Compounds 1 and 2 were prepared through typical wet-chemistry solution methods with the same precursor materials in different mixed solvents. Based on single-crystal X-ray diffraction, crystal structures of 1 and 2 crystallize in the monoclinic space groups of P2₁/n and C2/c containing [Sb₄Cl₁₆]⁴⁻ tetramer and [SbCl₅]²⁻ square pyramid as basic building units, respectively. As shown in Fig. 1A, the asymmetric unit of 1 contains two crystallographically independent Sb atoms, both of which are coordinated by five Cl atoms in square pyramidal coordination geometries with Sb–Cl bond distances in the large range of 2.3818(5) to 2.9693(5) Å. [Sb(1)Cl₅] and [Sb(2)Cl₅] units are condensed via edge-sharing to form [Sb₂Cl₈] dimer, which is further interconnected via weak Sb–Cl bonds (3.165 Å) into [Sb₄Cl₁₆]⁴⁻ tetramer charge balanced by [Ph₃EtP]⁺ cation. The crystal structure of compound 2 contains discrete [SbCl₅]²⁻ square pyramids with normal Sb–Cl bond distances of 2.3688(4) to 2.6209(3) Å, which are spatially separated by [Ph₃EtP]⁺ cation with guest EtOH locating in the free space (Fig. 1B). Differently, single crystal of compound 3 was obtained from the SC-SC transformation of 2 via solid-state recrystallization reaction. Along with the lost of EtOH molecule, the structural reassembly leads to another type of packing arrangement of [SbCl₅]²⁻ and [Ph₃EtP]⁺ blocks in triclinic space group (P-1) accompanied by variant Sb–Cl bond distances (2.240-2.691 Å) (Fig. 1C). Overall, all the compounds 1 to 3 belong to typical 0D hybrid structures with discrete [Sb₄Cl₁₆]⁴⁻ or [SbCl₅]²⁻ as luminescent center periodically embedding in the organic [Ph₃EtP]⁺ matrices. As a consequence, different structural assembly characteristics including geometry configuration, distortion degree, and stacking manner of anionic species synergistically induce the distinct PL performance of these 0D hybrid halides as following detailed discussion (Figs. S1 to S3). The sample purity of compounds 1 to 3 was validated by coincident powder X-ray diffraction (PXRD) patterns and simulated data from single-crystal X-ray diffractions (Fig. S4). The elemental mapping from scanning electron microscopy (SEM) demonstrates the presence and uniform distribution of Sb, Cl, C, and P elements on the surface of block single crystals for all compounds (Figs. S5, S15, and S16).

