Construction of 2D-3D Halide Perovskite Heterojunctions with the Amine Vapor Replacement Reaction

Abstract

2D/3D perovskite-based X-ray detectors have emerged as one of the most competitive candidates for future medical imaging owing to their ease of fabrication via solution processes and outstanding performance. Nevertheless, despite significant efforts in innovating materials and device structures, it remains challenging to achieve high-quality 2D/3D heterojunctions with a single and effective strategy. In this study, we propose a gas-phase replacement reaction between propylamine (PA) and MAPbBr₃ for the construction of the 2D (PA)₂PbBr₄–3D MAPbBr₃ heterojunction. The growth of the (PA)₂PbBr₄ layer was influenced by the diffusion of PA. The quality of (PA)₂PbBr₄ could be controlled by finely choosing reaction conditions. X-ray detectors based on optimized 2D–3D perovskite heterojunctions showed a sensitivity of 25,143 μC Gy_air ^–1 cm^–2, lowest limit of detection of 1.33 μGy_air s^–1, dark current drift value of 2.27 × 10^–3 nA cm^–1 s^–1 V^–1, and outstanding performance, which are all better than those of detectors based on MAPbBr₃ single crystals. This research unveils the growth mechanism of 2D (PA)₂PbBr₄ perovskite via gas-phase replacement reactions, which are beneficial for the construction of high-quality 2D/3D perovskite-based heterojunctions.

1. Introduction

The 2D/3D heterojunctions are steadily emerging as an indispensably pivotal element within the rapidly evolving realm of halide perovskite-based optoelectronic devices. − The rationally designed 2D/3D heterojunctions can take the advantages of both 2D and 3D perovskites, such as decent stability, slow ion movement, and efficient carrier transport. ,− These heterojunctions, with their unique structural and electronic properties, are enabling a new paradigm of device performance, opening up unprecedented opportunities for enhanced functionality and efficiency in applications such as exceptionally sensitive detectors, bright light-emitting diodes, and high-performance solar cells. ,− The phenethylammonium lead bromide–cesium lead bromide ((PEA)₂PbBr₄–CsPbBr₃) heterojunction-based X-ray detectors exhibited a high sensitivity of 34,496 μC Gy_air ^–1 cm^–2 in 120 keV hard X-ray detection and a low dark current drift of 2.49 × 10^–4 nA cm^–1 s^–1 V^–1. The propylammonium lead bromide–methylammonium lead bromide ((PA)₂PbBr₄-MAPbBr₃) heterojunction-based X-ray detectors exhibited a high sensitivity of 38,147 μC Gy_air ^–1 cm^–2 at an X-ray tube voltage of 70 keV and significantly suppressed ion drift.

The quality of the 2D/3D perovskite heterojunctions could significantly influence their optoelectronic properties as well as the performance of the devices. ,, The solution-based methods have been widely utilized to construct high-quality 2D/3D heterojunctions. , With the precisely controlled growth temperatures and precursor concentrations, (PEA)₂PbBr₄ could epitaxially grow on CsPbBr₃ single crystals (SCs). The (PEA)₂PbBr₄ of the heterojunctions showed an impressive quality, of which the full width at half-maxima (FWHM) for its (005) plane was in the same order of magnitude as the FWHM of (PEA)₂PbBr₄ SC. In an alternative method, (PA)₂PbBr₄ precursors were directly dropped onto the polished MAPbBr₃ SCs and then covered by a glass slide. As the solvent evaporated, orientated (PA)₂PbBr₄ layers formed on the MAPbBr₃ SCs, which were compact and free of pores. Vapor-based methods have been demonstrated to be a versatile strategy for the growth of high-crystallinity 2D organic–inorganic halide perovskites, which can realize epitaxial growth on various substrates and even lead-free tin-based 2D perovskites, thus circumventing the inherent limitations of solution-based methods such as solubility limit and orthogonal solvent requirement. The phenethylammonium lead iodide ((PEA)₂PbI₄) could be deposited via the cosublimation of phenethylammonium iodide (PEAI) and lead iodide (PbI₂). With the calibration of the deposition rate for each material, pure-phase 2D (PEA)₂PbI₄ thin films could be synthesized directly on the surface of 3D methylammonium lead iodide (MAPbI₃) perovskite with preferential orientation along the [00l] direction. Performance enhancement resulting from 2D–3D heterojunctions was scarcely observable. This was due to the trade-off between the reduction in trap states and the impediment to charge extraction. However, a systematic investigation into the quality of 2D–3D heterojunctions fabricated via the vapor-based method had not been carried out, as well as its influence on the device performance.

