Highly Stable and Photoluminescent CsPbBr3/Cs4PbBr6 Composites for White-Light-Emitting Diodes and Visible Light Communication

Inorganic lead halide perovskite is one of the most excellent fluorescent materials, and it plays an essential role in high-definition display and visible light communication (VLC). Its photochromic properties and stability determine the final performance of light-emitting devices. However, efficiently synthesizing perovskite with high quality and stability remains a significant challenge. Here, we develop a facile and environmentally friendly method for preparing high-stability and strong-emission CsPbBr3/Cs4PbBr6 composites using ultrasonication and liquid paraffin. Tuning the contents of liquid paraffin, bright-emission CsPbBr3/Cs4PbBr6 composite powders with a maximum PLQY of 74% were achieved. Thanks to the protection of the Cs4PbBr6 matrix and liquid paraffin, the photostability, thermostability, and polar solvent stability of CsPbBr3/Cs4PbBr6-LP are significantly improved compared to CsPbBr3 quantum dots and CsPbBr3/Cs4PbBr6 composites that were prepared without liquid paraffin. Moreover, the fabricated CsPbBr3/Cs4PbBr6-LP-based WLEDs show excellent luminescent performance with a power efficiency of 129.5 lm/W and a wide color gamut, with 121% of the NTSC and 94% of the Rec. 2020, demonstrating a promising candidate for displays. In addition, the CsPbBr3/Cs4PbBr6-LP-based WLEDs were also demonstrated in a VLC system. The results suggested the great potential of these high-performance WLEDs as an excitation light source to achieve VLC.


Introduction
With the intensification of competition since 2015, the large-scale application of semiconductor light-emitting diode (LED) devices in low-value fields such as lighting has seen industry profits drop below 10%. The key to solving this crisis is to develop the application of LED in high-value fields, such as high-definition displays and visible light communication [1,2]. Their technical requirements are mainly focused on high quality and reliability. The high quality of the devices depends on the photochromic performance of luminescent materials such as quantum dots and phosphors [3,4]. In contrast, the high reliability of devices is mainly affected by the stability of fluorescent materials, such as light, heat, and solvent stability [5].

Synthesis of CsPbBr3/Cs4PbBr6 Microcrystals
The highly green-emitting CsPbBr3/Cs4PbBr6 microcrystals were prepared according to our previous method by the facile ultrasonic method shown in Figure 1. First, 14.58 mmol CsBr, 2.43 mmol PbBr2, 0.8 mL liquid paraffin (LP), and 3.2 mL DMSO were mixed in a glass bottle. Then, the hybrid solution was sonicated for 30 min at 90 W of ultrasound power using an ultrasonic processor. Subsequently, unreacted precursors were removed by centrifugation at a speed of 10,000 rpm for 5 min. Then, the residue was re-dissolved in 4.0 mL of n-hexane. After that, the solution was centrifuged at a rate of 12,000 rpm for 5 min. Finally, the green powders were obtained by vacuum drying. In addition, the effect of LP content on the properties of the prepared samples was investigated, and experiments were carried out as described above.

Synthesis of CsPbBr3 Quantum Dots with a Room-Temperature Supersaturated Recrystallization Strategy
As a reference, CsPbBr3 quantum dots (QDs) were synthesized using the method reported by Li and coworkers with modifications [32]. A total of 0.8 mmol CsBr, 0.8 mmol PbBr2, and 20 mL DMSO were added to a reagent bottle. Then, 2.0 mL OA and 1.0 mL OAm were quickly injected to stabilize the precursor solution. After that, 2.0 mL of the mixed solution were slowly dropped into 20 mL of n-hexane while vigorously stirring. Then, bright-green-emission CsPbBr3 QDs were achieved.

Fabrication of Perovskite-based WLED Devices
The perovskite-based WLED devices consist of a green-emitting CsPbBr3/Cs4PbBr6 film, a red phosphor film, and a commercial blue LED chip. The CsPbBr3/Cs4PbBr6 film was fabricated by adding the as-prepared CsPbBr3/Cs4PbBr6 solid powders to PDMS. The mixture was stirred with a vacuum homogenizer for 12 min to degas bubbles. Subsequently, the mixture was injected into a Teflon mold and heated at 120 °C for one hour. The red phosphor film was obtained using the above procedures. Eventually, the WLED

Synthesis of CsPbBr 3 Quantum Dots with a Room-Temperature Supersaturated Recrystallization Strategy
As a reference, CsPbBr 3 quantum dots (QDs) were synthesized using the method reported by Li and coworkers with modifications [32]. A total of 0.8 mmol CsBr, 0.8 mmol PbBr 2 , and 20 mL DMSO were added to a reagent bottle. Then, 2.0 mL OA and 1.0 mL OAm were quickly injected to stabilize the precursor solution. After that, 2.0 mL of the mixed solution were slowly dropped into 20 mL of n-hexane while vigorously stirring. Then, bright-green-emission CsPbBr 3 QDs were achieved.

