A novel apatite, Lu₅(SiO₄)₃N:(Ce,Tb), phosphor material: synthesis, structure and applications for NUV-LEDs

Abstract

The lutetium containing nitride apatite Lu₅(SiO₄)₃N was prepared by a solid state reaction at high temperature for the first time. Rietveld refinement indicated that the Lu₅(SiO₄)₃N compound has a hexagonal space group of P6₃/m with cell parameters a = b = 9.700 Å and c = 7.238 Å. Additionally, the results revealed that there are two distinct lutetium sites in the Lu₅(SiO₄)₃N host lattice, i.e. a Lu(1) site with nine coordination (Wyckoff site 4f) and a Lu(2) site with seven coordination (Wyckoff site 6h). Furthermore, the ratio of the number of Lu atoms in Lu(1) and Lu(2) sites is 3 : 2. The band gap for Lu₅(SiO₄)₃N was determined to be 4.12 eV based on the density functional theory (DFT). In the Ce³⁺ doped Lu₅(SiO₄)₃N:0.03Ce³⁺ compound, the emission peak centered at 462 nm was observed with the Commission International de I'Eclairage (CIE) coordinates of (0.148, 0.184), indicating blue-emission. Remarkably, in Ce³⁺ and Tb³⁺ co-doped Lu_4.97−y(SiO₄)₃N:0.03Ce³⁺,yTb³⁺ compounds, the color-tunability was observed with increasing Tb³⁺ co-doping rate on moving from blue at Tb³⁺ = 0.00 to green at Tb = 0.09, due to the energy transfer from Ce³⁺ to Tb³⁺ ions being matched well with the decay curve results. Under the excitation at 359 nm, the absolute quantum efficiency (QE) for Lu₅(SiO₄)₃N:0.03Ce³⁺ was determined to be 42.13%. This phosphor material could be a platform for modeling a new phosphor and application in the solid-state lighting field.

---

1. Introduction

Apatite compounds have been investigated as effective luminescent host materials in recent years because of their physical and chemical stability. In the past decades, they have been widely used in lighting and display areas.^1–4 Apatite belongs to the hexagonal crystal system represented generally as M₁₀(XO₄)₆Z₂ (M = Ca²⁺, Sr²⁺, Ba²⁺, K⁺, Na⁺, Ce³⁺, La³⁺, Y³⁺, etc., X = P⁵⁺, Si⁴⁺, S⁶⁺, V⁵⁺, etc., and Z = O²⁻, F⁻, Cl⁻, Br⁻, OH⁻, S²⁻, etc.).^5,6 In general, there are two kinds of non-equivalent crystallographic sites in apatite compounds, that is, a nine-coordinated 4f site M(I) with C₃ point symmetry and a seven-coordinated 6h site M(II) with C_S point symmetry, respectively.^7,8 Thus, it is interesting and significant to investigate the relationship between the structure (substitution) and the luminescence properties of an apatite structure. Accordingly, many efforts have been made to obtain novel phosphors on the basis of different chemical compositions, such as Sr₁₀(PO₄)₆O:Eu²⁺,⁹ Sr₃GdNa(PO₄)₃F:Eu²⁺,Mn²⁺,¹⁰ and Ca₉Mg(PO₄)₆F₂:Eu²⁺,Mn²⁺.¹¹

Meanwhile, the wavelength of excitation and emission bands for phosphors with an oxonitridosilicate structure is even longer than that of other phosphors because of their high covalency and strong crystal fields. (Si[O/N]₄) units are stacked together by sharing their corners and edges forming condensed (Si[O/N]₄) frameworks, making oxonitridosilicates excellent hosts for rare earth ions. Therefore, oxonitridosilicates have been used as hosts to obtain phosphors, such as: Y₁₀(Si₆O₂₂N₂)O₂:Ce³⁺,Mn²⁺,¹² Ca₃Si₂O₄N₂:Eu²⁺,¹³ and Y₄Si₂O₇N₂:Eu²⁺.¹⁴ Besides, the luminescence properties of Ce³⁺ ions can be strongly affected by their surrounding crystal field environment, due to their 5d electrons being unshielded from the crystal field by the 5s and 5p electrons in the excited state. Ce³⁺ doped oxonitridosilicates have been widely investigated, such as Ca₃Si₂O₄N₂:Ce³⁺,¹⁵ Y₅(SiO₄)₃N:Ce³⁺,¹⁶ Sr₃Si₂O₄N₂:Ce³⁺,¹⁷ and Gd₅Si₃O₁₂N:Ce³⁺.¹⁸ Among oxonitridosilicates, Lu–Si–O–N quaternary system compounds have been investigated in the previous literature, such as Lu₄Si₂O₇N₂.^19,20 However, to the best of our knowledge, the Lu₅(SiO₄)₃N compound (oxonitridosilicate) has not been identified and reported before.

In light of the above, a novel oxonitridosilicate Lu₅(SiO₄)₃N with an apatite structure was obtained for the first time under 0.5 MPa atmospheric pressure. The crystal structure properties of Lu₅(SiO₄)₃N were studied in detail. Rietveld refinement indicated that the Lu₅(SiO₄)₃N compound has a hexagonal space group of P6₃/m with cell parameters a = b = 9.700 Å and c = 7.238 Å. Two distinct lutetium sites in the Lu₅(SiO₄)₃N host lattice, i.e. a Lu(1) site with nine coordination (Wyckoff site 4f) and a Lu(2) site with seven coordination (Wyckoff site 6h), were revealed, and the ratio of the number of Lu atoms in Lu(1) and Lu(2) sites is 3 : 2. The band gap for Lu₅(SiO₄)₃N was determined to be 4.12 eV based on the density functional theory (DFT). In addition, the luminescence properties and energy transfer behaviors between Ce³⁺ and Tb³⁺ were investigated systematically. The results indicated that the Lu₅(SiO₄)₃N phosphor material could be a platform for modeling a new phosphor and application in the solid-state lighting field.

2. Experimental procedure

2.1 Materials and synthesis

Lu₅(SiO₄)₃N hosts and Lu₅(SiO₄)₃N:Ce³⁺,Tb³⁺ phosphors were prepared by a high temperature solid-state reaction. SiO₂ (Aldrich, 99.9%), α-Si₃N₄ (Aldrich, 99.9%), Lu₂O₃ (Aldrich, 99.995%), CeO₂ (Aldrich, 99.995%), and Tb₄O₇ (Aldrich, 99.995%) were weighed and used as starting materials according to the given composition of Lu₅(SiO₄)₃N:Ce³⁺,Tb³⁺. The raw materials were fully mixed and ground in an agate mortar, and the mixture was placed into an alumina crucible and was heated at 1500 °C for 5 h in a reducing atmosphere with flowing gas (10% H₂ + 90% N₂) and 0.5 MPa atmospheric pressure. Eventually, the obtained samples were cooled to room temperature and ground again into powder for subsequent measurements.

