Vacancy-ordered Te⁴⁺-doped Rb₂ZrCl₆ double perovskite microcrystals for solid-state lighting and non-contact optical thermometry

In recent years, lead-based perovskite materials have attracted great interest in many fields, such as solar cells,¹ photodetectors,^2–4 x-ray scintillators,^5,6 and white light-emitting diodes (WLEDs),^7–9 for their remarkable light-harvesting and emission characteristics. However, lead-based perovskites have two major problems that need to be solved, one is the toxicity of lead ions, which is extremely hazardous, and the other is their poor stability and susceptibility to decomposition under humid or high temperature environment. To solve these issues, the first idea is to replace the noxious Pb²⁺ ions with Sn²⁺ or Ge²⁺ ions, which has low toxicity and similar ion radius. Sn²⁺ is considered as a promising candidate to replace Pb²⁺ because it has similar electronic configurations and the same valence state; however, Sn²⁺ is easily oxidized to Sn⁴⁺. The change in the crystal structure can be understood as the replacement of two Pb²⁺ by a tetravalent cation M^IV (Sn⁴⁺, Zr⁴⁺, Te⁴⁺), resulting in a 50% periodic vacancy in the octahedron, forming the double A₂M^IVX₆ type perovskite with a vacancy-ordered structure.¹⁰ After continuous exploration, the double perovskites family has been expanded, and it has been found that A₂M^IM^IIIX₆ type double perovskites are formed when two Pb²⁺ are replaced by the combination of a monovalent (Ag⁺, Na⁺) and trivalent (Bi³⁺, Sb³⁺, In³⁺) cation,^11,12 respectively. The doping of metal ions in the host is an effective means to improve its optical properties, providing new ideas to study emerging optical and optoelectronic applications. In 2017, Tang et al. reported that Bi³⁺-doped Cs₂SnCl₆ showed blue self-trapped exciton (STE) emission, photoluminescence quantum yield (PLQY) close to 80%, and displayed impressive thermal and water stability.¹³ In 2019, Xia et al. synthesized Sb³⁺-doped Cs₂SnCl₆ nanocrystals exhibiting bimodal broad-band emission at 348 and 615 nm, and they attributed the broad-band emission to the ³P_n–¹S₀ transition, providing insight into the transitional mechanism of the doping-induced emission centers.¹⁴ The B-site is gradually replaced by other metal ions besides Sn⁴⁺, such as Zr⁴⁺¹⁵ or Hf⁴⁺,¹⁶ but the compounds formed by them usually have large bandgap and require deep UV light to excite, which limits their application prospects of Zr/Hf-based vacancy-ordered perovskites to some extent.^17,18 Hence, the mismatch between phosphors and commercial UV chips with high excitation in the field of next-generation solid-state lighting is one of the key problem that urgently needs to be solved.¹⁹ Additionally, vacancy-ordered double perovskites are strong candidate for non-contact optical thermometry. Optical thermometry is usually divided into fluorescence intensity ratio (FIR)¹² and fluorescence lifetime (FL)²⁰ thermometry, where the thermal stability of material is the basis for wide-range thermometry, and dramatic change in PL intensity or FL with temperature is the critical for high sensitivity.²¹ Developing multiple applications for materials is an effective approach to cost reduction, so we applied the RZCT to both WLED and non-optical thermometry.

