Effective Control of the Growth and Photoluminescence Properties of CsPbBr3/Cs4PbBr6 Nanocomposites by Solvent Engineering.

Metal halide perovskites exhibit small exciton binding energy, which leads to a low electron-hole capture rate for radiative recombination and accordingly decreases the luminescence efficiency. Reducing the thickness of the perovskite film or the size of the perovskite crystal is found to be an effective method to spatially confine the electrons and holes to promote the bimolecular radiative recombination. Here, we fabricate CsPbBr3/Cs4PbBr6 nanocomposites, applicable for light emission diodes, by a simple self-assembly method. We effectively reduce the critical size of the CsPbBr3 nanocrystals in the CsPbBr3/Cs4PbBr6 nanocomposites by adding a certain amount of dimethyl sulfoxide into the perovskite precursor solution. Accordingly, the photoluminescence quantum yield of the CsPbBr3/Cs4PbBr6 nanocomposites increased from 56 to 91% due to the quantum size effect. In situ observation of the growth of CsPbBr3/Cs4PbBr6 nanocomposites reveals that the reduction of the CsPbBr3 crystal size is due to the change of the chemical reaction speed during the two-step growth process of the CsPbBr3/Cs4PbBr6 nanocomposites.

Introduction

Metal halide perovskites attract great attention owing to their facile solution processing, high color purity, tunable bandgap, and large and balanced electron and hole mobilities.[1−6] These unique properties make them promising in photoelectronics, such as color-conversion phosphors in white light emission diodes (LEDs), electroluminescent materials in LEDs, and as gain materials in lasers.[7−9] In recent years, the external quantum efficiency (EQE) of perovskite LEDs has reached more than 20%.[10−12]

Two strategies have been developed for fabricating perovskite emission devices i.e., (1) spin-coating and (2) deposition of bulk perovskite films. In the first method, highly luminescent presynthesized colloidal perovskite nanocrystals (NCs) are utilized to deposit the film by the spin-coating technique. In the second approach, perovskite precursor solutions are required for the deposition of bulk perovskite films. In both the strategies, decreasing the grain size of perovskite materials is important due to the small exciton binding energy in three-dimensional perovskites. Thus, reducing the grain size of ultrathin perovskite films or small perovskite NCs can spatially confine the electrons and holes to promote the electron–hole capture rate of radiative recombination. Wang et al. synthesized thin granular CsPbBr3 films by introducing trifluoroacetate anions such that an external quantum efficiency (EQE) of 10.5% was achieved.[13] In addition, the value of EQE is enhanced due to the confining growth rate of low leveled crystallites (∼10 nm) by adding ammonium halides to the perovskite precursor solution.[14]

Like controlling grain size, surface passivation is another important method to reduce nonradiative recombination to enhance the EQE of the device. Lin et al prepared the CsPbBr3/MABr quasi-core/shell structure film. The MABr shell effectively passivated the nonradiative defects and improved the performance of the device.[10] In the case of the perovskite-nanocrystal-based device, Zeng and co-workers controlled the perovskite nanocrystal surface through employing organic–inorganic hybrid ligands and demonstrated CsPbBr3 nanocrystal-based LEDs with an impressive peak EQE of over 16%.[15] The perovskite NCs can also be passivated by coating a shell material on the surface, such as poly(methyl methacrylate), polystyrene, and silica.[16−19] In our previous work, we prepared CsPbBr3/Cs4PbBr6 nanocomposites by a simple self-assembly method, in which Cs4PbBr6 played an important role in passivation and developed a white LED using the nanocomposites.[20]

In this paper, we effectively reduced the critical size of the CsPbBr3 NCs in the CsPbBr3/Cs4PbBr6 nanocomposites by adding an appropriate amount of dimethyl sulfoxide (DMSO) into the perovskite precursor solution. Utilizing the effect of DMSO on the solubility of lead (II) bromide (PbBr2) and cesium bromide (CsBr) in the solvent, the growth of CsPbBr3/Cs4PbBr6 nanocomposites can be precisely controlled. Accordingly, the photoluminescence quantum yield (PLQY) of the CsPbBr3/Cs4PbBr6 nanocomposites increased from 56 to 91% by reducing the size of embedded CsPbBr3 in host Cs4PbBr6 materials.

