Fine-Tunable Self-Activated Luminescence in Apatite-Type (Ba,Sr)5

Sep 12, 2018 - For a neutral defect (q = 0), the last item in eq 1 is omitted. ... Because the ionic radius of Sr2+ (CN = 9, 1.31 Å; CN = 8, 1.26 Å)...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/IC

Cite This: Inorg. Chem. XXXX, XXX, XXX−XXX

Fine-Tunable Self-Activated Luminescence in Apatite-Type (Ba,Sr)5(PO4)3Br and the Defect Process Jing Zhang,†,⊥ Tingting Zhang,‡,⊥ Zhongxian Qiu,*,† Shubin Liu,*,§ Jilin Zhang,† Wenli Zhou,† Liping Yu,† and Shixun Lian*,†

Inorg. Chem. Downloaded from pubs.acs.org by UNIV OF SOUTH DAKOTA on 09/12/18. For personal use only.



Key Laboratory of Chemical Biology & Traditional Chinese Medicine Research (Ministry of Education), Key Laboratory of Sustainable Resources Processing and Advanced Materials of Hunan Province, College of Chemistry and Chemical Engineering, Hunan Normal University, Changsha, Hunan 410081, China ‡ College of Chemistry and Molecular Engineering, Peking University, Beijing 100871, China § Research Computing Center, University of North Carolina, Chapel Hill, North Carolina 27599-3420, United States S Supporting Information *

ABSTRACT: Intrinsic defect-related luminescence has recently been attracting more research interest for the modification of phosphors. However, the connection between defect formation and crystal structure has never been considered. In this work, we report that in the absence of an impurity activator, under a reducing atmosphere, apatite-type compound M5(PO4)3X (M = Ca, Sr, or Ba; X = F, Cl, or Br) can emit tunable colors ranging from blue to orange depending on the content of M and X. To better understand the cause, Ba5−mSrm(PO4)3Br (BSPOB; m = 0−5) solid solutions were analyzed in detail. The dependency of self-activated luminescence on atmospheric conditions and solid solution compositions was investigated by combining experimental characterizations and theoretical calculations using density functional theory. Crystal structures of these solid solutions were verified by X-ray diffraction patterns as well as Rietveld refinements. With the defect formation energy and electron paramagnetic resonance measurement, we propose that an oxygen vacancy (VO) should be mainly responsible for the peculiar super wide band emission. Moreover, the enhanced distortion of solid solution crystal structures augments VO concentrations and leads to luminescence intensities in solid solutions that are higher than that in end point compounds. Variations of the electronic structure of BSPOB matrices with gradual tuning of the Sr/Ba ratio were also investigated. As a result, the introduction of VO defect levels within the band gap leads to the formation of donors and acceptors, allowing for a modulation of the photoluminescence throughout the visible part of the spectrum. As the first report in the literature to demonstrate finetunable emissions over a wide wavelength range as a consequence of native defective levels in a series of continuous apatite-type solid solutions, our results illustrate the feasibility of defect-meditated systems by carefully tailoring defect chemistry and nonstoichiometric chemical composition under controlled conditions to engineer phosphor properties.

1. INTRODUCTION

ions are relatively expensive, not to mention the high cost of environmental pollution during their preparation and purification. The other type of luminescent materials is called a selfactivated phosphor without any doping ions. This second type has recently been attracting more research interest in the literature because of its outstanding advantages of less toxicity, inexpensive cost, and high efficiency. It is also our focus in this work. In general terms, the self-activated luminescence originates from charge transfer in either an anionic complex or an intrinsic defective center. The former is common in ABO4-type compounds (A = Ca, Zn, Y, or La; B = W, Mo, V, Nb, etc.).7−9 The latter includes vacancies, self-interstitials, and antisites and

The technology of phosphor-converted white light-emitting diodes (pc-WLEDs) for general lighting is attracting extensive research and commercial interests and has been extensively integrated into our daily lives because of its environmentally friendly nature, robustness, and long lifetime.1 Phosphors with versatile luminescent properties are vitally important to the application of WLED devices. In general, there are two major classes of phosphors. One is composed of an optically inactive host lattice and impurity activators with low and optimized concentrations. Rare earth and transition metal ions are usually used as activators to generate multicolor lights in many phosphors like yellow Y3Al5O12:Ce3+ phosphors,2 blue BaMgAl10O17:Eu2+ (BAM:Eu2+),3 green β-SiAlON:Eu2+,4 red CaAlSiN3:Eu2+,5 warm white light-emitting Ca5(PO4)3(F,Cl):Sb3+,Mn2+,6 etc. However, these activator © XXXX American Chemical Society

Received: July 25, 2018

A

DOI: 10.1021/acs.inorgchem.8b02105 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

been investigated. A one- or two-stage surface annealing treatment was used at different atmospheres to explore the influence of the oxygen partial pressure on the self-activated luminescence behavior. Density functional theory (DFT) calculations were performed to estimate the possibilities of different intrinsic defects. The evolution of the band structures of BSPOB solid solution compounds was taken into account for the gradual spectral blue shift with the continuously increased Sr/Ba ratio. The variation of fine structures and degrees of distortion in defective models were also further examined to understand the relationship of the defect formation energy with local crystalline structures. The purpose of this study is to elucidate the intrinsic defects responsible for the self-activated luminescence in MPOX host materials and the formation process during the high-temperature reaction. This work should provide highly desirable insights and much needed perspectives for defect-related luminescence in haloapatite materials and for the development of novel defectmeditated phosphors by carefully tailoring the chemical composition and surface defect chemistry.

