A single Eu²⁺-activated high-color-rendering oxychloride white-light phosphor for white-light-emitting diodes

Abstract

Single-phased, high-color-rendering index (CRI) white-light phosphors are emerging as potential phosphor-converted white-light-emitting diodes (WLEDs) and as an alternative to blends of tricolor phosphors. However, it is a challenge to create a high CRI white light from a single-doped activator. Here, we present a high CRI (Ra = 91) white-light phosphor, Sr₅(PO₄)_3-x(BO₃)_xCl:Eu²⁺, composed of Sr₅(PO₄)₃Cl as the beginning member and Sr₅(BO₃)₃Cl as the end member. This work utilized the solid-solution method, and tunable Eu²⁺ emission was achieved. Color-tunable Eu²⁺ emissions in response to structural variation were observed in Sr₅(PO₄)_3-x(BO₃)_xCl solid solutions. This was further confirmed using X-ray Rietveld refinement, electron paramagnetic resonance spectroscopy, and in the photoluminescence spectra. The color-tunable emissions included the white light that originated from the combination of the blue emission of Sr₅(PO₄)₃Cl:Eu²⁺ and an induced Eu²⁺ yellow emission at approximately 550 nm in the solid solution. Importantly, the white-light phosphors showed a greater R9 = 90.2 under excitation at 365 nm. This result has rarely been reported in the literature and is greater than that of (R9 = 14.3) commercial Y₃A_l5O₁₂:Ce³⁺-based WLEDs. These findings demonstrate the great potential of Sr₅(PO₄)_3-x(BO₃)_xCl:0.04Eu²⁺ as a white-light phosphor for near-UV phosphor-converted WLEDs. These results also provide a shortcut for developing a high CRI white-light phosphor from a single Eu²⁺-doped compound.

Introduction

White-light-emitting diodes (WLED) fabricated with near ultraviolet LED (NUVLED) and tricolor phosphors are a potential substitute for current lighting sources^1,2,3. However, these types of WLED devices are rather complicated, and the color balance is difficult to control⁴. Recently, single-phased, white-light phosphors have attracted more attention because they avoid most of these problems⁴. A conventional strategy using single-phased white-light phosphors is to co-dope ion pairs based on the energy transfer principle, e.g., Eu²⁺→+Mn²⁺, Eu²⁺→Tb³⁺/Mn²⁺, Ce³⁺→Mn²⁺, and Ce³⁺→Tb³⁺/Mn²⁺ pairs^{5,6,7,8,9,10,11,12,13,14}. In these systems, the common feature is that Mn²⁺ is the red-emitting activator. However, the weak Mn²⁺ d-d transitions usually lead to additional excitation quenching paths to reduce the quantum yield (QY) and thermal stability^15,16,17,18. A single-doping activator-activated white-light phosphor that suppresses the energy loss related to energy transfer is highly desired for phosphor-converted WLEDs. However, it is difficult to achieve white light from a compound with a single-doped activator that must also cover the entire visible lighting range from 400 nm to 700 nm.

Solid-solution phosphors are an efficient tool with which to explore white-light phosphors activated by a single-doped activator because this phosphor has a variable composition. This variable composition can simultaneously alter multiple parameters of the host lattice and thereby induce multiple crystal field environments imposed on the emission centers. By utilizing this strategy, some novel phosphors and new luminescence phenomena have been identified, such as tunable emission color, efficient luminous output, and improved thermal/chemistry stability^{19,20,21,22,23,24,25}. For solid-solution phosphors, the end members usually must have high crystal chemical flexibility, or the end members are nearly isotypic despite crystallizing in structures with distinct space groups. Sr₅(PO₄)₃Cl is an important member of the apatite-type halophosphates M₅(XO₄)₃Cl compounds (M = Ca, Sr, Ba, X = P, V, Si) that has been widely investigated for nearly five decades^{8,26,27,28,29}. The compound crystallizes in a hexagonal structure with a P63/m space group, can incorporate various foreign ions, and can provide abundant crystal field environments^26,27,28,29. Sr₅(BO₃)₃Cl crystallizes in an orthorhombic structure with a C2221 space group that has stoichiometrically mimicked apatite-type compounds and has been studied for almost two decades³⁰. Sr₅(BO₃)₃Cl also presents a rich crystal chemistry due to the wide range of possible cationic substitution³⁰. Therefore, a novel oxychloride solid solution with a solid-solution containing the two traditional Sr₅(PO₄)₃Cl and Sr₅(BO₄)₃Cl compounds is possible based on their crystal chemistry similarity. Sr₅(PO₄)₃Cl:Eu²⁺ is a highly efficient blue-emitting phosphor for WLEDs⁸, whereas Sr₅(BO₄)₃Cl:Eu²⁺ phosphor presents a broadband red emission near 615 nm³¹. Therefore, a color-tunable emission that includes white light is expected from this oxychloride solid-solution compound.

