English

• Hybrid lead halide perovskites APbX₃ (A = CH₃NH₃⁺, HC(NH₂)₂⁺; X = Cl⁻, Br⁻, I⁻, Fig. S1a in Supporting information) had been the keystones of the next-generation solar cells as light absorbers and attracted worldwide attention in last decades [1–3]. However, concerns about the toxicity and stability of the lead-containing materials have stimulated a great deal of interest in the discovery of nontoxic and stable perovskites that might be equally effective as the parent lead halides [4–7]. One of the strategies for achieving this has been to synthesize cesium-based double perovskite (DP), of general formula Cs₂M^ⅠM^ⅢX₆ (where M^Ⅰ and M^Ⅲ are univalent and trivalent metals at the Pb site, and X is a halide anion, Fig. S1b in Supporting information), including examples such as Cs₂AgBiCl₆ [4,6,8,9], Cs₂AgSbBr₆ [10] and Cs₂NaBiCl₆ [11–13]. Among these Cs-based DPs, the Bi-containing materials display band structures broadly similar to those of the lead halide perovskites, since the substituted Bi(Ⅲ) has 6s² electronic configuration similar to Pb(Ⅱ). Therefore, Bi-based DPs are seen as promising alternatives for lead halide perovskite.

  Apart from solar absorbers, these halide DPs have favorable optoelectronic performances for applications such as photodetectors [8,14], scintillators [15,16], and phosphors [17–21]. Thus, the interest in the halide DPs, especially the Bi−based materials, has been continuously growing [4,6,8,12,22–25]. Although far more Bi-based halide DPs were predicted by theoretical studies, only three chloride Cs₂M^ⅠBiCl₆ (M^Ⅰ = Li, Na, and Ag) and one bromide Cs₂AgBiBr₆ have been isolated experimentally as stable crystalline phases so far. The Bi-based iodide DPs are unstable under normal conditions [26]. For instance, the Cs₂AgBiI₆ DP can only be prepared as kinetically stabilized nanocrystals by post-synthetic modification of the corresponding chloride and bromide [27]. The reports of Cs₂NaBiI₆ DP have not been substantiated by definitive structural characterization [28].

  In addition, the Bi-based chloride DPs suffer from instability issues. For example, the Cs₂CuBiCl₆ compositions cannot form perovskite structures in contrast with its Ag analogs. Such difference in stability comes from the fourfold coordination preference of Cu⁺ ion with halide ions, dissatisfying the sixfold coordination requirement in the DPs [29]. Another example is extremely moisture-sensitive Cs₂LiBiCl₆ DP. This instability is related to small-sized Li⁺ cation, which degrades from sixfold to lower coordination irreversibly upon the moisture attack [25].

  Although the rare-earth-based DPs Cs₂KM^ⅢCl₆ (M^Ⅲ = Sc, Eu, Lu) were first reported in the 1990s [30,31], little attention was paid to the Bi-based halide DPs containing a large alkali metal cation K⁺ (Cs₂KBiCl₆) except a pioneer experimental study on the nanocrystals of Cs₂KBiCl₆ DP showing anisotropic polarization luminescence phenomenon [32]. However, the crystal structure of bulk Cs₂KBiCl₆ is still an open question, as nanoscale materials often consist of surface defects in large amounts and take metastable structures differently from the bulk. This raises concerns about the structure and stability of Cs₂KBiCl₆. One might wonder whether Cs₂KBiX₆ DP is unstable like Cs₂AuBiX₆ given the similar ionic size and identical charge of K⁺ and Au⁺ (d¹⁰) ions regardless of their different electronic configurations [33]. If Cs₂KBiX₆ DP is stabilized, it would be a good host material for Mn²⁺-activated luminescence similar to Cs₂NaBiCl₆ DP [12]. Motivated by such an idea, we started an investigation on the Cs₂KBiCl₆ DPs. Here we report the isolation, crystal structure, and Mn²⁺-activated luminescent of a new halide DP Cs₂KBiCl₆.

  The initial synthesis attempt performed at ~503 K using the nominal composition of Cs₂KBiCl₆, led to a poor-crystalline DP with broad reflections in addition to secondary phases Cs₃BiCl₆ and KCl (Fig. S2 in Supporting information). Increasing the temperature above 573 K significantly improved the crystallinity of DP phase and reduced the secondary phases. Although the Cs₃BiCl₆ phase completely disappear at 853 K, a minor (~3%) KCl phase was still observable (Fig. S3 in Supporting information) and further confirmed through the segregation of potassium on the scanning electron microscope (SEM) image (Fig. S4 in Supporting information). By reducing 10% KCl in raw materials (Fig. S5 in Supporting information), the pure-phase DP sample was finally isolated at the nominal composition of Cs₂K_0.9BiCl_5.9 (or Cs_2.03K_0.92Bi_1.02Cl₆ equally).

