cannot be well separated. [22b,148] Second, as the number of layers decreases, the bandgap gradually widens, so the absorption of light is not ideal, thus resulting in a decrease in efficiency. Therefore, it is very important to achieve a balance between efficiency and stability by adjusting orientation and number of layers.
The first 2D layered perovskite solar cells based on (PEA) 2 (MA) 2 Pb 3 I 10 were reported to have a PCE of 4.7%. [39] Compared to MAPbI 3 , the 2D perovskite is more resistant to moisture, and due to the wider bandgap, the 2D structure is also suitable for the higherbandgap absorber in the dual absorber devices. Moreover, in terms of material optimization, 2D perovskite structure presents greater tunability at the molecular level. To date, a large number of 2D perovskite absorbers have been synthesized with significantly improved eff iciencies. [35,64a,149] Sargent's group reported a PCE of 15.3% for quasiperovskite PEA 2 (CH 3 NH 3 ) n−1 Pb n I 3n+1 (n = 60). It shows an excellent stability with the efficiency remaining at about 13% after 2 weeks in a humid environment, while the efficiency of 3D MAPbI 3 perovskite decreases from 16.6% to 4.3% in 3 days. [149d] (BA) 2 (MA) 2 PbI 3 based solar cell with PCE of 4.02% was obtained by Cao et al. [15a] Although the introduction of BA organic cation promotes resistance to moisture more than the 3D counterparts, it also causes outofplane charge transport inhibition. Notably, Tsai et al. overcame this disadvantage and achieved a vertical orientation of perovskite layers to the sub strate by means of a hotcasting deposition method. From the synchronous diffraction data, the main growth direction of perovskite is along (101) plane parallel to the q z direction. This unique orientation enables the photogenerated electrons and holes to move along the inorganic layer to the device's elec trodes, respectively, avoiding the inhibition of organic layers. Such an efficient charge transport results in a PCE of 12.5% for (BA) 2 (MA) 3 Pb 4 I 13 absorber solar cells. [35] The introduction of 2D RPPs into 3D perovskites has been proved to guarantee a high efficiency and improved stability of the solar cells. [149d] For example, Liu's group reported a high PCE of 20.62% for 2D/3D heterostructure solar cells. The devices demonstrated significant longterm ambient sta bility and worked for more than 2880 h when the efficiency dropped to 80% of the initial value without encapsulation. [150] The introduction of BA changes the crystallization kinetics and controls the morphology of the film, resulting in larger particle size and improved film quality. The highest PCE (>22%) for 2D/3D perovskite solar cell was achieved by Grät zel's group, and the solar cells showed remarkable stability with 90% efficiency of 1000 h in moist air under simulated sunlight. The excellent performance comes from the for mation of ultrathin, ultrahydrophobic, and highly uniform 2D (FEA) 2 PbI 4 (FEA = phenylethylammonium) perovskite film casted on the 3D perovskite layer. The incorporation of (FEA) 2 PbI 4 not only protects FAPbI 3 film from the influ ence of moisture due to its hydrophobicity of fluoroarene, but also promotes the hole transfer from perovskite layer to spiroOMeTAD. [151] Although 2D perovskite solar cells are at their startup stage, and the dielectric and quantum confinement effect plus carrier transport limit the PCE, they present excellent environmental stability far beyond 3D perovskites because of their unique lay ered structure. So, 2D perovskites lay the foundation for 2D/3D hybrid optoelectronic devices and will have a great potential for the solar cell commercialization.

Summary and Outlook
2D Ruddlesden-Popper perovskites have received wide spread attention as promising materials for optoelectronic devices especially in recent years, due to their unique prop erties of large exciton binding energy, strong quantum confinement effect, and stability. Here, we reviewed the stateoftheart 2D perovskites, with their synthesis methods for powders and thin films, including singlecrystal growth, colloidal synthesis, spin coating, and vaporphase deposition, and analyzed the possible growth kinetics, various properties in optoelectronic devices, and applications in LEDs and solar cells.
Quasi2D perovskites have larger exciton binding energy, which is more conducive to radiative recombination. The EQE of LEDs based on quasi2D perovskites has reached 20.1% in nearinfrared emission, 15.5% in green light emission, and 6.2% in blue light emission. In order to achieve higher EQE and more stable LEDs, a few immediate issues need to be addressed. 1) The phase impurity and lowtemperature solution method for quasi2D perovskites often increase the defect density. So, the precise regulation of components and improvement of film quality are needed. 2) The poor trans portation of charges in the organic amine spacing layers and the charge trapping in a surface trap limit the charge injection and reduce EQE. 3) The equilibrium between mobility and exciton binding energy should be further optimized for effec tive LEDs. Therefore, the design of heterostructures and effi cient charge transfer channels are worth studying for quasi 2D perovskites.
The solar cells based on 2D layered perovskites have demon strated excellent PCE and superior stability. The highest PCE of 2D/3D perovskites has been over 22% and it can be main tained in humid air for more than 1000 h under simulated sun light, while PCE has fallen by only 10%. It offers a route toward efficient and stable perovskite solar cells. However, there are still some problems to be solved before commercialization. 1) More efforts need to be made to fully understand their crystal growth mechanism and to further improve the quality and morphology of the films. 2) Similarly, the defects caused by the phase impurity of 2D perovskites will capture the charges and lead to nonradiative recombination, thus inhibiting the charge extraction. So, strictly following the stoichiometric ratio of the reaction to precisely control the value of n is desired. 3) In order to improve the charge transfer process in devices, thin films with outofplane orientation are desired and the technique of preparing controllable vertically oriented thin films needs to be improved. 4) Leadfree perovskites have not yet achieved com petitive device efficiencies, and their stability also needs to be improved.