2.2 Characterization

Phase structures of the as-prepared samples were checked using an X-ray powder diffractometer (D/max-rA 12kw, Japan) with Cu Kα radiation (λ = 1.5418 Å) from 10° to 70° (2θ). The XRD data for the Rietveld analysis were collected in a step scanning mode with a step size of 0.02° and 3 s counting time per step from 10° to 75° (2θ). Rietveld refinement was performed using the computer software: general structure analysis system (GSAS).²¹ Computations were performed within the density functional theory (DFT) framework, using the Vienna ab initio software package (VASP), within the generalized gradient approximation by Perdew–Burke–Ernzher (GGA-PBE) and projector augmented wave (PAW) pseudopotentials. A cutoff energy of 520 eV and an appropriate k-point grid were chosen to ensure that total energies converged within 10⁻⁴ eV per formula unit. The supercell of Lu₅(SiO₄)₃N used in this study includes 42 atoms. When the structures were optimized, all atoms were allowed to relax. Since the optimized lattices of the systems were determined using several structure parameters, it is time consuming to get the theoretical value of these parameters through the traditional DFT calculation. In this work, the structures were optimized until all residual forces were below 0.001 eV Å⁻¹. Brillouin-zone integrations were performed using Monkhorst–Pack grids of special points with a mesh (2 × 2 × 3).²² The room-temperature excitation and emission spectra were measured on a Hitachi F-4500 spectrophotometer with a 150 W Xe lamp as the excitation lamp. The luminescence decay curves were obtained on a spectro-fluorometer (HORIBA JOBIN YVON FL3-21), and the 370 nm pulse laser radiation (370 nm Nano LED, model number 08254) was used as the excitation source. The particle morphology and microstructure of the Lu₅(SiO₄)₃N and Lu_4.97(SiO₄)₃N:0.03Ce³⁺ samples were examined by scanning electron microscopy (SEM) using a HitachiS-520 instrument with an accelerating voltage of 25 kV. The results of elemental analysis were gotten from an energy dispersive X-ray spectrometer (EDS) attached to the scanning electron microscope. The quantum efficiency (QE) was obtained using an absolute photoluminescence quantum yield measurement system (FluoroLog-3-2-iHR320, Horiba Jobin Yvon) with a F-3018 integral sphere at room temperature.

3. Results and discussion

3.1 Insight into Lu₅(SiO₄)₃N

Fig. 1(a) shows the Rietveld structural refinement of the XRD pattern of the Lu₅(SiO₄)₃N host using the GSAS program. As depicted in Fig. 1(a), the red solid line, black square points, green short vertical and blue lines represent the calculated pattern, the observed pattern from experiment, the Bragg positions of the calculated pattern, and the difference between the observed and calculated patterns, respectively. It is obvious that almost all the peaks of Lu₅(SiO₄)₃N can be indexed by the hexagonal cell (P6₃/m). In addition, Lu₅(SiO₄)₃N has similar XRD patterns to La₅(SiO₄)₃N (apatite structure, ICSD no. 91850), which have been widely investigated.^23–25 Therefore, La₅(SiO₄)₃N was taken as a starting model structure for Rietveld refinement. The refinement was stable and gives low R-factors (Fig. 1(a) and Table 1), and the final refinement residual factors are R_wp = 8.88%, R_p = 5.99%, and χ² = 1.63. Besides, the refinement results further confirm that the single-phase compound Lu₅(SiO₄)₃N can be indexed in the hexagonal space group of P6₃/m with cell parameters a = b = 9.7005 Å, c = 7.2383 Å and V = 589.87 ų. Furthermore, the fractional atomic coordinates, isotropic displacement parameters (Ų), as well as the main bond lengths (Å) for Lu₅(SiO₄)₃N are illustrated in Tables 2 and 3. According to Table 2, O1 is located on site 2a with −6 site symmetry, while Si1, O2, and O3 are located on the mirror planes with site symmetry m (site 6h). O4 and N4 are on a general position (site 12i). Based on the results from Table 3, the average bond length of Lu(1)–O/N can be calculated to be 2.706 Å, and 2.543 Å for Lu(2) sites. Besides, the crystallographic information files (CIFs) of Lu₅(SiO₄)₃N are also illustrated in the ESI.†Fig. 1(b) illustrates the crystal structure of Lu₅(SiO₄)₃N viewed along the c-axis, and the coordination environment of cation sites for activators was also given. In the crystal lattice of this compound, Si atoms are tetrahedrally coordinated to form [SiO₄] groups, which are isolated from each other. Lu₅(SiO₄)₃N contains two non-equivalent lutetium sites, where Lu(1) is located on a 3-fold axis (Wyckoff symbol 4f), while Lu(2) is located on a mirror plane (site 6h). Lu(1) is nine coordinated by O/N atoms forming a tricapped trigonal-prismatic geometry, but Lu(2) is seven coordinated by O/N atoms forming an irregular polyhedron, respectively. Furthermore, the ratio of the number of Lu atoms in the two sites is 3 : 2.

Fig. 1 Powder XRD patterns for the Rietveld refinement analysis of Lu₅(SiO₄)₃N. Solid red lines are calculated intensities, the black square points are the observed intensities, and blue solid lines stand for the difference between the observed and the calculated intensities, and short green vertical lines below the profiles show the position of Bragg reflections of the calculated pattern (a); crystal structure diagrams of Lu₅(SiO₄)₃N viewed in the c-direction and the coordination environment of cation sites for activators (b); SEM image of Lu₅(SiO₄)₃N (c); calculated band structure of Lu₅(SiO₄)₃N (d); electronic densities of states for Lu₅(SiO₄)₃N, both total (TDOS) and partial (PDOS) onto the atomic orbital (e).