  SEM images and elementary analysis were recorded on the pellet Cs₂K_0.9BiCl_5.9 samples (Fig. S6 in Supporting information), showing homogeneous elements distributions. The Cs: K: Bi: Cl ratio of 2.37(7): 0.94(6): 1.05(6): 5.64(1) in the sample essentially matches with the value in the stoichiometric DP phase Cs₂KBiCl₆. The excessive CsCl and BiCl₃ in the initial composition were either lost or formed amorphous during the reaction.

  The XRD data of Cs₂KBiCl₆ sample can be well fitted with the Cs₂NaBiCl₆ structure model in Pawley mode (R_wp ~ 4.48%, Fig. S7a in Supporting information), indicating that Cs₂KBiCl₆ likely adopts an ideal DP structure as its Na analogue. However, the Rietveld refinement using the ideal DP structure model led to either extremely large B_eq (12.0–19.0 Ų) for all atoms or large residue intensities (R_wp = 15.23%, Fig. S7b in Supporting information) with the constrained atomic B_eq ≤ 3.0 Ų. This indicates that these atoms deviate from the high-symmetric positions of ideal DP. Therefore, we amended the DP structural model slightly and allowed atomic shifts from ideal high-symmetric sites as well as fractional occupancy in the Rietveld refinement. This led to satisfactory fitting results with R_wp = 6.83% (Fig. 1a).

  Figure 1

  

  Figure 1.  Rietveld plots of (a) XRD data and (b) neutron data for Cs₂KBiCl₆. The inset to (a) shows a disordered DP structural model with the color code: blue sphere (Cs), red sphere (Cl), yellow sphere (K) and purple sphere (Bi).

  DownLoad: Full-Size Img PowerPoint

  Due to their weak X-ray scattering ability, the light K⁺ and Cl⁻ ions are not well located with X-ray diffractions in the presence of heavy elements such as bismuth and cesium. Neutron diffraction (ND) is more sensitive to identifying K⁺ and Cl⁻ ions because the neutron scattering lengths of potassium (3.67 fm) and chloride (9.58 fm) are distinct from those of Cs (5.42 fm) and Bi (8.53 fm) [34]. Therefore, to find the accurate position of K⁺ and Cl⁻, neutron diffraction data were collected on the Cs₂KBiCl₆ sample. The weak diffused peaks on the ND data (Fig. 1b), i.e., 3.0, 1.5 and 0.8 Å, indicate the likely presence of local ordering or amorphous phase in the Cs₂KBiCl₆ sample.

  The Rietveld refinements were carried out against the XRD and neutron data simultaneously, using the disordered DP structural model. During the refinements, the preferred orientation of the sample was described with the spherical harmonic functions. The lattice parameters, atomic coordinates and B_eq parameters were refined. The final refinement converged to R_wp = ~6.83% and ~5.19% for XRD and ND data, as shown in Figs. 1a and b, respectively, and the refined structural parameters are listed in Table 1.

  Table 1

  

  Table 1.  Final refined structural parameters for Cs₂KBiCl₆.

  DownLoad: CSV

  In the crystal structure of Cs₂KBiCl₆, the K, Bi, Cs, and Cl elements are shifted from the ideal DP sites with fractional occupancies in compensations (Fig. 1a inset). Among all the ions, the K⁺ cation has the largest relative shift of ~7%, while Bi³⁺ has the smallest shift of ~3%. This indicates K⁺ is an important factor for structural disorder as discussed in the following sections. The disordering of all ions in Cs₂KBiCl₆ prevents further meaningful structural analysis.

  Quadrupolar ¹³³Cs (nuclear spin, I = 7/2, Qm = −0.34 fm²) is a highly sensitive nuclear magnetic resonance (NMR) nucleus with excellent resolution, as it behaves as a pseudo spin-½ nucleus, rendering it ideal to investigate the local coordination environment. The ¹³³Cs magic angle spinning (MAS) NMR spectra for the Cs₂KBiCl₆ sample (Fig. 2a) consist of four resolved NMR resonances centered at 56(2), 121(2), 180(4), and 253(4) ppm with a ratio of 10:49:35:6 (fitted peak area). This indicates four distinct local coordination environments of Cs in the Cs₂KBiCl₆ material. Although only one 12-coordinated Cs site is present in the average structure, the fractional occupancy of Cs, Bi, K and Cl sites in Cs₂KBiCl₆ are compatible with different local environments of Cs. Due to their similar ionic size, the substitution of K⁺ and Cs⁺ is evitable in Cs₂KBiCl₆ DP and could lead to local connections and coordination numbers (CN) such as 7, 8 and 9 for Cs⁺ ions similar to that in CsK₂BiCl₆ and Cs₃BiCl₆ materials (Fig. S8 in Supporting information). Furthermore, the ¹³³Cs peaks of Cs₂KBiCl₆ DP have full width at half maximum (FWHM) of 25.7–48.2 ppm and are much broader than those (FWHM = 1 ppm) of well-ordered DP Cs₂AgInCl₆ and Cs₂AgBiCl₆ [35]. This indicates the disorders of Cs ions in Cs₂KBiCl₆ are close to that in amorphous materials to some extent.

