Synergistic Halide- and Ligand-Exchanges of All-Inorganic Perovskite Nanocrystals for Near-Unity and Spectrally Stable Red Emission

All-inorganic perovskite nanocrystals (NCs) of CsPbX3 (X = Cl, Br, I) are promising for displays due to wide color gamut, narrow emission bandwidth, and high photoluminescence quantum yield (PLQY). However, pure red perovskite NCs prepared by mixing halide ions often result in defects and spectral instabilities. We demonstrate a method to prepare stable pure red emission and high-PLQY-mixed-halide perovskite NCs through simultaneous halide-exchange and ligand-exchange. CsPbBr3 NCs with surface organic ligands are first synthesized using the ligand-assisted reprecipitation (LARP) method, and then ZnI2 is introduced for anion exchange to transform CsPbBr3 to CsPbBrxI3−x NCs. ZnI2 not only provides iodine ions but also acts as an inorganic ligand to passivate surface defects and prevent ion migration, suppressing non-radiative losses and halide segregation. The luminescence properties of CsPbBrxI3−x NCs depend on the ZnI2 content. By regulating the ZnI2 exchange process, red CsPbBrxI3−x NCs with organic/inorganic hybrid ligands achieve near-unity PLQY with a stable emission peak at 640 nm. The CsPbBrxI3−x NCs can be combined with green CsPbBr3 NCs to construct white light-emitting diodes with high-color gamut. Our work presents a facile ion exchange strategy for preparing spectrally stable mixed-halide perovskite NCs with high PLQY, approaching the efficiency limit for display or lighting applications.


Introduction
With the growing development of high-definition displays, higher requirements were put forward for the luminescent materials in recent years. Semiconductor nanocrystals (NCs) show great application potential in next-generation display technologies due to their good optoelectronic properties, such as narrow full widths at half maximum (FWHM), high-photoluminescence quantum yield (PLQY) and tunable forbidden bandwidth [1,2]. Currently, cadmium chalcogenides are widely used compounds for the preparation of high performance NCs. However, in order to achieve good stability and high PLQY in cadmium chalcogenide NCs, a shell layer needs to be grown on top of the core to provide surface passivation, and the synthesis of core/shell structured NCs usually requires high-reaction temperature [3,4]. In this case, these stable and efficient core/shell NCs show high cost, which is not conducive to large-scale commercial application. Recently, the halide perovskite NCs, particularly the all-inorganic CsPbX 3 (X = Cl, Br, I), are favored by researchers The excellent spectral stability should be attribute to the inhibition of halide segregation in the CsPbBr x I 3−x NCs. More importantly, the PLQY of the pure red perovskite NCs was achieved to be close to 100%, suggesting the significant suppression of non-radiative recombination. This work provides a facile and effective method to prepare efficient and stable red perovskite NCs and paves the way for wide color gamut vivid display applications.

Synthesis of CsPbBr 3 NCs
The green CsPbBr 3 NCs were synthesized by a LARP method at room temperature [34]. Firstly, the solutions of Cs + (0.05 M Cs 2 CO 3 dissolved in OTAC) and Pb 2+ (0.05 M PbBr 2 and 0.1 M TOAB dissolved together in toluene) were prepared. Then, 280 µL Cs + solution was swiftly added into 2.5 mL Pb 2+ solution with vigorous stirring for 30 s. Subsequently, 830 µL DDAB solution (12 mg/mL in toluene) was added and stirred for another 5 min to obtain CsPbBr 3 crude solution. Afterwards, 7 mL ethyl acetate was poured into the CsPbBr 3 crude solution, followed by centrifuging at 12,000 rpm for 5 min. The supernatant was discarded, and the precipitate was collected and dispersed in 2 mL toluene. After centrifuging at 4000 rpm for 5 min, the solution of CsPbBr 3 NCs was filtered by the 0.22 µm filter for next experiments and characterization.