forms defect levels within the band gap acting as recombination centers and giving rise to visible light emissions at specified wavelengths.10 Defect-related luminescence has been examined for decades in oxide-, chalcogenide-, and nitridebased semiconductors, such as ZnO,11 CdSe,12 GaN,13 Ca3B2N4,14 Mg2SnO4,15 etc., where the defect emissive centers have been extensively studied and widely accepted. However, in recent years, unique defect luminescence has been discovered in oxysalts such as oxyfluoride Sr3−xAxMO4F (A = Ca or Ba; M = Ga or Al),16 fluosilicate La3F3[Si3O9]17 and Ca4F2Si2O7,18 phosphate α-Zn2P2O719 and BPO4,20 apatitetype M5(PO4)3OH (M = Ca or Sr),21,22 Ca5(PO4)3F23 and Ca2Ba3(PO4)3Cl,24 etc. It is noteworthy that the wide band emitting property of the oxysalts mentioned above usually relies on synthetic conditions. With respect to the complex composition and crystalline structure, the defect process and luminescent mechanism remain unclear. Much effort has been devoted to determining the charge transition in a defective system. Carbon impurities, peroxyl radicals, and/or anion vacancy defects were proposed to be responsible for defect-related luminescence.25 Lately, more focus has been shifted to anion vacancies in high-temperature solid reactions.16,18 Oxygen and/or halogen atom vacancies lead to a deviation from the stoichiometry to some degree. The defective center can be inferred through experimental measurements via, e.g., X-ray photoelectron spectroscopy (XPS),26 electron paramagnetic resonance (EPR),27 positron annihilation spectroscopy (PAS),28 and electron energy loss spectroscopy (EELS) via atomic-level resolution of transmission electron microscopy (TEM),29 but how can we assign these charge transitions and ascertain the defect levels that are responsible for the emission? Inspired by recent studies in semiconductors, computational studies in combination with experimental efforts can yield insightful and quantitative details about the impact of point defects.10,13,30−34 The defectinduced electronic states and broad emission bands as a result of different chemical bonding environments in the ground and excited states can also be undertaken using this approach.16 Apatite-type compounds with excellent thermal and chemical stabilities are well-established as host materials for phosphors because of their compositional flexibility, high efficiency, and wide compatibility with doping ions.35 The apatite-type lattice usually possesses an indirect band structure. It is hard for electrons to achieve a direct transition between the valence band (VB) and conduction band (CB). The introduction of defect levels within the band gap by doping or intrinsic defects has proven to be an effective method for altering its electronic structure, thereby generating emission transitions. Intense bright blue and cyan emissions from Ca5(PO4)3F23 and Ca2Ba3(PO4)3Cl,24 respectively, have been reported. However, the specific assignment of defects responsible for visible light emission is still controversial. In this work, a series of halo-apatite compounds M5(PO4)3X (MPOX; M = Ca, Sr, or Ba; X = F, Cl, or Br) were synthesized via the conventional high-temperature solid method. Different atmospheric conditions were used, leading to color-tunable luminescence in MPOX under a reducing atmosphere in the absence of doped activators. The (Ba,Sr)5(PO4)3Br (BSPOB) system with heavier atoms was specially chosen to avoid the interference of vacancy defects caused by the volatilization of alkaline metal and halogen atoms. The dependencies of the fine-tuned spectrum and luminescence efficiency on synthetic conditions as well as on solid solution crystal structures have

2. EXPERIMENTAL SECTION 2.1. Materials and Syntheses. Nondoped MPOX (M = Ca, Sr, or Ba; X = F, Cl, or Br) materials were prepared by the conventional high-temperature solid-state reaction. The constituent raw materials SrCO3 (A.R.), BaCO3 (A.R.), CaCO3 (A.R.), NH4H2PO4 (A.R.), and MX2·xH2O were weighed out in stoichiometric proportions. The halogen source was added to a different excess ratio. To obtain wellcrystallized phosphor compounds, the two-step heating process was adopted. In the first step, the starting chemicals were thoroughly mixed and preheated at 500 °C for 4 h in air to ensure the completed decomposition of NH4H2PO4. Then, in the second step, after regrounding, the as-obtained precursor was heated in a 10% H2/90% N2 reducing atmosphere at 1050 °C for 5 h. For comparison, an air or nitrogen flow atmosphere was also employed in the second stage. Furthermore, a two-stage surface annealing treatment at 1050 °C was performed for Ba5−mSrm(PO4)3Br (BSPOB; m = 0−5). The assynthesized materials suffered a further thermal treatment under atmospheric conditions other than that in the second step and for different duration times. After firing, the obtained samples were cooled to room temperature before further experiments. It has to be made clear that no impurity was introduced into the host matrix during the entire preparation process. 2.2. Characterizations. X-ray powder diffraction (XRD) patterns were collected with a model XRD-6100 Shimadzu X-ray diffractometer with Cu Kα radiation at 40 kV and 30 mA with a scan speed of 4 deg/min. Structural refinements of X-ray diffractograms were made using the GSAS (general structure analysis system) program. Photoluminescence (PL) and photoluminescence excitation (PLE) spectra of the phosphors were recorded on an EDINBURGH FLS920 combined Fluorescence Lifetime & Steady State Spectrometer with a 450 W xenon lamp. The PL quantum yields (QYs) of the phosphors were obtained directly using an absolute PL quantum yield measurement system (C9920-02, Hamamatsu Photonics K. K.). Diffuse reflectance spectra were recorded on a Hitachi U-3010 spectrometer equipped with an integration sphere attachment using BaSO4 as the background. Infrared spectroscopy was performed with a Fourier transform infrared (IR) spectrometer (AVATAR 370, Nicolet Co.). Raman spectra were measured with a DXR Raman microscope with a 780 nm laser device. The electron paramagnetic resonance (EPR) spectra were recorded on a JES FA200 spectrometer without irradiation excitation. All the measurements described above were taken at room temperature. 2.3. Theoretical Calculations. In this study, DFT calculations were performed on model systems Ba 5 (PO 4 ) 3 Br (BPOB), Ba3Sr2(PO4)3Br (BSPOB), and Sr5(PO4)3Br (SPOB) as well as their defect structures. Three different arrangements of O vacancies B

DOI: 10.1021/acs.inorgchem.8b02105 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Figure 1. (a) XRD patterns of Ba5−mSrm(PO4)3Br (m = 0−5) prepared under a H2/N2 reducing atmosphere and the standard profiles of BPOB (PDF 70-5069) and SPOB (PDF 89-5876). (b) Magnified XRD patterns in the 2θ region between 28.5° and 32.0° and the inserted (hkl) indicators (red dashes for those of BPOB and purple dashes for SPOB). on three crystallographic sites with different atomic coordinates were considered, where the arrangements were named VO(1)−VO(3). Meanwhile, cation vacancy VBa and/or VSr and bromine vacancy (VBr) defects were also simulated. The DFT computations were performed using a generalized gradient approximation (GGA) density functional by Perdew, Burke, and Ernzerhof (PBE), as implemented in the Vienna Ab initio Simulation Package (VASP). Only valence electrons of Sr(4s24p65s2), Ba(5s25p66s2), Br(4s24p5), P(3s23p3), and O(2s22p4) were treated quantum mechanically. Their interactions with the respective cores were described by the projected augmented wave (PAW) method. All three compounds crystallize in the hexagonal P63/m phase, which contains 42 atoms in one primitive cell. For the nondoped cases, there were a total of 288 valence electrons. Their geometries were fully optimized until the total energies and the Hellmann−Feynman forces on the atoms converged to 10−4 eV and 0.01 eV Å−1, respectively. A 12 × 12 × 6 Monkhorst Pack-type k point system was used to sample the Brillouin zone, with the cutoff energy for the plane wave basis set to 520 eV. A 2 × 2 × 2 supercell containing 336 atoms was obtained from the optimized structures, and then we obtained the defect structures from these perfect supercells employed for the calculations of intrinsic defects. One k point Γ was used for the calculations of defect structures. The cutoff energy for the plane wave basis was set to 520 eV, and the PREC flag was set to Accurate. For defect D in charge state q, the standard formalism of the formation energy (ΔEf) is defined as follows33,34

performed using the GGA-PBE functional with a double-zeta numerical basis set from the Material Studio (MS) CASTEP package.