Results and Discussion

Microstructures and Optical Properties of CsPbBr3/Cs4PbBr6 Nanocomposites

The microstructures and absorption and photoluminescence (PL) spectra of CsPbBr3/Cs4PbBr6 nanocomposites were investigated. Figure S1 shows the microstructures and optical properties of CsPbBr3/Cs4PbBr6 nanocomposites formed in pure N,N-dimethylformamide (DMF). high-resolution transmission electron microscopy (HRTEM) images show the matrix structure clearly, with CsPbBr3 nanocrystals embedded in Cs4PbBr6 host materials (Figure S1a). The absorption spectrum exhibits a strong peak at 315 nm and an absorption edge at 505 nm (Figure S1b). The PL spectrum displays a weak peak at 340 nm and a strong narrow peak at 517 nm (Figure S1b). As discussed in previous work, the PL peak near 340 nm is related to Cs4PbBr6 and the strong PL peak near 517 nm originates from embedded CsPbBr3 NCs.[20]

The formation mechanism of the nanocomposites also follows the model proposed previously that it is grown by a two-step chemical reaction driven by the difference of ion concentrations of Cs+ and Pb2+ ions.[20] The solubility of CsBr is very low in the solvent of N,N-dimethylformamide (DMF), whereas it is much higher for PbBr2.[7] Therefore, the ion concentration of Pb2+ is much higher and the Pb-rich phase of CsPbBr3 is formed in the first stage when CsBr and PbBr2 (CsBr/PbBr2 = 4:1, molar ratio) are loaded into DMF solvent. With the consumption of PbBr2, Cs+ concentration is higher in the second stage and therefore CsPbBr3 is converted into the Cs-rich phase of Cs4PbBr6 and CsPbBr3/Cs4PbBr6 nanocomposites are produced when newly formed Cs4PbBr6 are grown on the surface of CsPbBr3 nanocrystals.

Effect of Solvent on the Growth of CsPbBr3/Cs4PbBr6 Nanocomposites

We first investigated the coordination and solubility behavior of the reactants in DMF and DMSO. Figure a displays the evolution of the X-ray diffraction (XRD) patterns of the solids obtained from the solvent with different DMSO ratios. To conveniently depict the ratio of DMSO and DMF, we named the mixed solvent DMSO-X in the following discussions, where X represents the particular volume of DMSO in the mixed solvent with a volume of 2 mL. In the case of DMSO-0, all of the diffraction peaks are associated with PbBr2(DMF).[21] The peaks of PbBr2(DMF) disappear upon increasing the DMSO ratio from 0% (DMSO-0) to 7.5% (DMSO-0.15), whereas the peaks associated with the PbBr2(DMSO)2 complex appear, and upon further increasing the ratio of DMSO to 15% (DMSO-0.3), there is no change observed in DMSO-0.15.[21] Compared to PbBr2(DMF), the complexity of PbBr2(DMSO)2 is preferentially formed in the mixed system of DMSO and DMF. Except for the larger coordination number in PbBr2(DMSO)2, the Pb–O bond length is 2.386 Å in PbBr2(DMSO)2, which is smaller than the value of 2.431 Å in PbBr2(DMF), indicating that DMSO has a stronger coordination ability with PbBr2 than DMF.[21] The appearance of a characteristic absorption peak of PbBr2(DMSO)2 and the blue shift of the absorption edge in solvent DMSO-0.08 (Figure S2) indicates that the PbBr2(DMSO)2 coordination is first formed in the solution.[21] In addition, we also studied the coordination behavior of CsBr in the mixed solvent using the same experimental method. It can be seen from Figure S3 that all of the diffraction peaks correspond to pure CsBr, which means that CsBr cannot coordinate with DMSO and DMF. However, the dissolution experiments show that the solubility of CsBr is gradually increased with increasing DMSO ratio (shown in Figure b). The different coordination behavior of PbBr2 and dissolution behavior of CsBr in the mixed precursor solutions must have an important influence on the chemical reaction during the preparation of CsPbBr3/Cs4PbBr6 nanocomposites.

Figure 1

(a) XRD patterns of pure PbBr2 or complex formed in different solvents. The data are obtained by sequentially dissolving PbBr2 into the DMF/DMSO solvent, evaporating the solvent, and measuring the crystal structure of the solids obtained. (b) Solubility curve of CsBr in mixed DMF–DMSO solution.