In this study, the Eu²⁺-activated Sr₅(PO₄)_3-x(BO₃)_xCl solid-solution phosphors comprised of Sr₅(PO₄)Cl as the oxychloride beginning member and Sr₅(BO₄)₃Cl as the end member were synthesized via a solid-state reaction. Their structures were investigated using powder X-ray diffraction (XRD), XRD Rietveld refinement analysis, and electron paramagnetic resonance (EPR). The optical properties of the phosphors are discussed in detail using crystal structure analysis, photoluminescence (PL) excitation (PLE), transition decays, and fluorescence microscopy images. As expected, a tunable emission including white light with a high color rendering index (CRI; Ra = 91, R9 = 90.2) was achieved from the single-Eu²⁺-doped Sr₅(PO₄)_3-x(BO₃)_xCl solid solution at an excitation of 365 nm. To the best of our knowledge, these are the greatest indices for white-light phosphors reported, especially for those obtained by single-Eu²⁺ doping of an oxychloride compound. This study provides a shortcut for developing a high CRI white-light phosphor from a single-Eu²⁺-doped compound. In addition, this work promotes the development of single-phased white-light phosphors.

Materials and Methods

Materials synthesis

The Sr₅(PO₄)_3-x(BO₃)_xCl powders were made using solid-state reaction routes. The starting materials were SrCl₂⋅6H₂O (99.95%), H₂BO₃ (99.99%), SrHPO₄ (99.95%), SrCO₃ (99.95%), and Eu₂O₃ (99.99%). The powder mixtures were quantitatively mixed using an agate mortar and pestle, and were subsequently sintered for 3 h at 1350 °C in a CO reducing atmosphere. The products were reground and fired again for 1 h at 1350 °C in a muffle furnace and were then ground again for further characterization.

Characterization methods

The structures of the samples were identified using a Rigaku D/max-2500 X-ray diffractometer (Rigaku Co., Japan) with Cu K_α radiation (λ = 0.15406 nm). The detector covers an angular range 10° < 2θ < 105° with a counting time of 5 s per step. The Rietveld refinement used a general structure analysis system (GSAS) program³². The peak shapes were handled using a pseudo-Voigt function with Finger-Cox-Jephcoat asymmetry to correct for axial divergence at low angles. The crystal constants of the samples were calculated using X-ray Rietveld refinement. The room temperature PL spectra were measured with a homemade PL measurement system consisting of a 365-nm NUVLED array as the excitation source and an Ocean Optics USB 4000 plug-and-play spectrometer. The PLE spectra were analyzed using a Horiba FL-3 fluorescence spectrophotometer equipped with a 450 W Xe lamp. Fluorescence microscope PL images (TE-2000 Instruments; Nikon, Japan) of the powder particles were recorded under dark-field mode using a 365-nm UV lamp excitation. Fluorescence decay time curves were measured using an Edinburgh FLS920 spectrophotometer with an nF900 flash lamp. The EPR spectroscopy was recorded using a JES-FA200 (JEOL, Japan) EPR spectrometer. The temperature-dependent luminescence was measured under temperature control (THMS-600) in combination with a homemade PL measurement system consisting of a 365-nm NUVLED array as the excitation source and a micro-Raman spectrometer (Jobin-Yvon HR800) as the detector. A full set of 14 CRIs and the average CRI (Ra) values of the powders were measured using a homemade PL measurement system with a 365-nm UV light source (LA-410, full width at half maximum (FWHM): 5 nm, Japan) and a band-pass filter (Asahi spectra, Japan) as well as the same miniature fiber optic spectrometer with an integrated sphere. The electroluminescent (EL) spectrum and the CRI values of the commercial WLED were also measured using the same miniature spectrometer. The quantum yield (QY) was measured using a PL quantum-yield measurement system (C9920-02, Hamamatsu Photonics).