Vacancy-ordered perovskite Rb₂ZrCl₆ (RZC) and Te⁴⁺-doped Rb₂ZrCl₆ (RZCT) microcrystals with the feeding ratio changes have been synthesized by a typical hydrothermal method, and the schematic diagram of the synthesis is shown in Fig. 1(a). After doping with Te⁴⁺, the Te atoms replace the Zr atoms to form [TeCl₆]²⁻ octahedra, and the crystal structure after doping is shown in Fig. 1(b). In order to determine the conformation and purity of the synthesized products, powder x-ray diffraction (PXRD) is performed. The PXRD patterns of RZC and RZCT with different doping levels are presented in Fig. 1(c). The PXRD pattern of the RZC sample matches very well with its standard PXRD pattern (JCPDS card 74–1000), and all the diffraction peaks are in good agreement with the standard pattern, no other impurity phases are observed, indicating that the synthesized samples are pure. It should be noted that after Te⁴⁺ ions doping, due to the increase in ion radius (Te⁴⁺ > Zr⁴⁺), the lattice will expand and the diffraction peaks will be shifted to a small angle. The Raman spectra of the samples are also measured at room temperature (RT) [Fig. 1(d)], and the Raman peaks of RZCT are shifted to lower wave numbers compared to RZC, and the shift increases with the increase in Te⁴⁺ content, which indicates the successful doping of Te⁴⁺.²² It is noteworthy that the Raman peaks attributed to Te–Cl bond vibrations²³ appear at 250 and 291 cm⁻¹ when the doping amount of Te⁴⁺ reaches 20%. The absorption spectra and steady-state photoluminescence (PL) spectra are used to explore photophysical features of RZCT. Following the previous literature,¹⁵ RZC emits blue light at a wavelength of about 460 nm due to STE emission under 240–260 nm laser excitation (Fig. S1). Meanwhile, Te⁴⁺ doping causes the host to generate extra absorption peaks, resulting in a reduced bandgap and thereby excitation of luminescence without deep UV light. As shown in Fig. 1(e), RZCT exhibits strong absorption in the range of 350–450 nm compared with RZC, with increasing Te⁴⁺ content, the absorption band edge is red-shifted and the absorption intensity is further enhanced.²⁴ With the addition of Te⁴⁺, the PL intensity shows a trend of increasing and then decreasing, and the PL intensity reaches the maximum at 5% Te⁴⁺ doping. (Normalized absorption and PL spectra of doped samples are presented in Fig. S1.) Therefore, in the following, RZCT all represent the doped sample with a feeding ratio of 5%. The bandgap decreases from 3.61 to 3.07 eV as the amount of Te⁴⁺ substitution increases [Fig. 1(f)].

As shown in Fig. 2(a), the high-resolution x-ray photoelectron spectroscopy (XPS) exhibit two peaks at 588 and 574 eV, which are attributed to the 3d_5/2 and 3d_3/2 levels of Te⁴⁺, respectively. These peaks indicate the presence of Te⁴⁺ in the doped sample and suggest that Te is a positive 4-valent state. Figures S2 and S3 represent the high-resolution XPS enlargements of RZC and RZCT, respectively, from which it can be seen that the peaks of Rb⁺ do not move, but the peaks of Zr⁴⁺ and Cl⁻ are shifted to the lower energy direction, further indicating that Te⁴⁺ replaces the original Zr⁴⁺ element.²⁵ Scanning electron microscopy (SEM) and energy dispersive spectroscopy (EDS) are used to study the elemental distribution and micromorphological structure of RZCT. The SEM image in Fig. 2(b) clearly depicts the RZCT structure in a regular, defined shape. Furthermore, the accompanying EDS mapping images illustrate that the Cs, Zr, Te, and Cl elements exhibit uniform distribution throughout the crystal. To further explore the luminescent mechanism of RZCT, we collected its PL and PLE spectra of wavelength-dependent, as shown in Figs. 2(c) and 2(d). The PLE color map shows two different excitation peaks at 330 and 380 nm, which can be attributed to the ¹S₀ → ¹P₁ and ¹S₀ → ³P₀ transitions¹⁵ and ¹S₀ → ³P₀ dominating the excitation absorption.²⁶ In addition, the wavelength-dependent PLE and PL spectra both show negligible variations, which suggests that the yellow emission is an intrinsic property from RZCT. The 405 nm excitation power-dependent spectrum is measured, and the PL showed a strong linear increasing trend with the increase in pump energy, indicating that the PL emission is also not attributed to intrinsic defect luminescence [Fig. 2(e)],²⁰ and the measured PLQY can reach 34.6% [Fig. 2(f)].