We investigate the effect of solvent on the growth of CsPbBr3/Cs4PbBr6 nanocomposites by in situ observing the growth of CsPbBr3/Cs4PbBr6 nanocomposites in the solvent of DMSO-X and characterizing the products obtained with different reaction times. Figure shows the photographs taken at different times during the reaction process and the XRD patterns of the solid products obtained by centrifuging the corresponding dispersion. Here, the sample (DMSO-0) exhibits an orange-colored dispersion and precipitates with a reaction time of 3 min. XRD data obtained for the precipitate reveal that the product consists of CsPbBr3 and undissolved CsBr (Figure b). Upon further increasing the reaction time to 5 min, a light yellow precipitate starts to appear at the bottom of the tube associated with green emission under ultraviolet (UV) irradiation and is shown in Figure c. The luminescent precipitate is determined to be CsPbBr3/Cs4PbBr6 nanocomposites based on the XRD results (shown in Figure d) and the research in our previous work.[20] The production of CsPbBr3/Cs4PbBr6 nanocomposites is enhanced, and no observable diffraction peaks of CsPbBr3 have been noticed in the XRD patterns, when the reaction time is more than 30 min. The results demonstrate the two-step formation process of CsPbBr3/Cs4PbBr6 nanocomposites, as revealed in our previous work.[20] By comparing the photographs and XRD patterns for the samples prepared with different DMSO contents but with the same reaction time, we find that the DMSO can effectively accelerate the transformation speed from CsPbBr3 to Cs4PbBr6. A drastic color transition from the orange CsPbBr3 precipitate to the light yellow CsPbBr3/Cs4PbBr6 precipitate is observed for the case of DMSO-0.15 when the reaction time is within 3 min, which is faster than DMSO-0. The phase transition occurs even faster, and there is no CsPbBr3 observed in the photograph and XRD patterns after 3 min, when the content is increased to DMSO-0.3. The improved transformation speed is ascribed to the increased solubility of CsBr by DMSO.

Figure 2

Photograph of the product (under natural light (top) and UV light (bottom)) when the reaction time is 3 min (a), 5 min (c), and 30 min (e) in the case of DMSO-0, DMSO-0.15, and DMSO-0.3. XRD of the product when the reaction time is 3 min (b), 5 min (d), and 30 min (f) in the case of DMSO-0, DMSO-0.15, and DMSO-0.3; the purple diamonds, orange stars, and black dots represent characteristic peaks of Cs4PbBr6, CsPbBr3, and CsBr, respectively.

To further reveal how the solvent affects the growth of CsPbBr3/Cs4PbBr6 nanocomposites, we have reduced the speed of the chemical reaction by removing the ultrasonication and observed the first-step reaction process. In Video S1, PbBr2 solution dissolved in DMSO-X (X = 0, 0.1, 0.2, 0.3 from left to right, respectively) was added into the vessel with preloaded CsBr powders. The ratio of CsBr-to-PbBr2 was kept at 4:1, and the contents of the reactants were the same for each one. It can be seen that the chemical reaction between the CsBr and PbBr2 complex occurred and an orange precipitate was formed at the bottom sequentially from left to right after the addition of PbBr2 solution. XRD measurement of the orange precipitate (Figure S4) confirms that the precipitate composes of CsPbBr3 and unreacted CsBr. It indicates that DMSO can effectively retard the formation of CsPbBr3 during the first-step reaction. The aforesaid results suggest that the reduced speed is ascribed to the stronger coordination between PbBr2 and DMSO than DMF, which retards the release of PbBr2 and is important for the chemical reaction between PbBr2 and CsBr.

To confirm the explanation above, we keep on increasing the contents of DMSO during the fabrication of CsPbBr3/Cs4PbBr6 nanocomposites and see how it will affect the final product. Figure S5a displays the evolution of the XRD patterns of the final products with different DMSO ratios. When the ratio is below DMSO-1, the XRD patterns show similar diffraction peaks, and all of the diffraction peaks correspond to Cs4PbBr6 except for a small diffraction peak at 29.5° from CsBr (denoted by the dot). With the ratio being above DMSO-1, the diffraction intensity of CsBr increases obviously. By considering the fixed value of contents of the reactant in all cases, there must be abundant unreacted PbBr2 remaining in the solution with higher DMSO contents. To verify this, we collected the supernatant after the reaction was complete in the cases of DMSO-1, DMSO-1.5, and DMSO-2, evaporated the solvent, and measured the XRD patterns of the solids obtained. The results in Figure S5b show two new diffraction peaks indexed to the PbBr2(DMSO)2 complex (located at 10.8 and 14.3°) in the cases of DMSO-1.5 and DMSO-2. This supports the hypothesis well that the DMSO retards the release of PbBr2 and suppresses the chemical reaction between PbBr2 and CsBr.