Fabrication of Perovskite-based WLED Devices
The perovskite-based WLED devices consist of a green-emitting CsPbBr 3 /Cs 4 PbBr 6 film, a red phosphor film, and a commercial blue LED chip. The CsPbBr 3 /Cs 4 PbBr 6 film was fabricated by adding the as-prepared CsPbBr 3 /Cs 4 PbBr 6 solid powders to PDMS. The mixture was stirred with a vacuum homogenizer for 12 min to degas bubbles. Subsequently, the mixture was injected into a Teflon mold and heated at 120 • C for one hour. The red phosphor film was obtained using the above procedures. Eventually, the WLED devices were fabricated by coating green CsPbBr 3 /Cs 4 PbBr 6 film and red phosphor film layers on the surface of the blue LED chip.

Measurements of the Visible Light Communication System
The modulations of the bandwidths of the samples were tested by a facile visible light communication (VLC) system [33]. The transmitter of the VLC system mainly includes four parts: an arbitrary waveform generator (AFG, SIGLENT, SDG 6052X-E), a power amplifier (PA, mini circuits, ZHL-6A-S+), a direct-current power supply (DC, Keithley, 2231A), and a bias tee (mini circuit, ZFBT-4R2GW+). A sinusoidal radio frequency (RF) signal is generated by the AFG, which is combined with the DC bias using a bias tee to drive the 450 nm laser diode (LD, YuLiGuangZhou, 450 nm, 15 W). So far, the electric signal has been transformed into a modulated optical signal. Then, divergent and convex lenses were applied to scatter and collate the blue light from the LD. After that, the collimated light excited the green-emitting CsPbBr 3 /Cs 4 PbBr 6 film and red phosphor film to generate white light. A convex lens is used on the receiver side to gather the modulated white light from the transmitter. An avalanche photodiode (APD, Meno System APD210) was used to convert the optical signal into an electrical signal and amplify it. Finally, the output of the APD was further sampled by an oscilloscope (OSC, SIGLENT, SDS 2000X Plus), which recorded the signals for analysis.

Characterization
The products' X-ray diffraction (XRD) patterns were determined via an XRD-D8-ADVANCE (Bruker, Bremen, Germany) with a Cu-Kα radiation source. The surface morphology of the samples was characterized by a field-emission scanning electron microscope (FE-SEM, Merlin, Forchtenberg, Germany), and atomic-resolution chemical mapping was achieved using energy-dispersive spectroscopy (EDS) in the FE-SEM. Energy-dispersive X-ray (EDX) spectroscopy (JEOL, Tokyo, Japan) was also carried out to observe element distribution. X-ray photoelectron spectroscopy (XPS) spectra were obtained using a Thermo Scientific (Thermo K-Alpha) (Waltham, MA, USA) machine with a monoatomic Al-Kα excitation source (1486.6 eV). Absorption spectra were performed using a UV-Vis spectrometer (UV-Vis: Tu-1901, Purkinje, Beijing, China). Photoluminescence spectroscopy was implemented by an RF-600 fluorescence spectrophotometer (Shimadzu, Kyoto, Japan), using a xenon lamp as an excitation source. The PLQY measurement of the products was conducted on the Hamamatsu Quantum Yield Measurement System (C9920-02G, Hamamatsu, Japan) under an excitation wavelength of 365 nm. Time-resolved fluorescence spectra were collected by an FLS 980 fluorescence lifetime spectrofluorometer (Edinburgh Instrument, Edinburgh, UK). The PL decay curves obtained are fitted with the multiple exponential functions given in the expression below [30].
where A(t) represents the PL intensity at time t; A i denotes the relative weights of the lifetime components at time t = 0; and τ i represents the decay time for the lifetime components. The average decay lifetime τ avg. is calculated via the following expression [33]: Nanomaterials 2023, 13, 355 5 of 16

