Luminescence of Ce - ETIS

Luminescence of Ce - ETIS

Article pubs.acs.org/JPCC Luminescence of Ce3+-Doped MB2Si2O8 (M = Sr, Ba): A Deeper Insight into the Effects of Electronic Structure and Stokes Shift...

852KB Sizes 0 Downloads 3 Views

Article pubs.acs.org/JPCC

Luminescence of Ce3+-Doped MB2Si2O8 (M = Sr, Ba): A Deeper Insight into the Effects of Electronic Structure and Stokes Shift Qi Peng,† Chunmeng Liu,† Dejian Hou,† Weijie Zhou,† Chong-Geng Ma,*,‡ Guokui Liu,§ Mikhail G. Brik,‡,∥,⊥,# Ye Tao,¶ and Hongbin Liang*,† †

MOE Key Laboratory of Bioinorganic and Synthetic Chemistry, KLGHEI of Environment and Energy Chemistry, School of Chemistry and Chemical Engineering, Sun Yat-sen University, Guangzhou 510275, P. R. China ‡ College of Sciences, Chongqing University of Posts and Telecommunications, Chongqing 400065, P. R. China § Chemical Sciences and Engineering Division, Argonne National Laboratory, Argonne, Illinois 60439, United States ∥ Institute of Physics, University of Tartu, Ravila 14C, Tartu 50411, Estonia ⊥ Institute of Physics, Jan Dlugosz University, Armii Krajowej13/15, PL-42200 Czestochowa, Poland # Institute of Physics, Polish Academy of Sciences, al. Lotników 32/46, 02-668 Warsaw, Poland ¶ Beijing Synchrotron Radiation Facility, Institute of High Energy Physics, Chinese Academy of Sciences, Beijing 100039, P. R. China S Supporting Information *

ABSTRACT: A series of Sr 1−2x Ce x Na x B 2 Si 2 O 8 and Ba1−2xCexKxB2Si2O8 (x = 0.005, 0.01, 0.02, 0.04, 0.06, 0.08) samples were synthesized by a high-temperature solid-state reaction. The low temperature excitation, emission, and fluorescence decay spectra together demonstrated that all spectral bands arise from the Ce3+ ions located at only one kind of lattice site. The first-principles calculations of the structural and electronic properties of pure and Ce3+-doped MB2Si2O8 (M = Sr, Ba) were performed, and the obtained results were used for understanding the structural changes after doping and identification of the observed position of the host absorption bands. The measured 4f-5d excitation and emission spectra of Ce3+ ions doped in MB2Si2O8 were analyzed and simulated in the framework of the crystal-field (CF) theory. The electron−phonon coupling effect generally ignored in most studies published so far was also taken into account by applying the configurational coordinate model. The validity of such a combined insight into the 5d CF energy level positions and the Stokes shift has been confirmed by analyzing the dependence of the Ce3+ spectroscopic properties on the dopant concentration. In addition, the influence of temperature on the luminescent properties of the studied samples was also explored and is discussed. numerous research works have been published.3−5 Tuning of the 4f-5d transition energy is a key technological problem for the design and exploration of Ce3+-doped optical materials for desired applications. From a theoretical point of view, this controlling technique can be realized by adjusting the combination of 5d crystal-field (CF) energy level positions and strength of the electron−phonon coupling effect (i.e., the Stokes shift). Most of the previous studies focused on the former,5,6 probably because the latter effect was not predominant in those studied cases. However, it is important to understand why the electron−phonon coupling can be ignored (and for which systems) and whether there is a competition between them in some particular hosts. This is the

1. INTRODUCTION Ce3+ ions show 4f-5d transitions from the vacuum ultraviolet (VUV) to the visible range of the spectrum. Because these are the parity-allowed transitions, they are characterized by a fast decay. Phosphors doped with Ce3+ ions are used in important applications. For example, the well-known phosphor Y3Al5O12:Ce3+ can act as a yellow-emitting component in InGaN/GaN-based LEDs (light-emitting diodes).1 The commercially available scintillator Lu2SiO5:Ce3+ can be used in medical imaging detectors for PET (positron emission tomography).2 When Ce3+ ions enter a specific lattice site, their spectroscopic properties and emission lifetime are to a large extent determined by the coordination environment. Understanding the relationship between structural and optical properties of phosphors can help us acquire insights into the methods of developing those materials with specific optical properties. As far as the Ce3+-doped phosphors are concerned, © 2015 American Chemical Society

Received: October 22, 2015 Revised: December 18, 2015 Published: December 18, 2015 569

DOI: 10.1021/acs.jpcc.5b10355 J. Phys. Chem. C 2016, 120, 569−580

Article

The Journal of Physical Chemistry C

first-principles and CF models have provided good support to our experimental analysis and clearly demonstrate the relationship between the local geometry structures around the Ce3+ dopants and the observed 4f and 5d energy level structures of the Ce3+ ions. The present study gives an overall insight into the understanding of the shift of the 5d-4f emission of the Ce3+ ions. We emphasize that the electron−phonon coupling effect previously very often ignored is given more attention in the present study.

primary motivation that prompted us to look into some isostructural compounds with different cation host components. For such compounds, the electron−phonon coupling effect can be gradually enhanced with the cation host components becoming much heavier owing to softening of the host lattices, and thus one can expect to see the competition from the Stokes shift variation with the change in 5d CF energy levels.7,8 The series of danburite MB2Si2O8 (M = Ca, Sr, and Ba) are naturally occurring alkali borosilicate minerals that can be described by the orthorhombic crystallographic system with a Pnma space group (no. 62).9−11 Figure 1 shows a general view

2. EXPERIMENTAL SECTION 2.1. Materials and Synthesis. Series of polycrystalline samples with nominal chemical formulas Sr1−2xCexNaxB2Si2O8 (x = 0.005, 0.01, 0.02, 0.04, 0.06, 0.08; Sr series) and Ba1−2xCexKxB2Si2O8 (x = 0.005, 0.01, 0.02, 0.04, 0.06, 0.08; Ba series) were prepared by a solid-state reaction route at high temperature. When substituting Ce3+ for Sr2+ or Ba2+, the alkali Na+ or K+ ions were added to maintain charge neutrality of whole compounds in consideration of their matching ionic radii,20 respectively. The starting materials SrCO3 [analytical reagent (A.R.)], BaCO3 (A.R.), K2CO3 (A.R.), Na2CO3 (A.R.), H3BO3 (A.R.), SiO2 (A.R.), and CeO2 (99.99%) were weighed in stoichiometric amounts, and then the mixtures were adequately ground in an agate mortar by adding ethanol (A.R.). After being preheated in air at 500 °C for 5 h, the mixtures were spontaneously cooled to room temperature (RT) and thoroughly ground again. Finally, the sintering of the powders was maintained for 10 h in a thermal-carbon reducing atmosphere at 1000 °C for the Sr series and 800 °C for the Ba series, and the products were ground to fine powders for subsequent characterization after cooling to RT. 2.2. Characterization Method. The X-ray diffraction (XRD) data were recorded on a Bruker D8 ADVANCE model powder X-ray diffractometer with Cu Kα radiation (λ = 1.5418 Å) at RT. The vacuum ultraviolet (VUV) excitation and emission spectra were measured at the VUV spectroscopy experimental station on beamline 4B8 of Beijing Synchrotron Radiation Facility (BSRF) under a dedicated synchrotron mode (2.5 GeV, 150−60 mA).21 The UV−vis luminescence spectra and the luminescence decay curves were measured with an Edinburgh FSP920 combined fluorescence lifetime and steadystate spectrometer, which was equipped with a CTI-Cryogenics temperature control system. A 450 W xenon lamp was used as the excitation source for the steady-state spectra, while a 150 W nF900 ns flash lamp with a pulse width of 1 ns and pulse repetition rate of 40 kHz was employed as the excitation source for the luminescence decay curves. 2.3. Computational Method. 2.3.1. First-Principles Calculations. The first-principles calculations of the structural and electronic properties of pure and Ce3+-doped MB2Si2O8 (M = Sr, Ba) were performed by using the periodic ab initio CRYSTAL09 code based on the linear combination of the atomic orbitals method.22 The closed-shell and spin-polarized DFT calculation forms were applied for the pure and doped systems, respectively. The dopant effect was modeled with a 2 × 2 × 1 supercell containing 208 atoms, in which one Sr2+ or Ba2+ ion was replaced by a Ce3+ ion and then the nearest Sr2+ or Ba2+ site around the substituting Ce3+ ion was occupied by a Na+ or K+ ion due to consideration of the local charge compensation. The hybrid exchange-correlation functional WC1PBE consisting of a PBE correlation part and a Wu− Cohen exchange part with a fractional mixing (16%) of the nonlocal Hartree−Fock exchange was employed in this work to

