Water-induced MAPbBr3@PbBr(OH) with enhanced luminescence and stability

Poor stability has long been one of the key issues that hinder the practical applications of lead-based halide perovskites. In this paper, the photoluminescence (PL) quantum yield (QY) of bromide-based perovskites can be increased from 2.5% to 71.54% by introducing water, and the PL QY of a sample in aqueous solution decreases minimally over 1 year. The enhanced stability and PL QY can be attributed to the water-induced methylamino lead bromide perovskite (MAPbBr3)@PbBr(OH). We note that this strategy is universal to MAPbBr3, formamidine lead bromide perovskite (FAPbBr3), inorganic lead bromide perovskite (CsPbBr3), etc. Light-emitting devices (LEDs) are fabricated by using the as-prepared perovskite as phosphors on a 365 nm UV chip. The luminance intensity of the LED is 9549 cd/m2 when the driven current is 200 mA, and blemishes on the surface of glass are clearly observed under the illumination of the LEDs. This work provides a new strategy for highly stable and efficient perovskites.

The origin of the instability of LHPs is generally attributed to their low formation energy, which makes these materials vulnerable to light, heat, oxygen, and moisture, especially when water is involved [26][27][28][29] . In addition, the ions of LHPs with discrete components are exchanged with each other quickly, which results in severe shifts in the emissions. Strategies for improving the stability of LHPs can be classified into three types: compositional engineering, surface engineering, and matrix encapsulation 30 . The all-inorganic lead perovskite (CsPbBr 3 ), where methylamino (MA) ions are replaced with Cs ions, shows much higher stability, including thermal and environmental stability, than organic-inorganic perovskite 31 . Although the thermal and environmental stability of this perovskite has been improved through compositional engineering, poor moisture stability is still a serious problem. The stability of LHPs improves greatly when they are covered with long-chain surface agents, which has been widely demonstrated by researchers. However, LHPs usually lose surface ligands and then reunite and lose their colloidal stability during purification 32 . Matrix encapsulation has been applied to effectively enhance the stability and PL efficacy of perovskite. For example, Jia et al. prepared CsPbX 3 /Cs 4 PbX 6 core/shell perovskite nanocrystals by applying a seeded growth approach, which provided improved PL QY. Wang et al. produced a Rb 4 PbBr 6 shell to protect CsPbX 3 through the rubidium oleate posttreatment method. Similarly, Tang et al. reported a kind of CsPbBr 3 /CdS core/shell structure using a hot-injection method. Embedding CsPbBr 3 into SiO 2 or the pores of mesoporous silica can also improve the stability of the perovskite [33][34][35][36][37] . Although most inorganic matrices are dense and thermally stable, it is difficult to controllably form an inorganic protective layer covering perovskite quantum dots (QDs) due to the high synthetic temperature and relatively complicated synthesis conditions. Furthermore, the introduction of expensive or toxic elements further hinders the applications of perovskites. Moreover, the phase separation between LHPs and protection media still remains, resulting in size variation, low loading, broad spectra, and low PL QY. Some researchers have demonstrated the positive role of water in the synthesis of LHPs [38][39][40] . For example, Andrey et al. synthesized stable CsPbBr 3 nanocrystals by introducing a suitable amount of water into the reaction mixture 41 . It is important to note that when a large amount of water was introduced in this system, the perovskite still decomposed. Atanu and Kwang demonstrated the aqueous synthesis of various hybrids and all-inorganic halide perovskites in acidic and basic media 42 . The perovskites were stable in water for more than 6 months, but the PL QY of the as-prepared methylamino lead bromide perovskites (MAPbBr 3 ) was only 11.7%. These observations indicate that there exists a solubility equilibrium between the crystallization of perovskite and its saturated ionic components in water. Thus, the luminescence and stability of LHPs may be enhanced by water by choosing appropriate pH, metal halide salts, and different organic components.
In this work, we show that the PL QY and stability of LHPs can be greatly enhanced by adding water, and the PL QY of the LHPs can be increased from 2.50% to 71.54%, while that of a sample in aqueous solution decreases minimally after 1 year. This strategy is universal to MAPbBr 3 , formamidine lead bromide perovskites (FAPbBr 3 ), and CsPbBr 3 . UV-pumped LEDs have been fabricated by using the prepared perovskites as phosphors, and blemishes including scratches, dust, and fingerprints on the surface of glass can be observed clearly under the illumination of the LEDs, indicating that the LEDs are suitable for manual defect detection.