  Figure 2

  

  Figure 2.  (a) ¹³³Cs MAS NMR spectra (solid lines) and its overall best Gaussian fits (dotted lines) for Cs₂KBiCl₆. Spectra were acquired at a magnetic field strength of 9.4 T with a spinning frequency of 12 kHz. (b) TGA (black) and DTA (red) data for the Cs₂KBiCl₆ sample from room temperature (RT) to 973 K. The dashed baselines are shown in black and blue to guide the eyes for TGA and DTA, respectively.

  DownLoad: Full-Size Img PowerPoint

  Thermal analysis and variable temperature X-ray diffraction (VT-XRD) experiments were carried out to examine the thermal stability of Cs₂KBiCl₆ in the air. During the thermal analysis, the Cs₂KBiCl₆ sample kept a constant weight below 573 K, as shown in the thermo gravimetric analysis (TGA) curve (Fig. 2b), which is consistent with VT-XRD data (Fig. S9 in Supporting information) showing that Cs₂KBiCl₆ was stable up to 523 K. In the TGA data, the initial weight loss starts at ~573 K and turns abrupt at ~800 K. These two temperatures are close to the melting point 506 K and boiling point 723 K of BiCl₃ [36], suggesting that the weight change was due to the loss of BiCl₃. The endothermic peak at 673–897 K is probably related to the decomposition, consistent with the arising of KCl reflections at 523–773 K in VT-XRD data. The decomposition was completed at 773 K, resulting in KCl and amorphous phases, as evidenced by the broad peaks in the 773 K XRD data. Upon further temperature increase from 773 K, this mixture transformed into other intermediate phases. This could explain the exothermic peak above 897 K, which was also associated with the significant loss of BiCl₃.

  Low-temperature XRD patterns were collected to explore any possible phase change of Cs₂KBiCl₆ (Fig. S10a in Supporting information). No phase transition exists in the Cs₂KBiCl₆ material down to 5 K. Cs₂KBiCl₆ has a constant 3% volume decrease upon cooling from RT to 5 K. A smaller volume change (~2.5%) occurred when the sample was warmed from 5 K to RT at 1 K/min, although reversible volume increase was obtained at a warming rate of ~0.2 K/min (Fig. S10b in Supporting information). This history-dependent volume change indicates the slow relaxations in the Cs₂KBiCl₆ sample, also keeping with its disordered nature.

  UV–vis diffuse reflectivity data (Fig. 3a) were collected and transformed into an optical absorbance coefficient F(R) using the Kubelka-Munk equation; F(R) = (1 - R)²/2R, where R is the reflectance [37]. Similar to that of Cs₂NaBiCl₆, the absorption curve of Cs₂KBiCl₆ consists of three distinct peaks at 270, 321 and 349 nm, all associated with a 6s²–6s¹p¹ transition of the localized [BiCl₆]³⁻ octahedron. Precisely, the peak at a shorter wavelength ~270 nm (band 1) is assigned to the ¹S₀ → ³P₂ partially allowed transition; The double peaks centered at 321 and 349 nm (bands 2 and 3) are related to the ¹S₀ → ³P₁ spin-forbidden transition.

  Figure 3

  

  Figure 3.  (a) UV–vis absorption spectra and (b) the Tauc plot of Cs₂KBiCl₆ and Cs₂NaBiCl₆.

  DownLoad: Full-Size Img PowerPoint

  The absorption edge of Cs₂KBiCl₆ DP around 349 nm are sharp and intense, indicating a probable direct band-gap, which can be obtained on the Tauc plot as the intercept from the linear fits to the [F(R)E]². From the corresponding Tauc plot (Fig. 3b), the direct gaps of Cs₂KBiCl₆ can be estimated to be ~3.35 eV, agreeing with theoretical value (~3.18 eV) of Cs₂KBiCl₆ (Fig. S11 in Supporting information). While the fitting to [F(R)E]^1/2 leads to an indirect gap of 3.1 eV for Cs₂NaBiCl₆, matching well with the reported value (3.41 and 3.19 eV) [38–40]. An extrinsic defect-related subband gap is estimated to be ~3.02 eV for Cs₂NaBiCl₆ [25,38].