All the transparent bulk crystals of compounds 1, 2, and 3 are colorless under ambient sunlight, but 1 is completely nonluminescent, while 2 and 3 exhibit strong yellow and red light emissions under irradiation of 365 nm ultraviolet (UV) light as shown in Fig. 1D to F. To obtain the specific comprehension, the optical properties of compounds 1 to 3 are characterized systematically based on solid-state absorption, excitation, and emission spectra. As shown in Fig. 1G to I and Fig. S6, the UV–visible (UV–Vis) absorption spectra of compounds 1 to 3 show multiple absorbance bands, which can be attributed to the electron conversion between the sp excited state and s² ground state, including the low-energy spin-forbidden transition of ¹S₀-³P₁, the intermediate-energy spin-forbidden transition of ¹S₀-³P₂, and the high-energy spin-allowed transition of ¹S₀-¹P₁ [45]. The absorption features for 1 to 3 are similar to those of (C₆H₅CH₂NH₃)₃SbBr₆ and [(C₆H₂₂N₄)₂(Sb₂Cl₁₀)(SbCl₆)(Cl)₂(H₃O)]·(3H₂O) [46,47]. Best fitting based on Tauc function gives direct bandgaps of 3.26, 2.95, and 2.82 eV for 1 to 3, respectively (Fig. S6). All the PL excitation (PLE) spectra of compounds 1 to 3 feature two dominant bands corresponding to electronic transition of ¹S₀ → ¹P₁ (below 300 nm) and ¹S₀ → ³P₁ (above 300 nm), respectively, which are in accord with the UV−vis absorption spectra [48]. Subsequently, the strongest excitation wavelength is selected as light source to investigate the PL emission properties. Despite the invisible luminescence of compound 1 by the naked eye, spectroscopy analysis gives a very weak broadband green light emission at about 530 nm (CIE chromaticity coordinates: 0.29, 0.43) with negligible PLQY of ~0% (Fig. 1G). Based on previously reported work, the weak luminescence of 1 is mainly related to the higher nonradiative transition rate due to strong concentration-caused quenching effect in [Sb₄Cl₁₆]⁴⁻ tetramer [29,49–52]. Specifically, the shorter Sb···Sb distance (4.021 and 4.148 Å) within [Sb₄Cl₁₆]⁴⁻ tetramer enhances the electronic coupling and nonradiative energy transfer between adjacent [SbCl₅] center, and therefore, larger concentration of optically active centers quenches the radiative recombination. Distinctly, compound 2 displays a perfect Gauss-shaped broadband yellow light emission centered at 598 nm with Stokes shift of 228 nm, full width at half maximum (FWHM) of 123 nm, and CIE chromaticity coordinates of (0.52, 0.47) (Fig. 1H). Comparing with the negligible PLQY of compound 1, the strong yellow emission of compound 2 is verified by near-unity PLQY of 91.5%, which represents a high value in all yellow light-emitting perovskites (Fig. S8 and Table S1). To reveal the dynamic photophysical process, the time-resolved PL spectrum monitoring the maximum emission wavelength is performed at 300 K. The PL decay curve of compound 1 can be fitted by a double-exponential function giving an average lifetime of 3.45 μs (Fig. S7). For compound 2, the PL decay curve delivers single exponential function fitting with lifetime of 3.527 μs (Fig. 2A). Relative to yellow-emissive compound 2, the PL spectrum of compound 3 displays a significant red-shift with maximum emission wavelength of 642 nm, Stokes shift of 277 nm, FWHM of 146 nm, and CIE chromaticity coordinates of (0.60, 0.39) corresponding to red light (Fig. 1I). The larger Stokes shift and wider FWHM of 3 than 2 can be interpreted by the higher distortion level of anionic [SbCl₅]²⁻ units as illustrated in Table S6, and such phenomenon has been confirmed in many 0D hybrid halide perovskites [53]. The strong red light emission of compound 3 is proved by promising PLQY of 64.47%, which is comparable with the vast majority of red light-emitting 0D hybrid metal halides [53–55]. Based on time-resolved spectrum, the PL decay curve is fitted by a single exponential function giving a lifetime of 4.364 μs at 300 K, which is slightly longer than that of compound 2 due to the larger distortion level of [SbCl₅]²⁻ unit (Fig. 2G).