Herein, we investigate a replacement reaction between MAPbBr₃ and propylamine (PA), which was used for the synthesis of (PA)₂PbBr₄. By variation of the reaction time, compact and uniform (PA)₂PbBr₄ layers could directly form on the top of MAPbBr₃ SCs. While overreactions could yield porous and poor-quality capping layers. X-ray detectors based on optimized (PA)₂PbBr₄–MAPbBr₃ structures exhibited higher sensitivity, lower lowest limit of detection (LoD), and better stability, compared with X-ray detectors made with MAPbBr₃ SCs.

2. Experimental Section

2.1. Materials

Lead bromide (PbBr₂, >98%) was purchased from Alfa Aesar. N,N-dimethylformamide (DMF, >99.5%) and isopropyl alcohol (IPA, ≥99.7%) were purchased from Greagent. Methylammonium bromide (MABr, ≥99%) was purchased from Great Cell Solar. Propylamine (PA, ≥99%) was purchased from Adamas. All chemicals were used as received without further purification.

2.2. Preparation of MAPbBr₃ SC, (PA)₂PbBr₄ Capping Layers

2.2.1. Preparation of MAPbBr₃ SC

1.388 g (0.0125 mol) of MABr, 4.590 g (0.0125 mol) of PbBr₂, and 10 mL of DMF were mixed in a 40 mL beaker. Then, the beaker was sealed with a parafilm and aluminum foil. The mixed solution was stirred on a magnetic stirrer at 500 rpm for 4 h. The well-mixed solution was filtered with a 0.45 μm PTFE filter, and then 100 μL of IPA was added. To obtain seed crystals, the solution was directly placed on a hot plate at 85 °C for 30 min. The as-prepared seed crystals were collected and dried for storage. For the single-crystal growth, one seed crystal was added to the as-prepared precursor solution. Then, the precursor solution with seed crystals was heated to 44 °C in an oil bath. During the oil bath, the temperature needs to be maintained for 24 h and then it should be increased by 1 °C every 24 h.

2.2.2. Preparation of (PA)₂PbBr₄ Capping Layers

An SC with dimensions of approximately 10 × 10 × 2 mm was cut into two pieces by using a cutting machine with a 0.25 mm diamond wire. The surface of the SCs was manually polished using 7000-grit sandpaper. Then, the polished crystals were annealed at 100 °C for 1 h in nitrogen. For PA 200 and PA 400 capping layer synthesis, 50 μL of PA was added to a 100 mL glass bottle. For PA 1000 capping layer synthesis, 125 μL of PA was added to a 100 mL glass bottle and fully evaporated at room temperature. Then, SCs were covered by the bottles with PA vapor to enable the reaction between PA and MAPbBr₃. The reaction was not heated and was carried out under ambient temperature. No postannealing step was performed for the heterojunctions. The thickness of (PA)₂PbBr₄ capping layers could be tuned by adjusting the reaction time.

2.3. Device Fabrication

To fabricate an X-ray detector, a 100 nm thick Au electrode was thermally evaporated onto the samples in a vacuum chamber with a pressure less than 4 × 10^–4 Pa. The exposure area to X-ray was 0.01365 cm², and the area of the gold electrode was 0.0085 cm². The spacing between each finger is 0.15 mm. To enable sufficient X-ray absorption, the thickness of the MAPbBr₃ SCs is set to 1 mm.