Figure 1. Schematic representations for the unit cell of pure MB2Si2O8 (M = Ca, Sr, and Ba) crystal and the local coordination structure around the Ce3+ ion doped into the M2+ site. Drawn with VESTA.19

of the MB2Si2O8 crystal structure with one unit cell. The M2+ ions 10-fold-coordinated by the oxygen ions are embedded in the corner-shared BO45− and SiO44− tetrahedrons, which form the framework of four- and eight-membered ring interconnected layers, as reported in previous research.12−14 The Wyckoff position analysis for each ion in the crystal reveals that there are eight general atom positions, i.e., M(4c), B(8d), Si(8d), O1(8d), O2(8d), O3(8d), O4(4c), and O5(4c) sites. Because the 4c and 8d Wyckoff positions correspond to the local site symmetries Cs and C1, respectively, the ten oxygen ligands surrounding the M2+ ions can be symmetry-related by a mirror reflection operator and further classified into two groups O(8d) and O(4c) exhibiting two kinds of electron cloud overlap with the M2+ ion. This local coordination environment of the M2+ site can have an important influence on the luminescence properties of Ce3+-doped MB2Si2O8 if the doping Ce3+ ions can enter the M2+ site. The photoluminescence properties of Ce3+-doped CaB2Si2O8 were reported in the last century15 and later referred to in several other reports.16−18 There are many difficulties in the artificial synthesis of this danburite series.12 Moreover, the relatively small Stokes shift for Ce3+-doped CaB2Si2O8 in those reports has left a deep impression upon future work because it is known that the smaller Stokes shift indicates a stiffer host lattice and, as a result, less nonradiative relaxation after the doped ions are excited.13 In this work, we continue this line of research to investigate the VUV−vis photoluminescence properties of the Ce3+-doped MB2Si2O8 (M = Sr, Ba) for the first time. The luminescent properties were examined by considering both the 5d CF effect and the Stokes shift and is discussed in detail under experimental conditions of different measurement temperatures and doping levels. The theoretical calculations employing the 570

DOI: 10.1021/acs.jpcc.5b10355 J. Phys. Chem. C 2016, 120, 569−580

Article

The Journal of Physical Chemistry C yield better estimation on the host band gaps.23 The local Gaussian-type basis sets (BSs) were chosen as follows: the allelectron BSs in the form of 86-311d1G, 86-511G, 8-511G, 8411d1G, and 6-21d1G were used for Si,24 K,25 Na,25 O,26 and B27 atoms, respectively; for Sr and Ba atoms, the Hay−Wadt small-core pseudopotentials and the related valence BSs from ref 28 were adopted; for Ce atom, the inner electrons [Ar]3d10 were described by the scalar-relativistic effective-core pseudopotential (ECP) developed by the Stuttgart/Cologne group,29 whereas the other electrons 4s24p64d105s25p64f16s25d1 were explicitly treated by the valence BS (12s12p9d8f)/[8s7p4d4f]30 related to the ECP that we used. The Monkhorst−Pack schemes for 4 × 4 × 4 and 2 × 2 × 2 k-point meshes in the Brillouin zone were applied to the pure and doped cases, respectively. The truncation criteria for bielectronic integrals (Coulomb and HF exchange series) were set to 8, 8, 8, 8, and 16, and a predefined “extra large” pruned DFT integration grid was adopted together with much higher DFT density and grid weight tolerances (values 8 and 16). The tolerance of the energy convergence on the self-consistent field iterations was set to 10−8 hartree. The Anderson mixing scheme with 60% mixing was employed to achieve the converged solutions for all the calculation cases, and the total spin of the doped systems was locked to a value of 1/2 in the first five cycles due to the single electron occupation on the 4f orbits of the Ce3+ ions. In the geometry optimization with full structural relaxation, the convergence criteria on the root-mean-square of the gradient and the nuclear displacement were set to 0.00006 hartree/bohr and 0.00012 bohr, respectively. 2.3.2. Crystal-Field Analysis. The CF analysis of the 4f-5d excitation and emission spectra of Ce3+ ions in MB2Si2O8 (M = Sr or Ba) was implemented by employing the combinational theoretical scheme of the parametrized CF Hamiltonian model for 4f1 and 5d1 electronic configurations and the exchange charge model (ECM). The details about the scheme itself can be found in our past work,31 and the related calculation parameter setting for the present work is given in Supporting Information.

Figure 2. XRD patterns of (a) Sr1−2xCexNaxB2Si2O8 and (b) Ba1−2xCexKxB2Si2O8 with x = 0.01, 0.04, 0.08; expanded XRD patterns in the region 31.0−33.5° for (c) Sr1−2xCexNaxB2Si2O8 and (d) Ba1−2xCexKxB2Si2O8. The simulated XRD patterns of SrB2Si2O8 and BaB2Si2O8 are for comparison.

that the defect substitution molarity up to 8% cannot induce a significant deformation of the orthorhombic crystallographic phase. There are some very tiny impurity diffraction peaks observed at about 2θ = 21° for Sr0.98Ce0.01Na0.01B2Si2O8 and about 2θ = 28° for Ba0.98Ce0.01K0.01B2Si2O8, respectively. Fortunately, those impurity phases cannot have a very evident influence on the luminescent properties of the synthesized samples because there is only one kind of optical site found in our samples, as described at the end of section 3.3. Herein the Ce3+ ions can be assumed to be incorporated into the hosts by replacing the Sr2+ or Ba2+ ions based on the consideration of their similar ionic radii.20 A more detailed inspection of Figure 2 shows that the diffraction peaks shift slightly toward higher angles with the increasing doping content of Ce3+. This phenomenon is consistent with the substitutions of slightly smaller Ce3+ (1.25 Å) for Sr2+ (1.36 Å) or Ba2+ (1.52 Å) ions at 10-fold coordination. Furthermore, because of the ionic radii difference between the Ce3+ and Ba2+ ions being much larger than that between the Ce3+ and Sr2+ ions, the shift of diffraction peaks in the Ba series is displayed more obviously than those in the Sr series as shown in Figure 2c,d. This observation suggests that the shortening of the average (Ce, Ba)−O distance is more obvious than that of the average (Ce, Sr)−O distance. This experimental evidence has been supported by geometry optimization results of the local coordination environments of Ce3+ ions doped in hosts given by our first-principles calculations: the calculated average Ce3+− O2− bond lengths are much smaller than those of Sr2+−O2− and Ba2+−O2−, and the bond length contraction degree in BaB2Si2O8 is about three times than that in SrB2Si2O8 (as shown in Table 1). In addition, the introduction of the Na+ (1.39 Å) and K+ (1.59 Å) ions into the corresponding alkalineearth cation sites cannot result in a significant change in the host lattices due to the almost negligible ionic radii difference between the substituted and substituting ions, which has also been confirmed by the calculated average Na−O and K−O bond lengths in Table 1. 3.2. Electronic Properties of Pure and Ce3+-Doped MB2Si2O8. Figure 3a,c shows the calculated band structures of