To further explore the PL emission mechanism of RZCT, we performed temperature-dependent PL spectra measurements. Figure 3(a) presents the PL spectrum at 80–340 K, which can be found that the intensity of PL gradually decreases with increasing temperature and it is accompanied by a broadening of the full width at half maximum (FWHM), which can usually be attributed to the increased non-radiative relaxation and the enhanced electron–phonon coupling at high temperature.^27, Figure 3(b) is the pseudo-color plot of normalized PL intensity and wavelength with temperature, with almost invariant peaks in the temperature range of 80–340 K. As shown in Fig. 3(c), the calculated values of S and ℏω_phonon are 18.39 and 27.32 meV, which indicate the presence of strong electron–phonon coupling in the RZCT.²⁸ In total, such large S values, broad emission spectra and Stokes shifts suggest that its PL emission is attributed to STE emission, caused by the strong Jahn–Teller distortion of the [TeCl₆]²⁻ octahedron.^23,24 The E_b of RZCT be calculated is 137.9 meV [Fig. 3(d)]. The binding energy is much higher than that of lead-based perovskites suggesting the structural stability of RZCT and forming the stable STE emission with highly localized behavior.²⁴ To investigate the effect of temperature on the crystal structure, we studied the dominant Raman vibrational modes at different temperatures of RZC (Fig. S4) and RZCT. Two clear Raman vibrational peaks are observed consistently about 158 to 162 cm⁻¹ for bending vibrations (T_2g) and 328 to 336 cm⁻¹ for stretching vibrations (A_1g)²⁹ of RZCT in the temperature range of 80–480 K [Fig. 3(e)]. No further feature peaks appear indicating that the sample is not involved in phase transformation. As shown in Fig. 3(f), we have performed a detailed analysis of the peak positions and FWHM. As the temperature increases, the Raman peak is mildly shifted to lower wave numbers and is accompanied by a broadening of the FWHM due to an increase in the phonon scattering process.³⁰ 

In order to gain better understanding of the PL properties and emission mechanisms, we conducted measurements of the decay curves of RZCT. The PL lifetimes of samples with different levels of Te⁴⁺ at RT are displayed in Fig. 4(a). A definite relationship between PL lifetime and Te⁴⁺ concentration can be seen, with gradual decrease in PL lifetime as Te⁴⁺ increases. These results suggest that Te⁴⁺ may be the source of the localized emission that generates the PL emission observed in RZCT²⁹ (fitting data are shown in Table S1). At higher Te⁴⁺ concentration, the fully separated [TeCl₆]²⁻ octahedra will be converted to [TeCl₆]²⁻ and the local excitons will be reduced, which leads to self-absorption and self-quenching,³¹ thus making the PL lifetime shorter. Figure 4(b) shows pseudo-color map of the PL lifetime of RZCT at RT, and the inset is the decay of the PL spectrum from 1.06 to 2.68 μs. The temperature dependent time-resolved PL (TRPL) spectra of the RZCT are measured between 80 and 340 K. Figure 4(c) displays that the PL lifetime notably increases with the decrease in temperature. Moreover, Fig. 4(d) displays the TRPL pseudo-color plot as a function of temperature, meanwhile the inset depicts the correlation between PL lifetime and temperature (fitting data are shown in Table S2). At low temperatures owing to the dominance of radiative recombination, thus having longer decay lifetimes, as the temperature increases, non-radiative recombination will dominate leading to a decrease in PL lifetime.³² It should be noted that besides the large Stokes shift, the microsecond decay time is also one of the features of the STE emission,²⁰ thus further identifying the broadband emission of RZCT from the SET emission due to the deformation of [TeCl₆]²⁻. In Fig. 4(e), it can be observed that the PL lifetime remains almost unchanged after four cycles, spanning 80 to 340 K, suggesting that RZCT holds excellent thermometry repeatability. Hence, this property makes it an ideal candidate for optical fluorescence lifetime thermometry. The property of temperature sensing is usually referred to as the sensitivity of optical thermometers, and relative sensitivity (S_R) and absolute sensitivity (S_A) are crucial parameters for evaluating the performance of thermometric materials.³³ As can be seen from Fig. 4(f), the maximum values of S_R and S_A are as high as 0.89% and 4.76 × 10⁻³ K⁻¹, respectively, which is significantly higher than other thermometry materials (Table S3), demonstrating that RZCT has remarkable thermometric performance and outstanding application prospects in the field of non-contact thermometry.