Table 1 Refinement parameters for Lu₅(SiO₄)₃N obtained from the Rietveld refinement using X-ray powder diffraction data at room temperature

Sample	Lu5(SiO4)3N
Space group	P63/m(176)
a, Å	9.7005(17)
c, Å	7.2383(3)
V, Å^3	589.87(4)
2θ-interval, °	10–75
R wp, %	8.88
R p, %	5.99
χ ^2	1.63

---

Table 2 Structural parameters for Lu₅(SiO₄)₃N obtained from the Rietveld refinement using X-ray powder diffraction data at room temperature

	Wyck	x	y	z	U (Å^2)	Occupation
Lu1	4f	0.2310(8)	−0.0153(9)	1/4	0.0158	1
Lu2	6h	1/3	2/3	0.0007(4)	0.0249	1
Si1	6h	0.3932(5)	0.3788(1)	1/4	0.0017	1
O1	2a	0	0	1/4	0.0568	1
O2	6h	0.3308(8)	0.4682(7)	1/4	0.0143	1
O3	6h	0.5871(9)	0.4731(6)	1/4	0.0014	1
O4	12i	0.3443(2)	0.2570(5)	0.0769(2)	0.0027	5/6
N4	12i	0.3443(2)	0.2570(5)	0.0769(2)	0.0027	1/6

---

Table 3 Main bond lengths (Å) of Lu₅(SiO₄)₃N from Rietveld structure analysis, and the symmetry codes of the matching O/N element

Bond	Lengths (Å)	Symmetry codes	Bond	Lengths (Å)	Symmetry codes
Lu2–O1	2.3198(1)	x, y, z	Lu1–O3	2.6027(1)	x − y, x, −0.5 + z
Lu2–O2	2.7223(1)	−x + y, −x, z	Lu1–O3	2.6027(1)	1 − x, 1 − y, −0.5 + z
Lu2–O3	2.4911(1)	1 − y, x − y, z	Lu1–O3	2.6027(1)	y, 1 − x + y, −0.5 + z
Lu2–O4/N4	2.6187(1)	x, y, z	Lu1–O4/N4	2.8857(1)	1 − x, 1 − y, −z
Lu2–O4/N4	2.5149(1)	y, −x + y, 0.5 + z	Lu1–O4/N4	2.8857(1)	y, 1 − x + y, −z
Lu2–O4/N4	2.5149(1)	y, −x + y, −z	Lu1–O4/N4	2.8857(1)	x − y, x, −z
Lu2–O4/N4	2.6187(1)	x, y, 0.5 − z	Si1–O2	1.2823(1)	x, y, z
Lu1–O2	2.6294(1)	x, y, z	Si1–O3	1.6295(1)	x, y, z
Lu1–O2	2.6295(1)	1 − y, 1 + x − y, z	Si1–O4/N4	1.6215(0)	x, y, z
Lu1–O2	2.6294(1)	−x + y, 1 − x, z	Si1–O4/N4	1.6215(0)	x, y, 0.5 − z

---

The morphology of the Lu₅(SiO₄)₃N host prepared via solid state reaction is shown in Fig. 1(c). The SEM result shows that particles of the as-prepared sample are not uniform, which was caused by the agglomeration during sample sintering. Additionally, the SEM image indicated that the sample has a well-crystallized micropore with the diameter of about 3–6 μm. In order to confirm the chemical composition of the Lu₅(SiO₄)₃N, the as-prepared sample was analysed by energy dispersive spectroscopy (EDS) analysis. Table 4 shows the results of EDS elemental analysis (wt%) within the rectangle area in Fig. 1(c). As illustrated in Table 4, Lu, Si, O, and N exist in the host Lu₅(SiO₄)₃N. Besides, the weight ratio of each element in Lu₅(SiO₄)₃N is Lu 70.80%, Si 7.06%, O 19.21% and N 2.93% according to the EDS results in Table 4.

Table 4 The results of energy dispersive spectroscopy (EDS) elemental analysis (wt%) of the Lu₅(SiO₄)₃N sample within the rectangle area in Fig. 1(c) and Lu₅(SiO₄)₃N:0.03Ce³⁺ within the rectangle area in Fig. 3(d)

Lu5(SiO4)3N			Lu4.97(SiO4)3N:0.03Ce^3+
Element	Weight (%)	Atomic (%)	Element	Weight (%)	Atomic (%)
Lu	70.80	23.13	Lu	70.63	23.07
Si	7.06	12.20	Ce	0.15	0.22
O	19.21	56.32	Si	7.24	12.51
N	2.93	8.35	O	18.96	55.59
			N	3.02	8.61
Total	100.00	100.00	Total	100.00	100.00

---

The band structure of pure Lu₅(SiO₄)₃N calculated by using VASP is presented in Fig. 1(d). The results reveal that the valence band maximum (VBM) and the conduction band minimum (CBM) are at different points, indicating that Lu₅(SiO₄)₃N has an indirect broad band gap. In addition, the band gap was determined to be approximately 4.12 eV. As we know, electronic densities of states analysis were considered an effective way to assign the band structure and analyze the bonding interactions in the crystal. Therefore, the electronic densities of states of Lu₅(SiO₄)₃N, both total (TDOS) and partial (PDOS), onto the atomic orbital were performed as illustrated in Fig. 1(e). It can be seen that the valence bands in the range of −19.89 to −17.35 eV originate from O, Si, and Lu. The peaks at about −15.93 eV correspond to Lu and O state bands. Furthermore, the valence bands close to the Fermi level (−0.61 to 0 eV) are dominated by the states of O and N. Finally, the conduction band (below 6.5 eV) is dominated by states of Lu. Therefore, the band gap of Lu₅(SiO₄)₃N is determined by O/N and Lu atoms.