Results and Discussion

Figure 1 presents the structural differences between Sr₅(PO₄)₃Cl (SPO_Cl) and Sr₅(BO₄)₃Cl (SBO_Cl) and a coordinated polyhedron of anions/cations. SPO_Cl crystallizes as a hexagonal structure and has two cationic sites, referred to as Sr1 and Sr2. These sites reside in the 4f and 6h sites (Figure 1a)⁸. The coordination polyhedron of Sr1/O is composed of six O atoms, and the coordination polyhedron of Sr2/O is composed of two Cl and five O atoms. The P ions are connected with four O ions, forming the PO₄ tetrahedron structure⁸. The crystallization of SBO_Cl in the orthorhombic structure is a complex three-dimensional framework of 7-, 8-, and 9-coordinate Sr-centered polyhedral and BO₃ triangular planes (Figure 1b). There are four cationic sites, which are referred to as Sr1, Sr2, Sr3, and Sr4, that reside in the 4a, 8c, 4b, and 4b sites, respectively³⁰. The B ions are surrounded by three O atoms forming a BO₃ triangular plane with O-B-O bond angles near 120°.

Figure 1

Crystal structures of Sr₅(PO₄)₃Cl (a) and Sr₅(BO₃)₃Cl (b) viewed from the z-direction and the coordination of the Sr/P and B/Sr sites in the end members, respectively.

The calculated and experimental results as well as their differences in the XRD refinement of SPO_Cl are shown in Figure 2a. The initial structural mode was established by the standard crystallographic data of the SPO_Cl (ICSD#0993). The zero point error, scale factor, and background lattice and parameters were refined first followed by the profile parameters (i.e., W, V, U, X, and Y.). Finally, the structural parameters were refined. Based on the Finger-Cox-Jephcoat profile function model, the refinement of the low-angle asymmetry significantly improved the reliability of the pattern fitting. The final reliability factors for the complete pattern were R_wp = 8.46% and R_p = 5.46%. The refined lattice parameters were a = b = 9.872(5) Å, c = 7.187(9) Å, and V = 606.7(3) ų. These results agree well with the results previously reported³².

Figure 2

(a) Observed (crosses) and calculated (solid lines) XRD patterns of the Rietveld refinement of the as-prepared Sr₅(PO₄)₃Cl powder. (b) XRD patterns of the SPBO_{_}Cl:0.04Eu²⁺ (0 ≤ x ≤ 0.5). As a reference, the standard XRD data for Sr₅(PO₄)₃Cl is shown. (c) Magnified XRD patterns in the region between 30° and 32° for the SPBO_{_}Cl:0.04Eu²⁺ (0 ≤ x ≤ 0.5). (d) Unit cell parameters of the SPBO_{_}Cl:0.04Eu²⁺ (0 ≤ x ≤ 0.5) show a contraction in the lattice parameters a and b but only a slight increase in lattice parameter c.

Figure 2b shows the XRD patterns of the Sr₅(PO₄)_3-x(BO₃)_xCl:0.04Eu²⁺ (0 ≤ x ≤ 0.5) powders. The XRD patterns of the Sr₅(PO₄)_3-x(BO₃)_xCl:0.04Eu²⁺ (SPBO_Cl:0.04Eu²⁺) solid solution all maintain the characteristic pattern of the hexagonal structure of the SPO_Cl end member. These are consistent with literature values (Sr₅(PO₄)₃Cl, PDF card No. 16-0666)³³. No detectable XRD peaks corresponding to impurities or other phases were observed, indicating that solid solutions were obtained. When x > 0.5, some by-products appeared in the power sample due to the large mismatch (35%) of the ionic radii of P⁵⁺ (r = 0.17 Å, CN = 4) and B³⁺ (r = 0.11 Å, CN = 4). We therefore believe that the Sr₅(PO₄)_3-x(BO₃)_xCl compounds for 0 ≤ x ≤ 0.5 were primarily solid solutions. The XRD peaks located at 30.3° and 31.4° correspond to the (211) and (300) crystal planes, respectively. These shift slightly to the higher scattering angle with an increase in the B³⁺ ion concentrations, whereas the XRD peak at 30.7° shifts to the lower scattering angle side (Figure 2c). The XRD peaks shift toward the higher angle side. This trend is also seen on almost all of the other diffractive peaks except for the XRD peaks corresponding to the (002), (102), (222), (213), and (004) crystal planes. These observations indicate the substitution of smaller B³⁺ (r = 0.11 Å, CN = 4), for P⁵⁺ (r = 0.17 Å, CN = 4) changes the lattice parameters of the host. As shown in Figure 2d, the calculated lattice constants a and b decrease linearly as x increases, whereas the lattice constant c is mostly constant. The simultaneous expansion and contraction of the unit cell in different directions suggests that the anionic polyhedron substitution of PO₄ with BO₃ causes a greater distortion of the host lattice end member.

