The chiral ZIF was synthesized by the mixed-ligand coassembly approach. For example, L-His and Hmim were dissolved in mixed solvent of methanol and water. Then, a small amount of triethylamine (TEA) was added into solution under stirring. After stirring for 10 minutes, the mixed-ligand solution was gradually added to the methanol solution of Zn(NO₃)₂·6H₂O. Upon stirring at room temperature for 24 h, the colorless product was washed with large amounts of water and methanol, then collected by centrifugation and dried in a vacuum. First of all, powder X-ray diffraction (PXRD) was used to confirm the crystal structure of L-ZIF. As showed in Figure S1a, the PXRD patterns of L-ZIF were in consistence with those of ZIF-8 with identical positions, which indicated that the ZIF-8-like architecture with SOD topology formed in the framework of L-ZIF. However, the first-order diffraction peak of L-ZIF exhibited a slight shift to a larger angle which suggested a slightly smaller unit cell than that of ZIF-8 after incorporating with L-His (Figure S1b). Meanwhile, scanning electron microscope (SEM) images of L-ZIF exhibited similar rhombic dodecahedral shape with ZIF-8. To confirm the successful incorporation of L-His into the L-ZIF structure, Fourier-transform infrared (FT-IR) spectroscopy and X-ray photon spectroscopy (XPS) were systematically measured. As for FT-IR spectroscopy, L-ZIF showed a wide peak at 1629 cm^-1 which could ascribe to the C=O stretch in the carboxyl group from L-His. Additionally, vibration bands were also found around 1116 cm^-1 and 1058 cm^-1 in L-ZIF and L-His which were not observed in ZIF-8. XPS analysis identified the binding energy of the carboxylic group (-COOH) of L-ZIF at 288.75 eV from the C(1s) spectrum, while it was not presented in ZIF-8 (Figure S2). These results clearly indicated the formation of mixed-ligand chiral ZIF. To examine the chiral ligand loading efficiency, various ingredient molar ratios between L-His and Hmim from 1 : 7 to 1 : 2 were investigated. The final average molar ratio of these two ligands was determined by ¹H NMR measurements (Figures S2c and S3). The results illustrated that although different amounts of L-His were added, a certain molar ratio between L-His and Hmim is 0.1 in final products. Considering the decreased crystallization caused by the large amount of L-His during the synthetic process (Figure S4), the molar ratio of n_{L‐\text{His}}/n_{\text{Hmim}} was fixed to be 1/7 in the subsequent experiments. Comprehensively considering the above results, we argue that L-His and Hmim are most likely randomly distributed in the extended polymeric framework skeletons of L-ZIF. According to the crystal structure of ZIF-8, it is clear to deduce that 1 unit cell contains 2 cages with 24 Hmim ligands, while every cage would combine 2-3 L-His as linkers (Figure S5). In addition, D-ZIF consisted of Zn ion, D-His, and Hmim showed similar properties as L-ZIF (Figures S6–S8). The chirality of obtained mixed-ligand ZIF was tested by circular dichroism (CD) measurement. As shown in Figure S9a, L-His exhibited a positive signal in methanol. Interestingly, a negative signal was observed in L-ZIF with a larger dissymmetry factor g_{\text{CD}} (Figures S9b and S9c). The same results were also obtained between D-ZIF and D-His. We reason that the observed reversal CD signal of the chiral ZIF should be due to the supramolecular chirality induced by the histidine unit. A similar phenomenon can be found widely in supramolecular self-assembled systems [55]. Thus, the combination of chiral ligand would endow the ZIF-8 with chirality.