2.4. Material Characterizations

The XRD patterns are obtained using a Rigaku SmartLab X-ray diffractometer with Cu–Kα radiation. Scanning is performed from 5° to 80° at 45 kV and 200 mA with a scanning rate of 10°/min. SEM images were obtained via a field-emission scanning electron microscope (Gemini SEM 300). Time-resolved photoluminescence (TRPL) spectra were determined with a time-correlated single-photon counting system (Edinburgh Instruments, FLS1000), using a pulsed laser (375 nm) as the excitation source. UV–vis absorption was performed using Thermo Fisher’s Evolution 220, with a scanning wavelength range of 750 to 400 nm. PL test was monitored by F-2700 (Hitachi, Japan). To measure the photoluminescence (PL) peak of MAPbBr₃, the wavelength of excitation light was 405 nm. The start wavelength of emission spectra was 410 nm, and the end wavelength was 700 nm.

2.5. Measurement of X-ray Detector Performance

The X-ray detection performance was measured in a dark box made of lead to minimize electromagnetic and ambient light disturbance. A tungsten anode X-ray tube (Varex 160, VAREX, USA) was used as the X-ray source. The X-ray tube was operated under a fixed 60 kV voltage, and the current was adjusted from 6.4 to 0.1 mA to change the X-ray dose rate. Several 0.5 mm thick steel sheets were used as the attenuator to generate low X-ray dose rates. The X-ray dose rate was carefully calibrated by using the AT1123 radiation dosimeter manufactured by ATOMTEX (Belarus). During the performance measurements, the X-ray response current was collected by using a precision source meter (Keithley 2400, Tektronix, USA).

All of the above measurements of the X-ray detectors are performed in the ambient atmosphere environment (temperature ∼ 25 °C and relative humidity ∼ 40% RH), and none of the detectors in this work were encapsulated.

The signal-to-noise ratio (SNR) was calculated with the following equation:

$$\mathrm{SNR}=\frac{I_{\mathrm{s}}}{I_{\mathrm{n}}}$$ 1

where I _s represents the disparity between the average current under X-ray (I _p) and the average dark current, while I _n denotes the noise current. The latter is quantified as the standard deviation of the dark current, which can be mathematically expressed as

$$I_{\mathrm{n}}=\sqrt{\frac{1}{N}\sum _i^N(I_i-I_{\mathrm{d}}{)}^2}$$ 2

where I _i represents the measured dark current and I _d represents the average dark current.

2.6. Measurement of the Ion-Migration Activation Energy

To assess the ion-migration level in this work, temperature-dependent conductivity was measured in the dark. Generally, electronic conduction is the main cause of conductivity at low temperature, as the ion migration is suppressed. In this work, the ion-migration rate was characterized by the activation energy (E _a) of ionic conductivity, calculated using the Nernst–Einstein equation from the temperature-dependent measurements:

$$\sigma (T)=\frac{{\sigma }_0}{T}\exp \left(\frac{-E_{\mathrm{a}}}{k_{\mathrm{B}}T}\right)$$ 3

where k _B is the Boltzmann constant, σ₀ is a constant, and T is the temperature. The experimental setup used in this work consisted of a Tektronix 2450 source meter and a Lakeshore cryoprobe station with a Lakeshore 336 temperature controller. The conductivities are recorded in the temperature range of 120 to 300 K, in 10 K steps. The conductivities of the devices were extracted by measuring the I– t curve at 0.5 V.