The distortion of the host lattice has a significant effect on Eu²⁺ luminescence. As reported, Eu²⁺ occupied the Sr2 sites in the SPO_Cl host lattice more often than the Sr1 sites. Therefore, only one emission band in the blue spectral region was observed^28,29. In our case, SPO_Cl:yEu²⁺ (0.02 ≤ y ≤ 0.1) exhibited an asymmetric blue emission at approximately 446 nm under excitation at 365 nm (see Supplementary Fig. S1). As the Eu²⁺ content increased, the luminous intensity was intensified and reached a maximum at y = 0.04 (see Supplementary Fig. S2). Herein, we consider that most Eu²⁺ ions occupy the Sr2 sites (446 nm) rather than Sr1 sites (462 nm) in the SPO_Cl host, which supports the findings of previous reports^28,29.

Interestingly, an additional distinct emission peak at approximately 550 nm was observed under excitation at the same wavelength when doped with B³⁺ ions (see Supplementary Fig. S1). This observation suggests that the introduction of B³⁺ ions induces a distinct Eu²⁺ center that originates from the Eu²⁺ occupying a distorted crystal field environment. The emission spectra can be deconvoluted into three Gaussian components with peaks centered at approximately 446 and 462 nm (Eu(1)) and 550 nm (Eu(2)), as shown in Supplementary Fig. S1. We assign the higher-energy emission (446 and 462 nm) to Eu(1), which occupies Sr2/Sr1 sites with a weak crystal field. From the X-ray data measured here, it is difficult to determine which sites the induced Eu²⁺ ions can occupy. Therefore, we tentatively assigned the lower-energy emission (550 nm) to Eu(2) occupying the perturbed Sr2 with a strong crystal field. With an increase in the B³⁺-doped content, the QY values of the corresponding samples obviously decrease from 82% at x = 0 to 53% at x = 0.5 at an excitation of 365 nm, which should indicate the formation of more defects in the host lattice. Optimizing the synthesis process is expected to increase the QY.

Color tunability is very significant for phosphors because it facilitates WLED color tuning³⁴. The normalized PL spectra (λ_ex = 365 nm) of the SPBO_Cl:0.04Eu²⁺ powders (x = 0, 0.1, 0.2, 0.3, 0.4, and 0.5) are presented in Figure 3a. The relative intensity of the yellow to blue emission increases gradually with an increase in the B³⁺ content. This reaches a maximum at x = 0.5, whereas the intensity of the blue emission steadily decreases (Figure 3b). Therefore, a tunable color spectrum including white light can be obtained by combining the two emission bands located at approximately 446 and 550 nm in this SPBO_Cl:0.04Eu²⁺ solid-solution phosphors. The change in the emission ratio may be attributed to the higher number ratio of the Eu(2) relative to the Eu(1) sites. These are dependent on the structural variation in the solid-solution phosphor. Conversely, the energy transfer from the Eu(1) to Eu(2) sites also contributes to the change in the emission ratio. This will be shown and discussed in detail in the lifetime decay section.

Figure 3

(a) Normalized PL spectra of SPBO_{_}Cl:0.04Eu²⁺ (0 ≤ x ≤ 0.5) under excitation at 365 nm. (b) The integrated PL intensity of the higher-energy emission and the lower-energy emission of SPBO_{_}Cl:0.04Eu²⁺ as a function of the B³⁺ content (0 ≤ x ≤ 0.5) under excitation at 365 nm. (c), (d), and (e) the fluorescence microscope PL images for SPBO_{_}Cl:0.04Eu²⁺ (x = 0, 0.2, and 0.5) under excitation at 365 nm.