In order to confirm the PL nature of these broadband light emissions, varied wavelength-dependent spectroscopy was performed for compounds 2 and 3 as examples due to the highly efficient emitting performance. Apparently, there is no obvious difference in the PLE and PL spectra under varied emission and excitation wavelength for compound 2, respectively, corresponding to one dominant emitting center in 3D PL map, which suggests the yellow light emission originates from the relaxation of single excited state (Fig. 2B and C and Fig. S9). Compound 3 also delivers the same dominant red light emission in the whole UV light spectral range accompanied by one additional weak higher-energy emission at about 500 nm under the excitation of higher-energy UV light (<350 nm) (Figs. S10 to S12). We further compare the PL emission spectra of bulk crystals and microscale powders for both compounds considering the fact that the surface defect-associated broadband light emission shows a strong dependence on particle size and is readily quenched by particle aggregation (Figs. S13 and S14). The identical PL emission spectral profiles with slightly decreased intensities of microscale powders exclude the possibility of surface defect-generated broadband light emissions (Figs. S15 and S16) [56]. Additionally, the PL emission intensities of bulk crystals exhibit linear dependence on the excitation power, which rule out the emissions from permanent defects (Fig. 2D and Fig. S17) [57]. To further shed light on the underlying mechanism of broadband light emission, temperature-dependent PL emission spectra were carried out from 80 to 300 K (Fig. 2E and H). In the cooling process, there is not any splitting behavior in the emission spectra, further indicating that the broadband light emissions should originate from the single excited state. Along with the temperature decreasing, the emission intensity of compound 2 initially increases from 300 to 240 K and then slowly decreases until 80 K (Fig. S18). The decreased trend of emission intensity from 240 to 80 K is termed as anti-thermal quenching, which is related to the participation of shallow trapped states arising from the structural defects based on the previously reported work, such as [DMPZ]₂SbCl₇∙2H₂O (DMPZ = N,N′-dimethylpiperazine) and (C₉NH₂₀)₂SnBr₄, Mn@CsPbCl₃ [49,58,59]. For compound 3, the dominant emission wavelength exhibits a slight red-shift monotonically from 642 to 664 nm with temperature decreasing, which can be attributed to the electron–phonon interaction (Fig. S19). That is, the electron–phonon interaction weakens as the temperature decreases, resulting in a reduction of bandgap and red-shift of emission wavelength. At the same time, the emission intensity gradually increases due to the inhibited nonradiative recombination. Based on the theoretical fitting of temperature-dependent emission intensity according to famous Arrhenius-type equation, the exciton activation energy (E_a) is calculated to be 52.93 meV, which is much larger than the room-temperature thermal ionization energy (~26 meV) [60]. Higher E_a illustrates that excited electrons difficultly overcome the energy barrier of nonradiative transitions, thus generating stable STEs and strong luminescence at room temperature. Notably, both the PL emission spectra of compounds 2 and 3 broaden gradually with temperature increasing maybe due to the enhanced electron–phonon coupling effect in soft crystal lattice. To quantify the contribution of electron–phonon coupling, the carrier–phonon coupling constant (Γ_phonon) and the corresponding phonon energy (E_LO) can be estimated by theoretically fitting the temperature-dependent FWHM according to the longitudinal optical phonon broadening model:where Γ₀ is the FWHM of 0 K and E_LO and E_b represent the energy of the longitudinal–optical phonon and average binding energy of the trap states, respectively. Γ_phonon and Γ_inhomo represent the relative contributions of electron–phonon coupling and inhomogeneous broadening induced by trapped states, respectively [61,62]. The best fittings give Γ_phonon of 119 and 93.1 meV for compounds 2 and 3, respectively, which are comparable with most recently reported 0D halides demonstrating predominant contribution of electron–phonon interaction for the PL emission broadening (Fig. 2F and I). In addition, calculated phonon frequency energies (E_LO) of 17.5 and 16.9 meV are well consistent with the intense stretching bands of 144 and 139 cm⁻¹ in Raman spectra, respectively, which further attest the strong electron–phonon interaction in deformable crystal lattice (Fig. S20). Based on the above systematical investigations, these broadband light emissions with large Stokes shifts, wide FWHM, and long lifetime can be ascribed to the radiative recombination of STEs due to the strong electron–phonon coupling arising from local deformation of crystal lattice upon photoexcitation.