3. RESULTS AND DISCUSSION 3.1. Structural Analysis and Phase Identification. The phase purity of the synthesized phosphors BSPOB was identified by XRD. As shown in Figure 1a, one can see that the end point compounds consist of pure BPOB and SPOB and the entire observed diffraction peaks match well with those of the standard data cards of PDF 70-5069 and 89-5876, respectively. With increased m values, all diffraction peaks of the as-prepared BSPOB phosphors were shifted to higher angles because of the substitution of smaller Sr2+ ions for big Ba2+ ones, which implies the formation of a series of continuous solid solutions. A further support is shown in the partly magnified XRD patterns in Figure 1b. In addition, no trace amount of impurity is determined. The strong and sharp diffraction peaks shown in the Figure also indicate the high crystallization of the phosphors. Rietveld refinement based on the GSAS program was performed. The crystal structure of Sr2.54Ba2.45Eu0.01(PO4)3Cl (ICSD 83254) was taken as the starting model. The distributions of Sr and Ba atoms in Ba3Sr2(PO4)3Br and Ba2Sr3(PO4)3Br were simulated. As the chloro-apatite, Sr5(PO4)3Br and Ba5(PO4)3Br are isostructural with each other. Their crystal structures are shown in Figure S1. There are two crystallographically distinct Sr or Ba atoms. Sr(1)/ Ba(1) atoms are coordinated with nine O atoms to form an irregular Sr(1)/Ba(1)O9 polyhedron, and Sr(2)/Ba(2) atoms are coordinated with six O and two Br atoms to exhibit a distorted Sr(2)/Ba(2)O6Br2 polyhedron. As shown in Table S2, the average bond length of the former is shorter than that of the latter. Because the ionic radius of Sr2+ (CN = 9, 1.31 Å; CN = 8, 1.26 Å) is smaller than that of Ba2+ (CN = 9, 1.47 Å; CN = 8, 1.42 Å), it is expected that when Sr2+ is introduced into the Ba5(PO4)3Br matrix, Sr2+ ions tend to occupy the smaller Sr(1)/Ba(1) sites rather than the other sites. Furthermore, the atom site symmetry multiplicities of Sr(1)/ Ba(1) and Sr(2)/Ba(2) are 4 and 6, respectively, and different possible occupancies of Sr 2+ in Ba 3 Sr 2 (PO 4 ) 3 Br and Ba2Sr3(PO4)3Br were examined. For example, Ba3Sr2(PO4)3Br, namely Ba6Sr4(PO4)6Br2, prefers only Sr(1)/Ba(1) sites, so only Sr(2)/Ba(2) or a spontaneous distribution of Sr(1)/ Ba(1) and Sr(2)/Ba(2) was simulated. The total site occupancy of the solid solution crystals at the selected

ΔEf [Dq ] = Etot[Dq ] − Etot[perfect] +

∑ ΔnA μA A

+ q(εVBM + E F)

(1)

where Etot[Dq] and Etot[perfect] are the total energies of the defective and perfect supercells, respectively, ΔnA is the number of species A (Ba, Sr, O, or Br) removed from the perfect supercell to introduce point defects, and μA is the corresponding atomic chemical potential. To obtain atomic chemical potential μA, we made a 20 Å × 20 Å × 20 Å supercell of atom A. EF is the Fermi level measured from the valence band maximum (VBM) (εVBM), which is aligned with that of the perfect system by the macroscopic averaging approach. For a neutral defect (q = 0), the last item in eq 1 is omitted. The values of μA can be determined under thermodynamic equilibrium conditions of various phases containing the corresponding atomic species and depending on the surface-annealing conditions.34 Experimental atmospheres of air, N2, and H2/N2 flow were simulated by oxygen-rich (O → ∞), oxygen-poor (O → 0), and hydrogen-rich (H → ∞) conditions, respectively. Computational details about the μA calculations and the associated experimental details are shown in the S1 paragraphs of the Supporting Information. To appreciate the electronic structure of the phosphors studied in this work, as a comparison, first-principles DFT calculations were also C

DOI: 10.1021/acs.inorgchem.8b02105 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Figure 2. Rietveld refinement of XRD patterns of (a) Ba3Sr2(PO4)3Br and (b) Ba2Sr3(PO4)3Br: experimental data (crosses), calculated patterns (red line), differences between the observed and calculated values (blue line), and positions of Bragg reflections (magenta sticks).

composition was set as nSr/nBa = 4/6 or 6/4 for m = 2 or 3, respectively. However, only the as-obtained converged refinement parameters in the hypothesis of a preference of Sr2+ for Sr(1)/Ba(1) sites presented a reasonable total atom occupancy as Sr4Ba6P6O24Br2 or Sr6Ba4P6O24Br2. Therefore, just like that in Sr2.54Ba2.45Eu0.01(PO4)3Cl, in the bromo-apatite solid solution crystalline structure, we speculate that Sr2+ ions prefer to occupy the Sr(1)/Ba(1)O9 site rather than the Sr(1)/ Ba(1)O9 and Sr(2)/Ba(2)O6Br2 sites. The observed, calculated, difference, and Bragg positions as well as the final refined residual factors are summarized in Figure 2 and Table 1. All the refined crystallographic

components as shown in Figure S2. No luminescence was observed for the powders sintered in air or N2 flow. The atmosphere condition-dependent luminescence suggests that native defects, especially anion vacancies, should be responsible for the charge transfer resulting in different color emissions, which will be shown more in detail below. To minimize the volatilizations of alkaline earth metal and halogen atoms, BSPOB solid solutions with relatively heavy M and X atoms were specially chosen and quantitatively analyzed in detail. Figure 3a presents the normalized PL and PLE spectra of the phosphors prepared under the H2/N2 reducing atmosphere. It is well-known that a series of continuous solid solutions show the same structural type, and therefore, the photoluminescence spectra generally possess features that can be superimposed along with a smooth variation of structural parameters with respect to composition variation. It can be observed that all samples exhibit broad emission bands from 350 to 800 nm under excitation by high-energy ultraviolet (UV) light from 215 to 330 nm. A nearly linear dependence of emission wavelengths on m values is illustrated in Figure 3b along with the similar progressive blue shift of the excitation band. With m increasing from 0 to 5, the change in color from orange to yellow, then to green, and finally to cyan is clearly depicted in a Commission Internationale de L’Eclairage (CIE) chromaticity diagram. The digital photograph of the sample exposed to 254 nm irradiation provides a further verification of the gradual emission shift as shown in the inset of Figure 3d. The fluorescence decay curves of BSPOB under 254 nm excitation are depicted in Figure 3c, all of which fit well with the first-order exponential decay model of eq 2:

Table 1. Main Processing and Refinement Parameters of Ba3Sr2(PO4)3Br and Ba2Sr3(PO4)3Br from the Rietveld Structure Analysis space group symmetry a = b (Å) c (Å) V (Å3) A (deg) B (deg) Γ (deg) Z χ2 Rwp (%) Rp (%) RF (%)