Preparation of ZnI 2 and ZnBr 2 Precursor Solutions
The precursor solutions of ZnI 2 were prepared by dissolving ZnI 2 , TBPO and OAm in toluene. The concentration of TBPO was kept at 0.5 M, and the volume ratio of OAm to toluene was maintained at 3%. For preparing the ZnBr 2 precursors, the process is the same, only replaced ZnI 2 with ZnBr 2 .

Anion Exchange Processes
A 500 µL ZnI 2 precursor solution with various ZnI 2 content was swiftly added into 2 mL solution of CsPbBr 3 NCs with vigorous stirring for a desired time. Immediately, 7 mL ethyl acetate was poured into the solution following by centrifuging at 12,000 rpm for 8 min. The supernatant was discarded, and the precipitate was dispersed in 2 mL toluene and centrifuged at 5000 rpm for 5 min. Eventually, the resultant solution was filtered by 0.22 µm filter. Notably, the whole process was carried out in a N 2 -filled glove box.

Fabrication of WLED Device
The as-prepared green and red perovskite NCs were mixed with PMMA, respectively. The mixture of the pristine CsPbBr 3 NCs and PMMA was first coated onto the 460 nm blue LED chip and allowed to dry. Subsequently, the mixture of the red CsPbBr x I 3−x NCs (0.1 M ZnI 2 NCs) and PMMA was coated on top of the green layer and then dried again.

Results and Discussions
The green colloidal CsPbBr 3 NCs were firstly synthesized by the method of LARP under room-temperature with octanoic acid (OTAC) and didodecyldimethylammonium bromide (DDAB) as the ligands [34], and were dispersed in toluene for the subsequent ion exchange (see the experimental section for details). Figure 1a shows the schematic diagram of the exchange process. Here, n-Octylamine (OAm) from the ZnI 2 precursor solution plays the role of carriers for halide ions [35,36]. After mixing the colloidal CsPbBr 3 dispersion and ZnI 2 precursor, the OAm carried I − can dynamically approach to the surface of the perovskite NCs, and the exchange between Br − and I − will occur spontaneously to form the mixed-halide CsPbBr x I 3−x NCs (Figure 1a). The PL peaks of the CsPbBr x I 3−x NCs were precisely tunable between 518 nm and 640 nm by varying the concentration of ZnI 2 in the precursor solutions ( Figure 1b and Table S1). After the exchange, the spectra of the resultant perovskite NCs remain in a single PL peak with narrow FWHM, indicating the formation of CsPbBr x I 3−x NCs with the homogeneous phase structure. When the ZnI 2 concentration is up to 0.08 M and 0.1 M, the CsPbBr x I 3−x NCs reach the red emission region with PL peaks of 619 nm and 640 nm, respectively. The transmission electron microscope (TEM) measurements were performed to reveal the effect of ZnI 2 on the sizes and shapes of the CsPbBr x I 3−x NCs. From the TEM images, it is found that the average size of the red perovskite NCs exhibits an expansion from 12.9 ± 1.9 nm (pristine CsPbBr 3 ) to 13.2 ± 1.8 nm (0.08 M ZnI 2 ) and 13.7 ± 2.0 nm (0.1 M ZnI 2 ), as shown in Figure 1c-e, and their morphologies show ignorable change. Furthermore, the crystallographic characteristics of those perovskite NCs (pristine CsPbBr 3 , 0.08 M ZnI 2 and 0.1 M ZnI 2 ) were studied by the X-ray diffraction (XRD) measurements, and the corresponding XRD patterns are displayed in Figure 1f. The conventional cubic crystal structure is present in all the perovskite NCs, suggesting that the exchange process did not change the crystallographic feature which is consistent with the previous reports [23]. The diffraction peaks of the red CsPbBr x I 3−x perovskite NCs are consistently shifted to the lower 2θ angles, indicating an increase in cell volume as the radius of I anion is greater than that of the Br anion. These results show that ZnI 2 -dominated anion exchange can achieve the gradual red shift of the PL spectra without destructing the perovskite structure.