Structural and morphology characterization
The schematic synthesis process for the MAPbBr 3 @PbBr(OH) is illustrated in Fig. 1. The pH value of N,N-dimethylformamide (DMF) was preadjusted to 9.0 through dropwise addition of ammonium hydroxide; a mixture of PbBr 2 and MABr with a mole ratio of 1.05:1 was dissolved in DMF solvent under continuous stirring until a white precipitate was formed. Subsequently, the precipitate was placed into an oven at 70°C to obtain dried MAPbBr 3 perovskite, as shown in the middle of Fig. 1 (Fig. 2c) reveals that the diffraction peaks slightly shift to higher angles by~0.1°compared with the standard data of PbBr(OH), which might be a result of the lattice mismatch between PbBr(OH) and MAPbBr 3 . To further investigate the structural evolution of the samples before and after adding water, scanning electron microscopy (SEM) images and transmission electron microscopy (TEM) images were taken. Before adding water, MA-d is a typical cubic structure (Fig. 2d), which matches well with its XRD results. To further explore the microscopic morphology of the samples, the MA-d powders were broken by an ultrasonic cell crusher in toluene. In the TEM image of the crushed MA-d ( Supplementary Fig. 4), monodisperse and uniform QDs (10 nm in size) with cubic shapes can be observed, and the QDs tend to agglomerate on the TEM grids, as previously reported 44 Fig. 2f shows the corresponding laser confocal fluorescence microscope image, and uniform green fluorescence can be observed along the rod structure. Elemental mapping of MA-h is shown in Fig. 2g, and the elements O, Pb, and Br are uniformly distributed along the rod structure, which indicates the formation of PbBr(OH). To explore the inner structure of the rod, the MA-h powders were broken in a cell crusher. Some spherical QDs with an outer shell can be observed (Fig. 2h), and the lattice spacing of the QDs is 0.29 nm, corresponding to the d-spacing of the (200) crystal planes of MAPbBr 3 (Fig. 2i). The above results confirm that the MAPbBr 3 QDs were coated by PbBr(OH). Thus, one can conclude that the bulk cubic shape of MAPbBr 3 changes to a rod-like shape through the addition of water, and the MAPbBr 3 QDs are embedded into PbBr(OH). In addition, X-ray photoelectron spectroscopy (XPS) spectra were collected to detect the surface chemistry of MA-d and MA-h, as shown in Supplementary Fig. 6. The Br/Pb atomic ratios calculated from the XPS spectra are summarized in Supplementary Table 2. From the table, one can see that the Br/Pb atomic ratios of MA-d and MA-h are 3.06 and 1.11, respectively, confirming that these materials are MAPbBr 3 and PbBr(OH).