Ba3Sr2(PO4)3Br

Ba2Sr3(PO4)3Br

P63/m Hexagonal 10.15915(6) 7.52193(6) 672.318(10) 90 90 120 2 1.242 6.78 5.13 2.65

P63/m hexagonal 10.08969(7) 7.42413(7) 654.533(11) 90 90 120 2 1.768 8.28 6.29 2.94

I(t ) = I0 + A expt / τ

parameters well satisfy the reflection conditions, and good fits were obtained with χ2 values of 1.242 and 1.768, which can validate the phase purity of the as-prepared samples. The obtained cell parameters of the median solid solutions are listed in Table 1. Both Ba3Sr2(PO4)3Br and Ba2Sr3(PO4)3Br remain isostructural with the apatite BPOB or SPOB. The reasonable lattice constants present a visual sense of the formation of the solution crystals. 3.2. Self-Activated Photoluminescence Properties. A series of undoped MPOX compounds were prepared under air, N2, or H2/N2 flow atmosphere conditions. Only those synthesized in a H2-containing reducing flow show controlled color-tunable emissions depending on the M and X

(2)

where I and I0 are the luminescence intensity at time t and 0, respectively, A is a constant, t is the time, and τ is the decay time for an exponential component. The determined lifetime values are 10.98, 11.60, 11.59, 11.22, 9.50, and 6.73 μs for m = 0−5, respectively. It is noticeable that the luminescence decay times are quite different from the common broad emission due to impurity activator Ce3+ (nanoseconds), Eu2+ (≈1 μs), or Mn2+ (milliseconds). It provides a solid support to our hypothesis that the reducing atmosphere-dependent emission from BSPOB should be ascribed to intrinsic defect-induced self-activated luminescence rather than impurity defect-related transitions. D

DOI: 10.1021/acs.inorgchem.8b02105 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Figure 3. Luminescence properties of nondoped Ba5−mSrm(PO4)3Br (BSPOB; m = 0−5). (a) Normalized PLE and PL spectra. (b) Dependence of excitation (λex) and emission (λem) wavelengths and emission intensity on Sr content (m). (c) Luminescence decay curves. (d) CIE chromaticity coordinate diagram. The inset is a digital photograph of the phosphors being exposed to 254 nm irradiation (m increasing from right to left).

phenomenon in which Ca2Ba3(PO4)3Cl host material showed bright cyan emission centered near 480 nm after treatment under a H2-containing reducing atmosphere. They also considered that the possible luminescence mechanism was attributed to the anion-defective cluster A-O and A-Cl (A = Ca/Ba) groups in the host lattice.24 However, Arunkumar et al. proposed that the Ca4F2Si2O7 host synthesized under reduction conditions exhibited violet emission due to the formation of fluorine-deficient nonstoichiometric Ca4F2−δSi2O7+δ/2 species.18 Although the exact mechanism is still unclear, consensus that the anion deficiency as a result of H2 absorbing from the surface is the key to the self-activated luminescence has been reached. The associated reactions can be described as follows.

In addition, as depicted in Figure 3b, the composition of the solid solutions not only affects the PL and PLE wavelength but also changes the luminescence intensity. Under the excitation of their respective maximal wavelength, the solid solution BSPOB, especially around the middle cationic ratio, presents luminescence that is stronger than that of the end point compounds. The relative emission intensity of Ba2Sr3(PO4)3Br (m = 3) is almost 2 times higher than that of SPOB (m = 5) and >9 times higher than that of BPOB (m = 0). We believe that the intensified disturbed lattice structure generated by multicationic disordered occupancies should be the primary cause of the stronger self-activated luminescence in the solid solution crystal. A further illustration is given by the variation of the emission colors of BSPOB phosphors with a fixed Eu2+ doping concentration. As usual, an increased level of substitution for the larger Ba2+ sites with smaller Sr2+ cations leads to an enlarged crystal splitting of 5d energy levels. A gradual red shift of the emission wavelength from 4f−5d transitions of Eu2+ is expected. However, as shown in Figure S3, in the series of BSPOB:0.03Eu2+ species synthesized under a H2/N2 flow, an integral spectral red shift first occurs and then a blue shift follows along with the increased m value, although a continuous red shift of the Eu2+ blue emission can be detected. Accordingly, the unexpected saltation of integral emission color is determined by the enhancement of selfactivated emission from the host in a lower-energy region in the intermediate solid solution phosphors. 3.3. Intrinsic Defect Formation Energies in BSPOB and the Electronic Structure of BSPOB. Several groups have reported on the emission of undoped oxysalt phosphors. Park and Vogt suggested that the anion-deficient nonstoichiometry in (Sr3−xAx)1−α−2δMO4−αF1−δ (A = Ca or Ba; M = Ga or Al) was the origin of broad band photoluminescence when the oxyfluoride compounds were exposed to reducing gases. 16 Shang et al. reported a similar

× × H 2 + OO → VO + H 2O

(3)

H 2 + 2X×X → 2V×X + 2HX (X represents a halogen atom) (4)

To investigate the intrinsic defect responsible for selfactivated emission in halo-apatite compounds, DFT calculations have been performed for model systems of BPOB, BSPOB, and SPOB. A single defect formation energy was calculated by eq 1. Note that a neutral oxygen vacancy defect, labeled as V0O or V×O, is formed without any electrons being lost spontaneously. Here, only neutral vacancy defects are taken into consideration, so that the charge states of the defects are omitted. Meanwhile, bromine vacancy VBr defects as well as cation vacancies VBa and VSr are also simulated. The single defect formation energy and its dependence on the atmospheric condition are explored in the BPOB system. Air, N2, and H2/N2 conditions were simulated. As displayed in Table S3, the Br vacancy defect is easiest to form under both O-rich and O-poor conditions, where no self-activated luminescence is detected in such situations. However, the E

DOI: 10.1021/acs.inorgchem.8b02105 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

3.4. Confirmation of Oxygen Vacancy by EPR. To confirm the proposition that the existence of oxygen vacancies is correlated with luminescence, a one- or two-stage surface annealing treatment was performed on Ba2Sr3(PO4)3Br powder at different oxygen partial pressures. As shown in Table 3, for the one-stage process (samples A, E, and F), only

reasonable formation energy indicates easy formation of the O vacancy defect despite the occupancy sites under a H2-rich atmosphere. As shown in Table 2, under a reducing Table 2. Defect Formation Energies of Single Neutral Atom Vacancy Defects in a Ba5−mSrm(PO4)3Br Solid under a H2Rich Atmosphere

Table 3. Ba2Sr3(PO4)3Br Samples Synthesized by One- or Two-Stage Surface Annealing Treatments and Their Corresponding Luminescence Characteristics

ΔEf (eV) single vacancy defect

m=0

m=2

m=5

VSr(1)/Ba(1) VSr(2)/Ba(2) VBr VO(1) VO(2) VO(3)