  The conduction band minimum (CBM) is split off from other conduction bands due to the strong SOC interactions of the Bi 6p states. While the valence band maximum (VBM) is originated from the hybridization of the Bi 6s and Cl 3p states, as seen from the corresponding charge densities shown in Fig. S12 (Supporting information).

  Cs₂KBiCl₆ samples were doped by Mn²⁺ according to two Mn replacing K and Bi (i.e., Cs₂K_1-xBi_1-xMn_2xCl₆). All the Mn-doped samples Cs₂K_1-xBi_1-xMn_2xCl₆ (x = 0, 0.004, 0.007, 0.01, 0.02, 0.025, 0.05) exhibit similar XRD patterns (Fig. S13 in Supporting information). The Rietveld refinement against the XRD data of the highly Mn-doped composition x = 0.025 was performed with different structural models. The best fit was obtained with the model in which Mn equally substituted on both the K and Bi sites (Fig. S14 in Supporting information, R_wp = 6.50%), compared with those where Mn substituted on the Bi or K site only (R_wp = 6.64% and 6.80% respectively). This indicates that Mn(Ⅱ) ions prefer entering both K and Bi sites. Magnetization measurements were performed on the x = 0.025 sample, which was paramagnetic and consistent with the diluted manganese ions in the host lattice. The Curie-Weiss fitting was performed on the 2–12 K data with the most pronounced paramagnetic signal (Fig. S15 in Supporting information). The fits resulted in an effective moment of 5.7(1) μ_B per Mn, essentially identical to that of 5.9 μ_B for a high spin (HS) Mn²⁺ ion, thus confirming that the incorporated Mn is Mn²⁺ in HS state.

  The Mn²⁺-doped Cs₂KBiCl₆ sample shows orange-red photoluminescence with a medium-wide emission band centered at 600 nm (Fig. 4a). The emission originates from a ⁴T₁→⁶A₁ transition of octahedrally coordinated Mn²⁺ centers and shows CIE coordinates of x = 0.53 and y = 0.42, which are slightly red-shifted from those of the Na counterpart (x = 0.51 and y = 0.44). The change in chromaticity can be observed in the CIE diagram plotted in Fig. S16 (Supporting information). When no dopant is added, the Cs₂KBiCl₆ material exhibits very weak luminescence centered at ~615 nm from the Bi³⁺ ions located near the defect site. The emission intensity of Cs₂K_1-xBi_1-xCl₆ increases with the increasing Mn content within the range of x = 0–0.025 (Fig. 4b).

  Figure 4

  

  Figure 4.  (a) Excitation (blue) and emission (red) spectra for the x = 2.5% Mn-doped Cs₂KBiCl₆ sample, with its photograph taken under 365 nm illumination shown in the inset. (b) Emission spectra of Cs₂K_1-xBi_1-xMn_2xCl₆ (x = 0–0.05) samples under an excitation wavelength of 363 nm. (c) Luminescent decay curves of the ⁴T₁ → ⁶A₁ transition of Mn²⁺ in the Cs₂K_1-xBi_1-xMn_2xCl₆ samples. The symbols correspond to the experimental data, while the lines show the fitted double-exponential dependence. The inset shows the average lifetime for all the Mn-doped Cs₂KBiCl₆ samples.

  DownLoad: Full-Size Img PowerPoint

  The excitation spectrum for the x = 0.025 member (Fig. 4a) shows peaks centered at 297, 363, and 425 nm, slightly red-shifted compared with its Na analogue. These bands do not match the absorption bands observed in the UV–vis spectra of the parent material (Fig. S17 in Supporting information), so we assign them to the direct excitation of the electrons from the Mn²⁺ ground state ⁶A₁(6S) to higher levels of the 3d manifold, i.e., ⁴T₁(4P), ⁴T₂(4D), and ⁴T₂(4 G). Thus, the absence of absorption bands due to the spin-forbidden ¹S₀→³P_{1, 2} transitions of Bi³⁺ in the excitation spectrum, where the emission of Mn²⁺ is being monitored, proves that the resulting emission is not sensitized by the host.

  To find out the optimal dopant concentration, the decay of the emission originated from the ⁴T₁→⁶A₁ transition (at ~600 nm) was recorded for all the Mn²⁺−doped samples (Fig. 4c). The PL decay curves have been described by the following biexponential function:

  (1)

  where I(t) is the luminescence intensity, t is the time after excitation, and τ (i = 1, 2) is the decay time of the i^th component, with intensity I. This is reasonable since Mn simultaneously replaces both K and Bi sites, which are crystallographically inequivalent and may lead to two different luminescence relaxations, similar to that in the Cs₂LiBiCl₆: Mn materials. The average decay times 〈τ〉, calculated according to the following equation (Eq. 2) have been plotted in Fig. 4c inset.

