Recent advances on two-dimensional metal halide perovskite x-ray detectors

In recent years, two-dimensional metal halide perovskites (MHPs) have attracted increased attention for radiation detection and imaging. Their detection efficiencies are almost comparable to three-dimensional (3D) perovskites. Meanwhile, they demonstrate superior stability to 3D perovskites. The pursuit of high-quality, phase-pure and lead-free two-dimensional MHP materials and large-area fabrication capability for x-ray detectors are among the research hotspots. In this review, we first give a brief introduction of the crystallographic structure, optoelectronic characteristics and preparation methods of high-quality two-dimensional perovskites. In addition, we overview the general working principles of direct and indirect x-ray detection processes and the corresponding performance metrics from the perspective of detection and imaging. Furthermore, we provide a comprehensive discussion on the recent advances in 2D perovskite x-ray detectors and imaging devices. Finally, we pinpoint several major obstacles of 2D x-ray detectors that should be overcome in the near future.

7 Mingyue Han and Yingrui Xiao contributed equally to this work. * Authors to whom any correspondence should be addressed.
Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.

Introduction
X-rays are high-energy photons (100 eV < hν < 100 keV) with superior penetration capability and have been broadly used in medical diagnostics, food safety, nondestructive probing and scientific research, to name a few [1][2][3][4]. Due to the different attenuation coefficients of different materials, x-ray radiation is attenuated proportionally after penetrating an object, which is recorded by x-ray detectors, thus leaving corresponding images with high contrast. State-of-the-art xray detection process can be divided into two types, i.e. direct detection and indirect detection (also called a scintillator type). In the case of direct detection, electron-hole pairs are generated in a semiconductor detector under x-ray radiation. Subsequently, with the assistance of a bias, they are separated and transported to the electrodes and recorded by an electrical circuit. At present, commercial direct x-ray detectors such as amorphous Se (α-Se) [5] suffer from poor x-ray absorption coefficient (α), insufficient charge transport characteristics and low sensitivity. On the other hand, indirect detection uses a scintillator to convert high-energy x-ray photons into low-energy photons, which are then detected by arrays of photodetectors (such as amorphous silicon photodiodes or photomultipliers) for imaging. Commercial scintillation-type detectors such as CsI: Tl [6], CdWO 4 [7] and Bi 4 Ge 3 O 12 [8] are limited by relatively low x-ray stopping power, small light yield, long response time and low spatial resolution.
Metal halide perovskites (MHPs) have manifested great potential in x-ray detection because of strong x-ray stopping power [9] and tunable bandgaps (from 1.6 eV to 3.0 eV) for absorption and subsequent carrier creation. Besides, a low defect density and high carrier mobility result in a large µτ product (µ is the carrier mobility, τ is the carrier lifetime) of 10 −4 -10 −2 cm 2 V −1 and long carrier diffusion length [10], offering MHPs excellent charge transport characteristics [11]. Generally, the structural formula of three-dimensional (3D) metal halide perovskite is ABX 3 , where A is Cs + , CH 3 NH 3 + (MA + ), [HC(NH 2 ) 2 ] + (FA + ), B is Pb 2+ , Ge 2+ , Sn 2+ and X is Cl − , Br − , I − [12]. The application of MHPs in x-ray detectors was first reported by Mercouri G. Kanatzidis's group. CsPbBr 3 's auspicious photoelectric properties endow itself with a hole corresponding µτ value ten times larger than Cd 0.9 Zn 0.1 Te [13]. This opens up an avenue for employing MHPs as a new generation of high-energy radiation detectors. Motivated by this pioneering study, 3D MHPs with different compositions and device structures were investigated. Astonishing detection performance with a high sensitivity of 5.2 × 10 6 µC Gy air −1 cm −2 and a low detection limit of 0.1 nGy air s −1 has been obtained in MAPbI 3 single-crystal detector [14]. However, some intrinsic drawbacks of 3D perovskites, which were revealed and documented by the perovskite photovoltaics community, have also become evident for x-ray detection. Major concerns include their structural instability and severe ion migration upon exposure to moisture, light and heat [15]. These drawbacks deteriorate the performance and stability of MHPs x-ray detectors and hinder their commercialization.
As an alternative, 2D perovskites with a general formula of (A ′ ) m (A) n-1 B n X 3n+1 , where A' is monovalent (m = 2) or divalent cations (m = 1), exhibiting superb structural stability and suppressed ion migration, which may hold the key to achieve longer operational stability of the x-ray detectors under a higher working bias. In addition, 2D perovskites feature stronger self-trapping than 3D perovskites and unique anisotropic charge transport characteristics. In recent years, great breakthroughs in 2D perovskite optoelectronic devices were reported such as solar cells [16][17][18], light-emitting diodes (LEDs) [19,20], photodetectors [21,22] and x-ray detectors [23,24], etc. In this review, we introduce the crystallographic structure, optoelectronic characteristics of 2D perovskites and preparation methods to obtain high-quality 2D perovskites in the form of bulk materials, membranes and nanocrystals. In addition, we overview the general working principles of direct and indirect x-ray detection processes and the corresponding performance metrics from the perspective of detection and imaging. Furthermore, we give a comprehensive discussion on the recent advances in 2D perovskite x-ray detectors and imaging devices. Finally, we pinpoint several major obstacles of 2D x-ray detectors that should be overcome in the near future.