The PLQYs of the perovskite NCs were measured for comparison, and the results are shown in Figure 2a and Table S1. After the exchange of green CsPbBr 3 NCs with 0.04 M ZnI 2 , the PLQY shows significant reduction from 72.3% to 21.2% and the PL peak changes from 518 nm to 556 nm. Intriguingly, when further increasing the ZnI 2 concentration, the PLQY of the CsPbBr x I 3−x NCs is improved gradually with the red shift of PL spectra and approaches unity at 640 nm. The change of PLQY suggests that the suppression of non-radiative loss in the CsPbBr x I 3−x NCs is strongly dependent on the added ZnI 2 during the exchange process. To obtain insight into the recombination mechanism in the red perovskite NCs of 0.08 M and 0.1 M ZnI 2 , the time-resolved PLs were measured. In Figure 2b, it is observed that both 0.08 M and 0.1 M ZnI 2 NCs exhibit bi-exponential decay curves with one fast component and one slow component. The fitted parameters for these decay curves are listed in Table S2 in the Supporting Information. It is shown that the average lifetime (τ avg ) of the 0.1 M ZnI 2 NCs (68.73 ns) is longer than that of the 0.08 M ZnI 2 NCs (56.62 ns). This suggests that there are fewer nonradiative channels in the 0.1 M ZnI 2 NCs [37,38], indicating a reduction in defect states of CsPbBr x I 3−x NCs as the ZnI 2 concentration increases. Furthermore, the carrier dynamics and the excitons recombination process were investigated by the femtosecond transient absorption (fs-TA) measurements. As shown in Figure S1a,b, the photo-induced bleaching (PB) signals of 0.08 M ZnI 2 and 0.1 M ZnI 2 NCs were located at 611 nm and 630 nm, respectively, which are in agreement with the absorption results as shown in Figure 1b. The characteristic decay-associated spectra were then extracted through a global fitting procedure (Figure 2c,d), with τ 1 , τ 2 , and τ 3 components identified as being responsible for intraband hot-exciton relaxation, exciton trapping to the band gap trap states, and exciton recombination, respectively, as schematically illustrated in Figure 2e [38][39][40]. Results from the global fitting demonstrate that the exciton trapping process of 0.1 M ZnI 2 NCs was slower compared to that of 0.08 M ZnI 2 NCs. Additionally, a significant delay in kinetic recombination was observed for 0.1 M ZnI 2 NCs, which suggests the suppression of non-radiative recombination with an increase in ZnI 2 precursor solution concentration (Figure 2f). played in Figure 1f. The conventional cubic crystal structure is present in all the perovskite NCs, suggesting that the exchange process did not change the crystallographic feature which is consistent with the previous reports [23]. The diffraction peaks of the red CsP-bBrxI3−x perovskite NCs are consistently shifted to the lower 2θ angles, indicating an increase in cell volume as the radius of I anion is greater than that of the Br anion. These results show that ZnI2-dominated anion exchange can achieve the gradual red shift of the PL spectra without destructing the perovskite structure. The PLQYs of the perovskite NCs were measured for comparison, and the results are shown in Figure 2a and Table S1. After the exchange of green CsPbBr3 NCs with 0.04 M ZnI2, the PLQY shows significant reduction from 72.3% to 21.2% and the PL peak changes from 518 nm to 556 nm. Intriguingly, when further increasing the ZnI2 concentration, the PLQY of the CsPbBrxI3−x NCs is improved gradually with the red shift of PL spectra and approaches unity at 640 nm. The change of PLQY suggests that the suppression of nonradiative loss in the CsPbBrxI3−x NCs is strongly dependent on the added ZnI2 during the identified as being responsible for intraband hot-exciton relaxation, exciton trapping to the band gap trap states, and exciton recombination, respectively, as schematically illustrated in Figure 2e [38][39][40]. Results from the global fitting demonstrate that the exciton trapping process of 0.1 M ZnI2 NCs was slower compared to that of 0.08 M ZnI2 NCs. Additionally, a significant delay in kinetic recombination was observed for 0.1 M ZnI2 NCs, which suggests the suppression of non-radiative recombination with an increase in ZnI2 precursor solution concentration (Figure 2f).  