11.2648 11.4602 1.8379 −2.1632 −2.2563 −2.3583

12.0882 11.9913 1.8303 −2.1935 −2.2435 −2.4435

12.0677 12.0695 1.7438 −2.0875 −2.1316 −2.2941

synthetic atmospheric condition stage one

sample atmosphere A B C D E F

atmosphere, VO defects are dominant with respect to VBr or cation vacancies in the BSPOB solid solution series with different m values. Therefore, it can be inferred that the oxygen vacancy should be primarily responsible for the self-activated emissions of BSPOB compounds under the reducing atmospheric conditions. Further support can be provided by the projected density of states (DOS) of defective supercells. As shown in Figure 4,

H2/N2 N2 air H2/N2 N2 air

stage two time (h) 5 3 3 5 5 5

atmosphere

time (h)

emit (√) or not (×)

EPR response

− H2/N2 H2/N2 air − −

− 2 2 2 − −

√ √ √ × × ×

g = 1.999 g = 2.000 g = 2.001 − − −

those synthesized under a reducing atmosphere (A) present luminescence characteristics, which disappeared after a twostage treatment in air (D). It is possible that there was a passivation effect resulting from the repadding of VO defects in sample D by oxygen in air during the heat treatment.36 However, the host materials, which possess no self-activated luminescence first synthesized in N2 flow (E) or air (F), shine brightly after being reannealed under a H2/N2 reducing atmosphere, i.e., B and C. Hence, no matter how many stages of the treatment were employed, the process that ended up with the H2-containing atmosphere always led to the defectrelated luminescence. This result further proves that the removal of oxygen atoms under the reducing condition is the key to achieving defect-related luminescence in such haloapatite compounds.36,37 As we all know, EPR is a powerful technology for characterizing defects. EPR of representative Ba2Sr3(PO4)3Br samples was performed. As shown in Figure 5a, for samples A− C with emitting features, one can observe a g ≈ 2.000 signal, which corresponds to oxygen vacancy V+O or V·O.36 On the other hand, an almost undetectable response could be recognized in nonluminous samples D−F. The PL spectra of all six samples are shown in Figure S4. Bright greenish-yellow emissions are observed in samples A−C, while no luminescence is detected in samples D−F. Therefore, luminescence is closely related to oxygen vacancy defects in the samples. In addition, the influence of different surface annealing treatments on optical properties was investigated. As shown in Figure S4, the PL spectral profile of the self-activated luminescence in Ba2Sr3(PO4)3Br remains unchanged after different surface treatments, which is a further support for the same defect luminescence. Among the three kinds of host materials, as shown in Figure 5b, the one directly synthesized under H2/N2 flow has the highest QE, which reaches 93%, a surprising value for such a super-broad band emission. Unfortunately, QEs of those after a two-stage treatment drop to approximately >70%, which may be due to the interactions between multi-intrinsic defects formed on the crystal surfaces during the surface annealing treatment under different

Figure 4. Total density of states of Ba5(PO4)3Br. BPOB represents the perfect crystal, BPOB-VBr the defective supercell with neutral Br vacancy, and BPOB-VO the defective supercell with neutral O vacancy. The magenta dotted line aligned with the zero potential energy is labeled as the Fermi level. The conduction band minimum (CBM) and valence band maximum (VBM) are also marked.

although the overall DOS profiles for the defective BPOB are quite similar to those of the perfect one, VO or VBr induces defect states in the band gap. The defect level resulted from VBr located near the bottom of the CBM of BPOB. The small energy difference between VBr and CBM leads to strong coupling between CB and the bromine vacancy state, which implies a possible delocalized shallow level. In general, shallow donors prefer to enhance the electron carrier density in CB instead of serving as the luminescence center.10,30 On the other hand, the VO level in BPOB is more localized. Oxygen vacancies are known to be one of the most common defects in oxides and usually act as radiative centers in luminescence processes. Therefore, it can be believed that VO acts as an emissive recombination center in BSPOB halo-apatite compounds. F

DOI: 10.1021/acs.inorgchem.8b02105 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Figure 5. (a) EPR spectra and (b) comparison of the PL intensities and quantum efficiencies (QEs) of Ba2Sr3(PO4)3Br host materials prepared with different surface annealing treatments.

Table 4. PO43− Anion Polyhedron Structural Information for Ba5−mSrm(PO4)3Br Compounds bond length (Å) m

P−O(1)

P−O(1)

P−O(2)

P−O(3)

P−O (average)

D

0 2 3 5

1.55782 1.54881 1.50820 1.55455

1.55783 1.54881 1.50820 1.55454

1.56343 1.56135 1.54499 1.56137

1.56952 1.56630 1.49731 1.56651

1.56215 1.55632 1.52115 1.55924

0.002769 0.004824 0.012093 0.003013

Figure 6. (a) Infrared and (b) Raman spectra of Ba5−mSrm(PO4)3Br compounds.

According to the XRD refinement, the P−O bond lengths in the PO43− group are summarized in Table 4. Nonlinear changes in the average bond length are found with an increasing m value. A shorter average bond length in BSPOB than in the end point compounds is observed. Combined with the data in Table 2, for a certain m value, it is seen that the longer the P−Ox (x = 1, 2, or 3) bond length, the weaker the interaction between the P and Ox, leading to the lower VOx formation energy. However, considering the average bond length with different m values, it seems to be in violation with the variation of the defect concentration and PL intensity in solid solutions. As we know, multicationic disordered occupancies will lead to intensified disturbances of the crystal lattices. The polyhedral distortion index, D, can be used as a reference to evaluate the symmetry of the polyhedron. D can be calculated following39

atmospheres, as shown in Table S3. The effect of the concentration and distribution of VO in the host crystals on the defect-related optical properties is still unknown. 3.5. Effect of the Solid Solution Structure on the Formation of VO. As mentioned in Figure 3b, the intermediate solid solution phosphors BSPOB show defect luminescence that is more intense than that of the end point compounds BPOB and SPOB. On the basis of the discussions presented above, it can be deduced that the higher VO defect concentration in BSPOB is responsible for the higher PL intensity. The new question is what the relationship between the solid solution crystal structure and the formation of VO should be. The BSPOB crystal is characterized by strong covalent bonds within the PO4 tetrahedra and weak ionic bonding of an alkaline earth cation to O and Br.38 Because of the longer Ba/ Sr−O bond and smaller overlap between their orbitals, we assume that we should pay special attention to the bond length and interaction between P and O in the apatite structures.