To confirm this hypothesis, we performed first-principles calculations on compounds 2 and 3, simultaneously aiming to get deeper insight into the PL properties. As shown in Figs. S21 and S22, the calculated band structures give direct energy bandgaps of 2.64 and 2.56 eV, which are smaller than the experimental values from absorption spectra (2.95 and 2.82 eV) due to the limitation of software. Both the valance and conduction bands are fairly flat at ground states, indicating the absence of electronic interactions among adjacent metal halide species. The flat frontier orbitals illustrate highly localized carriers with large effective hole and electron masses, which facilitate the formation of bond excitons [49,63]. Based on the total and partial density of states, both the conduction band minimum and valence band maximum are mainly composed of Cl-3p and Sb-5p orbitals from anionic metal halide species. Although [Ph₃EtP]⁺ participates in frontier orbitals, the strong coulomb binding leads to the localization of the carriers within anionic metal halide species [51,64]. The photo-excited electrons significantly induce the dynamic structural deformation of discrete [SbCl₅]²⁻ units laying the foundation of lower-energy triplet STE-related broadband light emission. Subsequently, the first-principles geometry optimizations at the ground and excited state were performed to evaluate the structural deformability of anionic [SbCl₅]²⁻ models in order to unveil their distinct PL emission efficiency [65,66]. Upon the photo-excitation, both the [SbCl₅]²⁻ units deliver structural deformation at excited states for compounds 2 and 3 as demonstrated by the variant bond lengths and angles in Fig. 3A and B and Tables S4 and S5. For the purpose of quantitative comparison, the structural deformability is evaluated by the change of distortion degrees based on varied Sb–Cl bond lengths (Δd) and Cl–Sb–Cl bond angles (σ²) based on the following equations:

where d_n is the individual Sb−Cl bond length and d represents the average bond length, and θ_n denotes the bond angle of each Cl−Sb−Cl, respectively. Remarkably, the Δd and σ² values of 1.53 × 10⁻³ and 3.935 at ground state dramatically change to 3.52 × 10⁻³ and 198.169 at the excited state, indicating the higher structural deformability of discrete [SbCl₅] unit in compound 2 (Table S6). Relatively speaking, both the [SbCl₅]²⁻ units in compound 3 feature smaller structural deformability manifested by slightly changed geometries at excited state. Specifically, negligible structural deformation is found for [Sb(1)Cl₅]²⁻ unit, and Δd and σ² values of [Sb(2)Cl₅]²⁻ unit change from 3.21 × 10⁻³ and 9.025 to 2.10 × 10⁻³ and 40.989 upon the photo-excitation, respectively. Obviously, the structural deformability of [SbCl₅]²⁻ unit in compound 2 is greatly larger than those of compound 3, giving rise to enhanced radiative recombination rate of STEs and increased PLQY from 64.47% to 91.5%. The positive relationship between structural deformability and emission efficiency adequately validates that these broadband light emissions originate from the intrinsic STEs due to strong electron–phonon coupling in deformable crystal lattice, and the detailed PL mechanism is illustrated in the configuration coordination diagram of Fig. 3C and D. Upon the photo-excitation, the electrons are excited from the ground state (¹S₀) to excited state (¹P₁) accompanied by structural deformation, resulting in the formation of STE through intersystem crossing from singlet state (¹P₁) to triplet state (³P_n). Consequently, the radiative recombination of STE gives rise to broadband light emission with large Stokes shifts [48–52].