Structural Phase and Morphological Characterizations of As-Prepared Samples
To identify the crystalline phase of the as-prepared samples, XRD measurements were used. As shown in Figure [35]. Furthermore, we compared the as-synthesized samples with standard CsBr and PbBr 2 XRD spectra and observed whether these phases existed. After careful observation, soft peaks of unreacted PbBr 2 are observed. Furthermore, no other XRD patterns corresponding to CsBr or other perovskite compounds were detected. Overall, XRD characterization implies that this sample mainly includes CsPbBr 3 and Cs 4 PbBr 6 structural phases.
The FE-SEM characterization was performed to observe the morphological features of the samples. As shown in Figure 2b,c, the primary morphology of the sample reaches a micrometer scale. It presents a rhombohedral shape, and its outer surface is embedded with small particles, similar to the previously reported structure of CsPbBr3 embedded in Cs4PbBr6 crystals [30,36]. This result indicates that the sample is CsPbBr3/Cs4PbBr6 composites. In order to quantitatively determine the content of CsPbBr3 and Cs4PbBr6, EDX mapping and EDS characterization were conducted. As shown in Figure 2d, the Cs, Pb, and Br elements are uniformly distributed. Combined with the EDS spectrum, the molar ratio of CsPbBr3 and Cs4PbBr6 is 1:8.46, strongly confirming the co-existence of CsPbBr3 and Cs4PbBr6. To sum up, we successfully prepared CsPbBr3/Cs4PbBr6 composites.
.  The FE-SEM characterization was performed to observe the morphological features of the samples. As shown in Figure 2b,c, the primary morphology of the sample reaches a micrometer scale. It presents a rhombohedral shape, and its outer surface is embedded with small particles, similar to the previously reported structure of CsPbBr 3 embedded in Cs 4 PbBr 6 crystals [30,36]. This result indicates that the sample is CsPbBr 3 /Cs 4 PbBr 6 composites. In order to quantitatively determine the content of CsPbBr 3 and Cs 4 PbBr 6 , EDX mapping and EDS characterization were conducted. As shown in Figure 2d, the Cs, Pb, and Br elements are uniformly distributed. Combined with the EDS spectrum, the molar ratio of CsPbBr 3 and Cs 4 PbBr 6 is 1:8.46, strongly confirming the co-existence of CsPbBr 3 and Cs 4 PbBr 6 . To sum up, we successfully prepared CsPbBr 3 /Cs 4 PbBr 6 composites.
To investigate the optical properties of the CsPbBr 3 /Cs 4 PbBr 6 composites, UV-Vis absorption spectroscopy and fluorescence spectroscopy were performed, as shown in Figure 3a-c. The CsPbBr 3 /Cs 4 PbBr 6 powders demonstrate a strong absorption peak at about 311 nm and a broad absorption band with an absorption edge at 510 nm ( Figure 3a). The strong absorption peak located at about 311 nm is agrees well with previous reports on bulk Cs 4 PbBr 6 powders, further confirming that the isolated octahedral PbBr 6 4was formed in Cs 4 PbBr 6 [37]. Another broad absorption at 510 nm differs from previous reports in that the characteristic absorption band of CsPbBr 3 QD is usually located at 505 nm [38]. This result can explain, to a certain extent, the fact that CsPbBr 3 has been embedded in Cs 4 PbBr 6 , thereby affecting its absorption. The PLE spectrum of CsPbBr 3 /Cs 4 PbBr 6 also shows a difference from pure CsPbBr 3 or Cs 4 PbBr 6 . The fluorescence intensity of the PLE spectrum is relatively low in the short wavelength region (307-325 nm) when the PL emission peak is set at 518 nm and rapidly increases in the 325-345 nm region, extending to longer wavelengths. ports in that the characteristic absorption band of CsPbBr3 QD is usually located at 505 nm [38]. This result can explain, to a certain extent, the fact that CsPbBr3 has been embedded in Cs4PbBr6, thereby affecting its absorption. The PLE spectrum of CsPbBr3/Cs4PbBr6 also shows a difference from pure CsPbBr3 or Cs4PbBr6. The fluorescence intensity of the PLE spectrum is relatively low in the short wavelength region (307-325 nm) when the PL emission peak is set at 518 nm and rapidly increases in the 325-345 nm region, extending to longer wavelengths.
Further observation found that the PLE spectrum has a similar change with the absorption spectrum in the range of 345 nm to 505 nm, which is possibly related to the fact that CsPbBr3 absorbs the excitation photons and generates fluorescence; thus, a change in the absorption spectrum leads to a corresponding trend in the PLE spectrum. However, in the range of 307 nm to 325 nm, due to CsPbBr3 nanocrystals embedded in the Cs4PbBr6 matrix, the excitation photons are almost entirely absorbed by the Cs4PbBr6 matrix but not by the internal CsPbBr3 nanocrystals. This phenomenon is similar to the pure Cs4PbBr6 feature [39]. In addition, the PL spectrum of CsPbBr3/Cs4PbBr6 under different excitation wavelengths (340-480 nm) was investigated. As presented in Figure 3b, as the excitation wavelengths increased, the maximum PL emission peak of the CsPbBr3/Cs4PbBr6 showed no changes, indicating high and stable PL emission at 518 nm. An intrinsic emission most likely caused this excitation-independent characteristic. Moreover, as the excitation wavelengths increased, the PL intensity of the sample increased and then decreased, reaching its maximum when the excitation wavelength was 360 nm. This phenomenon can also be observed in the three-dimensional excitation-emission fluorescence spectrum, as shown in Figure 3c. Therefore, combining the PLE spectrum with the three-dimensional excitation-emission fluorescence spectrum, we can conclude that the excitation wavelength of the most substantial emission peak is between 345 nm and 360 nm.  Further observation found that the PLE spectrum has a similar change with the absorption spectrum in the range of 345 nm to 505 nm, which is possibly related to the fact that CsPbBr 3 absorbs the excitation photons and generates fluorescence; thus, a change in the absorption spectrum leads to a corresponding trend in the PLE spectrum. However, in the range of 307 nm to 325 nm, due to CsPbBr 3 nanocrystals embedded in the Cs 4 PbBr 6 matrix, the excitation photons are almost entirely absorbed by the Cs 4 PbBr 6 matrix but not by the internal CsPbBr 3 nanocrystals. This phenomenon is similar to the pure Cs 4 PbBr 6 feature [39]. In addition, the PL spectrum of CsPbBr 3 /Cs 4 PbBr 6 under different excitation wavelengths (340-480 nm) was investigated. As presented in Figure 3b, as the excitation wavelengths increased, the maximum PL emission peak of the CsPbBr 3 /Cs 4 PbBr 6 showed no changes, indicating high and stable PL emission at 518 nm. An intrinsic emission most likely caused this excitation-independent characteristic. Moreover, as the excitation wavelengths increased, the PL intensity of the sample increased and then decreased, reaching its maximum when the excitation wavelength was 360 nm. This phenomenon can also be observed in the three-dimensional excitation-emission fluorescence spectrum, as shown in Figure 3c. Therefore, combining the PLE spectrum with the three-dimensional excitationemission fluorescence spectrum, we can conclude that the excitation wavelength of the most substantial emission peak is between 345 nm and 360 nm.
XPS was applied to identify the chemical bonding and compositions to further explore the chemical structure of the CsPbBr 3 /Cs 4 PbBr 6 microcrystals. The CsPbBr 3 /Cs 4 PbBr 6 and Nanomaterials 2023, 13, 355 7 of 16 CsPbBr 3 QDs synthesized without liquid paraffin were used as reference, and CsPbBr 3 /Cs 4 PbBr 6 synthesized with liquid paraffin was named CsPbBr 3 /Cs 4 PbBr 6 -LP. Figure 4a-d shows the XPS full-scan spectra of the CsPbBr 3 / Cs 4 PbBr 6 -LP, CsPbBr 3 /Cs 4 PbBr 6 , and pure CsPbBr 3 QD powders and their corresponding high-resolution spectra of Cs-3d, Pb-4f, and Br-3d. Figure 4a demonstrates that three samples are composed of the Cs-3d, Pb-4f, and Br-3d bands. Figure 4b shows that all the Cs-3d spectra possess two peaks with two binding energies, which can be assigned to Cs-3d 5/2 and Cs-3d 3/2 , respectively. Similarly, the Pb-4f spectra reveal two separated peaks corresponding to the Pb-4f 7/2 and Pb-4f 5/2 levels, as shown in Figure 4c. For the Br-3d spectra, all three samples have a broad characteristic peak. After careful comparison, we found that the peaks of Cs-3d, Pb-4f, and Br-3d bands of the CsPbBr 3 /Cs 4 PbBr 6 -LP and CsPbBr 3 /Cs 4 PbBr 6 shift slightly toward higher binding energies (Figure 4d), which might be attributed to the coating of the Cs 4 PbBr 6 matrix. Furthermore, compared to the Cs-3d, Pb-4f, and Br-3d bands of the CsPbBr 3 /Cs 4 PbBr 6 , the corresponding peaks in the CsPbBr 3 /Cs 4 PbBr 6 -LP shift to lower binding energies again, suggesting the successful coating of the CsPbBr 3 /Cs 4 PbBr 6 -LP surface with liquid paraffin.
XPS was applied to identify the chemical bonding and compositions to further plore the chemical structure of the CsPbBr3/Cs4PbBr6 microcrystals. The CsPbBr3/Cs4Pb and CsPbBr3 QDs synthesized without liquid paraffin were used as reference, and C bBr3/Cs4PbBr6 synthesized with liquid paraffin was named CsPbBr3/Cs4PbBr6-LP. Fig  4a-d shows the XPS full-scan spectra of the CsPbBr3/Cs4PbBr6-LP, CsPbBr3/Cs4PbBr6, a pure CsPbBr3 QD powders and their corresponding high-resolution spectra of Cs-3d, 4f, and Br-3d. Figure 4a demonstrates that three samples are composed of the Cs-3d, 4f, and Br-3d bands. Figure 4b shows that all the Cs-3d spectra possess two peaks w two binding energies, which can be assigned to Cs-3d5/2 and Cs-3d3/2, respectively. Si larly, the Pb-4f spectra reveal two separated peaks corresponding to the Pb-4f7/2 and 4f5/2 levels, as shown in Figure 4c. For the Br-3d spectra, all three samples have a bro characteristic peak. After careful comparison, we found that the peaks of Cs-3d, Pb-4f, a Br-3d bands of the CsPbBr3/Cs4PbBr6-LP and CsPbBr3/Cs4PbBr6 shift slightly tow higher binding energies (Figure 4d), which might be attributed to the coating of Cs4PbBr6 matrix. Furthermore, compared to the Cs-3d, Pb-4f, and Br-3d bands of the C bBr3/Cs4PbBr6, the corresponding peaks in the CsPbBr3/Cs4PbBr6-LP shift to lower bind energies again, suggesting the successful coating of the CsPbBr3/Cs4PbBr6-LP surface w liquid paraffin. To analyze PL dynamics for the CsPbBr3, CsPbBr3/Cs4PbBr6, and CsPbBr3/Cs4PbB LP, the time-resolved PL decay curves of these three samples were collected using a nm pulse laser as an excitation source. According to Equations (1) and (2), each PL de curve can be accurately fitted by a triple exponential function, as shown in Figure 5 and Table 1. From Figure 5a, the PL lifetimes of the CsPbBr3/Cs4PbBr6 and the C bBr3/Cs4PbBr6-LP are relatively prolonged compared with pure CsPbBr3. In detail, av age PL decay times (τavg.) of 12.00, 29.50, and 42.63 ns for the CsPbBr3, CsPbBr3/Cs4Pb and CsPbBr3/Cs4PbBr6-LP, respectively, imply that the Cs4PbBr6 matrix can passivate CsPbBr3 and liquid paraffin can modify surface defects of CsPbBr3/Cs4PbBr6-LP. To analyze PL dynamics for the CsPbBr 3 , CsPbBr 3 /Cs 4 PbBr 6 , and CsPbBr 3 /Cs 4 PbBr 6 -LP, the time-resolved PL decay curves of these three samples were collected using a 375 nm pulse laser as an excitation source. According to Equations (1) and (2), each PL decay curve can be accurately fitted by a triple exponential function, as shown in Figure 5a-d and Table 1. From Figure 5a, the PL lifetimes of the CsPbBr 3 /Cs 4 PbBr 6 and the CsPbBr 3 /Cs 4 PbBr 6 -LP are relatively prolonged compared with pure CsPbBr 3 . In detail, average PL decay times (τ avg .) of 12.00, 29.50, and 42.63 ns for the CsPbBr 3 , CsPbBr 3 /Cs 4 PbBr 6 , and CsPbBr 3 /Cs 4 PbBr 6 -LP, respectively, imply that the Cs 4 PbBr 6 matrix can passivate the CsPbBr 3 and liquid paraffin can modify surface defects of CsPbBr 3 /Cs 4 PbBr 6 -LP.
factors A1 and A2, we found that A1 for CsPbBr3 is more significant than that of CsP-bBr3/Cs4PbBr6 and CsPbBr3/Cs4PbBr6-LP, while A2 for CsPbBr3 is smaller than that of CsP-bBr3/Cs4PbBr6 and CsPbBr3/Cs4PbBr6-LP, indicating that the surface defects of the CsPbBr3 are passivated by the well-matched lattice of the Cs4PbBr6 matrix, thereby increasing the probability of radiative recombination [41]. Furthermore, A1 for CsPbBr3/Cs4PbBr6 is more significant than A1 for CsPbBr3/Cs4PbBr6-LP, while A2 for CsPbBr3/Cs4PbBr6 is smaller than CsPbBr3/Cs4PbBr6-LP, further confirming the passivation of surface defects by liquid paraffin. This result also agrees with the above result that CsPbBr3/Cs4PbBr6-LP possesses a higher PLQY than CsPbBr3/Cs4PbBr6. Therefore, combining Cs4PbBr6 with liquid paraffin is suitable for better photoluminescence performance.   Furthermore, we found that three characteristic time constants τ 1 , τ 2 , and τ 3 existed in fitting curves, suggesting that the three samples contain more than one emission center with different recombination rates [26]. As previously reported, τ 1 is attributed to exciton recombination involving surface states and defects, τ 2 is assigned to radiative recombination, and τ 3 is related to non-radiative recombination [40]. The amplitude values A 1 , A 2 , and A 3 are considered the weighing factors [30]. As summarized in Table 1, the time constants τ 1 , τ 2 , and τ 3 for the CsPbBr 3 are 1.82, 6.53, and 22.17 ns, respectively. The respective amplitudes A 1 , A 2 , and A 3 are 0.56, 0.34, and 0.10, respectively. For CsPbBr 3 /Cs 4 PbBr 6 , the time constants τ 1 , τ 2 , and τ 3 for the CsPbBr 3 are 2.72, 10.71, and 55.05 ns, and the respective amplitudes A 1 , A 2 , and A 3 are 0.44, 0.47, and 0.09, respectively. A similar test procedure has been applied to the CsPbBr 3 /Cs 4 PbBr 6 -LP microcrystals. The results for time constants τ 1 , τ 2 , and τ 3 for the CsPbBr 3 are 2.27, 12.98, and 67.91 ns, and the respective amplitudes A 1 , A 2 , and A 3 are 0.38, 0.48, and 0.13, respectively. In terms of corresponding weighing factors A 1 and A 2 , we found that A 1 for CsPbBr 3 is more significant than that of CsPbBr 3 /Cs 4 PbBr 6 and CsPbBr 3 /Cs 4 PbBr 6 -LP, while A 2 for CsPbBr 3 is smaller than that of CsPbBr 3 /Cs 4 PbBr 6 and CsPbBr 3 /Cs 4 PbBr 6 -LP, indicating that the surface defects of the CsPbBr 3 are passivated by the well-matched lattice of the Cs 4 PbBr 6 matrix, thereby increasing the probability of radiative recombination [41]. Furthermore, A 1 for CsPbBr 3 /Cs 4 PbBr 6 is more significant than A 1 for CsPbBr 3 /Cs 4 PbBr 6 -LP, while A 2 for CsPbBr 3 /Cs 4 PbBr 6 is smaller than CsPbBr 3 /Cs 4 PbBr 6 -LP, further confirming the passivation of surface defects by liquid paraffin. This result also agrees with the above result that CsPbBr 3 /Cs 4 PbBr 6 -LP possesses a higher PLQY than CsPbBr 3 /Cs 4 PbBr 6 . Therefore, combining Cs 4 PbBr 6 with liquid paraffin is suitable for better photoluminescence performance.