  (2)

  As observed, 〈τ〉 slightly decreases monotonously from 339 µs to 290 µs when the Mn²⁺ content increases from 0.4% to 2.5%, although a clear quenching effect cannot be observed within the compositional range analyzed. Similar magnitude of lifetime has been reported in halide perovskite [41]. Photoluminescence quantum yield (PLQY) measurements were then performed at an excitation wavelength of 363 nm and indicated a maximum of ~15.8% in the x = 0.02 sample (Table S1 in Supporting information), similar to that in Cs₂NaBiCl₆: Mn materials. The x = 0.004 sample has the minimum PLQY of ~4.7%, close to that of Cs₂AgInCl₆: Mn materials. Although the more efficient phosphor has been found for the x = 0.02 composition, from a practical view, the most interesting composition is the one doped with 2.5 mol% Mn²⁺ as it exhibits the highest emission intensity (Fig. 4b) owing to the larger number of emission centers compensate for the concentration quenching effect.

  Temperature-dependent emission spectra (Fig. S18a in Supporting information) were collected for the x = 0.025 sample upon 363 nm excitation to evaluate the luminescence thermal quenching behavior of the material. The decrease in the emission intensity as well as a blue shift of the emission is observed with increasing temperature, which is normally ascribed to the enhanced nonradiative transition probability and crystal lattice strain. When the temperature rose to 200 and 300 K, the emission intensity of the orange emission can maintain ~65% and ~9% of its initial value (at 95 K), showing obvious thermal quenching behavior (Fig. S18b in Supporting information). The emission almost disappears at 360 K, indicating that the Cs₂KBiCl₆: Mn materials may not be suitable for application in pc-LED which typically achieve 400–450 K near the LED chip [42].

  Two empirical criteria, the Goldschmidt tolerance factor (TF) [43] and the octahedral factor (µ) are popular to assess the stability of the perovskite halides. They are defined as TF = (R_A + R_X)/ (R_M + R_X) and µ = R_M/R_X, where R_A, R_M, and R_X represent the ionic radii of A, M and X ions. Statistical analysis on simple perovskites indicates that stable halide perovskites have TF and µ in the range of 0.87–1.1 and 0.414–0.732 respectively [44]. This empirical relationship can be extended to the Cs₂M^ⅠM^ⅢCl₆ DPs when the average ionic radii of M^Ⅰ and M^Ⅲ are adopted as R_M. Employing the Shannon ionic radii of 12-coordinated Cs⁺ (1.88 Å) and 6-coordinated M^Ⅰ, M^Ⅲ, and Cl⁻ (1.81 Å) ions [45], the calculated TF and µ values of all the experimentally stabilized DPs Cs₂M^ⅠM^ⅢCl₆ (M^Ⅰ = Li⁺, [25,30,46,47] Na⁺, [11,47–51] Ag⁺ [4,23,52–54], Au⁺ [54,55], Tl⁺ [56] and K⁺ [30,31], M^Ⅲ = In³⁺, Tl³⁺, Sb³⁺, Au³⁺, Bi³⁺ and RE) falls in the empirical stable area, as listed in Table 2. However, for larger M^Ⅰ cations such as K⁺, Tl⁺ and Au⁺, which prefer to take more than six Cl⁻ ions, the octahedral factor of the M^ⅠCl₆ octahedron (0.76–0.83) is out of the empirical criterion, and the TF of Cs₂M^ⅠM^ⅢCl₆ (0.87–0.90) is on the lower limit of the empirical criterion. Therefore, Cs₂M^ⅠM^ⅢCl₆ (M^Ⅰ = K⁺, Tl⁺ and Au⁺) often adopt distorted DP structures in low symmetry rather than ideal DP structures as other Cs₂M^ⅠM^ⅢCl₆ (M^Ⅰ = Li⁺, Na⁺ and Ag⁺) DPs. Indeed, the low-symmetric structural distortion helps to meet the higher-coordination requirement of large M^Ⅰ cations. Meanwhile, due to the similar radius, Cs⁺ and K⁺ (1.64 Å, CN = 12) could substitute each other inevitably in the Cs₂KBiCl₆ DP. However, the substitution could destabilize the double perovskite phase and form a local non-perovskite connection, as indicated by the fact that 50% of Cs replaced by K leads to cryolite CsK₂BiCl₆, while the replacement of K by Cs results in cryolite Cs₃BiCl₆. Furthermore, due to the 6s² lone pair, the Bi³⁺ ion prefers local asymmetric coordination rather than octahedron in DP halides, which further enhances the structural distortion. This often leads to large atomic thermal displacements (B_eq = 0.9–3.1 Ų) in the cubic structure of Bi-based halides Cs₂M^ⅠBiCl₆ (M^Ⅰ = Li⁺, Na⁺ and Ag⁺) [4,11,25]. Overall, both factors large-sized K⁺ and Bi³⁺ with 6s² lone pair contribute to the enhanced structural distortion of Cs₂KBiCl₆ DP, forming a disordered cubic crystal structure with complex local coordination environments.