3.2 Luminescence properties of Lu₅(SiO₄)₃N:Ce³⁺ phosphor

Fig. 2 depicts the photoluminescence excitation (PLE) and emission (PL) spectra of Lu_5−x(SiO₄)₃N:xCe³⁺ (a), and the relative emission intensity and emission peak maximum (nm) as a function of Ce³⁺ concentration (b). Under the excitation at 359 nm, Lu_5−x(SiO₄)₃N:xCe³⁺ exhibit a broad blue emission band peaking at about 460 nm. The PLE spectrum monitored at 462 nm shows a broad absorption from 245 to 400 nm, including two obvious broad bands centered at 314 and 359 nm attributed to the 4f → 5d transition of Ce³⁺.^26,27 As depicted in Fig. 2(a) and (b), the emission intensity of Ce³⁺ increased firstly with the Ce³⁺ concentration x, and achieved the maximum value at x = 0.03 (mol), and then it decreased dramatically with further increase of concentration, which may be attributed to the concentration quenching effect. Therefore, the optimum doping concentration x of Ce³⁺ is 0.03. Moreover, the emission peak shows a slight red shift from 454 to 464 nm with increasing Ce³⁺ concentration, which may be due to the variation of crystal field strength. In addition, the correlation between the crystal field splitting (D_q) and the shape and the size of the polyhedron is given by the following equation:²⁸where D_q refers to the energy level separation, z is the charge or valence of the anion ligand, e stands for the charge of an electron, r represents the radius of the d wave function, and R represents the bond length. Assuming that z, e and r are the same for the d(Ce–O)-orbital and the d(Ce–N)-orbital, it can be considered that D_q is approximately proportional to 1/R⁵. Thus, if R is smaller, D_q has a larger value, and finally results in the red-shift of the emission peak. As the average bond length of the Lu1 site (2.706 Å, CN = 9) is somewhat longer than that of the Lu2 site (2.543 Å, CN = 7), it is presumed that the larger Ce³⁺ ion relative to the Lu³⁺ ion preferably enters the Lu1 site (more space with CN = 9) rather than the Lu2 site (less space with CN = 7). It is well-known that there are several factors that affect the extent of splitting of the d orbitals by ligands. Among several factors, the number and geometry of ligands have an important effect on the magnitude of the crystal field splitting (D_q). In comparison between an octahedral complex (Oh, CN = 6) and a tetrahedral complex (Td, CN = 4), it can be shown that for a point-charge model:²⁹

10D_q(Td) = 4/9[10D_q(Oh)] (2)

According to eqn (2), it is evident that the larger Ce³⁺ ion relative to the Lu³⁺ ion preferably enters the Lu1 site (more space, CN = 9) and the energy splitting of 5d orbitals of Ce³⁺ ions is larger than that in the Lu2 site (less space, CN = 7), implying that the emission band is gradually shifted to the red wavelength region with increasing Ce³⁺ concentration in Lu_5−xCe_x(SiO₄)₃N phosphors.

Fig. 2 The PLE and PL spectra of Lu_5−x(SiO₄)₃N:xCe³⁺ (a); the relative emission intensity and emission peak maximum (nm) as a function of Ce³⁺ concentration (b); PL spectrum of Lu_4.97Ce_0.03(SiO₄)₃N fitted by Gaussian functions under the excitation at 359 nm (c).

Furthermore, the PL spectrum of Lu_4.97Ce_0.03(SiO₄)₃N can be divided into two dotted bands centered at about 456 and 495 nm by using Gaussian fitting, as depicted in Fig. 2(c). In general, the emission of Ce³⁺ ions should contain two bands because of the spin–orbit splitting of the ground state (²F_7/2 and ²F_5/2 states) with an energy difference of about 2000 cm⁻¹.³⁰ However, the energy difference between 453 and 493 nm was calculated to be 1817 cm⁻¹. Thus, these components can be ascribed to the contributions of Ce³⁺ ions at two different luminescent centers, which is inconsistent with the structure of Lu₅(SiO₄)₃N. As far as we know, the emission position of Ce³⁺ ions is strongly affected by their local crystal environment, and the following equation can be used to determine the sites of Ce³⁺:³¹

in which E represents the position of the Ce³⁺ ion emission peak (cm⁻¹), Q is the energy position of the lower d-band edge for the free ions (50 000 cm⁻¹ for Ce³⁺), V stands for the valence of Ce³⁺ (V = 3 for Ce³⁺), n is the number of anions in the immediate shell around the Ce³⁺ ion, r is the effective radius of the host cations (La³⁺) replaced by the Ce³⁺ ion (Å), and E_a is the electron affinity of the atoms that form anions in eV, which is constant in the same host. The effective ionic radius of Lu³⁺ with nine-coordination is larger than that with seven-coordination. Therefore, the emission peaks at 453 nm and 555 nm are attributed to Ce³⁺ ions with nine-coordination and seven-coordination as shown in the inset of Fig. 2(c).

3.3 Crystal structure of Lu₅(SiO₄)₃N:Ce³⁺,Tb³⁺ phosphors

With 4f–4f forbidden transitions, Tb³⁺ can be effectively sensitized by Ce³⁺, which can make the emission color of phosphors tunable. Besides, phase purity is an important factor for phosphors to have stable luminescence properties. Thus, some Lu_5−x−y(SiO₄)₃N:xCe³⁺,yTb³⁺ phosphors were selected for XRD measurements. Fig. 3(a) shows the XRD patterns of Lu_4.995(SiO₄)₃N:0.005Ce³⁺, Lu_4.97(SiO₄)₃N:0.03Ce³⁺, Lu_4.96(SiO₄)₃N:0.03Ce³⁺,0.01Tb³⁺, Lu_4.92(SiO₄)₃N:0.03Ce³⁺,0.05Tb³⁺, Lu_4.88(SiO₄)₃N:0.03Ce³⁺,0.09Tb³⁺, as well as the simulated Lu₅(SiO₄)₃N obtained from the cell refinement. It is shown that all XRD peaks of the samples are consistent with the simulated data for Lu₅(SiO₄)₃N, indicating that all the samples are single-phase and the doped Ce³⁺ or Tb³⁺ ions were successfully introduced into the Lu₅(SiO₄)₃N host lattice. In order to better understand the phase structure and the purity of the Ce³⁺/Tb³⁺ doped phosphors, Rietveld refinements were taken for Lu_4.97Ce_0.03(SiO₄)₃N and Lu_4.90Ce_0.03Tb_0.07(SiO₄)₃N phosphors. The Ce³⁺ or Tb³⁺ ions were supposed to occupy both sites of Lu ions. The refinements were stable and gave low R-factors shown in Fig. 3(a) and (b), and Table 5. Besides, the lattice constants of Lu_4.97Ce_0.03(SiO₄)₃N and Lu_4.90Ce_0.03Tb_0.07(SiO₄)₃N are close to that of Lu₅(SiO₄)₃N, indicating that the doped Ce³⁺ or Tb³⁺ ions did not cause any significant changes in the crystal lattice of Lu₅(SiO₄)₃N.