Nitrogen adsorption analysis results showed that the Brunauer-Emmett-Teller (BET) surface area of L-ZIF was 1267.9 m² g^-1 with a micropore volume of 0.37 cm³ g^-1, indicating that the L-ZIF is of a typical porous structure with a high surface area. By applying an in situ synthetic approach in which the dye molecules were simultaneously encapsulated into the framework during the synthesis of L-ZIF, dye-loaded chiral ZIF could be obtained. After stirring at room temperature for 24 h, the crude products were collected by centrifugation and extensively washed with large amounts of good solvent of dyes. Then, the residues were centrifuged and washed by methanol for three times followed by drying in a vacuum. From the color change or emission behavior under UV light of the ZIF powder, we could easily and quickly select appropriate guest molecules. A great number of screening tests suggested that the cationic dyes like Rhodamine B, thioflavin t (ThT), and 4-(4-dimethylaminostyryl)-1-methylpyridinium (DASPI) were difficult to be encapsulated into the cationic L-ZIF. On the contrary, the linear guest molecules such as anionic stilbene 420 (S 420), neutral coumarin 6 (C6), and 4-(dicyanomethylene)-2-methyl-6-(4-dimethylaminostyryl)-4H-pyran (DCM) could be loaded into L-ZIF easily. Therefore, the anionic or neural dyes with suitable size were required to be encapsulated into the L-ZIF framework. Taking DCM as an example, it was found that after loading into the L-ZIF, the BET surface area of L‐\text{ZIF}\supset \text{DCM} exhibited an obvious decrease to 1153.9 m² g^-1 (Figure 1(a)). This demonstrated that dye molecules were successfully encapsulated into the cavity of L-ZIF. In addition, the loading amount of DCM was 0.04 wt% determined by fluorescence spectroscopy (Figure S27). The PXRD patterns of L‐\text{ZIF}\supset \text{DCM} were similar to those of L-ZIF, which indicated that the introduction of guest dyes did not damage the crystalline structure. In addition, the slight shift to a small angle at the first-order diffraction peak demonstrated that the loading DCM enlarged the size of the unit cell due to large molecule size (Figure 1(b)). As showed in the SEM image, the morphology of L‐\text{ZIF}\supset \text{DCM} has no obvious change compared with that of L-ZIF. With or without guest DCM molecules, only uniform rhombic dodecahedron structures could be observed (Figure 1(c)). Thus, DCM molecules were effectively encapsulated into the L-ZIF framework interior leading to the stable and clean morphologies. The test of L‐\text{ZIF}\supset \text{DCM} in solid powder by confocal laser scanning microscopy (CLSM) could further confirm that DCM molecules were effectively encapsulated into the L-ZIF framework (Figure S10).

It was well known that DCM was a kind of aggregation-caused quenching (ACQ) dye; the photophysical properties of DCM and DCM encapsulated into L-ZIF in the solid state were studied. The fluorescence spectra of the L‐\text{ZIF}\supset \text{DCM} showed a large hypochromatic shift from 640 nm to 578 nm compared with the one of DCM powder (Figure 1(d)), which clearly indicated the good molecular distribution in L-ZIF cages. Furthermore, a high quantum yield (0.43) of L‐\text{ZIF}\supset \text{DCM} in the solid state was detected, while a poor quantum yield (0.02) was obtained in DCM powder (Table S1). On the one hand, the molecular distribution of dye molecules in isolated L-ZIF cages can avoid intermolecular interactions, which will prohibit severe emission quenching. On the other hand, the good confinement of the DCM molecules within the size-matched cages of L-ZIF can effectively restrain the intramolecular torsional motion and increase the conformational rigidity of the dyes. Thus, the decreased ACQ effect and the improved radiative decay pathway result to the high emission efficiency of L‐\text{ZIF}\supset \text{DCM}. This could be further confirmed by the fluorescence lifetime testing. As showed in Figure 1(e), the fluorescence lifetime spectrum of DCM in the solid state was double-exponential decay with 1.8 ns. However, after being confined into the L-ZIF framework, the fluorescence decay was single-exponential mode up to 2.3 ns.