3. RESULTS AND DISCUSSION 3.1. Phase Identification and Structure Analysis. The calculated crystal structural parameters of the pure hosts SrB2Si2O8 and BaB2Si2O8 are summarized in Table S1 (as shown in Supporting Information) together with the experimental ones. The agreement between the experimental and calculated structural characteristics is very good. These calculated structural parameters can also be used to successfully reproduce the following observations about the local ligand environments of alkaline-earth cation sites and various interionic distances: there is only one kind of ten-coordinated Sr2+ or Ba2+ site with point symmetry Cs; the average Sr2+−O2− and Ba2+−O2− bond distances are ∼2.74 Å and ∼2.86 Å, respectively; the nearest adjacent distance of Sr2+−Sr2+ is ∼4.59 Å, while that of Ba2+−Ba2+ is ∼5.06 Å.13,32 The representative XRD patterns of Sr1−2xCexNaxB2Si2O8 and Ba1−2xCexKxB2Si2O8 with x = 0.005, 0.01, 0.02, 0.04, 0.06, and 0.08 are illustrated in Figure 2. The XRD pattern simulations obtained by using the Crystallographic Information Framework (CIF) files of the two compounds10,11 are also presented for the sake of comparison. One can find that the diffraction patterns of Sr 1 − 2 x Ce x Na x B 2 Si 2 O 8 and Ba1−2xCexKxB2Si2O8 with x = 0.01, 0.04, and 0.08 match well with the simulated ones, respectively. Thus, it can be concluded 571

DOI: 10.1021/acs.jpcc.5b10355 J. Phys. Chem. C 2016, 120, 569−580

Article

The Journal of Physical Chemistry C

Table 1. Calculated Distances (M−O, Å) from the Dopant Site to the Ten Nearest Oxygen Ligands before and after Ce3+ and Na+ (or K+) Codoping in MB2Si2O8 (M = Sr or Ba)a SrB2Si2O8 Sr, Cs O1 O2 O3 O4 O5 O6 O7 O8 O9 O10 Ravg

BaB2Si2O8

CeSr, C1

NaSr, C1

Ba, Cs

CeBa, C1

KBa, C1

exptl

calcd

calcd

calcd

exptl

calcd

calcd

calcd

2.5092(4c) 2.5890(8d) 2.5890(8d) 2.5850(8d) 2.5850(8d) 2.6081(8d) 2.6081(8d) 3.0817(8d) 3.0817(8d) 3.3700(4c) 2.7607

2.4955 2.5882 2.5882 2.5788 2.5788 2.6012 2.6012 3.0596 3.0596 3.4118 2.7563

2.3980 2.5026 2.5028 2.5580 2.5581 2.5614 2.5617 2.9625 2.9633 3.4495 2.7018 (−1.98%)

2.4512 2.6128 2.6143 2.4767 2.4774 2.6395 2.6411 3.1101 3.1117 3.3635 2.7498 (−0.24%)

2.6232(4c) 2.7937(8d) 2.7937(8d) 2.7521(8d) 2.7521(8d) 2.7628(8d) 2.7628(8d) 3.0602(4c) 3.1626(8d) 3.1626(8d) 2.8626

2.6552 2.8051 2.8051 2.7815 2.7815 2.7707 2.7707 3.0291 3.1689 3.1689 2.8737

2.4197 2.6419 2.6421 2.6517 2.6519 2.6620 2.6623 2.8543 2.9641 2.9646 2.7115 (−5.64%)

2.6748 2.7452 2.7776 2.7786 2.7685 2.7452 2.7460 2.8811 3.2083 3.2095 2.8535 (−0.70%)

The columns entitled as “exptl” contain the experimental data of the M−O bond lengths in the pure hosts, and the site symmetries of the ten oxygen ligands are also labeled after those data. The defect-induced average bond length changes with respect to the calculated pure host cases are given in the last row.

a

Figure 3. Calculated band structure (a, c) and DOS/PDOS (b, d) diagrams for pure MB2Si2O8 (M = Sr or Ba) crystals. The letters Γ, Z, T, Y, S, X, U, R, and LD stand for the high-symmetry k points (0,0,0), (0,0,1/2), (0,1/2,1/2), (0,1/2,0), (1/2,1/2,0), (1/2,0,0), (1/2,0,1/2), (1/2,1/2,1/2), and (0,0,z) (0 < z < 1/2), respectively.

(BaB2Si2O8). Composition of the calculated electronic bands can be analyzed with the help of the density of states (DOS) and partial density of states (PDOS) diagrams shown in Figure 3b,d. The valence bands in the two cases are formed by the O2p states, whereas the conduction bands are basically composed of the d states of alkaline-earth metal atoms (the Sr-4d or Ba-5d

pure SrB2Si2O8 and BaB2Si2O8 crystals, respectively. The common feature is that the valence bands are rather flat, which suggests a weak dispersion of holes in the reciprocal space, whereas the conduction bands are more dispersive (higher mobility of electrons than holes). The calculated band gaps are indirect and equal to 7.71 eV (SrB2Si2O8) and 7.96 eV 572

DOI: 10.1021/acs.jpcc.5b10355 J. Phys. Chem. C 2016, 120, 569−580

Article

The Journal of Physical Chemistry C

Figure 4. Calculated DOS/PDOS diagrams for Ce3+-doped MB2Si2O8 (M = Sr (a) or Ba (b)) with Na+ or K+ ions as local charge compensators. The two purple dashed vertical lines denote the positions of the top of the host’s valence band and the bottom of the host’s conduction band, respectively.

Table 2. Effective Mulliken Charge Q (e) of Individual Ions in Local Clusters Composed of the Dopant Site and Its Nearest Ten Oxygen Ligands before and after Ce3+ and Na+ (or K+) Codoping in MB2Si2O8 (M = Sr or Ba) SrB2Si2O8 Q(M2+) Q(O12−) Q(O22−) Q(O32−) Q(O42−) Q(O52−) Q(O62−) Q(O72−) Q(O82−) Q(O92−) Q(O102−) Q(O)avg

BaB2Si2O8

Sr

CeSr

NaSr

Ba

CeBa

KBa

1.875 −1.206 −1.234 −1.234 −1.169 −1.169 −1.145 −1.145 −1.234 −1.234 −1.206 −1.198

2.245 −1.169 −1.123 −1.123 −1.154 −1.154 −1.121 −1.121 −1.131 −1.131 −1.204 −1.143

0.968 −1.094 −1.131 −1.131 −1.062 −1.062 −1.028 −1.028 −1.123 −1.123 −1.184 −1.097

1.808 −1.218 −1.114 −1.114 −1.119 −1.119 −1.195 −1.195 −1.218 −1.195 −1.195 −1.168

2.326 −1.227 −1.135 −1.135 −1.125 −1.125 −1.127 −1.127 −1.247 −1.116 −1.116 −1.148

0.999 −1.147 −1.070 −1.069 −1.035 −1.035 −1.116 −1.116 −1.194 −1.127 −1.127 −1.104

occupation is empty for the 5d excitation state of Ce3+ ions, as Du et al. regarded the Ce3+ 5d excitation state as a Ce4+ ion plus a free electron trapped into its 5d ground-state orbital.33 Hence, the calculated 5d states here can be approximately thought to be the Ce 4f15d1 states if Ce3+ becomes Ce2+ by capturing a free electron into those empty 5d states. Inspection of the application diagram of the double zigzag model of Dorenbos to YPO4 supports such a hypothesis because the lowest 4f15d1 state of Ce2+ ion can lie in the band gap, whereas the lowest 4f2 state falls into the conduction band.34 Moreover, there should be hybridization of the Ce-5d states and the conduction band states, and thus a strict border between the two kinds of states does not exist. Therefore, those so-called Ce 4f15d1 states can lead to a red shift of the host absorption appearing in the excitation spectra of the two doped systems. The energy levels of the doping Na+ or K+ ions are not found inside the band gap, and it seems that they only play a role of charge compensator. However, the first-principles study of near-edge gap states in CaF2 with Na+ impurities reveals that the Na+ doping can slightly lift the top of the valence band by changing the 2p electron cloud distribution of the nearestneighbor F− ions around the Na+ ion due to the charge mismatch between the substituting and substituted ions.35 Inspection of Figure 4 shows that the position of the top of the valence band of Ce3+/Na+-codoped SrB2Si2O8 is 0.13 eV higher than that of the pure host, whereas there is also an energy shift