  Table 2

  

  Table 2.  Experimentally stabilized Cs₂M^ⅠM^ⅢCl₆ DPs showing tolerance factor TF and octahedral factor μ.

  DownLoad: CSV

  In summary, a new bismuth-based DP halide Cs₂KBiCl₆ was synthesized successfully. From combined Rietveld refinement against the XRD and NPD data, its crystal structure consists of shifted Cs, K, Bi, and Cl sites from the ideal positions with fractional occupancy in compensation. Four distinct local coordination of Cs⁺ ions are revealed in Cs₂KBiCl₆ by ¹³³Cs solid-state nuclear magnetic resonance spectroscopy. The Cs₂KBiCl₆ DP displays thermal-history-dependent volume change at the 5–298 K range. The Cs₂KBiCl₆ has a direct band gap of 3.35(2) eV and red-shift luminescence around 600 nm upon Mn doping compared with the Na analogue. The disordered structure of Cs₂KBiCl₆ is stabilized via two factors, large K⁺ cations which prefer to bond with more than six Cl⁻, and the Bi³⁺ with 6s² lone pair which has a preference for a local asymmetric environment.

  Declaration of competing interest

  The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

  Acknowledgments

  The authors thank the National Science Foundation of China (Nos. 22090043 and 22161014), Guangxi Natural Science Foundation (Nos. 2019GXNSFGA245006 and 2020GXNSFAA297220), and the Foundation of Guilin University of Technology (No. GUTQDJJ2018115) for the financial support.

  Supplementary materials

  CSD 2,160,488 contains the supplementary crystallographic data for this paper. The data can be obtained free of charge via www.ccdc.cam.ac.uk/data_request/cif, or by emailing datarequest@ccdc.cam.ac.uk, or by contacting The Cambridge Crystallographic Data centre, 12 Union Road, Cambridge CB2 1EZ, U.K.; fax: + 44 1223 336,033.

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2023.108641.

  
  
• 1. [1]

     A. Kojima, K. Teshima, Y. Shirai, et al., J. Am. Chem. Soc. 131 (2009) 6050–6051. doi: 10.1021/ja809598r

  2. [2]

     X. Sun, D. Zhao, Z. a. Li, Chin. Chem. Lett. 29 (2018) 219–231. doi: 10.1016/j.cclet.2017.09.038

  3. [3]

     A.K. Jena, A. Kulkarni, T. Chem. Rev. 119 (2019) 3036–3103. doi: 10.1021/acs.chemrev.8b00539

  4. [4]

     E.T. McClure, M.R. Ball, et al., Chem. Mater. 28 (2016) 1348–1354. doi: 10.1021/acs.chemmater.5b04231

  5. [5]

     C.N. Savory, A. Walsh, D.O. Scanlon, ACS Energy Lett. 1 (2016) 949–955. doi: 10.1021/acsenergylett.6b00471

  6. [6]

     A.H. Slavney, T. Hu, A.M. Lindenberg, et al., J. Am. Chem. Soc. 138 (2016) 2138–2141. doi: 10.1021/jacs.5b13294

  7. [7]

     R.L. Hoye, L. Eyre, F. Wei, et al., Adv. Mater. Interfaces 5 (2018) 1800464. doi: 10.1002/admi.201800464

  8. [8]

     W. Pan, H. Wu, J. Luo, et al., Nat. Photonics 11 (2017) 726–732. doi: 10.1038/s41566-017-0012-4

  9. [9]

     S. Wang, Y. Xie, W. Jiang, et al., Chin. Chem. Lett. 34 (2023) 108521.

  10. [10]

      F. Wei, Z. Deng, S. Sun, et al., Chem. Commun. 55 (2019) 3721–3724. doi: 10.1039/c9cc01134j

  11. [11]

      L.R. Morss, W.R. Robinson, Acta Cryst. B 28 (1972) 653–654.

  12. [12]

      J.D. Majher, M.B. Gray, T.A. Strom, et al., Chem. Mater. 31 (2019) 1738–1744. doi: 10.1021/acs.chemmater.8b05280

  13. [13]

      H. Yang, Y. Guo, G. Liu, et al., Chin. Chem. Lett. 33 (2022) 537–540. doi: 10.1016/j.cclet.2021.05.071

  14. [14]

      J. Luo, S. Li, H. Wu, et al., ACS Photonics 5 (2018) 398–405. doi: 10.1021/acsphotonics.7b00837

  15. [15]

      J. Glodo, R. Hawrami, K.S. Shah, J. Cryst. Growth 379 (2013) 73–78. doi: 10.1016/j.jcrysgro.2013.03.023

  16. [16]