Remember that, due to the introduction of chiral histidine into the ZIF framework, the induced chirality of loaded DCM was expected. The CD spectra of L‐\text{ZIF}\supset \text{DCM} showed a negative signal at the absorption band of DCM, which followed the chirality of L-ZIF (Figure S11a). In addition, DCM-loaded D-ZIF was also successfully achieved using the same way (Figure S12). Interestingly, D‐\text{ZIF}\supset \text{DCM} exhibited a positive CD signal, which was opposite to L‐\text{ZIF}\supset \text{DCM}. This indicated the authentic induced chirality from chiral confined space to achiral dyes. The guest dye DCM-loaded L-/D-ZIF showed significant fluorescence enhancement which encourages us to investigate the CPL activity. Amazingly, strong CPL signals were detected from these host-guest systems with a completely opposite curve (Figure 1(f)), while no obvious signal could be observed in \text{ZIF}‐8\supset \text{DCM} (Figure S11b). The magnitude of CPL can be evaluated according to the dissymmetric factor (g_{\text{lum}}), which is defined as g_{\text{lum}}=2\times \left(I_{\text{L}}‐I_{\text{R}}\right)/\left(I_{\text{L}}+I_{\text{R}}\right), in which I_{\text{L}} and I_{\text{R}} represent the intensity of left- and right-handed circularly polarized light, respectively. The maximum g_{\text{lum}} value ranges from +2 to -2. The calculated value of g_{\text{lum}} from L‐/D‐\text{ZIF}\supset \text{DCM} is about \pm 1.2\times {10}^{-3}, which is a comparable value for organic small molecules. It must be mentioned that the induced chirality of DCM could be regulated by the chirality of L-/D-ZIF matrixes. These results revealed that the initial chirality of chiral matrixes played a vital role for the chiral transfer.

Encouraged by these results, the encapsulation of the blue- and green-emitting guest molecules into L-ZIF was further investigated. Fluorescent brightener (S420) and the coumarin derivative (C6) in methanol solution display intense blue and green emission bands centered at about 423 nm and 508 nm, respectively (Figure S13). Fortunately, they could also be easily loaded into the cavities of L-/D-ZIF, while the emission properties were similar to the above-demonstrated L‐/D‐\text{ZIF}\supset \text{DCM} (Figures S14–S22). Considering that the anionic S420 may coordinate with Zn ion during the synthetic process, we have carefully investigated the additive amount of S420 in the reaction mixtures. It was found that high dye concentration (over 2 mM) not only affected the chirality of final L‐\text{ZIF}\supset \text{S}420 but also led to the damage of morphology as well as crystallinity of L‐\text{ZIF}\supset \text{S}420 (Figures S16 and S17). Thus, after optimizing the usage of S420 in the fabrication process, we fixed the concentration to 1 mM. Under UV irradiation, the powder of L‐\text{ZIF}\supset \text{S}420, L‐\text{ZIF}\supset \text{C}6, and L‐\text{ZIF}\supset \text{DCM} exhibited bright blue, green, and orange emission, respectively (Figures 2(a) and 2(b)). Furthermore, the fluorescence spectrum of L‐\text{ZIF}\supset \text{S}420 showed blue emission bands centered at 426 nm with enhanced emission efficiency from 37% to 59% compared with S420 powder (Table S1). A similar phenomenon was observed from chiral \text{ZIF}\supset \text{C}6, and the highest emission enhancement was observed in D‐\text{ZIF}\supset \text{C}6 with quantum yield 76%. These further indicated that the guest dyes were homogeneously distributed inside the L-ZIF framework without severe aggregation. Additionally, induced CPL could be observed in all the tested chiral \text{ZIF}\supset \text{dyes} and were found to exhibited mirror-image signal from 426 nm to 578 nm (Figure 2(c)). The \mid g_{\text{lum}}\mid values obtained from chiral ZIF-encapsulated S420 and C6 were 9.0\times {10}^{-4} and 3.0\times {10}^{-4}, respectively. In addition, the Commission International de l’Éclairage (CIE) coordinates calculated from L‐\text{ZIF}\supset \text{S}420, L‐\text{ZIF}\supset \text{C}6, and L‐\text{ZIF}\supset \text{DCM} were (0.15, 0.07), (0.19, 0.59), and (0.49, 0.50), respectively (Figure S23), which covered the blue via green to orange circularly polarized emission. To date, this kind of full-color covering CPL systems was rarely reported in the solid state especially in these chirality-induced host-guest systems.