Fig. 3 The XRD patterns of Lu_4.995(SiO₄)₃N:0.005Ce³⁺, Lu_4.97(SiO₄)₃N:0.03Ce³⁺, Lu_4.96(SiO₄)₃N:0.03Ce³⁺,0.01Tb³⁺, Lu_4.92(SiO₄)₃N:0.03Ce³⁺,0.05Tb³⁺, Lu_4.88(SiO₄)₃N:0.03Ce³⁺,0.09Tb³⁺, as well as the simulated Lu₅(SiO₄)₃N obtained from the structure refinement (a); powder XRD patterns for Rietveld structure analysis of the selected Lu_4.97Ce_0.03(SiO₄)₃N (b) and Lu_4.90Ce_0.03Tb_0.07(SiO₄)₃N (c) phosphors based on the simulated Lu₅(SiO₄)₃N phase model. Solid red lines are calculated intensities, the black circles are the observed intensities, and blue solid lines for the difference between the observed and the calculated intensities, and short green vertical lines below the profiles show the position of Bragg reflections of the calculated pattern; SEM image of Lu_4.97(SiO₄)₃N:0.03Ce³⁺ (d).

Table 5 Refinement parameters for Lu_4.97Ce_0.03(SiO₄)₃N and Lu_4.90Ce_0.03Tb_0.07(SiO₄)₃N from the Rietveld refinement using X-ray powder diffraction data at room temperature

Samples	Lu4.97Ce0.03Si3O12N	Lu4.90Ce0.03Tb0.07Si3O12N
Space group	P63/m(176)	P63/m(176)
a, Å	9.6989(33)	9.69885(25)
c, Å	7.2489(27)	7.2592(03)
V, Å^3	590.35(5)	591.99(24)
2θ-interval, °	10–75	10–75
R wp, %	9.56	9.40
R p, %	6.73	7.09
χ ^2	1.85	2.43

---

The SEM image of Lu_4.97(SiO₄)₃N:0.03Ce³⁺ is demonstrated in Fig. 3(d), and the phosphor was also analyzed by EDS (Table 4). It is found that the particles have a smooth morphology with diameters of the particles being between 1 and 6 μm. In addition, the EDS results indicated the presence of Ce³⁺ in the sample.

3.4 Energy transfer investigations in Lu_4.97−y(SiO₄)₃N:0.03Ce³⁺,yTb³⁺

Fig. 4 presents the PLE and PL spectra of Lu_4.97(SiO₄)₃N:0.03Ce³⁺ (a), Lu_4.97(SiO₄)₃N:0.03Tb³⁺ (b), and Lu_4.94(SiO₄)₃N:0.03Ce³⁺,0.03Tb³⁺ (c). The PLE spectrum of Lu_4.97(SiO₄)₃N:0.03Tb³⁺ presents a broad band from 225 to 314 nm with a maximum at 270 nm corresponding to the 4f⁸–4f⁷5d transition of Tb³⁺. As illustrated in the inset of Fig. 4(b), there are some weak lines at 341, 353, and 376 nm, corresponding to the transitions from the ⁷F₆ ground state to the excited states (⁵D₁₀, ⁵G_J, and ⁵L₁₀ levels) of Tb³⁺ ions. Under 270 nm excitation, Lu_4.97(SiO₄)₃N:0.03Tb³⁺ can emit strong green light, and the obtained emission spectrum gives typical characteristic transitions of Tb³⁺ ions, which are situated at about 417 nm (⁵D₃–⁷F₅), 440 nm (⁵D₃–⁷F₄), 491 nm (⁵D₄–⁷F₆), 547 nm (⁵D₄–⁷F₅), 589 nm (⁵D₄–⁷F₄), and 623 nm (⁵D₄–⁷F₃).^32,33 As shown in Fig. 4(c), it includes three peaks located at 270, 314, and 359 nm in the excitation spectrum of Lu_4.94(SiO₄)₃N:0.03Ce³⁺,0.03Tb³⁺ phosphor monitored with 462 nm. According to the above discussions, the peak at 265 nm belongs to the 4f⁸–4f⁷5d¹ transition of Tb³⁺, while the peaks at 306 and 365 nm correspond to the 4f–5d transition of Ce³⁺. Upon excitation at 359 nm, both the emission of Ce³⁺ and Tb³⁺ can be observed in the PL spectrum of the co-doped Lu_4.94(SiO₄)₃N:0.03Ce³⁺,0.03Tb³⁺ sample as depicted in Fig. 4(c). Based on the Dexter theory, energy transfer requires a spectral overlap between the donor emission band and the acceptor excitation band.³⁴ The inset of Fig. 4(c) shows the PLE spectrum of Lu_4.97(SiO₄)₃N:0.03Tb³⁺ and the PL spectrum of Lu_4.97(SiO₄)₃N:0.03Ce³⁺ ranging from 380 to 400 nm. It is clear that there is a spectral overlap in the range from 386 to 396 nm in combination with the emission band of Ce³⁺ (donor) and the excitation of Tb³⁺ (acceptor). Therefore, the energy transfer from Ce³⁺ to Tb³⁺ may occur in the Lu₅(SiO₄)₃N host.

Fig. 4 The PL excitation and emission spectra of Lu_4.97(SiO₄)₃N:0.03Ce³⁺ (a); the PL excitation and emission spectra of Lu_4.97(SiO₄)₃N:0.03Tb³⁺ (b), and the inset of (b) shows PLE spectra of Lu_4.97(SiO₄)₃N:0.03Tb³⁺ ranging from 330 nm to 390 nm; the PL excitation and emission spectra of Lu_4.94(SiO₄)₃N:0.03Ce³⁺,0.03Tb³⁺ (c), and the inset of (c) illustrates the PLE spectrum of Lu_4.97(SiO₄)₃N:0.03Tb³⁺ and the PL spectrum of Lu_4.97(SiO₄)₃N:0.03Ce³⁺ in the range from 380 to 400 nm; the emission spectra of Lu_4.97−y(SiO₄)₃N:0.03Ce³⁺,yTb³⁺ (y = 0, 0.01, 0.03, 0.05, 0.07, and 0.09) phosphors with an excitation wavelength of 359 nm (d); relative emission intensities of Ce³⁺ and Tb³⁺ ions in Lu_4.97−y(SiO₄)₃N:0.03Ce³⁺,yTb³⁺ (e).