Crystallographic structures and optoelectronic characteristics
2D perovskites i.e. (A ′ ) m (A) n−1 B n X 3n+1 can be obtained through the dimensional reduction of a 3D crystal lattice, where A ′ represents monovalent (m = 2) or divalent (m = 1) long-chain organic cations and the A, B, X ions are the same as 3D perovskites [25,26]. Based on the octahedral connectivity mode, 2D perovskites can be divided into (100)-oriented, (110) -oriented, and (111) -oriented ones. So far, (100)-oriented 2D perovskites have been intensively studied for x-ray detectors. According to the molecular structure of A ′ , they can be categorized into three types, i.e. the Ruddlesden−Popper (RP) phase [27,28], the Dion−Jacobson (DJ) phase [29,30] and alternating cation in the interlayer space (ACI) phase [31,32]. To give a clear view of crystal structures, the RP, DJ, and ACI phase homologous series with n = 3 are shown in figure 1. For the RP phases like (PEA) 2 (MA) 2 Pb 3 I 10 (PEA + = C 6 H 5 C 2 H 4 NH 3 + ) or (BA) 2 (MA) 2 Pb 3 I 10 (BA + = CH 3 (CH 2 ) 3 NH 3 + ), the layers are offset by one octahedral unit, showing a (1/2,1/2) in-plane displacement. The bilayer spacer organic cations of the RP phase cause large interlayer distances, which inevitably increase the layer separation from 1.5 times the cation length when interleaved (as in BA) to more than twice without overlap (as in PEA). The layers in the DJ phases such as (3AMP)(MA) 2 Pb 3 I 10 align perfectly with (0,0) displacement. The stacking mode of the ACI layer combines the characteristics of the RP phase and DJ phase, showing a (1/2,0) displacement [33].
2D perovskites can also be classified by the n value. As shown in figure 2(a), 2D perovskite has a multilayer structure in which a BX 6 octahedral inorganic layer is sandwiched between two organic cation layers. The value n represents the number of the BX 6 octahedral inorganic layers. A 2D perovskite is considered a pure 2D perovskite when n = 1 and a quasi-2D perovskite when 2 ⩽ n < ∞ [34]. When n = ∞, it becomes a 3D perovskite. The optical properties of 2D perovskites change with the number of BX 6 octahedral inorganic layers. For example, the color of RP-type perovskite crystals (i.e. (CH 3 (CH 2 ) 3 NH 3 ) 2 (CH 3 NH 3 ) n-1 Pb n I 3n+1 ) changes from red to dark black with the increase of n value [35,36] (figure 2(b)), accompanied by the red-shift of the PL emission peaks (figure 2(c)). This means that the optical bandgap of the 2D perovskite becomes larger with the decrease of BX 6 octahedral inorganic layers (figure 2(d)) [37]. The change of bandgap is attributed to the formation of a 2D multiple-quantum-well (MQW) structure where the semiconducting inorganic layers act as a potential 'well' and the insulating large organic layers act as a potential 'barrier' (figure 2(e)). Such inherent nature of the MQW structure endows 2D perovskites with a larger exciton binding energy compared to 3D counterparts, which enhances the exciton radiative recombination efficiency and anisotropic transport of charge carriers [38,39].
In addition to the intriguing optical properties, 2D perovskites also offer the promising advantage of inhibiting ion migration, which prevents device failure due to ion diffusioninduced corrosion of the perovskite layer and the metal electrode. Ion migration such as iodide vacancies (V I ) and methylammonium vacancy (V MA ) in 2D perovskites is suppressed due to the increase of V MA and V I formation energies, resulting in a lower density of point defects (figure 2(f)). In addition, the unique structure of 2D perovskite may also contribute to inhibiting ion migration. The suppressed ion migration of 2D perovskites can also be confirmed by conductivity measurement. Huang et al [40] studied the in-plane conductivity of the quasi-2D (BA) 2 (MA) 2 Pb 3 I 10 (n = 3) single crystals (SCs) at varied temperatures. Compared to 3D perovskites, which show a clear electronic to ionic conductivity transition at 280 K both in the dark and under illumination, the quasi-2D (BA) 2 (MA) 2 Pb 3 I 10 perovskite samples are absent of such conductivity transition up to 350 K. Moreover, 2D perovskites are endowed with superior humidity stability [41][42][43], structural stability [44][45][46], and thermal stability [47][48][49] due to the hydrophobic nature of organic cations and strong van der Waals forces between the organic layers.

Materials and preparation methods
Preparing high-quality 2D perovskite materials is essential for the fabrication of radiation detectors. In this chapter, various methods for preparing 2D perovskites in the form of bulk SCs, polycrystalline pellets, thin films and nanocrystals are introduced.