Materials with white light emission are particularly important, and plenty of studies on achieving white emission MOFs have been reported [38, 56, 57]. However, to date, MOFs with white-color circularly polarized light emission have never been reported. Based on the previous results, it was found that L‐\text{ZIF}\supset \text{S}420, L‐\text{ZIF}\supset \text{C}6, and L‐\text{ZIF}\supset \text{DCM} showed blue, green, and orange emission, respectively. Encouraged by these results, if S420, C6, and DCM are encapsulated into L-ZIF compatibly, white light can be realized through balancing the contents of their emissions. Encouragingly, the L-ZIF composite, L‐\text{ZIF}\supset \text{S}420/\text{C}6/\text{DCM}, was obtained simply by adding mixed solution of S420, C6, and DCM during the synthesis process. According to the screening tests, white light-emitting L‐\text{ZIF}\supset \text{S}420/\text{C}6/\text{DCM} was achieved under excitation at 370 nm (Figure 2(d)). The fluorescence spectrum of L‐\text{ZIF}\supset \text{S}420/\text{C}6/\text{DCM} covered almost the whole visible spectral region, with the CIE chromaticity coordinates measured as (0.33, 0.33), the same as that of the ideal white light (Figure 2(e)). Furthermore, the white light is kept when varying the excitation wavelength from 335 nm to 380 nm (Figure S24a). Most importantly, for the first time, we succeeded in achieving circularly polarized white light emission in a luminescent MOF system (Figure S24b). A reasonably large spectral width of 400 to 750 nm was observed in the CPL spectra. After converting to g_{\text{lum}} spectra, nearly parallel diagrams were obtained (Figure 2(f)). The luminescence dissymmetry factor \mid g_{\text{lum}}\mid was round 1\times {10}^{-3}, which was the same as those of the individual chiral \text{ZIF}\supset \text{dye}. This method thus provides a simple and general approach for fabricating luminescent MOFs that emit white circularly polarized light.

Besides organic dyes, the incorporation of semiconductor QDs in chiral ZIFs has been also achieved in this work. For example, during the preparation of L‐\text{ZIF}\supset \text{QD}533 (with green-color emission) composites, the functionalization of the QD surface with polyvinylpyrrolidone (PVP) was firstly carried on. Then, L‐\text{ZIF}\supset \text{QD}533 composites were prepared by the crystallization of L-ZIF in methanol at room temperature in the presence of PVP-modified QD533. The optimizing concentration of QD533 during the synthesis process was 0.3 mg/mL (Figure S30a). PXRD indicated that the crystal structure of L-ZIF remained after the incorporation with QDs. Although the CD signal of L‐\text{ZIF}\supset \text{QD}533 was not detected due to the strong scattering, the intense CPL signal of L‐\text{ZIF}\supset \text{QD}533 can be observed at the corresponding emission region (Figure S28). Interestingly, the CPL of QD533 in L-ZIF exhibited a positive signal with g_{\text{lum}} value 4.6\times {10}^{-3} at 533 nm, an opposite and relatively larger value than that of L‐\text{ZIF}\supset \text{dye} (Table S2). Considering that the size of QD533 was much larger than the cage size of L-ZIF, QDs were beyond the cages and dispersed into the crystals of L-ZIF. Therefore, the inverted CPL signals resulted from the induction outside of the surrounding cages. Otherwise, D‐\text{ZIF}\supset \text{QD}533 composites were fabricated, and a negative sign of CPL was detected with the same order of magnitudes of g_{\text{lum}} value -4.3\times {10}^{-3} as L‐\text{ZIF}\supset \text{QD}533. This indicated that the induced chirality of QDs could be controlled by the inherent chirality of chiral ZIFs. In addition, the wavelength of CPL could be tuned by simply changing the type of QDs (Figures S29 and S30). As shown in Figures 3(a) and 3(b) and S29b, the color of CPL from chiral \text{ZIF}\supset \text{QD} composites could cover the range from blue via cyan, green, and orange to red in the solid state. To date, this kind of full-color covering CPL-active systems from semiconductor QDs in the solid state has rarely been reported due to the tedious synthesis. To further confirm the origin of chirality of QDs, transmission electron microscopy (TEM) imaging of \text{ZIF}\supset \text{QD}533 was performed (Figure 3(c) and S30c). The TEM images showed that QD533 was encapsulated inside the L-ZIF particles while the high-resolution TEM (HRTEM) image indicated that the original crystal structure of QD533 was preserved. These results illustrated that the detected CPL signal of chiral \text{ZIF}\supset \text{QD}533 resulted from the induced chirality by confined chiral environment from chiral ZIFs [58, 59]. In addition, through tuning the mass ratio of five kinds of QDs loaded in chiral ZIF, chiral \text{ZIF}\supset \text{QDs} with white circularly polarized light emission also could be obtained in the solid state. As showed in Figures 3(d)–3(f) and S31, white circularly polarized light-emitting chiral ZIFs showed mirror-image CPL signal from 400 nm to 700 nm with a \mid g_{\text{lum}}\mid value around 5.0\times {10}^{-3}.