D= G

1 n

n

∑ i=1

|li − lav| lav

(5) DOI: 10.1021/acs.inorgchem.8b02105 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry where li is the distance from the central atom to the ith coordinating atom and lav is the average bond length. Our results (Table 4) reveal a shaper increase in the distortion of PO43− tetrahedra in a BSPOB solid solution with m = 2 or 3 than in the single Ba- or Sr-containing BPOB and SPOB, which implies a lower symmetry of PO4 groups from pure Ba or Sr crystals than from solid solution structures. The symmetry of the PO4 group in the solid solution compound is reduced because of the perturbation of the crystal structure of the solid solution. The lower symmetry of the tetrahedra leads to an easier escape of oxygen atoms from PO4 groups. As a result, a more intensive luminescence was detected in the solid solution phosphors than in the end point compounds. The variation tendency of D values is consistent with the PL intensity evolution of the BSPOB host materials with a corresponding m value. Figure 6a provides the FT-IR spectra of BSPOB compounds. The P−O stretching vibrations (in the region between 900 and 1100 cm−1) appear with three well-resolved bands for all compositions. These are assigned to symmetric P−O stretching vibration ν1 (lying between 931 and 946 cm−1) and the triply degenerate ν3 (lying between 970 and 1100 cm−1) antisymmetric P−O stretching vibration of the PO43− groups.40 The strong bands at 1004 and 1041 cm−1 in BPOB moved to 1027 and 1064 cm−1 in SPOB, respectively. On one hand, the entire absorption band maxima red shift toward a higher wavenumber moving from Ba to Sr in the BSPOB compounds, which is attributed to the decrease in the unit cell volumes.35 A smooth shift with increasing m values in Ba5−mSrm(PO4)3Br (m = 0−5) verifies a characteristic of these solid solutions. The continuous shift was also found in the Raman spectra as shown in Figure 6b. On the other hand, asynchronous variation between ν1 and ν3 modes is observed. The ν1 symmetry stretching vibration of PO43− groups in BSPOB solid solutions is stronger than in BPOB and SPOB, while the ν3 antisymmetry stretching vibration is reversed. This result suggests a lower symmetry of the PO4 group in the former than in the latter, in line with the inference based on deformation index D. Moreover, a decrease in the P−O average bond length in BSPOB coincides with the stronger bands in the solid solutions than the end point compounds at around 931−946 cm−1 associated with symmetry stretching vibrations. Consequently, the correlation between the VO defect formation energy and the defect concentration and crystal structure mainly depends on the P−O interaction and PO43− group symmetry. Generally, a longer and looser P−O bond is consistent with a lower VO formation energy. In addition, the decrease in the tetrahedral symmetry of the PO43− anionic group, associated with crystal structure perturbations, leads to an increase in the concentration of VO defects and also a higher defect-related luminescence intensity. It is the first time that the possible luminescence mechanism has been deduced from oxygen vacancies based on the valence bond theory and the novel possibility of investigating the correlation between radiation transitions and lattice structures has been provided. 3.6. Possible Luminescence Mechanism. For a neutral V0O in the BSPOB system, two electrons are localized in the O vacancy maintaining the charge that was at that site in the pure crystal. According to our EPR characterization, the oxygen vacancy at a single positively charged state, V+O, was confirmed. The defect centers become stable through the redistribution of the electrons bound to the original oxygen atom. One of the

two electrons is transferred to a nearby P atom, resulting in the formation of negatively charged PO32− reduced from PO43−.41 For the +1 charge oxygen vacancy, the removal of an electron reduces the occupancy of the filled gap states in the neutral case from two electrons to one electron and produces a new empty state in the gap. V0O plays a role as the donor, and V+O as the acceptor. As a result, if the host material is exposed to irradiation, the defect is transferred between the neutral charge state and the +1 charge state, which is thought to be responsible for the photoemission. Because of the geometry relaxation around the vacancy site, the defective lattice can bear a long excited-state lifetime, which is in accordance with the magnitude of the luminescence decay estimated in Figure 3c. Although a more accurate determination should require excited-state calculations, it has been commonly accepted that the location of the V0/+ O charge transition level (CTL) below the CBM in the band gap determines the transition energy.33 The optical transition energy can be calculated by using the total energies of the defective supercells in the 0 and +1 charge states. Because of the approximate hypotheses of a rigid band structure and a similar defect formation energy with changing m values, the CTL should be confined to a rather narrow energy range in the series BSPOB wide band gap compounds if the VBM is aligned.30 Therefore, the energy difference depends on the location of the CBM, in other words, the band gap width with VBM aligned to E = 0 eV. Then, tunable emissions could be achieved by selectively controlling the Sr/ Ba solid solution ratio. To appreciate the electronic structure of the phosphors studied in this work, first-principles DFT calculations were performed for BPOB, BSPOB (m = 2), and SPOB using the GGA-PBE density functional. The computationally optimized crystal structure yielded lattice constants of a = 10.15920 Å and c = 7.52190 Å for Ba3Sr2(PO4)3Br, which are consistent with those determined by XRD analyses. As shown in Figures S5 and S6, the results suggest an indirect band gap of 4.824 eV between the VBM at the K point and the CBM at the G point. The CBM consists of the Ba 5d orbital, and the O 2p orbital constitutes the VBM. In addition, the s−p hybrid orbitals also contribute to the CB. Likewise, the magnitude of the indirect band gap is 5.040 eV for SPOB. The computational band gap width is in reasonably good agreement with experimental data determined by spectroscopic measurements (Figure S7). We can believe that the band gap of BSPOB widens with the increased Sr content, which is consistent with the gradual blue shift of the self-activated photoemissions. The proposed luminescence mechanism is depicted in Figure 7.

Figure 7. Schematic diagram of the VO defect-induced luminescence mechanism and the spectral shift-dependent band gap width. H

DOI: 10.1021/acs.inorgchem.8b02105 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry ORCID

4. CONCLUSIONS Apatite structure-type M5(PO4)3X (MPOX; M = Ca, Sr, or Ba; X = F, Cl, or Br) systems, especially the middle solid solution ratio adjacent compounds, demonstrate intensive self-activated luminescence under UV excitation. Gradual spectral shifts from blue to orange are observed with continuous variation of the composition. In this work, the defect formation process and defect-related luminescence mechanism have been investigated in combination with experimental characterization and theoretical calculations. With relatively low volatilities, Ba5−mSrm(PO4)3Br (BSPOB; m = 0−5) species are specially chosen and quantitatively analyzed in detail. Their continuous solid solution compositions and crystal structures are affirmed by XRD and GASA refinement. According to the defect formation energy as well as EPR technology, an oxygen vacancy with bound electrons should be mainly responsible for their broad band emissions. This hypothesis has further been supported by the two-stage surface annealing treatment under different atmospheres. The localization of the V0/+ O defect level contributes to the recombination of electron−hole pairs and thus leads to visible light emission. The evolution of compositions and solid solution ratios results in the change in the band gaps, making it possible to finely tune emissions with different energies. Close to the middle solid solution ratio, the randomness of the cations and the disturbance of the crystal structure possibly improve the intrinsic defect concentrations and dramatically boost the defect-related luminescence more than for the end point compounds. As a result, the quantum efficiency of yellowish green emitting Ba2Sr3(PO4)3Br under UV irradiation reaches 93%. The native defect-induced tunable emission from halo-apatite solid solutions reported in this work should provide a novel strategy for controlling the optical properties for both self-activated and impurity-activated phosphors.