As discussed above, the optimum doping concentration of Ce³⁺ is 0.03 (mol). In order to further investigate the energy transfer process between the Ce³⁺ and Tb³⁺ ions in the Lu₅(SiO₄)₃N host, a series of Lu_4.97−y(SiO₄)₃N:0.03Ce³⁺,yTb³⁺ (y = 0, 0.01, 0.03, 0.05, 0.07, and 0.09) phosphors were prepared. Fig. 4(d) presents the emission spectra of Lu_4.97−y(SiO₄)₃N:0.03Ce³⁺,yTb³⁺ phosphors with an excitation at 359 nm. It is clear that both the emission peaks of Ce³⁺ and Tb³⁺ can be observed in the as-prepared Lu_4.97−y(SiO₄)₃N:0.03Ce³⁺,yTb³⁺ phosphors except for Lu_4.97(SiO₄)₃N:0.03Ce³⁺. As illustrated in Fig. 4(e), the relative emission intensities of the Ce³⁺ ions decreased remarkably, whereas the relative emission intensities of Tb³⁺ increased first with the Tb³⁺ concentration. Then, the emission intensity of Tb³⁺ in Lu_4.97(SiO₄)₃N:0.03Ce³⁺,0.09Tb³⁺ decreased slightly with further increasing concentration of Tb³⁺, ascribed to the concentration quenching effect.

To further confirm the energy transfer between Ce³⁺ and Tb³⁺, the decay curves of Lu_4.97−y(SiO₄)₃N:0.03Ce³⁺,yTb³⁺ (y = 0, 0.01, 0.03, and 0.05) were collected (λ_ex = 370 nm and λ_em = 462 nm), which are shown in Fig. 5(d). Generally, energy transfer from Ce³⁺ to Tb³⁺ belongs to multipolar interaction between donor and acceptor ions. According to Dexter theory, the energy transfer probability through multipolar interaction can be calculated by eqn (4):³⁵

in which b and c are constant, P represents the energy transfer probability, τ_D is the decay time of the donor (activator Ce³⁺) emission, Q_A stands for the total absorption cross-section of the acceptor, and R is the distance between the donor (Ce³⁺) and the acceptor (Tb³⁺). Thus, the energy-transfer probability P is inversely proportional to the decay time τ_D. Additionally, the decay curves for all the phosphors can be well fitted with a second-order exponential decay mode using eqn (5):³⁶

I(t) = A₁ exp(−t/τ₁) + A₂ exp(−t/τ₂) (5)

where I(t) is the luminescence intensities at times t, A₁ and A₂ are fitting constants, and τ₁ and τ₂ stand for the decay times for the exponential components. Based on eqn (6), the average lifetimes (τ*) can be calculated using the parameters in eqn (5).³⁷

τ* = (A₁τ₁² + A₂τ₂²)/(A₁τ₁ + A₂τ₂) (6)

As illustrated in the inset of Fig. 5(d), all the R² values are higher than 98%, which indicated that the fitting is relatively good. Therefore, τ* values for Lu_4.97−y(SiO₄)₃N:0.03Ce³⁺,yTb³⁺ (y = 0, 0.01, 0.03, and 0.05) were determined to be 4.41, 3.74, 2.10, and 1.67 ns, respectively. Obviously, the decay time decreases gradually with the increase of Tb³⁺ concentration, which strongly demonstrates the energy transfer from Ce³⁺ to Tb³⁺.

Fig. 5 Dependence of I_S0/I_S of Ce³⁺ on C^6/3 (a); C^8/3 (b); C^10/3 (c); decay curves of Ce³⁺ emission in Lu_4.97−y(SiO₄)₃N:0.03Ce³⁺,yTb³⁺ (y = 0, 0.01, 0.03, and 0.05) phosphors under excitation at 370 nm monitored at 462 nm at room temperature, and the inset depicts the fitted lifetimes, fitting parameters, as well as the energy transfer efficiency calculated on the basis of lifetimes (d); the relative emission intensities centered at 462 nm (Ce³⁺) and 547 nm (Tb³⁺) of the selected sample as a function of temperature ranging from 25 °C to 300 °C (e).

In general, the energy transfer efficiency η_T from a sensitizer to an activator can also be expressed using the following equation by Paulose et al.:^38,39

η_T = 1 − τ_S/τ₀ (7)

where τ_S and τ₀ represent the luminescence lifetimes of the sensitizer Ce³⁺ ion in the presence and the absence of the activator Tb³⁺ ion, respectively. On the basis of the above calculated average values of the lifetimes (τ*), the energy transfer efficiency η_T can be calculated to be 15.19%, 52.39%, and 62.13% as given in Fig. 5(d). With the increase of Tb³⁺ concentration, the value of η_T increased gradually and reached the maximum of 62.13% when y equals to 0.05. Thus, an efficient energy transfer from Ce³⁺ to Tb³⁺ must have occurred in Lu_4.97−y(SiO₄)₃N:0.03Ce³⁺,yTb³⁺.

As we know, there are two main aspects that are responsible for the resonance energy transfer mechanism: one is the multipolar interaction and the other is the exchange interaction. Meanwhile, the critical distance between the sensitizer and the activator should be shorter than 5 Å for the exchange interaction. To further determine the energy transfer mechanism in Lu₅(SiO₄)₃N, the critical distance (R_C) between Ce³⁺ and Tb³⁺ ions was calculated using the concentration quenching method proposed by Blasse.^40,41

where x_c is the critical concentration, V corresponds to the volume of the unit cell, and N stands for the number of host cations in the unit cell. For the Lu₅(SiO₄)₃N host, N = 10, V is estimated to be 589.87 ų as shown in Table 1. Besides, the critical concentration is about 0.10 (mol) from the overall concentration of Ce³⁺ (0.03) and Tb³⁺ (0.07). Therefore, the critical distance R_C of energy transfer can be estimated to be about 10.405 Å (x_c = 0.10) on the basis of the above equation. The value of R_C is much larger than 5 Å, which indicated that the electric multipolar interaction can take place for the energy transfer between Ce³⁺ and Tb³⁺ in the Lu₅(SiO₄)₃N host lattice. In addition, the multipolar interaction can be determined by Dexter's energy-transfer expressions and Reisfeld's approximation given as follows:^42,43

η₀/η ∝ C^n/3 (9)

where η₀ and η stand for the luminescence quantum efficiency of Ce³⁺ in the absence and the presence of Tb³⁺, C corresponds to the total content of Ce³⁺ and Tb³⁺, and n = 6, 8, and 10, respectively. The value η₀/η is approximately calculated by the ratio of related luminescence intensities as (I₀/I_S). Besides, n = 6, 8, and 10 correspond to the electric multipolar interaction of the electric dipole–dipole, dipole–quadrupole, and quadrupole–quadrupole, respectively. The I₀/I_S − C^n/3 plots are further illustrated in Fig. 5(a)–(c). By comparing the values of the fitting factors R², the optimal linear relationship is obtained when n = 10 (R² = 0.9894). This indicates that the quadrupole–quadrupole interaction should be mainly responsible for the energy transfer from Ce³⁺ to Tb³⁺ ions in Lu_4.97−y(SiO₄)₃N:0.03Ce³⁺,yTb³⁺.