Photophysical property
To evaluate the stability of the as-prepared perovskites, a repetitive hydrate-dehydrate cycle test was carried out, and the change in the PL intensity was monitored, as shown in Fig. 3a. The PL intensity of the MA-d perovskite increases instantly with the addition of water, accompanied by a color change from yellow to greenish. The perovskite gradually turns back to a yellow color after drying at 60°C. As the hydrating-dehydrating cycle repeats, the PL intensity of MA-h decreases, while the color of the sample gradually stabilizes at greenish under ambient conditions. Additionally, the PL intensity of MAd increases after several hydrating-dehydrating cycles, with its yellow color changing to a greenish color. The corresponding evolution of the PL spectra and the color of both samples are shown in Supplementary Figs. 7 and 8. In the first round, the PL QY of MA-d is only 2.5% and increases to 71.54% after the addition of water. The sample is held in water for 1 year, and the PL QY of the MA-h is still at 89.9% of its initial value, i.e., 64.28%, which is one of the most stable fluorescence perovskite powders ever reported to the best of our knowledge. A slight hypsochromic shift of PL is observed during the hydrating-dehydrating cycles, as evidenced by the normalized PL spectra shown in Fig. 3b. The continuous hypsochromic shift in the PL spectra might be a consequence of the decrease in the perovskite QD size during cycling. In addition, traditional LHPs are highly sensitive to polar solvents due to their inherent ionic crystal structure, and the optical properties and even structural integrity of these materials usually deteriorate in polar organic solvents 30 . However, in our work, the MA-h powders show ultrastability to many common organic solvents, such as DMSO, dimethyl formamide, ethanol, acetone, ethylacetate, and benzene. The PL spectra of the MA-h powders immersed in these organic solvents are shown in Fig. 3c, d, and the corresponding images under UV illumination are shown in the insets of Fig. 3c, d. The emission of the MA-h powders is stable in the different organic solvents, and no obvious emission peak shift is observed. In Supplementary Fig. 9, no obvious structural or morphological changes are observed in DMF or DMSO, indicating that PbBr(OH) is almost insoluble in common solvents, which is why MA-h can maintain good stability. Thermal stability is another important indicator for luminescent materials 45 , especially for LHPs. Thermoinduced particle regrowth would induce PL quenching, and the crystal structure of the perovskites would collapse directly under exceedingly high temperatures. Additionally, high temperature might accelerate the rates of oxidation and hydration, which means that the oxygen-and moisture-induced decomposition would be amplified, leading to more rapid PL quenching 30 . Fig. 3e shows the thermal stability of MA-h in the temperature range from 20°C to 180°C, and the corresponding PL spectra are shown in Supplementary Fig. 10. It is worth noting that the entire test process was carried out in an autoclave filled with water. It can be clearly seen that the PL intensity of MA-h decreases slowly as the temperature increases in the range of 20-100°C, indicating good thermal stability. Upon heating above 100°C, the PL intensity decreases sharply and eventually vanishes when the temperature reaches 180°C. The PL decrease is due to the destruction of dense matrix PbBr(OH), and the corresponding XRD and SEM images are shown in Supplementary Fig. 11. In addition, the PL intensity of MA-h is still at 78.7% of the initial intensity upon heating from 20°C to 100°C. The photostability of MA-h was also investigated. Figure 3f shows the PL spectra of MA-h under continuous irradiation with a UV lamp (365 nm, 0.15 mW/cm 2 ), and the corresponding PL spectra are shown in Supplementary Fig. 12. The intensity decreases sharply during the first 24 h and maintains 50% of the initial intensity after 60 h of UV illumination. Supplementary Fig. 13 shows the XRD patterns and SEM images of MA-h under UV irradiation, and no obvious structural or morphological changes can be observed. According to previous studies, a possible reason for the PL decrease under UV illumination is photoassisted ionization or photoinduced defects 46,47 . The above results indicate that MA-h has good ambient, thermal, and photostability.