3. Results and Discussion

A replacement reaction between MAPbBr₃ and PA was utilized to produce (PA)₂PbBr₄:

$$2\mathrm{PA}(\mathrm{g})+2{\mathrm{MAPbBr}}_3(\mathrm{s})={(\mathrm{PA})}_2{\mathrm{PbBr}}_4(\mathrm{s})+{\mathrm{PbBr}}_2(\mathrm{s})+2\mathrm{MA}(\mathrm{g})$$ 4

In such reaction, PA and methylamine (MA) are in the gas phase. The MAPbBr₃, (PA)₂PbBr₄, and PbBr₂ are in the solid phase. As PA can be easily evaporated at room temperature, this reaction is able to perform at ambient atmosphere, which does not require the ultrahigh vacuum environment. The MA can also be easily evaporated at room temperature, which escapes from the crystals after the reaction. The (PA)₂PbBr₄ perovskite and PbBr₂ ought to form at the surface of MAPbBr₃ SCs.

The thickness of the (PA)₂PbBr₄ layers could be tuned by varying the reaction time. After 20 s of reaction, a 200 nm thick capping layer formed at the surface of MAPbBr₃ SCs (Figure a,b). Compared with the inner MAPbBr₃ SCs, the capping layer was rough and polycrystalline. By further increasing the reaction time to 40 and 80 s, the capping layers’ thickness became ∼300 and ∼400 nm, respectively (Figures S1a and c). The growth speed of the 2D perovskite layer decreased as the reaction prolonged. At the first 20 s, the average growth speed was around 10 nm/s. From 20 to 40 s, the average growth speed decreased to ∼5 nm/s. From 40 to 80 s, the average growth speed further decreased to ∼2.5 nm/s. Figure S2 demonstrates the relationship between the growth rate of (PA)₂PbBr₄ and the reaction time: the growth rate decreases with the extension of time. The growth of (PA)₂PbBr₄ became extremely slow and nearly halted after 100 s, which was a typical characteristic of self-limited growth. Given that the concentration of the gas-phase precursor remained identical, the reduction in growth rate could be ascribed to the impeded diffusion of PA from the surface to the inner part of crystals. As the thickness of the 2D perovskite layer increased, PA encountered greater difficulty in migrating to reach MAPbBr₃. Consequently, this not only restricted the reaction rate but also slowed the growth rate of the capping layers. To achieve a thicker capping layer, the concentration of PA vapor should be increased to reach a higher growth rate. When the concentration of PA was increased to 2.5-fold its initial value, the thickness of the capping layer reached 1000 nm (Figure d). However, nanopores emerged, which should be detrimental to the optoelectronic properties of 2D–3D heterojunctions. For simplicity, the MAPbBr₃ SC with 200 nm (PA)₂PbBr₄ layers, 400 nm (PA)₂PbBr₄ layers, and 1000 nm (PA)₂PbBr₄ layers were named PA 200, PA 400, and PA 1000, respectively.

1.

Side-view scanning electron microscopy (SEM) images of (a) the cleaved MAPbBr₃ SC, (b) 200 nm (PA)₂PbBr₄ on a MAPbBr₃ SC (PA 200), (c) 400 nm (PA)₂PbBr₄ on a MAPbBr₃ SC (PA 400), and (d) 1000 nm (PA)₂PbBr₄ on a MAPbBr₃ SC (PA 1000). The scale bar is 1 μm. Top-view SEM images of (e) the cleaved MAPbBr₃ SC, (f) PA 200, (g) PA 400, and (h) PA 1000. The scale bar is 2 μm. (i) XRD patterns of the MAPbBr₃ SC (yellow) and PA 400 (green). (j) Enlarged XRD pattern of PA 400. The XRD peaks of (PA)₂PbBr₄ are indicated by hash signs.