As described above, the ZnI 2 -based exchange can be used to prepare highly efficient red perovskite NCs. To further study the effect of exchange process on the formation of red perovskite NCs, the in situ PL spectroscopy was employed to monitor the spectral evolution after dropping the ZnI 2 precursor into the pristine CsPbBr 3 NCs. Figure 3a,b shows the change of PL spectra during the exchange with 0.1 M ZnI 2 . The PL peak of the perovskite NCs presents negligible redshift during the first second (Figure 3a), suggesting that the halide-exchange is very slight in this period. After then, the PL peak sharply changes to around 600 nm and then gradually redshifts to 640 nm (Figure 3a,b), which clearly indicates that the halide-exchange is very fast at the beginning and then slows. After around 50 s, no change of the PL peak is found in the CsPbBr x I 3−x NCs, meaning that the equilibrium of halide-exchange is reached. Notably, the PL intensity decreases dramatically during the first second and then gradually increases, and ultimately remains unchanged after around 90 s. The evolution of the PL intensity suggests that the abundant non-radiative channels are quickly formed and then gradually eliminated. When dropping ZnI 2 precursor, the non-radiative channels could be mainly originated from the rapidly emerged defects located at the surface of the perovskite NCs, leading to the great decrease of PL intensity. We speculate that the surface defects might be caused by the detachment of the original ligands (OTAC and DDAB) and the formation of halogen vacancies. And the abundant halogen vacancies on the surface can trigger the halide-exchange when OAm carried I − approached. Actually, the desorption and absorption of ligands can remain a dynamic balance in the solution of the pristine CsPbBr 3 NCs. However, this balance is broken by adding the ZnI 2 precursor, and the ligand-exchange immediately occurs at the surface of NCs. It's worth noting that the PL intensity continues growing in tens of seconds after the halide-exchange equilibrium. During this period, the reduction of surface defects should be mainly attributed to the effect of ligand-exchange. Finally, the equilibrium of ligand-exchange is achieved and the suppression of surface defects stops, resulting in no change of the PL intensity. More importantly, all the PL spectra in Figure 3b show the single-peak characteristics, implying that the perovskite NCs keeps the homogeneous phase structure during the whole exchange process. And the similar evolution of PL spectra was also observed in the exchange process for the 0.08 M ZnI 2 NCs, but the establish times for the equilibriums of halide-exchange and ligand-exchange became shorter than those of the 0.1 M ZnI 2 NCs, as shown in Figure 3c,d. Color scale bar is provided on the right side for the intensity of the PL.
In the ZnI2 precursor solutions, OAm and tributylphosphine oxide (TBPO) were used to improve the solubility of ZnI2 in the solvent of toluene. Actually, OAm and TBPO can act as the ligands and would contribute to the exchange process. To study the effect of OAm and TBPO, we performed the in situ PL measurement to probe the emission change of the pristine CsPbBr3 NCs after dropping the precursor without ZnI2. As shown in Figure   Figure 3. PL evolution during the exchange process by adding ZnI 2 precursor into the as-prepared CsPbBr 3 NCs. The fitted 2D PL spectra for the 0.1 M ZnI 2 NCs (a,b) and the 0.08 M ZnI 2 NCs (c,d).
Color scale bar is provided on the right side for the intensity of the PL.