Next, confocal fluorescence microscopy PL images were recorded using the 365-nm UV lamp excitation to demonstrate the PL characteristics of the phosphor particles. The emission color of these microcrystalline powders was visible to the naked eye and was tunable from blue to blue-greenish and ultimately to white luminescence (Figure 3c–3e). The calculated CIE color coordinates can be used to characterize the emission spectra (Figure 7 and Supplementary Table S1). These values range from (0.162, 0.062) at x = 0 to (0.327, 0.361) at x = 0.5. The color tones are in accordance with the results of the fluorescence microscopy PL images.

In addition, the FWHM of the PL spectra was quite broad for samples x ≥ 0.3 and covers a wide range of the visible spectrum. This is helpful for creating high CRI phosphor-converted WLEDs. For example, the composition at x = 0.5 can achieve greater CRI white-light versus commercial YAG:Ce³⁺-based WLEDs. This will be shown and discussed in Figure 8. A continuous redshift is also observed in the two emission bands with an increase in the B³⁺ content (see Supplementary Table S1). In our case, there are two possible causes for the PL redshift: (1) reabsorption; and (2) increased crystal field splitting of Eu²⁺. In reabsorption, the greater energy emission generally resonates with the lower energy part of the excitation spectra. This results in the partial reabsorption of the high-energy emission and a shift in the emission to the red. Herein, it is clear that the higher-energy emission overlaps the low energy PLE spectra (λ_em = 550 nm). The extent of the overlap becomes more prominent with an increase in the B³⁺ content. In the case of increased Eu²⁺ crystal fields, the lowest 5d energy level of Eu²⁺ is strongly influenced by the crystal field splitting. Therefore, the PL spectrum will redshift when the splitting of the Eu²⁺ 5d energy levels increases. As shown in the XRD section, the anionic polyhedron PO₄ units are successively replaced with BO₃ by increasing the B³⁺ content. This reduces the unit cell volume and decreases the Sr-O (Eu-O) bond length. A decreased Eu-O bond length possibly increases the crystal field splitting of the Eu²⁺ ions in the 5d energy level. This causes a continuous redshift in the emission spectra.

To validate this hypothesis, the EPR data were used to investigate the surroundings of the Eu²⁺ ions because the EPR signals from the Eu²⁺ ions are very sensitive to the area surrounding the Eu²⁺ ions (4f⁷, S = 7/2, L = 0). Figure 4 shows the EPR spectra of end member SPO_Cl and SPBO_Cl:0.04Eu²⁺ powders (x = 0, 0.2, 0.4, and 0.5), respectively. At x = 0, there is a very weak EPR signal for the SPO_Cl end member (no Eu) indicating that paramagnetic impurities exist in the host. These impurities may originate from sample defects. The weak EPR signal corresponding to the defects can be neglected compared with the EPR signal of the SPO_Cl:0.04Eu²⁺. Accordingly, the EPR signals of the SPO_Cl:0.04Eu²⁺ can be totally attributed to the unpaired electrons in the Eu²⁺. We consider that the local coordination environments of Eu²⁺ can be divided into two sections: with and without B⁵⁺ (x = 0 and x = 0.2–0.5, respectively). These sections yield different emissions, including Eu(1) at 446 nm and Eu(2) at 550 nm. In contrast, the difference in the EPR signals (indicated by *) is likely due to changes in the Eu²⁺ sites in the host lattice with the addition of B⁵⁺. This result indicates that the introduction of B⁵⁺ induces a distinct Eu²⁺ site that may occupy the perturbed Sr site in the host lattice. Similar results have also been reported in the literature.³⁵ By increasing B⁵⁺ content, we found that the FWHM of the EPR spectra of SPBO_Cl:0.04Eu²⁺ (x = 0.2, 0.4, and 0.5) is gradually broadened. This result presents evidence that the introduction of B³⁺ significantly intensifies the interaction between Eu²⁺ centers and the host lattice³⁶, leading to an increase in the energy-level splitting of Eu²⁺ 5d. This redshifts the PL spectra. In addition, the Stokes shift for both emission bands changes with an increase in the B³⁺ content, but this change is not significant (see Supplementary Table S1). Therefore, we consider the redshift of the PL spectra to be primarily due to reabsorption and increased crystal filed splitting of Eu²⁺. Of course, the contribution of the Stokes shift to the observed redshift cannot be ruled out.