Effect of the Liquid Paraffin Concentration on the Optical Properties of the CsPbBr 3 /Cs 4 PbBr 6
In the preparation process, we found that the contents of liquid paraffin significantly affected the optical properties of CsPbBr 3 /Cs 4 PbBr 6 -LP, so we further investigated this factor. As shown in Figure 6a, as the contents of liquid paraffin increased, the PL intensity of CsPbBr 3 /Cs 4 PbBr 6 -LP increased first. Then, it decreased and was at its maximum when the content of liquid paraffin was 20%, keeping the intensity of the characteristic absorption peak (510 nm) constant. This phenomenon is related to the solubility of CsBr and PbBr 2 in DMSO and liquid paraffin. Increasing the content of liquid paraffin can improve the protection of samples. However, since CsBr and PbBr 2 are more difficult to dissolve in liquid paraffin than in DMSO, less CsPbBr 3 /Cs 4 PbBr 6 is synthesized, and the fluorescence efficiency is lower. In addition, a PL blue shift of the samples, from 518 to 505 nm, was found as the content of liquid paraffin increased (Figure 6b). The reason for this phenomenon is that under different liquid paraffin contents, CsPbBr 3 /Cs 4 PbBr 6 crystals' crystallization speed is different, leading to different sizes of CsPbBr 3 /Cs 4 PbBr 6 composites. When liquid paraffin's content increases, perovskite's particle size decreases, and the energy gap widens, so the PL emission wavelength blue-shifts, similar to what has been previously reported [39,42]. Although the content of liquid paraffin is constantly changing, the FWHM of the samples is all less than 30 nm (Figure 6c), which helps to obtain LED devices with high color gamut values. Moreover, the PLQY of the samples varied with the content of liquid paraffin. When the liquid paraffin concentration increased, the PLQY first increased rapidly and then decreased slowly, reaching a maximum of 74% at a liquid paraffin concentration of 20%, which is consistent with the change in PL intensity, as shown in Figure 6d. Therefore, selecting the appropriate liquid paraffin concentration is crucial to obtaining high optical performance.