As we know, the thermal stability of phosphor is one of the most important technological parameters for its application. Therefore, Lu_4.90(SiO₄)₃N:0.03Ce³⁺,0.07Tb³⁺ was selected for PL measurements under 359 nm excitation at different temperatures ranging from 25 °C to 300 °C. The relative emission intensities centered at 462 nm (Ce³⁺) and 547 nm (Tb³⁺) of the selected sample as a function of temperature are shown in Fig. 5(e). The relative PL intensities centered at 462 nm (Ce³⁺) and 547 nm (Tb³⁺) decrease to 80% and 70% of the initial values (25 °C) with the temperature increasing to 150 °C, implying that the sample has high thermal stability.

3.5 CIE chromaticity diagram and quantum efficiency (QE) of Lu_4.97−y(SiO₄)₃N:0.03Ce³⁺,yTb³⁺

Fig. 6 depicts the CIE chromaticity diagram of Lu_4.97−y(SiO₄)₃N:0.03Ce³⁺,yTb³⁺ (y = 0, 0.01, 0.03, 0.05, 0.07, and 0.09) phosphors with different Tb³⁺ concentrations under 359 nm excitation calculated by PL spectroscopy, and the insets of Fig. 6 show the digital photos of the selected samples (y = 0, 0.03, 0.05, and 0.09) under 365 nm UV lamp excitation. The CIE coordinates of the Lu_4.97−y(SiO₄)₃N:0.03Ce³⁺,yTb³⁺ phosphors shifted from the blue (0.1480, 0.1842) to green (0.2249, 3713) region with Tb³⁺ concentration, as indicated in Table 6. Therefore, the emission color of the phosphors can be tuned by adjusting the doping content of Tb³⁺ ions. In addition, the quantum efficiency of phosphors is another important factor for their application in LEDs. Therefore, the absolute quantum efficiency (QE) of the Lu_4.97(SiO₄)₃N:0.03Ce³⁺ sample was investigated. Under the excitation at 359 nm, the absolute QE for the phosphor was determined to be 42.13%.

Fig. 6 The CIE chromaticity diagram of Lu_4.97−y(SiO₄)₃N:0.03Ce³⁺,yTb³⁺ (y = 0, 0.01, 0.03, 0.05, 0.07, and 0.09) phosphors under 359 nm excitation, and the insets of show the digital photos of the selected samples (y = 0, 0.03, 0.05, and 0.09) under 365 nm UV lamp excitation.

Table 6 CIE coordinates for Lu_4.97−y(SiO₄)₃N:0.03Ce³⁺,yTb³⁺ (y = 0, 0.01, 0.03, 0.05, 0.07, and 0.09) (λ_ex = 359 nm)

Samples	CIE (x, y)
y = 0	(0.1480, 0.1842)
y = 0.01	(0.1586, 0.1937)
y = 0.03	(0.1696, 0.2239)
y = 0.05	(0.1783, 0.2483)
y = 0.07	(0.2188, 0.3549)
y = 0.09	(0.2249, 0.3713)

---

4. Conclusions

In summary, novel single-phase Lu₅(SiO₄)₃N compounds and Lu₅(SiO₄)₃N:Ce³⁺,Tb³⁺ phosphors with an apatite structure were obtained for the first time via a high temperature solid state method with 0.5 MPa atmospheric pressure. The results indicated that Lu₅(SiO₄)₃N crystallized in the hexagonal space group of P6₃/m with cell parameters a = b = 9.7005 Å and c = 7.2383 Å. Two non-equivalent lutetium sites exist in Lu₅(SiO₄)₃N, that is, Lu(1) with nine coordination and Lu(2) with seven coordination. The ratio of the number of Lu atoms in Lu(1) and Lu(2) sites is 3 : 2. The band gap for Lu₅(SiO₄)₃N was 4.12 eV. In the Ce³⁺ doped Lu₅(SiO₄)₃N:0.03Ce³⁺ compound, the emission peak centered at 462 nm was observed with the Commission International de I'Eclairage (CIE) coordinates of (0.148, 0.184), indicating blue-emission. Remarkably, in Ce³⁺ and Tb³⁺ co-doped Lu_4.97−y(SiO₄)₃N:0.03Ce³⁺,yTb³⁺ compounds, the color-tunability was observed with increasing Tb³⁺ co-doping rate on moving from blue at Tb³⁺ = 0.00 to green at Tb = 0.09, due to the energy transfer from Ce³⁺ to Tb³⁺ ions being matched well with the decay curve results. The energy transfer process from the Ce³⁺ to Tb³⁺ ions occurs in the host lattice by the quadrupole–quadrupole interaction mechanism. The excitation spectra of the Lu₅(SiO₄)₃N:Ce³⁺,Tb³⁺ phosphors give a broad excitation band in the wavelength range from 235 nm to 435 nm. The absolute quantum efficiency (QE) for Lu₅(SiO₄)₃N:0.03Ce³⁺ was determined to be 42.13%. All the above results demonstrate that Lu₅(SiO₄)₃N could be a platform for modeling a new phosphor and application in the solid-state lighting field and that Lu₅(SiO₄)₃N:Ce³⁺,Tb³⁺ samples could be used as blue-green phosphors.

Acknowledgements

This work is supported by the National Natural Science Foundations of China (Grant No. 41172053) and Chinese Scholarship Council.