Zhongxian Qiu: 0000-0002-3277-9245 Shubin Liu: 0000-0001-9331-0427 Jilin Zhang: 0000-0001-7235-341X Wenli Zhou: 0000-0002-6975-2206 Shixun Lian: 0000-0001-6524-2703 Author Contributions ⊥

Jing Zhang and Tingting Zhang contributed equally to this work. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work is partially supported by the National Natural Science Foundation of China (Grants 21805082, 21571059, 21501058, 21471055, and 51402105), the Hunan Provincial Natural Science Foundation of China (Grant 2017JJ3200), and the National Key Research and Development Program (2016YFB0302403).



ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.inorgchem.8b02105. Different atomic chemical potentials (μA) depending on the surface annealing conditions (Table S1), crystal structures of Ba 5 (PO 4 ) 3 Br, Ba 3 Sr 2 (PO 4 ) 3 Br, and Sr5(PO4)3Br (Figure S1), average M−L (M = Sr or Ba; L = O or Br) bond lengths in BSPOB crystals (Table S2), PLE and PL spectra of MPOX phosphors (Figure S2), PL spectra of BSPOB:0.03Eu2+ phosphors (Figure S3), defect formation energies of single atom vacancy defects in BPOB under different atmospheres (Table S3), comparison of PL spectra of Ba2Sr3(PO4)3Br with different surface annealing treatments (Figure S4), total and partial densities of states of BPOB (Figure S5), band structures of BPOB and SPOB (Figure S6), and diffuse reflection spectra of BSPOB and the calculated Eg (Figure S7) (PDF)



REFERENCES

(1) Reineke, S. Complementary LED technologies. Nat. Mater. 2015, 14, 459. (2) Lin, C. C.; Liu, R.-S. Advances in phosphors for light-emitting diodes. J. Phys. Chem. Lett. 2011, 2 (11), 1268−1277. (3) Kim, K.-B.; Kim, Y.-I.; Chun, H.-G.; Cho, T.-Y.; Jung, J.-S.; Kang, J.-G. Structural and Optical Properties of BaMgAl10O17:Eu2+ Phosphor. Chem. Mater. 2002, 14 (12), 5045−5052. (4) Li, S.; Wang, L.; Tang, D.; Cho, Y.; Liu, X.; Zhou, X.; Lu, L.; Zhang, L.; Takeda, T.; Hirosaki, N.; Xie, R.-J. Achieving high quantum efficiency narrow-band β-Sialon:Eu2+ phosphors for highbrightness LCD backlights by reducing the Eu3+ luminescence killer. Chem. Mater. 2018, 30 (2), 494−505. (5) Xie, R. J.; Hirosaki, N. Silicon-based oxynitride and nitride phosphors for white LEDsA review. Sci. Technol. Adv. Mater. 2007, 8 (7−8), 588−600. (6) Apple, E. F. Observations on the Formation of Color Centers in Calcium Halophosphates. J. Electrochem. Soc. 1963, 110 (5), 374− 380. (7) Blasse, G.; Brixner, L. H. Ultraviolet emission from ABO4-type niobates, tantalates and tungstates. Chem. Phys. Lett. 1990, 173 (5-6), 409−411. (8) Liu, X.; Chen, C.; Li, S.; Dai, Y.; Guo, H.; Tang, X.; Xie, Y.; Yan, L. Host-Sensitized and Tunable Luminescence of GdNbO4:Ln3+ (Ln3+ = Eu3+/Tb3+/Tm3+) Nanocrystalline Phosphors with Abundant Color. Inorg. Chem. 2016, 55 (20), 10383−10396. (9) Bharat, L. K.; Jeon, S.-K.; Krishna, K. G.; Yu, J. S. Rare-earth free self-luminescent Ca2KZn2(VO4)3 phosphors for intense white lightemitting diodes. Sci. Rep. 2017, 7, 42348. (10) Alkauskas, A.; McCluskey, M. D.; Van de Walle, C. G. Tutorial: Defects in semiconductorsCombining experiment and theory. J. Appl. Phys. 2016, 119 (18), 181101. (11) Ruane, W. T.; Johansen, K. M.; Leedy, K. D.; Look, D. C.; von Wenckstern, H.; Grundmann, M.; Farlow, G. C.; Brillson, L. J. Defect segregation and optical emission in ZnO nano- and microwires. Nanoscale 2016, 8 (14), 7631−7637. (12) Pu, C.; Qin, H.; Gao, Y.; Zhou, J.; Wang, P.; Peng, X. Synthetic Control of Exciton Behavior in Colloidal Quantum Dots. J. Am. Chem. Soc. 2017, 139 (9), 3302−3311. (13) Mattila, T.; Zunger, A. Deep electronic gap levels induced by isovalent P and As impurities in GaN. Phys. Rev. B: Condens. Matter Mater. Phys. 1998, 58 (3), 1367−1373. (14) Ding, J.; Wu, Q.; Li, Y.; Long, Q.; Wang, Y.; Ma, X.; Wang, Y. Self-Activated Yellow Light Emitting Phosphors of α, β-Ca3B2N4 with Long Afterglow Properties. Inorg. Chem. 2016, 55 (21), 10990− 10998.

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. *E-mail: [email protected]. I

DOI: 10.1021/acs.inorgchem.8b02105 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

accurate descriptors of oxygen vacancy formation energetics. Energy Environ. Sci. 2014, 7 (6), 1996−2004. (33) Ning, L.; Cheng, W.; Zhou, C.; Duan, C.; Zhang, Y. Energetic, Optical, and Electronic Properties of Intrinsic Electron-Trapping Defects in YAlO3: A Hybrid DFT Study. J. Phys. Chem. C 2014, 118 (34), 19940−19947. (34) Zhang, S. B.; Northrup, J. E. Chemical potential dependence of defect formation energies in GaAs: Application to Ga self-diffusion. Phys. Rev. Lett. 1991, 67 (17), 2339−2342. (35) Kim, D.; Kim, S. C.; Bae, J. S.; Kim, S.; Kim, S. J.; Park, J. C. Eu2+-Activated Alkaline-Earth Halophosphates, M5(PO4)3X:Eu2+ (M = Ca, Sr, Ba; X = F, Cl, Br) for NUV-LEDs: Site-Selective Crystal Field Effect. Inorg. Chem. 2016, 55 (17), 8359−8370. (36) Drouilly, C.; Krafft, J.-M.; Averseng, F.; Casale, S.; Bazer-Bachi, D.; Chizallet, C.; Lecocq, V.; Vezin, H.; Lauron-Pernot, H.; Costentin, G. ZnO Oxygen Vacancies Formation and Filling Followed by in Situ Photoluminescence and in Situ EPR. J. Phys. Chem. C 2012, 116 (40), 21297−21307. (37) Wang, T.; Radovanovic, P. V. In situ enhancement of the blue photoluminescence of colloidal Ga2O3 nanocrystals by promotion of defect formation in reducing conditions. Chem. Commun. 2011, 47 (25), 7161−7163. (38) Junhui, Z.; Yonghua, D.; Lishi, M.; Runyue, L. Structural properties, electronic structure and bonding of Ba10(PO4)6X2 (X = F, Cl and Br). J. Alloys Compd. 2016, 680, 121−128. (39) Baur, W. H. The geometry of polyhedral distortions. Predictive relationships for the phosphate group. Acta Crystallogr., Sect. B: Struct. Crystallogr. Cryst. Chem. 1974, 30, 1195−1215. (40) Nakamoto, K. Infrared and Raman Spectra of Inorganic and Coordination Compounds, Theory and Applications in Inorganic Chemistry, 6th ed.; John Wiley & Sons: Hoboken, NJ, 2009; pp 180−184. (41) Liu, C. S.; Hou, C. J.; Kioussis, N.; Demos, S. G.; Radousky, H. B. Electronic structure calculations of an oxygen vacancy in KH2PO4. Phys. Rev. B: Condens. Matter Mater. Phys. 2005, 72 (13), 134110.