Stability of CsPbBr 3 /Cs 4 PbBr 6 Microcrystal
The performance of perovskite crystals is severely affected by the surrounding environment, such as heat, ultraviolet light, and polar solvents. The stability of perovskite crystals is crucial for their practical applications. Here, we systematically investigated the UV photostability, thermotolerance, storage stability, and water stability of CsPbBr 3 /Cs 4 PbBr 6 powders. To evaluate the UV light resistance of the as-obtained samples, CsPbBr 3 , CsPbBr 3 /Cs 4 PbBr 6 , and CsPbBr 3 /Cs 4 PbBr 6 -LP powders were irradiated under continuous 365 nm UV light with an optical power density of 16 mW/cm 2 for 50 h, as shown in Figure 7a. The normalized PL intensity of CsPbBr 3 /Cs 4 PbBr 6 -LP presented only a 13.4% decrease after the irradiation of 50 h, while CsPbBr 3 and CsPbBr 3 /Cs 4 PbBr 6 dropped 68.9% and 26.3% at the same measurement conditions, respectively. This excellent photostability of CsPbBr 3 /Cs 4 PbBr 6 -LP is mainly due to the double protection of liquid paraffin and the Cs 4 PbBr 6 matrix. In addition, the thermotolerance of CsPbBr 3 , CsPbBr 3 /Cs 4 PbBr 6 , and CsPbBr 3 /Cs 4 PbBr 6 -LP was investigated. Figure 7b shows the normalized PL intensity change of three sample powders heated at 100 • C within 120 h. The results show that the CsPbBr 3 /Cs 4 PbBr 6 -LP decays only 12.5%, while the CsPbBr 3 is nearly fluorescence-quenched, and the CsPbBr 3 /Cs 4 PbBr 6 drops 38.4% after heating for 120 h, which further verifies the satisfactory thermotolerance of CsPbBr 3 /Cs 4 PbBr 6 -LP.