2D perovskite bulk materials
X-rays have a strong penetration capability. To this end, 2D perovskite bulk materials such as SCs and polycrystalline pellets with a sufficient thickness are favorable to completely absorb and utilize x-ray photon energy. Solution crystallization is mostly used to obtain 2D perovskite SCs with low trap density [50], long carrier lifetime [51] and high carrier mobility [52]. These methods include the temperature-cooling solution crystallization [23,53], controlled evaporation [54,55], antisolvent vapor-assisted crystallization (AVC) [56,57], surface tension-controlled crystallization [58], and the spaceconfined method [40]. 2D perovskite bulk materials can be obtained by gradually decreasing the temperature of an HXbased (X = Cl, Br, I) perovskite precursor solution from a high temperature (>90 • C) to room temperature. Shen et al [23] prepared (BDA)PbI 4 (BDA = NH 3 C 4 H 8 NH 3 ) SCs by mixing (BDA)I 2 and PbO in concentrated aqueous HI at 90 • C, followed by a gradual cooling to room temperature (25 • C) at a rate of 1 • C h −1 . Centimeter-sized crystals without obvious twinning defects were obtained at a slower ramp rate of temperature. Xiao et al [53] employed two kinds of organic cation (linear butylamine (BA) and branched iso-butylamine (i-BA)) to tailor CsPbBr 3 SCs via the temperature-cooling solution crystallization method. As illustrated in figure 3(a), the temperature of the solution was first kept 5-10 K above the oversaturated temperature for several hours (State I). Nuclei started to generate at the solution surface when decreasing temperature below the oversaturated temperature to overcome the nucleation barrier (State II). The as-prepared seeds grew larger with further decreasing temperature (State III). Meanwhile, some new seeds appeared at the bottom of the solution (State IV). Finally, SCs of two shapes (i.e. long rod-like shape and block shape) were obtained ( figure 3(a)). The cooling rate should be low enough to ensure a suitable growth rate and to obtain high-quality transparent bulk SCs. Liu's group [58] successfully synthesized high-quality 36 mm sized 2D (PEA) 2 PbI 4 perovskite SCs via a surface tension-controlled crystallization method. The solubility of (PEA) 2 PbI 4 in γ-butyrolactone (GBL) decreased with cooling temperature and small crystallites grew at the solution surface due to higher nucleation probability [59,60]. Since the net force of the buoyant force (F b ) and the surface tension (F s ) of the precipitated crystal nuclei on the solution surface was greater than the gravitational force (F g ), the small single crystal could remain floating on the solution surface for a while. As the temperature dropped, the crystals grew bigger until the weight exceeded buoyancy and sunk to the bottom of the solution. Liu et al [61] found that the nucleation of 2D perovskite occurred at the solution-air interface when growing on a hot plate due to a temperature gradient of the precursor solution from bottom to top. The grown crystals suffered from unregular shapes and poor crystallinity. By using a heating element with double heating sources, they avoided the temperature gradient during nucleation and growth and prepared mixed-cation BA 2 FA x MA 1−x Pb 2 I 7 SCs with well-controlled shapes and improved crystal quality. Yang's group [55] grew 2D layered lead-free perovskite, i.e. (NH 4 ) 3 Bi 2 I 9 using a facile low-temperature solution method (figure 3(b)). The precursor was dissolved in hydroiodic acid and stirred at room temperature overnight. The obtained solution was evaporated and concentrated at 110 • C on a heat plate followed by quickly transferred to a heating oven for crystal growth at 60 • C. The largest single crystal with a size of 21 × 20 × 7 mm 3 was achieved in 5 d. It should be noted that massive nucleation is necessary to be avoided. In addition to changing the temperature, saturated solutions can be obtained by solvent engineering. Ren's group [56] achieved millimeter-sized 2D (PEA) 2 PbBr 4 perovskite SCs at room temperature by using a modified AVC method. They dissolved phenethylammonium bromide (PEABr) and PbBr 2 in a solvent with high solubility and moderate coordination ability (i.e. N, N ′ -dimethylformamide, DMF) and used chlorobenzene as an insoluble antisolvent. Millimeter-sized 2D (PEA) 2 PbBr 4 perovskite SCs were grown by controlling the diffusion rate of the antisolvent vaporization ( figure 3(c)). An optimal vaporization rate of the antisolvent is critical for the high crystal quality, while too slow a vaporization rate can prolong the growth time. Thin SCs with tuned thickness are directly obtained by space-confined method. Huang's group [40] synthesized quasi-2D perovskite SCs BA 2 MA 2 Pb 3 I 10 by cooling-induced supersaturation in HI solution to suppress the in-plane ion migration.
Although considerable progress has been made in the solution-prepared 2D perovskite SCs, challenges remain in the synthesis of homologous 2D perovskites with high quality over a large scale at a fast speed. To overcome this obstacle, researchers have paid much attention to the development of 2D perovskite tablet pellets. Yang's group [62] fabricated an oriented 2D perovskite pellet by tableting 2D crystal powder at 280 MPa (figure 3(d)). A dark red 2D (F-PEA) 3 BiI 6 perovskite pellet exhibited a smooth and shiny surface. Depending on the mold used, a pellet with a diameter of 13 mm and an area of 1.33 cm 2 was obtained.