      H. Shi, M.H. Du, Phys. Rev. Appl. 3 (2015) 054005. doi: 10.1103/PhysRevApplied.3.054005

  17. [17]

      J. Luo, X. Wang, S. Li, et al., Nature 563 (2018) 541–545. doi: 10.1038/s41586-018-0691-0

  18. [18]

      A. Zhang, Y. Liu, G. Liu, et al., Chem. Mater. 34 (2022) 3006–3012. doi: 10.1021/acs.chemmater.1c03878

  19. [19]

      Y. Liu, X. Rong, M. Li, et al., Angew. Chem. Int. Ed. 59 (2020) 11634–11640. doi: 10.1002/anie.202004562

  20. [20]

      Y. Liu, M.S. Molokeev, Z. Xia, Energy Mater. Adv. 2021 (2021) 2585274.

  21. [21]

      Q. Jia, T. Shao, L. Tong, et al., Chin. Chem. Lett. 34 (2023) 107539. doi: 10.1016/j.cclet.2022.05.053

  22. [22]

      A.H. Slavney, L. Leppert, D. Bartesaghi, et al., J. Am. Chem. Soc. 139 (2017) 5015–5018. doi: 10.1021/jacs.7b01629

  23. [23]

      T.T. Tran, J.R. Panella, J.R. Chamorro, et al., Mater. Horiz. 4 (2017) 688–693. doi: 10.1039/C7MH00239D

  24. [24]

      K.P. Lindquist, S.A. Mack, A.H. Slavney, et al., Chem. Sci. 10 (2019) 10620–10628. doi: 10.1039/c9sc02581b

  25. [25]

      Y. Sun, A.J. Fernández-Carrión, Y. Liu, et al., Chem. Mater. 33 (2021) 5905–5916. doi: 10.1021/acs.chemmater.1c00854

  26. [26]

      P. Vishnoi, R. Seshadri, A.K. Cheetham, J. Phys. Chem. C. 125 (2021) 11756–11764. doi: 10.1021/acs.jpcc.1c02870

  27. [27]

      S.E. Creutz, E.N. Crites, M.C. De Siena, Nano Lett. 18 (2018) 1118–1123. doi: 10.1021/acs.nanolett.7b04659

  28. [28]

      C. Zhang, L.G. Gao, S. Teo, et al., Sustain. Energy Fuels 2 (2018) 2419–2428. doi: 10.1039/c8se00154e

  29. [29]

      Z.W. Xiao, K.Z. Du, W.W. Meng, et al., Angew. Chem. Int. Ed. 56 (2017) 12107–12111. doi: 10.1002/anie.201705113

  30. [30]

      G. Meyer, S.J. Hwu, J.D. Corbett, Z. Anorg. Allg. Chem. 535 (1986) 208–212. doi: 10.1002/zaac.19865350423

  31. [31]

      M. Villafuerte-Castrejón, M. Estrada, J. Gomez-Lara, et al., J. Solid State Chem. 132 (1997) 1–5. doi: 10.1006/jssc.1997.7382

  32. [32]

      C. Wang, M. Sun, H. Wang, et al., J. Phys. Chem. Lett. 14 (2023) 164–169. doi: 10.1021/acs.jpclett.2c03482

  33. [33]

      C.J. Bartel, J.M. Clary, C. Sutton, et al., J. Am. Chem. Soc. 142 (2020) 5135–5145. doi: 10.1021/jacs.9b12440

  34. [34]

      V.F. Sears, Neutron News 3 (1992) 26–37. doi: 10.1080/10448639208218770

  35. [35]

      A. Karmakar, G.M. Bernard, A. Meldrum, et al., J. Am. Chem. Soc. 142 (2020) 10780–10793. doi: 10.1021/jacs.0c02198

  36. [36]

      P. Patnaik, Handbook of Inorganic Chemicals, McGraw-Hill Companies, Inc., New York, 2002, pp. 1–1125.

  37. [37]

      D.B. Judd, J. Res. Nat. Bur. Standards 13 (1934) 281–291. doi: 10.6028/jres.013.021

  38. [38]

      S. Wu, W. Li, J. Hu, et al., J. Mater. Chem. C. 8 (2020) 13603–13611. doi: 10.1039/d0tc03003a

  39. [39]

      J. Zhou, X.M. Rong, P. Zhang, et al., Adv. Opt. Mater. 7 (2019) 1801435. doi: 10.1002/adom.201801435

  40. [40]

      X.G. Zhao, J.H. Yang, Y.H. Fu, et al., J. Am. Chem. Soc. 139 (2017) 2630–2638. doi: 10.1021/jacs.6b09645

  41. [41]