Stability of CsPbBr3/Cs4PbBr6 Microcrystal
The performance of perovskite crystals is severely affected by the surrounding environment, such as heat, ultraviolet light, and polar solvents. The stability of perovskite crystals is crucial for their practical applications. Here, we systematically investigated the UV photostability, thermotolerance, storage stability, and water stability of CsP-bBr3/Cs4PbBr6 powders. To evaluate the UV light resistance of the as-obtained samples, CsPbBr3, CsPbBr3/Cs4PbBr6, and CsPbBr3/Cs4PbBr6-LP powders were irradiated under continuous 365 nm UV light with an optical power density of 16 mW/cm 2 for 50 h, as shown in Figure 7a. The normalized PL intensity of CsPbBr3/Cs4PbBr6-LP presented only a 13.4% decrease after the irradiation of 50 h, while CsPbBr3 and CsPbBr3/Cs4PbBr6 dropped 68.9% and 26.3% at the same measurement conditions, respectively. This excellent photostability of CsPbBr3/Cs4PbBr6-LP is mainly due to the double protection of liquid paraffin and the Cs4PbBr6 matrix. In addition, the thermotolerance of CsPbBr3, CsP-bBr3/Cs4PbBr6, and CsPbBr3/Cs4PbBr6-LP was investigated. Figure 7b shows the normalized PL intensity change of three sample powders heated at 100 °C within 120 h. The results show that the CsPbBr3/Cs4PbBr6-LP decays only 12.5%, while the CsPbBr3 is nearly fluorescence-quenched, and the CsPbBr3/Cs4PbBr6 drops 38.4% after heating for 120 h, which further verifies the satisfactory thermotolerance of CsPbBr3/Cs4PbBr6-LP.  Water and air resistance are also essential for perovskite materials. We further study the storage and polar solvent stability of CsPbBr3, CsPbBr3/Cs4PbBr6, and CsP-bBr3/Cs4PbBr6-LP. As shown in Figure 7c, the PL intensity of CsPbBr3/Cs4PbBr6-LP dropped only 9.1% after 120 days of storage under ambient conditions (HH 80% and HT 25 °C), while CsPbBr3 and CsPbBr3/Cs4PbBr6 attenuated 99% and 30.8%, suggesting that Water and air resistance are also essential for perovskite materials. We further study the storage and polar solvent stability of CsPbBr 3 , CsPbBr 3 /Cs 4 PbBr 6 , and CsPbBr 3 /Cs 4 PbBr 6 -LP. As shown in Figure 7c, the PL intensity of CsPbBr 3 /Cs 4 PbBr 6 -LP dropped only 9.1% after 120 days of storage under ambient conditions (HH 80% and HT 25 • C), while CsPbBr 3 and CsPbBr 3 /Cs 4 PbBr 6 attenuated 99% and 30.8%, suggesting that our CsPbBr 3 /Cs 4 PbBr 6 -LP possesses stable green emission ability. Additionally, the polar solvent stability of three samples was evaluated, as shown in Figure 7d. Surprisingly, the CsPbBr 3 /Cs 4 PbBr 6 -LP keeps 78% of its initial PL intensity after being soaked in water (30 mg/mL) for 16 days, while CsPbBr 3 decay is swift, and it eventually loses its luminescence. The CsPbBr 3 /Cs 4 PbBr 6 maintains 67% of its initial PL intensity. Therefore, these results demonstrate that our CsPbBr 3 /Cs 4 PbBr 6 -LP possesses excellent photostability, thermal stability, storage stability, and polar solvent stability, showing great potential as high-quality optoelectronic devices under harsh conditions.