The surface morphology of the capping layer also changed as the reaction time varied. The cleaved MAPbBr₃ SC showed an ultrasmooth surface (Figure e). In stark contrast, the polished MAPbBr₃ SC showed scratches of over 10 μm lengths and grains of around 200 nm diameters (Figure S1b). After reacting with PA, the grain size significantly increased to around 500 nm and pinholes appeared for the PA 200 sample (Figure f). For the PA 400 sample, the capping layer became dense and free of pinholes (Figure g). Besides, the surface of PA 400 was flatter than the surface of PA 200. For the PA 1000 sample, the capping layer was full of cracks and plates, which showed a big difference from the PA 200 and PA 400 samples (Figure h). The X-ray diffraction (XRD) patterns of polished SC only showed peaks of MAPbBr₃, which indicated that polishing treatment did not produce new phases (Figure i, yellow patterns). Thus, it could be concluded that the polishing process only changed the surface morphology, but not the phase of MAPbBr₃ SC. The XRD pattern of PA 400 showed peaks at 7.04°, 14.10°, 21.22°, and 28.38°, which corresponded to the (002), (004), (006), and (008) crystal planes of (PA)₂PbBr₄ (Figure i,j, green patterns). At the sample time, diffraction peaks at 14.9° and 30.1° of the (100) and (200) crystal planes of MAPbBr₃ were also observed, which indicated the successful construction of (PA)₂PbBr₄-MAPbBr₃ heterojunctions. Compared with PA 400, the XRD pattern of PA 1000 showed new peaks at 14.50°, 18.54°, 21.40°, 21.60°, and 24.20° apart from the peaks of (PA)₂PbBr₄ and MAPbBr₃ (Figure S3). These peaks corresponded to the (101), (002), (011), (102), and (111) planes of PbBr₂, which indicated its formation after the reaction with PA.

To clarify its location within the heterojunction, we conducted two rounds of mild polishing on the PA 1000 sample and collected the corresponding XRD patterns (Figure S4a–c), which correspond to the unpolished (PA 1000-0), once-polished (PA 1000-1), and twice-polished (PA 1000-2) surfaces of the samples, respectively. Distinct characteristic diffraction peaks of PbBr₂ (matching its (101), (011), and (102) crystal planes) were detected on the unpolished surface (Figure S4a). The diffraction peak intensity of the (PA)₂PbBr₄ (006) plane was comparable to that of the PbBr₂ (011) plane. After one round of polishing (Figure S4b), the diffraction peak intensity of the PA₂PbBr₄ (006) plane became significantly larger than that of the PbBr₂ (011) plane, indicating the decreased portion of PbBr₂. After two rounds of polishing (Figure S4c), only the diffraction peaks of MAPbBr₃ remained. This is because the nanoscale PA₂PbBr₄ layer can be completely removed, even by mild polishing. These results confirm that the PbBr₂ formed during the over-reaction stage of the vapor-phase reaction and is located within the PA₂PbBr₄ layer. This conclusion is further corroborated by the top-view optical images of the PA 1000 samples recorded during the polishing process (Figure S4d–f). White powdery aggregates (identified as PbBr₂) were observed on the surface of the unpolished device (Figure S4d). Partial residues of these white powdery aggregates remained after one round of polishing (Figure S4e). The white powder disappeared entirely after two rounds of polishing (Figure S4f)a phenomenon consistent with the removal of the top PbBr₂ phase.

At the initial stage, PA vapor reacts with MAPbBr₃ to produce (PA)₂PbBr₄, while little PbBr₂ forms. However, chemical eq is unbalanced in this case. Thus, we proposed a modified equation for the chemical reaction at the initial stage:

$$2\mathrm{PA}(\mathrm{g})+{\mathrm{MAPbBr}}_3(\mathrm{s})={(\mathrm{PA})}_2{\mathrm{PbBr}}_4(\mathrm{s})+\mathrm{MA}(\mathrm{g})+{\mathrm{V}}_{\mathrm{Br}}$$ 5

where the V_Br is the bromide vacancy. In this balanced equation, the replacement reaction between PA and MAPbBr₃ is proposed to produce (PA)₂PbBr₄, MA gas, and bromide vacancies. This assumption is reasonable, considering the bromide ions could migrate within the lattice of MAPbBr₃ SCs. −

To intuitively investigate whether bromine vacancies (V_Br) are introduced during the first stage of the gas-phase reaction (eq ), we characterized the interfaces before and after the gas-phase treatment using X-ray photoelectron spectroscopy (XPS). As shown in the XPS spectra in Figure S5, the pristine MAPbBr₃ sample exhibits characteristic Br 3d peaks at 68.5 and 69.4 eV. For the PA 200 sample after gas-phase reaction, the Br 3d peaks are located at 67.7 and 68.6 eV. A distinct shift of the Br 3d peaks toward lower binding energies can be observed, which is ascribed to the formation of V_Br introduced by eq .