In the ZnI 2 precursor solutions, OAm and tributylphosphine oxide (TBPO) were used to improve the solubility of ZnI 2 in the solvent of toluene. Actually, OAm and TBPO can act as the ligands and would contribute to the exchange process. To study the effect of OAm and TBPO, we performed the in situ PL measurement to probe the emission change of the pristine CsPbBr 3 NCs after dropping the precursor without ZnI 2 . As shown in Figure S2a, the PL intensity of the CsPbBr 3 NCs also experienced a decrease followed by an increase, which is consistent with the ligand-exchange process as discussed above. But the decrease of the PL intensity is not significant, suggesting that the effect of ligand-exchange on the surface structure of CsPbBr 3 NCs is limited when adding OAm and TBPO without ZnI 2 . And the absorption, PL, TEM, PLQY, XRD and X-ray photoelectron spectroscopy (XPS) results demonstrate that the exchanged perovskite NCs show the similar characteristics as the pristine CsPbBr 3 NCs (Figures S3-S7). At the same time, we also studied the effect of Zn 2+ without I − on the ligand-exchange by adding OAm and TBPO with ZnBr 2 . Figure S2b shows the change of the PL spectra after dropping the precursor solution containing 0.1 M ZnBr 2 into the pristine CsPbBr 3 dispersion. It is shown that the decrease of PL intensity becomes more significant and faster (in a few seconds), which is accompanied with the blue-shift of PL peaks. This indicates that the presence of Zn 2+ can encourage the change of surface structure by facilitating the detachment of the initial ligands from the pristine CsPbBr 3 NCs. And the spectral blue-shift should be attributed to the reduced size of the perovskite NCs, which was confirmed by the TEM result ( Figure S4). The extensive detachment of ligands would cause the disruption of the surface lattice and thus decrease the size of the CsPbBr 3 NCs, and the Zn 2+ , detected by XPS as shown in Figure S7, can combine with Br − to be anchored on the CsPbBr 3 NCs as the inorganic ligands to "healing" the defective surface, leading to a small increase in the PLQY ( Figure S5). The above results demonstrate that the Zn 2+ plays a very important role during the ligand-exchange process, and its effect on the surface properties of NCs is more effective than OAm and TBPO.
To obtain more insight into the change of ligands between the pristine CsPbBr 3 NCs and the obtained red CsPbBr x I 3−x NCs, Fourier transform infrared (FTIR) and XPS measurements were performed. As depicted in Figure 4a, the FTIR spectrum of the pristine perovskite NCs clearly showed the resonances of organic groups (ν s (C-H x ) = 2700-3000 cm −1 , ν s (-COOH) = 1680-1700 cm −1 ). The intensities of these peaks greatly dropped for the 0.08 M ZnI 2 NCs and 0.1 M ZnI 2 NCs, especially the C=O stretch peaks (ν s (-COOH)), suggesting that the OTAC ligands were mostly removed from the surface of perovskite NCs. From the XPS results ( Figure S8b), the lowered oxygen contents further verified the reduction of OTAC in the exchanged perovskite NCs. In addition, nitrogen is only present in the ligand of DDAB for the pristine perovskite NCs, and would originate from both of DDAB and OAm for the red NCs. The decrease of N content confirmed the significant reduction of DDAB after exchange (Figure 4b). The pristine CsPbBr 3 NCs are covered by the ligands of OTAC and DDAB, wherein OTAC is located on the NCs surface through the binding of Pb 2+ with COO − , and DDAB exists on the NCs surface through the interaction of halogen anion and tetra-ammonium cation [34]. One of the main functions of these organic ligands is to passivate the surface defects and thus suppress non-radiative recombination in the perovskite NCs. The reduction of these organic ligands should result in more defects, which act as the non-radiative recombination centers, on the surface of the perovskite NCs. However, the high PLQYs clearly indicate that the non-radiative recombination of the red perovskite NCs is not enhanced with significantly decreasing the amount of organic ligands. It is believed that the Zn 2+ should make up the loss of organic ligands and thus greatly suppress the non-radiative recombination. Mixed halide perovskites are known to undergo photo-induced halide segregation, which will result in the spectral instability [42][43][44]. In order to study the photostability of the red CsPbBrxI3−x NCs, we measured the PL spectra of the perovskite films under intense illumination of ultraviolet light. A simplified schematic of the measurement equipment is shown in Figure 5a. The PL spectra and their peak intensities of the 0.1 M ZnI2 NCs did not exhibit any change after 5 h of irradiation, indicating the