Mechanism for enhanced efficiency
To further investigate the high PL QY and the luminescence mechanisms of the as-prepared perovskite in the presence of water, the steady-state PL spectra, absorption spectra, time-resolved spectra, and temperaturedependent PL spectra of the as-prepared perovskite were recorded, as shown in Fig. 4. The PL, PL excitation (PLE), and absorption spectra of the as-prepared MA-d perovskite are shown in Fig. 4a, and the samples have strong absorption at 516 nm. Figure 4b shows the PL, PLE, and absorption spectra of MA-h, from which semblable PL and excitation characteristics can be observed, while obvious differences are observed in the absorption spectra. Compared with MA-h, MA-d has a significant absorption peak at 516 nm that absorbs the emissions of MA-d. The absorption peak centered at 310 nm with an absorption edge of 360 nm can be ascribed to the absorption of PbBr(OH). To further understand the luminescence mechanism of MA-h, temperature-dependent PL measurements were carried out with temperatures from 10 K to 300 K. In Fig. 4c, the PL peak (523 nm) at 10 K is as narrow as 6.8 nm, which originates from the emission of a typical strong excitonic   31 . With the increasing temperature, the intensity of the peak centered at 530 nm (at 10 K) gradually decreases and ultimately disappears at 60 K. Additionally, a new emission peak appears in the shorter wavelength range that corresponds to free exciton emissions. By tracking the PL evolution versus temperature, we clearly reveal that the PL of MA-h at room temperature is dominated by free exciton emissions. The lifetimes of the as-prepared perovskites with and without water were recorded, as shown in Fig. 4d Figure 4f shows the 3D plot of the PL spectra at various times, revealing an essentially time-independent spectral shape of PL emissions. In addition, the FWHM and emission peak are recorded in Supplementary Fig. 14  To further investigate the formation mechanism of MAh in the presence of water, we studied the morphological evolution of MA-h, and a possible mechanism was proposed. The morphological evolution of MA-h during the drying process was observed by optical microscopy and fluorescence microscopy. In Supplementary Fig. 15, the MA-h self-assembles into a cubic structure with loss of water, and then the rod structure forms again after water is added. Figure 4g shows a schematic illustration of the formation process of the as-prepared sample in the presence of water. Initially, yellow bulk perovskite with NH 4+ and OH − attached to the surface is produced. With the addition of water, the yellow bulk perovskite is destroyed and decomposes from the outside to inside, and PbBr (OH) is simultaneously formed with the aid of OH − during this process. A portion of the decomposed small crystals (QDs) are wrapped into rod-shaped PbBr(OH) by oriented self-assembly. The sample exhibits improved PL QY and ultrahigh stability due to the formation of MAPbBr 3 QDs embedded into the PbBr(OH) matrix. With the volatilization of water during the drying process, PbBr(OH) reacts with precipitated MABr (MABr is soluble in water) to form cubic perovskite. Correspondingly, the color of the sample changes from greenish to yellow, and the fluorescence intensity decreases after drying. Subsequently, each cycle (adding water and drying) produces a similar phenomenon until all the MABr is taken away by water and the perovskite QDs are entirely coated by PbBr(OH). Therefore, the color of the sample gradually changes to greenish in ambient light, and ultrastable MAPbBr 3 @PbBr(OH) is formed. The energy level of the MAPbBr 3 QDs and PbBr(OH) was studied by ultraviolet photoelectron spectroscopy (UPS), which can provide both the ionization potential and the valence band (VB) level of the sample, as well as its work function (φ). By combining UPS (Supplementary Fig. 16) and absorption spectra, we can establish the energy level. In Fig. 4h, the CB of the matrix (PbBr(OH)) is higher than that of the QDs, while the VB of the matrix is lower than that of the QDs. As a result, the surface defects are passivated by PbBr(OH), and both electrons and holes are confined in the QDs. The band alignment between the inner QDs and the outer matrix can guarantee exciton generation and high-rate radiative recombination of the QDs, thus resulting in a sharp PL QY.

Application to LEDs
Interestingly, this strategy is universal to other bromide perovskites, including FAPbBr 3 perovskites and CsPbBr 3 . The PL spectra of FAPbBr 3 are shown in Fig. 5a. The corresponding optical properties, structure, and morphology are shown in Supplementary Figs. 17 and 18.  Figure 5b shows the PL spectra of CsPbBr 3 , and the corresponding optical properties, structure, and morphology are shown in Supplementary Figs. 19 and 20. In addition, MAPb(Br/I) 3 @Pb(Br/I)(OH) was prepared by this method, and the corresponding spectrum and image are shown in Supplementary Fig. 21. The fluorescence peak redshifts to~630 nm, indicating that the emissions can be tuned by changing the X site anion. Notably, this method is simple and facile, and the samples can be prepared on a large scale by proportionally enlarging the precursors, as shown in Fig. 5c. Three grams of the powders are obtained in one synthesis process, and no obvious differences are observed from batch to batch. A Chinese character was written in the samples using water as ink; the regions with water are greenish in color and show bright green fluorescence under UV illumination. As discussed above, compared to traditional perovskites, synthetic powders have improved ambient, thermal, and photostabilities, and the reported perovskites with improved stability are summarized in Table 1. High PL QY and stability can be achieved simultaneously for the first time in this work. Based on these merits, this material can be employed as a promising phosphor in LEDs. UVpumped LEDs were fabricated by coating the MA-h and PDMS mixture onto 365 nm UV chips. Bright green light with color coordinates of (0.21, 0.52) can be observed ( Supplementary Fig. 22), and the emission intensity of the LEDs increases as the drive current increases (Supplementary Fig. 23). The light output of the LED versus time was measured, as shown in Supplementary Fig. 24. The initial luminous intensity of the LED is 1061 cd/m 2 , and the intensity decreases with running time. The luminous intensity of the LED settles at~400 cd/m 2 after 480 min (8 h) of running. The luminescence intensity of the asprepared LEDs can reach 9549 cd/m 2 when the driving current increases to 200 mA. The corresponding EL spectra of the LEDs under different drive currents from 10 mA to 200 mA are shown in Fig. 5d. In addition, an LED with bright green light can improve the contrast of illuminated objects due to the sensitivity of human eyes to green light. Thus, the as-prepared LEDs are suitable for manual defect detection. As illustrated in Fig. 5e, the light of the LED shines on the surface of the object; thus, the blemishes on the surface of the object (such as scratches, dust, and fingerprints) can be clearly observed. The different defects on the surface of substrates under illumination with a white LED and the as-prepared LED are shown in Supplementary Figs. 25 and 26. Under irradiation with green light, the fingerprint on the glass is extracted (Fig. 5f), and the fingerprint can be observed clearly. The corresponding pseudocolour map and gray values along the lines in Fig. 5f (bottom images) further indicates the resolution of the image.
In summary, we synthesized bromide-based perovskites whose PL QY can be increased from 2.5% to 71.54% by introducing water and decreases minimally in aqueous solution in 1 year. In addition, the as-synthesized MAPbBr 3 @PbBr(OH) can maintain their luminescence in many kinds of solvents and also exhibit excellent ambient, thermal, and photostabilities. The enhanced