Figure 4

EPR spectra of the end member Sr₅(PO₄)₃Cl (SPO_{_}Cl) and SPBO_{_}Cl:0.04Eu²⁺.

The normalized PLE spectra for Sr₅(PO₄)_3-x(BO₃)_xCl:0.04Eu²⁺ (x = 0, 0.1, 0.2, 0.3, 0.4, and 0.5) samples were obtained using the maxima value of each PL band. Monitoring the blue emission showed that the PLE spectrum has a broad band from 250 nm to 440 nm (Figure 5a). Similarly, when monitoring yellow emission, broad PLE spectra were observed in the range of 250–475 nm (Figure 5b). Both PLE spectra are ascribed to the 4f→5d transitions of Eu²⁺. Nevertheless, different spectral profiles for the two PLE bands suggest that there is a difference in the splitting of the 4f⁶5d¹ excited states for the two sites. This concurs with the PL and EPR results.

Figure 5

(a) Normalized PLE spectra of the SPBO_{_}Cl:0.04Eu²⁺ (0 ≤ x ≤ 0.5) obtained by monitoring the higher-energy emission. (b) Normalized PLE spectra of the SPBO_{_}Cl:0.04Eu²⁺ (0 ≤ x ≤ 0.5) obtained by monitoring the lower-energy emission. Additionally, the PL spectrum of Sr₅(PO₄)₃Cl:0.04Eu²⁺ is also shown as a reference.

When increasing the B³⁺ content, the Eu²⁺ 5d bands in both PLE spectra shift to the lower-energy side along with profile changes. This is a consequence of the stronger crystal field splitting of the Eu²⁺ ions’ 5d energy level as discussed previously. In addition, the bandgaps (E_g) of SPO_Cl, SPBO_Cl, and SPBO_Cl:0.04Eu²⁺ are calculated from their reflectance spectra. The undoped sample shows no absorption in the wavelength range greater than 375 nm and is thus white, which agrees with the PLE results. The E_g of SPO_Cl is estimated to be 3.836 eV. The SPBO_Cl:0.04Eu²⁺ has a strong absorption from 250 nm to 475 nm, and the color turns light yellow (see Supplementary Figure S3). This result is attributed to the 4f–5d transitions at the Eu²⁺ center. The E_g of SPBO_Cl:0.04Eu²⁺ is estimated to be 2.850 eV. The PL spectra (λ_ex = 365 nm) of SPO_Cl:0.04Eu²⁺ is also shown in Figure 5b as a reference. Obviously, there is a significant spectral overlap between the blue emission located at 446 nm (Eu(1)) and the PLE spectra monitored by the lower-energy emissions of Eu(2) from 410 nm to 470 nm. This result implies the possibility of energy transfer from the Eu(1) to the Eu(2) site.

To further investigate the dynamic luminescence process between the higher-energy emission sites and the lower-energy emission sites, the lifetime decay of the higher-energy emission sites in the SPBO_Cl:0.04Eu²⁺ powders (x = 0, 0.1, 0.3, and 0.5) was recorded (λ_ex = 365 nm; Figure 6a). At x = 0, the decay curve of the higher-energy emission at 446 nm presents a single exponential decay behavior. However, the fluorescence decay of the higher-energy emission becomes faster with an increase in the B³⁺ content, and the decay curves are non-exponential. This suggests that the incorporation of B⁵⁺ ions changes the luminescence dynamic process of Eu²⁺. The average lifetime (τ) of these samples can be estimated using Equation (1)³⁷:

Figure 6

(a) The lifetime decay curves for the Eu(1) in SPBO_{_}Cl:0.04Eu²⁺ phosphors (x = 0, 0.1, 0.3, and 0.5) excited at 365 nm and monitored at 446 nm. The inset shows the calculated energy transfer efficiency (η_ET) from Eu(1) to Eu(2) with a varying B³⁺ doping content. (b) The lifetime decay curves for Eu²⁺ in the SPBO_{_}Cl:0.04Eu²⁺ (x = 0.1) phosphor excited at 275 nm and monitored at 448 and 564 nm. The lifetime decay curves were fitted using the double-exponential equation I(t) = I₀ + A₁exp(−t/τ₁) + A₂exp(−t/τ₂).