A density functional theory (DFT) is used to calculate the energy bands of RZC and RZCT as well as the density of states (DOS). As shown in Figs. 5(a) and 5(b), the calculated bandgap values are 3.56 and 2.93 eV for the pure and doped structures, respectively, which are consistent with our experimental values of 3.61 and 3.07 eV. As shown by the DOS of Figs. S5 and S6, for RZC, its valence band minimum (VBM) consists mainly of Zr⁴⁺ 4d and Cl⁻ 3p orbital hybridization, while the conduction band maximum (CBM) is composed of Cl⁻ 3p orbitals. For RZCT, its CBM is almost constant, but the bandgap value decreases due to the addition of Te⁴⁺ ions forming a new VBM′ above the original VBM contributed by Te⁴⁺ 5s orbitals and Cl⁻ 3p orbitals hybridization. It can be seen more visually from the VBM′ and CBM charge densities distributed within the [TeCl₆]²⁻ and [ZrCl₆]²⁻ octahedra³⁴ [Figs. 5(c) and 5(d)]. The electron cloud prefers to be localized in octahedra, and there is no overlap between them with less energy loss, which is favorable for the formation of STE.³⁵ It is pointed out from the calculations that before and after the doping of Te⁴⁺, the CBM and VBM are always at the same K point and the bandgap type is always direct bandgap, which does not require phonon assistance during the electron leap, which may be the reason for the higher PLQY.³⁶ Therefore, we obtain a schematic representation of the RZCT orbital energy mechanism as shown in Fig. 5(e). In the overview, we can obtain the PL emission mechanism diagram as shown in Fig. 5(f). Similar to other ns2 electron configuration ions,^37,38 Te⁴⁺ ion has a total of five energy levels, namely, the ground state ¹S₀, the excited singlet state ¹P₁, and the triplet state ¹P₁/³P_n (n = 0, 1, 2). When it is excited, electrons are pumped from the ground state to the ¹P₁ and ³P₁ states, and since the ¹P₁ to ¹S₁ leap is forbidden,³⁹ the electrons of ¹P₁ will fall back to ³P₁ and then progress to the STE state and go back to the ground state to produce bright yellow emission.

Stability is an important indicator used to evaluate the potential for applications in the lighting field. Hence, the photostability, environmental stability, and thermal stability of RZCT are investigated. From Fig. 6(a), the spectral shape and peak position of the PL spectrum did not change significantly after 5 h of continuous irradiation, and the intensity change was less than 8% (Fig. S7). The PXRD measured did not change significantly after 2 weeks of storage [Fig. 6(b)]. The thermogravimetric analysis (TGA) results in Fig. 6(c) demonstrate the excellent thermal stability of the sample. Non-toxic, stable, and high PLQY broadband emission properties of RZCT indicate that it has the potential to be used as a phosphor for solid-state lighting. Figures 6(d) and 6(e) show the emission spectrum of the RZCT excited by the 395 nm UV chip with different current values and the inset shows the picture of the LED state when corresponding to the input current. In addition, by varying the current level, the LED exhibits linear evolution of the CIE coordinates for the regulation of the warm and cold white light. Figure 6(e) shows the evolution of the CIE coordinates for currents of 30–150 mA. Table S4 displays the correlated color temperature (CCT), CIE coordinates, and the light emitting efficiency of WLED with different current values. Comparison of light efficiency of other similar materials is presented in Table S5.

In conclusion, we synthesized vacancy-ordered RZCT perovskite microcrystals by a hydrothermal method, which exhibits a broadband yellow emission with a peak at 573 nm under 405 nm laser excitation. The source of the broadband emission was investigated systematically by measuring temperature-dependent spectra and TRPL spectra and finally attributed to the deformation caused by Te⁴⁺ ion doping, which leads to broad-band STE emission. Furthermore, RZCT is used as a solid-state lighting phosphor to produce WLED devices with highly efficient luminescence that can be converted from warm to cold white light by adjusting the input current. In addition, the PL lifetime of RZCT was used for optical thermometry, and the PL lifetime of RZCT showed a significant decrease in the temperature range of 80–340 K, with S_R and S_A reaching a maximum of 0.89% K⁻¹ at 220 K and 4.76 × 10⁻³ K⁻¹ at 180 K, respectively. Our work provides a useful strategy for the development of high-efficiency solid-state lighting and optical thermometric materials.