Synthesis of methylammonium bromide
Methylammonium bromide (CH 3 NH 3 Br) was prepared by slowly mixing methylamine with HBr at a 1:1 molar ratio under continuous stirring for 2 h at 0°C. CH 3 NH 3 Br was then crystallized by removing the solvent from an evaporator. Then, the CH 3 NH 3 Br was washed with diethyl ether three times. White powders were obtained by recrystallization with ethanol. Subsequently, the powders were dried in vacuum for 24 h and stored in a dark and dry environment for further use.

Synthesis of MAPbBr 3 perovskites
The pH of DMF was preadjusted with ammonium hydroxide, and the pH was~9. Then, 2.31 g PbBr 2 and 0.67 g MABr (the molar ratio of PbBr 2 to MABr was 1.05) were dissolved in 6 ml alkaline DMF mixed solution to form a white viscous solution, and then the solution was continuously stirred until a white-colored precipitate was formed. Subsequently, the precipitate was centrifuged at 7500 rpm, and the precipitate was placed into an oven at 70°C to obtain MAPbBr 3 perovskite powders. MAPbBr 3 can be converted to bright emissive MAPbBr 3 @PbBr(OH) through the addition of water.

Synthesis of MAPbBr 3 @PbBr(OH)
The synthesis details of MAPbBr 3 @PbBr(OH) are as follows: 1 g of the as-prepared MAPbBr 3 powder was added to 10 ml of water, and then the mixture was stirred for 1 min. The process was replicated several times until the yellow MAPbBr3 powder transformed into greenish MAPbBr 3 @PbBr(OH) powder.

Fabrication of LEDs
First, the as-prepared MA-h phosphors were uniformly mixed with PDMS. The mixture of MA-h and PDMS was debubbled, and coated onto a 365 nm UV chip and then cured in an oven at 70°C for 1 h to obtain UVpumped LEDs.

Characterization
The XRD patterns were obtained by an X'Pert Pro diffractometer. The TEM images were taken by a transmission electron microscope (JEM-2010), and the SEM images were obtained on a JEOL JSM6700F field-emission scanning electron microscope. The Fourier transform infrared (FTIR) spectra of the samples were recorded on a Thermo Scientific Nicolet iS10 FTIR spectrometer. The PL QYs of the samples were measured by an Edinburgh fluorescence spectrometer (FLS980). The X-ray photoelectron spectra of the samples were collected by a Thermo Fisher Scientific ESCALAB 250Xi spectrometer equipped with an Al Ka X-ray radiation source, and the XPS binding energy was internally referenced to the C 1 s peak (BE = 284.8 eV). The PL spectra were measured by a Hitachi F-7000 spectrophotometer, and the UV-vis absorption spectra were characterized by a UH4150 spectrophotometer.

Computational methods
All the first-principle calculations were performed utilizing the plane-wave pseudopotential approach within density functional theory as implemented in the Vienna Ab initio Simulation Package. The 1 s for H, 2s 2 2p 2 for C, 2s 2 2p 3 for N, 2s 2 2p 4 for O, 6s 2 6p 2 for Pb, and 4s 2 4p 5 for Br were treated as valence electrons, and the interactions between nucleus and valence electrons were described by the projected augmented wave method. The electronic wave functions were expanded in plane-wave basis sets with a kinetic energy of 400 eV, and Brillouin zone sample meshes with 2π × 0.02 Å -1 were utilized to ensure the energy convergence of our calculations.