states) and the Si-3p states with the admixture of the B-2p states. Introduction of Ce3+ and alkali-metal monovalent ions can considerably change the electronic properties of hosts. The calculated DOS and PDOS diagrams for Ce3+/Na+-doped SrB2Si2O8 and Ce3+/K+-doped BaB2Si2O8 are illustrated in Figure 4. The calculated band structure diagrams are also shown in Figure S1 as a reference to confirm our DOS and PDOS analysis. The 4f states of the Ce3+ ions are clearly seen in the band gap. The position of the Ce3+ 4f occupied state with respect to the top of the host’s valence band can be evaluated as 1.76 eV (SrB2Si2O8) and 1.19 eV (BaB2Si2O8). Such an energy separation can be related to the Coulomb interaction strength between the Ce3+ 4f electron and the O2− 2p electrons, and thus one can conclude that the shorter the Ce−O bond average length, the higher the Ce3+ 4f occupied energy level position. The calculated two Ce3+ 5d states appear below the bottom of the conduction band, and their energy positions relative to the top of the host’s valence band can be evaluated as 7.05 and 7.53 eV for SrB2Si2O8, and 7.13 and 7.63 eV for BaB2Si2O8, whereas others are fully resonant in the conduction band. Because the 4f electron occupation can affect the 5d level positions due to the screening effect by the 4f electron, it should be kept in mind that the Ce3+ 5d energy level positions given by the present DFT ground-state calculations are not the real 5d energy levels corresponding to the 4f-5d transitions because the 4f 573

DOI: 10.1021/acs.jpcc.5b10355 J. Phys. Chem. C 2016, 120, 569−580

Article

The Journal of Physical Chemistry C of 0.25 eV for the case of Ce3+/K+-codoped BaB2Si2O8. The dependence of such a slight shift of the top of the valence band on the supercell size was checked, and the same results were obtained in the case of the 2 × 2 × 2 supercell. Mulliken population analysis was performed for individual ions in those local clusters composed of the dopant site and its nearest ten oxygen ligands before and after Ce3+ and Na+ (or K+) codoping in MB2Si2O8. The calculation results given in Table 2 suggest that the O2− 2p electrons tend to go back to those alkalineearth cation sites after Ce3+ and Na+ (or K+) codoping, and thus the 2p occupied orbitals in comparison with the case of the pure hosts are more dispersive so that their energies in the band structure diagrams will shift to the higher energy direction as shown by the rise of the top of the valence band in Figure 4. The energy level positions of Ce impurity ions with respect to the top of the valence band before and after doping are plotted in Figure 5 based on the calculation results above and

Figure 6. Measured excitation/emission and simulated 4f-5d absorption/emission spectra of Sr0.99Ce0.005Na0.005B2Si2O8 at 15 K. The light green dotted lines represent the Gaussian fitting of the excitation spectrum, and the dark green vertical bars show the calculated energies and relative line strengths of the 4f-5d transition ZPLs.

Figure 5. Energy level positions of Ce impurity ions with respect to the top of the valence band (VB) before and after doping. The red horizontal dashed bars represent the 5d energy levels whereas the blue horizontal dotted bars stand for the predicted 4f15d1 energy levels (only those below the conduction band (CB) are shown).

Figure 7. Measured excitation/emission and simulated 4f-5d absorption/emission spectra of Ba0.99Ce0.005K0.005B2Si2O8 at 15 K.

3.3. Luminescence of Sr0.99Ce0.005Na0.005B2Si2O8 and Ba0.99Ce0.005K0.005B2Si2O8. Figure 6 shows the VUV-UV excitation and emission spectra of Sr0.99Ce0.005Na0.005B2Si2O8 at 15 K. In the excitation spectrum, monitoring the emission at 345 nm, there are three intense bands ((A, B), (C, D), and F) and a weak band E. Each of the first two intense bands can be decomposed into two sub-bands by using the Gaussian multiple peak fitting technique as shown in Figure 6. Following our electronic structure calculation results in the last section, the host absorption onset of Sr0.99Ce0.005Na0.005B2Si2O8 can be predicted to be around 179 nm, and thus the higher-lying band F can be attributed to the host absorption. There is no doubt that the bands A to E can be assigned to the 4f-5d excitation transitions of Ce3+ ions in the host if one refers to the 4f-5d excitation data of Ce3+ ions doped in various oxides as summarized by Dorenbos in ref 36. In the emission spectrum obtained by using excitation at 318 nm, there are two broad bands located around 345 and 370 nm, respectively. The energy difference between them is 1958 cm−1, which is a very typical spin−orbit splitting between 2F5/2 and 2F7/2 multiplets of 4f1 configuration. Therefore, the two emission bands can be attributed to the transitions from the lowest 5d CF energy level to the 2F5/2 and 2F7/2 multiplets of 4f1 configuration,

Table 3. Measured and Calculated 5d Energy Level Positions, 5d CF Splitting, and 5d Energy Centroid of Ce3+ Doped in MB2Si2O8 (M = Sr, Ba) (all in cm−1)a SrB2Si2O8 E(5d1) E(5d2) E(5d3) E(5d4) E(5d5) CFS centroid

BaB2Si2O8

exptl

calcd

exptl

pred

calcd

30100 31490 39530 41300 45820 15720 37648

30462 32357 37552 41864 46007 15545 37648

30350 31730 35440 40590 44250 13900 36472

30675 32412 34483 39485 43470 12795 36105

30247 32637 34687 40539 44249 14002 36472

The column entitled as “pred” contains the predicted values obtained by using the ECM G parameters of Ce3+ ions in SrB2Si2O8.

a

the observed 4f-5d transition energies listed in Table 3. The energies of the host absorption onsets can be estimated as 6.92 and 6.88 eV for Ce3+/Na+-doped SrB2Si2O8 and Ce3+/K+doped BaB2Si2O8, respectively. The theoretical predictions are in good agreement with the observations of the excitation spectra of the doped samples (as shown in Figures 6 and 7). 574