2D perovskite thin films
Compared to bulk materials, it is easier to fabricate 2D perovskite thin films over a large area. In view of the crystal structure characteristics, 2D perovskite thin films can be divided into polycrystalline and single crystalline. Gu's group [24] prepared a polycrystalline (PEA) 2 SnBr 4 −polymethyl methacrylate (PMMA) composite film via a spin-coating method (figure 4(a)). In short, the 2D (PEA) 2 SnBr 4 perovskites were ultrasonically treated in an ice bath to reduce the grain size and dissolved in a mixed solution of PMMA/trichloromethane. Then, a dynamic spin-coating method was used, followed by a drying process to achieve a 100 µm-thick crack-free composite film with a size of 100 × 100 mm.
However, polycrystalline thin films usually have high defect density and low carrier diffusion length, leading to unsatisfactory x-ray detection performance. To this end, a series of methods including induced peripheral crystallization (IPC) [63,64], that is, nucleation and crystallization preferentially grew on the edge of the glass with small lateral area on the upper surface by maintaining the temperature. Anti-solvent vapor-assisted capping crystallization (AVCC) [65] and surface tension-controlled crystallization [60] were developed to pursue high-quality, flexible and large-area 2D single-crystal thin films. The IPC growth method is illustrated in figure 4(b). The hot perovskite precursor solution was dripped on a preheated glass substrate. A smaller glass slide was covered on top of the solution to form a sandwich structure, which was maintained at a high temperature for solvent evaporation. Nucleation occurred at the edges of the top glass slide after a period. After that, the temperature was decreased to induce crystallization. Using this method, Lédée et al grew a large (PEA) 2 PbI 4 single-crystal thin films (area: 73 × 35 mm 2 ) with a controllable thickness of approximately 0.6 µm. Deleporte's group [65] developed the AVCC method by combining the AVC process and the cast-capping crystallization process to synthesize a high aspect ratio (PEA) 2 PbI 4 single-crystal thin films. The (PEA) 2 PbI 4 crystals grown on a clean quartz substrate with the AVC method (figure 4(c)) were covered by another sheet and pressed with a vial of dichloromethane. In a few minutes, regular crystals appeared between the two substrates. Over time, its lateral dimensions could reach 2000 × 700 µm 2 while the thickness can be controlled to 5 µm, which is an order of magnitude higher than the AVC method ( figure 4(d)). Priya's group [60] proposed the surface tension-controlled crystallization method to quickly synthesize large-area single-crystal thin films at the water−air interface. The nucleation mechanism of (BA) n (MA) n−1 Pb n I 3n+1 perovskite at the water−air interface is shown in figure 4(e). The molecules at the water−air interface encountered not only the intermolecular forces but also tensile elastic stress. This increased the energy of precursor molecules and lowered the nucleation barrier, leading to a greatly enhanced probability of nucleation at the water−air interface. The preparation process was relatively simple. In short, the precursor solutions with different solute mole ratios were prepared in small vials at a high temperature, followed by slowly cooling to obtain 2D perovskites of different n values (figures 4(f) and (g)).

2D perovskite Nanocrystals
2D perovskite nanocrystals featuring the structural stability of 2D perovskites and the high photoluminescence quantum yield of nanocrystals are especially suitable for preparing scintillator devices. Colloidal solutions of 2D nanocrystals were fabricated by solution techniques including the simple roomtemperature antisolvent precipitation method [66], the hotinjection method [67] and the ligand-assisted reprecipitation (LARP) method [68], etc. Large-area 2D nanocrystal films could be prepared by further processing of the colloidal solutions. Lee's group [66] prepared 2D perovskite nanocrystals by the antisolvent precipitation method without the addition of capping ligands. PEAI and PbI 2 were mixed in DMF for 1 h to form the precursor solution, which was dripped in chlorobenzene with continuous stirring. After centrifugation at 1000 rpm for 10 min to discard the precipitates, the bright yellow supernatant was collected to obtain the colloidal solution. Huang's group [67] produced CsPbI 3 quantum dots (QDs) of different sizes by the modified hot-injection method. They used the glove box system to create an anhydrous and oxygenfree environment. First, they mixed Cs 2 CO 3 , oleic acid (OA) and octadecene (ODE) to prepare the Cs-Oleate Precursor. Then ODE and PbI 2 were added and dried at 120 • C for 1 h. Next, the OA and Oleyamine (OAM) (preheated at 70 • C) were mixed. The solution became clear after heating up to 185 • C. Finally, Cs-Oleate precursor was swiftly injected into the solution. After 5 s, the solution was cooled by immediately immersing into an ice bath. The LARP technique was similar to the previous method except that a small number of long-chain ligands were introduced to the precursor solution. These ligands were anchored on the surface of nanoparticles to control their size and morphology and prevented the nanoparticles from aggregation [69]. Dong's group [68] fabricated CH 3 NH 3 PbX 3 QDs at room temperature using the LARP synthesis. They simply mixed a DMF solution of CH 3 NH 3 PbBr 3 with a poor solvent (toluene, hexane, etc) containing longchain organic ligands by vigorously stirring, resulting in the controlled crystallization of precursors into colloidal nanoparticles. In view of x-ray detection, 2D nanocrystal films need to be thick enough because of the strong x-ray penetration. Liu's group [70] reported large-area thin films prepared by self-assembly of nanosheets. These scintillators' solutionprocessable nanocrystals could be assembled into uniform and dense high-quality films over an area of 8.5 × 8.5 cm 2 , achieving high resolution (<0.21 mm) imaging. In addition, Bao's group [71] fabricated large-area QDs film arrays via inkjet printing on various substrates. The CsPbBr 3 QDs ink was printed in the middle of the counter electrode pairs with an inkjet printer. The approach showed the advantage of mass production of flexible x-ray detectors with high spatial resolution.