The closely related crystal structures but distinct PL performance inspire us to explore the feasible phase and PL transformation to realize stimuli-responsive PL switching between these 0D hybrid halides. Significantly, the crystal of nonluminescent compound 1 displays gradually enhanced yellow light emission once exposed to EtOH validated by the evolution of photo-images upon the 365 nm UV light irradiation (Fig. 4A). The detailed in situ PL spectral evolution also gives enhanced broadband yellow light emission centered at 598 nm, which is completely identical to that of compound 2 (Fig. 4B). The final experimental PXRD pattern after immersion in EtOH is consistent with the simulated data of compound 2, indicating the synchronous SC-SC transformation from 1 to 2 (Fig. S23). Moreover, the XRD patterns of compound 1 and 2 after transformation were also fitted with the simulated XRD data of compounds 2 and 3 by Pawley refinement, and the fitting results gave the refinement parameters of R_wp = 7.97% and 5.54%, R_p = 5.19% and 3.58% for compounds 1 and 2 after transformation, respectively, further proving the successful structural transformations of compounds 1 and 2. In addition, we performed X-ray single-crystal diffraction tests for the crystal after transformation to prove its SC-SC transformation. As shown in Fig. S24, the appearance of compounds 1 and 2 after the structural transformation is clearly presented on the single-crystal diffractometer with independent and strong diffraction points. The independent diffraction points, high indexation rate (96.57%), and good refinement results of obtained single-crystal 2 indicate the successful SC-SC transformation from compound 1 to 2 (Fig. S25). As a consequence, controllable turn-on switchable PL vapochromism can be realized through external stimulus-responsive SC-SC transformation, which is rarely found in 0D hybrid metal halides despite frequently reported emitting color change in this field [36,40,67–69]. Comparing with the impressible PL transformation performance of compound 1, compound 2 displays higher structural and luminescent stability toward common organic vapor and humid air validated by the unchanged PXRD patterns and PL emission spectra after long-term exposure in these chemical environments (Figs. S26 and S27). The highly efficient and stable broadband yellow light emission endows compound 2 with potential application in solid-state white light-emitting diode (Fig. S28). However, compound 2 delivers lower thermal stability than 1 based on thermogravimetric analysis (TGA) curves mainly due to the gradual losing of guest EtOH molecule, which corresponds to the mass loss of 4.89% (Fig. S29). Interestingly, the lost of EtOH molecule in the heating process of compound 2 results in synchronous structural reassembly and SC-SC transformation to 3 accompanied by PL transformation from yellow to red color, which can also return to original yellow color after EtOH impregnation (Fig. 4C). The phase and PL transformation between 2 and 3 can be validated by the reversible evolution of PXRD patterns and in situ PL emission spectra (Fig. 4D and Fig. S23B). In addition, this reversible transformation exhibits excellent repeatability with almost the same emission intensity and PXRD patterns after at least 5 consecutive cycles, demonstrating fatigueless reversibility and high stability (Fig. 4E and Fig. S30). Finally, triple-mode controllable PL off–on^I–on^II switching is realized in this family of 0D halide through dual external stimuli-responsive structural reassembly and stepwise SC-SC transformation, which is reported in 0D hybrid halides (Fig. 4F). In order to further explore the formation mechanism of structural transformation, the possible transformation reactions are proposed according to their chemical compositions. We then calculated the Gibbs free energy of compounds 1 to 3 and the small molecules involved to speculate on the reaction process. As shown in Table S12, the possible reaction equations have been given as Eqs. 1 to 3:

Compound 1 can be easily and quickly reconstructed under the stimulation of EtOH to form compound 2 and a free antimony trichloride molecule is released, because the Gibbs free energy change (ΔG) is −4.73 × 10⁻⁴ kcal·mol⁻¹ for this reaction. In contrast, the transition from compound 2 to 3 is relatively more difficult at room temperature, because of the positive ΔG of 2.53 × 10⁻⁵ kcal·mol⁻¹. Therefore, compound 2 can only be converted to compound 3 with the additional promotion of heating treatment, which is consistent with the experimental phenomenon. Conversely, the transition from compound 3 to compound 2 becomes extremely easy with negative ΔG of −2.53 × 10⁻⁵ kcal·mol⁻¹, which can occur only under the action of EtOH at room temperature. The possible reaction mechanism will help us to understand in depth the formation process of structural transformation and provide a guide to rationally design new smart PL switching in 0D hybrid metal halides.