DOI: 10.1021/acs.jpcc.5b10355 J. Phys. Chem. C 2016, 120, 569−580

Article

The Journal of Physical Chemistry C respectively. Moreover, the Stokes shift of Ce3+ in the host can be estimated as 2264 cm−1 by evaluating the difference between the lowest 4f-5d excitation maximum and the highest 5d-4f emission maximum of Ce3+ ions. Following Reid’s assumption about the 5d electronic origin position,33one can calculate all the 4f-5d transition zero-phonon lines (ZPLs) for the present case by subtracting one-half of the Stokes shift (i.e., Eshift) from the maximum of those corresponding 4f-5d vibronic bands. The estimated 5d energy level positions are collected in Table 3. The 5d CF calculation was performed by fitting them, and thus the values of the ECM fitting parameters G(O(8d)) and G(O(4c)) were determined to be 1.2754 and 2.2470, respectively. It is easily seen from Table 3 and Figure 6 that there is very good agreement between the experimental and theoretical energy band positions. A slight disagreement between the calculated and observed band intensities can be also found in Figure 6 (similarly in Figure 7), but such a disagreement is caused by the difference between the simulated absorption and experimental excitation spectra, as the quantum efficiency of the relaxation behavior from those excited energy levels to the monitoring luminescent level in the excitation spectra should never reach 100%. In addition, the calculated emission spectrum reproduces very well the doublet emission band caused by the 4f spin−orbit splitting of Ce3+ ions. This second agreement further confirms the validity of our CF calculations. The synchrotron radiation VUV-UV excitation spectrum and corresponding emission spectrum of Ba0.99Ce0.005K0.005B2Si2O8 at 15 K are displayed in Figure 7. Two broad bands were recorded in the excitation spectrum when monitoring the emission at 343 nm. To focus our study on the 4f-5d transitions of Ce3+ ions in BaB2Si2O8, the band F around 172 nm was first separated from our analysis scope due to its character of host absorption. The electronic structure calculation results in the last section also support the present assignment because the calculated host absorption onset is located at 180 nm. The broad band stretching from 210 to 335 nm can be attributed to the 4f-5d excitation transitions of Ce3+ ions in BaB2Si2O8 if we make the comparison with the 4f-5d excitation energy region of Ce3+ ions in SrB2Si2O8. Unfortunately, the band is very poorly resolved so that five 5d CF components cannot be clearly distinguished. This is probably because the spectral broadening behavior caused by the partial melting of the prepared sample can easily blur the border of each 5d CF component. When checking the sample carefully, we indeed found that a small part of the final Ba0.99Ce0.005K0.005B2Si2O8 product was slightly melted under the present synthetic conditions. It is very hard to avoid such sample melting because the formation of the pure phase requires that the sintering temperature and time are not lower than what we used in the work. A similar phenomenon occurs in the 5d-4f emission spectrum under excitation at 306 nm. The 4f spin−orbit splitting of Ce3+ ions is washed off due to the spectral broadening, and thus there are no 5d-4f doublet emission bands observed even at 15 K. However, the partial melting character is not found in the isomorphic compound Sr0.99Ce0.005Na0.005B2Si2O8 although the preparation temperature of Sr0.99Ce0.005Na0.005B2Si2O8 (1000 °C) is far higher than that of Ba0.99Ce0.005K0.005B2Si2O8 (800 °C). The reason for such a melting phenomenon will be further investigated. Due to the isostructural character with SrB2Si2O8, the ECM G parameter values extracted from the 4f-5d excitation spectrum of Ce3+ ions in SrB2Si2O8 can be used to predict the 5d CF splitting pattern of Ce3+ ions in BaB2Si2O8. All the

5d energy level positions can be tentatively determined by combining the predicted 5d CF splitting pattern and the lowest 5d CF energy level position estimated from the intersection point (∼326 nm) of the measured excitation and emission spectra.6 Following the consideration, the predicted 5d energy level positions are collected in Table . The comparison of the predicted results with the observed 4f-5d transition broad band from 210 to 335 nm suggests that the band can be decomposed into five Gaussian-type sub-bands if we adopt the Gaussian multiple peak fitting technique, and the position of the maximum of each sub-band can be initialized by using those predicted energy level values. The Gaussian fitting to the experimentally observed 4f-5d transition band was implemented to obtain the position of the maximum of each Gaussian-type sub-band, and then all the 5d electronic origins were deduced by subtracting the value of the parameter Eshift from those maximum positions. The final results are listed in Table 3 as the experimental values of the 5d CF fitting calculation. The 5d CF fitting calculation was performed, and the values of the ECM fitting parameters G(O(8d)) and G(O(4c)) were evaluated as 2.1531 and 3.6301, respectively. Inspection of Table 3 and Figure 7 shows that the agreement between the calculated and experimental 4f-5d excitation spectra is very good. A similar decomposition can be carried out on the 5d-4f emission spectrum. The 5d-4f emission spectrum contains two Gaussian-type sub-bands whose maxima are located at 343 and 368 nm, respectively. The energy difference between the two sub-bands (∼1980 cm−1) is the typical value of the 4f spin−orbit splitting of Ce3+ ions, which confirms the validity of this Gaussian-type decomposition. The calculated 5d-4f emission spectrum also indicates a good agreement with the experimental analyses as shown in Figure 7. The Stokes shift of Ce3+ ions in the host can be estimated as 2391 cm−1 by inspecting the positions of the lowest 4f-5d excitation maximum and the highest 5d-4f emission maximum in Figure 7. For Ce3+ ions in a specific site, the coordination environment has a large influence on the 5d crystal field splitting (CFS) and centroid energy which together decide the positions of 5d states. The CFS is defined as the difference between the highest and lowest 5d states, and the centroid energy is defined as the average position of the 5d states. Based on the calculated and measured 5d energy level positions, the 5d CFS values and centroid estimated energies are listed in Table 3. A smaller CFS and lower centroid energy can be found in Ce3+-doped BaB2Si2O8, which is due to the influence of the bigger site size. Furthermore, the lowest 5d state of Ce3+ in BaB2Si2O8 is slightly higher than that in SrB2Si2O8. This is related to the combined factors of the smaller CFS and lower centroid energy in Ba samples, and therefore the CFS should predominate in this case. The monitoring wavelength dependences on the excitation and emission spectra and the luminescence decay curves of samples Sr0.99Ce0.005Na0.005B2Si2O8 and Ba0.99Ce0.005K0.005B2Si2O8 at 15K were checked to confirm whether there is only one kind of occupied site for the doping Ce3+ ions, as shown in Figure S2, Supporting Information. It can be easily seen from Figure S2 that the energy-selective and time-resolved excitation and emission do not result in changes in the spectroscopic properties of the studied two samples. Thus, one can conclude from the observation that Ce3+ ions must only enter the M2+ site. 575

DOI: 10.1021/acs.jpcc.5b10355 J. Phys. Chem. C 2016, 120, 569−580

Article

The Journal of Physical Chemistry C

Figure 8. Temperature-dependent luminescence decay curves of Sr0.99Ce0.005Na0.005B2Si2O8 (a) and Ba0.99Ce0.005K0.005B2Si2O8 (b); insets show the lifetimes of Ce3+ emissions as a function of temperature and the fitting curves.

Figure 9. Concentration-dependent excitation and emission spectra of Sr1−2xCexNaxB2Si2O8 (x = 0.005−0.08) (a, c) and Ba1−2xCexKxB2Si2O8 (x = 0.005−0.08) (b, d) at RT. The insets show the integrated Ce3+ emission intensities as a function of x.

3.4. Influence of Temperature on Luminescence. The temperature-dependent synchrotron radiation VUV-UV excitation spectra and corresponding emission spectra of Sr0.99Ce0.005Na0.005B2Si2O8 and Ba0.99Ce0.005K0.005B2Si2O8 were measured and are shown in Figure S3, Supporting Information. With increasing temperature, the positions and shapes of these excitation and emission bands are nearly invariant. This implies that the 5d crystal field splitting (CFS), centroid energy, and Stokes shift are invarant in the testing temperature range (15− 296 K). The temperature-dependent luminescence decay curves of Sr0.99Ce0.005Na0.005B2Si2O8 and Ba0.99Ce0.005K0.005B2Si2O8 are shown in Figure 8a,b, respectively. The decay curves gradually deviate from monoexponential decay with increasing temperature. This effect is induced by the thermal quenching effect

which may relate to several different mechanisms, such as the thermally activated multiphonon relaxation, intersystem crossing, and ionization to the conduction band.37,38 The theoretical analysis of the temperature dependence of the lifetime of the Ce3+ 5d emission state was implemented below to understand the possible mechanism of the thermal quenching effect. The average fluorescence lifetime (τ) of the Ce3+ emission can be estimated by using the formula39 ∞

τ=

∫0 tI(t )dt ∞

∫0 I(t )dt

(1)

where I(t) is the emission intensity after pulsed excitation and I(0) is the emission intensity at t = 0. From the obtained τ, as 576

DOI: 10.1021/acs.jpcc.5b10355 J. Phys. Chem. C 2016, 120, 569−580

Article

The Journal of Physical Chemistry C

and Stokes shifts as the doping concentration of Ce3+ ions increases. As shown in Figure 10, the concentration-dependent

shown in the insets of Figure 8a,b, we calculated the activation energies using eq 2,40 τT =

τ0 E

( )