To understand the influence of DMSO on CsPbBr3/Cs4PbBr6 nanocomposites, we examined the morphology of the nanocomposites by using TEM. Figure shows the TEM images of the CsPbBr3/Cs4PbBr6 nanocomposites prepared in the solvents of DMSO-0 and DMSO-1. Both matrix structures clearly show that CsPbBr3 nanoparticles are embedded in the host Cs4PbBr6 materials. The average size of CsPbBr3 is 20 nm in the case of pure DMF, whereas it is reduced to 6 nm in the case of DMSO-1. The average size is 8 nm for DMSO-0.5 (not shown). Reduction in the nanoparticle size indicates that the controlled growth by DMSO strongly influences the microstructure of CsPbBr3/Cs4PbBr6 nanocomposites.

Figure 3

TEM image of CsPbBr3/Cs4PbBr6 nanocomposites prepared in the solutions of (a) DMSO-0 and (b) DMSO-1.

Based on the above discussion, we summarize that DMSO delays the formation of CsPbBr3 (step 1) but it accelerates the transformation from CsPbBr3 to Cs4PbBr6 (step 2) during the preparation of CsPbBr3/Cs4PbBr6 nanocomposites. Such modifications significantly influence the critical size of the CsPbBr3 nanoparticles. This effect can be well illustrated by the model below (Figure ) based on the two-step growth theory. In the first step, the CsPbBr3 crystals are formed with a typical size of micrometers due to the higher Pb2+ concentration. The concentration of Cs+ is inversely higher at the second stage with the consumption of Pb2+. Then, the CsPbBr3 crystals start to dissolve and react with CsBr to produce Cs4PbBr6. The CsPbBr3/Cs4PbBr6 nanocomposites are produced when the Cs4PbBr6 crystals are grown on the surface of CsPbBr3 nanoparticles, which are not fully dissolved.[20] Due to the strong coordination between PbBr2 and DMSO, the release of PbBr2 from the coordination state PbBr2(DMSO)2 is difficult such that the formation speed of CsPbBr3 particles is decelerated when DMSO is included. Moreover, the transformation of CsPbBr3 to Cs4PbBr6 is accelerated, induced by the higher concentration of Cs+ due to the higher solubility of CsBr in DMSO. Therefore, the size of CsPbBr3 NCs embedded in host Cs4PbBr6 materials is smaller with the inclusion of DMSO.

Figure 4

Schematic illustration for the formation mechanism of CsPbBr3/Cs4PbBr6 nanocomposites in the cases of DMSO-0 (a) and DMSO-1 (b).

Effect of Solvent on the Optical Properties of CsPbBr3/Cs4PbBr6 Nanocomposites

Figure shows the absorbance and PL spectra of the CsPbBr3/Cs4PbBr6 nanocomposites prepared in DMSO-X. As the DMSO ratio is increased, the intensity of the absorption peak around 505 nm related to CsPbBr3 NCs gradually decreases and a blue shift of the absorption edge is also observed in Figure a. The PL spectrum (Figure b, excitation: 295 nm) indicates that the products exhibit a strong narrow emission peak at around 517 nm (full width at half-maximum = 20 nm). A blue shift of the PL spectrum appears as the DMSO ratio is increased, which is consistent with the absorption spectrum. These phenomena are ascribed to the quantum confinement effect of CsPbBr3 NCs, consistent with the TEM results. Figure c shows that the introduction of DMSO-X can significantly enhance the PLQY of CsPbBr3/Cs4PbBr6 nanocomposites. The maximum value of PLQY in the case of DMSO-1 reaches 83%. The reduction of CsPbBr3 size can spatially confine electrons and holes,[22,23] which effectively improves its fluorescence efficiency. When the content of DMSO is higher than 50% (DMSO-1), the PLQY starts to decrease. We suggest that it is induced by the unreacted CsBr remaining in the product since the content of CsBr is obviously increased when the content is higher than DMSO-1 in Figure S5a. To verify it, we intentionally reduced the amount of CsBr reactant to the ratio of 3.7 (CsBr/PbBr2) in the case of DMSO-1. Figure S6 shows the XRD patterns of the products for the ratios of 3.7 and 4. The characteristic peaks of CsBr disappeared, and all of the diffraction peaks are indexed to Cs4PbBr6 with the ratio of 3.7. The PLQY of the CsBr-free CsPbBr3/Cs4PbBr6 nanocomposites was measured to be 91%, which is much higher than the CsPbBr3/Cs4PbBr6 nanocomposites coexisting with the CsBr remnant. This confirms that the lowered PLQY for high DMSO contents is induced by the excessive CsBr remnant. By reducing/removing the CsBr remnant, the CsPbBr3/Cs4PbBr6 nanocomposites (DMSO-1.5) are also purified and the PLQY is optimized to be 88% when the CsBr/PbBr2 ratio is 3.5. The corresponding XRD patterns and PLQY of each sample are shown in Figure S7 and Table S1 in the Supporting Information. By comparing the PLQYs of DMSO-1 and DMSO-1.5 samples, we suggest that the ratio of CsPbBr3 to Cs4PbBr6 is probably another factor affecting the PLQY of the nanocomposite since the CsPbBr3 size of DMSO-1.5 is even smaller but exhibits a lower PLQY. Further research work is needed to confirm this in the future.