Considering the importance of information security for commercial and military domains, the unique multiple external stimuli-responsive PL switching performance of this 0D hybrid halide family inspires us to explore the advanced applications in anti-counterfeiting and information storage as multistage security technology, after all such a three-fold mode PL switch is very rare and highly difficult to replicate (Table S2). To demonstrate the feasibility of this triplet PL switching as anti-fake technology, a series of luminescence security patterns including designed pictures and Chinese character based on compound 1 were deposited by filling powders into designed patterned grooves (Fig. 5A). All these luminescence security patterns are nearly colorless in daylight and invisible under UV light irradiation, ensuring the complete concealment of actual information. After spraying the EtOH solvent on substrates as stimuli, the invisible patterns gradually appear as yellow color with high resolution under the excitation of 365 nm UV light corresponding to SC-SC transformation from 1 to 2 (Fig. 5A). Subsequently, the yellow-emissive security patterns turn into red color after heating treatment and double anti-counterfeiting technology is successfully achieved with high security level. In addition, these luminescence security patterns also exhibit excellent repeatability and reversibility between yellow and red color with high stability. By virtue of these stimuli-responsive structural and PL transformation, single- and double-mode information encryption–decryption technologies can be achieved based on combined phases of 1+2 and 1+2+3 as luminescence materials, respectively, and the detailed running principles are depicted in Figs. S31 and S32. In order to enhance the decoding complexity of cryptosystem, we further propose a triplet information encryption–decryption strategy, in which only the red-emissive data were designated as correct information and others were blended as interference single. Specifically, we fabricate a combined luminescence security pattern including 3 parts of flower (A), leaf (B), and bud (C) made by 1, 2, and 2@PTFE composite (PTFE = polytetrafluoroethylene), respectively (Fig. 5C). Herein, the 2@PTFE composite is fabricated by using PTFE as protection shell to encapsulate microscale powder of 2 due to the multiple merits of high thermal stability, excellent organic solvent resistance, good transparency, and mechanical property of PTFE. As expected, the 2@PTFE composite displays unchanged yellow light emission upon heating treatment (Fig. S33). Without the knowledge of correct decryption code, the direct excitation of UV light gives the false information revealed by B and C parts. Only under the successive EtOH trigger and heating treatment as combined decoding keys, the true information composed by A and B parts can be correctly recognized under irradiation of 365-nm UV light. These multi-fold information encryption–decryption techniques make this family of 0D halides as an ideal anti-counterfeiting platform for optical information protection and storage applications.

The multiple stimuli-responsive PL switching also makes this halide platform suitable for the construction of molecular-based optical logic gate used in photonic integrated circuit. Herein, the various external stimuli were defined as inputs and responsive PL switching was set as the output based on respective 0D hybrid halides. For example, we set EtOH trigger as input value and emission intensity change (ΔI) of halide 1 as output single (ΔI > 0 as “1” and ΔI ≤ 0 as “0”). According to the binary system, once the EtOH is impregnated as input single, the output is “1” corresponding to EtOH-induced yellow light emission; otherwise, the output is “0” representing the absence of EtOH (Fig. 6A). Similarly, by using temperature or EtOH as individual input signal and change of emission wavelength (Δλ) as output value based on compound 2 or 3, respectively, emission wavelength resolved binary optical logic gate can also be realized by virtue of the reversible PL transformation between yellow-emissive 2 and red-emissive 3 (Fig. 6B and C). In addition, by simultaneously combining the responsive PL switching of distinct luminescent materials, multicoding logic gates could be constructed based on both external stimuli and samples as dual input signals. Specifically, a pair of hybrid halides is defined as input 1 (compound 1 as “1” and compound 3 as “0”) and stimulus of EtOH is assigned as input 2 (presence as “1” and absence as “0”), and change of emission wavelength (Δλ) is set as output single (Δλ > 0 as “1” and Δλ ≤ 0 as “0”) (Fig. 6D). As shown in the truth table of Fig. 6E, only in the case of input 1 = “1” and input 2 = “1”, the output value will be “1” and a 2-input “AND” logic gate is activated, which is related to the activated yellow light emission from compound 1 in the stimulus of EtOH. For other case, the output value will become “0” corresponding to unchanged or decreased PL emission wavelength. In a similar manner, compound 2 and 2@PTFE composite can also be combined as input 3 representing signal of “1” and “0,” respectively, and the external temperature is defined as input 4, in which the temperature threshold is set as 100 °C (above 100 °C as “1” and below 100 °C as “0”). As a result, similar “AND” logic gate also happens for output 1 under the induction of inputs 3 and 4, where the true output of “1” is corresponding to the emission wavelength red-shift (Δλ > 0) of compound 2 under the constant heating treatment (Fig. 6F). These single- and multi-input logic gates provide great potential of smart 0D hybrid halide luminescence materials in molecular recognition, sensing, information storage, keypad lock, memory devices, and so on.