1 + A exp − kTa

(2)

where Ea is the activation energy for the thermal quenching, τ0 is the initial lifetime of the phosphor, and τT is the lifetime at a given temperature T. k is the Boltzmann constant (8.625 × 10−5 eV·K−1), and A is the rate constant for the thermally activated escape. The values of Ea were obtained to be 0.21 eV f o r S r 0 . 9 9 Ce 0 . 0 0 5 N a 0 . 0 0 5 B 2 S i 2 O 8 an d 0 . 19 e V f or Ba0.99Ce0.005K0.005B2Si2O8, respectively. Due to the large energy difference between the lowest 5d CF energy level and the 2FJ (J = 5/2, 7/2) multiplet of 4f1 configuration, the multiphonon relaxation between the 5d and 4f energy states is not predominant. It can also be clearly found that the energy differences between the lowest 5d level and the bottom of the conduction band are much larger than the values of Ea (0.21 or 0.19) if we refer to Figure 5, and therefore the thermally activated ionization cannot be the leading mechanism. The most possible reason in this case should be the thermally induced nonradiative relaxation to other kinds of defects or the process of the second type of charge-transfer vibronic exciton.41 3.5. Influence of Doping Concentration on Luminescence. The normalized excitation spectra of Sr1−2xCexNaxB2Si2O8 (λem = 370 nm) and Ba1−2xCexKxB2Si2O8 (λem = 360 nm) with different doping concentrations (x) are shown in Figure 9a,b, respectively. With the increase of x, the excitation band positions are apparently unchanged except for a slight red shift in sample Ba1−2xCexKxB2Si2O8 (x = 0.08). The highest intensity normalized emission spectra of Sr1−2xCexNaxB2Si2O8 (λex = 320 nm) and Ba1−2xCexKxB2Si2O8 (λex = 300 nm) are also displayed in Figure 9 as a function of x at RT. There are significant red shifts in the Ce3+ emission bands of Ba1−2xCexKxB2Si2O8 with increasing x, but the red shift is smaller for the Ce 3 + emission bands of Sr1−2xCexNaxB2Si2O8. The dependence of integrated Ce3+ emission intensity of the Sr series and the Ba series is shown in the insets of Figure 9c,d, respectively. The maximum emission intensity of Sr1−2xCexNaxB2Si2O8 occurs at x = 0.02, whereas the luminescence intensity reaches the highest value at x = 0.04 in Ba1−2xCexKxB2Si2O8. The emission intensities in both cases subsequently decrease after the maxima. This is due to the concentration quenching effect caused by the energy transfer between Ce3+ ions. Moreover, because the distance between the nearest adjacent Sr2+−Sr2+ (∼4.59 Å) is shorter than that of Ba2+−Ba2+ (∼5.06 Å), the energy transfer between Ce3+ ions for the case of SrB2Si2O8 can be more efficient, which has been confirmed by the lower quenching concentration in SrB2Si2O8 as shown in Figure 9c. In the case of a single type of crystallographic site, due to energy migration to perturbed Ce3+ ions, the emission tends to occur from Ce3+ having relative lower energy and results in inhomogeneous broadening; therefore, the emission seems to shift to the long-wavelength side.42 Because the Ce3+−O2− bond length contraction degree in BaB2Si2O8 is larger than that in SrB2Si2O8 (as shown in Table 1), the site-perturbed phenomenon should be stronger in Ba samples than that in Sr samples. Thus, the red shift of the emission spectrum is larger in Ba samples, as pointed out previously. Herein we used the excitation and emission spectra of Ba1−2xCexKxB2Si2O8 (x = 0.005−0.08) samples to study the energy changes of the ZPLs

Figure 10. Concentration-dependent ZPL positions and 5d1 → 2FJ (J = 5/2, 7/2) transitions obtained by Gaussian multipeak fitting of the emission spectra of samples Ba1−2xCexKxB2Si2O8 (x = 0.005−0.08).

ZPLs can be taken from the intersection point of the excitation and emission spectra.43 Meanwhile, following the configurational coordinate model,44 the value of the Stokes shift (Δ(Stokes)) can be estimated from the energy difference between the maxima of the lowest 4f-5d excitation and the highest 5d-4f emission bands, or roughly predicted by the double of the energy difference between the ZPL and the 5d1 → 2F5/2 emission band maximum. The contribution to the Stokes shift from the 5d-4f emission transition is not equal to that from the 4f-5d excitation transition, and thus the predicted value of the Stokes shift can be only used as a reference when evaluating our experimental estimation values. Because the 5d → 4f emission bands are so broad as to consist of two subbands to 2F5/2 and 2F7/2 multiplets, the Gaussian multiple peak fitting technique should be applied to obtain the real position of the band maximum of the 5d1 → 2F5/2 emission. All the spectroscopic parameters mentioned here are collected in Table 4. The red shift of the ZPL positions can be clearly observed as shown in Figure 10 and Table 4. This can be correlated to the decreased average Ce3+−O2− bond length in the Ba series as the doping level increases. The observed Stokes shift shows an increasing trend with the change in x, but such a change is more dramatic than the shift of the ZPL, as shown in Table 4. Table 4. Observed ZPL Energy Positions, Emission and Excitation Maxima, and Stokes Shift of the 5d1-2F5/2 Transitions of Samples Ba1‑2xCexKxB2Si2O8 (x = 0.005-0.08) (all in cm−1)a

x x x x x x

= = = = = =

0.005 0.01 0.02 0.04 0.06 0.08

ZPL

exptl (5d1→2F5/2)

30675 30395 30303 30030 29762 29240

28902 28249 28169 27624 26954 26316

exptl (2F5/2→ 5d1) 32395 32241 32199 32120 32057 32001

(32448) (32541) (32437) (32436) (32570) (32164)

Δ(Stokes) 3493 3992 4030 4496 5103 5685

(3546) (4292) (4268) (4812) (5616) (5848)

a

The values in the parentheses are the theoretical ones and were obtained by using the two formulas: Δ(Stokes)calcd ≈ 2 × (ZPLexptl(5d1 → 2F5/2)) and calcd(2F5/2 → 5d1) = exptl(5d1 → 2F5/2)+Δ(Stokes)calcd. 577

DOI: 10.1021/acs.jpcc.5b10355 J. Phys. Chem. C 2016, 120, 569−580

Article

The Journal of Physical Chemistry C

Figure 11. Concentration-dependent luminescence decay curves of Sr1−2xCexNaxB2Si2O8 (x = 0.005−0.08) (a) and Ba1−2xCexKxB2Si2O8 (x = 0.005−0.08) (b) at RT; insets show the lifetimes as a function of x.

present study of 5d electronic structure and the Stokes shift could also be applied to Eu2+-doped systems, and details will be presented elsewhere.

Therefore, one can conclude that although the energy position of the 5d1 → 2F5/2 emission maximum is determined by both the ZPL and the Stokes shift, the Stokes shift in Ba1−2xCexKxB2Si2O8 looks more dominant. The luminescence decay curves are shown in Figure 11. The average fluorescence lifetimes of Ce3+ with different x in Sr1−2xCexNaxB2Si2O8 and Ba1−2xCexKxB2Si2O8 are shown in the insets of Figure 11a,b, respectively. The lifetimes of Ce3+ are slightly changed in Sr1−2xCexNaxB2Si2O8 (x ≤ 0.02) and Ba1−2xCexKxB2Si2O8 (x ≤ 0.04), mainly becaused of nonradiative loss at the perturbed sites42 or the emission from selfabsorption6 of Ce3+ ions. At higher x, the decays become faster and deviate from monoexponential gradually, which is related to the concentration quenching effects.



ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpcc.5b10355. Calculation details of the CF analysis, calculated and experimental structural data for pure MB2Si2O8 (M = Sr or Ba) crystals (Table S1), the 4f and 5d energy parameters of Ce3+ ions doped in MB2Si2O8 (M = Sr and Ba) in the framework of ECM (unit: cm−1) (Table S2), calculated band structure diagrams for Ce3+/Na+-doped SrB2Si 2O8 and Ce3+/K+-doped BaB2 Si2O 8 crystals (Figure S1), energy-selective excitation/emission spectra and luminescence decay curves of Sr0.99Ce0.005Na0.005B2Si2O8 and Ba0.99Ce0.005K0.005B2Si2O8 at 15 K (Figure S2), and temperature-dependent excitation and emission spectra of Sr0.99Ce0.005Na0.005B2Si2O8 and Ba0.99Ce0.005K0.005B2Si2O8 (Figure S3). Full list of authors for ref 22 (PDF)

4. CONCLUSION In this paper, a series of Ce3+-doped MB2Si2O8 (M = Sr, Ba) samples were synthesized by solid-state reaction, and their VUV−vis luminescence properties were studied for the first time by using a synchrotron radiation light source. The observed results of the energy-selective excitation/emission spectra and luminescence decay curves confirm that there is only one Ce3+ luminescent center. The measurement of the temperature effect on the Ce3+ 4f-5d luminescence properties indicates that the temperature increase may accelerate the nonradiative relaxation rates of the 5d CF energy levels, but it does not affect the 5d CF energy levels and the Stokes shift. The structural change after Ce3+-doping and the host absorption energy position were predicted by employing the first-principles method, and the results obtained are in reasonable agreement with the experimental data. The CF analyses of the 4f-5d excitation and emission bands of the synthesized samples at 15 K were performed, and a clear link between the local geometry structures around Ce3+ dopants and the observed 4f and 5d energy level structures of Ce3+ ions was demonstrated by the ECM. The Stokes shift ignored by most papers was highlighted and given more attention, and its combination with the 5d CF energy level positions in the framework of the configurational coordinate model was used to analyze the 4f-5d spectroscopic dependence on the Ce3+doping concentration effect. The calculation results reveal that the change in Stokes shift is comparable with the shift in 5d CF energy levels as the doping concentration increases, and thus the Stokes shift should not be ignored when analyzing the shift of the 5d-4f emission band of Ce3+ ions. Insight from the



AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]; tel +86 13647659358; fax +86 23 62471791. *E-mail: [email protected]; tel +86 20 84113695; fax +86 20 84111038. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The work is financially supported by the National Natural Science Foundation of China (21171176, U1232108, and U1432249), the Guangzhou Science and Technology Project (2013Y2-00118), and the Science and Technology Project of Guangdong Province (2013B010403010). C.-G. Ma acknowledges the financial support from National Natural Science Foundation of China (grant no. 11204393), Natural Science Foundation Project of Chongqing (grant no. CSTC2014JCYJA50034), and National Training Program of Innovation and Entrepreneurship for Undergraduates (grant no. 201410617001). M.G. Brik acknowledges the support from 578

DOI: 10.1021/acs.jpcc.5b10355 J. Phys. Chem. C 2016, 120, 569−580

Article

The Journal of Physical Chemistry C

BaSiF6 and Selecting Host Lattices with 1S0 Pr3+ Emission. J. Lumin. 2002, 97, 212−223. (18) Dorenbos, P. f→d Transition Energies of Divalent Lanthanides in Inorganic Compounds. J. Phys.: Condens. Matter 2003, 15, 575−594. (19) Momma, K.; Izumi, F. VESTA 3 for Three-Dimensional Visualization of Crystal, Volumetric and Morphology Data. J. Appl. Crystallogr. 2011, 44, 1272−1276. (20) Shannon, R. D. Revised Effective Ionic Radii and Systematic Studies of Interatomic Distances in Halides and Chalcogenides. Acta Crystallogr., Sect. A: Cryst. Phys., Diffr., Theor. Gen. Crystallogr. 1976, 32, 751−767. (21) Tao, Y.; Huang, Y.; Gao, Z. H.; Zhuang, H.; Zhou, A. Y.; Tan, Y. L.; Li, D. W.; Sun, S. S. Developing VUV Spectroscopy for Protein Folding and Material Luminescence on Beamline 4B8 at the Beijing Synchrotron Radiation Facility. J. Synchrotron Radiat. 2009, 16, 857− 863. (22) Dovesi, R.; Saunders, V. R.; Roetti, C.; Orlando, R.; ZicovichWilson, C. M.; Pascale, F.; Civalleri, B.; Doll, K.; Harrison, N. M.; Bush, I. J.; et al. CRYSTAL09 User’s Manual; University of Torino: Torino, Italy, 2009. (23) Demichelis, R.; Civalleri, B.; Ferrabone, M.; Dovesi, R. On the Performance of Eleven DFT Functionals in the Description of the Vibrational Properties of Aluminosilicates. Int. J. Quantum Chem. 2010, 110, 406−415. (24) Pascale, F.; Zicovich-Wilson, C. M.; Orlando, R.; Roetti, C.; Ugliengo, P.; Dovesi, R. Vibration Frequencies of Mg3Al2Si3O12 Pyrope. An ab Initio Study with the CRYSTAL Code. J. Phys. Chem. B 2005, 109, 6146−6152. (25) Civalleri, B.; Ferrari, A. M.; Llunell, M.; Orlando, R.; Merawa, M.; Ugliengo, P. Cation Selectivity in Alkali-Exchanged Chabazite: An ab Initio Periodic Study. Chem. Mater. 2003, 15, 3996−4004. (26) Bredow, T.; Jug, K.; Evarestov, R. A. Electronic and Magnetic Structure of ScMnO3. Phys. Status Solidi B 2006, 243, R10−R12. (27) Maslyuk, V. V.; Islam, M. M.; Bredow, T. Electronic Properties of Compounds of the Li2O-B2O3 System. Phys. Rev. B: Condens. Matter Mater. Phys. 2005, 72, 125101. (28) Piskunov, S.; Heifets, E.; Eglitis, R. I.; Borstel, G. Bulk Properties and Electronic Structure of SrTiO3, BaTiO3, PbTiO3 Perovskites: An ab Initio HF/DFT Study. Comput. Mater. Sci. 2004, 29, 165−178. (29) Dolg, M.; Stoll, H.; Preuss, H. Energy-Adjusted ab Initio Pseudopotentials for the Rare Earth Elements. J. Chem. Phys. 1989, 90, 1730−1734. (30) Cao, X.; Dolg, M. Segmented Contraction Scheme for SmallCore Lanthanide Pseudopotential Basis Sets. J. Mol. Struct.: THEOCHEM 2002, 581, 139−147. (31) Przybylińska, H.; Ma, C.-G.; Brik, M. G.; Kamińska, A.; Szczepkowski, J.; Sybilski, P.; Wittlin, A.; Berkowski, M.; Jastrzębski, W.; Suchocki, A. Evidence of Multicenter Structure of Cerium Ions in Gadolinium Gallium Garnet Crystals Studied by Infrared Absorption Spectroscopy. Phys. Rev. B: Condens. Matter Mater. Phys. 2013, 87, 045114. (32) Wang, Y. H.; Zhang, Z. Y.; Zhang, J. C.; Lu, Y. H. Electronic Properties and Rare-Earth Ions Photoluminescence Behaviors in Borosilicate: SrB2Si2O8. J. Solid State Chem. 2009, 182, 813−820. (33) Biswas, K.; Du, M.-H. Energy Transport and Scintillation of Cerium-Doped Elpasolite Cs2LiYCl6: Hybrid Density Functional Calculations. Phys. Rev. B: Condens. Matter Mater. Phys. 2012, 86, 014102. (34) Poolton, N. R. J.; Bos, A. J. J.; Jones, G. O.; Dorenbos, P. Probing Electron Transfer Processes in YPO4: Ce, Sm by Combined Synchrotron-Laser Excitation Spectroscopy. J. Phys.: Condens. Matter 2010, 22, 185403. (35) Bermudez, V. M. Hartree-Fock Study of Near-Edge Gap States in CaF2 with Na+, Cl− or Sr2+ Impurities. Comput. Mater. Sci. 2002, 24, 501−512. (36) Dorenbos, P. 5d-Level Energies of Ce3+ and the Crystalline Environment. III. Oxides Containing Ionic Complexes. Phys. Rev. B: Condens. Matter Mater. Phys. 2001, 64, 125117.