absence of phase separation caused by photo-induced halide segregation, as depicted in Figure 5b. Encouraged by the exceptional photostability and high PLQY of the obtained red perovskite NCs, we prepared the white light-emitting diode (LED) by combining the 0.1 M ZnI2 NCs and pristine green CsPbBr3 NCs with a commercial blue LED chip. In Figure 5c, it is evident that there are three distinct emission peaks which are assigned to the corresponding blue chip, green CsPbBr3 NCs and red 0.1 M ZnI2 NCs. Notably, there is no intermediate emission peak detected, indicating that the perovskite NCs were effectively separated in the PMMA matrix to block the anion exchange between the green and red NCs. Figure 5d displays the CIE chromaticity coordinates of the blue chip, pristine green CsPbBr3 NCs and 0.1 M ZnI2 NCs. The selected CIE color coordinate triangle, including the corresponding blue, green and red monochromatic emission, encompasses 120% of the National Television System Committee's (NTSC) latest standard (Rec. 2020). This demonstrates that these green and red perovskite NCs, which are easily obtained by LARP and ion-exchange, respectively, can be used as the promising low-cost emitters in wide color gamut display devices. To investigate the role of Zn 2+ , density functional theory (DFT,) simulations were employed to evaluate the impact of Zn 2+ doping into the lattices of the CsPbBr 3 and CsPbI 3 NCs. Our calculations show that the formation energies of doping Zn 2+ at the B-site of cubic CsPbBr 3 and CsPbI 3 are around 0.8 eV ( Figure S9 and Table S3). This finding suggests that the doping of Zn 2+ into the perovskite NCs is an energy absorbing process. However, our exchange reaction was carried out at room temperature and the reaction time was short, which did not allow for the doping of Zn 2+ into the lattice. Therefore, Zn 2+ detected by XPS after the anion exchange should be present on the surface of the perovskite NCs as an inorganic ligand, which is consistent with the previous report [41]. The entire exchange processes from the pristine CsPbBr 3 NCs to the CsPbBr x I 3−x NCs are illustrated in Figure 4c. After the treatment of CsPbBr 3 NCs with ZnI 2 precursors, the exchange reaction can be divided into three steps. Firstly, the addition of ZnI 2 precursor disrupts the balance of ligands on the pristine green CsPbBr 3 NCs, and massive desorption of the original ligands results in the destruction of lattice structure at the NCs surface, leading to abundant surface defects, including halogen vacancies. Secondly, the formed defective surface triggers the rapid halide-exchange to produce CsPbBr x I 3−x NCs; simultaneously, the ligand-exchange is taking place, and the Zn 2+ combined with halogen ions can be anchored as the inorganic ligands to passivate the surface defects. Finally, the halide-exchange slows and then reaches an equilibrium followed by the ligand-exchange equilibrium in the CsPbBr x I 3−x NCs.
Mixed halide perovskites are known to undergo photo-induced halide segregation, which will result in the spectral instability [42][43][44]. In order to study the photostability of the red CsPbBr x I 3−x NCs, we measured the PL spectra of the perovskite films under intense illumination of ultraviolet light. A simplified schematic of the measurement equipment is shown in Figure 5a. The PL spectra and their peak intensities of the 0.1 M ZnI 2 NCs did not exhibit any change after 5 h of irradiation, indicating the absence of phase separation caused by photo-induced halide segregation, as depicted in Figure 5b. Encouraged by the exceptional photostability and high PLQY of the obtained red perovskite NCs, we prepared the white light-emitting diode (LED) by combining the 0.1 M ZnI 2 NCs and pristine green CsPbBr 3 NCs with a commercial blue LED chip. In Figure 5c, it is evident that there are three distinct emission peaks which are assigned to the corresponding blue chip, green CsPbBr 3 NCs and red 0.1 M ZnI 2 NCs. Notably, there is no intermediate emission peak detected, indicating that the perovskite NCs were effectively separated in the PMMA matrix to block the anion exchange between the green and red NCs. Figure 5d displays the CIE chromaticity coordinates of the blue chip, pristine green CsPbBr 3 NCs and 0.1 M ZnI 2 NCs. The selected CIE color coordinate triangle, including the corresponding blue, green and red monochromatic emission, encompasses 120% of the National Television System Committee's (NTSC) latest standard (Rec. 2020). This demonstrates that these green and red perovskite NCs, which are easily obtained by LARP and ion-exchange, respectively, can be used as the promising low-cost emitters in wide color gamut display devices.