X-ray detection and imaging mechanisms and performance metrics
The current mainstream processes of x-ray detection include direct and indirect (scintillator) detection. In this chapter, we will introduce the working mechanisms of direct-and indirecttype x-ray detectors in detail. In addition, some key parameters of both types of detectors are discussed.

Working mechanisms of direct and indirect x-ray detection and imaging
The biggest difference between the mechanisms of direct detection and scintillation lies in the collection of carriers at the last stage. The direct-type generated electron-hole pairs can be directly collected as electrical signals under an electric field. Therefore, direct x-ray detectors have unique advantages, such as high sensitivity and energy resolution, convenient integration with readout electronics and excellent spatial resolution of x-ray imaging [72,73]. Direct-type x-ray detectors are divided into current and pulse modes according to their operating principles. In the current mode, a detector is subjected to a high flux of high-energy photons. A current signal is generated when charge carriers are collected by two electrodes of the detector under a certain bias ( figure 5(a)). In contrast to the current-mode detector, the detector receives a low flux of high-energy photons during operation and a voltage pulse is output and recorded in the pulse mode. In the fields of radiography imaging and computed tomography, currentmode detectors achieving fast image collection with high spatial resolution are widely used [74].
Indirect detection, the scintillation process in inorganic scintillators can be divided into three main stages including conversion, transport and luminescence ( figure 5(b)). (I) Conversion. High-energy electron-hole pairs are firstly created upon x-ray exposure through the photoelectric effect and Compton scattering at the radiation energy regime of below 1 MeV.
These high-energy carriers are converted to secondary electrons or holes by electron-electron scattering and Auger processes before losing energy. Subsequently, low-kinetic-energy electrons and holes can be generated by interacting with the phonons and gradually accumulate at the conduction band and valence band, respectively. The process of their interaction with photons in the material until the energy disappears is called 'thermalization'. (II) Transport of carriers toward the luminescence center. In this process, charge carriers are likely to be captured by defects or vacancies in the material. (III) Luminescence [75][76][77][78][79][80][81]. Electrons and holes radiatively combine at the luminescence center to emit light in the UVvisible region, which is detected by photodetectors. The scintillation process of organic or noncrystalline scintillators can be more complex.

Direct detection.
Carrier generation, transport and collection are the main processes of direct x-ray detection.
Several key parameters affecting these processes can determine the detector's performance. (a) First, the x-ray attenuation ability can be determined by the absorption coefficient (α) which is proportional to Z 4 /E 3 , where Z is atomic number and E is the photon energy. As a consequence, materials with heavy atoms such as Pb, Bi, I, etc, are beneficial for improving x-ray absorption efficiency. (b) In addition, the µτ product is usually used to demonstrate the transport capacity of charge carriers under the electric field, which indicates the average migration distance of carriers before recombination and it can be derived by a modified Hecht Equation [82]: where I and I 0 are the measured photocurrent and saturated photocurrent, respectively, V is the voltage, L is the thickness and s is the surface recombination velocity, which is inversely proportional to the carrier surface lifetime. (c) Sensitivity (S) is another important performance metric for direct x-ray detectors which can be calculated by equation (2) [73]: where I R is the response photocurrent, D is the dose rate and A is the active area of devices. Sensitivity describes the ability of the detector to convert x-ray photons to electronic signals. Higher sensitivity means the detector can achieve a larger response current at a lower dose rate. (d) Besides, resistivity and the lowest limit of detection (LoD) also determine the performance of detectors. Large bulk resistivity is required to reduce the leakage current and achieve the low noise of the detector. The lowest LoD is the lowest dose rate at which the detector produces a signal-to-noise (SNR) of 3 as defined by the International Union of Pure and Applied Chemistry (IUPAC).