Encouraged by the results from QDs, one type of lanthanide-loaded upconversion nanoparticles (UCNPs)—NaYF₄:Yb, Er—was also encapsulated into the chiral ZIFs by a similar strategy. Excitingly, the upconverted CPL (UC-CPL) could be observed under the excitation of 980 nm laser (Figure 4(a)). Firstly, the photoluminescence of UCNP:Er was not changed after loading into L-ZIF, and the crystallinity of L-ZIF was reserved (Figure S32). Additionally, the upconverted emission intensity of L‐\text{ZIF}\supset \text{UCNP}:\text{Er} was found to be closely dependent on the excitation power density with the slope of 2.0 at 541 nm, which indicated a two-photon absorption-based upconversion (Figure S33). As showed in Figure 4(c), the mirror-image UC-CPL could be observed from 390 nm (navy) to 700 nm (red). More intriguingly, due to the relatively large g_{\text{lum}} value (\text{~}\pm 1.2\times {10}^{-2}) of UC-CPL [12], even some emission peaks showed lower intensity under the same excitation, a distinct CPL signal could be detected (Figure 4(c) and Table S3). Similar to the QD-loaded ZIF, TEM images showed that UCNP:Er was distributed inside the ZIF particles (Figures 4(b) and S34), which indicated that the induced UC-CPL was resulting from the chiral confinement effect of chiral ZIFs.

All reagents and solvents were used as received, unless otherwise indicated. Zinc nitrate hexahydrate and 2-methylimidazole were purchased from Sigma-Aldrich. Polymethylpyrrolidone (PVP, \text{Mw}=10000), L-histidine, and D-histidine were purchased from TCI (Shanghai) Development Co., Ltd. YbCl₃·6H₂O (99.9%), YCl₃·6H₂O (99.99%), and ErCl₃ (99.9%) were purchased from Innochem (Beijing) Co., Ltd. All the used quantum dots were purchased from Suzhou Xingshuo Nanotech Co., Ltd.

The ¹H NMR spectra were recorded on a Bruker Fourier 400 (400 MHz) spectrometer. X-ray diffraction (XRD) was achieved on a Rigaku D/Max-2500 X-ray diffractometer (Japan) with Cu/Kα radiation (\lambda =1.5406 Å), and XPS was measured on an ESCALAB250XI spectrometer. Nitrogen adsorption isotherm was tested on a ASAP2420-4 instrument at 77 K. Transmission electron microscopy (TEM) and high-resolution transmission electron microscopy (HRTEM) images were taken on a Tecnai G2 F20 U-TWIN microscope (200 kV); before testing, the samples were suspended on carbon-coated Cu grids. Scanning electron microscopy (SEM) was performed on a Hitachi S-4800 FE-SEM with an accelerating voltage of 10 kV. Before SEM proceeded, the samples on silicon wafers were coated with a thin layer of Pt to increase the contrast. FT-IR spectra were collected on the S-ONE spectrometer utilizing an HATR mode. UV-vis spectra were obtained using a Hitachi U-3900 spectrophotometer, and fluorescence spectra were measured on a Zolix Omni-λ500i monochromator with photomultiplier tube PMTH-R 928 using a Xe lamp as the excitation source. CD spectra were measured on a JASCO J-815 spectrophotometer. CPL spectra were obtained using a JASCO CPL-200 spectrophotometer. Confocal laser scanning microscopy was recorded on the Olympus FV1000-IX81. The absolute fluorescence quantum yield was measured by using an absolute PL quantum yield spectrometer (Edinburg FLS-980 fluorescence spectrometer) with a calibrated integrating sphere, and fluorescence lifetime measurements were recorded on a NanoLog spectrometer using e chiral confinement effect of chiral.

Further details of the materials and methods are included in Supplementary Materials (available here).