      C. Wang, L. Ma, S. Wang, et al., J. Phys. Chem. Lett. 12 (2021) 12129–12134. doi: 10.1021/acs.jpclett.1c03583

  42. [42]

      P.F. Smet, A.B. Parmentier, D. Poelman, J. Electrochem. Soc. 158 (2011) R37–R54. doi: 10.1149/1.3568524

  43. [43]

      V.M. Goldschmidt, Naturwissenschaften 14 (1926) 477–485. doi: 10.1007/BF01507527

  44. [44]

      C. Li, X. Lu, W. Ding, et al., Acta Cryst. B 64 (2008) 702–707. doi: 10.1107/S0108768108032734

  45. [45]

      R.D. Shannon, Acta Cryst. A 32 (1976) 751–767. doi: 10.1107/S0567739476001551

  46. [46]

      C. Combes, P. Dorenbos, C. Van Eijk, et al., J. Lumin. 82 (1999) 299–305. doi: 10.1016/S0022-2313(99)00047-2

  47. [47]

      C. Reber, H.U. Guedel, G. Meyer, et al., Inorg. Chem. 28 (1989) 3249–3258. doi: 10.1021/ic00315a034

  48. [48]

      P. Han, X. Mao, S. Yang, et al., Angew. Chem. Int. Ed. 58 (2019) 17231–17235. doi: 10.1002/anie.201909525

  49. [49]

      L.R. Morss, M. Siegal, L. Stenger, et al., Inorg. Chem. 9 (1970) 1771–1775. doi: 10.1021/ic50089a034

  50. [50]

      M. Spirlet, J. Rebizant, J. Fuger, Acta Cryst. C 44 (1988) 1300–1301. doi: 10.1107/S0108270188002288

  51. [51]

      L.R. Morss, J. Fuger, Inorg. Chem. 8 (1969) 1433–1439. doi: 10.1021/ic50077a013

  52. [52]

      G. Volonakis, A.A. Haghighirad, R.L. Milot, et al., J. Phys. Chem. Lett. 8 (2017) 772–778. doi: 10.1021/acs.jpclett.6b02682

  53. [53]

      A.H. Slavney, L. Leppert, A.S. Valdes, et al., Angew. Chem. Int. Ed. 57 (2018) 12765–12770. doi: 10.1002/anie.201807421

  54. [54]

      N. Elliott, L. Pauling, J. Am. Chem. Soc. 60 (1938) 1846–1851. doi: 10.1021/ja01275a037

  55. [55]

      X.J. Liu, K. Matsuda, Y. Moritomo, et al., Phys. Rev. B 59 (1999) 7925–7930. doi: 10.1103/PhysRevB.59.7925

  56. [56]

      M. Retuerto, T. Emge, J. Hadermann, et al., Chem. Mater. 25 (2013) 4071–4079. doi: 10.1021/cm402423x

• 

  Figure 1  Rietveld plots of (a) XRD data and (b) neutron data for Cs₂KBiCl₆. The inset to (a) shows a disordered DP structural model with the color code: blue sphere (Cs), red sphere (Cl), yellow sphere (K) and purple sphere (Bi).

  下载: 全尺寸图片 幻灯片
  

  Figure 2  (a) ¹³³Cs MAS NMR spectra (solid lines) and its overall best Gaussian fits (dotted lines) for Cs₂KBiCl₆. Spectra were acquired at a magnetic field strength of 9.4 T with a spinning frequency of 12 kHz. (b) TGA (black) and DTA (red) data for the Cs₂KBiCl₆ sample from room temperature (RT) to 973 K. The dashed baselines are shown in black and blue to guide the eyes for TGA and DTA, respectively.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  (a) UV–vis absorption spectra and (b) the Tauc plot of Cs₂KBiCl₆ and Cs₂NaBiCl₆.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  (a) Excitation (blue) and emission (red) spectra for the x = 2.5% Mn-doped Cs₂KBiCl₆ sample, with its photograph taken under 365 nm illumination shown in the inset. (b) Emission spectra of Cs₂K_1-xBi_1-xMn_2xCl₆ (x = 0–0.05) samples under an excitation wavelength of 363 nm. (c) Luminescent decay curves of the ⁴T₁ → ⁶A₁ transition of Mn²⁺ in the Cs₂K_1-xBi_1-xMn_2xCl₆ samples. The symbols correspond to the experimental data, while the lines show the fitted double-exponential dependence. The inset shows the average lifetime for all the Mn-doped Cs₂KBiCl₆ samples.

  下载: 全尺寸图片 幻灯片

  Table 1.  Final refined structural parameters for Cs₂KBiCl₆.

  下载: 导出CSV

  Table 2.  Experimentally stabilized Cs₂M^ⅠM^ⅢCl₆ DPs showing tolerance factor TF and octahedral factor μ.

  下载: 导出CSV
•