Based on the above discussion, the growth process of (PA)₂PbBr₄ layers can be separated into two stages (Scheme S1). At stage I, (PA)₂PbBr₄ is formed through a replacement reaction between MAPbBr₃ and PA. Besides, its nucleation sites are distributed on the surface of MAPbBr₃ SCs at certain intervals. Subsequently, pinholes are present in the (PA)₂PbBr₄ layers. As the growth continues, the grains of (PA)₂PbBr₄ grow larger and coalesce. In this manner, the pinholes become smaller and may even vanish. Consequently, a compact (PA)₂PbBr₄ layer is formed, which impedes the invasion of the PA vapor. As a result, the growth rate of the (PA)₂PbBr₄ layer decreases as it becomes more compact. At stage I, the replacement reaction between MAPbBr₃ and PA can be described by eq . When the growth of (PA)₂PbBr₄ layers goes into stage II, PbBr₂ forms and the replacement reaction can be described by eq . Simultaneously, the compact (PA)₂PbBr₄ layer thickens and then collapses. The capping layer evolves into a mixture of (PA)₂PbBr₄ and PbBr₂, featuring numerous nanopores.

The bandgaps of MAPbBr₃, PA 200, and PA 400 were 2.16 eV (Figures a and S6), while the bandgap of PA 1000 was 2.11 eV. As PA₂PbBr₄ layers of PA 200 and PA 400 were thin and uniform, they could hardly show a perceivable influence on the bandgaps. The trap states of semiconductors can remarkably influence their PL. − For the MAPbBr₃ SC, the steady-state PL peak was located at 545 nm with an intensity of ∼230 counts (Figure b, yellow spectrum). For the PA 200 and PA 400, the PL peaks of MAPbBr₃ were located at 533 nm with an intensity of ∼2575 and 3375 counts, respectively (Figure b, purple and green spectra). With the (PA)₂PbBr₄ capping layers, the PL peaks were blue-shifted and their intensities increased significantly, which implied the trap states of MAPbBr₃ were largely reduced. − For PA 1000, the PL peak intensity of MAPbBr₃ was too low for detection (Figure b, blue spectrum). The TRPL lifetime of MAPbBr₃ SC was only 112 ns. With the (PA)₂PbBr₄ capping layers, TRPL lifetimes of the PA 200 and PA 400 were 346 and 465 ns, respectively (Figure c, purple and green spectra). The TRPL lifetime decreases as the trap state density increases in semiconductors. ,, Thus, the remarkably extended TRPL lifetime further corroborated that the trap states of MAPbBr₃ SCs have been diminished due to the presence of (PA)₂PbBr₄ capping layers.

2.

(a) Tauc plots of MAPbBr₃ (yellow), PA 200 (purple), PA 400 (green), and PA 1000 (blue). (b) Steady-state PL spectra and (c) TRPL curves of MAPbBr₃, PA 200, and PA 400.

The halide perovskite-based X-ray detectors had a lateral structure (Figure a). When the detectors were working, carriers passed through the layers with a sequence: Au → (PA)₂PbBr₄ layer → MAPbBr₃ → (PA)₂PbBr₄ layer → Au. In this way, the device structure should be described as Au/(PA)₂PbBr₄/MAPbBr₃/(PA)₂PbBr₄/Au, which was symmetric (Figure S7). The dark current was reduced with the (PA)₂PbBr₄ layer (Figure b). When not exposed to X-ray, the dark current of PA 200 and PA 400 was notably lower than the dark current of MAPbBr₃ SCs in the voltage range from 0 to 50 V (Figure b, yellow, purple and green curves). The (PA)₂PbBr₄ layer reduced the Shockley–Read–Hall (SRH) recombination within materials and increased the resistivity of the materials, which led to the reduction in the dark current. However, the dark current of PA 1000 was comparable with that of MAPbBr₃ SCs, indicating the failure of such porous capping layers.