(15) Xue, F.; Hu, Y.; Fan, L.; Ju, G.; Lin, X.; Lv, Y.; Mu, Z. Reversible white-brown photochromism in a self-activated longpersistent phosphor Mg2SnO4. Opt. Mater. Express 2017, 7 (3), 1014−1021. (16) Park, S.; Vogt, T. Defect Monitoring and Substitutions in Sr3−xAxAlO4F (A = Ca, Ba) Oxyfluoride Host Lattices and Phosphors. J. Phys. Chem. C 2010, 114 (26), 11576−11583. (17) Oskam, K. D.; Kaspers, K. A.; Meijerink, A.; Muller-Bunz, H.; Schleid, T. Luminescence of La3F3[Si3O9]:Ce3+. J. Lumin. 2002, 99, 101−105. (18) Arunkumar, P.; Kim, Y. H.; Im, W. B. Versatile Ca4F2Si2O7 Host from Defect-Induced Host Emission to White-Light-Emitting Ce3+-Doped Ca4F2Si2O7 Phosphor for Near-UV Solid-State Lighting. J. Phys. Chem. C 2016, 120 (8), 4495−4503. (19) Gupta, S. K.; Ghosh, P. S.; Yadav, A. K.; Jha, S. N.; Bhattacharyya, D.; Kadam, R. M. Origin of Blue-Green Emission in αZn2P2O7 and Local Structure of Ln3+ Ion in α-Zn2P2O7:Ln3+ (Ln = Sm, Eu): Time-Resolved Photoluminescence, EXAFS, and DFT Measurements. Inorg. Chem. 2017, 56 (1), 167−178. (20) Zhang, C.; Lin, C.; Li, C.; Quan, Z.; Liu, X.; Lin, J. Enhanced Luminescence of BPO4 by Mixing with SiO2 and Al2O3. J. Phys. Chem. C 2008, 112 (6), 2183−2192. (21) Zhang, C.; Cheng, Z.; Yang, P.; Xu, Z.; Peng, C.; Li, G.; Lin, J. Architectures of Strontium Hydroxyapatite Microspheres: Solvothermal Synthesis and Luminescence Properties. Langmuir 2009, 25 (23), 13591−13598. (22) Zhang, C.; Yang, J.; Quan, Z.; Yang, P.; Li, C.; Hou, Z.; Lin, J. Hydroxyapatite Nano- and Microcrystals with Multiform Morphologies: Controllable Synthesis and Luminescence Properties. Cryst. Growth Des. 2009, 9 (6), 2725−2733. (23) Zhang, C.; Huang, S.; Yang, D.; Kang, X.; Shang, M.; Peng, C.; Lin, J. Tunable luminescence in Ce3+, Mn2+-codoped calcium fluorapatite through combining emissions and modulation of excitation: a novel strategy to white light emission. J. Mater. Chem. 2010, 20 (32), 6674−6680. (24) Shang, M.; Geng, D.; Yang, D.; Kang, X.; Zhang, Y.; Lin, J. Luminescence and energy transfer properties of Ca2Ba3(PO4)3Cl and Ca2Ba3(PO4)3Cl:A (A = Eu2+/Ce3+/Dy3+/Tb3+) under UV and lowvoltage electron beam excitation. Inorg. Chem. 2013, 52 (6), 3102− 3112. (25) Zhang, C.; Lin, J. Defect-related luminescent materials: synthesis, emission properties and applications. Chem. Soc. Rev. 2012, 41 (23), 7938−7961. (26) Nandy, S.; Thapa, R.; Kumar, M.; Som, T.; Bundaleski, N.; Teodoro, O. M. N. D.; Martins, R.; Fortunato, E. Efficient Field Emission from Vertically Aligned Cu2O1‑δ(111) Nanostructure Influenced by Oxygen Vacancy. Adv. Funct. Mater. 2015, 25 (6), 947−956. (27) Tan, H.; Zhao, Z.; Zhu, W.-b.; Coker, E. N.; Li, B.; Zheng, M.; Yu, W.; Fan, H.; Sun, Z. Oxygen Vacancy Enhanced Photocatalytic Activity of Pervoskite SrTiO3. ACS Appl. Mater. Interfaces 2014, 6 (21), 19184−19190. (28) Jiang, X.; Zhang, Y.; Jiang, J.; Rong, Y.; Wang, Y.; Wu, Y.; Pan, C. Characterization of Oxygen Vacancy Associates within Hydrogenated TiO2: A Positron Annihilation Study. J. Phys. Chem. C 2012, 116 (42), 22619−22624. (29) Luna, L. E.; Gardner, D.; Radmilovic, V. R.; Maboudian, R.; Carraro, C. Atomic-Scale Electronic Characterization of Defects in Silicon Carbide Nanowires by Electron Energy Loss Spectroscopy. J. Phys. Chem. C 2018, 122 (22), 12047−12051. (30) Linderalv, C.; Lindman, A.; Erhart, P. A Unifying Perspective on Oxygen Vacancies in Wide Band Gap Oxides. J. Phys. Chem. Lett. 2018, 9 (1), 222−228. (31) Yin, W.-J.; Wei, S.-H.; Al-Jassim, M. M.; Yan, Y. Prediction of the chemical trends of oxygen vacancy levels in binary metal oxides. Appl. Phys. Lett. 2011, 99 (14), 142109. (32) Deml, A. M.; Stevanovic, V.; Muhich, C. L.; Musgrave, C. B.; O’Hayre, R. Oxide enthalpy of formation and band gap energy as J

DOI: 10.1021/acs.inorgchem.8b02105 Inorg. Chem. XXXX, XXX, XXX−XXX