where I(t) is the luminous intensity at time t. According to Equation (1), the average lifetimes of higher-energy emission are 0.56, 0.47, 0.39, and 0.32 μs for B³⁺ values of 0, 0.1, 0.3, and 0.5, respectively. The decrease in the lifetime of the higher-energy emission with an increase in the B³⁺ content strongly demonstrates an energy transfer from the higher-energy emission sites to the lower-energy emission sites. In this case, the calculated lifetime values corresponding to the higher-energy emission cannot be completely assigned to the 446-nm Eu(1) sites because the 446- and 462-nm emissions from the Eu(1) site are simultaneously recorded (Supplementary Fig. S1). With regard to the lifetime of the Eu(2) acceptor, an initial increasing trend should be observed if the energy transfer from the higher-energy emission to the lower-energy emission sites definitely occurs. Figure 6b shows the decay profile of the lower-energy emission (λ_em = 550 nm) for x = 0.1 samples at an excitation of 275 nm. This clearly shows that the decay profile is nonexponential with a prominent initial increasing trend. The calculated average lifetime using Equation (1) is 1.44 µs with a buildup characterized by a constant ∼0.37 µs. In contrast, the decay monitored the higher-energy emission (λ_em = 446 nm) is multi-exponential with the fastest component at ∼0.38 µs. The similarity in this time constant for the fastest decay component of the higher-energy emission as well as the increase in the lower-energy emission also suggests the occurrence of energy transfer from the higher-energy emission sites to the lower-energy emission sites.

The different decay profiles and lifetime values for the blue and yellow emissions further support the notion of the two emission bands from two distinct Eu²⁺ centers. In addition, the energy transfer efficiency (η) from the higher-energy emission to the lower-energy emission sites can be estimated using Equation (2)^38,39:

where τ_so and τ_s are the lifetimes of the higher-energy emission sites in the absence and in the presence of the lower-energy emissions, respectively. As shown in the inset of Figure 6b, η increases gradually with an increase in the B³⁺ content. More precisely, the energy-transfer efficiencies, η, are calculated to be 0%, 16.9%, 26.7%, 31.2%, 41.7%, 42.1%, and 44.2% for SPBO_Cl:0.04Eu²⁺ when x = 0, 0.1, 0.2, 0.3, 0.4, 0.5, and 0.6, respectively.

To evaluate the potential application of an as-prepared sample, the PL spectrum of the SPBO_Cl:0.04Eu²⁺ phosphor (x = 0.5) at a 365-nm excitation is recorded and presented in Figure 8. White light can clearly be generated by combining the blue emission located at approximately 455 and the yellow emission at 550 nm. Interestingly, the white-light emission profile of the as-prepared sample is very similar to the emission profile of the commercial WLED (blue LED chip + Y₃A_l5O₁₂:Ce³⁺ (YAG:Ce³⁺) yellow-emitting phosphor).

Figure 7

The CIE chromaticity coordinates the SPBO_{_}Cl:0.04Eu²⁺ phosphors (x = 0, 0.1, 0.2, 0.4, and 0.5) under excitation at 365 nm.

Figure 8

The PL spectrum of Sr₅(PO₄)_2.5(BO₃)_0.5Cl:0.04Eu²⁺ phosphor under excitation at 365 nm (red line). The electroluminescence spectrum of the commercial YAG:Ce³⁺-based WLED driven by forward-bias current at 20 mA is also plotted as a reference.