3.

(a) Schematic illustration of the structure of halide perovskite-based X-ray detectors. (b) Dark current vs voltage curves of MAPbBr₃ (yellow), PA 200 (purple), PA 400 (green), and PA 1000 (blue). (c) Sensitivity and (d) SNR of MAPbBr₃, PA 200, PA 400, and PA 1000. Current vs time curves of (e) MAPbBr₃, (f) PA 200, (g) PA 400, and (h) PA 1000 under different dose rates.

The X-ray sensitivity (S) is defined as the collected charge (C) per unit area (A) per unit of radiation exposure (D), which presents the efficiency of a detector for converting an X-ray to a photocurrent (eq ).

$$S=\frac{C}{A\times D}=\frac{\mathrm{Photo}\mathrm{current}\mathrm{density}}{\mathrm{Dose}\mathrm{rate}}$$ 6

Under 267 V mm^–1 electric field, the sensitivity of MAPbBr₃, PA 200, PA 400, and PA 1000 was 5002 μC Gy_air ^–1 cm^–2, 20,941 μC Gy_air ^–1 cm^–2, 25,143 μC Gy_air ^–1 cm^–2, and 6184 μC Gy_air ^–1 cm^–2, respectively (Figure c). Compared to MAPbBr₃, both PA 200 and PA 400 showed largely increased sensitivity, which indicated the benefit of (PA)₂PbBr₄ layers. The reason was that (PA)₂PbBr₄ layers reduced the trap states of the MAPbBr₃ SCs. The sensitivity of PA 1000 was only slightly higher than that of MAPbBr₃.The X-ray responses of MAPbBr₃, PA 200, PA 400, and PA 1000 were stable and linear to the dose rate of the X-ray (Figure e–g). For optimal performance, the thickness ratio between the (PA)₂PbBr₄ layer and the MAPbBr₃ layer was 1:2500.

According to the definition of the International Union of Pure and Applied Chemistry, the detection limit is the equivalent dose rate to generate a signal larger than three times of the signal noise. Accordingly, LoD was defined as the dose rate yielding a SNR value of 3. By extrapolating from existing data, the LoD of MAPbBr₃, PA 200, PA 400, and PA 1000 was calculated to be 19.87 μGy_air s^–1, 1.97 μGy_air s^–1, 1.33 μGy_air s^–1, and 13.33 μGy_air s^–1 (Figure d). Among them, PA 400 showed the lowest value, which suggested its potential for low-dose X-ray imaging.

In addition, the surface of PA 400 showed fewer pinholes than the surface of PA 200. As pinholes in the 2D layer act as charge leakage channels, allowing direct contact between the electrode and the underlying 3D perovskite. PA 400 could reduce more trap states than PA 200 did, which resulted in a higher sensitivity of PA 400. Along this line, those pinholes could increase the dark current and reduce the SNR of the X-ray detectors, whereas the suppression of such pinholes in PA 400 resulted in a decreased dark current and an increased SNR compared to PA 200.