Indirect detection.
As for indirect x-ray detection, the key parameters mainly include light yield, respond-decay time and maximum emission wavelength. (a) Light yield determines the conversion efficiency and detection limit of the x-ray detector. The number of emitted photons (N ph ) produced in the scintillation conversion per each MeV energy can be calculated by using equation (3) [82] where E is the energy of the incoming x-ray photon, E g is the bandgap of the scintillator material, S and Q are the quantum efficiencies in the transport and luminescence stages, respectively, β is the average energy to generate one thermalized electron-hole pair. (b) Respond-decay time can reflect dynamic real-time x-ray imaging ability. The light response dynamics of the scintillator are indicated in figure 5(b). In the transportation phase, the optical response of the nonexponential slow component of the material under high-energy excitation gets complicated due to the recapture process. The decay time of the scintillator can be calculated to provide a simple 1/e or 1/10 decay time when the light intensity after high-energy excitation falls to 1/e or 1/10 of the initial intensity. (c) Maximum emission wavelength needs to be adjusted by changing the E g of scintillator materials to maximize light yield.

X-ray imaging.
To realize a flat panel x-ray imaging (FPXI), one needs to integrate pixelated arrays of detectors into a readout circuit (such as a thin-film transistor) [83][84][85]. In the case of x-ray images with sufficient contrast and dynamic range, the spatial resolution of an x-ray detector is limited by the pixel size and charge crosstalk. For direct and indirect x-ray imagers, the modulation transfer function (MTF) and detective quantum efficiency (DQE) are commonly used to evaluate imaging quality. The MTF, defined as the output-toinput contrast ratio as a function of spatial frequency (ƒ), is used to assess the spatial resolution of an image. The MTF is described in equation (4) [73].
where ƒ stands for spatial frequency and is usually expressed as the number of discernable black and white 'line pairs' per unit length (line pairs/mm) [86]. The DQE represents the efficiency of an x-ray detector in converting x-ray photons, which reflects the combined effects of the signal and noise features of an imaging system.

Recent advances in 2D perovskite x-ray detectors and imaging
Given the extraordinary optoelectronic properties and moisture tolerance, 2D perovskites are expected to be excellent potential candidates for radiation detectors. In this chapter, we mainly introduce the recent research progress of 2D perovskites in direct-type and scintillator-type radiation detectors, respectively. Tables 1 and 2 show the core performance parameters of the recently reported direct 2D perovskite x-ray detectors and 2D scintillators, respectively.

Direct 2D perovskite detector and imaging
In 2019, Prof. Luo's group [87] reported a highperformance x-ray detector based on 2D hybrid perovskite BA 2 EA 2 Pb 3 Br 10 grown by the slow evaporation of concentrated hydrobromic acid. Owing to the ferroelectric polarization of BA 2 EA 2 Pb 3 Br 10 single crystal ( figure 6(a)), the carrier mobility-lifetime product (µτ ) was anisotropic along different crystallographic axes. It exhibited a high sensitivity of 6.8 × 10 3 µC Gy air −1 cm −2 along the crystallographic c-axis and a low LoD of 5.5 µGy air s −1 under a low voltage (figures 6(b) and (c)). Although Pb-based perovskites exhibit excellent x-ray detection performance, the toxicity of lead should not be ignored in view of potential leakage during long-term operation. Because of the similar electronic structure of Bi 3+ and Pb 2+ , Bi is considered a promising alternative to Pb as a low-toxic element. Zhuang and coworkers [55] reported a sensitive x-ray detector made of an NH 4 Bi 2 I 9 single crystal. NH 4 Bi 2 I 9 has a 2D layered structure and provides unique anisotropic x-ray detecting performance ( figure 6(d)). Specifically, the NH 4 Bi 2 I 9 single crystal has µτ products of 1.1 × 10 −2 cm 2 V −1 in the parallel direction and 4.0 × 10 −3 cm 2 V −1 in the perpendicular directions respectively. Due to the large µτ products of the NH 4 Bi 2 I 9 single crystal, the sensitivity of the parallel direction device exhibited a larger value of 8.2 × 10 3 µC Gy air −1 cm −2 (figure 6(e)). The LoD of 55 nGy air s −1 was achieved in the perpendicular direction (figure 6(f)). Although perovskite SCs usually have better performance than polycrystals, the fabrication of large-size 2D SCs remains a significant challenge. Li and coworkers [62] fabricated an oriented lead-free 2D perovskite (F-PEA) 3 BiI 6 pellet with an area of 1.33 cm 2 by a fast tableting strategy. The oriented 2D pellet showed different anisotropy resistivities of 5 × 10 10 Ω cm and 2 × 10 11 Ω cm in the lateral and vertical directions respectively (figure 6(g)). The device exhibited a sensitive x-ray response and the LoD of 30 nGy air s −1 (figures 6(h) and (i)).
2D perovskites typically contain insulated long-chain organic cations that can improve hydrophobicity and bulk resistivity, resulting in lower noise and enhanced stability than 3D perovskites. For example, Prof. Wei's group [88] reported a sensitive 2D (F-PEA) 2 PbI 4 perovskite single-crystal hardx-ray detector. Compared with (PEA) 2 PbI 4 , the introduction of deficient F atoms enhanced supramolecular interactions between organic spacers and inorganic [PbI 6 ] 4− octahedral sheets and suppressed ion migration under a large bias of 200 V. Consequently, the sensitivity of 3402 µC Gy air −1 cm −2 under 120 keV p (the accelerated voltage of the x-ray tube is 120 kV, so the peak photon energy is 120 keV p ) and the LoD of 23 nGy air s −1 were achieved. Meanwhile, the detector showed high operation stability under high-flux x-ray irradiation (figures 7(a) and (b)). Different from A 2 A' n−1 M n X 3n+1 type 2D perovskite SCs reported above, Liu and co-workers [23] successfully synthesized a centimeter-sized AA' n−1 M n X 3n+1 type 2D perovskite single crystal of BDAPbI 4 by a modified temperature crystallization method ( figure 7(c)). The BDAPbI 4 single-crystal x-ray detector presented an excellent sensitivity of 242 µC Gy air −1 cm −2 ( figure 7(d)) under a bias of 10 V and a low LoD of 430 nGy air s −1 . The device also exhibited excellent stability due to the strong interaction between the diammonium cation and the inorganic layer (figure 7(e)).