Application in WLEDs and Visible Light Communication Devices
Benefiting from the good optical properties and excellent stability of CsPbBr 3 /Cs 4 PbBr 6 -LP, its application in display and visible light communication devices is promising. Firstly, we fabricated a WLED device with a commercial blue LED chip, green-emissive CsPbBr 3 /Cs 4 PbBr 6 -LP film, and red phosphor film; the schematic and physical diagram of WLEDs are shown in Figure 8a. When the CsPbBr 3 /Cs 4 PbBr 6 -LP powder concentration was adjusted from 1.5 wt% to 10 wt%, the correlated color temperature (CCT) of the WLED devices ranged from 4100 K to 6500 K, and their color rendering index (CRI) was more than 85. In addition, with increased CsPbBr 3 /Cs 4 PbBr 6 -LP powder concentration, the CCT value changes along the Planckian locus line, demonstrating a promising candidate for display. The driving current-dependent luminous flux and efficacy of the CsPbBr 3 /Cs 4 PbBr 6 -LP WLED device are presented in Figure 8b. The luminous efficacy curve shows that the luminous efficacy of the remote excitation LED device reaches 129.5 lm/W before aging, which is comparable to the highest efficiency record of the traditional ligand-protected CsPbBr 3 QDs. Compared with the expected standard, the CsPbBr 3 /Cs 4 PbBr 6 -LP-based WLEDs have a wide color gamut (solid red triangle), with 121% of the NTSC (white dashed line) and 94% of the Rec. 2020 (blue dashed line), as demonstrated in Figure 8c.
In addition to analyzing the CsPbBr 3 /Cs 4 PbBr 6 -LP-based WLED device's optical performance, we also evaluated its reliability. The LED industry usually adopts the double 85 standards. The product is aged in an environment with a high temperature of 85 • C and a high humidity of 85% (85 • C/85% RH). The performance changes before and after aging are compared to determine the product's heat and humidity resistance. Here, we put the WLED device in an environment of 85 • C for thermal reliability experiments. The changes in optical performance during the aging process are shown in Figure 8d. After aging for 50 h, the EL spectrum shape remains almost unchanged, and the intensity of the green light spectrum decreases slowly. It can be seen from the normalized data that although the green light decreases after aging (Figure 8e), the aging time for its spectral intensity to decay to 90% of the initial intensity is as long as 50 h and gradually tends to be stable. In contrast, under the same conditions, WLED devices using CsPbBr 3 decayed to about 70% of their initial intensity in less than an hour. In comparison, WLED devices containing CsPbBr 3 /Cs 4 PbBr 6 decayed to 68% of their initial intensity in 50 h, which shows that CsPbBr 3 /Cs 4 PbBr 6 -LP-based WLED devices have good thermal reliability.
CsPbBr3/Cs4PbBr6-LP WLED device are presented in Figure 8b. The luminous efficacy curve shows that the luminous efficacy of the remote excitation LED device reaches 129.5 lm/W before aging, which is comparable to the highest efficiency record of the traditional ligand-protected CsPbBr3 QDs. Compared with the expected standard, the CsP-bBr3/Cs4PbBr6-LP-based WLEDs have a wide color gamut (solid red triangle), with 121% of the NTSC (white dashed line) and 94% of the Rec. 2020 (blue dashed line), as demonstrated in Figure 8c. In addition to analyzing the CsPbBr3/Cs4PbBr6-LP-based WLED device's optical performance, we also evaluated its reliability. The LED industry usually adopts the double 85 standards. The product is aged in an environment with a high temperature of 85 °C and a high humidity of 85% (85 °C/85% RH). The performance changes before and after aging are compared to determine the product's heat and humidity resistance. Here, we put the WLED device in an environment of 85 °C for thermal reliability experiments. The changes in optical performance during the aging process are shown in Figure 8d. After aging for 50 h, the EL spectrum shape remains almost unchanged, and the intensity of the green light spectrum decreases slowly. It can be seen from the normalized data that although the green light decreases after aging (Figure 8e), the aging time for its spectral intensity to decay to 90% of the initial intensity is as long as 50 h and gradually tends to be stable. In contrast, under the same conditions, WLED devices using CsPbBr3 decayed to about 70% of their initial intensity in less than an hour. In comparison, WLED devices Besides the display, CsPbBr 3 /Cs 4 PbBr 6 -LP-based WLEDs can also be used as optical sources to transmit data in VLC systems. Here, we investigated the communication performance of CsPbBr 3 /Cs 4 PbBr 6 -LP-based WLEDs using the measurement setup shown in Figure 9a. The electrical-optical-electrical frequency response of the device can be obtained using such a test system. CsPbBr 3 /Cs 4 PbBr 6 -LP-based WLEDs were tested at a direct current bias of 3.0 V. From Figure 9b, it can be seen that these devices exhibit a typical low-pass frequency response, corresponding to a −3 dB bandwidth of about 3.7 MHz. Compared with the conventional phosphor white light system, the PL lifetime of CsPbBr 3 /Cs 4 PbBr 6 -LP (nanoseconds) is much shorter than phosphor (microseconds). According to previous reports, the −3 dB bandwidth of the CsPbBr 3 /Cs 4 PbBr 6 -LP-based WLEDs also can be calculated by Equation (3) [7,43]: Thus, the bandwidth estimation of the CsPbBr3/Cs4PbBr6-LP-based WLEDs is 3.733 MHz, consistent with the result collected from Figure 9b. In addition, the time-dependent −3 dB bandwidth of the CsPbBr3/Cs4PbBr6-LP-based WLEDs was measured after exposure to the natural environment. Figure 9c exhibits that the CsPbBr3/Cs4PbBr6-LP-based WLED device shows almost no decay after 15 days in the air (20-28 °C), suggesting that it has good stability and is promising in communication applications.