Conclusions
In this study, we synthesized efficient and spectrally stable red CsPbBrxI3−x NCs by a facile ion-exchange technique of post-processing CsPbBr3 NCs with ZnI2 in solution. The in situ PL measurements were conducted to probe the exchange procedures. It was ob-

Conclusions
In this study, we synthesized efficient and spectrally stable red CsPbBr x I 3−x NCs by a facile ion-exchange technique of post-processing CsPbBr 3 NCs with ZnI 2 in solution. The in situ PL measurements were conducted to probe the exchange procedures. It was observed that the halide-exchange will rapidly arise and then gradually decelerate to achieve an equilibrium. And the original organic ligands were found to partially desorb from the surface of the perovskite NCs to trigger the halide-exchange in a very short time, and simultaneously facilitate the ligand-exchange. More importantly, zinc halides can be anchored on the surface of perovskite NCs as the inorganic ligands to reduce the vacancy defects of halide anions formed during the ion-exchange process, which is beneficial to stabling the emission spectra. The synergistic effect of the halide-exchange and ligandexchange enabled the mixed-halide red perovskite NCs with the hybrid organic/inorganic ligands to achieve a stable emission peak at 640 nm with a narrow FWHM of 33 nm and a PLQY close to unity. Ultimately, we combined the exchanged red perovskite NCs and the pristine green perovskite NCs with a commercial blue LED chip to fabricate the white device with wide color gamut and high-color saturation. Our study presents a cost-effective strategy to develop efficient and stable perovskite NCs for applications in full-color displays.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/nano13162337/s1, Figure S1. TA features for (a) 0.08 M ZnI 2 NCs and (b) 0.1 M ZnI 2 NCs. Figure S2. The fitted 2D PL spectra of the treated perovskite NCs over time, with w/o M ZnX 2 (0-120 s) (a), with 0.1 M ZnBr 2 (0-60 s) (b). Color scale bar is provided on the right side for the intensity of the PL. Figure S3. UV-vis absorption and PL spectra of the pristine CsPbBr 3 NCs and the exchanged CsPbBr 3 NCs treated with precursor solutions of without Zn 2+ (w/o ZnX 2 ) and with ZnBr 2 (0.1 M ZnBr 2 ). Figure S4. Size histogram of the exchanged CsPbBr 3 NCs: (a) w/o ZnX 2 and (b) 0.1 M ZnBr 2 (inset: TEM images of the perovskite NCs). Figure S5. PLQY spectra of the pristine CsPbBr 3 NCs and the exchanged CsPbBr 3 NCs of w/o ZnX 2 and 0.1 M ZnBr 2 in solutions. Figure S6. XRD spectra of the films of the pristine CsPbBr 3 and the exchanged CsPbBr 3 NCs of w/o ZnX 2 and 0.1 M ZnBr 2 . Figure S7. XPS spectra of the perovskite NCs powders of the pristine CsPbBr 3 and the exchanged CsPbBr 3 NCs of w/o ZnX 2 and 0.1 M ZnBr 2 : survey spectra (a), Zn 2p spectra (b). Figure S8. XPS spectra of perovskite NCs powders: survey spectra (a), N/Pb atomic ratios in the perovskite NCs calculated from the XPS data (b), Zn 2p spectra (c), I 3d spectra (d). Figure S9. Schematic of the crystal structure of CsPbBr 3 and CsPbI 3 NCs before and after Zn 2+ doping. Table S1: Summary of the optical properties of the different NCs. Table S2: Fitted PL lifetimes of different NCs solution measured by using the bi-exponential decay: I (t) = A 1 exp(−t/τ 1 ) + A 2 exp(−t/τ 2 ). Table S3: The calculated formation energies (∆E) of the Zn-doping perovskite NCs based on the crystal structure as shown in Figure S9.

Data Availability Statement:
The data that support the findings of this study are available from the corresponding author upon reasonable request.