Indirect 2D perovskite detector and imaging
Unlike most semiconductors, the self-assembled quantum well structure of 2D perovskite enables a high exciton binding energy due to the quantum confinement effect [92], leading to high luminescence efficiency even at a high temperature. Furthermore, the fast response time [93], rapid decay and high light yield make 2D perovskite one of the most promising scintillators [94]. Recent advances in 2D scintillators mainly focus on the fabrication of high-quality scintillator materials. Chemical doping of the scintillator host materials created luminescent centers and provided rapid energy transfer pathways [95,96]. Xie et al [97] introduced Li + into the precursor solution of the PEA 2 PbBr 4 and prepared high-quality crystals through a slow evaporation approach. A dopinginduced broadening of the x-ray emission spectrum (between 450 and 750 nm) was observed at 10 K and assigned to selftrapping exciton (STE) emission. Compared to the bump of pristine scintillator crystals, Li doping with a molar ratio of 1:1 can also survive up to at 200 K (figures 8(a) and (b)). This ruled out the possibility that the Li doping quenched the emission of free excitons (FE) and STE and indicated that it can increase the FE and STE at the same time. Furthermore, they observed a stronger thermomolar emission peak after 60 min for crystals with higher Li doping ratios ( figure 8(c)). These results confirmed that Li doping can improve the luminescence properties of 2D PEA 2 PbBr 4 scintillators while broadening the luminescence wavelength under x-ray irradiation. The incorporation of different equivalent cations such as Sr 2+ , Cd 2+ , Ba 2+ , Mn 2+ , Fe 2+ , etc into the inorganic layer was reported to improve the performance of the scintillators. Nakauchi et al [98] partially replaced Pb 2+ with Ba 2+ to prepare the 2D perovskite (i.e. (C 6 H 5 (CH 2 ) 2 NH 3 ) 2 Ba x Pb 1−x Br 4 ). When x is 0.25, the quantum yield (QY) of the scintillator reached a maximum value of 31.7% and the luminescence yield (LY) reached 19 000 photons MeV −1 , which was 1.4 times than that of the undoped one ( figure 8(d)). Afterglow of the x = 0.5 doped sample is lower than that of the control Ce: GSO scintillator (figure 8(e)), which increases the possibility for the practical application of 2D perovskite scintillator.    [98]. © 2020 Japanese Journal of Applied Physics. (f) Radioluminescence intensity of (C 8 H 17 NH 3 ) 2 SnBr 4 −PMMA as a function of radiation dose. The inset shows the radioluminescence profile at low dose rates. (g) Photograph of a bar pattern phantom (left) and the corresponding x-ray image (right). (h) Photographs of target materials (a capsule containing a spring inside, a circuit board and a crab) and the obtained x-ray images. (f)-(h) Reproduced with permission [24]. Copyright 2020, American Chemical Society.
In addition, 2D perovskites with lead-free B-site cations such as Sn 2+ , Mn 2+ , Ba 2+ and heterovalent metals such as Ag + , Bi 3+ are currently under the spotlight due to reduced toxicity concerns [99,100]. Gu's group [24] used an improved spin-coating method to prepare uniform high-quality Sn-based 2D perovskite (i.e. (C 8 H 17 NH 3 ) 2 SnBr 4 ) thick-film. The Snbased 2D perovskite scintillator exhibited an effective light conversion efficiency at a low dose of 104.23 µGy s −1 (figure 8(f) and inset). The spatial resolution of the tin-based 2D scintillator was investigated by x-ray imaging of a barpattern phantom. A slit as thin as 0.2 mm could be displayed and distinguished ( figure 8(g)). To demonstrate the application of 2D perovskite scintillators in non-destructive inspection and imaging of biological tissues, they also succeeded in obtaining clear images of three representative target materials, i.e. opaque capsules with metal springs inside, circuit boards with complex structures and a live crab ( figure 8(h)).
To improve the environmental stability of scintillator devices, it is ideal to replace the volatile organic cations at the A site with inorganic components. Kuang's group [101] for the first time prepared all-inorganic Mn-based 2D perovskite single crystal scintillators. The crystal structure of Cs 4 MnBi 2 Cl 12 with n = 3 was shown in figure 9(a). The [MnCl 6 ] 4− octahedron layer was sandwiched between the [BiCl 6 ] 3− octahedron layers through corner-sharing and the gap was filled with Cs + to achieve charge balance. The introduction of the [BiCl 6 ] 3− octahedra induced a strong absorption in the ultraviolet region ( figure 9(b)). In addition, the irradiation-generated excitons were immediately transferred from the [BiCl 6 ] 3− to the [MnCl 6 ] 4− octahedron through the  [101], Copyright (2020), with permission from Elsevier. 4 T 1g to 6 A 1g transition process (figure 9(c)). The unique optical management characteristics enabled Cs 4 MnBi 2 Cl 12 a superior radiative luminescence than a commercial NaI(TI) scintillator in the x-ray range (10-35 keV) (figure 9(d)). They fabricated a large-area (100 square centimeters) module with ground Cs 4 MnBi 2 Cl 12 single crystal powders for x-ray imaging (figure 9(e)). The structure of a metal core inside of a plastic capsule was recorded (figure 9(f)).