Figure 5

(a) UV–vis absorption spectra, (b) PL spectra, and (c) PLQY of the product obtained in mixed DMF–DMSO solution.

Conclusions

We fabricate CsPbBr3/Cs4PbBr6 nanocomposites by a simple self-assembly method. CsPbBr3/Cs4PbBr6 nanocomposites are formed by a two-step reaction driven by the ion concentration. CsPbBr3 is formed first in the Pb-rich environment, and it converts to Cs4PbBr6 in the second step when the solution is Cs-rich, which leads to the formation of CsPbBr3/Cs4PbBr6 nanocomposites. When the DMSO is introduced to the perovskite precursor solution, it decelerates the first reaction speed due to the strong coordination between Pb2+ and DMSO and accelerates the second step owing to the larger solubility of CsBr in DMSO than in DMF. Consequently, the critical size of CsPbBr3 NCs is obviously decreased in the CsPbBr3/Cs4PbBr6 nanocomposites and the PLQY is increased from 56 to 91%.

Experimental Section

Materials

Commercially available lead (II) bromide (PbBr2 99.999%), cesium bromide (CsBr 99.999%), N,N-dimethylformamide (DMF), and dimethyl sulfoxide (DMSO) were purchased from Aldrich.

Synthesis of CsPbBr3/Cs4PbBr6 Nanocomposites

For this, 0.4 mmol PbBr2 was dissolved into the mixed solvent of DMF and DMSO with different volume ratios (the total volume is 2 mL) in a centrifuge tube with a volume of 5 mL. Then, the precursor is mixed with 1.6 mmol CsBr (CsBr/PbBr2 molar ratio of 4:1). The centrifuge tube was ultrasonicated for about 30 min until a light yellowish precipitate appears. The precipitate was collected via centrifugation (TGL-16G) at 5000 rpm for 5 min and then dried in a vacuum oven at 70 °C to obtain CsPbBr3/Cs4PbBr6 nanocomposite powders.

Characterization

The crystal structure was determined using an X-ray diffractometer (Bruker D8 Advance) with Cu Kα radiation (λ = 1.5406 Å). TEM images were captured on an FEI Tecnai G2 F30 transmission electron microscope operated at 300 kV. Absorption spectra were obtained using a UV–visible–near infrared (NIR) spectrophotometer (Shimadzu UV-3600). Photoluminescence (PL) spectra were measured on a PerkinElmer LS55 fluorescence spectrometer. Absolute PLQYs were measured by a Quantaurus-QY spectrometer (C11347-11, Hamamatsu, Japan) with an excitation wavelength of 450 nm.

Experiments on the Coordination between Reactants and Solvent

To investigate the coordination and dissolution characteristics of reactants in the mixed solvents, we sequentially dissolved 0.04 mmol PbBr2 powder into DMF/DMSO solvent, measured the UV–visible–NIR spectra of the solution, heated the solution at 60 °C for 24 h in a vacuum to evaporate the solvent, and investigated the crystal structure of the obtained solids by XRD. The coordination behavior of CsBr in the mixed solvent was investigated using the same experimental method.