The chemical materials and regents were commercially purchased from Aladdin Chemical Company and directly used in the preparation reaction without any further purification or other physical process. SbCl₃ (99.99%), triphenylethylphosphonium chloride ([Ph₃EtP]Cl, 99.5%), hydrochloric acid (HCl, 37%), ethyl ether (Et₂O), and EtOH were used.

The mixture of SbCl₃ (1 g, 4.38 mmol) and [Ph₃EtP]Cl (1.5 g, 4.59 mmol) was dissolved in a mixed solution of EtOH (2.5 ml) and hydrochloric acid (7.5 ml). The suspension was heated to completely dissolve and then allowed to stand at room temperature. After 1 day, lots of colorless crystals were filtrated and washed with Et₂O three times (yield: 86% based on Sb). The structure was subsequently determined to be C₄₀H₄₀P₂Sb₂Cl₈ by using the single-crystal X-ray diffraction. Elemental analysis (EA) calculated for C₄₀H₄₀P₂Sb₂Cl₈: C, 43.29%; H, 3.63%; found: C, 43.32%; H, 3.68%.

[Ph₃EtP]Cl (0.287 g, 0.88 mmol) and SbCl₃ (0.1 g, 0.44 mmol) were dissolved in a mixed solution of EtOH (4 ml) and hydrochloric acid (0.5 ml). The suspension was constantly transferred into a 20 ml glass vial, which was then sealed and heated to a constant temperature of 80 °C for 5 days. After the reaction, lots of colorless block crystals were filtrated from the vial and washed with EtOH three times (yield: 78% based on Sb). The structure was subsequently determined to be C₄₂H₄₅OP₂SbCl₅ by using the single-crystal X-ray diffraction. EA calculated for C₄₂H₄₅OP₂SbCl₅: C, 54.43%; H, 4.89%; found: C, 54.42%; H, 4.87%.

Compound 2 powder samples (0.2 g) were placed in a glass bottle and then heated to 140 °C under vacuum conditions for 2 h. After slow cooling to room temperature, lots of colorless block crystals were obtained (yield: 92% base on compound 2). The structure was subsequently determined to be C₄₀H₄₀P₂SbCl₅ by using the single-crystal X-ray diffraction. EA is calculated for C₄₀H₄₀P₂SbCl₅: C, 54.49%; H, 4.57%; found: C, 54.50%; H, 4.54%.

Compound 2 powders (0.5 g) were dispersed into 3 ml of EtOH to obtain a homogeneous suspension. Then, 0.1 ml of PTFE solution was placed in the solution and stirred at room temperature for 8 h. White powder of 2@PTFE was obtained by evaporation solvent.

The PXRD of compounds 1 to 3 were performed on a Rigaku Dmax 2500 and Rigaku SmartLab (Dtex250, Japan) X-ray powder diffraction meter equipped with Cu-Kα radiation (λ = 1.5418 Å). EA of C, H, and N was carried out on elemental analyzer of Vario EL-Cube. The solid-state UV–Vis optical absorption spectrum were performed on a PE Lambda 900 UV/Vis spectrophotometer. The thermal stabilities were tested by using the TGA, which were carried out on a Netzsch STA449C thermal analyzer in the temperature range of 25 to 800 °C under the constant protection of N₂ atmosphere flow with a heating rate of 10 °C·min⁻¹. The Raman measurements were performed by using Horiba Scientific LabRam HR Evolution under 532-nm excitation wavelength. SEM images were obtained by Zeiss Sigma 500. The PL spectra, time-resolved decay data, and PLQYs were performed on an Edinbergh FLS1000 fluorescence spectrometer.