Challenges and outlook
The unique crystallographic structure and high bulk resistivity offer 2D perovskites excellent optoelectronic properties and superior environmental stability, making them very promising candidates for the fabrication of next-generation x-ray detectors and imaging devices. On the other hand, the investment in 2D perovskite x-ray detectors is at the initial stage from the technology readiness level viewpoint [113,114]. Great challenges remain to be solved before technology commercialization can be considered. In this section, we pinpointed several major obstacles of 2D perovskite x-ray detectors that should be overcome in the near future (figure 10).

Fabricating high-quality 2D perovskites
A thick layer of 2D perovskite (e.g. at least hundreds of micrometers) is essential to completely absorb high-energy x-ray photons. Because 2D perovskites tend to grow along the x and y axes to form a thin layered structure, it is a nontrivial task to fabricate 2D perovskite SCs with sufficient thickness and regular shapes. In addition, the n value determines the crystal structure of a 2D perovskite (A ′ ) m (A) n−1 B n X 3n+1, and therefore their optoelectronic characteristics. Previous studies indicated that solution-processed 2D perovskites usually exhibit a mixed n value, which may limit their performance. Special care such as the precursor concentration and growth temperature should be taken to better control the thermodynamic and kinetic processes during the 2D perovskite growth. This may hold the key to obtaining high-quality 2D perovskites with sufficient thickness and well-defined structure in a reproducible manner.

Improving the sensitivity of direct-type 2D perovskite detectors
The 2D perovskite direct-type detectors can maintain a stable baseline and manifest excellent on-off cycle stability. However, the poor carrier transport dynamics compared to the 3D perovskite detectors severely limit their sensitivity. Compared to pure 2D or 3D perovskites, a rational combination of 2D and 3D perovskites to form a multilayered structure or 2D/3D heterojunctions may hold the advantages of low ion migration in 2D perovskites and high sensitivity in 3D perovskites [115], thus resulting in excellent x-ray detection performance.

Enhancing the light yield for 2D perovskite scintillators
Due to inherent small stokes shift and serious self-absorption behavior (the excited photons reabsorbed by the crystals) [116], 2D halide perovskites suffer from unsatisfactory light Wiley-VCH GmbH]. Reproduced with permission [53]. Copyright 2021, Royal Society of Chemistry. Reprinted from [58], Copyright (2019), with permission from Elsevier. Reprinted from [101], Copyright (2020), with permission from Elsevier. Reprinted from [123], Copyright (2022), with permission from Elsevier. yield and optical coupling efficiency. The increase in the Stokes shift contributes to a small self-absorption. Some measures can be taken to increase the Stokes shift and restrain the self-absorption effect. The relatively thin scintillator is a benefit for improving efficiency and minimizing the response time of self-absorption effects. On the other hand, proper doping of elements is also a powerful strategy to increase transmittance and reduce self-absorption effects.

Mitigating lead toxicity concerns of 2D perovskite detectors
State-of-the-art 2D perovskite x-ray detectors usually contain a lead element, which is highly toxic to humans, animals and ecosystems [117]. Mitigation of the potential lead leakage risk from damaged perovskite detectors is an essential task to be solved before technology commercialization. Fortunately, serious life cycle assessment and significant efforts have been made by the perovskite photovoltaic community. Effective strategies including element substitution and encapsulation have been developed [118][119][120], which could be borrowed by the perovskite x-ray detector community with possible adoptions.