For comparison, the EL spectrum of the commercial WLED driven by a forward bias at 20 mA is also presented in Figure 8. The full set of 8 CRIs and the average CRI (Ra) of the as-prepared sample as well as the commercial WLEDs are listed in Supplementary Table S2, respectively. All 8 CRIs are greater than the YAG:Ce³⁺-based WLED. The as-prepared sample has a greater Ra value (∼91.0) versus the YAG:Ce³⁺-based WLED (∼80.2). Ra is widely used in the lighting community to evaluate the performance of artificial lighting including WLEDs^40,41,42; however, a high Ra does not completely guarantee the good saturation of colors of illuminated objects⁴³ because the Ra is calculated only using the first eight (R1–R8) values of the CIE from the recommended 14 test color samples⁴³. Therefore, the special CRI R9 value, which denotes the color reproduction in the strong red region and is very important for biomedical and painting^{44,45,46,47,48}, is critical for high-quality color renditions of the light source. Therefore, we found that the R9 of the as-prepared sample is greater (∼90.2) than the R9 of the current available lamp sources, e.g., halo-fluorescent lamps (∼–81)⁴⁷, tri-phosphor fluorescent lamps (∼8)⁴⁷, and commercial WLEDs (∼14.3) (see Supplementary Table S3). Compared with the YAG:Ce³⁺-based commercial WLED, the as-prepared sample has greater Ra and R9 values. This is consistent with the fact that there was greater number of blue-greenish (470–515 nm) and deep-red components in the emission spectrum and a wider spectral distribution (∼137 nm) at approximately 550 nm compared with the YAG:Ce³⁺ phosphor (∼104 nm).

Thus far, there have been many investigations of single Eu²⁺/Ce³⁺-activated white-light phosphors, e.g., (BaSrMg)(PO₄)₂:Eu²⁺ (Ra = 87)¹⁹, (Na_1−xCa_x)(Sc_1−xMg_x)Si₂O₆:Eu²⁺ (Ra = 72)²⁰, (Gd_1−xSr_2+x)AlO_5-xF_x:Eu²⁺ (Ra = 53)²¹, Sr₂Ba(AlO₄F)_1−x(SiO₅)_x:Ce²⁺ (Ra = 70)²², (La_0.075-xCe_0.025Sr_2+x)(Al_1−xSi_xO₅) (Ra = 70)²³, and (Sr_2.975Ce_0.025)Al_1−xSi_xO_4+xF_1−x (Ra = 69)²⁴. However, indices greater than Ra = 91.0 and R9 = 90.2 have not yet been reported. These results indicate that this single-phased SPBO_Cl:0.04Eu²⁺ white-light phosphor is a promising candidate for NUVLED-excited devices for general illumination purposes.

Thermal stability is an important technological parameters for a WLED phosphor. The thermal quenching behavior of the SPBO_Cl:0.04Eu²⁺ phosphor (x = 0.5) in the temperature range of 20–200 °C under an excitation at 365 nm is shown in Supplementary Fig. S4. The emission intensity of Eu²⁺ decreased with an elevated temperature (Supplementary Fig. S4a). The emission intensity decreased to 50% of the initial values at 120 °C (Supplementary Fig. S4b). Furthermore, the thermal quenching temperatures for both emission bands of the SPBO_Cl:0.04Eu²⁺ phosphor were different (Supplementary Fig. S4c). The thermal quenching temperature for the yellow emission was 150 °C, whereas that for blue was 90 °C. Different thermal quenching temperatures for the blue and the yellow emissions may originate from the distinct Eu²⁺ ions with different rigid environments. As for the solid-solution white-light phosphor (x = 0.5), the thermally activated energy was calculated in terms of the Arrhenius equation I(T) ≍ I₀[1+A exp (–ΔE/k_BT)]^–1 (where I₀ is the initial intensity, I(T) is the intensity at a given temperature T, A is a constant, ΔE is the activation energy for the thermal quenching, and k_B is Boltzmann’s constant). The large decrease in luminescence might be attributed to the smaller barrier energy (∼0.269 eV).

Conclusions

The Sr₅(PO₄)_3-x(BO₃)_xCl:0.04Eu²⁺ phosphors demonstrate a continuously tunable emission and were obtained using a solid-solution method and a solid-state reaction. Based on the Rietveld and EPR analysis, substituting PO₄ with BO₃ anionic polyhedrons contracts and distorts the lattice of the end member Sr₅(PO₄)₃Cl of the solid solution. This induces a new Eu²⁺ yellow emission at approximately 550 nm. The distinct Eu²⁺ center originates from the Eu²⁺, occupying a different crystal field site. Combining a blue emission near 446 nm and an induced yellow emission at 550 nm results in a tunable emission including white light via 365-nm excitation. Significantly, the white-light phosphor presents greater Ra and R9 values (91 and 90.2, respectively) relative to commercial YAG:Ce³⁺-based WLEDs (Ra = 80 and R9 = 14.3). This study provides an important shortcut for developing a high CRI white-light phosphor from single-Eu²⁺-doped compounds.