The stability of halide perovskite-based X-ray detectors was tested under an electric field of 267 V mm^–1, and the total X-ray dose was 2.15 Gy. Compared to MAPbBr₃ and PA 1000, PA 200, and PA 400 showed more stable X-ray response with a small fluctuation in current (Figures a–d and S8). After operation, the sensitivity of MAPbBr₃, PA 200, PA 400, and PA 1000 maintained 93.37, 95.03, 97, and 94.32% of their initial value, respectively. Compact (PA)₂PbBr₄ layers could significantly improve the stability of MAPbBr₃-based X-ray detectors. The dark current drift was performed under a 267 V mm^–1 electric field. The MAPbBr₃, PA 200, PA 400, and PA 1000 had a dark current drift value of 6.06 × 10^–2, 8.83 × 10^–3, 2.27 × 10^–3, and 4.41 × 10^–2 nA cm^–1 s^–1 V^–1, respectively (Figure e–h). Among them, PA 400 exhibited the lowest dark current drift, which indicated that its (PA)₂PbBr₄ layer could suppress ion migration most effectively. To accurately separate the contributions of ionic and electronic conductivities to the total current and ensure the reliability of ion-migration activation energy (E _a) calculations, we adopted a standard protocol for ionic transport characterization in halide perovskites, as reported in the literature. Specifically, we analyzed the dynamic characteristics of temperature-dependent current–time (I– t) curves, leveraging the distinct time-scale differences between ionic migration and electronic transport to decouple the two components. , Under a constant bias of 0.5 V, the instantaneous current in the initial stage of the I– t curve is dominated by electronic conductivity, while the subsequent slow current decay originates from ionic migration and accumulation at the electrodes. The pure ionic conductivity contribution was extracted by fitting the slow-decay component of the steady-state current, and the ion-migration activation energy was calculated based on the Arrhenius equation (Figures S9 and S10).

4.

Long-term current vs time curves of (a) MAPbBr₃, (b) PA 200, (c) PA 400, and (d) PA 1000 with a total X-ray dose of 2.15 Gy. Dark current drift curves of (e) MAPbBr₃, (f) PA 200, (g) PA 400, and (h) PA 1000.

4. Conclusions

In this study, a gas-phase replacement reaction was developed to synthesize 2D (PA)₂PbBr₄ capping layers on top of 3D SCs. This replacement reaction between PA and MAPbBr₃ was self-limited, and the reaction speed decreased as the reaction prolonged. Through the optimization of the reaction conditions, the ideal thickness of (PA)₂PbBr₄ layers was experimentally determined to be 400 nm. The MAPbBr₃ SCs with 400 nm (PA)₂PbBr₄ layers showed the highest PL intensity and longest TRPL lifetime, which indicated their low defect densities. Thereafter, X-ray detectors based on PA 400 showed a sensitivity of 25,143 μC Gy_air ^–1 cm^–2, LoD of 1.33 μGy_air s^–1, and dark current drift value of 2.27 × 10^–3 nA cm^–1 s^–1 V^–1, which are all better than those of detectors based on MAPbBr₃ SCs.

The research data underlying this article, including experimental characterization data, spectral curves, and device performance test results, are fully presented in the main manuscript figures/tables and the accompanying Supporting Information. All relevant raw data are available from the corresponding author upon reasonable request.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.5c11585.

• Side-view scanning electron microscopy (SEM) images of 300 nm (PA)₂PbBr₄ on a MAPbBr₃ SC (PA 300); growth rate of (PA)₂PbBr₄ on MAPbBr₃; XRD patterns of the PA 1000 sample; XRD patterns of the PA 1000 device surface under different polishing conditions and the top-view photographs of PA 1000; schematic diagram of the growth mechanism of the capping layers; XPS of the MAPbBr₃ and PA 200; UV–vis absorption spectra of four samples; schematic illustration of the current flow inside detectors with capping layers; performance degradation in stability tests and temperature-dependent current of four samples; and temperature-dependent conductivity measurements of PA 200 and PA 400 (PDF)

Conceptualization, X.W. and S.X.; methodology, B.Z., C.H., S.Z., B.X., X.X., T.Z., and S.X.; formal analysis, S.X., B.Z., C.H., X.X., B.Y., and X.Z.; investigation, B.Z., C.H., B.Y., and S.Z.; writingoriginal draft preparation, S.X., C.H., and B.Z.; writingreview and editing, S.X., X.W., and T.Z.; supervision, X.W. and S.X.; and funding acquisition, S.X., X.W., and T.Z. All authors have read and agreed to the published version of the manuscript. B.Z. and C.H. contributed equally to this work.

The authors declare no competing financial interest.