Notes and references

1. Y. Li, W. J. Liu, X. C. Wang, G. Zhu, C. Wang and Y. H. Wang, Dalton Trans., 2015, 44, 13196–13203.
2. D. Kim, D. Park, N. Oh, J. Kim, E. D. Jeong, S. J. Jim, S. Kim and J. C. Park, Inorg. Chem., 2015, 54, 1325–1336.
3. F. Lu, X. Song and Q. Liu, J. Alloys Compd., 2011, 509, 2099–2104.
4. L. K. Bharat, J. Y. Park and J. S. Yu, Chem. Eng. J., 2014, 240, 179–186.
5. M. F. Zhang, Y. J. Liang, S. Y. Xu, Y. L. Zhu, X. Y. Wu and S. Q. Liu, CrystEngComm, 2016, 18, 68–76.
6. Z. L. Dong, T. J. White, K. Sun, L. M. Wang and R. C. Ewing, J. Am. Chem. Soc., 2015, 88, 184–190.
7. E. N. Bulanov, J. X. Wang, A. V. Knyazev, T. White, M. E. Manyakina, T. Baikie, A. N. Lapshin and Z. L. Dong, Inorg. Chem., 2015, 54, 11356–11361.
8. K. Fukuda, T. Asaka, S. Hara, M. Oyabu, A. Berghout, E. Bechade, O. Masson, I. Julien and P. Thomas, Chem. Mater., 2013, 25, 2154–2162.
9. Q. F. Guo, L. B. Liao, L. F. Mei and H. K. Liu, J. Solid State Chem., 2015, 226, 107–113.
10. M. M. Jiao, Y. C. Jia, W. Lu, W. Z. Lv, Q. Zhao, B. Q. Shao and H. P. You, J. Mater. Chem. C, 2014, 2, 90–97.
11. K. Li, D. L. Geng, M. M. Shang, Y. Zhang, H. Z. Lian and J. Lin, J. Phys. Chem. C, 2014, 118, 11026–11034.
12. D. L. Geng, H. Z. Lian, M. M. Shang, Y. Zhang and J. Lin, Inorg. Chem., 2014, 53, 2230–2239.
13. Y. C. Chiu, T. J. Lee, W. R. Liu, Y. T. Yeh, S. M. Jang and R. S. Liu, Opt. Express, 2011, 19, 331–339.
14. G. Z. Chen, L. J. Yin, J. T. Dong, Y. Y. Feng, Y. Gao, W. D. He, Y. Jia and H. T. Hintzen, J. Am. Chem. Soc., 2016, 99, 183–190.
15. C. H. Huang, Y. C. Chiu and W. R. Liu, Eur. J. Inorg. Chem., 2014, 3674–3680.
16. J. W. H. van Krevel, H. T. Hintzen, R. Metselaar and A. Meijerink, J. Alloys Compd., 1998, 268, 272–277.
17. X. M. Wang, C. H. Wang, X. J. Kuang, R. Q. Zou, Y. X. Wang and X. P. Jing, Inorg. Chem., 2012, 51, 3540–3547.
18. F. C. Lu, L. J. Bai, Z. P. Yang and X. N. Han, Mater. Lett., 2015, 151, 9–11.
19. X. Xu, C. Cai, L. Y. Hao, Y. C. Wang and Q. X. Li, Mater. Chem. Phys., 2009, 118, 270–272.
20. J. Takahashi, H. Yamane, M. Shimad, Y. Yamamoto, N. Hirosaki, M. Mitomo, K. Oikawa, S. Torii and T. Kamiyama, J. Am. Chem. Soc., 2002, 85, 2072–2077.
21. A. C. Larson and R. B. Von Dreele, Los Alamos National Laboratory, Report LAUR: Los Alamos, 1986, 86–748.
22. G. Kresse and G. Furthmuller, Comput. Mater. Sci., 1996, 6, 15–50.
23. F. C. Lu and L. J. Bai, J. Am. Ceram. Soc., 2014, 97, 2397–2400.
24. B. Dierre, R. J. Xie, N. Hirosaki and T. Sekiguchi, J. Mater. Res., 2007, 22, 1933–1941.
25. S. Titeux, M. Gervais, P. Verdier and Y. Laurent, Mater. Sci. Forum, 2000, 325–326, 17–20.
26. M. Yang, H. P. You, K. Liu, Y. H. Zheng, N. Guo and H. J. Zhang, Inorg. Chem., 2010, 49, 4996–5002.
27. X. H. Gong, J. H. Huang, Y. J. Chen, Y. F. Lin, Z. D. Luo and Y. D. Huang, Inorg. Chem., 2014, 53, 6607–6614.
28. C. L. Zhao, Z. G. Xia and S. X. Yu, J. Mater. Chem. C, 2014, 2, 6032–6039.
29. J. E. Huheey, E. A. Keiter and R. L. Keiter, Inorganic Chemistry: Principles of Structure and Reactivity, Harper Collins College Publishers, New York, 1993.
30. H. B. Liang, H. H. Lin, G. B. Zhang, P. Dorenbos and Q. Su, J. Lumin., 2011, 131, 194–198.
31. L. G. Van Uitert, J. Lumin., 1984, 29, 1–9.
32. T. Grzyb, K. Kubasiewicz, A. Szczeszak and S. Lis, Dalton Trans., 2015, 44, 4063–4069.
33. L. Liu, Y. M. Liang, L. L. Li, L. C. Zou and S. C. Gan, CrystEngComm, 2015, 17, 7754–7761.
34. Y. Chen, Y. Li, J. Wang, M. M. Wu and C. X. Wang, J. Phys. Chem. C, 2014, 118, 12494–12499.
35. D. L. Dexter, J. Chem. Phys., 1953, 21, 836–850.
36. D. Geng, M. Shang, Y. Zhang, H. Lian and J. Lin, Dalton Trans., 2013, 42, 15372–15380.
37. C. H. Huang, T. M. Chen, W. R. Liu, Y. C. Chiu, Y. T. Yeh and S. M. Jang, ACS Appl. Mater. Interfaces, 2010, 2, 259–264.
38. P. I. Paulose, G. Jose, V. Thomas, N. V. Unnikrishnan and M. K. R. Warrier, J. Phys. Chem. Solids, 2003, 64, 841–846.
39. W. Z. Lv, M. M. Jiao, Q. Zhao, B. Q. Shao, W. Lu and H. P. You, Inorg. Chem., 2014, 53, 11007–11014.
40. G. Blasse, J. Solid State Chem., 1986, 62, 207–211.
41. G. Blasse, Phys. Lett. A, 1968, 28, 444–445.
42. L. H. Liu, L. Wang, C. N. Zhang, Y. J. Cho, B. Dierre, N. Hirosaki, T. Sekiguchi and R. J. Xie, Inorg. Chem., 2015, 54, 5556–5565.
43. C. H. Huang, T. W. Kuo and T. M. Chen, ACS Appl. Mater. Interfaces, 2010, 2, 1395–1399.

---

Footnote

† Electronic supplementary information (ESI) available: Crystallographic information files (CIFs) of the Lu₅(SiO₄)₃N compound. See DOI: 10.1039/c6cp01512c

---