National Recruitment Program of High-end Foreign Experts (grant no. GDW20145200225), Program for the Foreign Experts offered by Chongqing University of Posts and Telecommunications, European Regional Development Fund (Center of Excellence ‘Mesosystems: Theory and Applications’, TK114), Marie Curie Initial Training Network LUMINET, grant agreement no. 316906, and Ministry of Education and Research of Estonia, Project PUT430.



REFERENCES

(1) Jang, H. S.; Im, W. B.; Lee, D. C.; Jeon, D. Y.; Kim, S. S. Enhancement of Red Spectral Emission Intensity of Y3Al5O12:Ce3+ Phosphor via Pr co-Doping and Tb Substitution for the Application to White LEDs. J. Lumin. 2007, 126, 371−377. (2) Rack, P. D.; Peak, J. D.; Melcher, C. L.; Fitz-Gerald, J. M. Scanning Electron and Cathodoluminescence Imaging of Thin Film Lu2SiO5:Ce Scintillating Materials. Appl. Phys. Lett. 2007, 91, 244102. (3) Liu, Y.; Zhang, X.; Hao, Z.; Wang, X.; Zhang, J. Tunable FullColor-Emitting Ca3Sc2Si3O12: Ce3+, Mn2+ Phosphor via Charge Compensation and Energy Transfer. Chem. Commun. 2011, 47, 10677−10679. (4) George, N. C.; Pell, A. J.; Dantelle, G.; Page, K.; Llobet, A.; Balasubramanian, M.; Pintacuda, G.; Chmelka, B. F.; Seshadri, R. Local Environments of Dilute Activator Ions in the Solid-State Lighting Phosphor Y3‑xCexAl5O12. Chem. Mater. 2013, 25, 3979−3995. (5) Ding, X. M.; Liang, H. B.; Hou, D. J.; Su, Q.; Dorenbos, P.; Sun, S. S.; Tao, Y. Ultraviolet-Vacuum Ultraviolet Photoluminescence and X Ray Radioluminescence of Ce3+-Doped Ba3MgSi2O8. J. Appl. Phys. 2011, 110, 113522. (6) Zhou, L.; Liang, H. B.; Tanner, P. A.; Zhang, S.; Hou, D. J.; Liu, C. M.; Tao, Y.; Huang, Y.; Li, L. Luminescence, Cathodoluminescence and Ce3+→Eu2+ Energy Transfer and Emission Enhancement in the Sr5(PO4)3Cl: Ce3+, Eu2+ Phosphor. J. Mater. Chem. C 2013, 1, 7155− 7165. (7) Xia, Z. G.; Ma, C. G.; Molokeev, M. S.; Liu, Q. L.; Rickert, K.; Poeppelmeier, K. R. Chemical Unit Cosubstitution and Tuning of Photoluminescence in the Ca 2 (Al 1‑x Mg x ) (Al 1‑x Si 1+x )O 7 : Eu 2+ Phosphor. J. Am. Chem. Soc. 2015, 137, 12494−12497. (8) Xia, Z. G.; Molokeev, M. S.; Im, W. B.; Unithrattil, S.; Liu, Q. L. Crystal Structure and Photoluminescence Evolution of La5(Si2+xB1‑x)(O13‑xNx): Ce3+ Solid Solution Phosphors. J. Phys. Chem. C 2015, 119, 9488−9495. (9) Sugiyama, K.; Takeuchi, Y. Unusual Thermal Expansion of a B-O Bond in the Structure of Danburite CaB2Si2O8. Z. Kristallogr. 1985, 173, 293−304. (10) Berger, T.; Range, K.-J. Hochdrucksynthese und Strukturverfeinerung von Strontium-Danburit, SrB2Si2O8. Z. Naturforsch., B: Chem. Sci. 1996, 51, 172−174. (11) Pautov, L. A.; Agakhanov, A. A.; Sokolova, E.; Hawthorne, F. C. Maleevite, BaB2Si2O8, and Pekovite, SrB2Si2O8, New Mineral Species from the Dara-i-Pioz Alkaline Massif, Northern Tajikistan: Description and Crystal Structure. Can. Mineral. 2004, 42, 107−119. (12) Verstegen, J. M. P. J.; Ter Vrugt, J. W.; Wanmaker, W. L. Luminescence of Eu2+-Activated SrB2Si2O8. J. Inorg. Nucl. Chem. 1972, 34, 3588−3589. (13) Saradhi, M. P.; Boudin, S.; Varadaraju, U. V.; Raveau, B. A New BaB2Si2O8: Eu2+/Eu3+, Tb3+ Phosphor-Synthesis and Photoluminescence Properties. J. Solid State Chem. 2010, 183, 2496−2500. (14) Downs, J. W.; Swope, R. J. The Laplaclan of the Electron Density and the Electrostatic Potential of Danburite, CaB2Si2O8. J. Phys. Chem. 1992, 96, 4834−4840. (15) Gaft, M. L.; Gorobets, B. S. Luminescence of Bivalent RareEarth Ions in Natural Borosilicates. J. Appl. Spectrosc. 1979, 31, 1488− 1491. (16) Dorenbos, P. The 5d Level Position of the Trivalent Lanthanides in Inorganic Compounds. J. Lumin. 2000, 91, 155−176. (17) van der Kolk, E.; Dorenbos, P.; van Eijk, C. W. E.; Vink, A. P.; Fouassier, C.; Guillen, F. VUV Excitation and 1S0 Emission of Pr3+ in 579

DOI: 10.1021/acs.jpcc.5b10355 J. Phys. Chem. C 2016, 120, 569−580

Article

The Journal of Physical Chemistry C (37) Birkel, A.; Denault, K. A.; George, N. C.; Doll, C. E.; Héry, B.; Mikhailovsky, A. A.; Birkel, C. S.; Hong, B.-C.; Seshadri, R. Rapid Microwave Preparation of Highly Efficient Ce3+-Substituted Garnet Phosphors for Solid State White Lighting. Chem. Mater. 2012, 24, 1198−1204. (38) Ueda, J.; Dorenbos, P.; Bos, A. J. J.; Meijerink, A.; Tanabe, S. Insight into the Thermal Quenching Mechanism for Y3Al5O12: Ce3+ through Thermoluminescence Excitation Spectroscopy. J. Phys. Chem. C 2015, 119, 25003−25008. (39) Wang, L.; Zhang, X.; Hao, Z. D.; Luo, Y. S.; Zhang, J. H.; Wang, X. J. Inerionic Energy Transfer in Y3Al5O12: Ce3+, Pr3+ Phosphor. J. Appl. Phys. 2010, 108, 093515. (40) Shalapska, T.; Stryganyuk, G.; Gektin, A.; Demchenko, P.; Voloshinovskii, A.; Dorenbos, P. Crystal Structure and Luminescence Properties of LiYP4O12: Ce3+ Phosphor. J. Phys.: Condens. Matter 2010, 22, 485503. (41) Zhai, D. Y.; Ning, L. X.; Huang, Y. C.; Liu, G. K. Ce-O Covalence in Silicate Oxyapatites and Its Influence on Luminescence Dynamics. J. Phys. Chem. C 2014, 118, 16051−16059. (42) Bachmann, V.; Ronda, C.; Oeckler, O.; Schnick, W.; Meijerink, A. Color Point Tuning for (Sr, Ca, Ba)Si2O2N2: Eu2+ for White Light LEDs. Chem. Mater. 2009, 21, 316−325. (43) Ogieglo, J. M.; Zych, A.; Justel, T.; Meijerink, A.; Ronda, C. R. Luminescence and Energy Transfer in Lu3Al5O12 Scintillators coDoped with Ce3+ and Pr3+. Opt. Mater. 2013, 35, 322−331. (44) Henderson, B.; Imbusch, G. F. Optical Spectroscopy of Inorganic Solids; Clarendon Press: Oxford, U.K., 1989.

580

DOI: 10.1021/acs.jpcc.5b10355 J. Phys. Chem. C 2016, 120, 569−580