Conclusions
In summary, we demonstrated a facile and effective strategy to enhance the performance of CsPbBr3/Cs4PbBr6, which was performed by applying ultrasonication and liquid paraffin. By applying XRD, SEM, EDX, EDS, Abs/PL/PLE, XPS, and PL decay lifetime characterizations, all these results provide solid evidence supporting the formation of CsPbBr3/Cs4PbBr6 composites. Changing the content of liquid paraffin, bright-emission CsPbBr3/Cs4PbBr6-LP solid powders with a maximum PLQY of 74% and a narrow FWHM of about 27 nm were achieved. Thanks to the protection of the Cs4PbBr6 matrix and liquid paraffin, the PL intensity of CsPbBr3/Cs4PbBr6-LP dropped only 13.4% after continued irradiation by 365 nm UV light for 50 h and decayed only 12.5% at 100 °C within 120 h. Moreover, the CsPbBr3/Cs4PbBr6-LP powder shows superior stability with minimal degradation after 120 days of storage under ambient conditions. Even after soaking in a polar solvent (water) for 16 days, its PL intensity remained at about 85% of the initial value. The fabricated CsPbBr3/Cs4PbBr6-LP-based WLEDs show excellent luminescent performance, with a power efficiency of 129.5 lm/W and a wide color gamut, with 121% of the NTSC

Conclusions
In summary, we demonstrated a facile and effective strategy to enhance the performance of CsPbBr 3 /Cs 4 PbBr 6 , which was performed by applying ultrasonication and liquid paraffin. By applying XRD, SEM, EDX, EDS, Abs/PL/PLE, XPS, and PL decay lifetime characterizations, all these results provide solid evidence supporting the formation of CsPbBr 3 /Cs 4 PbBr 6 composites. Changing the content of liquid paraffin, bright-emission CsPbBr 3 /Cs 4 PbBr 6 -LP solid powders with a maximum PLQY of 74% and a narrow FWHM of about 27 nm were achieved. Thanks to the protection of the Cs 4 PbBr 6 matrix and liquid paraffin, the PL intensity of CsPbBr 3 /Cs 4 PbBr 6 -LP dropped only 13.4% after continued irradiation by 365 nm UV light for 50 h and decayed only 12.5% at 100 • C within 120 h. Moreover, the CsPbBr 3 /Cs 4 PbBr 6 -LP powder shows superior stability with minimal degradation after 120 days of storage under ambient conditions. Even after soaking in a polar solvent (water) for 16 days, its PL intensity remained at about 85% of the initial value. The fabricated CsPbBr 3 /Cs 4 PbBr 6 -LP-based WLEDs show excellent luminescent performance, with a power efficiency of 129.5 lm/W and a wide color gamut, with 121% of the NTSC and 94% of the Rec. 2020, suggesting they represent a promising candidate for displays. In addition, the CsPbBr 3 /Cs 4 PbBr 6 -LP-based WLEDs were also demonstrated in a VLC system. The results suggested the great potential of these high-performance WLEDs as an excitation light source to achieve visible light communication.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.

Conflicts of Interest:
The authors declare no conflict of interest.